Association of Albuminuria with Preserved eGFR and Mortality in Pooled NHANES 2007–2018: Conventional and Machine Learning Survival Analyses

Association of Albuminuria with Preserved eGFR and Mortality in Pooled NHANES 2007–2018: Conventional and Machine Learning Survival Analyses | 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 Research Article Association of Albuminuria with Preserved eGFR and Mortality in Pooled NHANES 2007–2018: Conventional and Machine Learning Survival Analyses Sai Tharun Reddy Gopannagari, Kesava Manikanta Achuta, Divya Ganisetti This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-9420411/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted 6 You are reading this latest preprint version Abstract Background Albuminuria is an established marker of kidney damage and cardiovascular risk, but its prognostic significance in individuals with preserved estimated glomerular filtration rate (eGFR) remains incompletely characterized. We evaluated the association of albuminuria with preserved eGFR and all-cause mortality in pooled National Health and Nutrition Examination Survey (NHANES) 2007–2018 data and compared conventional and machine learning survival models. Methods We performed a retrospective cohort study of adults in the pooled NHANES 2007–2018 with linked mortality follow-up. Albuminuria was defined as a urinary albumin-creatinine ratio at least 30 mg/g, and preserved kidney function as eGFR ≥ 60 mL/min/1.73 m². Cox proportional hazards models evaluated the association between albuminuria with preserved eGFR and all-cause mortality. Sensitivity analyses included age-group and survey-weighted Cox models. Secondary analyses compared standard Cox, least absolute shrinkage and selection operator (LASSO) Cox, Elastic Net Cox, and Random Survival Forest models. Results The final cohort included 32,290 participants, including 3,090 with albuminuria and preserved eGFR; 2,768 deaths occurred during follow-up. In the fully adjusted pooled model, albuminuria with preserved eGFR was independently associated with higher all-cause mortality (hazard ratio, 1.57; 95% CI, 1.42–1.73; P < .001). Findings were consistent in age-group and survey-weighted sensitivity analyses. In secondary analyses, Random Survival Forest showed the best predictive performance, but gains over standard and penalized Cox models were modest. Conclusions Albuminuria, despite preserved eGFR, was independently associated with higher all-cause mortality in pooled NHANES 2007–2018. Although Random Survival Forest performed best, conventional Cox-based models remained informative for mortality risk stratification. Albuminuria Estimated glomerular filtration rate Mortality NHANES Machine learning Figures Figure 1 Figure 2 Figure 3 Figure 4 Introduction Albuminuria is an important marker of kidney damage and a broader indicator of cardiorenal risk. Current Kidney Disease: Improving Global Outcomes (KDIGO) guidance emphasizes that kidney disease risk should be assessed using both estimated glomerular filtration rate (eGFR) and albuminuria, as these measures capture different but complementary aspects of kidney health[ 1 ]. Prior studies have also shown that albuminuria is associated with adverse outcomes, including all-cause and cardiovascular mortality, independent of eGFR[ 2 , 3 ]. Together, these observations suggest that clinically meaningful kidney and vascular injury may already be present even when filtration function appears preserved. Despite its clinical importance, albuminuria remains underdetected in routine practice, particularly outside recognized chronic kidney disease (CKD) populations[ 4 ]. Recent United States (US) data suggest that albuminuria testing is performed in only a minority of at-risk adults, raising concern that clinically relevant kidney disease may go unrecognized when assessment relies mainly on serum creatinine or eGFR[ 5 – 7 ]. This gap in recognition is especially relevant in population-based studies, where many individuals may have abnormal albumin excretion despite preserved filtration function. In this setting, better characterization of the mortality implications of albuminuria with preserved eGFR could help refine risk identification beyond traditional creatinine-based assessment alone. Against this background, population-level data are needed to better define the prognostic significance of albuminuria in adults without reduced eGFR. The National Health and Nutrition Examination Survey (NHANES), with linked mortality follow-up, provides a useful setting for this question because it combines a nationally representative sample of US adults with standardized laboratory data and longitudinal outcomes[ 8 , 9 ]. It also offers an opportunity to examine whether machine learning–based survival models provide meaningful predictive benefit beyond conventional Cox analysis. In this study, we evaluated the association of albuminuria with preserved eGFR and all-cause mortality in pooled NHANES 2007–2018 data linked to mortality follow-up. We also assessed the consistency of this association in sensitivity analyses and compared the predictive performance of conventional Cox-based survival modeling with machine learning–based survival approaches. Methods Study Design and Data Source We performed a retrospective cohort study using pooled NHANES 2007–2018 data with linked mortality follow-up. NHANES is a nationally representative survey of the noninstitutionalized US population that integrates interview, examination, and laboratory data. Mortality follow-up data were obtained from the NHANES Linked Mortality Files. This study used publicly available, de-identified NHANES data with linked mortality follow-up; therefore, additional institutional review board approval and informed consent were not required for this secondary analysis. Study Population Participants aged 18 years or older were eligible if urinary albumin, urine creatinine, and serum creatinine data were available. Participants were excluded if mortality linkage eligibility, mortality status, follow-up time, or key demographic variables were missing. After pooling all cycles, the final analytic cohort included 32,290 participants. Exposure Definition The primary exposure was albuminuria with preserved kidney function. UACR was obtained directly when available and was calculated manually in 2007–2008 from urine albumin and urine creatinine. Albuminuria was defined as UACR at least 30 mg/g. Kidney function was estimated using the 2021 Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) creatinine equation, and preserved kidney function was defined as eGFR ≥ 60 mL/min/1.73 m². The exposure group, therefore, comprised participants with both albuminuria and preserved eGFR; all others served as the comparison group. Covariates Covariates were selected a priori based on clinical relevance and availability across cycles and included age, sex, race/ethnicity, body mass index, hypertension, diabetes, smoking status, hemoglobin A1c, and NHANES cycle. Because race/ethnicity coding differed across cycles, a harmonized pooled variable was created: Mexican American, Other Hispanic, Non-Hispanic White, Non-Hispanic Black, and other race/ethnicity. Hypertension and diabetes were derived from questionnaire variables, and smoking status was categorized as current/some days, former, or never. Outcome Outcome The primary outcome was all-cause mortality. Mortality status and follow-up time in months were obtained from the NHANES Linked Mortality Files. Follow-up time was defined using person-months from examination, and death status was defined from linked mortality indicators. Survival time was analyzed from examination to death or censoring. Data Management and Harmonization Cycle-specific demographic, laboratory, examination, and questionnaire files were merged by participant identifier (SEQN), harmonized across cycles, and pooled into a single 2007–2018 analytic dataset, including race/ethnicity harmonization and cycle-specific handling of 2007–2008 UACR. Statistical Analysis Baseline characteristics were summarized by exposure group as mean (SD) or n (%). Group comparisons used Wilcoxon rank-sum tests for continuous variables and Pearson chi-square tests for categorical variables. Overall survival was evaluated using Kaplan–Meier methods. Associations between albuminuria with preserved eGFR and all-cause mortality were examined using Cox proportional hazards regression with 3 models: unadjusted; adjusted for age, sex, pooled race/ethnicity, and NHANES cycle; and additionally adjusted for body mass index, hypertension, diabetes, and smoking status. Hazard ratios with 95% confidence intervals were reported. Sensitivity Analyses and Assessment of Proportional Hazards Proportional hazards assumptions were assessed using Schoenfeld residual testing with cox.zph, focusing on the primary exposure. Because adjusted pooled models showed mild global nonproportionality driven mainly by age, a sensitivity analysis replaced continuous age with categories of 18–39, 40–59, 60–74, and ≥ 75 years. Survey-weighted Cox models were also fit using pooled 12-year Mobile Examination Center (MEC) examination weights, masked strata, and primary sampling units with Taylor linearized variance estimation; the same sequence of unadjusted, demographic- and cycle-adjusted, and fully adjusted models was applied. Machine Learning Survival Modeling As a secondary predictive analysis, survival machine learning models were developed in the pooled complete-case dataset (n = 31,060; deaths = 2,661), split into training (70%; n = 21,742) and held-out test (30%; n = 9,318) sets. Candidate predictors included age, sex, pooled race/ethnicity, body mass index, hypertension, diabetes, smoking status, hemoglobin A1c, eGFR, log-transformed UACR, albuminuria with preserved eGFR phenotype, and NHANES cycle. A small positive offset was applied to zero follow-up times to permit penalized Cox modeling. Evaluated models were standard Cox, least absolute shrinkage and selection operator (LASSO) Cox, Elastic Net Cox, and Random Survival Forest. Penalized Cox models used cross-validation to select tuning parameters, Random Survival Forest used 500 trees, and model performance was assessed on the held-out test set using Harrell’s C-index, Brier scores at 60 and 120 months, and integrated Brier score through 120 months. XGBoost survival modeling was explored but not retained because of unstable performance and poor discrimination. Software All analyses were performed in R using RStudio. Major packages included tidyverse, haven, janitor, survival, survminer, gtsummary, flextable, officer, survey, glmnet, pec, randomForestSRC, and