HEART rate variability biofeedback breathing programme for LOng Covid dysautonomia (HEARTLOC): results of a feasibility study

Introduction Post -covid- 19 syndrome, or Long covid ( LC) refers to symptoms persi sting 12 weeks after Covid-19 infection. LC comprises a wi de range of dysautonomia symptoms, including fatigue, breathlessness, palpitations, dizziness, pain and brain fog. Thi s study tested the feasibili ty and estimated the efficacy, of a Heart Rate Variability Biofeedback (HRV-B) technique via a standardised slow diaphragmat ic breat hing programme in individuals with LC.

and Analysis LC patients underwent a 4-week HRV-B intervention fo r 10 minutes twice daily for a total of 4 weeks using the Polar H10 ECG (Elect rocardiogram) chest strap and Elite H RV phone application. Outcome measures C19- YRSm ( Yorkshire Rehabilitation Scale modified), EQ5D- 5L (Eur oQol 5 Dimensions) , Composit e Autonomic Symptom Score (COMPASS- 31), W HO Disability Assessment Schedule (WHODAS) , and Root Mean Square of Successive Dif ferences between heartbeat s ( RMSSD) using a Fit bit device were completed before and after t he interven tion. The study was pre-registered at clinicaltrials.gov N CT0 5 2 286 65 . Resul ts 13 participants (54% female, 46% mal e) completed th e study with high levels of data complet eness and adherence. There was a stat ist ically significant improvement in C19YRS-m ( p=0.001), EQ5D Global Health Score (p=0.009) , CO M PASS-31 (p=0.007) , RMSSD (p=0.047) and EHODAS (p=0.02). Q ualitative f eedback suggested participants were able to use it independently, were sati sf ied with th e interv ention, and reported beneficial effects from th e interven tion.

This i s the first study in the litera ture to report that HRV-B is a feasible inte rvention for LC and seems to be potentially improvin g symptoms of LC and dysautonomia. All rights reserved. No reuse allowed without permission. preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for thisthis version posted September 12, 2023. ; https://doi.org/10.1101/2023.09.09.23295208doi: medRxiv preprint 3 Ethics and Dissemination The study received ethical approval from Health Research Authority ( HR A) Leicester South Research Et hics Commit tee (21/EM/0271).

Post -covid- 19 condition, post-covid-19 syndrome, dysautonomia, aut onom ic dysfunct ion, HRV biofeedback, sympathetic, parasympathetic, rehabilitation, technology Article summary Strengths and limitations of the study • To our knowledge, this is the first study of HRV- B in long covid and dysautonomia and has shown the feasibility of a novel technology- based intervention in a ho me setting. • There was a statisticall y signif icant improvement in LC symptoms, functional ability, quality of life and dysautonomia scor es. • The stud y provides an estimation of efficacy which w ill determine th e sample size for a larger controlled trial in LC and dysautonomia. • The main limitation of this study i s the small sample size and uncontrolled design which might not give us an accurate estimate of eff icacy. • The sample was predominantly Caucasian participants, 54% of w hom were female. • The take- up for HRV -B in those with a lack of experience in using digital technology and t hose f rom less privileged backgrounds is unknown.

Post -covid- 19 syndrome or Long covid ( LC) refers to persistent sympt oms 12 weeks af ter SARS-COV2 infection and includes symptoms of physi cal f atigue, cognitive f atigue or “brain f og”, breathlessness, pain and psychological distress. 1 2 An estimated 1.9 m illion p eople are reported to be affected by LC in the UK alone. 3 A GP at an average sized practice in the UK can expect to have 65 patients with Long covid. 4 The condition can be highly debilitating f or some, particularly middle-aged individuals who were previously functioning at a high level All rights reserved. No reuse allowed without permission. preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for thisthis version posted September 12, 2023. ; https://doi.org/10.1101/2023.09.09.23295208doi: medRxiv preprint 4 and in demanding vocational ro les. 5 Many will experience signif icant disruption to employment, social and caregiving roles, and participation in society. Many LC symptoms such as palpitations, di zziness, fatigue, pain and breathlessness can be explained by dysf unction of the Autonomic Nervous System ( ANS) or dysaut onomia. 6-10 Estimates of prevalence of dysautonomia in LC range from 2.5 to 67%. 8 10 11 Usually the ANS can maint ain a finely tuned stat e of homeostasi s or mount an appropriate stress response wher e necessar y. 10 However, in dysautonomia, there is episodic dysregulation in the ANS, typically with sympathetic overdrive. Dysautonomia also plays a si gnif icant role in the symptomology of many long-term conditions including mult iple sclerosi s, Parkinson’ s disease, diabet es mellitus, fibromyalgia, chronic fat igue syndrome and migraine. 