Wearable-derived Sleep Measurements are Associated with Long-COVID in the RECOVER Adult Cohort

Wearable-derived Sleep Measurements are Associated with Long-COVID in the RECOVER Adult Cohort | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Article Wearable-derived Sleep Measurements are Associated with Long-COVID in the RECOVER Adult Cohort Sairam Parthasarathy, Shari Brosnahan, Solveig Sieberts, Elias Neto, and 49 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-7422764/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted You are reading this latest preprint version Abstract Wearables yield a wide array of sleep-related measures that are relevant to Long COVID. We leveraged wearables-derived sleep measures (WDSM) to identify differences between individuals with Long COVID (LC) versus individuals with possible or no LC in the RECOVER adult cohort. We found significant associations between LC and reduced heart rate variability measured during sleep and increased nightly variability in sleep duration after adjusting for confounders. Moreover, LC was independently associated with lower sleep efficiency, greater variability of nighttime sleep timing, higher resting heart rate, lower respiratory rate during rapid eye movement (REM) sleep, prolonged REM sleep onset latency, worse global physical and mental health. Cluster analysis identified distinct multidimensional patterns of WDSM that are associated with LC and quality of life. Together, the strong association between WDSM, or WDSM clusters, with LC provides a potential biomarker for future validation efforts to detect LC and monitor treatment effectiveness. Health sciences/Diseases Health sciences/Diseases/Infectious diseases PASC Long COVID Wearables Sleep Digital Health Heart Rate Variability Figures Figure 1 Figure 2 Introduction Long COVID (LC) is estimated to affect 6.9% of the U.S. population and costs an estimated $ 3.7 trillion over 5-years and continues to compound the health and economic burden of the nation 1 – 4 . However, there is marked under-detection of LC and one of the prominent reasons includes the lack of a clear diagnostic test 5 , 6 . In fact, little correlation between standard clinical laboratory tests and patient reported symptoms of LC has been observed 7 . Conceivably, diagnostic tests fail to demonstrate consistent associations with LC due to the heterogeneity of this condition that involves multiple organ systems and multiple pathobiological processes 8 , 9 . Sleep-related symptoms are the third most common symptom of LC and in some reports nearly 40% of patients with LC report sleep disturbances 10 , 11 . Moreover, sleep, like LC, is interconnected with multiple organ systems and biological processes such as cognition, immune regulation, and anti-inflammatory processes which are dysregulated in individuals with LC 12 – 14 . This is further supported by evidence suggesting that preexisting sleep issues may play a role in the development of LC 15 – 17 . Sleep can be readily measured by wearables 18 – 21 . Considering the cross-cutting and interconnected nature of sleep and LC, it is conceivable that wearables-derived sleep measures (WDSM; Table 1 ) may provide a digital indicator of LC. WDSM could further our understanding of the relationship between LC and sleep by enabling longitudinal collection in sufficiently powered cohorts. Accordingly, we investigated whether WDSM could identify differences between individuals with a high burden of LC symptoms versus those that had low or no burden of LC symptoms. As a secondary objective we investigated whether there are differences in WDSM for participants with and without individual symptoms of LC. Finally, we conducted cluster analysis of WDSM to identify multidimensional patterns of sleep changes that are associated with LC and health-related quality of life. Table 1 Wearables Derived Sleep Measures (WDSMs) of Multi-Dimensional Model of Sleep Dimension Metric Description Quantity Sleep Duration (mins) Average duration-quantity- of sleep during the main sleep period. Sleep duration is measured as the sum of all periods during the main sleep period spent asleep (vs wake). Sleep Depth and Architecture Minutes in Light Sleep Estimated time in stages I and 2 sleep summarized across the entire sleep period. Minutes in Deep Sleep Estimated time in stage 3 (deep) sleep summarized across the entire sleep period. Minutes in REM Sleep Estimated time in Rapid Eye Movement (REM) sleep summarized across the entire sleep period. Sleep Quality Sleep Efficiency (%) Percentage time during the overnight sleep (bedtime) period when estimated to be asleep (vs sleep plus wake) WASO (Wake After Sleep Onset, minutes) Total duration in minutes spent awake during the overnight (bedtime) period REM fragmentation Number of times periods of sleep are interrupted by wakefulness Sleep Regularity Standard Deviation (SD) of Sleep Duration (mins) The night-to-night regularity in sleep patterns is described by measuring the extent to which sleep duration and sleep timing varies across time periods of 5 of more consecutive days SD of Mid-Sleep (mins) Mid-Sleep is measured as the clock time mid- way between sleep onset and wake time. Sleep-Disordered Breathing REM Sleep Breathing Rate (per minute) Estimated rate of breathing during REM sleep. SpO2 (%) Average oxygen saturation during the sleep period. Sleep-Related Cardiac Activity Heart Rate Variability (HRV; ms) Beat to beat variation in heart rate during the sleep period. Resting Heart Rate (bpm; beats per minute) Average heart rate during the sleep period. To do so, we leveraged the WDSM in adult participants in the RECOVER Adult Cohort Study who were also enrolled in the RECOVER Digital Health Platform: A total of 1,230 individuals with WDSM over 6 months (minimum of 5 days of valid measurements) for a total of 151,045 nights of measurements (Fig. 1 ). The WDSM were complemented by two symptom-surveys over a 6-month period in the RECOVER Adult Observational cohort that yielded information regarding LC status based on the 2024 RECOVER Long COVID Research Index (LCRI) that classifies highly symptomatic LC (“likely LC”) as Long COVID Research Score ≥ 11 as a weighted index of LC symptoms. 22 WDSM includes sleep time, as well as metrics of sleep quality and sleep stages that are estimated by monitoring heart rate variability (HRV), the variation in time intervals between heartbeats, movement patterns and oxygen saturation. 20 , 23 We analyzed 14 Wearables-derived variables in relation to the LC status (any LCRI score ≥ 11 vs LCRI score < 11 at all timepoints) measured over 6-months to test the robustness of the observed associations. Although some smaller studies have assessed some of the wearables-derived data in patients with Long COVID 24 , 25 they have largely focused on activity-related measures. A comprehensive assessment of WDSMs in relation to well-defined LC status has not yet been performed 24 , 26 . Association between WDSM and LC, if present, has wide generalizability and a patient-powered approach for potentially detecting and monitoring the time-based or treatment-driven trajectories of LC. Results Cohort demographics Among the 1,262 sleep digital health participants, 433 were classified as Likely LC (LCRS ≥ 11) for at least one of the two timepoints, 529 had lower symptom burden at both timepoints (possible LC; 11 > LCRS > 0), and 299 had no LC symptoms (LCRS = 0) at either timepoint (Table 2 ). Female participants were more likely to be in the LC group (81% female in likely LC and 74% in both the possible LC and no LC symptoms group). White, non-Hispanic participants were more likely to be in the LC group (70%, 68% and 63% for the likely, possible and no LC groups, respectively). In contrast, Asian, non-Hispanic participants were less likely to be in the LC likely group (3%, 5.9% and 11% for the likely, possible and no LC symptoms groups, respectively). The median age at enrollment was similar across groups: 46 years (IQR: 37–56) for the likely LC group, 44 years (IQR: 34–59) for the possible LC group, and 43 years (IQR: 28–58) for the no LC group. Table 2 Study demographics Characteristic Highly Symptomatic LCRS ≥ 11 N = 433 1 Possible LC 11 > LCRS > 0 N = 529 1 None LCRS = 0 N = 299 1 Unknown N = 1 1 Sex assigned at birth Male 81 (19%) 136 (26%) 79 (26%) 0 (0%) Female 351 (81%) 388 (74%) 220 (74%) 1 (100%) Intersex 0 (0%) 2 (0.4%) 0 (0%) 0 (0%) Unknown 1 3 0 0 Age at enrollment 46 (37, 57) 44 (34, 58) 45 (33, 61) 63 (63, 63) Race and ethnicity Asian, Non-Hispanic 13 (3.0%) 31 (5.9%) 33 (11%) 0 (0%) Black, Non-Hispanic 34 (7.9%) 46 (8.7%) 19 (6.4%) 1 (100%) Hispanic 56 (13%) 65 (12%) 42 (14%) 0 (0%) White, Non-Hispanic 304 (70%) 358 (68%) 189 (63%) 0 (0%) Mixed race/Other/Missing 26 (6.0%) 29 (5.5%) 16 (5.4%) 0 (0%) 1 n (%); Median (Q1, Q3) Association between Long COVID and Sleep We examined the association between LCRI group and the digital sleep measures using two primary models: age, sex and race/ethnicity (model 2) and age, sex, race/ethnicity and BMI (model 3). We additionally examined an unadjusted model (model 1) and a full model which adjusted for age, sex, race/ethnicity, BMI, alcohol consumption, smoking, seasonality and two data quality metrics (see Methods). The results for these models were qualitatively similar to those for models 2 and 3, respectively. Under model 2 (age, sex and race/ethnicity), likely LC was associated with decreases in sleep efficiency, sleep duration, REM sleep breathing rate, SpO2, HRV, and increases in resting HR, SD of sleep duration, SD of mid-Sleep, and REM onset latency (corrected p-value ≤ 0.05) Table 3 ; Table S2; Figure S1 ; Fig. 2 (A) ). The associations of LCRI with sleep duration and SpO2 become non-significant (adjusted p-value = 0.160 and 0.077, respectively) when further including BMI in the model (model 3), implying that these effects may be at least partially explained by BMI ( Table S3; Figure S1 ; Fig. 2 (B), Figure S1 ). The association of SD of sleep duration and SD of mid-sleep suggest that participants with likely LC show more between-night variability in sleep (or irregular sleep) patterns. A sensitivity analysis with and without weekend days showed similar results ( Figure S5 ) suggesting that these effects are not driven by social jetlag differences (i.e., differences in sleep timing between weekday and weekends) but instead represent a broader pattern of sleep irregularity. A sensitivity analysis examined the effect of increasing the minimum available data to 14 days (versus 5 in the primary analyses). The results were qualitatively similar to the primary analysis ( Figures S3 & S4 ). Table 3 Distribution of wearables derived sleep measures (WDSMs) by Long COVID Research Index (LCRI) group Characteristic Long COVID N = 477 1 Negative Long COVID N = 853 1 Sleep Efficiency (%) 93.35 (91.01, 94.95) 93.78 (92.08, 95.17) Sleep Duration (mins) 349.19 (300.27, 391.13) 358.90 (312.96, 399.69) Minutes in Deep Sleep 60.72 (50.26, 73.45) 63.47 (52.53, 73.37) Minutes in REM Sleep 82.68 (67.78, 96.71) 83.00 (68.94, 95.36) REM Sleep Breathing Rate (per minute) 14.19 (12.97, 15.63) 14.51 (13.26, 15.79) SpO2 (%) 94.75 (93.80, 95.58) 95.07 (94.18, 95.78) Heart Rate Variability (ms) 22.84 (16.98, 33.02) 27.85 (20.75, 40.78) Resting Heart Rate (bpm) 68.76 (63.30, 74.20) 66.30 (60.54, 72.09) SD of Sleep Duration (mins) 112.30 (91.62, 139.09) 98.50 (75.75, 123.03) SD of Mid-Sleep (mins) 24.12 (17.58, 34.07) 19.79 (13.91, 29.19) REM Onset Latency (mins) 119.50 (99.04, 141.67) 110.80 (94.33, 131.61) REM Fragmentation Index 3.33 (2.92, 3.99) 3.38 (3.00, 3.96) Minutes in Light Sleep 248.81 (218.90, 275.71) 247.48 (221.20, 272.90) WASO (mins) 47.22 (38.54, 55.10) 48.86 (40.59, 57.11) 1 Median (Q1, Q3) Associations among LC Symptoms and Sleep We also examined the LC symptoms associated with the WDSM. The measures included were: brain fog, chest pain, chronic cough, dizziness, fatigue, gastrointestinal (GI) symptoms, head pain, palpitations, post exertional malaise (PEM), shortness of breath (SOB), sleep apnea, sleep disturbance, loss of smell and/or taste, and thirst as described earlier 22 , 27 . In models 2 and 3, all LC symptoms were associated with at least one sleep metric: brain fog (9), sleep disturbance (9), dizziness (8), fatigue (8), PEM (7), SOB (6), GI symptoms (7), thirst (6), sleep apnea (5), head pain (3), loss of smell or taste (3), chest pain (3), chronic cough (3), and palpitations (3) (Fig. 2 A, 2 B, S9, S10, S12-S25). Overall, decreased sleep efficiency and HRV, as well as increased SD of sleep duration and longer REM onset latency were associated with almost all symptoms examined (Fig. 2 A, 2 B; S9, S10, S12-S25). Complete tables by symptom and each model can be found in tables S9-S64. Increased brain fog was associated with a decrease in sleep efficiency, sleep duration, minutes in deep sleep, minutes in REM sleep, REM sleep breathing rate, HRV and an increase in SD of sleep duration, SD of mid-sleep, and REM onset latency. Increased resting HR was significant in model 2 but lost significance with the addition of BMI perhaps as higher BMI is associated with higher resting HR ( Fig. 2 A, 2 B, S15; Tables S22, S23). Dizziness was associated with a decrease in sleep efficiency, minutes in deep sleep, minutes in REM sleep, REM sleep breathing rate, HRV and increase in SD of sleep duration, REM onset latency and minutes in light sleep. For dizziness the associations of several WDSM increased with subsequent regression models ( Figure S20; Table S41-S44 ). Self-reported sleep disturbance was associated with a decrease in sleep efficiency, sleep duration, minutes in REM, REM sleep breathing rate, and HRV, and an increase in resting HR, SD of sleep duration, SD of mid-sleep, and REM onset latency ( Tables S61-S64 ). Fatigue was associated with a decrease in sleep efficacy, sleep duration, minutes in deep sleep, REM sleep breathing rate, HRV and in increased resting HR, SD of sleep duration, and REM onset latency. Similar to brain fog, in both sleep disturbance and fatigue, SpO2 does not retain its significance between model 2 and 3 related to BMI correction, while the REM sleep breathing rate association strengthens in significance with the inclusion of additional covariates ( Table S37-S40) . Sleep and health-related quality of life We also examined the association between the WDSMs and two measures of quality of life (QoL) from the Patient-Reported Outcome Measurement Information System (PROMIS) Global Health scale, the Global Physical Health (GPH) index T-scores (4 items on overall physical health, physical function, pain, and fatigue) and Global Mental Health (GMH) T-scores (4 items on quality of life, mental health, satisfaction with social activities, and emotional problems). 28 For each index, better QoL was associated with better sleep efficiency, increased sleep duration, more time in deep sleep, higher HRV, a lower resting HR, more sleep regularity (lower SD of sleep duration and SD of mid-sleep), and shorter REM onset latency. Wake after sleep onset (WASO) and SpO2 were significant in model 2 but did not retain significance when controlled for BMI in both scores as well. Increased time in REM sleep was associated with GPH but not GMH (Fig. 2 A, 2 B; Figures S26 & S27; Tables S65-S72 ). Cluster analysis of digital sleep measures To understand participant patterns with respect to the WDSM, we performed hierarchical clustering based on a subset of digital sleep measures (sleep efficiency, sleep duration, minutes in deep sleep, minutes in REM sleep, REM onset latency, REM sleep breathing rate, SD of sleep duration, HRV and resting HR) ( Fig. 2 (C); Table S73) . These were chosen as the most consistently correlated with likely LC and presence of LC symptoms. Visual inspection of the scree plot ( Figure S28 ) identified four primary clusters (Fig. 2 (C) ). Cluster 1 consists of participants manifesting better cardiovascular fitness with a lower baseline heart rate (mean(SD) = 67.94(7.28)), but lower HRV (mean(SD) = 25.54(10.19)), with an increased REM Breathing Rate (mean(SD) = 14.18(1.70)). Sleep efficiency, duration and REM Onset Latency were high (mean(SD) = 94.16(1.9), 381.79(48.74) and 124.25(34.91), respectively), and SD of sleep duration was low (mean(SD) = 93.11 (29.09)). Cluster 2 represents a group with higher efficiency, longer sleep duration and lower variability (mean(SD) = 94.00(2.30), 366.12(55.66) and 97.62(31.8), respectively). Participants in this cluster displayed an increase both in minutes in REM sleep (mean(SD) = 85.13 (20.02)) and a decrease in REM onset latency (mean(SD) = 106.82 (24.93)). They showed the highest HRV (mean(SD) = 52.04 (22.60)) and lowest resting HR (mean(SD) = 61.74(7.15)). The cluster 3 group was characterized by a short sleep time (mean(SD) = 287.10(58.26)), reduced minutes in both deep and REM sleep (mean(SD) = 54.19(15.96) and 71.72(17.53), respectively), and greater SD of sleep duration (mean(SD) = 143.73(43.11)), as well as a higher resting HR (71.36(8.25)). Cluster 4 shares the feature of sleep duration < 6h with cluster 3 but is further characterized by low sleep efficiency (mean(SD) = 63.36(9.68)) but moderate sleep duration (mean(SD) = 337.73(75.04)). Cluster 4 showed the highest proportion of LC positive participants (48%; Table 4 ). It also showed lower GPH and GMH scores (median(IQR) = 45 (39–51) and 45 (39–52), for Physical and Mental, respectively) ( Fig. 2 (D)) . It also showed a higher proportion of females (84% vs 79%, 72% and 73% for clusters 1, 2, and 3, respectively). In contrast, cluster 2, showed the lowest proportion of likely LC positive participants (20%) with high GPH (median(IQR) = 52 (46–58)) and GMH (median(IQR) = 50 (44–55)). This cluster had younger participants than the other three (median age = 36 vs 48, 49, and 47 for clusters 1, 3 and 4, respectively) ( Fig. 2 (D)) . Clusters 1 and 3 showed a moderate proportion of likely LC participants (35% and 41%, respectively). Cluster 1 showed higher GPH scores (median(IQR) = 49 (42–54) versus 46 (39–52) for cluster 3). It also had a higher proportion of female participants (79% versus 73%). Table 4 Demographic and symptom characteristics by cluster Characteristic Cluster 1 N = 459 1 Cluster 2 N = 266 1 Cluster 3 N = 253 1 Cluster 4 N = 150 1 LCRI Category Positive 160 (35%) 54 (20%) 104 (41%) 72 (48%) Negative 299 (65%) 212 (80%) 149 (59%) 78 (52%) Sex assigned at birth Male 96 (21%) 74 (28%) 69 (27%) 24 (16%) Female 361 (79%) 190 (72%) 184 (73%) 124 (84%) Intersex 1 (0.2%) 1 (0.4%) 0 (0%) 0 (0%) Unknown 1 1 0 2 Age at enrollment 48 (36, 60) 36 (30, 48) 49 (38, 60) 47 (38, 57) Race and ethnicity Non-Hispanic Asian 18 (3.9%) 29 (11%) 22 (8.7%) 3 (2.0%) Non-Hispanic Black 19 (4.1%) 22 (8.3%) 37 (15%) 9 (6.0%) Hispanic 58 (13%) 35 (13%) 30 (12%) 21 (14%) Non-Hispanic White 345 (75%) 160 (60%) 145 (57%) 110 (73%) Mixed race/Other/Missing 19 (4.1%) 20 (7.5%) 19 (7.5%) 7 (4.7%) Sleep Apnea 115 (25%) 35 (13%) 94 (38%) 54 (37%) Unknown 4 4 3 4 Sleep Disturbance 63 (14%) 24 (9.2%) 54 (22%) 41 (28%) Unknown 4 4 3 4 PROMIS Global Physical (t-score) 49 (42, 54) 52 (46, 58) 46 (39, 52) 45 (39, 51) PROMIS Global Mental (t-score) 47 (40, 53) 50 (44, 55) 45 (39, 51) 45 (39, 52) 1 n (%); Median (Q1, Q3) Profile Analysis In order to further understand the patterns of WDSMs associated with participant self-reports of sleep apnea and sleep disturbance, as well as self-reported quality of life measures (i.e. PROMIS GPH and GMH Indices), we employed Criterion Profile Analysis (CPA) 29 . Sleep apnea was associated with a decrease in REM sleep breathing rate and increase of variability of sleep duration, while sleep disturbance was associated with both those factors, as well as a decrease in sleep efficiency, minutes in REM sleep and an increase in resting HR ( Figure S29 ). Lower scores for the GPH and GMH measures were associated with decreases in sleep efficiency and minutes in deep sleep, as well as lower HRV ( Figure S30 ). Lower scores are also associated with increases in REM onset latency and SD of sleep duration. Additionally, an increase in resting HR was associated with lower GPH but did not reach significance for a change of GMH. Likewise, a lower REM sleep breathing rate was associated with lower GMH but did not reach significance for a change of GPH. By comparing the absolute values of CPA scores, we found that variability of sleep duration generally played a more important role than other digital sleep measures in predicting sleep problems and physical/mental health. Significant pattern effects were found for all outcome variables (sleep apnea: F-statistic(17, 1085) = 7.69, p < 0.001; sleep disturbance: F-statistic(17, 1085) = 3.48, p < 0.001; GPH: F-statistic(17, 1100) = 15.30, p < 0.001; GMH: F-statistic(17, 1100) = 9.14, p < 0.001). Stratified Analyses We performed stratified analyses by age ( 65) ( Figure S31 ), sex ( Figure S32 ) and race and ethnicity ( Figure S33 ). With respect to age, the associations and directions of effect were similar for most of the variables which were significant in the main analysis (i.e. sleep efficiency, HRV, resting HR, SD of sleep duration, SD of mid-sleep, REM onset latency) ( Figure S31 ). Even though some associations showed discordant directions none reached statistical significance, with none of the associations reaching statistical significance in the older group (over 65). Given the reduced sample sizes in some of these age groups (n = 611, 474, and 145 for the 65 age groups, respectively), some reduction in power was to be expected. Interestingly, we did observe a significant positive association between likely LC and minutes spent in light sleep in the youngest subgroup, which was not observed in either subgroup or the combined analyses but may be due to reduced power. The patterns of change in WDSM with likely LC were generally similar across both sexes ( Figure S32 ). However, the magnitudes and statistical significance did vary across the sex groups. Both male and females with LC showed significant decrease in HRV, increased in resting HR, SD sleep duration and SD of mid-sleep, while only females showed a significant decrease in sleep efficiency, SpO2 and increase in REM onset latency which was not seen as significant in the males. Likewise, males had a significant decrease in sleep duration and WASO that was not observed in women. Discussion In this work involving the largest cohort of wearables measurement in individuals with LC, we found biologically plausible differences in WDSM between individuals with likely LC and those with low or no LC symptom burden over a six-month measurement period. Specifically, the strongest association was observed for reduced HRV measured during sleep and increased variability in sleep duration. HRV measures the beat-by-beat variability that reflects the dynamic interaction between sympathetic and parasympathetic tone within the autonomic nervous system and reflects overall health with higher HRV associated with cardiovascular health and better response to stress. There are numerous prior reports of reduced HRV – measured by conventional electrocardiogram – in individuals with LC that could exemplify the underlying autonomic instability that may be triggered or exacerbated by a COVID infection or underlie and/or drive LC presentation and/or symptoms as dysautonomia, Postural Orthostatic Tachycardia Syndrome (POTS), and PEM 30 , 31 . While many studies have previously used electrocardiograms, Holter monitors, and cardiac belts, the photoplethysmography-based wearables used in our study to yield reliable HRV data poses a promising alternative and practical approach 32 . Such opportunity has been capitalized by others to demonstrate a reduction in HRV during acute SARS-CoV-2 infection but, to our knowledge, there are no prior wearables-based studies demonstrating the association between reduced HRV and LC 33 . The observed association between wearables-derived reduced HRV during sleep and LC further supports autonomic imbalance as one of the mechanistic underpinnings of LC. Additionally, it lends itself for longitudinal monitoring of patients undergoing treatments that targets the autonomic imbalance such as slow-paced breathing. 