riskRegression. Results A total of 32,290 participants from pooled NHANES 2007–2018 met criteria for the final analytic cohort, including 3,090 with albuminuria and preserved eGFR and 29,200 comparison participants. Across the pooled cohort, 2,768 deaths occurred during follow-up. Cohort contributions by cycle were 5,450 in 2007–2008, 5,886 in 2009–2010, 5,090 in 2011–2012, 5,534 in 2013–2014, 5,297 in 2015–2016, and 5,033 in 2017–2018. At baseline, participants with albuminuria and preserved eGFR were older than all other participants (mean age, 54.05 vs 47.50 years; P < .001), had higher mean body mass index (30.12 vs 28.97 kg/m²; P < .001), and were more likely to have hypertension (1,591/3,090 (52%) vs 9,549/29,200 (33%); P < .001) and diabetes (879/3,090 (28%) vs 3,153/29,200 (11%); P < .001). Mean hemoglobin A1c (HbA1c) was also higher in the phenotype group (6.43% vs 5.69%; P < .001), and differences were also observed in sex, race/ethnicity, and smoking status (Table 1 ). Table 1 Baseline characteristics of participants in the pooled NHANES 2007–2018 cohort, stratified by albuminuria with preserved eGFR phenotype. Characteristic All other participants N = 29,200 Albuminuria + preserved eGFR N = 3,090 p-value Age, years 47.50 (18.33) 54.05 (18.25) < 0.001 Sex 0.008 Female 14,911 (51%) 1,656 (54%) Male 14,289 (49%) 1,434 (46%) Race/ethnicity < 0.001 Mexican American 4,530 (16%) 562 (18%) Non-Hispanic Black 6,006 (21%) 664 (21%) Non-Hispanic White 11,957 (41%) 1,136 (37%) Other Hispanic 3,101 (11%) 341 (11%) Other race/ethnicity 3,606 (12%) 387 (13%) Body mass index, kg/m² 28.97 (6.87) 30.12 (7.76) < 0.001 Unknown 266 67 Hypertension 9,549 (33%) 1,591 (52%) < 0.001 Unknown 36 4 Diabetes 3,153 (11%) 879 (28%) < 0.001 Unknown 18 5 Smoking status < 0.001 Current/Some days 5,614 (20%) 662 (22%) Former 6,596 (23%) 765 (25%) Never 16,258 (57%) 1,581 (53%) Unknown 732 82 HbA1c, % 5.69 (0.93) 6.43 (1.83) < 0.001 Unknown 46 2 Values are presented as mean (SD) or n (%). P values were calculated using the Wilcoxon rank-sum test for continuous variables and the Pearson chi-square test for categorical variables. Abbreviations: eGFR, estimated glomerular filtration rate; HbA1c, hemoglobin A1c. Kaplan–Meier analysis showed lower overall survival among participants with albuminuria and preserved eGFR than among all other participants, with clear separation of curves over follow-up (Fig. 1 ). In unadjusted Cox analysis, albuminuria with preserved eGFR was associated with increased all-cause mortality (HR, 2.47; 95% CI, 2.25–2.71; P < .001). The association remained significant after adjustment for age, sex, pooled race/ethnicity, and NHANES cycle (HR, 1.77; 95% CI, 1.61–1.95; P < .001), and in the fully adjusted model including body mass index, hypertension, diabetes, and smoking status (HR, 1.57; 95% CI, 1.42–1.73; P < .001) (Table 2 ). In the fully adjusted pooled model, older age, male sex, non-Hispanic Black race, non-Hispanic White race, hypertension, diabetes, and smoking status were also associated with mortality, whereas NHANES cycle was not (Table 2 ). There was no evidence of proportional hazards violation for the primary exposure in any pooled model. In the primary fully adjusted model, the proportional hazards test for albuminuria with preserved eGFR was not significant (P = .865). Although the global test was significant in adjusted pooled models, this appeared to be driven mainly by age. In sensitivity analysis replacing continuous age with age groups, the association between albuminuria with preserved eGFR and mortality remained similar (HR, 1.58; 95% CI, 1.43–1.74; P < .001), and the proportional hazards test for the primary exposure remained non-significant (P = .715). Table 2 Association of albuminuria with preserved eGFR and all-cause mortality in pooled NHANES 2007–2018 Cox proportional hazards models. Variable Model 1: Unadjusted Model 2: Demographic + cycle adjusted Model 3: Clinical + cycle adjusted HR 95% CI p-value HR 95% CI p-value HR 95% CI p-value Albuminuria + preserved eGFR phenotype 2.47 2.25–2.71 < 0.001 1.77 1.61–1.95 < 0.001 1.57 1.42–1.73 < 0.001 Age, years 1.09 1.09–1.09 < 0.001 1.09 1.09–1.10 < 0.001 Sex Female — — — — Male 1.52 1.41–1.64 < 0.001 1.39 1.28–1.50 < 0.001 Race/ethnicity Mexican American — — — — Non-Hispanic Black 1.62 1.39–1.89 < 0.001 1.45 1.24–1.70 < 0.001 Non-Hispanic White 1.65 1.43–1.90 < 0.001 1.65 1.43–1.90 < 0.001 Other Hispanic 1.06 0.87–1.29 0.6 1.09 0.89–1.33 0.4 Other race/ethnicity 1.05 0.85–1.29 0.7 1.06 0.85–1.31 0.6 NHANES cycle 2007–2008 — — — — 2009–2010 0.96 0.87–1.06 0.4 0.95 0.85–1.05 0.3 2011–2012 1.05 0.93–1.18 0.4 1.02 0.90–1.15 0.7 2013–2014 1.08 0.95–1.23 0.2 1.06 0.93–1.21 0.4 2015–2016 1.04 0.88–1.22 0.6 1.02 0.86–1.20 0.8 2017–2018 1.00 0.79–1.25 0.97 0.95 0.75–1.20 0.7 Body mass index, kg/m² 0.99 0.98–1.00 0.010 Hypertension 1.28 1.17–1.39 < 0.001 Diabetes 1.47 1.35–1.62 < 0.001 Smoking status Current/Some days — — Former 0.56 0.51–0.63 < 0.001 Never 0.45 0.40–0.50 < 0.001 Model 1 was unadjusted. Model 2 was adjusted for age, sex, race/ethnicity, and NHANES cycle. Model 3 was additionally adjusted for body mass index, hypertension, diabetes, and smoking status. Hazard ratios are presented with 95% confidence intervals. Abbreviations: CI, confidence interval; eGFR, estimated glomerular filtration rate; HR, hazard ratio. Participants with albuminuria despite preserved kidney function had lower overall survival than all other participants over follow-up. In survey-weighted sensitivity analyses accounting for the complex NHANES sampling design, the association between albuminuria with preserved eGFR and all-cause mortality remained significant. Weighted hazard ratios were 2.80 (95% CI, 2.43–3.24; P < .001) in the unadjusted model, 1.92 (95% CI, 1.69–2.19; P < .001) in the demographic- and cycle-adjusted model, and 1.68 (95% CI, 1.47–1.93; P < .001) in the fully adjusted model (Supplementary Table 1). These findings support the consistency of the primary pooled analyses. In the secondary machine learning analysis, the complete-case modeling cohort included 31,060 participants, of whom 2,661 died during follow-up; the training and test sets included 21,742 and 9,318 participants, respectively. All evaluated survival models demonstrated good discrimination on the held-out test set (Table 3). The standard Cox model achieved a C-index of 0.846, while LASSO Cox and Elastic Net Cox achieved 0.847 each; Random Survival Forest showed the highest discrimination (C-index, 0.851). In the secondary machine learning analysis, the complete-case modeling cohort included 31,060 participants, of whom 2,661 died during follow-up; the training and test sets included 21,742 and 9,318 participants, respectively. All evaluated survival models demonstrated good discrimination on the held-out test set (Table 3). The standard Cox model achieved a C-index of 0.846, while LASSO Cox and Elastic Net Cox achieved 0.847 each; Random Survival Forest showed the highest discrimination (C-index, 0.851). Table 3 Predictive performance of survival models for all-cause mortality in pooled NHANES 2007–2018 Model C-index Brier score, 60 mo Brier score, 120 mo IBS, 0–120 mo Standard Cox 0.846 0.0429 0.0796 0.0215 LASSO Cox 0.847 0.0427 0.0792 0.0213 Elastic Net Cox 0.847 0.0427 0.0792 0.0214 Random Survival Forest 0.851 0.0414 0.0779 0.0207 Model performance was evaluated on the held-out test set. A higher C-index indicates better discrimination, whereas lower Brier scores and integrated Brier scores indicate lower prediction error. Abbreviations: IBS, integrated Brier score. Prediction error analyses showed the same pattern. At 60 months, Brier scores were 0.0429 for standard Cox, 0.0427 for LASSO Cox, 0.0427 for Elastic Net Cox, and 0.0414 for Random Survival Forest. At 120 months, the corresponding Brier scores were 0.0796, 0.0792, 0.0792, and 0.0779. Integrated Brier score through 120 months was lowest for Random Survival Forest (0.0207), followed by LASSO Cox (0.0213), Elastic Net Cox (0.0214), and standard Cox (0.0215) (Table 3, Fig. 2). Overall, Random Survival Forest performed best, but absolute gains over standard and penalized Cox models were modest. Lower Brier scores indicate lower prediction error. Random Survival Forest showed the lowest prediction error across most follow-up time points, whereas standard Cox, LASSO Cox, and Elastic Net Cox performed similarly overall. Variable importance analysis from the Random Survival Forest model identified age, eGFR, and log-transformed urinary albumin-creatinine ratio as the strongest contributors to mortality prediction (Fig. 3). A heatmap of predicted risk further demonstrated higher estimated mortality risk at lower eGFR levels and higher log-transformed UACR values (Fig. 4). The figure shows the top predictors ranked by variable importance in the Random Survival Forest model. Age, eGFR, and log-transformed urinary albumin-creatinine ratio were the strongest contributors to mortality prediction. The heatmap shows relative predicted mortality risk across combinations of eGFR and log-transformed urinary albumin-creatinine ratio (UACR), with other covariates fixed at typical values. Higher predicted risk was observed at lower eGFR and higher log-transformed UACR values. Overall, albuminuria despite preserved eGFR was independently associated with higher all-cause mortality in pooled NHANES 2007–2018, while more complex machine learning approaches provided only limited incremental predictive benefit beyond conventional Cox-based survival modeling. Discussion In this pooled analysis of NHANES 2007–2018 data linked to mortality follow-up, albuminuria in the setting of preserved eGFR was independently associated with higher all-cause mortality. This association persisted after adjustment for demographic and clinical covariates and remained unchanged in age-group and survey-weighted sensitivity analyses, supporting the consistency of the primary findings. All evaluated survival models demonstrated good predictive performance, with Random Survival Forest achieving the best overall discrimination and lowest prediction error; however, the absolute improvement over standard and penalized Cox models was modest. Together, these findings suggest that albuminuria identifies an increased risk of mortality even when kidney function appears preserved by creatinine-based estimation. To our knowledge, this is one of the few contemporary population-based analyses to specifically evaluate albuminuria in the setting of preserved eGFR while directly comparing conventional and machine learning survival approaches. Albuminuria likely