12 One way of est imating and measuring autonomic f unct ion is through heart rate variability ( HRV), as cardiac rate and rhythm are controlled largely by the ANS. A low HRV is associated with sympathetic nervous system activation, also described as a state of ‘f ight or flight’. Higher HRV corresponds with parasympathet ic nervous syst em act ivat ion and is believed to reflect a state of rest and recovery. Lower HRV has been observed to be associated with f atigue and pain sympt oms of chronic fatigue syndrome/myal gic encephalomyelitis ( ME/CFS) and f ibromyalgia 13-15 , as well as other chronic physical and mental health pathologies including asthma, anxiety and st ress. 13-17 When physiological parameters such as HRV are monito red in real-time with self-regulatio n t echniques such as breat hing techniques t o increase parasympat hetic activity through vagus nerve activation to in fluence the parameters, this is known as biofeedback. 18 19 To the best of our knowledge, there have not yet been any studies of HRV-B in LC. How ever, HRV- B using breathing techniques has been test ed in other chronic condit ions such as asthma 16 , depression 20 , fibromya lgia 15 and post -traumatic st ress 21 and a 2021 systematic review summarises these comprehensively. 13 The optimal breath ing frequency to produce maximal increase in HRV varies for each individual but on average is between 5.5 and 6 breat hs per minute and is known as resonant breat hing. 16 22 23 Resonant breathing helps to restore autonomic balance due to increased baroreflex gain and vagal activation. 16 22-24 All rights reserved. No reuse allowed without permission. preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for thisthis version posted September 12, 2023. ; https://doi.org/10.1101/2023.09.09.23295208doi: medRxiv preprint 5 The aim of this study is to determine the feasibility and impact of a structured HRV-B regime incorporat ing diaphragmatic breathing exerci se, on LC and dysautonomia symptoms. Aims and objectives The aim of this study was: To assess the feasibility of a 4- w eek HRV-B structured breathing programm e in individuals with LC. The objectives include: 1. Do breat hing exercises t hrough HRV-B increase HRV amongst part icipant s with LC? 2. Are consumer grade monitors approp riate technology to use for HRV-B in the domiciliary setting? 3. Does regular HRV-B have any effect on LC and dysaut onomia symptoms?

Study design This was a phase 2 uncontrolled open-label feasibility study of a home technology- based HRV- B in 15 individuals with LC. Participants were identif ied through the L eeds covid-19 Rehabili tation Service, based at Leeds Community Healthcare NHS Trust. The total study period was 6 weeks f or each participant. The inclusion criteria were: age ≥ 18 year s, conf irmed LC diagnosis as per the NI CE criteria for post -covid syndrome 1 , self-rating of at least ‘moderate’ or ’severe’ on dysautonomia questions of palp itations or dizziness on the C19-YRSm 25 ; and abnormal NASA Lean Test ( N LT) 26-28 (HR increase of 30bpm or ≥120bpm, OR, BP decrease of 20mmHg systolic or 10 mmHg diastolic in the first 3 minutes of standing). NLT i s an accepted measure of cardiovascular instability and is conducted at initial assessment clinic for all LC service users in the Leeds covid reha bilita tion se rvice. The patient lies down for 2 to 5 minu tes prior to the test with HR and BP taken each minute to calculate average supine values. They then stand with heels 6 inches from a wall and lean back against it w ith HR and BP taken each minute for 10 min. Ab normal results (as described above) are All rights reserved. No reuse allowed without permission. preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for thisthis version posted September 12, 2023. ; https://doi.org/10.1101/2023.09.09.23295208doi: medRxiv preprint 6 demonstrated t hrough orthostat ic hypot ension or tachycardia on standing which are hallmarks of dysautonomia and therefore objectively quantifiable. Exclusion criteria were inability to use the wearable o r smartphone app technology, cognitive dif ficulties or mental health disorders causing inability to consent, cardiac arrhythmia, unst able respiratory disease (except asthma management ) Equipm ent and intervention To collect medium- term HRV data, participants wore a Fitbit Charge 5 smartwatch for a total of 6 w eeks. The HRV-B element was conducted using a Polar H10 chest strap for 10 minutes twice daily for a total of 4 weeks (Figure 1). This connected via Bluetoo th to the Elite HRV smartphone app which was downloaded t o participants’ phones. Participants aimed to increase their HRV score as displayed in Elite HRV in real time using a diaphragmatic breathing technique ( as explained in the protocol paper 29 ). Omron M2 blo od pressure monitor (endorsed by the British