34 While many studies have previously used electrocardiograms, Holter monitors, and cardiac belts, the photoplethysmography-based wearables used in our study to yield reliable HRV data during sleep poses a promising alternative and practical approach and considering the relative immobility may provide a more accurate measure and minimize artifacts induced by activity 21 , 32 . We also find a strong association of LC with SD of sleep duration and to a lesser degree SD of mid-sleep, measures of sleep irregularity. Irregular sleep duration and timing (markers for circadian misalignment and its myriad effects of healthy physiology) have recently been identified to be associated with insulin resistance and dyslipidemia and to predict incident cardiovascular disease, cognitive decline, and all-cause mortality 35 , 36 . In the context of a multi-dimensional model for sleep health, increased irregularity in sleep-wake patterns has been shown to cluster with other sleep dimensions, including insufficient sleep duration, poor sleep efficiency, and sleep-disordered breathing 37 . Improving sleep regularity may be an important strategy for more broadly improving sleep health, and in fact is a core principle for sleep hygiene interventions. Interestingly, recent reports indicate increased prevalence of cardiovascular events among individuals hospitalized for COVID-19 or even in mild cases of SARS-CoV2 infection when associated with certain ABO blood types 38 . Coehlo et al have previously described marked instability in polysomnographically measured sleep patterns with both long and short bouts of objectively measured sleep in individuals with LC 39 . An ongoing RECOVER clinical trial in fact is targeting sleep regularity as a strategy for improving LC-related sleep disturbance. The observed variability in sleep duration and mid-sleep (timing of nighttime sleep) in our study may serve as a biomarker for further stratification of individuals with greater likelihood of adverse long-term outcomes of LC and presents additional opportunity to augment current risk scores for predicting cardiovascular events in these individuals 38 . We observed that LC was also associated with lower sleep efficiency and a tendency for reduced sleep duration. This contrasts with a prior report of similar wearables-measured sleep quantity between individuals with LC and those without LC following a known SARS-CoV-2 infection 24 . In this prior study by Radin et al, participants with LC (n = 279) and without LC (n = 274) were followed up to one year but sleep measures besides sleep quality were unfortunately not reported 24 . Conceivably, better case identification using the LCRI and larger sample size of our study (n = 1,230 including 417 likely LC cases) may have resulted in greater power to observe statistical differences, including lower sleep efficiency and tendency for reduced sleep duration in participants with persistent LC compared to those without LC. Similar to the findings in our study, Radin et al found more variable sleep duration and higher resting HR in LC which may signify sleep fragmentation, reduced parasympathetic tone, possibly in response to nocturnal hypoxia and sleep apnea. In our study, reflectance oximetry levels (SpO2) measured by the wearables were lower in individuals with likely LC when compared to those with low or no LC symptoms. Nocturnal hypoxia has been previously described in individuals with LC and may reflect either the greater prevalence of obstructive sleep apnea (OSA) in these individuals with LC, sleep-related hypoventilation, or underlying pulmonary complications of LC 15 , 40 – 42 . Interestingly, nocturnal hypoxemia can be relatively asymptomatic in acute COVID due to posited alterations in chemosensitivity and control of breathing, but importantly may contribute to delayed detection that could be mitigated by wearables 40 . Such a phenomenon of altered chemosensitivity may persist in individuals with LC but this is unclear 40 . Obesity may further contribute to sleep-related hypoventilation and consequently lower oxygen saturation, which, in turn may contribute to fatigue and cognitive problems 43 . In our study, after adjustment for body mass index (BMI), the association between lower SpO2 and LC was lost suggesting that BMI (and comorbid OSA and sleep-related hypoxia or restrictive pulmonary phenotype) may explain the observed association between oxygen levels and LC 44 – 46 . Additionally, our observations of a lower respiratory rate during REM sleep further strengthens the possibility of comorbid OSA and sleep-related hypoxia playing a role in the LC pathophysiology 47 . Prolonged REM sleep onset latency in individuals with LC has been previously described 48 . The reasons for delayed REM sleep latency are uncertain, but may be due to comorbid OSA and/or medications (e.g., Low Dose Naltrexone, sedative-hypnotics such as zolpidem, and SSRI/SNRI antidepressants prescribed for their “purported” protective effects against LC) that may suppress REM sleep 49 . Prolonged REM sleep latency has been associated with cognitive dysfunction and biomarkers of Alzheimer’s disease 50 . In our study, WDSMs were strongly associated with worse GPH and GMH QoL ratings. Specifically, worse Mental and worse Physical scores were associated with multiple metrics of poor sleep including reduced sleep efficiency, reduced REM related breathing rate, reduced HRV during sleep, increased variability of mid-point sleep and sleep duration, prolonged REM latency, and higher resting HR during sleep. Importantly, such analysis of health-related quality of life measures, individual LC symptoms, and WDSM revealed marked internal consistency and external validity of the directionality of associations. The identification of associations of LC with multiple metrics of sleep is consistent with modern frameworks of sleep health that posit that sleep is a multidimensional construct composed of various interacting features of sleep that together better describe sleep health than individual metrics. The strongest associations between individual symptoms of LC and WDSM were observed for the most common and burdensome LC symptoms -- PEM, brain fog, fatigue, and dizziness. Cluster analysis revealed four distinct patterns indicative of various manifestations of sleep disorder symptoms. A cluster ( cluster 3; Fig. 2 b) with clear reduction in sleep duration, increased resting HR, increased variability of sleep duration and reduced time in REM and deep sleep may signify a phenotype characterized by increased arousal and shorter, lighter, and more variable sleep, which may reflect LC-associated “insomnia phenotype”. Whereas, another cluster (cluster 2) is delineated by higher HRV, higher REM breathing rate (i.e., less apnea in REM), and lower resting HR and appears to denote a more “normal” phenotype and likely consisted of a more resilient younger age group of participants (Fig. 2 b). While the findings represent the cross-cutting nature of sleep and sleep-related physiological changes (heart rate, oxygen saturation), these findings may also denote the bidirectional association between LC and sleep. Taken together, the strong association between WDSMs with LC status and symptoms support their utility as a possible digital biomarker that may be further validated for identifying subgroups of patients with the LC condition who may experience different health trajectories and may benefit from targeted interventions, as well as for monitoring disease course and treatment effectiveness. The findings of lower oxygen saturation levels, lower HRV, and variability in sleep duration and timing with LCRI status provide potential targets for future interventions using positive airway pressure therapy or supplemental oxygen, slow paced breathing/vagal stimulation, and behavioral and circadian (e.g., light) approaches and the very same wearables can assist with monitoring the corresponding treatment response. Finally, identifying clusters of sleep-related metrics that reflect diverse domains of sleep health supports both the importance of applying a multi-dimensional sleep health framework in patients with LC, and suggests that this framework include measures of rest/sleep HRV, sleep variability, sleep efficiency, sleep duration, and metrics related to breathing and oxygenation during sleep. Methods Institutional review boards at NYU Grossman School of Medicine, serving as a single Institutional Review Board, and other participating institutions reviewed and approved the protocol. All participants provided written informed consent to participate in research. Study Design In RECOVER adult study design, participants, aged 18 years or older, were recruited at 83 sites in 33 states plus Washington, DC, and Puerto Rico, regardless of prior infection with SARS-CoV-2 51 . All participants completed a baseline set of surveys, a focused physical examination, and standard laboratory test sample collection at enrollment. Participants were followed prospectively with survey completion every 3 months and laboratory sample collection at enrollment and 6, 12, 24, 36, and 48 months after infection, or in the case of uninfected controls, the date of a negative test result (index date; Fig. 2 ). Subjects could enter the study at different times in relation to their initial COVID infection and were defined into cohorts including the acute infected (enrolled during first infection), post-acute (enrolled after initial COVID infection) and uninfected cohort (who enrolled prior to first infection but could be included in this study group if they were infected after study enrollment). 51 Participants The Digital Health Program (DHP) is a component of RECOVER that was launched in 2023 to promote participant engagement and retention, and to collect new information that compliments and augments other RECOVER data including the collection of consumer-grade wearable “wrist-worn” device data (activity, heart rate, and sleep). All RECOVER Adult Cohort participants were invited to sign-up for DHP within an app that they downloaded to their device. They were asked to connect and share data from their own (bring your own (BYO) wearable devices (Fitbit, Apple Watch, Garmin, etc.) or order a free study device (Fitbit smartwatch, Sense 2 or Fitbit tracker, Charge 5). As of November 11, 2024, a total of 6,497 RECOVER Adult participants enrolled in DHP, of which 4,104 (63%) connected a wearable device. Comparable wearables-derived sleep measures were only transmitted in Fitbit monitors which accounted for 2,696 (66%) subjects. Of those previously infected participants with Fitbit sleep data, 1,262 (47%) had at least 5 nights of wearables-derived sleep measures concurrent with symptom surveys during the 6-month data collection period and were included in the cohort analysis. An additional 474 (18%) participants had sufficient data but were withheld for future validation of findings. Exposures and Outcomes Likely LC was defined by a Long COVID Research Index (LCRI) Score ≥ 11 at any point in time over a 6-month period post-infection. Possible LC was defined as LCRI score between 1–10, and the no LCRI symptom group had a LCRI score = 0 22 . The scores were based on the RECOVER data-derived LCRI from 2024 that used Lasso models to discriminate between SARS-CoV2 infected and non-infected participants in the RECOVER Adult Cohort (Geng 2024). Possible LC and no LC was collapsed into an LC indeterminate group 22 . Further details for defining LC status are provided in the statistical analysis section (see below). Sensitivity analysis was conducted using the LCRI Score as a continuous variable and referred to the LCRS. LC symptoms were included based on items in the scoring system for either the RECOVER LCRI (2024) or the 2024 NASEM LC Common Symptoms Definition 8 , 22 , 27 . LC outcome variables included: Long COVID status, Brain Fog, Chest Pain, Chronic Cough, Dizziness, Fatigue, Gastrointestinal (GI) Symptoms, Head Pain, Palpitations, Post Exertional Malaise (PEM), Shortness of Breath (SOB), Sleep Apnea, Sleep Disturbance, Loss of Smell and/or Taste, and Thirst. Measures of quality of life included PROMIS GMH (the sum value of the responses to items Global02, Global04, Global05, and Global10r) and PROMIS GPH (sum value of the responses to items Global03, Global06, Global07r, and Global08r) of the PROMIS Scale v1.2. 28 Wearables-Derived Sleep Measures (WDSM) We selected and summarized a set of Fitbit sleep measures and measures collected during sleep and summarized them over the six-month window. A summary of these sleep metrics, mapped to domains of sleep health, is provided in Table 1 . This includes the mean of the following measures (with units): sleep efficiency (percent), total sleep duration (minutes), minutes in sleep stages (deep sleep, REM sleep, light sleep), WASO, REM onset latency (minutes), REM fragmentation index, 52 resting heart rate (RHR, beats per minute during sleep and rest), heart rate variability (HRV) during sleep (ms), REM sleep breathing rate (breaths per minute), average oxygen saturation (SpO2; percent). We also examined across-night sleep variability metrics by including the within-individual standard deviation (SD) of sleep duration (minutes) and within-individual SD of sleep midpoint (SD of mid-sleep; minutes). For the standard deviation metrics, we performed a sensitivity analysis excluding weekend days, to examine the degree sleep variability was driven by weekday/weekend differences. For each measure, we required a minimum data availability of 5 days, with at least 3 hours of sleep wear time. Results did not qualitatively change with a minimum data availability of 14 days. Statistical Analysis To evaluate the association between WDSM and LC, we took a cross-sectional approach to compare a 6-month window of wearables data with two consecutive symptom surveys (Fig. 1 ). For each participant, we selected the first available symptom survey that followed the first availability of the wearables data and was collected at least 6 months post-infection. The subsequent survey timepoint was selected to be the next available survey that was > 1 month and < 6 months following the first one. If no such survey existed in that time window, the second timepoint was determined to be missing. The 6-month window of digital health data was selected such that it was centered around the two survey timepoints. If the second timepoint was missing, the window was selected to begin 45 days before the available symptom survey. Because each wearables measure had different patterns of data availability, this window selection was done independently by measure. Using the symptom surveys at the two timepoints, we combined the LCRI and LCRS for each participant at each timepoint. The LCRS was taken to be the average value at the two timepoints. A participant was defined as likely LC if their LCRI score was ≥ 11 at either timepoint. For participants with only the first timepoint, we classified them based only on the available data. To examine the implicit assumption “last observation carried forward”, specifically in the LC negative group (LCRI < 11), we performed a sensitivity analysis which excluded these participants. These analyses showed no qualitative differences to the primary analyses. Association between sleep measures and long COVID Logistic regression was used to assess the association between each sleep measure and LCRI or individual symptoms, with the sleep measures being the independent variable. Analogously, associations between sleep measures and continuous measures (LCRS, PROMIS GPH and PROMIS GMH) were assessed using linear regression with the sleep measure as the independent variable. For each comparison, four different models were considered. Our primary model (model 3) adjusted for known explanatory variables (age, sex, race and ethnicity, and BMI). Additionally, we explored additional variables (alcohol, smoking, seasonality, and time between surveys that may influence LC status: LCRI/LC Symptom ~ Digital Sleep Measure (1) LCRI/LC Symptom ~ Digital Sleep Measure + Age + Sex + Race/Ethnicity (2) LCRI/LC Symptom ~ Digital Sleep Measure + Age + Sex + Race/Ethnicity + BMI (3) LCRI/LC Symptom ~ Digital Sleep Measure + Age + Sex + Race/Ethnicity + BMI + Alcohol consumption + Smoking + Seasonality + time between surveys + n records, digital (Full) where n records, digital is the number of Fitbit sleep records in the time window. Because of the small number of participants who are intersex or of unknown sex, these individuals were grouped with the male sex group for the purpose of analysis. To account for multiple significance testing, p -values were adjusted using the Benjamini-Hochberg method within LCRI, LCRS and the set of symptoms separately for each model (q-value < 0.05). Clustering Participants were clustered into subgroups using hierarchical clustering based on a subset of digital sleep measures that were strongly correlated with the GPH and GMH variables (for external validity) but were relatively distinct (showing low correlations) from each other (for discriminant validity) 53 . Clusters were generated using the hclust() function in the R “stats” package using the Ward.D2 agglomeration method which minimizes the within-cluster variance. The resulting dendrogram was cut using a hybrid algorithm in the cutreeDynamic() function in the R “dynamicTreeCut” package. The optimal number of clusters was determined based on both the scree plot which shows how the within-cluster sum of squares varies with different numbers of clusters, and the Bayesian Information Criterion (BIC) plot generated using the mclustBIC() function in the R “mclust” package. We further investigated the demographics, presence of sleep-related symptoms, and distribution of quality of life measures within each cluster to identify distinct patterns in subgroups. Criterion Profile Analysis The criterion profile analysis (CPA) was used to determine whether there was a pattern of digital sleep measures associated with the presence of sleep-related symptoms or high scores on quality of life measures 29 . The relative importance of predictors (i.e., the digital sleep measures) was quantified using the criterion pattern score, which was defined as the standardized regression coefficient of a given predictor minus the average of all standardized regression coefficients in the model, multiplied by a positive constant, such that higher absolute criterion pattern scores indicate higher importance in predicting the outcome of interest. Under CPA, the predictor scores are decomposed into two components: a level component (defined as the mean of the predictor scores) and a pattern