reflects clinically meaningful kidney and vascular injury that may become apparent before a measurable decline in filtration function, which may help explain why the excess mortality risk remained evident even among participants with preserved eGFR. Prior work has shown that albuminuria is associated with adverse outcomes independently of eGFR and traditional cardiovascular risk factors[2]. Filtered albumin may also contribute directly to injury through proximal tubular uptake, inflammatory signaling, and tubulointerstitial fibrosis. More broadly, albuminuria also reflects disruption of the glomerular filtration barrier and systemic endothelial dysfunction[10,11]. In that context, our findings suggest that albuminuria is not simply an accompanying laboratory abnormality, but an early risk marker that may identify higher-risk individuals earlier than creatinine-based kidney function alone. This interpretation is also consistent with our machine learning results, in which both eGFR and log-transformed UACR ranked among the strongest contributors to mortality prediction, suggesting that kidney risk in this phenotype is better captured by combining filtration- and albumin-based information than by either measure in isolation[12,13]. These findings support broader use of albuminuria testing in individuals with preserved eGFR as part of routine risk stratification. An important additional finding of this study was that more complex survival modeling provided only a limited incremental benefit over conventional regression-based approaches. Although Random Survival Forest achieved the best overall performance, the absolute improvements in discrimination and prediction error over standard Cox, LASSO Cox, and Elastic Net Cox models were small[14]. This pattern suggests that in structured population health data, much of the prognostic signal may already be captured by a relatively small number of strong and clinically interpretable predictors, limiting the added value of more flexible machine learning methods. Similar benchmarking work has shown that no single survival method consistently dominates across datasets, and that classical Cox-based approaches often remain highly competitive for clinical prediction tasks[15–17]. At the same time, the strong contribution of eGFR and log-transformed UACR in the Random Survival Forest model reinforces the importance of kidney-related markers in mortality prediction and supports the view that albumin-based and filtration-based information remain central even in more advanced predictive frameworks. Another important aspect of our findings is their consistency across multiple analytic approaches. The association between albuminuria with preserved eGFR and all-cause mortality remained similar after multivariable adjustment, replacement of continuous age with age categories, and survey-weighted Cox modeling that accounted for the complex NHANES sampling design. In addition, there was no evidence that the proportional hazards assumption was violated for the primary exposure, supporting the stability of the main association over follow-up. This study has several limitations. Albuminuria was defined from a single UACR measurement, and because urinary albumin excretion can show substantial within-person variability, some participants may have been misclassified[18]. In addition, current KDIGO guidance emphasizes that confirmation of chronicity requires repeat assessment over time rather than a single abnormal measurement[1]. As an observational analysis, the study cannot establish causality, and residual confounding remains possible despite multivariable adjustment. Finally, although the machine learning models were evaluated on a held-out test set, they were not externally validated in an independent cohort, which limits assessment of generalizability[19,20]. Conclusion In conclusion, albuminuria in the setting of preserved eGFR was independently associated with higher all-cause mortality in pooled NHANES 2007–2018. This association remained consistent across sensitivity analyses, supporting the role of albuminuria as an early marker of risk even when kidney function appears preserved. Although Random Survival Forest demonstrated the best predictive performance, the incremental improvement over conventional Cox-based models was modest. These findings highlight the prognostic importance of albuminuria beyond eGFR and support the continued relevance of conventional survival models for population-level risk stratification. Declarations Funding No funding was received for conducting this study. Conflicts of interest/Competing interests The authors have no relevant financial or non-financial interests to disclose. Ethics approval This study used publicly available, de-identified NHANES data with linked mortality follow-up. Therefore, additional institutional review board approval was not required for this secondary analysis. Consent to participate Not applicable because this study was a secondary analysis of publicly available, de-identified data. Written Consent for publication Not applicable. Availability of data and material The datasets analysed in this study are publicly available from the National Health and Nutrition Examination Survey (NHANES) website and the NHANES Linked Mortality Files [8,9]. Code availability The analytic code used for data cleaning, statistical analysis, and machine learning modeling is available from the corresponding author on reasonable request. Large Language Model (LLM) No large language model was used for the generation of the study concept, data analysis, results, or scientific conclusions. Any language editing assistance, if used, did not affect the scientific content, and all authors take full responsibility for the final manuscript. Authors' contributions Sai Tharun Reddy Gopannagari contributed to the study concept and design, data curation, formal analysis, interpretation of the data, and drafting of the original manuscript. Divya Ganisetti and Kesava Manikanta Achuta contributed equally to the study concept and design, manuscript writing, critical revision, and final editing of the manuscript. All authors read and approved the final manuscript and agree to be accountable for all aspects of the work. References Levin A, Ahmed SB, Carrero JJ, et al. Executive summary of the KDIGO 2024 Clinical Practice Guideline for the Evaluation and Management of Chronic Kidney Disease: known knowns and known unknowns. Kidney Int. 2024;105(11):684–701. https://doi.org/10.1016/j.kint.2023.10.016 . Matsushita K, van der Velde M, Astor BC, et al. Association of estimated glomerular filtration rate and albuminuria with all-cause and cardiovascular mortality: a collaborative meta-analysis of general population cohorts. Lancet. 2010;375(9731):2073. https://doi.org/10.1016/S0140-6736(10)60674-5 . Astor BC, Hallan SI, Miller ER, Yeung E, Coresh J. Glomerular filtration rate, albuminuria, and risk of cardiovascular and all-cause mortality in the US population. Am J Epidemiol. 2008;167(10):1226–34. https://doi.org/10.1093/aje/kwn033 . Barzilay JI, Farag YMK, Durthaler J. Albuminuria: an underappreciated risk factor for cardiovascular disease. J Am Heart Assoc. 2024;13(2). https://doi.org/10.1161/JAHA.123.030131 . Chu CD, Powe NR, Shlipak MG, et al. Albuminuria testing and nephrology care among insured US adults with chronic kidney disease: a missed opportunity. BMC Prim Care. 2022;23(1). https://doi.org/10.1186/S12875-022-01910-9 . Chu CD, Xia F, Du Y, et al. Estimated prevalence and testing for albuminuria in US adults at risk for chronic kidney disease. JAMA Netw Open. 2023;6(7):e2326230. https://doi.org/10.1001/jamanetworkopen.2023.26230 . Albekery MA, Alhomoud IS, Alabdulathim LS, et al. Underutilization of albuminuria screening in adults with diabetes mellitus or hypertension: a systematic review and meta-analysis. BMC Nephrol. 2025;27(1). https://doi.org/10.1186/S12882-025-04672-5 . Questionnaires NHANES. Datasets, and Related Documentation. National Center for Health Statistics. https://wwwn.cdc.gov/nchs/nhanes/ Accessed 12 Apr 2026. National Death Index | Data Linkage | CDC. Centers for Disease Control and Prevention. https://www.cdc.gov/nchs/linked-data/mortality-files/index.html Accessed 12 Apr 2026. Salmon AH, Satchell SC. Endothelial glycocalyx dysfunction in disease: albuminuria and increased microvascular permeability. J Pathol. 2012;226(4):562–74. https://doi.org/10.1002/path.3964 . Molitoris BA, Sandoval RM, Yadav SPS, Wagner MC. Albumin uptake and processing by the proximal tubule: physiological, pathological, and therapeutic implications. Physiol Rev. 2022;102(4):1625. https://doi.org/10.1152/physrev.00014.2021 . Matsushita K, Coresh J, Sang Y, et al. Estimated glomerular filtration rate and albuminuria for prediction of cardiovascular outcomes: a collaborative meta-analysis of individual participant data. Lancet Diabetes Endocrinol. 2015;3(7):514–25. https://doi.org/10.1016/S2213-8587(15)00040-6 . Wright EE, Cebrian A, Ngui D. A primary care guide to the screening and pharmacologic management of chronic kidney disease in people living with type 2 diabetes. Clin Diabetes. 2025;43(4):531–44. https://doi.org/10.2337/cd25-0013 . Simon N, Friedman J, Hastie T, Tibshirani R. Regularization paths for Cox’s proportional hazards model via coordinate descent. J Stat Softw. 2011;39(5):1. https://doi.org/10.18637/jss.v039.i05 . Zhang Y, Wong G, Mann G, Muller S, Yang JYH. SurvBenchmark: comprehensive benchmarking study of survival analysis methods using both omics data and clinical data. Gigascience. 2022;11:1–13. https://doi.org/10.1093/gigascience/giac071 . van Belle V, Pelckmans K, van Huffel S, Suykens JAK. Improved performance on high-dimensional survival data by application of Survival-SVM. Bioinformatics. 2011;27(1):87–94. https://doi.org/10.1093/bioinformatics/btq617 . Herrmann M, Probst P, Hornung R, Jurinovic V, Boulesteix AL. Large-scale benchmark study of survival prediction methods using multi-omics data. Brief Bioinform. 