Hypertension Society) was used to conduct NLT in clinic and t he adapted Autonomic Profile ( aAP) 30 . Insert Fig 1 about here Participants were issued with a Fitbit Charge 5 device and smartphone application. They wore the Fitbit all the time during the 6 week period. This device collected nightly HRV data along with other m easures of sleep. They al so received a Polar H10 chest strap which was connected to the Elite HRV smartphone app f or the HRV-B element o f the intervention. This started after an initial 1 week baseline period in which HRV data was collected only by Fitbit to allow a baseline o f HRV data for comparison to post-interven tion. The breathing intervention lasted for 10 minutes tw ice daily for a total of 6 weeks. Measures Th e primar y o utcome measur e w as t he C19YRSm, a self-reported pat ient-report ed outcome measure to assess LC symptom severity, functional disability, and overall h ealth status. Secon d ary outcome m ea s ur es All rights reserved. No reuse allowed without permission. preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for thisthis version posted September 12, 2023. ; https://doi.org/10.1101/2023.09.09.23295208doi: medRxiv preprint 7 Fitbit Data measures: These included root mean square of successive differences between heartbeats (RMSSD), sleeping resting heart rate, non- rapid eye movement sleep heart rate (NREM HR), overall sleep score, composition score, revitalisation score, sleep duration score, deep sleep score Patient Reported Outcome Measures: • Composite Aut onomic Symptom Scor e (COMPASS 31) • World Health Organisat ion Disability Assessment Schedule ( W HO DAS) • EQ5D health related quality of l if e assessment (EQ-5D-5L) • NASA Lean Test ( NLT) – heart rate an d blood pressure data • adapted Autonomic Profile (aAP) – heart rate and b lood pressure data A summary of the schedule for the co mpletion of ou tcome measures is sho w n in Table 1. Table 1. Out come measures summary schedule I n i t i a l assessment Clinic Pre HRV-B phase (1 week) HRV-B phase (4 weeks) Post HRV-B phase (1 week) A utonomic screening (NLT) √ √ A utonomic function (COM PASS 31 ) √ √ H ome autonomic test (aAP) √ √ Fitbit wri st strap H RV, sleep data √ √ daily √ LC specific PROM C19-YRSm √ √ weekly √ WH O D A S √ √ Q uality of life (E Q5D - 5L) √ √ Anal ysi s Data from questionnaires were scored using standardised procedures f or each measure. Data f rom th e Fitbit was downloaded securely, cleaned and analysed using Python statistical platform. 2 participants were excluded f rom analysis, therefore a total of 1 3 individual All rights reserved. No reuse allowed without permission. preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for thisthis version posted September 12, 2023. ; https://doi.org/10.1101/2023.09.09.23295208doi: medRxiv preprint 8 datasets were analysed. Participant 6 was not able to complete the f ull 6 week study period due to unrelated health issues. Partici pant 12 was unw ell during the final assessment w ith an acut e respiratory infect ion. This was a cause extraneous t o the study process and encompassed by our exclusion criteria therefore th is participant was removed from data analysis. To explore the impact of the intervention, pre and post measures were compared on a within-participant basis using Wilcoxon signed rank tests. The pre-HRV phase was the baseline pre int ervention period. The post-HRV-B phase data was used for post -intervention ef fect. Fitbit data and Elite HRV data for HRV was analysed during the HRV-B intervention phase and compared t o pre-HRV - B data .A formal sample size cal culation was not required f or this feasibility study as it did not mimic a definitive randomised trial and aim was not to measure effect size. 31

15 participants were enrolled in the study, out of which 13 completed the study and their data was included for the final analysi s [ Table 2] . Table 3 shows the results of the N LT for the 13 participants. There was a 92% compliance rate with th e breath ing intervention as a total of 720 sessions were completed over the 4-week period. Table 2. Participant demographics No Age range Sex Ethnicity LC Duration (months) Comorbidities Medication Predominant LC symptoms Study completion 1 4 0 -5 0 M W hi t e 1 3 A n x i e t y D e pr e s s i on Amitri p ty line Code i ne S u m atri p ta n L o ra tad ine F atigue C om plete d 2 5 0- 6 0 F W h ite 1 5 Oste oarthri tis HRT Na p roxe n Brea thl e ssne ss Fa t i gu e He adache Co m p l e t e d 3 4 0 -5 0 M W hi t e 1 0 G as t r o e s o p h a ge al refl ux Om eprazole F atigue Brea thl e ssne ss Co m p l e t e d 4 4 0 -5 0 F M i x ed 2 0 A d e n o my o s i s Ha yf eve r P a i n Fa t i gu e Brea thl e ssne ss Co m p l e t e d 5 5 0 -6 0 M W hi t e 1 4 G as t r o e s o p h a ge al refl ux Asthma S kin d i sorder L ansopraz ole In h a l e rs Li mo cy c l in e Brea thl e ssne ss C om plete d 6 6 0- 