component (defined as the deviation of each predictor score from the mean predictor score), such that the total variance of the outcome is explained by both the level and the pattern of the predictors. The level and pattern effects can be tested via F-tests by comparing the full model and the model with only level/pattern components. The analysis was done using the cpa() function in the R “profileR” package 54 . Results were generated to show digital sleep measures associated with presence of sleep symptoms (e.g., apnea and sleep disturbance) and decreased global health scores. Declarations Data availability: The RECOVER Adult Cohort Observational Study remains in progress and therefore datasets are updated dynamically, individual participant-level questionnaire-based data including the data dictionary is available on the RECOVER website (https://recovercovid.org/). NHLBI has undertaken a significant effort to release harmonized data from all RECOVER observational cohort studies to the public via their BioData Catalyst platform (https://biodatacatalyst.nhlbi.nih.gov/). Datasets include participant-level data collected for the present study, as well as the technical infrastructure and analytic tools to evaluate the results and conclusions from this study. Conflicts of Interest/Disclosures: Dr. Wisnivesky received honorarium from Sanofi, Banook, PPD, and AMA and grants from Sanofi, Regeneron, Axella, and Boehringer Ingelheim. Dr. Mazziotti has been awarded devices from SleepImage, Inc. Dr. Raytselis has enrolled as a participant in the RECOVER Observational study and is a patient representative. Dr. Levitan has received research funding from Amgen and personal fees from the University of Pittsburgh for serving on a DSMB, unrelated to the current work. Dr. Dunn is a Scientific Advisor to Veri, Inc. and Biio, Inc. Dr. Verduzco-Gutierrez has an AHRQ grant for Long COVID. Dr. Peluso has received consulting fees from Gilead Sciences, AstraZeneca, BioVie, Apellis Pharmaceuticals, and BioNTech, travel support from Invivyd, and research support from Aerium Therapeutics and Shionogi, outside the submitted work. Dr. Upinder Singh received funding from Pfizer for studies on Long-Covid. Dr. McComsey served as research consultant for Merck, Gilead, and GlaxoSmithKline unrelated to the current work. Dr. Redline has received funding from NIH and consulting fees from Eli Lilly Inc. She is an unpaid board member of the Alliance of Sleep Apnea Partners and National Sleep Foundation, and is Editor-In-Chief of Sleep Health. Dr. Parthasarathy has received consulting fees from Abbvie, Inc., Apria Healthcare Inc., Zoll Medical, Inc., Jazz Pharmaceuticals, Inc., DynaMed, LLC, Fisher & Paykel, Inc., Apnimed, Inc., and has received grants to the institution from Phillips, Inc., Verily Lifesciences, Inc., Regeneron, Inc., US Biotest, Inc., WHOOP, Inc., and Sommetrics, Inc. Funding/Support: This research was funded by the NIH OT2HL161847 as part of the Researching COVID to Enhance Recovery (RECOVER) research program. Additional support for Dr. Parthasarathy from Arizona CEAL (OT2HL158287) and R25HL126140 during the writing of this manuscript. Role of the Funder/Sponsor: This content is solely the responsibility of the authors and does not necessarily represent the official views of the RECOVER Program, the NIH, or other funders. Acknowledgements: We would like to thank the National Community Engagement Group (NCEG), all patient, caregiver and community Representatives, and all the participants enrolled in the RECOVER Initiative. This study is part of the NIH Researching COVID to Enhance Recovery (RECOVER) Initiative, which seeks to understand, treat, and prevent the post-acute sequelae of SARS-CoV-2 infection (PASC). For more information on RECOVER visit https://recovercovid.org/ . References Ford, N. D. et al. Notes from the Field: Long COVID Prevalence Among Adults - United States, 2022. MMWR Morb Mortal Wkly Rep 73 , 135-136 (2024). https://doi.org/10.15585/mmwr.mm7306a4 Cutler, D. M. The Costs of Long COVID. JAMA Health Forum 3 , e221809 (2022). https://doi.org/10.1001/jamahealthforum.2022.1809 2025 Long COVID Fact Sheet , (2025). Cutler, D. M. The Economic Cost of Long COVID: An Update , 2025). Tsilingiris, D. et al. Laboratory Findings and Biomarkers in Long COVID: What Do We Know So Far? Insights into Epidemiology, Pathogenesis, Therapeutic Perspectives and Challenges. Int J Mol Sci 24 (2023). https://doi.org/10.3390/ijms241310458 National Center for Health Statistics. U.S. Census Bureau, Household Pulse Survey, 2022-2024. Long COVID. , (2025). Erlandson, K. M. et al. Differentiation of Prior SARS-CoV-2 Infection and Postacute Sequelae by Standard Clinical Laboratory Measurements in the RECOVER Cohort. Ann Intern Med 177 , 1209-1221 (2024). https://doi.org/10.7326/M24-0737 National Academies of Sciences, Engineering, and Medicine; Health and Medicine Division; Board on Global Health; Board on Health Sciences Policy; Committee on Examining the Working Definition for Long COVID . (National Academies Press (US), 2024). Perumal, R. et al. Long COVID: a review and proposed visualization of the complexity of long COVID. Front Immunol 14 , 1117464 (2023). https://doi.org/10.3389/fimmu.2023.1117464 Pena-Orbea, C. et al. Sleep Disturbance Severity and Correlates in Post-acute Sequelae of COVID-19 (PASC). J Gen Intern Med 38 , 2015-2017 (2023). https://doi.org/10.1007/s11606-023-08187-3 Tedjasukmana, R., Budikayanti, A., Islamiyah, W. R., Witjaksono, A. & Hakim, M. Sleep disturbance in post COVID-19 conditions: Prevalence and quality of life. Front Neurol 13 , 1095606 (2022). https://doi.org/10.3389/fneur.2022.1095606 Garbarino, S., Lanteri, P., Bragazzi, N. L., Magnavita, N. & Scoditti, E. Role of sleep deprivation in immune-related disease risk and outcomes. Commun Biol 4 , 1304 (2021). https://doi.org/10.1038/s42003-021-02825-4 Simon, K. C., Nadel, L. & Payne, J. D. The functions of sleep: A cognitive neuroscience perspective. Proc Natl Acad Sci U S A 119 , e2201795119 (2022). https://doi.org/10.1073/pnas.2201795119 Xie, L. et al. Sleep drives metabolite clearance from the adult brain. Science 342 , 373-377 (2013). https://doi.org/10.1126/science.1241224 H, L. M. et al. Risk of post-acute sequelae of SARS-CoV-2 infection associated with pre-coronavirus disease obstructive sleep apnea diagnoses: an electronic health record-based analysis from the RECOVER initiative. Sleep 46 (2023). https://doi.org/10.1093/sleep/zsad126 Schilling, C. et al. Pre-existing sleep problems as a predictor of post-acute sequelae of COVID-19. J Sleep Res 33 , e13949 (2024). https://doi.org/10.1111/jsr.13949 Quan, S. F. et al. Sleep and long COVID: preexisting sleep issues and the risk of post-acute sequelae of SARS-CoV-2 infection in a large general population using 3 different model definitions. J Clin Sleep Med 21 , 249-259 (2025). https://doi.org/10.5664/jcsm.11322 de Zambotti, M., Cellini, N., Goldstone, A., Colrain, I. M. & Baker, F. C. Wearable Sleep Technology in Clinical and Research Settings. Med Sci Sports Exerc 51 , 1538-1557 (2019). https://doi.org/10.1249/MSS.0000000000001947 Phillips, A. J. K. et al. Irregular sleep/wake patterns are associated with poorer academic performance and delayed circadian and sleep/wake timing. Sci Rep 7 , 3216 (2017). https://doi.org/10.1038/s41598-017-03171-4 Lee, T. et al. Accuracy of 11 Wearable, Nearable, and Airable Consumer Sleep Trackers: Prospective Multicenter Validation Study. JMIR Mhealth Uhealth 11 , e50983 (2023). https://doi.org/10.2196/50983 Berryhill, S. et al. Effect of wearables on sleep in healthy individuals: a randomized crossover trial and validation study. J Clin Sleep Med 16 , 775-783 (2020). https://doi.org/10.5664/jcsm.8356 Geng, L. N. et al. 2024 Update of the RECOVER-Adult Long COVID Research Index. JAMA (2024). https://doi.org/10.1001/jama.2024.24184 Haghayegh, S., Khoshnevis, S., Smolensky, M. H., Diller, K. R. & Castriotta, R. J. Accuracy of Wristband Fitbit Models in Assessing Sleep: Systematic Review and Meta-Analysis. J Med Internet Res 21 , e16273 (2019). https://doi.org/10.2196/16273 Radin, J. M. et al. Long-term changes in wearable sensor data in people with and without Long Covid. NPJ Digit Med 7 , 246 (2024). https://doi.org/10.1038/s41746-024-01238-x Kukreti, S., Lu, M. T., Yeh, C. Y. & Ko, N. Y. Physiological Sensors Equipped in Wearable Devices for Management of Long COVID Persisting Symptoms: Scoping Review. J Med Internet Res 27 , e69506 (2025). https://doi.org/10.2196/69506 Mekhael, M. et al. Studying the Effect of Long COVID-19 Infection on Sleep Quality Using Wearable Health Devices: Observational Study. J Med Internet Res 24 , e38000 (2022). https://doi.org/10.2196/38000 Thaweethai, T. et al. Development of a Definition of Postacute Sequelae of SARS-CoV-2 Infection. JAMA 329 , 1934-1946 (2023). https://doi.org/10.1001/jama.2023.8823 Hays, R. D., Bjorner, J. B., Revicki, D. A., Spritzer, K. L. & Cella, D. Development of physical and mental health summary scores from the patient-reported outcomes measurement information system (PROMIS) global items. Qual Life Res 18 , 873-880 (2009). https://doi.org/10.1007/s11136-009-9496-9 Davison, M. L. & Davenport, E. C., Jr. Identifying criterion-related patterns of predictor scores using multiple regression. Psychol Methods 7 , 468-484 (2002). https://doi.org/10.1037/1082-989x.7.4.468 Haischer, M. H. et al. Heart rate variability is reduced in COVID-19 survivors and associated with physical activity and fatigue. Physiol Rep 12 , e15912 (2024). https://doi.org/10.14814/phy2.15912 da Silva, A. L. G. et al. Impact of long COVID on the heart rate variability at rest and during deep breathing maneuver. Sci Rep 13 , 22695 (2023). https://doi.org/10.1038/s41598-023-50276-0 Georgiou, K. et al. Can Wearable Devices Accurately Measure Heart Rate Variability? A Systematic Review. Folia Med (Plovdiv) 60 , 7-20 (2018). https://doi.org/10.2478/folmed-2018-0012 Sanches, C. A., Silva, G. A., Librantz, A. F. H., Sampaio, L. M. M. & Belan, P. A. Wearable Devices to Diagnose and Monitor the Progression of COVID-19 Through Heart Rate Variability Measurement: Systematic Review and Meta-Analysis. J Med Internet Res 25 , e47112 (2023). https://doi.org/10.2196/47112 Mauro, M., Cegolon, L., Bestiaco, N., Zulian, E. & Larese Filon, F. Heart Rate Variability Modulation Through Slow-Paced Breathing in Health Care Workers with Long COVID: A Case-Control Study. Am J Med 138 , 870-883.e875 (2025). https://doi.org/10.1016/j.amjmed.2024.05.021 Leproult, R., Holmback, U. & Van Cauter, E. Circadian misalignment augments markers of insulin resistance and inflammation, independently of sleep loss. Diabetes 63 , 1860-1869 (2014). https://doi.org/10.2337/db13-1546 Huang, T., Mariani, S. & Redline, S. Sleep Irregularity and Risk of Cardiovascular Events: The Multi-Ethnic Study of Atherosclerosis. J Am Coll Cardiol 75 , 991-999 (2020). https://doi.org/10.1016/j.jacc.2019.12.054 Tracy, E. L. et al. Multidimensional Sleep and Self-Rated Physical Health and Depressive Symptoms Among Retired Older Adults: A Sex-Stratified Analysis. Behav Sleep Med , 1-12 (2025). https://doi.org/10.1080/15402002.2025.2493651 Hilser, J. R. et al. COVID-19 Is a Coronary Artery Disease Risk Equivalent and Exhibits a Genetic Interaction With ABO Blood Type. Arterioscler Thromb Vasc Biol 44 , 2321-2333 (2024). https://doi.org/10.1161/atvbaha.124.321001 Coelho, F. M. S. et al. Sleep disorder syndromes of post-acute sequelae of SARS-CoV-2 (PASC) / Long Covid. Sleep Med 123 , 37-41 (2024). https://doi.org/10.1016/j.sleep.2024.08.030 Tobin, M. J., Laghi, F. & Jubran, A. Why COVID-19 Silent Hypoxemia Is Baffling to Physicians. Am J Respir Crit Care Med 202 , 356-360 (2020). https://doi.org/10.1164/rccm.202006-2157CP Halawa, S. et al. Potential long-term effects of SARS-CoV-2 infection on the pulmonary vasculature: a global perspective. Nat Rev Cardiol 19 , 314-331 (2022). https://doi.org/10.1038/s41569-021-00640-2 Sun, H. et al. Facility-Measured Nocturnal Hypoxemia and Sleep Among Adults with Long COVID versus Age- and Sex-Matched Healthy Adults: A Preliminary Observational Study. SLEEP Advances (2025). https://doi.org/10.1093/sleepadvances/zpaf017 Subramanian, A. et al. Symptoms and risk factors for long COVID in non-hospitalized adults. Nat Med 28 , 1706-1714 (2022). https://doi.org/10.1038/s41591-022-01909-w Kendzerska, T., Leung, R. S., Gershon, A. S., Tomlinson, G. & Ayas, N. The Interaction of Obesity and Nocturnal Hypoxemia on Cardiovascular Consequences in Adults with Suspected Obstructive Sleep Apnea. A Historical Observational Study. Ann Am Thorac Soc 13 , 2234-2241 (2016). https://doi.org/10.1513/AnnalsATS.201604-263OC Gabbay, I. E., Gabbay, U. & Lavie, P. Obesity plays an independent worsening modifying effect on nocturnal hypoxia in obstructive sleep apnea. Sleep Med 13 , 524-528 (2012). https://doi.org/10.1016/j.sleep.2012.01.005 Guerra, S. et al. Morbidity and mortality associated with the restrictive spirometric pattern: a longitudinal study. Thorax 65 , 499-504 (2010). https://doi.org/10.1136/thx.2009.126052 Bonsignore, M. R. et al. REM sleep obstructive sleep apnoea. Eur Respir Rev 33 (2024). https://doi.org/10.1183/16000617.0166-2023 Rouen, A. et al. Polysomnographic parameters in long-COVID chronic insomnia patients. Dialogues Clin Neurosci 25 , 43-49 (2023). https://doi.org/10.1080/19585969.2023.2222714 Sramek, J. et al. The effect of naltrexone on sleep parameters in healthy male volunteers. J Clin Psychopharmacol 34 , 167-168 (2014). https://doi.org/10.1097/JCP.0b013e3182a607ff Jin, J. et al. Association of rapid eye movement sleep latency with multimodal biomarkers of Alzheimer's disease. Alzheimers Dement 21 , e14495 (2025). https://doi.org/10.1002/alz.14495 Horwitz, L. I. et al. Researching COVID to Enhance Recovery (RECOVER) adult study protocol: Rationale, objectives, and design. PLoS One 18 , e0286297 (2023). https://doi.org/10.1371/journal.pone.0286297 von Gall, C., Holub, L., Ali, A. A. H. & Eickhoff, S. Timing of Deep and REM Sleep Based on Fitbit Sleep Staging in Young Healthy Adults under Real-Life Conditions. Brain Sci 14 (2024). https://doi.org/10.3390/brainsci14030260 Little, T., Lindenberger, U. & Nesselroade, J. R. On selecting indicators for multivariate measurement and modeling with latent variables. Psychological Methods 4 (1999). Desjardins, C. D. & Bulut, O. profileR: An R package for profile analysis. Journal of Open Source Software 5 , 1941 (2020). https://doi.org/https://joss.theoj.org/papers/10.21105/joss.01941 Additional Declarations There is NO Competing Interest. Supplementary Files SleepManuscriptSUpplement20250611.docx Wearable-derived Sleep Measurements are Associated with Long-COVID in the RECOVER Adult Cohort Cite Share Download PDF Status: Under Review Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-7422764","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":505309969,"identity":"e9eff43c-9dd9-4aeb-b58b-ec7732af611c","order_by":0,"name":"Sairam Parthasarathy","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA50lEQVRIiWNgGAWjYLACxgYGHn4gfQDKNyCs5WADg4xkA6labAwOIPj4tcjPbn72+eMOOx7j42cfHvxRcUeegb15mwQ+LQZ3jhnPOHgmmcfsTLrBAYkzzwwbeI6V4dcikWDMcLCNmcfsQBrDAcO2w4wNEjlmeLXIz0j/DNRSz2Pc/4zhQGLbYfsG+Tf4tTDcyAHZcpjHQAJoC5CR2CDBg1+LwZ0zxQxn247zSNx4Bgy5M4eT23jSii3wOmx2+2aGyrZqe/7+NOaPPyoO2/azH954A6/DMNzAhlc5Vi2jYBSMglEwCtABAB1ATdsP6Az4AAAAAElFTkSuQmCC","orcid":"","institution":"University of Arizona","correspondingAuthor":true,"prefix":"","firstName":"Sairam","middleName":"","lastName":"Parthasarathy","suffix":""},{"id":505309970,"identity":"426673c9-72ac-4782-b023-05305a623bf3","order_by":1,"name":"Shari Brosnahan","email":"","orcid":"https://orcid.org/0000-0002-2092-3633","institution":"Division of Pulmonary and Critical Care Medicine, New York University Grossman School of Medicine, NYU Langone Health","correspondingAuthor":false,"prefix":"","firstName":"Shari","middleName":"","lastName":"Brosnahan","suffix":""},{"id":505309971,"identity":"b520a834-a6a3-47b9-a9a4-37902a287dec","order_by":2,"name":"Solveig Sieberts","email":"","orcid":"https://orcid.org/0000-0003-1033-0954","institution":"Sage Bionetworks","correspondingAuthor":false,"prefix":"","firstName":"Solveig","middleName":"","lastName":"Sieberts","suffix":""},{"id":505309972,"identity":"dcc4570c-d11f-4108-a98e-975149149f2e","order_by":3,"name":"Elias Neto","email":"","orcid":"","institution":"Sage Bionetworks","correspondingAuthor":false,"prefix":"","firstName":"Elias","middleName":"","lastName":"Neto","suffix":""},{"id":505309973,"identity":"98bfc74c-af09-4d8b-a852-9e0011729d31","order_by":4,"name":"Yanling Li","email":"","orcid":"","institution":"Pennsylvannia State University","correspondingAuthor":false,"prefix":"","firstName":"Yanling","middleName":"","lastName":"Li","suffix":""},{"id":505309974,"identity":"d4a9a8a4-db83-453d-b60b-d1fbd5f00182","order_by":5,"name":"Meghasyam Tummalacherla","email":"","orcid":"","institution":"Sage Bionetworks","correspondingAuthor":false,"prefix":"","firstName":"Meghasyam","middleName":"","lastName":"Tummalacherla","suffix":""},{"id":505309975,"identity":"04f02ea2-8db1-45b6-b324-e1ece5b51eb8","order_by":6,"name":"Heather-Elizabeth Brown","email":"","orcid":"","institution":"National Community Engagement Group","correspondingAuthor":false,"prefix":"","firstName":"Heather-Elizabeth","middleName":"","lastName":"Brown","suffix":""},{"id":505309976,"identity":"56deb5f3-a502-4e3a-8317-f50d99961d63","order_by":7,"name":"Sy-Miin Chow","email":"","orcid":"","institution":"The Pennsylvania State University","correspondingAuthor":false,"prefix":"","firstName":"Sy-Miin","middleName":"","lastName":"Chow","suffix":""},{"id":505309977,"identity":"312780b1-be20-45f2-aa62-9ca0d5681b7b","order_by":8,"name":"Jessilyn Dunn","email":"","orcid":"https://orcid.org/0000-0002-3241-8183","institution":"Duke University","correspondingAuthor":false,"prefix":"","firstName":"Jessilyn","middleName":"","lastName":"Dunn","suffix":""},{"id":505309978,"identity":"a3797c41-bc68-4546-bdad-6f40654ad7f1","order_by":9,"name":"Monika Haack","email":"","orcid":"","institution":"Beth Israel Deaconness Medical Center","correspondingAuthor":false,"prefix":"","firstName":"Monika","middleName":"","lastName":"Haack","suffix":""},{"id":505309979,"identity":"f967e0fc-3353-4ef5-b016-efc9fb70a38a","order_by":10,"name":"Shekh Md Islam","email":"","orcid":"","institution":"Duke University","correspondingAuthor":false,"prefix":"","firstName":"Shekh","middleName":"Md","lastName":"Is","suffix":"Md"},{"id":505309980,"identity":"b88d0d09-c052-4e12-9843-1fa34cdb7235","order_by":11,"name":"Marissa Jacobs-Diggs","email":"","orcid":"","institution":"Family Advisory Coucil, New York Presbyterian Morgan Stanley Children’s Hospital","correspondingAuthor":false,"prefix":"","firstName":"Marissa","middleName":"","lastName":"Jacobs-Diggs","suffix":""},{"id":505309981,"identity":"249ddfa4-ef00-422d-95cc-962f2f48e7b0","order_by":12,"name":"Yihang Jiang","email":"","orcid":"","institution":"Duke University","correspondingAuthor":false,"prefix":"","firstName":"Yihang","middleName":"","lastName":"Jiang","suffix":""},{"id":505309982,"identity":"0e531be3-f11d-4175-a6ea-d75d4748911f","order_by":13,"name":"Joe Kossowsky","email":"","orcid":"","institution":"Boston Children's Hospital","correspondingAuthor":false,"prefix":"","firstName":"Joe","middleName":"","lastName":"Kossowsky","suffix":""},{"id":505309983,"identity":"f6b475df-6b52-429a-ab28-be5f1bf85cbc","order_by":14,"name":"Aric Prather","email":"","orcid":"","institution":"University of California San Francisco","correspondingAuthor":false,"prefix":"","firstName":"Aric","middleName":"","lastName":"Prather","suffix":""},{"id":505309984,"identity":"a81b6d6a-860f-49b7-9b3b-afefd07a624c","order_by":15,"name":"Nadia Raytselis","email":"","orcid":"","institution":"Emory University","correspondingAuthor":false,"prefix":"","firstName":"Nadia","middleName":"","lastName":"Raytselis","suffix":""},{"id":505309985,"identity":"350f22a8-25d7-409e-93b6-d6fe2014ab78","order_by":16,"name":"Nima Salimi","email":"","orcid":"","institution":"Sage Bionetworks","correspondingAuthor":false,"prefix":"","firstName":"Nima","middleName":"","lastName":"Salimi","suffix":""},{"id":505309986,"identity":"0897fbe7-96e2-43e2-b0ed-8f0f0f77277e","order_by":17,"name":"Mirna Ayache","email":"","orcid":"","institution":"Case Western Reserve University","correspondingAuthor":false,"prefix":"","firstName":"Mirna","middleName":"","lastName":"Ayache","suffix":""},{"id":505309987,"identity":"b7fc7e4c-97d9-4d69-a934-b4b8c2f6cd2c","order_by":18,"name":"Logan Bartram","email":"","orcid":"","institution":"Icahn School of Medicine at Mount Sinai","correspondingAuthor":false,"prefix":"","firstName":"Logan","middleName":"","lastName":"Bartram","suffix":""},{"id":505309988,"identity":"0312785c-c796-4fed-8548-31a58192ad0e","order_by":19,"name":"Jacqueline Becker","email":"","orcid":"","institution":"Icahn School of Medicine at Mount Sinai","correspondingAuthor":false,"prefix":"","firstName":"Jacqueline","middleName":"","lastName":"Becker","suffix":""},{"id":505309989,"identity":"4535a1d5-bbce-4afb-bf43-f4ef3fc51ac7","order_by":20,"name":"Alicia Chung","email":"","orcid":"","institution":"NYU Grossman School of Medicine","correspondingAuthor":false,"prefix":"","firstName":"Alicia","middleName":"","lastName":"Chung","suffix":""},{"id":505309990,"identity":"ccf8dcb7-c257-4092-84a7-af1f42eb0d32","order_by":21,"name":"James