2021;22(3). https://doi.org/10.1093/bib/bbaa167 . Waikar SS, Rebholz CM, Zheng Z, et al. Biological variability of estimated GFR and albuminuria in CKD. Am J Kidney Dis. 2018;72(4):538. https://doi.org/10.1053/j.ajkd.2018.04.023 . Riley RD, Ensor J, Snell KIE, et al. External validation of clinical prediction models using big datasets from e-health records or IPD meta-analysis: opportunities and challenges. BMJ. 2016;353. https://doi.org/10.1136/bmj.i3140 . Collins GS, Moons KGM, Dhiman P et al. TRIPOD + AI statement: updated guidance for reporting clinical prediction models that use regression or machine learning methods. BMJ. 2024;385. https://doi.org/10.1136/bmj-2023-078378 Additional Declarations No competing interests reported. Supplementary Files SupplementaryTable1.docx Cite Share Download PDF Status: Under Review Version 1 posted Reviews received at journal 11 May, 2026 Reviewers agreed at journal 07 May, 2026 Reviewers invited by journal 21 Apr, 2026 Editor assigned by journal 20 Apr, 2026 Submission checks completed at journal 20 Apr, 2026 First submitted to journal 14 Apr, 2026 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-9420411","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":628380388,"identity":"8a241f40-5577-4902-9b7f-5025767d8d5d","order_by":0,"name":"Sai Tharun Reddy Gopannagari","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAAwElEQVRIiWNgGAWjYDADfgjFTIIWyQaStRgcIFaLvHvvwc+FO+zsNl87/kyCocI6sYGQFsMz55KlZ55JTt52O8dMguFMOhFaZuQYSPO2MSeb3c5hk2BsO0yUFuPfvG31ycaz059JMP4jQou8RI4Z0JbDdgbSCWYSjA1EaDHgOZdmzdt2PEHido6xRcKxdGPCtrT3Hr7N21Ztzz87/eGNDzXWsoRtOcADpiHuSSCkHGxLA0SLPTGKR8EoGAWjYIQCAEvPO/g4AEvnAAAAAElFTkSuQmCC","orcid":"","institution":"Garden City Hospital, Michigan State University","correspondingAuthor":true,"prefix":"","firstName":"Sai","middleName":"Tharun Reddy","lastName":"Gopannagari","suffix":""},{"id":628380389,"identity":"508c11ec-8d33-4e4f-95d2-f7dd9d1f8ca5","order_by":1,"name":"Kesava Manikanta Achuta","email":"","orcid":"","institution":"Garden City Hospital, Michigan State University","correspondingAuthor":false,"prefix":"","firstName":"Kesava","middleName":"Manikanta","lastName":"Achuta","suffix":""},{"id":628380390,"identity":"f650b121-ddc6-4277-b857-72dd21598840","order_by":2,"name":"Divya Ganisetti","email":"","orcid":"","institution":"Atal Bihari Vajpayee Institute of Medical Sciences \u0026 Dr Ram Manohar Lohia Hospital","correspondingAuthor":false,"prefix":"","firstName":"Divya","middleName":"","lastName":"Ganisetti","suffix":""}],"badges":[],"createdAt":"2026-04-15 01:38:32","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-9420411/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-9420411/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":108492299,"identity":"1217219a-acb7-40cb-b5d2-34ec8935dd2d","added_by":"auto","created_at":"2026-05-05 09:57:24","extension":"jpeg","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":336072,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eKaplan–Meier survival curves for all-cause mortality according to albuminuria with preserved eGFR status in pooled NHANES 2007–2018.\u003c/strong\u003e\u003c/p\u003e","description":"","filename":"floatimage1.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-9420411/v1/b096106739375d5ef5395730.jpeg"},{"id":109203961,"identity":"d981e378-a3ce-467e-8c58-62d547cab3d4","added_by":"auto","created_at":"2026-05-13 14:51:03","extension":"jpeg","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":240760,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eBrier score over time for survival models predicting all-cause mortality in pooled NHANES 2007–2018.\u003c/strong\u003e\u003c/p\u003e","description":"","filename":"floatimage2.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-9420411/v1/af17cc3ba9786522ab7dda37.jpeg"},{"id":108385080,"identity":"84596164-4ebc-48e6-8828-4667c86c93c3","added_by":"auto","created_at":"2026-05-04 06:01:11","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":33383,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eVariable importance in the Random Survival Forest model for all-cause mortality in pooled NHANES 2007–2018.\u003c/strong\u003e\u003c/p\u003e","description":"","filename":"floatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-9420411/v1/fa82a6b12542630a0a602b46.png"},{"id":108385082,"identity":"fad89b15-44e8-42f2-8248-1f95210a1393","added_by":"auto","created_at":"2026-05-04 06:01:11","extension":"jpeg","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":309379,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003ePredicted all-cause mortality risk by eGFR and log-transformed UACR in the Random Survival Forest model.\u003c/strong\u003e\u003c/p\u003e","description":"","filename":"floatimage4.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-9420411/v1/d755def587fde88f33e1d28f.jpeg"},{"id":109206055,"identity":"9e5205f6-2a86-42a1-965b-5aad4134a307","added_by":"auto","created_at":"2026-05-13 15:10:51","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1271549,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-9420411/v1/88ac4d0e-aaee-4658-b133-185864c2c1c1.pdf"},{"id":108385078,"identity":"433f7944-7412-43d0-b0aa-109d77966d75","added_by":"auto","created_at":"2026-05-04 06:01:11","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":14982,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementaryTable1.docx","url":"https://assets-eu.researchsquare.com/files/rs-9420411/v1/c69e41155a6fdb5dd5a41313.docx"}],"financialInterests":"No competing interests reported.","formattedTitle":"Association of Albuminuria with Preserved eGFR and Mortality in Pooled NHANES 2007–2018: Conventional and Machine Learning Survival Analyses","fulltext":[{"header":"Introduction","content":"\u003cp\u003eAlbuminuria is an important marker of kidney damage and a broader indicator of cardiorenal risk. Current Kidney Disease: Improving Global Outcomes (KDIGO) guidance emphasizes that kidney disease risk should be assessed using both estimated glomerular filtration rate (eGFR) and albuminuria, as these measures capture different but complementary aspects of kidney health[\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e]. Prior studies have also shown that albuminuria is associated with adverse outcomes, including all-cause and cardiovascular mortality, independent of eGFR[\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e, \u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e]. Together, these observations suggest that clinically meaningful kidney and vascular injury may already be present even when filtration function appears preserved.\u003c/p\u003e \u003cp\u003eDespite its clinical importance, albuminuria remains underdetected in routine practice, particularly outside recognized chronic kidney disease (CKD) populations[\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e]. Recent United States (US) data suggest that albuminuria testing is performed in only a minority of at-risk adults, raising concern that clinically relevant kidney disease may go unrecognized when assessment relies mainly on serum creatinine or eGFR[\u003cspan additionalcitationids=\"CR6\" citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e]. This gap in recognition is especially relevant in population-based studies, where many individuals may have abnormal albumin excretion despite preserved filtration function. In this setting, better characterization of the mortality implications of albuminuria with preserved eGFR could help refine risk identification beyond traditional creatinine-based assessment alone.\u003c/p\u003e \u003cp\u003eAgainst this background, population-level data are needed to better define the prognostic significance of albuminuria in adults without reduced eGFR. The National Health and Nutrition Examination Survey (NHANES), with linked mortality follow-up, provides a useful setting for this question because it combines a nationally representative sample of US adults with standardized laboratory data and longitudinal outcomes[\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e, \u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e]. It also offers an opportunity to examine whether machine learning\u0026ndash;based survival models provide meaningful predictive benefit beyond conventional Cox analysis.\u003c/p\u003e \u003cp\u003eIn this study, we evaluated the association of albuminuria with preserved eGFR and all-cause mortality in pooled NHANES 2007\u0026ndash;2018 data linked to mortality follow-up. We also assessed the consistency of this association in sensitivity analyses and compared the predictive performance of conventional Cox-based survival modeling with machine learning\u0026ndash;based survival approaches.\u003c/p\u003e"},{"header":"Methods","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003eStudy Design and Data Source\u003c/h2\u003e \u003cp\u003eWe performed a retrospective cohort study using pooled NHANES 2007\u0026ndash;2018 data with linked mortality follow-up. NHANES is a nationally representative survey of the noninstitutionalized US population that integrates interview, examination, and laboratory data. Mortality follow-up data were obtained from the NHANES Linked Mortality Files. This study used publicly available, de-identified NHANES data with linked mortality follow-up; therefore, additional institutional review board approval and informed consent were not required for this secondary analysis.\u003c/p\u003e \u003c/div\u003e\n\u003ch3\u003eStudy Population\u003c/h3\u003e\n\u003cp\u003eParticipants aged 18 years or older were eligible if urinary albumin, urine creatinine, and serum creatinine data were available. Participants were excluded if mortality linkage eligibility, mortality status, follow-up time, or key demographic variables were missing. After pooling all cycles, the final analytic cohort included 32,290 participants.\u003c/p\u003e\n\u003ch3\u003eExposure Definition\u003c/h3\u003e\n\u003cp\u003eThe primary exposure was albuminuria with preserved kidney function. UACR was obtained directly when available and was calculated manually in 2007\u0026ndash;2008 from urine albumin and urine creatinine. Albuminuria was defined as UACR at least 30 mg/g. Kidney function was estimated using the 2021 Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) creatinine equation, and preserved kidney function was defined as eGFR\u0026thinsp;\u0026ge;\u0026thinsp;60 mL/min/1.73 m\u0026sup2;. The exposure group, therefore, comprised participants with both albuminuria and preserved eGFR; all others served as the comparison group.\u003c/p\u003e\n\u003ch3\u003eCovariates\u003c/h3\u003e\n\u003cp\u003eCovariates were selected a priori based on clinical relevance and availability across cycles and included age, sex, race/ethnicity, body mass index, hypertension, diabetes, smoking status, hemoglobin A1c, and NHANES cycle. Because race/ethnicity coding differed across cycles, a harmonized pooled variable was created: Mexican American, Other Hispanic, Non-Hispanic White, Non-Hispanic Black, and other race/ethnicity. Hypertension and diabetes were derived from questionnaire variables, and smoking status was categorized as current/some days, former, or never.