7 0 F W h ite 1 2 Oste oarthri tis Nil F atigue Wi t h dr e w All rights reserved. No reuse allowed without permission. preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for thisthis version posted September 12, 2023. ; https://doi.org/10.1101/2023.09.09.23295208doi: medRxiv preprint 9 Table 3. Results of baseline NASA Lean test Par ti cipant no. Signi ficant postural hear t r at e r is e ? He ar t ra t e diff er ence (beat s per minut e) Sign ifican t pos tural bl ood pressure change? Blood pres sure diff er ence (mmHg ) 1 Y +36 N +27 systolic +26 di astoli c 2 Y +37 N -11 systol ic +7 dia s tolic 3 Y +36 Y -8 sys t oli c +3 dia s tolic 4 Y +38 Y -40 systolic -8 diastol ic 5 Y +35 Y -28 systolic -18 diastolic 7 Y +56 N - 1 s y s t o l i c + 12 dias tol ic 8 Y +35 N - 2 s y s t o l i c +14 di astoli c 9 Y +34 Y -34 systolic +9 dia s tolic 10 Y +33 N - 5 s y s t o l i c 7 2 0 -3 0 F W hi t e 7 H ay f e v e r B i s o p r ol o l B r e at h l es s n e s s Pain Co m p l e t e d 8 3 0- 4 0 F W h ite 2 5 Anxiety/ d e p r e s sio n Cita lopram F atigue Brea thl e ssne ss Sl ee p dist u rba nce Co m p l e t e d 9 3 0 -4 0 M W hi t e 2 5 A n x i e t y /d ep r es s i o n Cita lopram F atigue C om plete d 10 4 0- 5 0 M W h ite B rit ish 1 5 Nil Nil F atigue C om plete d 11 1 0- 2 0 F W h ite 1 2 Nil Ora l co ntrace p ti o n F atigue Na u se a Brea thl e ssne ss Co m p l e t e d 12 2 0- 3 0 F W h ite 2 0 Nil Amitri p ty line F atigue Exc l uded 13 4 0- 5 0 F 1 1 Crohn ’s d isea s e Migraine ADHD Aza thi op ri n e Amitri p ty line L i sdex am p heta mi n e Iva b ra dine Palp i ta tions Brea thl e ssne ss Co m p l e t e d 1 4 40 - 50 M Wh i t e 17 Asthma In h a l e rs Altere d tas te Fa t i gu e Altere d sm ell Co m p l e t e d 15 6 0- 7 0 F W h ite 7 Hy p ot h y ro id i sm H o dgk i ns Oste op oro s i s Iva b ra dine Th yr o xi n e Anastra zole A l e n d r oni c aci d Aspirin Fa t i gu e Brea thl e ssne ss Cough Co m p l e t e d All rights reserved. No reuse allowed without permission. preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for thisthis version posted September 12, 2023. ; https://doi.org/10.1101/2023.09.09.23295208doi: medRxiv preprint 10 +7 dia s tolic 11 Y +47 N U n a v a i l a b l e 13 Y +31 N - 4 s y s t o l i c +9 dia s tolic 14 N +26 Y -21 systolic -14 diastolic 15 Y +37 Y +5 systolic -19 diastolic St atist ically significant improvements were noted in Long Covid specif ic out come measure C19YRS-m for symptom severity, functional ability, overall h ealth score as well as total “other” symptoms, CO MPASS31, and EQ 5D-5L. The effect size for C19-YRSm was large. Table 4. Change and eff ect size for the pat ient-report ed outcome measure scores M e a n ( S D ) a t baseline pre HRV-B phase Mean SD at post HRV-B phase Change (compared to baseline) P value for paired pre/port diff er ence Effect Size (Cohen’s d) C19YRS- m symptom seve ri ty 19.54(4.37) 13.46(6.54) -5 0.0010 1.09316 C19YRS- m functional abi lity 7.615(3.885) 5.231(4.902) -2 0.0015 0.539022 C19YRS- m global health score 3.923(1.553) 5.007(1.801) +1 0.0137 0.644633 C19YRS- m total “other” LC symptoms 6.462(4.502) 4.000(4.690) -2 0.0059 0.535571 CO MPASS-31 38.55(18.10) 32.49(19.59) -7.6 0.0078 0.047074 EQ 5D-5L global health score 46.31(14.07) 56.69(19.86) +10 0.0083 0.007391 WH O D A S disabi lity scale 36.26(18.41) 29.38(20.22) -6.9 0.0215 2.029452 C 19 Y R S S ympto m s S everity – ran ge fro m 0- 3 0 , hi gher score = g re a ter se v e rity C 19 Y R S fun c tio n a l di s a b i l ity – range fro m 0-15, h i g h e r score = g re a ter d isa b ility C 1 9Y RS T ot a l ‘ o t h e r ’ s y m pt o m s – t o t a l of 4 0 ot h e r m i s c e ll a ne ou s s y m p t om s po s s i bl e C 19 yrs g lo b a l h e a lt h sc ore – ran ge fro m 0 - 10 - , higher score = greate r to ta l heal th p e rceptio n COM P ASS - 31 – range from 1 – 1 00, greater score = hi gh e r a u to n o m i c sy mpto m bu rden E Q5 D-L global health score – greater score = hi gh e r g l o bal h e a lth perce p tion There was a signif icant dif ference in RMSSD, a HRV measure ref lect ing parasympat hetic activity 13 , between preHRV-B and post HRV-B phases (Table 5) . All rights reserved. No reuse allowed without permission. preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for thisthis version posted September 12, 2023. ; https://doi.org/10.1101/2023.09.09.23295208doi: medRxiv preprint 11 Table 5. HRV data fro m Fitbit device M e a n ( S D ) a t baseline pre HRV-B phase Mean SD at post HRV-B phase P val ue for paired difference between post HRV - B and baseline Effect Size (Cohen’s d) Fitbit RM SSD Summaries 34.19 (19.65) 40.94 (29.36) 0.0479* 0.270202 Fitbit Sleeping Resting HR 67.84 (6.555) 67.70 (5.932) 0.8984 0.022395 Fitbit NREM HR 66.59 (7.606) 66.15 (7.124) 0.5417 0.05971 Fitbit O verall Sleep Score 76.78 (4.756) 76.32 (5.083) 0.8477 0.093454 Fitbit Composi tion Score 20.08 (1.288) 