DelAlcazar","email":"","orcid":"","institution":"Providence Swedish Medical Center","correspondingAuthor":false,"prefix":"","firstName":"James","middleName":"","lastName":"DelAlcazar","suffix":""},{"id":505309991,"identity":"93d42ed2-15fe-475d-9db8-0cf3f8550850","order_by":22,"name":"Valerie Flaherman","email":"","orcid":"","institution":"University of California San Francisco","correspondingAuthor":false,"prefix":"","firstName":"Valerie","middleName":"","lastName":"Flaherman","suffix":""},{"id":505309992,"identity":"726c908b-a049-452c-8e51-b141cc697b81","order_by":23,"name":"Kelly Gibson","email":"","orcid":"https://orcid.org/0000-0001-8605-7064","institution":"Metrohealth System","correspondingAuthor":false,"prefix":"","firstName":"Kelly","middleName":"","lastName":"Gibson","suffix":""},{"id":505309993,"identity":"a8e4f045-18d4-401d-8038-74e93c802580","order_by":24,"name":"Minjoung Go","email":"","orcid":"","institution":"Stanford University","correspondingAuthor":false,"prefix":"","firstName":"Minjoung","middleName":"","lastName":"Go","suffix":""},{"id":505309994,"identity":"7c2b659d-d921-4c73-b9db-d444ca23c9f9","order_by":25,"name":"Ramkiran Gouripeddi","email":"","orcid":"","institution":"University of Utah","correspondingAuthor":false,"prefix":"","firstName":"Ramkiran","middleName":"","lastName":"Gouripeddi","suffix":""},{"id":505309995,"identity":"fd9ea481-da3d-40d2-a569-2ca210457375","order_by":26,"name":"Jenny Han","email":"","orcid":"","institution":"Emory University","correspondingAuthor":false,"prefix":"","firstName":"Jenny","middleName":"","lastName":"Han","suffix":""},{"id":505309996,"identity":"4e0b1af5-c83c-4903-a016-24333f9b00d5","order_by":27,"name":"Mathew Hoffman","email":"","orcid":"","institution":"University of Utah","correspondingAuthor":false,"prefix":"","firstName":"Mathew","middleName":"","lastName":"Hoffman","suffix":""},{"id":505309997,"identity":"599031ab-3cf6-4656-ac6c-d350837cd258","order_by":28,"name":"Sarah Jolley","email":"","orcid":"","institution":"University of Colorado AMC","correspondingAuthor":false,"prefix":"","firstName":"Sarah","middleName":"","lastName":"Jolley","suffix":""},{"id":505309998,"identity":"fbd2e5ba-3b5f-4ee5-af31-af52e8357a6b","order_by":29,"name":"J. Kelly","email":"","orcid":"https://orcid.org/0000-0002-7616-0321","institution":"University of California, San Francisco","correspondingAuthor":false,"prefix":"","firstName":"J.","middleName":"","lastName":"Kelly","suffix":""},{"id":505309999,"identity":"aafd81f7-c2ff-47ed-b93d-cf1d06737459","order_by":30,"name":"Ziad Koberssy","email":"","orcid":"","institution":"Case Western Reserve University","correspondingAuthor":false,"prefix":"","firstName":"Ziad","middleName":"","lastName":"Koberssy","suffix":""},{"id":505310000,"identity":"0d857722-84a6-4d9e-9b02-eb293ba75e0a","order_by":31,"name":"Jerry Krishnan","email":"","orcid":"","institution":"Division of Pulmonary, Critical Care, Sleep and Allergy, University of Illinois at Chicago","correspondingAuthor":false,"prefix":"","firstName":"Jerry","middleName":"","lastName":"Krishnan","suffix":""},{"id":505310001,"identity":"b96d2cf9-30e8-4461-9a37-2cdfde3532cf","order_by":32,"name":"Adeyinka Laiyemo","email":"","orcid":"","institution":"Howard University College of Medicine","correspondingAuthor":false,"prefix":"","firstName":"Adeyinka","middleName":"","lastName":"Laiyemo","suffix":""},{"id":505310002,"identity":"81ea2390-f006-4d53-8cc9-83f6d735e1d0","order_by":33,"name":"Joyce Lee-Iannotti","email":"","orcid":"","institution":"Barrow Neurological Institute","correspondingAuthor":false,"prefix":"","firstName":"Joyce","middleName":"","lastName":"Lee-Iannotti","suffix":""},{"id":505310003,"identity":"26e8fe4d-722e-4ff3-92ae-000490a8350a","order_by":34,"name":"Emily Levitan","email":"","orcid":"","institution":"University of Alabama at Birmingham","correspondingAuthor":false,"prefix":"","firstName":"Emily","middleName":"","lastName":"Levitan","suffix":""},{"id":505310004,"identity":"2e6c17bd-6800-4559-8884-d61a8ead9ef4","order_by":35,"name":"Diego Mazzotti","email":"","orcid":"https://orcid.org/0000-0003-3924-9199","institution":"University of Kansas Medical Center","correspondingAuthor":false,"prefix":"","firstName":"Diego","middleName":"","lastName":"Mazzotti","suffix":""},{"id":505310005,"identity":"e946efc8-4634-439c-8725-1aefe302ac2b","order_by":36,"name":"Grace McComsey","email":"","orcid":"https://orcid.org/0000-0003-2690-8888","institution":"Case Western Reserve University","correspondingAuthor":false,"prefix":"","firstName":"Grace","middleName":"","lastName":"McComsey","suffix":""},{"id":505310006,"identity":"3153116b-2a17-472c-9b6e-60e952509c05","order_by":37,"name":"Alem Mehari","email":"","orcid":"","institution":"Howard University","correspondingAuthor":false,"prefix":"","firstName":"Alem","middleName":"","lastName":"Mehari","suffix":""},{"id":505310007,"identity":"23578121-bbb4-4b35-989c-10e7fcdd7f7c","order_by":38,"name":"Megumi Okomura","email":"","orcid":"","institution":"University of California San Francisco","correspondingAuthor":false,"prefix":"","firstName":"Megumi","middleName":"","lastName":"Okomura","suffix":""},{"id":505310008,"identity":"4a652516-5988-47b1-ad83-53133b176ab4","order_by":39,"name":"Thomas Patterson","email":"","orcid":"","institution":"University of Texas Health Science Center at San Antonio","correspondingAuthor":false,"prefix":"","firstName":"Thomas","middleName":"","lastName":"Patterson","suffix":""},{"id":505310009,"identity":"638203ea-ecbd-4212-9570-a59b86ddf27b","order_by":40,"name":"Michael Peluso","email":"","orcid":"https://orcid.org/0000-0003-0585-6230","institution":"Division of HIV, Infectious Diseases, and Global Medicine, University of California San Francisco, San Francisco, CA USA","correspondingAuthor":false,"prefix":"","firstName":"Michael","middleName":"","lastName":"Peluso","suffix":""},{"id":505310010,"identity":"724a8be7-ed1a-4155-9fb4-7e900a287d80","order_by":41,"name":"Bharati Prasad","email":"","orcid":"","institution":"University of Illinois","correspondingAuthor":false,"prefix":"","firstName":"Bharati","middleName":"","lastName":"Prasad","suffix":""},{"id":505310011,"identity":"bde53cbf-d431-4dcb-8b38-6e4baaeea94f","order_by":42,"name":"Orlando Quintero","email":"","orcid":"https://orcid.org/0000-0002-1355-6281","institution":"Stanford University","correspondingAuthor":false,"prefix":"","firstName":"Orlando","middleName":"","lastName":"Quintero","suffix":""},{"id":505310012,"identity":"e95a7363-6a97-4dbd-bb2c-17588c5eb172","order_by":43,"name":"A Ryerson","email":"","orcid":"","institution":"Kaiser Permanante Georgia","correspondingAuthor":false,"prefix":"","firstName":"A","middleName":"","lastName":"Ryerson","suffix":""},{"id":505310013,"identity":"61a090b8-cdca-43b9-aba1-d9e4b1f78621","order_by":44,"name":"Prachi Singh","email":"","orcid":"https://orcid.org/0000-0002-7392-9035","institution":"Pennington Biomedical Research Center","correspondingAuthor":false,"prefix":"","firstName":"Prachi","middleName":"","lastName":"Singh","suffix":""},{"id":505310014,"identity":"bd10fdbf-f790-4691-b1ee-dc32a8ee9704","order_by":45,"name":"Upinder Singh","email":"","orcid":"","institution":"Stanford University","correspondingAuthor":false,"prefix":"","firstName":"Upinder","middleName":"","lastName":"Singh","suffix":""},{"id":505310015,"identity":"f117087d-0a6c-4e39-b390-be533222baea","order_by":46,"name":"Monica Verduzco-Gutierrez","email":"","orcid":"","institution":"University of Texas San Antonio","correspondingAuthor":false,"prefix":"","firstName":"Monica","middleName":"","lastName":"Verduzco-Gutierrez","suffix":""},{"id":505310016,"identity":"fc6a7b45-b953-4ef3-981a-9c8ff4e0a2cb","order_by":47,"name":"Peter Whitesell","email":"","orcid":"","institution":"Howard University","correspondingAuthor":false,"prefix":"","firstName":"Peter","middleName":"","lastName":"Whitesell","suffix":""},{"id":505310017,"identity":"0db68439-8196-46cc-97f6-934cef171071","order_by":48,"name":"Natasha Williams","email":"","orcid":"","institution":"New York University","correspondingAuthor":false,"prefix":"","firstName":"Natasha","middleName":"","lastName":"Williams","suffix":""},{"id":505310018,"identity":"b227c40c-67ce-4a5b-baaa-5d528259b9bc","order_by":49,"name":"Juan Wisnivesky","email":"","orcid":"","institution":"Icahn School of Medicine at Mount Sinai","correspondingAuthor":false,"prefix":"","firstName":"Juan","middleName":"","lastName":"Wisnivesky","suffix":""},{"id":505310019,"identity":"d6b72e55-8423-4d84-96a4-b4c75fa5ccdb","order_by":50,"name":"Janet Mullington","email":"","orcid":"https://orcid.org/0000-0002-9156-9424","institution":"Beth Israel Deaconess Medical Center","correspondingAuthor":false,"prefix":"","firstName":"Janet","middleName":"","lastName":"Mullington","suffix":""},{"id":505310020,"identity":"19b65980-43b3-47da-8860-7aaf34830234","order_by":51,"name":"Susan Redline","email":"","orcid":"https://orcid.org/0000-0002-6585-1610","institution":"Brigham and Women's Hospital","correspondingAuthor":false,"prefix":"","firstName":"Susan","middleName":"","lastName":"Redline","suffix":""},{"id":505310021,"identity":"f704b777-daf4-425f-b833-e06b95f75cba","order_by":52,"name":"Elizabeth Karlson","email":"","orcid":"","institution":"Brigham and Women's Hospital,","correspondingAuthor":false,"prefix":"","firstName":"Elizabeth","middleName":"","lastName":"Karlson","suffix":""}],"badges":[],"createdAt":"2025-08-21 06:21:08","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-7422764/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-7422764/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":90501074,"identity":"ec95051f-2098-402b-8eed-80ad6ee18648","added_by":"auto","created_at":"2025-09-03 11:45:55","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":168256,"visible":true,"origin":"","legend":"\u003cp\u003eRECOVER participants were invited to enroll in the Digital Health Program starting in March 2023. For this analysis, we used the first 6-month window of digital health data that was at least 6-months post-infection and compared it to 2 consecutive symptom surveys.\u003c/p\u003e","description":"","filename":"1.png","url":"https://assets-eu.researchsquare.com/files/rs-7422764/v1/45a19fe8c67d8ac10dce10c5.png"},{"id":90501075,"identity":"d18963b3-a330-4912-9bf1-b6c26a02acc4","added_by":"auto","created_at":"2025-09-03 11:45:55","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":274717,"visible":true,"origin":"","legend":"\u003cp\u003eEffect sizes for WDSM associations with LCRI status (black) and individual symptoms of LCRI and LC adjusting for \u003cstrong\u003e(A)\u003c/strong\u003e age, sex and race/ethnicity (model 2) or \u003cstrong\u003e(B) \u003c/strong\u003eage, sex, race/ethnicity and BMI. \u003cstrong\u003e(C)\u003c/strong\u003e Hierarchical clustering of WDSM result in four clusters. The clusters are characterized by good sleep and average resting heart rate and HRV (cluster 1), low resting HR and high HRV and REM breathing rate (cluster 2), low sleep time and high variability (cluster 3) and low sleep efficiency (cluster 4). \u003cstrong\u003e(D)\u003c/strong\u003e The distribution of age, PROMIS Global Mental and Global Physical Health Scores.\u003c/p\u003e","description":"","filename":"2.png","url":"https://assets-eu.researchsquare.com/files/rs-7422764/v1/03b759ad5da28099782671cc.png"},{"id":90502216,"identity":"38d5d7b0-081d-4f25-ace5-e315e80e3c37","added_by":"auto","created_at":"2025-09-03 11:53:56","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":2058199,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7422764/v1/0e9ea60b-1331-464a-a93e-01334324bfc9.pdf"},{"id":90501076,"identity":"b97e9823-8f3a-41b3-94cf-c2a0a7760c05","added_by":"auto","created_at":"2025-09-03 11:45:55","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":13990267,"visible":true,"origin":"","legend":"Wearable-derived Sleep Measurements are Associated with Long-COVID in the RECOVER Adult Cohort","description":"","filename":"SleepManuscriptSUpplement20250611.docx","url":"https://assets-eu.researchsquare.com/files/rs-7422764/v1/855d5c8383e2ea8f728ebd01.docx"}],"financialInterests":"There is \u003cb\u003eNO\u003c/b\u003e Competing Interest.","formattedTitle":"Wearable-derived Sleep Measurements are Associated with Long-COVID in the RECOVER Adult Cohort","fulltext":[{"header":"Introduction","content":"\u003cp\u003eLong COVID (LC) is estimated to affect 6.9% of the U.S. population and costs an estimated \u003cspan\u003e$\u003c/span\u003e3.7 trillion over 5-years and continues to compound the health and economic burden of the nation \u003csup\u003e\u003cspan additionalcitationids=\"CR2 CR3\" citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e\u003c/sup\u003e. However, there is marked under-detection of LC and one of the prominent reasons includes the lack of a clear diagnostic test \u003csup\u003e\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e,\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e\u003c/sup\u003e. In fact, little correlation between standard clinical laboratory tests and patient reported symptoms of LC has been observed \u003csup\u003e\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e\u003c/sup\u003e. Conceivably, diagnostic tests fail to demonstrate consistent associations with LC due to the heterogeneity of this condition that involves multiple organ systems and multiple pathobiological processes \u003csup\u003e\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e,\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e\u003cp\u003eSleep-related symptoms are the third most common symptom of LC and in some reports nearly 40% of patients with LC report sleep disturbances \u003csup\u003e\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e,\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e\u003c/sup\u003e. Moreover, sleep, like LC, is interconnected with multiple organ systems and biological processes such as cognition, immune regulation, and anti-inflammatory processes which are dysregulated in individuals with LC\u003csup\u003e\u003cspan additionalcitationids=\"CR13\" citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e\u003c/sup\u003e. This is further supported by evidence suggesting that preexisting sleep issues may play a role in the development of LC\u003csup\u003e\u003cspan additionalcitationids=\"CR16\" citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e\u003c/sup\u003e. Sleep can be readily measured by wearables \u003csup\u003e\u003cspan additionalcitationids=\"CR19 CR20\" citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e\u003c/sup\u003e. Considering the cross-cutting and interconnected nature of sleep and LC, it is conceivable that wearables-derived sleep measures (WDSM; Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e) may provide a digital indicator of LC. WDSM could further our understanding of the relationship between LC and sleep by enabling longitudinal collection in sufficiently powered cohorts. Accordingly, we investigated whether WDSM could identify differences between individuals with a high burden of LC symptoms versus those that had low or no burden of LC symptoms. As a secondary objective we investigated whether there are differences in WDSM for participants with and without individual symptoms of LC. Finally, we conducted cluster analysis of WDSM to identify multidimensional patterns of sleep changes that are associated with LC and health-related quality of life.\u003c/p\u003e\u003cp\u003e\u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e\u003ccaption language=\"En\"\u003e\u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e\u003cdiv class=\"CaptionContent\"\u003e\u003cp\u003eWearables Derived Sleep Measures (WDSMs) of Multi-Dimensional Model of Sleep\u003c/p\u003e\u003c/div\u003e\u003c/caption\u003e\u003ccolgroup cols=\"3\"\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e\u003cthead\u003e\u003ctr\u003e\u003cth align=\"left\" colname=\"c1\"\u003e\u003cp\u003eDimension\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c2\"\u003e\u003cp\u003eMetric\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c3\"\u003e\u003cp\u003eDescription\u003c/p\u003e\u003c/th\u003e\u003c/tr\u003e\u003c/thead\u003e\u003ctbody\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e\u003cb\u003eQuantity\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eSleep Duration (mins)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eAverage duration-quantity- of sleep during the main sleep period. Sleep duration is measured as the sum of all periods during the main sleep period spent asleep (vs wake).\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\" morerows=\"2\" rowspan=\"3\"\u003e\u003cp\u003e\u003cb\u003eSleep Depth and Architecture\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eMinutes in Light Sleep\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eEstimated time in stages I and 2 sleep summarized across the entire sleep period.\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eMinutes in Deep Sleep\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eEstimated time in stage 3 (deep) sleep summarized across the entire sleep period.\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eMinutes in REM Sleep\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eEstimated time in Rapid Eye Movement (REM) sleep summarized across the entire sleep period.\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\" morerows=\"2\" rowspan=\"3\"\u003e\u003cp\u003e\u003cb\u003eSleep Quality\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eSleep Efficiency (%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003ePercentage time during the overnight sleep (bedtime) period when estimated to be asleep (vs sleep plus wake)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eWASO (Wake After Sleep Onset, minutes)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eTotal duration in minutes spent awake during the overnight (bedtime) period\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eREM fragmentation\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eNumber of times periods of sleep are interrupted by wakefulness\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003e\u003cb\u003eSleep Regularity\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eStandard Deviation (SD) of Sleep Duration (mins)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eThe night-to-night regularity in sleep patterns is described by measuring the extent to which sleep duration and sleep timing varies across time periods of 5 of more consecutive days\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eSD of Mid-Sleep (mins)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eMid-Sleep is measured as the clock time mid- way between sleep onset and wake time.\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003e\u003cb\u003eSleep-Disordered Breathing\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eREM Sleep Breathing Rate (per minute)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eEstimated rate of breathing during REM sleep.\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eSpO2 (%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eAverage oxygen saturation during the sleep period.\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003e\u003cb\u003eSleep-Related Cardiac Activity\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eHeart Rate Variability (HRV; ms)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eBeat to beat variation in heart rate during the sleep period.\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eResting Heart Rate (bpm; beats per minute)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eAverage heart rate during the sleep period.\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003c/tbody\u003e\u003c/colgroup\u003e\u003c/table\u003e\u003c/div\u003e\u003c/p\u003e\u003cp\u003eTo do so, we leveraged the WDSM in adult participants in the RECOVER Adult Cohort Study who were also enrolled in the RECOVER Digital Health Platform: A total of 1,230 individuals with WDSM over 6 months (minimum of 5 days of valid measurements) for a total of 151,045 nights of measurements (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e). The WDSM were complemented by two symptom-surveys over a 6-month period in the RECOVER Adult Observational cohort that yielded information regarding LC status based on the 2024 RECOVER Long COVID Research Index (LCRI) that classifies highly symptomatic LC (\u0026ldquo;likely LC\u0026rdquo;) as Long COVID Research Score\u0026thinsp;\u0026ge;\u0026thinsp;11 as a weighted index of LC symptoms.\u003csup\u003e\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e\u003c/sup\u003e WDSM includes sleep time, as well as metrics of sleep quality and sleep stages that are estimated by monitoring heart rate variability (HRV), the variation in time intervals between heartbeats, movement patterns and oxygen saturation.