\u003c/p\u003e\n\u003ch3\u003eOutcome\u003c/h3\u003e\n\u003cdiv class=\"Heading\"\u003eOutcome\u003c/div\u003e \u003cp\u003eThe primary outcome was all-cause mortality. Mortality status and follow-up time in months were obtained from the NHANES Linked Mortality Files. Follow-up time was defined using person-months from examination, and death status was defined from linked mortality indicators. Survival time was analyzed from examination to death or censoring.\u003c/p\u003e \u003cdiv id=\"Sec8\" class=\"Section2\"\u003e \u003ch2\u003eData Management and Harmonization\u003c/h2\u003e \u003cp\u003eCycle-specific demographic, laboratory, examination, and questionnaire files were merged by participant identifier (SEQN), harmonized across cycles, and pooled into a single 2007\u0026ndash;2018 analytic dataset, including race/ethnicity harmonization and cycle-specific handling of 2007\u0026ndash;2008 UACR.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec9\" class=\"Section2\"\u003e \u003ch2\u003eStatistical Analysis\u003c/h2\u003e \u003cp\u003eBaseline characteristics were summarized by exposure group as mean (SD) or n (%). Group comparisons used Wilcoxon rank-sum tests for continuous variables and Pearson chi-square tests for categorical variables. Overall survival was evaluated using Kaplan\u0026ndash;Meier methods. Associations between albuminuria with preserved eGFR and all-cause mortality were examined using Cox proportional hazards regression with 3 models: unadjusted; adjusted for age, sex, pooled race/ethnicity, and NHANES cycle; and additionally adjusted for body mass index, hypertension, diabetes, and smoking status. Hazard ratios with 95% confidence intervals were reported.\u003c/p\u003e \u003c/div\u003e\n\u003ch3\u003eSensitivity Analyses and Assessment of Proportional Hazards\u003c/h3\u003e\n\u003cp\u003eProportional hazards assumptions were assessed using Schoenfeld residual testing with cox.zph, focusing on the primary exposure. Because adjusted pooled models showed mild global nonproportionality driven mainly by age, a sensitivity analysis replaced continuous age with categories of 18\u0026ndash;39, 40\u0026ndash;59, 60\u0026ndash;74, and \u0026ge;\u0026thinsp;75 years. Survey-weighted Cox models were also fit using pooled 12-year Mobile Examination Center (MEC) examination weights, masked strata, and primary sampling units with Taylor linearized variance estimation; the same sequence of unadjusted, demographic- and cycle-adjusted, and fully adjusted models was applied.\u003c/p\u003e \u003cdiv id=\"Sec11\" class=\"Section2\"\u003e \u003ch2\u003eMachine Learning Survival Modeling\u003c/h2\u003e \u003cp\u003eAs a secondary predictive analysis, survival machine learning models were developed in the pooled complete-case dataset (n\u0026thinsp;=\u0026thinsp;31,060; deaths\u0026thinsp;=\u0026thinsp;2,661), split into training (70%; n\u0026thinsp;=\u0026thinsp;21,742) and held-out test (30%; n\u0026thinsp;=\u0026thinsp;9,318) sets. Candidate predictors included age, sex, pooled race/ethnicity, body mass index, hypertension, diabetes, smoking status, hemoglobin A1c, eGFR, log-transformed UACR, albuminuria with preserved eGFR phenotype, and NHANES cycle. A small positive offset was applied to zero follow-up times to permit penalized Cox modeling. Evaluated models were standard Cox, least absolute shrinkage and selection operator (LASSO) Cox, Elastic Net Cox, and Random Survival Forest. Penalized Cox models used cross-validation to select tuning parameters, Random Survival Forest used 500 trees, and model performance was assessed on the held-out test set using Harrell\u0026rsquo;s C-index, Brier scores at 60 and 120 months, and integrated Brier score through 120 months. XGBoost survival modeling was explored but not retained because of unstable performance and poor discrimination.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec12\" class=\"Section2\"\u003e \u003ch2\u003eSoftware\u003c/h2\u003e \u003cp\u003eAll analyses were performed in R using RStudio. Major packages included tidyverse, haven, janitor, survival, survminer, gtsummary, flextable, officer, survey, glmnet, pec, randomForestSRC, and riskRegression.\u003c/p\u003e \u003c/div\u003e"},{"header":"Results","content":"\u003cp\u003eA total of 32,290 participants from pooled NHANES 2007\u0026ndash;2018 met criteria for the final analytic cohort, including 3,090 with albuminuria and preserved eGFR and 29,200 comparison participants. Across the pooled cohort, 2,768 deaths occurred during follow-up. Cohort contributions by cycle were 5,450 in 2007\u0026ndash;2008, 5,886 in 2009\u0026ndash;2010, 5,090 in 2011\u0026ndash;2012, 5,534 in 2013\u0026ndash;2014, 5,297 in 2015\u0026ndash;2016, and 5,033 in 2017\u0026ndash;2018.\u003c/p\u003e\n\u003cp\u003eAt baseline, participants with albuminuria and preserved eGFR were older than all other participants (mean age, 54.05 vs 47.50 years; P \u0026lt; .001), had higher mean body mass index (30.12 vs 28.97 kg/m\u0026sup2;; P \u0026lt; .001), and were more likely to have hypertension (1,591/3,090 (52%) vs 9,549/29,200 (33%); P \u0026lt; .001) and diabetes (879/3,090 (28%) vs 3,153/29,200 (11%); P \u0026lt; .001). Mean hemoglobin A1c (HbA1c) was also higher in the phenotype group (6.43% vs 5.69%; P \u0026lt; .001), and differences were also observed in sex, race/ethnicity, and smoking status (Table \u003cspan class=\"InternalRef\"\u003e1\u003c/span\u003e).\u0026nbsp;\u003c/p\u003e\n\u003cdiv class=\"gridtable\"\u003e\n \u003ctable id=\"Tab1\" border=\"1\"\u003e\n \u003ccaption\u003e\n \u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e\n \u003cdiv class=\"CaptionContent\"\u003e\n \u003cp\u003eBaseline characteristics of participants in the pooled NHANES 2007\u0026ndash;2018 cohort, stratified by albuminuria with preserved eGFR phenotype.\u003c/p\u003e\n \u003c/div\u003e\n \u003c/caption\u003e\n \u003cthead\u003e\n \u003ctr\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eCharacteristic\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eAll other participants\u003c/p\u003e\n \u003cp\u003eN\u0026thinsp;=\u0026thinsp;29,200\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eAlbuminuria\u0026thinsp;+\u0026thinsp;preserved eGFR\u003c/p\u003e\n \u003cp\u003eN\u0026thinsp;=\u0026thinsp;3,090\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003ep-value\u003c/p\u003e\n \u003c/th\u003e\n \u003c/tr\u003e\n \u003c/thead\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003eAge, years\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e47.50 (18.33)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e54.05 (18.25)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\" char=\".\"\u003e\n \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003eSex\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"char\" char=\".\"\u003e\n \u003cp\u003e0.008\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eFemale\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e14,911 (51%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1,656 (54%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eMale\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e14,289 (49%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1,434 (46%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003eRace/ethnicity\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"char\" char=\".\"\u003e\n \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eMexican American\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e4,530 (16%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e562 (18%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eNon-Hispanic Black\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e6,006 (21%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e664 (21%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eNon-Hispanic White\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e11,957 (41%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1,136 (37%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eOther Hispanic\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e3,101 (11%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e341 (11%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eOther race/ethnicity\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e3,606 (12%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e387 (13%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003eBody mass index, kg/m\u0026sup2;\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e28.97 (6.87)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e30.12 (7.76)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\" char=\".\"\u003e\n \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eUnknown\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e266\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e67\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003eHypertension\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e9,549 (33%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1,591 (52%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\" char=\".\"\u003e\n \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eUnknown\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e36\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e4\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003eDiabetes\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e3,153 (11%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e879 (28%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\" char=\".\"\u003e\n \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eUnknown\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e18\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e5\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003eSmoking status\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"char\" char=\".\"\u003e\n \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eCurrent/Some days\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e5,614 (20%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e662 (22%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eFormer\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e6,596 (23%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e765 (25%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eNever\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e16,258 (57%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1,581 (53%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eUnknown\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e732\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e82\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003eHbA1c, %\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e5.69 (0.93)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e6.43 (1.83)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\" char=\".