19.70 (1.050) 0.4143 0.323393 Fitbit Revitalisation Score 19.16 (1.878) 18.75 (2.559) 0.3394 0.182671 Fitbit Sleep D uration Score 37.53 (3.266) 37.87 (2.932) 0.6848 0.109554 Fitbit D eep Sleep Score 72.10 (22.14) 69.20 (16.64) 0.1909 0.14808 Ro ot Mea n Squar e of S ucc e s s i v e Di ffer enc e s bet we e n heart beats ( RM S SD) N o n- Ra p i d Eye M o ve m e nt H eart R a t e ( NR EM H R) Participants were asked how they felt about the technology and its eff ect on their LC symptoms. Some of the patient quotes are summarised in Table 6. Table 6. Participant feedback upon st udy completion Participant No. Qualitative Feedback Highlights 1 “S le e p is im pro ved ” “I’ m E n j o y i ng th e exerc is e , it ’ s e a sier a nd mor e i nst inct i v e ov e r ti m e ” 2 “Wou l d c ontin ue d o in g the br ea th i ng ex erc i s e s ” “F eel mo r e r el ax e d” “Hel p ed me g e t to sl e e p at ni g ht” “Wou l d do t h e e x e r c ise s a t w o r k wh e n m y heart was g o in g” “He lp s to c ont ro l my p a nic” 3 “T ech no l o g y was u ser friendly” “I m p rov es s l ee p ” 4 “Huge im pro ve m e nt i n fat i g ue le vels , no t g ot the s a me lev e l s of f a t igu e a ny m o r e ” “No l on g er cr as hin g ” “Able to st a rt tr y i ng t o go to w o r k ” “T h e b r ea thing ex erc i s e s m a k e me feel li k e I c a n c op e w ith my si tu a ti on an d gi ve m e c on fi de nc e in addre ss ing m y f ati gue” 5 “Hel p s w he n I get palpi ta t i o ns or if I’ m s tr u g g l ing t o sle e p” “I do the breath ing e x e rc is es wh e n I’m g e tti n g s tre s se d ” “I do n’ t w ak e u p s tre ss ed as mu c h an ym o r e ” . All rights reserved. No reuse allowed without permission. preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for thisthis version posted September 12, 2023. ; https://doi.org/10.1101/2023.09.09.23295208doi: medRxiv preprint 12 “ I w il l carr y on doi ng th e breat hing ex ercis es, even if I get b et te r , I w ill al wa ys do t he m ” 7 “E q uipm ent w a s s tr ai g htf orw ard t o u se” “I w il l kee p do i ng t he b re at hi n g ex e rc is e s af te rwards” 8 “I feel a lot bett er, my fat i g ue is b e tt er, I ’m not f alling a sle e p as m uch, my s l e ep i ng p att er n s has i m pr o ved” “I’ m f eeling a l ot m or e pos iti ve, I’ m a bl e to enj oy wor k an d d o m y jo b again” “I feel li ke I' m bac k to w hat I w as p re co vid” 10 “I feel mas si vel y b e t ter, it h elps m e m anage m y tri gger of s tre ss” . “Physic ally a lo t bet ter now, I ca n thi nk abo ut do ing e x e r cis e ag a in” 11 “I do h a v e so m e more ener gy” “W h e n I d o a cti v ity , I am ab le to d o i t f o r a l on g e r pe r i od o f t im e ” 13 “It d i d hel p m e get t o sleep q uic ke r at n i gh t” 14 “ E q uipm ent w a s ea s y to us e ” “It definite l y he l ps a t n ight ti me, h el p e d m e s l e ep”

This study demonstrat ed that an HRV -B intervention using diaphragmat ic breathing, smartphone HRV application and an ECG chest strap was feasible for participants with LC in a home-based setting. This study represents the first use of HRV-B with diap hragmatic breathing modulation in a LC popu lation. The study adds weight to the notion that dysautonomia plays a critical role in LC symptom generation and interventions to target this may prove beneficial. It should be noted that unlike many potential LC interventions, diaphragmatic breathing is a non-pharmacological treatment, well tolerated and scalable with app-based technologies. The results of this study support a further controlled trial in LC to explore its actual ef ficacy on symptoms, particularly dysautonomia. Other st udies of HRV- B as an adjunct in chronic disease management have found similar positive results amongst other conditions. 13 In f ibro myalgia, Hassett et al d emonstrated significant improvement in fib romyalgia physi cal functioning scores w hich was evident at 3 months follow up (p -0.0022). 15 As demonstrated by the qualitative f eedback, participants were satisfied with the HRV- B practice which is comparable to other studies. 13 There are many indices for capturing HRV and most significant improvement was t he RMSSD met ric, reflecting greater parasympathetic activity. 13 This w as al so demonst rated by Shumann et al albeit in a different population of veterans with post-traumatic stress diso rder, although no p-values were pr ovided for compar ison. 21 All rights reserved. No reuse allowed without permission. preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for thisthis version posted September 12, 2023. ; https://doi.org/10.1101/2023.09.09.23295208doi: medRxiv preprint 13 All participants fed back that they appreciated the concept of the technology and particularly the N ASA Lean Test 27 and the adap ted Autonomic Profile (aAP) 30 assessment tools for dysautonomia. They f ound the tools easy t o use and empowering. O t her authors have not ed that testing for dysautonomia should be more accessible and holistic. 