\u003csup\u003e\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e,\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e\u003c/sup\u003e We analyzed 14 Wearables-derived variables in relation to the LC status (any LCRI score\u0026thinsp;\u0026ge;\u0026thinsp;11 vs LCRI score\u0026thinsp;\u0026lt;\u0026thinsp;11 at all timepoints) measured over 6-months to test the robustness of the observed associations. Although some smaller studies have assessed some of the wearables-derived data in patients with Long COVID \u003csup\u003e\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e,\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e\u003c/sup\u003e they have largely focused on activity-related measures. A comprehensive assessment of WDSMs in relation to well-defined LC status has not yet been performed \u003csup\u003e\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e,\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e\u003c/sup\u003e. Association between WDSM and LC, if present, has wide generalizability and a patient-powered approach for potentially detecting and monitoring the time-based or treatment-driven trajectories of LC.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e"},{"header":"Results","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e\u003ch2\u003eCohort demographics\u003c/h2\u003e\u003cp\u003eAmong the 1,262 sleep digital health participants, 433 were classified as Likely LC (LCRS\u0026thinsp;\u0026ge;\u0026thinsp;11) for at least one of the two timepoints, 529 had lower symptom burden at both timepoints (possible LC; 11\u0026thinsp;\u0026gt;\u0026thinsp;LCRS\u0026thinsp;\u0026gt;\u0026thinsp;0), and 299 had no LC symptoms (LCRS\u0026thinsp;=\u0026thinsp;0) at either timepoint (Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e). Female participants were more likely to be in the LC group (81% female in likely LC and 74% in both the possible LC and no LC symptoms group). White, non-Hispanic participants were more likely to be in the LC group (70%, 68% and 63% for the likely, possible and no LC groups, respectively). In contrast, Asian, non-Hispanic participants were less likely to be in the LC likely group (3%, 5.9% and 11% for the likely, possible and no LC symptoms groups, respectively). The median age at enrollment was similar across groups: 46 years (IQR: 37\u0026ndash;56) for the likely LC group, 44 years (IQR: 34\u0026ndash;59) for the possible LC group, and 43 years (IQR: 28\u0026ndash;58) for the no LC group.\u003c/p\u003e\u003cp\u003e\u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab2\" border=\"1\"\u003e\u003ccaption language=\"En\"\u003e\u003cdiv class=\"CaptionNumber\"\u003eTable 2\u003c/div\u003e\u003cdiv class=\"CaptionContent\"\u003e\u003cp\u003eStudy demographics\u003c/p\u003e\u003c/div\u003e\u003c/caption\u003e\u003ccolgroup cols=\"6\"\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e\u003cthead\u003e\u003ctr\u003e\u003cth align=\"left\" colname=\"c1\"\u003e\u003cp\u003eCharacteristic\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c2\"\u003e\u003cp\u003eHighly Symptomatic\u003c/p\u003e\u003cp\u003eLCRS\u0026thinsp;\u0026ge;\u0026thinsp;11\u003c/p\u003e\u003cp\u003e N\u0026thinsp;=\u0026thinsp;433\u003csup\u003e\u003cem\u003e1\u003c/em\u003e\u003c/sup\u003e\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c3\"\u003e\u003cp\u003ePossible LC\u003c/p\u003e\u003cp\u003e11\u0026thinsp;\u0026gt;\u0026thinsp;LCRS\u0026thinsp;\u0026gt;\u0026thinsp;0\u003c/p\u003e\u003cp\u003e N\u0026thinsp;=\u0026thinsp;529\u003csup\u003e\u003cem\u003e1\u003c/em\u003e\u003c/sup\u003e\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c4\"\u003e\u003cp\u003eNone\u003c/p\u003e\u003cp\u003eLCRS\u0026thinsp;=\u0026thinsp;0\u003c/p\u003e\u003cp\u003e N\u0026thinsp;=\u0026thinsp;299\u003csup\u003e\u003cem\u003e1\u003c/em\u003e\u003c/sup\u003e\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c5\"\u003e\u003cp\u003eUnknown\u003c/p\u003e\u003cp\u003e N\u0026thinsp;=\u0026thinsp;1\u003csup\u003e\u003cem\u003e1\u003c/em\u003e\u003c/sup\u003e\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colspan=\"1\" nameend=\"c6\" namest=\"c6\"\u003e\u0026nbsp;\u003c/th\u003e\u003c/tr\u003e\u003ctr\u003e\u003cth align=\"left\" colname=\"c1\"\u003e\u003cp\u003eSex assigned at birth\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/th\u003e\u003cth align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/th\u003e\u003cth align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/th\u003e\u003cth align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/th\u003e\u003cth align=\"left\" colspan=\"1\" nameend=\"c6\" namest=\"c6\"\u003e\u0026nbsp;\u003c/th\u003e\u003c/tr\u003e\u003c/thead\u003e\u003ctbody\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eMale\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e81 (19%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e136 (26%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e79 (26%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e0 (0%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colspan=\"1\" nameend=\"c6\" namest=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eFemale\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e351 (81%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e388 (74%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e220 (74%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e1 (100%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colspan=\"1\" nameend=\"c6\" namest=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eIntersex\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e0 (0%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e2 (0.4%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e0 (0%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e0 (0%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colspan=\"1\" nameend=\"c6\" namest=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eUnknown\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e1\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e3\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e0\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e0\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colspan=\"1\" nameend=\"c6\" namest=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e\u003cb\u003eAge at enrollment\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e46 (37, 57)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e44 (34, 58)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e45 (33, 61)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e63 (63, 63)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colspan=\"1\" nameend=\"c6\" namest=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e\u003cb\u003eRace and ethnicity\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colspan=\"1\" nameend=\"c6\" namest=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eAsian, Non-Hispanic\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e13 (3.0%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e31 (5.9%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e33 (11%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e0 (0%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colspan=\"1\" nameend=\"c6\" namest=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eBlack, Non-Hispanic\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e34 (7.9%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e46 (8.7%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e19 (6.4%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e1 (100%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colspan=\"1\" nameend=\"c6\" namest=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eHispanic\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e56 (13%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e65 (12%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e42 (14%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e0 (0%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colspan=\"1\" nameend=\"c6\" namest=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eWhite, Non-Hispanic\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e304 (70%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e358 (68%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e189 (63%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e0 (0%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colspan=\"1\" nameend=\"c6\" namest=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eMixed race/Other/Missing\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e26 (6.0%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e29 (5.5%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e16 (5.4%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e0 (0%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colspan=\"1\" nameend=\"c6\" namest=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colspan=\"6\" nameend=\"c6\" namest=\"c1\"\u003e\u003cp\u003e\u003csup\u003e\u003cem\u003e\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e\u003c/em\u003e\u003c/sup\u003e n (%); Median (Q1, Q3)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003c/tbody\u003e\u003c/colgroup\u003e\u003c/table\u003e\u003c/div\u003e\u003c/p\u003e\u003c/div\u003e\n\u003ch3\u003eAssociation between Long COVID and Sleep\u003c/h3\u003e\n\u003cp\u003eWe examined the association between LCRI group and the digital sleep measures using two primary models: age, sex and race/ethnicity (model 2) and age, sex, race/ethnicity and BMI (model 3). We additionally examined an unadjusted model (model 1) and a full model which adjusted for age, sex, race/ethnicity, BMI, alcohol consumption, smoking, seasonality and two data quality metrics (see Methods). The results for these models were qualitatively similar to those for models 2 and 3, respectively. Under model 2 (age, sex and race/ethnicity), likely LC was associated with decreases in sleep efficiency, sleep duration, REM sleep breathing rate, SpO2, HRV, and increases in resting HR, SD of sleep duration, SD of mid-Sleep, and REM onset latency (corrected p-value\u0026thinsp;\u0026le;\u0026thinsp;0.05) Table\u0026nbsp;\u003cspan refid=\"Tab3\" class=\"InternalRef\"\u003e3\u003c/span\u003e; \u003cb\u003eTable S2; Figure \u003cspan refid=\"MOESM1\" class=\"InternalRef\"\u003eS1\u003c/span\u003e\u003c/b\u003e; Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e\u003cb\u003e(A)\u003c/b\u003e). The associations of LCRI with sleep duration and SpO2 become non-significant (adjusted p-value\u0026thinsp;=\u0026thinsp;0.160 and 0.077, respectively) when further including BMI in the model (model 3), implying that these effects may be at least partially explained by BMI (\u003cb\u003eTable S3; Figure \u003cspan refid=\"MOESM1\" class=\"InternalRef\"\u003eS1\u003c/span\u003e;\u003c/b\u003e Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e\u003cb\u003e(B), Figure \u003cspan refid=\"MOESM1\" class=\"InternalRef\"\u003eS1\u003c/span\u003e\u003c/b\u003e). The association of SD of sleep duration and SD of mid-sleep suggest that participants with likely LC show more between-night variability in sleep (or irregular sleep) patterns. A sensitivity analysis with and without weekend days showed similar results (\u003cb\u003eFigure S5\u003c/b\u003e) suggesting that these effects are not driven by social jetlag differences (i.e., differences in sleep timing between weekday and weekends) but instead represent a broader pattern of sleep irregularity. A sensitivity analysis examined the effect of increasing the minimum available data to 14 days (versus 5 in the primary analyses). The results were qualitatively similar to the primary analysis (\u003cb\u003eFigures S3 \u0026amp; S4\u003c/b\u003e).\u003c/p\u003e\u003cp\u003e\u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab3\" border=\"1\"\u003e\u003ccaption language=\"En\"\u003e\u003cdiv class=\"CaptionNumber\"\u003eTable 3\u003c/div\u003e\u003cdiv class=\"CaptionContent\"\u003e\u003cp\u003eDistribution of wearables derived sleep measures (WDSMs) by Long COVID Research Index (LCRI) group\u003c/p\u003e\u003c/div\u003e\u003c/caption\u003e\u003ccolgroup cols=\"3\"\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e\u003cthead\u003e\u003ctr\u003e\u003cth align=\"left\" colname=\"c1\"\u003e\u003cp\u003eCharacteristic\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c2\"\u003e\u003cp\u003eLong COVID\u003c/p\u003e\u003cp\u003eN\u0026thinsp;=\u0026thinsp;477\u003csup\u003e\u003cem\u003e1\u003c/em\u003e\u003c/sup\u003e\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c3\"\u003e\u003cp\u003eNegative Long COVID\u003c/p\u003e\u003cp\u003eN\u0026thinsp;=\u0026thinsp;853\u003csup\u003e\u003cem\u003e1\u003c/em\u003e\u003c/sup\u003e\u003c/p\u003e\u003c/th\u003e\u003c/tr\u003e\u003c/thead\u003e\u003ctbody\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eSleep Efficiency (%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e93.35 (91.01, 94.95)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e93.78 (92.08, 95.17)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eSleep Duration (mins)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e349.19 (300.27, 391.13)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e358.90 (312.96, 399.69)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eMinutes in Deep Sleep\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e60.72 (50.26, 73.45)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e63.47 (52.53, 73.37)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eMinutes in REM Sleep\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e82.68 (67.78, 96.71)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e83.00 (68.94, 95.36)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eREM Sleep Breathing Rate (per minute)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e14.19 (12.97, 15.63)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e14.51 (13.26, 15.79)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eSpO2 (%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e94.75 (93.80, 95.58)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e95.07 (94.18, 95.78)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eHeart Rate Variability (ms)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e22.84 (16.98, 33.02)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e27.85 (20.75, 40.78)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eResting Heart Rate (bpm)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e68.76 (63.30, 74.20)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e66.30 (60.54, 72.09)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eSD of Sleep Duration (mins)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e112.30 (91.62, 139.09)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e98.50 (75.75, 123.03)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eSD of Mid-Sleep (mins)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e24.12 (17.58, 34.07)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e19.79 (13.91, 29.19)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eREM Onset Latency (mins)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e119.50 (99.04, 141.67)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e110.80 (94.33, 131.61)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eREM Fragmentation Index\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e3.33 (2.92, 3.99)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e3.38 (3.00, 3.96)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eMinutes in Light Sleep\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e248.81 (218.90, 275.71)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e247.48 (221.20, 272.90)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eWASO (mins)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e47.22 (38.54, 55.10)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e48.86 (40.59, 57.11)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e\u003csup\u003e\u003cem\u003e\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e\u003c/em\u003e\u003c/sup\u003e Median (Q1, Q3)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003c/tbody\u003e\u003c/colgroup\u003e\u003c/table\u003e\u003c/div\u003e\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\n\u003ch3\u003eAssociations among LC Symptoms and Sleep\u003c/h3\u003e\n\u003cp\u003eWe also examined the LC symptoms associated with the WDSM. The measures included were: brain fog, chest pain, chronic cough, dizziness, fatigue, gastrointestinal (GI) symptoms, head pain, palpitations, post exertional malaise (PEM), shortness of breath (SOB), sleep apnea, sleep disturbance, loss of smell and/or taste, and thirst as described earlier \u003csup\u003e\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e,\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e\u003c/sup\u003e. In models 2 and 3, all LC symptoms were associated with at least one sleep metric: brain fog (9), sleep disturbance (9), dizziness (8), fatigue (8), PEM (7), SOB (6), GI symptoms (7), thirst (6), sleep apnea (5), head pain (3), loss of smell or taste (3), chest pain (3), chronic cough (3), and palpitations (3) (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eA, \u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eB, S9, S10, S12-S25). Overall, decreased sleep efficiency and HRV, as well as increased SD of sleep duration and longer REM onset latency were associated with almost all symptoms examined (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eA, \u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eB; S9, S10, S12-S25). Complete tables by symptom and each model can be found in \u003cb\u003etables S9-S64.\u003c/b\u003e\u003c/p\u003e\u003cp\u003eIncreased brain fog was associated with a decrease in sleep efficiency, sleep duration, minutes in deep sleep, minutes in REM sleep, REM sleep breathing rate, HRV and an increase in SD of sleep duration, SD of mid-sleep, and REM onset latency. Increased resting HR was significant in model 2 but lost significance with the addition of BMI perhaps as higher BMI is associated with higher resting HR \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eA, \u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eB, S15; \u003cb\u003eTables S22, S23).\u003c/b\u003e\u003c/p\u003e\u003cp\u003eDizziness was associated with a decrease in sleep efficiency, minutes in deep sleep, minutes in REM sleep, REM sleep breathing rate, HRV and increase in SD of sleep duration, REM onset latency and minutes in light sleep. For dizziness the associations of several WDSM increased with subsequent regression models (\u003cb\u003eFigure S20; Table S41-S44\u003c/b\u003e).\u003c/p\u003e\u003cp\u003eSelf-reported sleep disturbance was associated with a decrease in sleep efficiency, sleep duration, minutes in REM, REM sleep breathing rate, and HRV, and an increase in resting HR, SD of sleep duration, SD of mid-sleep, and REM onset latency (\u003cb\u003eTables S61-S64\u003c/b\u003e).\u003c/p\u003e\u003cp\u003eFatigue was associated with a decrease in sleep efficacy, sleep duration, minutes in deep sleep, REM sleep breathing rate, HRV and in increased resting HR, SD of sleep duration, and REM onset latency. Similar to brain fog, in both sleep disturbance and fatigue, SpO2 does not retain its significance between model 2 and 3 related to BMI correction, while the REM sleep breathing rate association strengthens in significance with the inclusion of additional covariates (\u003cb\u003eTable S37-S40)\u003c/b\u003e.\u003c/p\u003e\n\u003ch3\u003eSleep and health-related quality of life\u003c/h3\u003e\n\u003cp\u003eWe also examined the association between the WDSMs and two measures of quality of life (QoL) from the Patient-Reported Outcome Measurement Information System (PROMIS) Global Health scale, the Global Physical Health (GPH) index T-scores (4 items on overall physical health, physical function, pain, and fatigue) and Global Mental Health (GMH) T-scores (4 items on quality of life, mental health, satisfaction with social activities, and emotional problems).\u003csup\u003e\u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e\u003c/sup\u003e For each index, better QoL was associated with better sleep efficiency, increased sleep duration, more time in deep sleep, higher HRV, a lower resting HR, more sleep regularity (lower SD of sleep duration and SD of mid-sleep), and shorter REM onset latency. Wake after sleep onset (WASO) and SpO2 were significant in model 2 but did not retain significance when controlled for BMI in both scores as well. Increased time in REM sleep was associated with GPH but not GMH (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eA, \u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eB; \u003cb\u003eFigures S26 \u0026amp; S27; Tables S65-S72\u003c/b\u003e).