\"\u003e\n \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eUnknown\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e46\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e2\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n \u003ctfoot\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"4\"\u003eValues are presented as mean (SD) or n (%). P values were calculated using the Wilcoxon rank-sum test for continuous variables and the Pearson chi-square test for categorical variables. Abbreviations: eGFR, estimated glomerular filtration rate; HbA1c, hemoglobin A1c.\u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tfoot\u003e\n \u003c/table\u003e\n\u003c/div\u003e\n\u003cp\u003eKaplan\u0026ndash;Meier analysis showed lower overall survival among participants with albuminuria and preserved eGFR than among all other participants, with clear separation of curves over follow-up (Fig. \u003cspan class=\"InternalRef\"\u003e1\u003c/span\u003e). In unadjusted Cox analysis, albuminuria with preserved eGFR was associated with increased all-cause mortality (HR, 2.47; 95% CI, 2.25\u0026ndash;2.71; P \u0026lt; .001). The association remained significant after adjustment for age, sex, pooled race/ethnicity, and NHANES cycle (HR, 1.77; 95% CI, 1.61\u0026ndash;1.95; P \u0026lt; .001), and in the fully adjusted model including body mass index, hypertension, diabetes, and smoking status (HR, 1.57; 95% CI, 1.42\u0026ndash;1.73; P \u0026lt; .001) (Table \u003cspan class=\"InternalRef\"\u003e2\u003c/span\u003e).\u003c/p\u003e\n\u003cp\u003eIn the fully adjusted pooled model, older age, male sex, non-Hispanic Black race, non-Hispanic White race, hypertension, diabetes, and smoking status were also associated with mortality, whereas NHANES cycle was not (Table \u003cspan class=\"InternalRef\"\u003e2\u003c/span\u003e). There was no evidence of proportional hazards violation for the primary exposure in any pooled model. In the primary fully adjusted model, the proportional hazards test for albuminuria with preserved eGFR was not significant (P = .865). Although the global test was significant in adjusted pooled models, this appeared to be driven mainly by age. In sensitivity analysis replacing continuous age with age groups, the association between albuminuria with preserved eGFR and mortality remained similar (HR, 1.58; 95% CI, 1.43\u0026ndash;1.74; P \u0026lt; .001), and the proportional hazards test for the primary exposure remained non-significant (P = .715).\u003c/p\u003e\n\u003cdiv class=\"gridtable\"\u003e\n \u003ctable id=\"Tab2\" border=\"1\"\u003e\n \u003ccaption\u003e\n \u003cdiv class=\"CaptionNumber\"\u003eTable 2\u003c/div\u003e\n \u003cdiv class=\"CaptionContent\"\u003e\n \u003cp\u003eAssociation of albuminuria with preserved eGFR and all-cause mortality in pooled NHANES 2007\u0026ndash;2018 Cox proportional hazards models.\u003c/p\u003e\n \u003c/div\u003e\n \u003c/caption\u003e\n \u003cthead\u003e\n \u003ctr\u003e\n \u003cth rowspan=\"2\" align=\"left\"\u003e\n \u003cp\u003eVariable\u003c/p\u003e\n \u003c/th\u003e\n \u003cth colspan=\"3\" align=\"left\"\u003e\n \u003cp\u003eModel 1: Unadjusted\u003c/p\u003e\n \u003c/th\u003e\n \u003cth colspan=\"3\" align=\"left\"\u003e\n \u003cp\u003eModel 2: Demographic\u0026thinsp;+\u0026thinsp;cycle adjusted\u003c/p\u003e\n \u003c/th\u003e\n \u003cth colspan=\"3\" align=\"left\"\u003e\n \u003cp\u003eModel 3: Clinical\u0026thinsp;+\u0026thinsp;cycle adjusted\u003c/p\u003e\n \u003c/th\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eHR\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003e95% CI\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003ep-value\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eHR\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003e95% CI\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003ep-value\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eHR\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003e95% CI\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003ep-value\u003c/p\u003e\n \u003c/th\u003e\n \u003c/tr\u003e\n \u003c/thead\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003eAlbuminuria\u0026thinsp;+\u0026thinsp;preserved eGFR phenotype\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\" char=\".\"\u003e\n \u003cp\u003e2.47\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\" char=\".\"\u003e\n \u003cp\u003e2.25\u0026ndash;2.71\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\" char=\".\"\u003e\n \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1.77\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1.61\u0026ndash;1.95\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\" char=\".\"\u003e\n \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1.57\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1.42\u0026ndash;1.73\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\" char=\".\"\u003e\n \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003eAge, years\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1.09\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1.09\u0026ndash;1.09\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\" char=\".\"\u003e\n \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1.09\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1.09\u0026ndash;1.10\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\" char=\".\"\u003e\n \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003eSex\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eFemale\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u0026mdash;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u0026mdash;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u0026mdash;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u0026mdash;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eMale\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1.52\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1.41\u0026ndash;1.64\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\" char=\".\"\u003e\n \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1.39\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1.28\u0026ndash;1.50\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\" char=\".\"\u003e\n \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003eRace/ethnicity\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eMexican American\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u0026mdash;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u0026mdash;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u0026mdash;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u0026mdash;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eNon-Hispanic Black\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1.62\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1.39\u0026ndash;1.89\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\" char=\".\"\u003e\n \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1.45\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1.24\u0026ndash;1.70\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\" char=\".\"\u003e\n \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eNon-Hispanic White\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1.65\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1.43\u0026ndash;1.90\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\" char=\".\"\u003e\n \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1.65\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1.43\u0026ndash;1.90\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\" char=\".\"\u003e\n \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eOther Hispanic\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1.06\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.87\u0026ndash;1.29\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\" char=\".\"\u003e\n \u003cp\u003e0.6\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1.09\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.89\u0026ndash;1.33\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\" char=\".\"\u003e\n \u003cp\u003e0.4\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eOther race/ethnicity\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1.05\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.85\u0026ndash;1.29\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\" char=\".\"\u003e\n \u003cp\u003e0.7\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1.06\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.85\u0026ndash;1.31\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\" char=\".\"\u003e\n \u003cp\u003e0.6\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003eNHANES cycle\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e2007\u0026ndash;2008\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u0026mdash;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u0026mdash;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u0026mdash;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u0026mdash;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e2009\u0026ndash;2010\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.96\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.87\u0026ndash;1.06\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\" char=\".\"\u003e\n \u003cp\u003e0.4\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.95\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.85\u0026ndash;1.05\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\" char=\".\"\u003e\n \u003cp\u003e0.3\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e2011\u0026ndash;2012\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1.05\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.93\u0026ndash;1.18\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\" char=\".\"\u003e\n \u003cp\u003e0.4\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1.02\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.90\u0026ndash;1.15\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\" char=\".\"\u003e\n \u003cp\u003e0.7\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e2013\u0026ndash;2014\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1.08\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.95\u0026ndash;1.23\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\" char=\".\"\u003e\n \u003cp\u003e0.2\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1.06\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.93\u0026ndash;1.21\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\" char=\".