10 The aAP is bot h a holistic and comprehensive test assessing different triggers of dysautonomia including position, meal, time of day and exercise. Thi s i s the first clinical research study featuring its use and we would encourage its use i n future research along with the NASA lean test for monitoring the condition and response to interven tions. HRV data collected in th is study was diverse and rich. The Elite HRV data w as collected only during the 10 minute b reathing in tervention. The Fitbit collected data during sleep offered a reasonable longer-term measure of HRV. There was a significant improvement in RMSSD f rom th e Fitbit data, however th e effect size was not as pronounced as clinical outcome measures. The reason for this may be due to the fact that HRV is impacted by many factors including nutrition, sleep, exercise and menstrual periods in f emales. 32-34 We d id n ot contr ol f or these factors and future trials should attemp t to contro l fo r them to truly quantif y ef fect of HRV-B on LC and dysaut onomia sympt oms. The stud y had several limitations. It was a small open-label uncontro lled feasibility study, and we need to bear caution abou t the generalisability of study f indings. The study has how ever showed that the technology can be used in a home sett ing independently by individuals with LC. The changes in physiologi cal parameters did not match the clinical outcome measures in terms of signif icance values and eff ect sizes. Thi s can be explained by the W orld Health Organi sation’ s International Classification of Functioning Disability and Health (ICF) f ramework which su ggest s non-linear changes in dif ferent domains of the health condition ( body function, functional activities and quality of life). The intervention d uration could be argued to be short t o have a reasonable effect on autonomic imbalance and future research could explore the relationship between dose (duration) and response. Finally, we did not undertake a long follow-u p after stopping the use of technology. This needs to be explored in future la rger scale studies. All rights reserved. No reuse allowed without permission. preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for thisthis version posted September 12, 2023. ; https://doi.org/10.1101/2023.09.09.23295208doi: medRxiv preprint 14 In summary, a diaphragmatic breathing technique using HRV-B is feasible to be used independently in a home setting by individuals with LC and the interventio n seems to have a potential effect on impro ving LC symptoms, particularly those related to d ysautonomia. Ethics and dissemination The stud y received ethical approval from Health Research Authorit y ( HR A) Leicester Sout h Research Et hics Commit tee (21/EM/0271). Conflicts of interest Manoj Sivan is an advisor to the World Health O rganisation ( W HO ) for the long covid policy in Europe.

The authors would like to thank all study participants and the Leeds LC Advisory G roup (PAG ) for the ir invo lvemen t in a ll stages of this study. Patient and Public Involvement Members of the PAG provided input to analyse and develop the research aims, objectives, and questions, ensuring these align with the key research priorities of those with LC and dysautonomia. All advisory group members have lived experience of LC. The PAG met quarterly to review progress, ensure the research continues to answ er relevant issues and that find ings can inf orm LC care. Funding statement This research was support ed by IAA EPSRC grant [Ref 112538] w it h the Universit y of Leeds as t he sponsor organi sation and the Leeds Communit y Healthcare NHS Trust Covid Rehabili tation service as the research site organisation. Author contributions All rights reserved. No reuse allowed without permission. preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for thisthis version posted September 12, 2023. ; https://doi.org/10.1101/2023.09.09.23295208doi: medRxiv preprint 15 MS and AC conceptualised t he st udy. MS, AC and RJ O C were awarded EPSRC IAA pump- priming grant f or the f easibility study with MS as the Principal Investigator. All au thors contributed to the study design and obtained eth ical approval. JC and NI perf ormed the initial statistical anal y si s independently and verified the results. JC w rote an initia l draft o f the paper by adapting the grant proposal, the ethics protocol, and the protocol paper. A ll authors approved the final manuscript. MS is the corresponding author and guarantor. Figure legends Fig 1. Heart Rate Variab ility Bio feedb ack (HRV-B) using a breathing techniq ue and chest strap f or real t ime HRV monitoring. Polar H 10 picture from Wikimedia commons, reprint ed under CC BY-SA 3.0 license. EliteHRV screenshot from W ikimedia commons, reprinted under CC BY- SA 4.0 license.