\u003c/p\u003e\n\u003ch3\u003eCluster analysis of digital sleep measures\u003c/h3\u003e\n\u003cp\u003eTo understand participant patterns with respect to the WDSM, we performed hierarchical clustering based on a subset of digital sleep measures (sleep efficiency, sleep duration, minutes in deep sleep, minutes in REM sleep, REM onset latency, REM sleep breathing rate, SD of sleep duration, HRV and resting HR) \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e\u003cb\u003e(C); Table S73)\u003c/b\u003e. These were chosen as the most consistently correlated with likely LC and presence of LC symptoms. Visual inspection of the scree plot (\u003cb\u003eFigure S28\u003c/b\u003e) identified four primary clusters (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e\u003cb\u003e(C)\u003c/b\u003e). Cluster 1 consists of participants manifesting better cardiovascular fitness with a lower baseline heart rate (mean(SD)\u0026thinsp;=\u0026thinsp;67.94(7.28)), but lower HRV (mean(SD)\u0026thinsp;=\u0026thinsp;25.54(10.19)), with an increased REM Breathing Rate (mean(SD)\u0026thinsp;=\u0026thinsp;14.18(1.70)). Sleep efficiency, duration and REM Onset Latency were high (mean(SD)\u0026thinsp;=\u0026thinsp;94.16(1.9), 381.79(48.74) and 124.25(34.91), respectively), and SD of sleep duration was low (mean(SD)\u0026thinsp;=\u0026thinsp;93.11 (29.09)). Cluster 2 represents a group with higher efficiency, longer sleep duration and lower variability (mean(SD)\u0026thinsp;=\u0026thinsp;94.00(2.30), 366.12(55.66) and 97.62(31.8), respectively). Participants in this cluster displayed an increase both in minutes in REM sleep (mean(SD)\u0026thinsp;=\u0026thinsp;85.13 (20.02)) and a decrease in REM onset latency (mean(SD)\u0026thinsp;=\u0026thinsp;106.82 (24.93)). They showed the highest HRV (mean(SD)\u0026thinsp;=\u0026thinsp;52.04 (22.60)) and lowest resting HR (mean(SD)\u0026thinsp;=\u0026thinsp;61.74(7.15)). The cluster 3 group was characterized by a short sleep time (mean(SD)\u0026thinsp;=\u0026thinsp;287.10(58.26)), reduced minutes in both deep and REM sleep (mean(SD)\u0026thinsp;=\u0026thinsp;54.19(15.96) and 71.72(17.53), respectively), and greater SD of sleep duration (mean(SD)\u0026thinsp;=\u0026thinsp;143.73(43.11)), as well as a higher resting HR (71.36(8.25)). Cluster 4 shares the feature of sleep duration\u0026thinsp;\u0026lt;\u0026thinsp;6h with cluster 3 but is further characterized by low sleep efficiency (mean(SD)\u0026thinsp;=\u0026thinsp;63.36(9.68)) but moderate sleep duration (mean(SD)\u0026thinsp;=\u0026thinsp;337.73(75.04)).\u003c/p\u003e\u003cp\u003eCluster 4 showed the highest proportion of LC positive participants (48%; Table\u0026nbsp;\u003cspan refid=\"Tab4\" class=\"InternalRef\"\u003e4\u003c/span\u003e). It also showed lower GPH and GMH scores (median(IQR)\u0026thinsp;=\u0026thinsp;45 (39\u0026ndash;51) and 45 (39\u0026ndash;52), for Physical and Mental, respectively) \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e\u003cb\u003e(D))\u003c/b\u003e. It also showed a higher proportion of females (84% vs 79%, 72% and 73% for clusters 1, 2, and 3, respectively). In contrast, cluster 2, showed the lowest proportion of likely LC positive participants (20%) with high GPH (median(IQR)\u0026thinsp;=\u0026thinsp;52 (46\u0026ndash;58)) and GMH (median(IQR)\u0026thinsp;=\u0026thinsp;50 (44\u0026ndash;55)). This cluster had younger participants than the other three (median age\u0026thinsp;=\u0026thinsp;36 vs 48, 49, and 47 for clusters 1, 3 and 4, respectively) \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e\u003cb\u003e(D))\u003c/b\u003e. Clusters 1 and 3 showed a moderate proportion of likely LC participants (35% and 41%, respectively). Cluster 1 showed higher GPH scores (median(IQR)\u0026thinsp;=\u0026thinsp;49 (42\u0026ndash;54) versus 46 (39\u0026ndash;52) for cluster 3). It also had a higher proportion of female participants (79% versus 73%).\u003c/p\u003e\u003cp\u003e\u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab4\" border=\"1\"\u003e\u003ccaption language=\"En\"\u003e\u003cdiv class=\"CaptionNumber\"\u003eTable 4\u003c/div\u003e\u003cdiv class=\"CaptionContent\"\u003e\u003cp\u003eDemographic and symptom characteristics by cluster\u003c/p\u003e\u003c/div\u003e\u003c/caption\u003e\u003ccolgroup cols=\"5\"\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e\u003cthead\u003e\u003ctr\u003e\u003cth align=\"left\" colname=\"c1\"\u003e\u003cp\u003eCharacteristic\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c2\"\u003e\u003cp\u003eCluster 1\u003c/p\u003e\u003cp\u003e N\u0026thinsp;=\u0026thinsp;459\u003csup\u003e\u003cem\u003e1\u003c/em\u003e\u003c/sup\u003e\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c3\"\u003e\u003cp\u003eCluster 2\u003c/p\u003e\u003cp\u003e N\u0026thinsp;=\u0026thinsp;266\u003csup\u003e\u003cem\u003e1\u003c/em\u003e\u003c/sup\u003e\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c4\"\u003e\u003cp\u003eCluster 3\u003c/p\u003e\u003cp\u003e N\u0026thinsp;=\u0026thinsp;253\u003csup\u003e\u003cem\u003e1\u003c/em\u003e\u003c/sup\u003e\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c5\"\u003e\u003cp\u003eCluster 4\u003c/p\u003e\u003cp\u003e N\u0026thinsp;=\u0026thinsp;150\u003csup\u003e\u003cem\u003e1\u003c/em\u003e\u003c/sup\u003e\u003c/p\u003e\u003c/th\u003e\u003c/tr\u003e\u003ctr\u003e\u003cth align=\"left\" colname=\"c1\"\u003e\u003cp\u003eLCRI Category\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/th\u003e\u003cth align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/th\u003e\u003cth align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/th\u003e\u003cth align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/th\u003e\u003c/tr\u003e\u003c/thead\u003e\u003ctbody\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003ePositive\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e160 (35%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e54 (20%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e104 (41%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e72 (48%)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eNegative\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e299 (65%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e212 (80%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e149 (59%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e78 (52%)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e\u003cb\u003eSex assigned at birth\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eMale\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e96 (21%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e74 (28%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e69 (27%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e24 (16%)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eFemale\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e361 (79%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e190 (72%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e184 (73%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e124 (84%)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eIntersex\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e1 (0.2%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e1 (0.4%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e0 (0%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e0 (0%)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eUnknown\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e1\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e1\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e0\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e2\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e\u003cb\u003eAge at enrollment\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e48 (36, 60)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e36 (30, 48)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e49 (38, 60)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e47 (38, 57)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e\u003cb\u003eRace and ethnicity\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eNon-Hispanic Asian\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e18 (3.9%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e29 (11%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e22 (8.7%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e3 (2.0%)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eNon-Hispanic Black\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e19 (4.1%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e22 (8.3%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e37 (15%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e9 (6.0%)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eHispanic\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e58 (13%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e35 (13%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e30 (12%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e21 (14%)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eNon-Hispanic White\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e345 (75%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e160 (60%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e145 (57%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e110 (73%)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eMixed race/Other/Missing\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e19 (4.1%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e20 (7.5%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e19 (7.5%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e7 (4.7%)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e\u003cb\u003eSleep Apnea\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e115 (25%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e35 (13%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e94 (38%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e54 (37%)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eUnknown\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e4\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e4\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e3\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e4\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e\u003cb\u003eSleep Disturbance\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e63 (14%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e24 (9.2%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e54 (22%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e41 (28%)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eUnknown\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e4\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e4\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e3\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e4\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e\u003cb\u003ePROMIS Global Physical (t-score)\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e49 (42, 54)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e52 (46, 58)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e46 (39, 52)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e45 (39, 51)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e\u003cb\u003ePROMIS Global Mental (t-score)\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e47 (40, 53)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e50 (44, 55)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e45 (39, 51)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e45 (39, 52)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colspan=\"5\" nameend=\"c5\" namest=\"c1\"\u003e\u003cp\u003e\u003csup\u003e\u003cem\u003e\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e\u003c/em\u003e\u003c/sup\u003e n (%); Median (Q1, Q3)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003c/tbody\u003e\u003c/colgroup\u003e\u003c/table\u003e\u003c/div\u003e\u003c/p\u003e\u003cdiv id=\"Sec8\" class=\"Section2\"\u003e\u003ch2\u003eProfile Analysis\u003c/h2\u003e\u003cp\u003eIn order to further understand the patterns of WDSMs associated with participant self-reports of sleep apnea and sleep disturbance, as well as self-reported quality of life measures (i.e. PROMIS GPH and GMH Indices), we employed Criterion Profile Analysis (CPA) \u003csup\u003e\u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e\u003c/sup\u003e. Sleep apnea was associated with a decrease in REM sleep breathing rate and increase of variability of sleep duration, while sleep disturbance was associated with both those factors, as well as a decrease in sleep efficiency, minutes in REM sleep and an increase in resting HR (\u003cb\u003eFigure S29\u003c/b\u003e). Lower scores for the GPH and GMH measures were associated with decreases in sleep efficiency and minutes in deep sleep, as well as lower HRV (\u003cb\u003eFigure S30\u003c/b\u003e). Lower scores are also associated with increases in REM onset latency and SD of sleep duration. Additionally, an increase in resting HR was associated with lower GPH but did not reach significance for a change of GMH. Likewise, a lower REM sleep breathing rate was associated with lower GMH but did not reach significance for a change of GPH. By comparing the absolute values of CPA scores, we found that variability of sleep duration generally played a more important role than other digital sleep measures in predicting sleep problems and physical/mental health. Significant pattern effects were found for all outcome variables (sleep apnea: F-statistic(17, 1085)\u0026thinsp;=\u0026thinsp;7.69, p\u0026thinsp;\u0026lt;\u0026thinsp;0.001; sleep disturbance: F-statistic(17, 1085)\u0026thinsp;=\u0026thinsp;3.48, p\u0026thinsp;\u0026lt;\u0026thinsp;0.001; GPH: F-statistic(17, 1100)\u0026thinsp;=\u0026thinsp;15.30, p\u0026thinsp;\u0026lt;\u0026thinsp;0.001; GMH: F-statistic(17, 1100)\u0026thinsp;=\u0026thinsp;9.14, p\u0026thinsp;\u0026lt;\u0026thinsp;0.001).\u003c/p\u003e\u003c/div\u003e\n\u003ch3\u003eStratified Analyses\u003c/h3\u003e\n\u003cp\u003eWe performed stratified analyses by age (\u0026lt;\u0026thinsp;45, 45 to 65 and \u0026gt;\u0026thinsp;65) (\u003cb\u003eFigure S31\u003c/b\u003e), sex (\u003cb\u003eFigure S32\u003c/b\u003e) and race and ethnicity (\u003cb\u003eFigure S33\u003c/b\u003e). With respect to age, the associations and directions of effect were similar for most of the variables which were significant in the main analysis (i.e. sleep efficiency, HRV, resting HR, SD of sleep duration, SD of mid-sleep, REM onset latency) (\u003cb\u003eFigure S31\u003c/b\u003e). Even though some associations showed discordant directions none reached statistical significance, with none of the associations reaching statistical significance in the older group (over 65). Given the reduced sample sizes in some of these age groups (n\u0026thinsp;=\u0026thinsp;611, 474, and 145 for the \u0026lt;\u0026thinsp;45, 45\u0026ndash;65, and \u0026gt;\u0026thinsp;65 age groups, respectively), some reduction in power was to be expected. Interestingly, we did observe a significant positive association between likely LC and minutes spent in light sleep in the youngest subgroup, which was not observed in either subgroup or the combined analyses but may be due to reduced power.\u003c/p\u003e\u003cp\u003eThe patterns of change in WDSM with likely LC were generally similar across both sexes (\u003cb\u003eFigure S32\u003c/b\u003e). However, the magnitudes and statistical significance did vary across the sex groups. Both male and females with LC showed significant decrease in HRV, increased in resting HR, SD sleep duration and SD of mid-sleep, while only females showed a significant decrease in sleep efficiency, SpO2 and increase in REM onset latency which was not seen as significant in the males. Likewise, males had a significant decrease in sleep duration and WASO that was not observed in women.\u003c/p\u003e"},{"header":"Discussion","content":"\u003cp\u003eIn this work involving the largest cohort of wearables measurement in individuals with LC, we found biologically plausible differences in WDSM between individuals with likely LC and those with low or no LC symptom burden over a six-month measurement period. Specifically, the strongest association was observed for reduced HRV measured during sleep and increased variability in sleep duration. HRV measures the beat-by-beat variability that reflects the dynamic interaction between sympathetic and parasympathetic tone within the autonomic nervous system and reflects overall health with higher HRV associated with cardiovascular health and better response to stress. There are numerous prior reports of reduced HRV – measured by conventional electrocardiogram – in individuals with LC that could exemplify the underlying autonomic instability that may be triggered or exacerbated by a COVID infection or underlie and/or drive LC presentation and/or symptoms as dysautonomia, Postural Orthostatic Tachycardia Syndrome (POTS), and PEM \u003csup\u003e\u003cb\u003e\u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e,\u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e31\u003c/span\u003e\u003c/b\u003e\u003c/sup\u003e. While many studies have previously used electrocardiograms, Holter monitors, and cardiac belts, the photoplethysmography-based wearables used in our study to yield reliable HRV data poses a promising alternative and practical approach \u003csup\u003e\u003cb\u003e\u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e\u003c/b\u003e\u003c/sup\u003e. Such opportunity has been capitalized by others to demonstrate a reduction in HRV during acute SARS-CoV-2 infection but, to our knowledge, there are no prior wearables-based studies demonstrating the association between reduced HRV and LC\u003csup\u003e\u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e33\u003c/span\u003e\u003c/sup\u003e. The observed association between wearables-derived reduced HRV during sleep and LC further supports autonomic imbalance as one of the mechanistic underpinnings of LC. Additionally, it lends itself for longitudinal monitoring of patients undergoing treatments that targets the autonomic imbalance such as slow-paced breathing. \u003csup\u003e\u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e\u003c/sup\u003e While many studies have previously used electrocardiograms, Holter monitors, and cardiac belts, the photoplethysmography-based wearables used in our study to yield reliable HRV data during sleep poses a promising alternative and practical approach and considering the relative immobility may provide a more accurate measure and minimize artifacts induced by activity \u003csup\u003e\u003cb\u003e\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e,\u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e\u003c/b\u003e\u003c/sup\u003e.\u003c/p\u003e\u003cp\u003eWe also find a strong association of LC with SD of sleep duration and to a lesser degree SD of mid-sleep, measures of sleep irregularity. Irregular sleep duration and timing (markers for circadian misalignment and its myriad effects of healthy physiology) have recently been identified to be associated with insulin resistance and dyslipidemia and to predict incident cardiovascular disease, cognitive decline, and all-cause mortality \u003csup\u003e\u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e35\u003c/span\u003e,\u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e36\u003c/span\u003e\u003c/sup\u003e. In the context of a multi-dimensional model for sleep health, increased irregularity in sleep-wake patterns has been shown to cluster with other sleep dimensions, including insufficient sleep duration, poor sleep efficiency, and sleep-disordered breathing \u003csup\u003e\u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e37\u003c/span\u003e\u003c/sup\u003e. Improving sleep regularity may be an important strategy for more broadly improving sleep health, and in fact is a core principle for sleep hygiene interventions. Interestingly, recent reports indicate increased prevalence of cardiovascular events among individuals hospitalized for COVID-19 or even in mild cases of SARS-CoV2 infection when associated with certain ABO blood types \u003csup\u003e\u003cb\u003e\u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e38\u003c/span\u003e\u003c/b\u003e\u003c/sup\u003e. Coehlo et al have previously described marked instability in polysomnographically measured sleep patterns with both long and short bouts of objectively measured sleep in individuals with LC \u003csup\u003e\u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e39\u003c/span\u003e\u003c/sup\u003e. An ongoing RECOVER clinical trial in fact is targeting sleep regularity as a strategy for improving LC-related sleep disturbance. The observed variability in sleep duration and mid-sleep (timing of nighttime sleep) in our study may serve as a biomarker for further stratification of individuals with greater likelihood of adverse long-term outcomes of LC and presents additional opportunity to augment current risk scores for predicting cardiovascular events in these individuals \u003csup\u003e\u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e38\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e\u003cp\u003eWe observed that LC was also associated with lower sleep efficiency and a tendency for reduced sleep duration. This contrasts with a prior report of similar wearables-measured sleep quantity between individuals with LC and those without LC following a known SARS-CoV-2 infection \u003csup\u003e\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e\u003c/sup\u003e. In this prior study by Radin et al, participants with LC (n = 279) and without LC (n = 274) were followed up to one year but sleep measures besides sleep quality were unfortunately not reported \u003csup\u003e\u003cb\u003e\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e\u003c/b\u003e\u003c/sup\u003e. Conceivably, better case identification using the LCRI and larger sample size of our study (n = 1,230 including 417 likely LC cases) may have resulted in greater power to observe statistical differences, including lower sleep efficiency and tendency for reduced sleep duration in participants with persistent LC compared to those without LC. Similar to the findings in our study, Radin et al found more variable sleep duration and higher resting HR in LC which may signify sleep fragmentation, reduced parasympathetic tone, possibly in response to nocturnal hypoxia and sleep apnea. In our study, reflectance oximetry levels (SpO2) measured by the wearables were lower in individuals with likely LC when