\"\u003e\n \u003cp\u003e0.4\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e2015\u0026ndash;2016\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1.04\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.88\u0026ndash;1.22\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\" char=\".\"\u003e\n \u003cp\u003e0.6\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1.02\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.86\u0026ndash;1.20\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\" char=\".\"\u003e\n \u003cp\u003e0.8\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e2017\u0026ndash;2018\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1.00\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.79\u0026ndash;1.25\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\" char=\".\"\u003e\n \u003cp\u003e0.97\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.95\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.75\u0026ndash;1.20\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\" char=\".\"\u003e\n \u003cp\u003e0.7\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003eBody mass index, kg/m\u0026sup2;\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.99\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.98\u0026ndash;1.00\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\" char=\".\"\u003e\n \u003cp\u003e0.010\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003eHypertension\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1.28\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1.17\u0026ndash;1.39\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\" char=\".\"\u003e\n \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003eDiabetes\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1.47\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1.35\u0026ndash;1.62\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\" char=\".\"\u003e\n \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003eSmoking status\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eCurrent/Some days\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u0026mdash;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u0026mdash;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eFormer\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.56\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.51\u0026ndash;0.63\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\" char=\".\"\u003e\n \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eNever\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.45\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.40\u0026ndash;0.50\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\" char=\".\"\u003e\n \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n \u003c/table\u003e\n\u003c/div\u003e\n\u003cdiv class=\"gridtable\"\u003e\n \u003cdiv class=\"colspec\" align=\"left\"\u003e\n \u003cp\u003eModel 1 was unadjusted. Model 2 was adjusted for age, sex, race/ethnicity, and NHANES cycle. Model 3 was additionally adjusted for body mass index, hypertension, diabetes, and smoking status. Hazard ratios are presented with 95% confidence intervals. Abbreviations: CI, confidence interval; eGFR, estimated glomerular filtration rate; HR, hazard ratio.\u003c/p\u003e\n \u003cp\u003eParticipants with albuminuria despite preserved kidney function had lower overall survival than all other participants over follow-up.\u003c/p\u003e\n \u003cp\u003eIn survey-weighted sensitivity analyses accounting for the complex NHANES sampling design, the association between albuminuria with preserved eGFR and all-cause mortality remained significant. Weighted hazard ratios were 2.80 (95% CI, 2.43\u0026ndash;3.24; P \u0026lt; .001) in the unadjusted model, 1.92 (95% CI, 1.69\u0026ndash;2.19; P \u0026lt; .001) in the demographic- and cycle-adjusted model, and 1.68 (95% CI, 1.47\u0026ndash;1.93; P \u0026lt; .001) in the fully adjusted model (Supplementary Table 1). These findings support the consistency of the primary pooled analyses.\u003c/p\u003e\n \u003cp\u003eIn the secondary machine learning analysis, the complete-case modeling cohort included 31,060 participants, of whom 2,661 died during follow-up; the training and test sets included 21,742 and 9,318 participants, respectively. All evaluated survival models demonstrated good discrimination on the held-out test set (Table 3). The standard Cox model achieved a C-index of 0.846, while LASSO Cox and Elastic Net Cox achieved 0.847 each; Random Survival Forest showed the highest discrimination (C-index, 0.851).\u003c/p\u003e\n \u003cp\u003eIn the secondary machine learning analysis, the complete-case modeling cohort included 31,060 participants, of whom 2,661 died during follow-up; the training and test sets included 21,742 and 9,318 participants, respectively. All evaluated survival models demonstrated good discrimination on the held-out test set (Table 3). The standard Cox model achieved a C-index of 0.846, while LASSO Cox and Elastic Net Cox achieved 0.847 each; Random Survival Forest showed the highest discrimination (C-index, 0.851).\u003c/p\u003e\n \u003c/div\u003e\n \u003ctable id=\"Tab3\" border=\"1\"\u003e\n \u003ccaption\u003e\n \u003cdiv class=\"CaptionNumber\"\u003eTable 3\u003c/div\u003e\n \u003cdiv class=\"CaptionContent\"\u003e\n \u003cp\u003ePredictive performance of survival models for all-cause mortality in pooled NHANES 2007\u0026ndash;2018\u003c/p\u003e\n \u003c/div\u003e\n \u003c/caption\u003e\n \u003cthead\u003e\n \u003ctr\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eModel\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eC-index\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eBrier score, 60 mo\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eBrier score, 120 mo\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eIBS, 0\u0026ndash;120 mo\u003c/p\u003e\n \u003c/th\u003e\n \u003c/tr\u003e\n \u003c/thead\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eStandard Cox\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\" char=\".\"\u003e\n \u003cp\u003e0.846\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\" char=\".\"\u003e\n \u003cp\u003e0.0429\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\" char=\".\"\u003e\n \u003cp\u003e0.0796\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\" char=\".\"\u003e\n \u003cp\u003e0.0215\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eLASSO Cox\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\" char=\".\"\u003e\n \u003cp\u003e0.847\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\" char=\".\"\u003e\n \u003cp\u003e0.0427\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\" char=\".\"\u003e\n \u003cp\u003e0.0792\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\" char=\".\"\u003e\n \u003cp\u003e0.0213\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eElastic Net Cox\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\" char=\".\"\u003e\n \u003cp\u003e0.847\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\" char=\".\"\u003e\n \u003cp\u003e0.0427\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\" char=\".\"\u003e\n \u003cp\u003e0.0792\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\" char=\".\"\u003e\n \u003cp\u003e0.0214\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eRandom Survival Forest\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\" char=\".\"\u003e\n \u003cp\u003e0.851\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\" char=\".\"\u003e\n \u003cp\u003e0.0414\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\" char=\".\"\u003e\n \u003cp\u003e0.0779\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\" char=\".\"\u003e\n \u003cp\u003e0.0207\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n \u003ctfoot\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"5\"\u003eModel performance was evaluated on the held-out test set. A higher C-index indicates better discrimination, whereas lower Brier scores and integrated Brier scores indicate lower prediction error. Abbreviations: IBS, integrated Brier score.\u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tfoot\u003e\n \u003c/table\u003e\n\u003c/div\u003e\n\u003cp\u003ePrediction error analyses showed the same pattern. At 60 months, Brier scores were 0.0429 for standard Cox, 0.0427 for LASSO Cox, 0.0427 for Elastic Net Cox, and 0.0414 for Random Survival Forest. At 120 months, the corresponding Brier scores were 0.0796, 0.0792, 0.0792, and 0.0779. Integrated Brier score through 120 months was lowest for Random Survival Forest (0.0207), followed by LASSO Cox (0.0213), Elastic Net Cox (0.0214), and standard Cox (0.0215) (Table 3, Fig. 2). Overall, Random Survival Forest performed best, but absolute gains over standard and penalized Cox models were modest.\u003c/p\u003e\n\u003cp\u003eLower Brier scores indicate lower prediction error. Random Survival Forest showed the lowest prediction error across most follow-up time points, whereas standard Cox, LASSO Cox, and Elastic Net Cox performed similarly overall.\u003c/p\u003e\n\u003cp\u003eVariable importance analysis from the Random Survival Forest model identified age, eGFR, and log-transformed urinary albumin-creatinine ratio as the strongest contributors to mortality prediction (Fig. 3). A heatmap of predicted risk further demonstrated higher estimated mortality risk at lower eGFR levels and higher log-transformed UACR values (Fig. 4).\u003c/p\u003e\n\u003cp\u003eThe figure shows the top predictors ranked by variable importance in the Random Survival Forest model. Age, eGFR, and log-transformed urinary albumin-creatinine ratio were the strongest contributors to mortality prediction.\u003c/p\u003e\n\u003cp\u003eThe heatmap shows relative predicted mortality risk across combinations of eGFR and log-transformed urinary albumin-creatinine ratio (UACR), with other covariates fixed at typical values. Higher predicted risk was observed at lower eGFR and higher log-transformed UACR values.\u003c/p\u003e\n\u003cp\u003eOverall, albuminuria despite preserved eGFR was independently associated with higher all-cause mortality in pooled NHANES 2007\u0026ndash;2018, while more complex machine learning approaches provided only limited incremental predictive benefit beyond conventional Cox-based survival modeling.