1. Nati onal I nstitute for Health and Care Excell ence (NICE) Scotti sh Int ercol legia te G ui delines Network (SI GN) and Royal College of Gener al Practitioner s (RCG P). COVI D-19 r apid guidel ine: managing the long- term ef fec t s of COVI D-19. London: NI CE 2022. 2. Crook H, Raza S, Nowel l J, et al . Long covid—mechani sms , ri sk fact ors, and management . bmj 2021;374 3. Office of National Stati stics. Preval ence of ongoi ng symptoms foll owi ng corona virus ( COVI D-19) infecti on in the UK: 1 June 2022. London: O NS 2022. 4. Gre enhalgh T, Sivan M , D elaney B, et al. Long covid-an update f or primary care. BMJ 2022;378:e072117. doi : 10.1136/bmj -2022-072117 [publi shed Online Fir st : 2022/09/23] 5. Stat istics O o N. Coronavir us and the soci al impacts of ‘long COVI D’ on people’ s l i ves in Great Brit ain: 7 April to 13 June 2021. 2021; https: //www.ons .gov.uk/peoplepopulati onandcommuni ty/heal thandsoci alcare/c onditionsa n ddisease s/ar ti cles/coronavirus andthesoc ialimpact sofl ongcovidonpeopleslivesingr eatbritain /7 aprilto13june2021 . (access ed 02. 05.2022). 6. Dani M, Dirksen A, Taraborr elli P, et al. Autonomic dysfunc tion in ' long COVI D': rational e, physiology and management st rat egie s. Cl in Med (L o nd ) 2021;21(1): e63-e67. doi: 10.7861/c linmed.2020-0896 [publi shed Online Fir st: 2020/11 /28] All rights reserved. No reuse allowed without permission. preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for thisthis version posted September 12, 2023. ; https://doi.org/10.1101/2023.09.09.23295208doi: medRxiv preprint 16 7. Shouman K, Va nichkachorn G, Che shir e W P, et al . Aut onomi c dysfuncti on following COVI D-19 infecti on: an early experi ence. C lin A ut on R e s 2021;31(3):385-94. doi : 10.1007/s10286-021- 00803-8 [publi sh ed Online First : 2021/04/17] 8. Lars en NW, Stil es LE, Shaik R, et al . Charact erization of autonomic symptom burden in l ong COVID: A global survey of 2, 314 adul ts. Fr ont N eur o l 2022;13:1012668. doi : 10.3389/f neur .2022.1012668 [ publi shed Online Fir st: 2022/11/11] 9. If tekhar N, Sivan M . Venous insuffi ciency and acrocyanosi s in l ong COVID: dysautonomia . Lanc et 2023;402(10401): e9. doi : 10. 1016/S0140 -6736(23)01461-7 [publi shed Online Fi r st: 2023/08/13] 10. Carmona-Torre F, Minguez-Ol aondo A , Lopez -Bravo A, et al . Dysautonomi a in COVI D-19 Pati ents : A Narr ative Review on Clinical Cour se, Di agnostic and Therapeutic Stra tegies. Fro nt Ne urol 2022;13:886609. doi: 10.3389/ fneur. 2022.886609 [publi shed O nl ine Fi rst: 2022/06/21] 11. Is aac RO, Cor rado J, Sivan M. Detecti ng Orthos tatic I ntol erance in Long COVI D i n a Cl inic Set ting. Int J Envi r on Res Public Healt h 2023;20(10) doi: 10.3390/i jerph20105804 [ publi shed Online First: 2023/05/27 ] 12. Zalewski P, Sl omko J, Zawadka-Kunik owska M. Autonomi c dysfunction and chronic disease. Br Me d B u l l 2018;128 (1): 61-74. doi : 10.1093/bmb/ldy036 [ publi shed O nli ne Fi rs t: 2018/11/13] 13. Fourni e C, Chouchou F, Dalleau G, et al. Hea rt rate vari abili ty biofeedback in chronic di sea se management : A systema ti c rev iew. C o mpleme nt Ther Me d 2021;60: 102750. doi: 10.1016/j.ctim.2021.102750 [publi sh ed O nline Firs t: 2021/06/13] 14. Escorihuela RM, Capdevil a L, Castro J R, et al . Reduced hear t rate variabi lity pr edict s fatigue severity in i ndividual s with chronic fatigue syndrome/myal gic encephalomyeliti s . J Transl Me d 2020;18(1) :4. doi: 10 .1186/ s12967-019-02184-z [publi shed Online Fir st: 2020/01 /08] 15. Hasset t AL, Radvanski DC, Vaschil lo EG, et al . A pilot study of the ef ficacy of heart ra te vari abi lity (HRV) biofeedback in pati ent s with fibromyalgia. Appl Psyc hop hysiol Bi of eed b ack 2007;32(1) :1-10. doi: 10.1007/ s10484-006-9028-0 [publi shed O nli ne Fir st: 2007/01/16] 16. Lehrer PM, Vaschill o E, Vaschillo B, et al. Biof eedback treatment f or asthma. Ch est 2004;126(2): 352-61. doi : 10.1378/chest. 126.2.352 [publis hed Onli ne Fir st: 2004/08/11] 17. Goes sl VC, Cur ti s s JE, Hofmann SG. The ef fec t of hear t rate vari abili ty biof eedback trai ning on st ress and anxi ety: a meta- analysi s. P syc hol Med 2017;47 (15) :2578-86. doi: 10.1017/S0033291717001003 [publi shed Online Fir st: 2017/05/10] 18. Lehrer P, Vaschil lo B, Zucker T, et al. Pro tocol f or hea rt r ate vari abili ty biofeedback trai ning. Biofe edback 2013;41(3):98-109. doi: ht tps: //doi. org/10. 5298/1081-5937-41.3.08 All rights reserved. No reuse allowed without permission. preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for thisthis version posted September 12, 2023. ; https://doi.org/10.1101/2023.09.09.23295208doi: medRxiv preprint 17 19. Gevir tz R. The promi se of hear t rate variability biofeedback: Evidence-based applicati on. Biofe edback 2013;41(3):110-20. doi : ht tps: //doi. org/10. 