compared to those with low or no LC symptoms. Nocturnal hypoxia has been previously described in individuals with LC and may reflect either the greater prevalence of obstructive sleep apnea (OSA) in these individuals with LC, sleep-related hypoventilation, or underlying pulmonary complications of LC \u003csup\u003e\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e,\u003cspan additionalcitationids=\"CR41\" citationid=\"CR40\" class=\"CitationRef\"\u003e40\u003c/span\u003e–\u003cspan citationid=\"CR42\" class=\"CitationRef\"\u003e42\u003c/span\u003e\u003c/sup\u003e. Interestingly, nocturnal hypoxemia can be relatively asymptomatic in acute COVID due to posited alterations in chemosensitivity and control of breathing, but importantly may contribute to delayed detection that could be mitigated by wearables \u003csup\u003e\u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e40\u003c/span\u003e\u003c/sup\u003e. Such a phenomenon of altered chemosensitivity may persist in individuals with LC but this is unclear \u003csup\u003e\u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e40\u003c/span\u003e\u003c/sup\u003e. Obesity may further contribute to sleep-related hypoventilation and consequently lower oxygen saturation, which, in turn may contribute to fatigue and cognitive problems \u003csup\u003e\u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e43\u003c/span\u003e\u003c/sup\u003e. In our study, after adjustment for body mass index (BMI), the association between lower SpO2 and LC was lost suggesting that BMI (and comorbid OSA and sleep-related hypoxia or restrictive pulmonary phenotype) may explain the observed association between oxygen levels and LC\u003csup\u003e\u003cb\u003e\u003cspan additionalcitationids=\"CR45\" citationid=\"CR44\" class=\"CitationRef\"\u003e44\u003c/span\u003e–\u003cspan citationid=\"CR46\" class=\"CitationRef\"\u003e46\u003c/span\u003e\u003c/b\u003e\u003c/sup\u003e. Additionally, our observations of a lower respiratory rate during REM sleep further strengthens the possibility of comorbid OSA and sleep-related hypoxia playing a role in the LC pathophysiology \u003csup\u003e\u003cspan citationid=\"CR47\" class=\"CitationRef\"\u003e47\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e\u003cp\u003eProlonged REM sleep onset latency in individuals with LC has been previously described \u003csup\u003e\u003cspan citationid=\"CR48\" class=\"CitationRef\"\u003e48\u003c/span\u003e\u003c/sup\u003e. The reasons for delayed REM sleep latency are uncertain, but may be due to comorbid OSA and/or medications (e.g., Low Dose Naltrexone, sedative-hypnotics such as zolpidem, and SSRI/SNRI antidepressants prescribed for their “purported” protective effects against LC) that may suppress REM sleep \u003csup\u003e\u003cspan citationid=\"CR49\" class=\"CitationRef\"\u003e49\u003c/span\u003e\u003c/sup\u003e. Prolonged REM sleep latency has been associated with cognitive dysfunction and biomarkers of Alzheimer’s disease \u003csup\u003e\u003cspan citationid=\"CR50\" class=\"CitationRef\"\u003e50\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e\u003cp\u003eIn our study, WDSMs were strongly associated with worse GPH and GMH QoL ratings. Specifically, worse Mental and worse Physical scores were associated with multiple metrics of poor sleep including reduced sleep efficiency, reduced REM related breathing rate, reduced HRV during sleep, increased variability of mid-point sleep and sleep duration, prolonged REM latency, and higher resting HR during sleep. Importantly, such analysis of health-related quality of life measures, individual LC symptoms, and WDSM revealed marked internal consistency and external validity of the directionality of associations. The identification of associations of LC with multiple metrics of sleep is consistent with modern frameworks of sleep health that posit that sleep is a multidimensional construct composed of various interacting features of sleep that together better describe sleep health than individual metrics. The strongest associations between individual symptoms of LC and WDSM were observed for the most common and burdensome LC symptoms -- PEM, brain fog, fatigue, and dizziness.\u003c/p\u003e\u003cp\u003eCluster analysis revealed four distinct patterns indicative of various manifestations of sleep disorder symptoms. A cluster (\u003cb\u003ecluster 3;\u003c/b\u003e Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eb) with clear reduction in sleep duration, increased resting HR, increased variability of sleep duration and reduced time in REM and deep sleep may signify a phenotype characterized by increased arousal and shorter, lighter, and more variable sleep, which may reflect LC-associated “insomnia phenotype”. Whereas, another cluster (cluster 2) is delineated by higher HRV, higher REM breathing rate (i.e., less apnea in REM), and lower resting HR and appears to denote a more “normal” phenotype and likely consisted of a more resilient younger age group of participants (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eb).\u003c/p\u003e\u003cp\u003eWhile the findings represent the cross-cutting nature of sleep and sleep-related physiological changes (heart rate, oxygen saturation), these findings may also denote the bidirectional association between LC and sleep. Taken together, the strong association between WDSMs with LC status and symptoms support their utility as a possible digital biomarker that may be further validated for identifying subgroups of patients with the LC condition who may experience different health trajectories and may benefit from targeted interventions, as well as for monitoring disease course and treatment effectiveness. The findings of lower oxygen saturation levels, lower HRV, and variability in sleep duration and timing with LCRI status provide potential targets for future interventions using positive airway pressure therapy or supplemental oxygen, slow paced breathing/vagal stimulation, and behavioral and circadian (e.g., light) approaches and the very same wearables can assist with monitoring the corresponding treatment response. Finally, identifying clusters of sleep-related metrics that reflect diverse domains of sleep health supports both the importance of applying a multi-dimensional sleep health framework in patients with LC, and suggests that this framework include measures of rest/sleep HRV, sleep variability, sleep efficiency, sleep duration, and metrics related to breathing and oxygenation during sleep.\u003c/p\u003e"},{"header":"Methods","content":"\u003cp\u003e Institutional review boards at NYU Grossman School of Medicine, serving as a single Institutional Review Board, and other participating institutions reviewed and approved the protocol. All participants provided written informed consent to participate in research.\u003c/p\u003e\u003ch2\u003eStudy Design\u003c/h2\u003e\u003cp\u003eIn RECOVER adult study design, participants, aged 18 years or older, were recruited at 83 sites in 33 states plus Washington, DC, and Puerto Rico, regardless of prior infection with SARS-CoV-2 \u003csup\u003e51\u003c/sup\u003e. All participants completed a baseline set of surveys, a focused physical examination, and standard laboratory test sample collection at enrollment. Participants were followed prospectively with survey completion every 3 months and laboratory sample collection at enrollment and 6, 12, 24, 36, and 48 months after infection, or in the case of uninfected controls, the date of a negative test result (index date; Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e). Subjects could enter the study at different times in relation to their initial COVID infection and were defined into cohorts including the acute infected (enrolled during first infection), post-acute (enrolled after initial COVID infection) and uninfected cohort (who enrolled prior to first infection but could be included in this study group if they were infected after study enrollment).\u003csup\u003e\u003cspan citationid=\"CR51\" class=\"CitationRef\"\u003e51\u003c/span\u003e\u003c/sup\u003e\u003c/p\u003e\u003ch2\u003eParticipants\u003c/h2\u003e\u003cp\u003eThe Digital Health Program (DHP) is a component of RECOVER that was launched in 2023 to promote participant engagement and retention, and to collect new information that compliments and augments other RECOVER data including the collection of consumer-grade wearable “wrist-worn” device data (activity, heart rate, and sleep). All RECOVER Adult Cohort participants were invited to sign-up for DHP within an app that they downloaded to their device. They were asked to connect and share data from their own (bring your own (BYO) wearable devices (Fitbit, Apple Watch, Garmin, etc.) or order a free study device (Fitbit smartwatch, Sense 2 or Fitbit tracker, Charge 5).\u003c/p\u003e\u003cp\u003eAs of November 11, 2024, a total of 6,497 RECOVER Adult participants enrolled in DHP, of which 4,104 (63%) connected a wearable device. Comparable wearables-derived sleep measures were only transmitted in Fitbit monitors which accounted for 2,696 (66%) subjects. Of those previously infected participants with Fitbit sleep data, 1,262 (47%) had at least 5 nights of wearables-derived sleep measures concurrent with symptom surveys during the 6-month data collection period and were included in the cohort analysis. An additional 474 (18%) participants had sufficient data but were withheld for future validation of findings.\u003c/p\u003e\u003ch2\u003eExposures and Outcomes\u003c/h2\u003e\u003cp\u003eLikely LC was defined by a Long COVID Research Index (LCRI) Score ≥ 11 at any point in time over a 6-month period post-infection. Possible LC was defined as LCRI score between 1–10, and the no LCRI symptom group had a LCRI score = 0 \u003csup\u003e22\u003c/sup\u003e. The scores were based on the RECOVER data-derived LCRI from 2024 that used Lasso models to discriminate between SARS-CoV2 infected and non-infected participants in the RECOVER Adult Cohort (Geng 2024). Possible LC and no LC was collapsed into an LC indeterminate group \u003csup\u003e\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e\u003c/sup\u003e. Further details for defining LC status are provided in the statistical analysis section (see below). Sensitivity analysis was conducted using the LCRI Score as a continuous variable and referred to the LCRS.\u003c/p\u003e\u003cp\u003eLC symptoms were included based on items in the scoring system for either the RECOVER LCRI (2024) or the 2024 NASEM LC Common Symptoms Definition\u003csup\u003e\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e,\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e,\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e\u003c/sup\u003e. LC outcome variables included: Long COVID status, Brain Fog, Chest Pain, Chronic Cough, Dizziness, Fatigue, Gastrointestinal (GI) Symptoms, Head Pain, Palpitations, Post Exertional Malaise (PEM), Shortness of Breath (SOB), Sleep Apnea, Sleep Disturbance, Loss of Smell and/or Taste, and Thirst. Measures of quality of life included PROMIS GMH (the sum value of the responses to items Global02, Global04, Global05, and Global10r) and PROMIS GPH (sum value of the responses to items Global03, Global06, Global07r, and Global08r) of the PROMIS Scale v1.2. \u003csup\u003e\u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e\u003c/sup\u003e\u003c/p\u003e\u003ch2\u003eWearables-Derived Sleep Measures (WDSM)\u003c/h2\u003e\u003cp\u003eWe selected and summarized a set of Fitbit sleep measures and measures collected during sleep and summarized them over the six-month window. A summary of these sleep metrics, mapped to domains of sleep health, is provided in Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e. This includes the mean of the following measures (with units): sleep efficiency (percent), total sleep duration (minutes), minutes in sleep stages (deep sleep, REM sleep, light sleep), WASO, REM onset latency (minutes), REM fragmentation index,\u003csup\u003e\u003cspan citationid=\"CR52\" class=\"CitationRef\"\u003e52\u003c/span\u003e\u003c/sup\u003e resting heart rate (RHR, beats per minute during sleep and rest), heart rate variability (HRV) during sleep (ms), REM sleep breathing rate (breaths per minute), average oxygen saturation (SpO2; percent). We also examined across-night sleep variability metrics by including the within-individual standard deviation (SD) of sleep duration (minutes) and within-individual SD of sleep midpoint (SD of mid-sleep; minutes). For the standard deviation metrics, we performed a sensitivity analysis excluding weekend days, to examine the degree sleep variability was driven by weekday/weekend differences. For each measure, we required a minimum data availability of 5 days, with at least 3 hours of sleep wear time. Results did not qualitatively change with a minimum data availability of 14 days.\u003c/p\u003e\u003ch2\u003eStatistical Analysis\u003c/h2\u003e\u003cp\u003eTo evaluate the association between WDSM and LC, we took a cross-sectional approach to compare a 6-month window of wearables data with two consecutive symptom surveys (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e). For each participant, we selected the first available symptom survey that followed the first availability of the wearables data and was collected at least 6 months post-infection. The subsequent survey timepoint was selected to be the next available survey that was \u0026gt; 1 month and \u0026lt; 6 months following the first one. If no such survey existed in that time window, the second timepoint was determined to be missing. The 6-month window of digital health data was selected such that it was centered around the two survey timepoints. If the second timepoint was missing, the window was selected to begin 45 days before the available symptom survey. Because each wearables measure had different patterns of data availability, this window selection was done independently by measure.\u003c/p\u003e\u003cp\u003eUsing the symptom surveys at the two timepoints, we combined the LCRI and LCRS for each participant at each timepoint. The LCRS was taken to be the average value at the two timepoints. A participant was defined as likely LC if their LCRI score was ≥ 11 at either timepoint. For participants with only the first timepoint, we classified them based only on the available data. To examine the implicit assumption “last observation carried forward”, specifically in the LC negative group (LCRI \u0026lt; 11), we performed a sensitivity analysis which excluded these participants. These analyses showed no qualitative differences to the primary analyses.\u003c/p\u003e\u003ch2\u003eAssociation between sleep measures and long COVID\u003c/h2\u003e\u003cp\u003eLogistic regression was used to assess the association between each sleep measure and LCRI or individual symptoms, with the sleep measures being the independent variable. Analogously, associations between sleep measures and continuous measures (LCRS, PROMIS GPH and PROMIS GMH) were assessed using linear regression with the sleep measure as the independent variable.\u003c/p\u003e\u003cp\u003eFor each comparison, four different models were considered. Our primary model (model 3) adjusted for known explanatory variables (age, sex, race and ethnicity, and BMI). Additionally, we explored additional variables (alcohol, smoking, seasonality, and time between surveys that may influence LC status:\u003c/p\u003e\u003cp\u003eLCRI/LC Symptom ~ Digital Sleep Measure (1)\u003c/p\u003e\u003cp\u003eLCRI/LC Symptom ~ Digital Sleep Measure + Age + Sex + Race/Ethnicity (2)\u003c/p\u003e\u003cp\u003eLCRI/LC Symptom ~ Digital Sleep Measure + Age + Sex + Race/Ethnicity + BMI (3)\u003c/p\u003e\u003cp\u003eLCRI/LC Symptom ~ Digital Sleep Measure + Age + Sex + Race/Ethnicity + BMI +\u003c/p\u003e\u003cp\u003eAlcohol consumption + Smoking + Seasonality +\u003c/p\u003e\u003cp\u003etime between surveys + n\u003csub\u003erecords, digital\u003c/sub\u003e (Full)\u003c/p\u003e\u003cp\u003ewhere n\u003csub\u003erecords, digital\u003c/sub\u003e is the number of Fitbit sleep records in the time window. Because of the small number of participants who are intersex or of unknown sex, these individuals were grouped with the male sex group for the purpose of analysis. To account for multiple significance testing, \u003cem\u003ep\u003c/em\u003e-values were adjusted using the Benjamini-Hochberg method within LCRI, LCRS and the set of symptoms separately for each model (q-value \u0026lt; 0.05).\u003c/p\u003e\u003ch2\u003eClustering\u003c/h2\u003e\u003cp\u003eParticipants were clustered into subgroups using hierarchical clustering based on a subset of digital sleep measures that were strongly correlated with the GPH and GMH variables (for external validity) but were relatively distinct (showing low correlations) from each other (for discriminant validity) \u003csup\u003e\u003cspan citationid=\"CR53\" class=\"CitationRef\"\u003e53\u003c/span\u003e\u003c/sup\u003e. Clusters were generated using the hclust() function in the R “stats” package using the Ward.D2 agglomeration method which minimizes the within-cluster variance. The resulting dendrogram was cut using a hybrid algorithm in the cutreeDynamic() function in the R “dynamicTreeCut” package. The optimal number of clusters was determined based on both the scree plot which shows how the within-cluster sum of squares varies with different numbers of clusters, and the Bayesian Information Criterion (BIC) plot generated using the mclustBIC() function in the R “mclust” package. We further investigated the demographics, presence of sleep-related symptoms, and distribution of quality of life measures within each cluster to identify distinct patterns in subgroups.\u003c/p\u003e\u003ch2\u003eCriterion Profile Analysis\u003c/h2\u003e\u003cp\u003eThe criterion profile analysis (CPA) was used to determine whether there was a pattern of digital sleep measures associated with the presence of sleep-related symptoms or high scores on quality of life measures\u003csup\u003e\u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e\u003c/sup\u003e. The relative importance of predictors (i.e., the digital sleep measures) was quantified using the criterion pattern score, which was defined as the standardized regression coefficient of a given predictor minus the average of all standardized regression coefficients in the model, multiplied by a positive constant, such that higher absolute criterion pattern scores indicate higher importance in predicting the outcome of interest. Under CPA, the predictor scores are decomposed into two components: a level component (defined as the mean of the predictor scores) and a pattern component (defined as the deviation of each predictor score from the mean predictor score), such that the total variance of the outcome is explained by both the level and the pattern of the predictors. The level and pattern effects can be tested via F-tests by comparing the full model and the model with only level/pattern components. The analysis was done using the cpa() function in the R “profileR” package\u003csup\u003e\u003cspan citationid=\"CR54\" class=\"CitationRef\"\u003e54\u003c/span\u003e\u003c/sup\u003e. Results were generated to show digital sleep measures associated with \u003cem\u003epresence\u003c/em\u003e of sleep symptoms (e.g., apnea and sleep disturbance) and \u003cem\u003edecreased\u003c/em\u003e global health scores.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cem\u003e\u003cu\u003eData availability:\u003c/u\u003e\u003c/em\u003e The RECOVER Adult Cohort Observational Study remains in progress and therefore datasets are updated dynamically, individual participant-level questionnaire-based data including the data dictionary is available on the RECOVER website (https://recovercovid.org/). NHLBI has undertaken a significant effort to release harmonized data from all RECOVER observational cohort studies to the public via their BioData Catalyst platform (https://biodatacatalyst.nhlbi.nih.gov/). Datasets include participant-level data collected for the present study, as well as the technical infrastructure and analytic tools to evaluate the results and conclusions from this study.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConflicts of Interest/Disclosures:\u003c/strong\u003e Dr. Wisnivesky received honorarium from Sanofi, Banook, PPD, and AMA and grants from Sanofi, Regeneron, Axella, and Boehringer Ingelheim. Dr. Mazziotti has been awarded devices from SleepImage, Inc. Dr. Raytselis has enrolled as a participant in the RECOVER Observational study and is a patient representative. Dr. Levitan has received research funding from Amgen and personal fees from the University of Pittsburgh for serving on a DSMB, unrelated to the current work. Dr. Dunn is a Scientific Advisor to Veri, Inc. and Biio, Inc. Dr. Verduzco-Gutierrez has an AHRQ grant for Long COVID. Dr. Peluso has received consulting fees from Gilead Sciences, AstraZeneca, BioVie, Apellis Pharmaceuticals, and BioNTech, travel support from Invivyd, and research support from Aerium Therapeutics and Shionogi, outside the submitted work. Dr. Upinder Singh received funding from Pfizer for studies on Long-Covid. Dr. McComsey served as research consultant for Merck, Gilead, and GlaxoSmithKline unrelated to the current work. Dr. Redline has received funding from NIH and consulting fees from Eli Lilly Inc. She is an unpaid board member of the Alliance of Sleep Apnea Partners and National Sleep Foundation, and is Editor-In-Chief of Sleep Health. Dr. Parthasarathy has received consulting fees from Abbvie, Inc., Apria Healthcare Inc., Zoll Medical, Inc., Jazz Pharmaceuticals, Inc., DynaMed, LLC, Fisher \u0026amp; Paykel, Inc., Apnimed, Inc., and has received grants to the institution from Phillips, Inc., Verily Lifesciences, Inc., Regeneron, Inc., US Biotest, Inc., WHOOP, Inc., and Sommetrics, Inc.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding/Support:\u003c/strong\u003e This research was funded by the NIH OT2HL161847 as part of the Researching COVID to Enhance Recovery (RECOVER) research program. Additional support for Dr. Parthasarathy from Arizona CEAL (OT2HL158287) and R25HL126140 during the writing of this manuscript.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eRole of the Funder/Sponsor:\u003c/strong\u003e This content is solely the responsibility of the authors and does not necessarily represent the official views of the RECOVER Program, the NIH, or other funders.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAcknowledgements:\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWe would like to thank the National Community Engagement Group (NCEG), all patient, caregiver and community Representatives, and all the participants enrolled in the RECOVER Initiative.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eThis study is part of the NIH Researching COVID to Enhance Recovery (RECOVER) Initiative, which seeks to understand, treat, and prevent the post-acute sequelae of SARS-CoV-2 infection (PASC). For more information on RECOVER visit https://recovercovid.org/ .\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eFord, N. D.\u003cem\u003e et al.\u003c/em\u003e Notes from the Field: Long COVID Prevalence Among Adults - United States, 2022. \u003cem\u003eMMWR Morb Mortal Wkly Rep\u003c/em\u003e \u003cstrong\u003e73\u003c/strong\u003e, 135-136 (2024). https://doi.org/10.15585/mmwr.mm7306a4\u003c/li\u003e\n\u003cli\u003eCutler, D. M. The Costs of Long COVID. \u003cem\u003eJAMA Health Forum\u003c/em\u003e \u003cstrong\u003e3\u003c/strong\u003e, e221809 (2022). https://doi.org/10.1001/jamahealthforum.2022.1809\u003c/li\u003e\n\u003cli\u003e\u003cem\u003e2025 Long COVID Fact Sheet\u003c/em\u003e, \u0026lt;https://patientresearchcovid19.com/2025-long-covid-fact-sheet/\u0026gt; (2025).\u003c/li\u003e\n\u003cli\u003eCutler, D. M. \u003cem\u003eThe Economic Cost of Long COVID: An Update\u003c/em\u003e, 2025).\u003c/li\u003e\n\u003cli\u003eTsilingiris, D.\u003cem\u003e et al.\u003c/em\u003e Laboratory Findings and Biomarkers in Long COVID: What Do We Know So Far? Insights into Epidemiology, Pathogenesis, Therapeutic Perspectives and Challenges. \u003cem\u003eInt J Mol Sci\u003c/em\u003e \u003cstrong\u003e24\u003c/strong\u003e (2023). https://doi.org/10.3390/ijms241310458\u003c/li\u003e\n\u003cli\u003e\u003cem\u003eNational Center for Health Statistics. U.S. Census Bureau, Household Pulse Survey, 2022-2024. Long COVID. \u003c/em\u003e, \u0026lt;https://www.cdc.gov/nchs/covid19/pulse/long-covid.htm\u0026gt; (2025).\u003c/li\u003e\n\u003cli\u003eErlandson, K. M.\u003cem\u003e et al.\u003c/em\u003e Differentiation of Prior SARS-CoV-2 Infection and Postacute Sequelae by Standard Clinical Laboratory Measurements in the RECOVER Cohort. \u003cem\u003eAnn Intern Med\u003c/em\u003e \u003cstrong\u003e177\u003c/strong\u003e, 1209-1221 (2024). https://doi.org/10.7326/M24-0737\u003c/li\u003e\n\u003cli\u003e\u003cem\u003eNational Academies of Sciences, Engineering, and Medicine; Health and Medicine Division; Board on Global Health; Board on Health Sciences Policy; Committee on Examining the Working Definition for Long COVID\u003c/em\u003e. (National Academies Press (US), 2024).\u003c/li\u003e\n\u003cli\u003ePerumal, R.\u003cem\u003e et al.\u003c/em\u003e Long COVID: a review and proposed visualization of the complexity of long COVID. \u003cem\u003eFront Immunol\u003c/em\u003e \u003cstrong\u003e14\u003c/strong\u003e, 1117464 (2023). https://doi.org/10.3389/fimmu.2023.1117464\u003c/li\u003e\n\u003cli\u003ePena-Orbea, C.\u003cem\u003e et al.\u003c/em\u003e Sleep Disturbance Severity and Correlates in Post-acute Sequelae of COVID-19 (PASC). \u003cem\u003eJ Gen Intern Med\u003c/em\u003e \u003cstrong\u003e38\u003c/strong\u003e, 2015-2017 (2023). https://doi.org/10.1007/s11606-023-08187-3\u003c/li\u003e\n\u003cli\u003eTedjasukmana, R., Budikayanti, A., Islamiyah, W. R., Witjaksono, A. \u0026amp; Hakim, M. Sleep disturbance in post COVID-19 conditions: Prevalence and quality of life. \u003cem\u003eFront Neurol\u003c/em\u003e \u003cstrong\u003e13\u003c/strong\u003e, 1095606 (2022). https://doi.org/10.3389/fneur.2022.1095606\u003c/li\u003e\n\u003cli\u003eGarbarino, S., Lanteri, P., Bragazzi, N. L., Magnavita, N. \u0026amp; Scoditti, E. Role of sleep deprivation in immune-related disease risk and outcomes. \u003cem\u003eCommun Biol\u003c/em\u003e \u003cstrong\u003e4\u003c/strong\u003e, 1304 (2021). https://doi.org/10.1038/s42003-021-02825-4\u003c/li\u003e\n\u003cli\u003eSimon, K. C., Nadel, L. \u0026amp; Payne, J. D. The functions of sleep: A cognitive neuroscience perspective. \u003cem\u003eProc Natl Acad Sci U S A\u003c/em\u003e \u003cstrong\u003e119\u003c/strong\u003e, e2201795119 (2022). https://doi.org/10.1073/pnas.2201795119\u003c/li\u003e\n\u003cli\u003eXie, L.\u003cem\u003e et al.\u003c/em\u003e Sleep drives metabolite clearance from the adult brain. \u003cem\u003eScience\u003c/em\u003e \u003cstrong\u003e342\u003c/strong\u003e, 373-377 (2013). https://doi.org/10.1126/science.1241224\u003c/li\u003e\n\u003cli\u003eH, L. M.\u003cem\u003e et al.\u003c/em\u003e Risk of post-acute sequelae of SARS-CoV-2 infection associated with pre-coronavirus disease obstructive sleep apnea diagnoses: an electronic health record-based analysis from the RECOVER initiative. \u003cem\u003eSleep\u003c/em\u003e \u003cstrong\u003e46\u003c/strong\u003e (2023). https://doi.org/10.1093/sleep/zsad126\u003c/li\u003e\n\u003cli\u003eSchilling, C.\u003cem\u003e et al.\u003c/em\u003e Pre-existing sleep problems as a predictor of post-acute sequelae of COVID-19. \u003cem\u003eJ Sleep Res\u003c/em\u003e \u003cstrong\u003e33\u003c/strong\u003e, e13949 (2024). https://doi.org/10.1111/jsr.13949\u003c/li\u003e\n\u003cli\u003eQuan, S. F.\u003cem\u003e et al.\u003c/em\u003e Sleep and long COVID: preexisting sleep issues and the risk of post-acute sequelae of SARS-CoV-2 infection in a large general population using 3 different model definitions. \u003cem\u003eJ Clin Sleep Med\u003c/em\u003e \u003cstrong\u003e21\u003c/strong\u003e, 249-259 (2025). https://doi.org/10.5664/jcsm.11322\u003c/li\u003e\n\u003cli\u003ede Zambotti, M., Cellini, N., Goldstone, A., Colrain, I. M. \u0026amp; Baker, F. C. Wearable Sleep Technology in Clinical and Research Settings. \u003cem\u003eMed Sci Sports Exerc\u003c/em\u003e \u003cstrong\u003e51\u003c/strong\u003e, 1538-1557 (2019). https://doi.org/10.1249/MSS.0000000000001947\u003c/li\u003e\n\u003cli\u003ePhillips, A. J. K.\u003cem\u003e et al.\u003c/em\u003e Irregular sleep/wake patterns are associated with poorer academic performance and delayed circadian and sleep/wake timing. \u003cem\u003eSci Rep\u003c/em\u003e \u003cstrong\u003e7\u003c/strong\u003e, 3216 (2017). https://doi.org/10.1038/s41598-017-03171-4\u003c/li\u003e\n\u003cli\u003eLee, T.\u003cem\u003e et al.\u003c/em\u003e Accuracy of 11 Wearable, Nearable, and Airable Consumer Sleep Trackers: Prospective Multicenter Validation Study. \u003cem\u003eJMIR Mhealth Uhealth\u003c/em\u003e \u003cstrong\u003e11\u003c/strong\u003e, e50983 (2023). https://doi.org/10.2196/50983\u003c/li\u003e\n\u003cli\u003eBerryhill, S.\u003cem\u003e et al.\u003c/em\u003e Effect of wearables on sleep in healthy individuals: a randomized crossover trial and validation study. \u003cem\u003eJ Clin Sleep Med\u003c/em\u003e \u003cstrong\u003e16\u003c/strong\u003e, 775-783 (2020). https://doi.org/10.5664/jcsm.8356\u003c/li\u003e\n\u003cli\u003eGeng, L. N.\u003cem\u003e et al.\u003c/em\u003e 2024 Update of the RECOVER-Adult Long COVID Research Index. \u003cem\u003eJAMA\u003c/em\u003e (2024). https://doi.org/10.1001/jama.2024.24184\u003c/li\u003e\n\u003cli\u003eHaghayegh, S., Khoshnevis, S., Smolensky, M. H., Diller, K. R. \u0026amp; Castriotta, R. J. Accuracy of Wristband Fitbit Models in Assessing Sleep: Systematic Review and Meta-Analysis. \u003cem\u003eJ Med Internet Res\u003c/em\u003e \u003cstrong\u003e21\u003c/strong\u003e, e16273 (2019). https://doi.org/10.2196/16273\u003c/li\u003e\n\u003cli\u003eRadin, J. M.\u003cem\u003e et al.\u003c/em\u003e Long-term changes in wearable sensor data in people with and without Long Covid. \u003cem\u003eNPJ Digit Med\u003c/em\u003e \u003cstrong\u003e7\u003c/strong\u003e, 246 (2024). https://doi.org/10.1038/s41746-024-01238-x\u003c/li\u003e\n\u003cli\u003eKukreti, S., Lu, M. T., Yeh, C. Y. \u0026amp; Ko, N. Y. Physiological Sensors Equipped in Wearable Devices for Management of Long COVID Persisting Symptoms: Scoping Review. \u003cem\u003eJ Med Internet Res\u003c/em\u003e \u003cstrong\u003e27\u003c/strong\u003e, e69506 (2025). https://doi.org/10.2196/69506\u003c/li\u003e\n\u003cli\u003eMekhael, M.\u003cem\u003e et al.\u003c/em\u003e Studying the Effect of Long COVID-19 Infection on Sleep Quality Using Wearable Health Devices: Observational Study. \u003cem\u003eJ Med Internet Res\u003c/em\u003e \u003cstrong\u003e24\u003c/strong\u003e, e38000 (2022). https://doi.org/10.2196/38000\u003c/li\u003e\n\u003cli\u003eThaweethai, T.\u003cem\u003e et al.\u003c/em\u003e Development of a Definition of Postacute Sequelae of SARS-CoV-2 Infection. \u003cem\u003eJAMA\u003c/em\u003e \u003cstrong\u003e329\u003c/strong\u003e, 1934-1946 (2023). https://doi.org/10.1001/jama.2023.8823\u003c/li\u003e\n\u003cli\u003eHays, R. D., Bjorner, J. B., Revicki, D. A., Spritzer, K. L. \u0026amp; Cella, D. Development of physical and mental health summary scores from the patient-reported outcomes measurement information system (PROMIS) global items. \u003cem\u003eQual Life Res\u003c/em\u003e \u003cstrong\u003e18\u003c/strong\u003e, 873-880 (2009). https://doi.org/10.1007/s11136-009-9496-9\u003c/li\u003e\n\u003cli\u003eDavison, M. L. \u0026amp; Davenport, E. C., Jr. Identifying criterion-related patterns of predictor scores using multiple regression. \u003cem\u003ePsychol Methods\u003c/em\u003e \u003cstrong\u003e7\u003c/strong\u003e, 468-484 (2002). https://doi.org/10.1037/1082-989x.7.4.468\u003c/li\u003e\n\u003cli\u003eHaischer, M. H.\u003cem\u003e et al.\u003c/em\u003e Heart rate variability is reduced in COVID-19 survivors and associated with physical activity and fatigue. \u003cem\u003ePhysiol Rep\u003c/em\u003e \u003cstrong\u003e12\u003c/strong\u003e, e15912 (2024). https://doi.org/10.14814/phy2.15912\u003c/li\u003e\n\u003cli\u003eda Silva, A. L. G.\u003cem\u003e et al.\u003c/em\u003e Impact of long COVID on the heart rate variability at rest and during deep breathing maneuver. \u003cem\u003eSci Rep\u003c/em\u003e \u003cstrong\u003e13\u003c/strong\u003e, 22695 (2023). https://doi.org/10.1038/s41598-023-50276-0\u003c/li\u003e\n\u003cli\u003eGeorgiou, K.\u003cem\u003e et al.\u003c/em\u003e Can Wearable Devices Accurately Measure Heart Rate Variability? A Systematic Review. \u003cem\u003eFolia Med (Plovdiv)\u003c/em\u003e \u003cstrong\u003e60\u003c/strong\u003e, 7-20 (2018). https://doi.org/10.2478/folmed-2018-0012\u003c/li\u003e\n\u003cli\u003eSanches, C. A., Silva, G. A., Librantz, A. F. H., Sampaio, L. M. M. \u0026amp; Belan, P. A. Wearable Devices to Diagnose and Monitor the Progression of COVID-19 Through Heart Rate Variability Measurement: Systematic Review and Meta-Analysis. \u003cem\u003eJ Med Internet Res\u003c/em\u003e \u003cstrong\u003e25\u003c/strong\u003e, e47112 (2023). https://doi.org/10.2196/47112\u003c/li\u003e\n\u003cli\u003eMauro, M., Cegolon, L., Bestiaco, N., Zulian, E. \u0026amp; Larese Filon, F. Heart Rate Variability Modulation Through Slow-Paced Breathing in Health Care Workers with Long COVID: A Case-Control Study. \u003cem\u003eAm J Med\u003c/em\u003e \u003cstrong\u003e138\u003c/strong\u003e, 870-883.e875 (2025). https://doi.org/10.1016/j.amjmed.2024.05.021\u003c/li\u003e\n\u003cli\u003eLeproult, R., Holmback, U. \u0026amp; Van Cauter, E. Circadian misalignment augments markers of insulin resistance and inflammation, independently of sleep loss. \u003cem\u003eDiabetes\u003c/em\u003e \u003cstrong\u003e63\u003c/strong\u003e, 1860-1869 (2014). https://doi.org/10.2337/db13-1546\u003c/li\u003e\n\u003cli\u003eHuang, T., Mariani, S. \u0026amp; Redline, S. Sleep Irregularity and Risk of Cardiovascular Events: The Multi-Ethnic Study of Atherosclerosis. \u003cem\u003eJ Am Coll Cardiol\u003c/em\u003e \u003cstrong\u003e75\u003c/strong\u003e, 991-999 (2020). https://doi.org/10.1016/j.jacc.2019.12.054\u003c/li\u003e\n\u003cli\u003eTracy, E. L.\u003cem\u003e et al.\u003c/em\u003e Multidimensional Sleep and Self-Rated Physical Health and Depressive Symptoms Among Retired Older Adults: A Sex-Stratified Analysis. \u003cem\u003eBehav Sleep Med\u003c/em\u003e, 1-12 (2025). https://doi.org/10.1080/15402002.2025.2493651\u003c/li\u003e\n\u003cli\u003eHilser, J. R.\u003cem\u003e et al.\u003c/em\u003e COVID-19 Is a Coronary Artery Disease Risk Equivalent and Exhibits a Genetic Interaction With ABO Blood Type. \u003cem\u003eArterioscler Thromb Vasc Biol\u003c/em\u003e \u003cstrong\u003e44\u003c/strong\u003e, 2321-2333 (2024). https://doi.org/10.1161/atvbaha.124.321001\u003c/li\u003e\n\u003cli\u003eCoelho, F. M. S.\u003cem\u003e et al.\u003c/em\u003e Sleep disorder syndromes of post-acute sequelae of SARS-CoV-2 (PASC) / Long Covid. \u003cem\u003eSleep Med\u003c/em\u003e \u003cstrong\u003e123\u003c/strong\u003e, 37-41 (2024). https://doi.org/10.1016/j.sleep.2024.08.030\u003c/li\u003e\n\u003cli\u003eTobin, M. J., Laghi, F. \u0026amp; Jubran, A. Why COVID-19 Silent Hypoxemia Is Baffling to Physicians. \u003cem\u003eAm J Respir Crit Care Med\u003c/em\u003e \u003cstrong\u003e202\u003c/strong\u003e, 356-360 (2020). https://doi.org/10.1164/rccm.202006-2157CP\u003c/li\u003e\n\u003cli\u003eHalawa, S.\u003cem\u003e et al.\u003c/em\u003e Potential long-term effects of SARS-CoV-2 infection on the pulmonary vasculature: a global perspective. \u003cem\u003eNat Rev Cardiol\u003c/em\u003e \u003cstrong\u003e19\u003c/strong\u003e, 314-331 (2022). https://doi.org/10.1038/s41569-021-00640-2\u003c/li\u003e\n\u003cli\u003eSun, H.\u003cem\u003e et al.\u003c/em\u003e Facility-Measured Nocturnal Hypoxemia and Sleep Among Adults with Long COVID versus Age- and Sex-Matched Healthy Adults: A Preliminary Observational Study. \u003cem\u003eSLEEP Advances\u003c/em\u003e (2025). https://doi.org/10.1093/sleepadvances/zpaf017\u003c/li\u003e\n\u003cli\u003eSubramanian, A.\u003cem\u003e et al.\u003c/em\u003e Symptoms and risk factors for long COVID in non-hospitalized adults. \u003cem\u003eNat Med\u003c/em\u003e \u003cstrong\u003e28\u003c/strong\u003e, 1706-1714 (2022). https://doi.org/10.1038/s41591-022-01909-w\u003c/li\u003e\n\u003cli\u003eKendzerska, T., Leung, R. S., Gershon, A. S., Tomlinson, G. \u0026amp; Ayas, N. The Interaction of Obesity and Nocturnal Hypoxemia on Cardiovascular Consequences in Adults with Suspected Obstructive Sleep Apnea. A Historical Observational Study. \u003cem\u003eAnn Am Thorac Soc\u003c/em\u003e \u003cstrong\u003e13\u003c/strong\u003e, 2234-2241 (2016). https://doi.org/10.1513/AnnalsATS.201604-263OC\u003c/li\u003e\n\u003cli\u003eGabbay, I. E., Gabbay, U. \u0026amp; Lavie, P. Obesity plays an independent worsening modifying effect on nocturnal hypoxia in obstructive sleep apnea. \u003cem\u003eSleep Med\u003c/em\u003e \u003cstrong\u003e13\u003c/strong\u003e, 524-528 (2012). https://doi.org/10.1016/j.sleep.2012.01.005\u003c/li\u003e\n\u003cli\u003eGuerra, S.\u003cem\u003e et al.\u003c/em\u003e Morbidity and mortality associated with the restrictive spirometric pattern: a longitudinal study. \u003cem\u003eThorax\u003c/em\u003e \u003cstrong\u003e65\u003c/strong\u003e, 499-504 (2010). https://doi.org/10.1136/thx.2009.126052\u003c/li\u003e\n\u003cli\u003eBonsignore, M. R.\u003cem\u003e et al.\u003c/em\u003e REM sleep obstructive sleep apnoea. \u003cem\u003eEur Respir Rev\u003c/em\u003e \u003cstrong\u003e33\u003c/strong\u003e (2024). https://doi.org/10.1183/16000617.0166-2023\u003c/li\u003e\n\u003cli\u003eRouen, A.\u003cem\u003e et al.\u003c/em\u003e Polysomnographic parameters in long-COVID chronic insomnia patients. \u003cem\u003eDialogues Clin Neurosci\u003c/em\u003e \u003cstrong\u003e25\u003c/strong\u003e, 43-49 (2023). https://doi.org/10.1080/19585969.2023.2222714\u003c/li\u003e\n\u003cli\u003eSramek, J.\u003cem\u003e et al.\u003c/em\u003e The effect of naltrexone on sleep parameters in healthy male volunteers. \u003cem\u003eJ Clin Psychopharmacol\u003c/em\u003e \u003cstrong\u003e34\u003c/strong\u003e, 167-168 (2014). https://doi.org/10.1097/JCP.0b013e3182a607ff\u003c/li\u003e\n\u003cli\u003eJin, J.\u003cem\u003e et al.\u003c/em\u003e Association of rapid eye movement sleep latency with multimodal biomarkers of Alzheimer\u0026apos;s disease. \u003cem\u003eAlzheimers Dement\u003c/em\u003e \u003cstrong\u003e21\u003c/strong\u003e, e14495 (2025). https://doi.org/10.1002/alz.14495\u003c/li\u003e\n\u003cli\u003eHorwitz, L. I.\u003cem\u003e et al.\u003c/em\u003e Researching COVID to Enhance Recovery (RECOVER) adult study protocol: Rationale, objectives, and design. \u003cem\u003ePLoS One\u003c/em\u003e \u003cstrong\u003e18\u003c/strong\u003e, e0286297 (2023). https://doi.org/10.1371/journal.pone.0286297\u003c/li\u003e\n\u003cli\u003evon Gall, C., Holub, L., Ali, A. A. H. \u0026amp; Eickhoff, S. Timing of Deep and REM Sleep Based on Fitbit Sleep Staging in Young Healthy Adults under Real-Life Conditions. \u003cem\u003eBrain Sci\u003c/em\u003e \u003cstrong\u003e14\u003c/strong\u003e (2024). https://doi.org/10.3390/brainsci14030260\u003c/li\u003e\n\u003cli\u003eLittle, T., Lindenberger, U. \u0026amp; Nesselroade, J. R. On selecting indicators for multivariate measurement and modeling with latent variables. \u003cem\u003ePsychological Methods\u003c/em\u003e \u003cstrong\u003e4\u003c/strong\u003e (1999).\u003c/li\u003e\n\u003cli\u003eDesjardins, C. D. \u0026amp; Bulut, O. profileR: An R package for profile analysis. \u003cem\u003eJournal of Open Source Software\u003c/em\u003e \u003cstrong\u003e5\u003c/strong\u003e, 1941 (2020). https://doi.org/https://joss.theoj.org/papers/10.21105/joss.01941\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"nature-portfolio","isNatureJournal":true,"hasQc":false,"allowDirectSubmit":false,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"","title":"Nature Portfolio","twitterHandle":"","acdcEnabled":false,"dfaEnabled":false,"editorialSystem":"ejp","reportingPortfolio":"","inReviewEnabled":true,"inReviewRevisionsEnabled":false},"keywords":"PASC, Long COVID, Wearables, Sleep, Digital Health, Heart Rate Variability","lastPublishedDoi":"10.21203/rs.3.rs-7422764/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-7422764/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eWearables yield a wide array of sleep-related measures that are relevant to Long COVID. \u0026nbsp;We leveraged wearables-derived sleep measures (WDSM) to identify differences between individuals with Long COVID (LC) versus individuals with possible or no LC in the RECOVER adult cohort. We found significant associations between LC and reduced heart rate variability measured during sleep and increased nightly variability in sleep duration after adjusting for confounders. Moreover, LC was independently associated with lower sleep efficiency, greater variability of nighttime sleep timing, higher resting heart rate, lower respiratory rate during rapid eye movement (REM) sleep, prolonged REM sleep onset latency, worse global physical and mental health. Cluster analysis identified distinct multidimensional patterns of WDSM that are associated with LC and quality of life. Together, the strong association between WDSM, or WDSM clusters, with LC provides a potential biomarker for future validation efforts to detect LC and monitor treatment effectiveness.\u003c/p\u003e","manuscriptTitle":"Wearable-derived Sleep Measurements are Associated with Long-COVID in the RECOVER Adult Cohort","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-09-03 11:45:51","doi":"10.21203/rs.3.rs-7422764/v1","editorialEvents":[],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"nature-communications","isNatureJournal":true,"hasQc":false,"allowDirectSubmit":false,"externalIdentity":"NCOMMS","sideBox":"Learn more about [Nature Communications](http://www.nature.com/ncomms/)","snPcode":"","submissionUrl":"https://mts-ncomms.nature.com/","title":"Nature Communications","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"ejp","reportingPortfolio":"Nature Communications","inReviewEnabled":true,"inReviewRevisionsEnabled":false}}],"origin":"","ownerIdentity":"f4bfa482-c310-4914-84ae-4c9312730f29","owner":[],"postedDate":"September 3rd, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[{"id":53671352,"name":"Health sciences/Diseases"},{"id":53671353,"name":"Health sciences/Diseases/Infectious diseases"}],"tags":[],"updatedAt":"2025-09-03T11:45:51+00:00","versionOfRecord":[],"versionCreatedAt":"2025-09-03 11:45:51","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-7422764","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-7422764","identity":"rs-7422764","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}