5298/1081-5937-41.3.01 20. Karavidas MK, Lehrer PM, Vaschil lo E, et al. Preliminary resul ts o f an open label study of heart rate variabi lity biofeedback for the t reatment of major depr ession. App l P sy cho physiol Biofe edback 2007;32(1):19-30. doi: 10.1007/s10484-006 -9029-z [publi shed Online First: 2007/03/03] 21. Schuman D L, Kill ian MO. Pilot Study of a Singl e Session Hear t Rat e Variabili ty Bi ofeedback Intervention on Vete rans' Post tr aumatic Stress Symptoms. Ap pl Psych op hysiol Bi o feed ba ck 2019;44(1) :9-20. doi: 10.1007/ s10484-018-9415-3 [publi shed O nli ne Fir st: 2018/09/20] 22. Lehrer PM, Gevirt z R. H eart r at e variabil ity biof eedback: how and why does it work? F ron t P sy chol 2014;5:756. doi: 10.3389/ fpsyg.2014.00756 [publi shed Online Fir st: 2014/08/08 ] 23. Vaschi llo EG, Vaschill o B, Lehr er PM. Characteristic s of resonance in heart rate variability stimul ate d by bi ofeedback. Ap p l P s yc h oph y s io l B iof eed ba ck 2006;31(2):129 -42. doi: 10.1007/ s10484-006-9009-3 [publ i shed O nl ine Fi rs t: 2006/07/14] 24. Pagaduan JC, Chen YS, Fell J W, et al. Can Heart Rate Vari ability Bi ofeedback I mprove Athl etic Per formance? A Syst ematic Revi ew. J Hum Ki net 2020;73:103-14. doi : 10. 2478/ hukin-2020- 0004 [publ i shed O nl ine Firs t: 2020/08/11] 25. Si van M, Preston N, Parki n A, et al . The modifi ed COVID-19 Yorkshire Rehabili tation Scal e ( C19- YRSm) pati ent-repor ted outc ome mea sure for L ong Covid or Pos t -COVID-19 syndr ome. J Me d Vi ro l 2022 doi: 10.1002/jmv.27878 [publi shed O nli ne Fi r st: 2022/05/24] 26. Bungo MW , Charl es JB, Johnson PC, J r. Cardiovascular deconditi oni ng during space fli ght and the use of saline as a count ermeasure t o ort hostatic intolerance. Av i at Spac e En viro n Med 1985;56(10): 985-90. [publ i shed O nline Fi rs t: 1985/10/01] 27. Lee J, Vernon SD, Jeys P, et al . Hemodynamics during the 10-minute NASA Lean Test: evidence of circul atory decompensati on in a subset of ME/ CFS patients . J Transl Med 2020;18( 1):314. doi : 10.1186/ s12967-020-02481-y [publi shed Online F ir st : 2020/08 /18] 28. Hyatt KH, J acobson LB, Schneider VS. Compari son of 70 degrees til t, LBNP, and passive sta nding a s measrue s of orthostatic tol erance. Av iat Spac e En viron Me d 1975;46(6):801-8. [publi shed O nli ne Fir s t: 1975/06/01] 29. Cor rado J , Halpin S, Preston N, et al . H EART rate vari ability biofeedback for long COVID symptoms (HEARTLOC): protoc ol for a fea sibility study. BMJ Ope n 2022;12(11 ): e066044. doi: 10.1136/bmjopen-2022-066044 [publi shed O nli ne Fi r st: 2022/11/22] All rights reserved. No reuse allowed without permission. preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for thisthis version posted September 12, 2023. ; https://doi.org/10.1101/2023.09.09.23295208doi: medRxiv preprint 18 30. Si van M, Corrado J, Mathias C. The adapted A utonomic Profile (aAP) Home- Bas ed Test f or the Evaluation of Neuro-Cardiova scular Autonomic Dysfunction. Ad v Clin N e u ro sci Rehabi l 2022 doi: ht tps:/ /doi .org/10.47795 /QKBU 6715 31. Whit ehead AL, Sully BG , Campbell MJ. Pilot and feasi bili ty studies: i s there a differe nce f rom each other and f rom a randomi sed contr olled trial? Contemp Clin Trial s 2014;38(1) :130-3. doi : 10.1016/j.cct.2014 .04.001 [publi shed Online Fir st: 2014/04/17] 32. Young HA, Benton D. Heart -ra te vari abi lity: a biomarke r to st udy the influenc e of nut rition on physiologi cal and psychological health? B e hav Ph arma col 2018;29(2 and 3-Spec I s sue):140- 51. doi: 10.1097/FBP. 0000000000000383 [publi shed Online Fi rs t: 2018/03/16] 33. Sajjadieh A, Shahsavari A, Safaei A, et al. The Associ ati on of Sl eep Duration and Q uali ty with Heart Rate Vari abil ity and Bl ood Pres sur e. T a na ffos 2020;19(2):135-43. [publ i she d O nli ne First: 2020/12/03] 34. Sandercock GR, Broml ey PD, Brodi e DA. Eff ect s of exerci se on hear t rat e vari ability: inf er ences from meta- analy si s. Me d Sci Sports Exerc 2005;37(3):433-9. doi : 10.1249/01.ms s.0000155388.39002.9d [ publ ished O nline First : 2005/03/03] All rights reserved. No reuse allowed without permission. preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for thisthis version posted September 12, 2023. ; https://doi.org/10.1101/2023.09.09.23295208doi: medRxiv preprint All rights reserved. No reuse allowed without permission. preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for thisthis version posted September 12, 2023. ; https://doi.org/10.1101/2023.09.09.23295208doi: medRxiv preprint