\u003c/p\u003e"},{"header":"Discussion","content":"\u003cp\u003eIn this pooled analysis of NHANES 2007–2018 data linked to mortality follow-up, albuminuria in the setting of preserved eGFR was independently associated with higher all-cause mortality. This association persisted after adjustment for demographic and clinical covariates and remained unchanged in age-group and survey-weighted sensitivity analyses, supporting the consistency of the primary findings. All evaluated survival models demonstrated good predictive performance, with Random Survival Forest achieving the best overall discrimination and lowest prediction error; however, the absolute improvement over standard and penalized Cox models was modest. Together, these findings suggest that albuminuria identifies an increased risk of mortality even when kidney function appears preserved by creatinine-based estimation. To our knowledge, this is one of the few contemporary population-based analyses to specifically evaluate albuminuria in the setting of preserved eGFR while directly comparing conventional and machine learning survival approaches.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eAlbuminuria likely reflects clinically meaningful kidney and vascular injury that may become apparent before a measurable decline in filtration function, which may help explain why the excess mortality risk remained evident even among participants with preserved eGFR. Prior work has shown that albuminuria is associated with adverse outcomes independently of eGFR and traditional cardiovascular risk factors[2]. Filtered albumin may also contribute directly to injury through proximal tubular uptake, inflammatory signaling, and tubulointerstitial fibrosis. More broadly, albuminuria also reflects disruption of the glomerular filtration barrier and systemic endothelial dysfunction[10,11]. In that context, our findings suggest that albuminuria is not simply an accompanying laboratory abnormality, but an early risk marker that may identify higher-risk individuals earlier than creatinine-based kidney function alone. This interpretation is also consistent with our machine learning results, in which both eGFR and log-transformed UACR ranked among the strongest contributors to mortality prediction, suggesting that kidney risk in this phenotype is better captured by combining filtration- and albumin-based information than by either measure in isolation[12,13]. These findings support broader use of albuminuria testing in individuals with preserved eGFR as part of routine risk stratification.\u003c/p\u003e\n\u003cp\u003eAn important additional finding of this study was that more complex survival modeling provided only a limited incremental benefit over conventional regression-based approaches. Although Random Survival Forest achieved the best overall performance, the absolute improvements in discrimination and prediction error over standard Cox, LASSO Cox, and Elastic Net Cox models were small[14]. This pattern suggests that in structured population health data, much of the prognostic signal may already be captured by a relatively small number of strong and clinically interpretable predictors, limiting the added value of more flexible machine learning methods. Similar benchmarking work has shown that no single survival method consistently dominates across datasets, and that classical Cox-based approaches often remain highly competitive for clinical prediction tasks[15–17]. At the same time, the strong contribution of eGFR and log-transformed UACR in the Random Survival Forest model reinforces the importance of kidney-related markers in mortality prediction and supports the view that albumin-based and filtration-based information remain central even in more advanced predictive frameworks.\u003c/p\u003e\n\u003cp\u003eAnother important aspect of our findings is their consistency across multiple analytic approaches. The association between albuminuria with preserved eGFR and all-cause mortality remained similar after multivariable adjustment, replacement of continuous age with age categories, and survey-weighted Cox modeling that accounted for the complex NHANES sampling design. In addition, there was no evidence that the proportional hazards assumption was violated for the primary exposure, supporting the stability of the main association over follow-up.\u003c/p\u003e\n\u003cp\u003eThis study has several limitations. Albuminuria was defined from a single UACR measurement, and because urinary albumin excretion can show substantial within-person variability, some participants may have been misclassified[18]. In addition, current KDIGO guidance emphasizes that confirmation of chronicity requires repeat assessment over time rather than a single abnormal measurement[1]. As an observational analysis, the study cannot establish causality, and residual confounding remains possible despite multivariable adjustment. Finally, although the machine learning models were evaluated on a held-out test set, they were not externally validated in an independent cohort, which limits assessment of generalizability[19,20].\u003c/p\u003e"},{"header":"Conclusion","content":"\u003cp\u003eIn conclusion, albuminuria in the setting of preserved eGFR was independently associated with higher all-cause mortality in pooled NHANES 2007–2018. This association remained consistent across sensitivity analyses, supporting the role of albuminuria as an early marker of risk even when kidney function appears preserved. Although Random Survival Forest demonstrated the best predictive performance, the incremental improvement over conventional Cox-based models was modest. These findings highlight the prognostic importance of albuminuria beyond eGFR and support the continued relevance of conventional survival models for population-level risk stratification.\u003c/p\u003e\n\u003cp\u003e\u003cbr\u003e\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eFunding\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNo funding was received for conducting this study.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConflicts of interest/Competing interests\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors have no relevant financial or non-financial interests to disclose.\u003cbr\u003e\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eEthics approval\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis study used publicly available, de-identified NHANES data with linked mortality follow-up. Therefore, additional institutional review board approval was not required for this secondary analysis.\u003cstrong\u003e\u003cbr\u003e\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConsent to participate\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNot applicable because this study was a secondary analysis of publicly available, de-identified data.\u003cstrong\u003e\u003cbr\u003e\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eWritten Consent for publication\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNot applicable.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAvailability of data and material\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe datasets analysed in this study are publicly available from the National Health and Nutrition Examination Survey (NHANES) website and the NHANES Linked Mortality Files [8,9].\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCode availability\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe analytic code used for data cleaning, statistical analysis, and machine learning modeling is available from the corresponding author on reasonable request.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eLarge Language Model (LLM)\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNo large language model was used for the generation of the study concept, data analysis, results, or scientific conclusions. Any language editing assistance, if used, did not affect the scientific content, and all authors take full responsibility for the final manuscript.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthors' contributions\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eSai Tharun Reddy Gopannagari contributed to the study concept and design, data curation, formal analysis, interpretation of the data, and drafting of the original manuscript. Divya Ganisetti and Kesava Manikanta Achuta contributed equally to the study concept and design, manuscript writing, critical revision, and final editing of the manuscript. All authors read and approved the final manuscript and agree to be accountable for all aspects of the work.\u003cstrong\u003e\u003cbr\u003e\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eLevin A, Ahmed SB, Carrero JJ, et al. Executive summary of the KDIGO 2024 Clinical Practice Guideline for the Evaluation and Management of Chronic Kidney Disease: known knowns and known unknowns. Kidney Int. 2024;105(11):684\u0026ndash;701. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.kint.2023.10.016\u003c/span\u003e\u003cspan address=\"10.1016/j.kint.2023.10.016\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMatsushita K, van der Velde M, Astor BC, et al. Association of estimated glomerular filtration rate and albuminuria with all-cause and cardiovascular mortality: a collaborative meta-analysis of general population cohorts. Lancet. 2010;375(9731):2073. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/S0140-6736(10)60674-5\u003c/span\u003e\u003cspan address=\"10.1016/S0140-6736(10)60674-5\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAstor BC, Hallan SI, Miller ER, Yeung E, Coresh J. Glomerular filtration rate, albuminuria, and risk of cardiovascular and all-cause mortality in the US population. Am J Epidemiol. 2008;167(10):1226\u0026ndash;34. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1093/aje/kwn033\u003c/span\u003e\u003cspan address=\"10.1093/aje/kwn033\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBarzilay JI, Farag YMK, Durthaler J. Albuminuria: an underappreciated risk factor for cardiovascular disease. J Am Heart Assoc. 2024;13(2). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1161/JAHA.123.030131\u003c/span\u003e\u003cspan address=\"10.1161/JAHA.123.030131\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChu CD, Powe NR, Shlipak MG, et al. Albuminuria testing and nephrology care among insured US adults with chronic kidney disease: a missed opportunity. BMC Prim Care. 2022;23(1). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1186/S12875-022-01910-9\u003c/span\u003e\u003cspan address=\"10.1186/S12875-022-01910-9\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChu CD, Xia F, Du Y, et al. Estimated prevalence and testing for albuminuria in US adults at risk for chronic kidney disease. JAMA Netw Open. 2023;6(7):e2326230. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1001/jamanetworkopen.2023.26230\u003c/span\u003e\u003cspan address=\"10.1001/jamanetworkopen.2023.26230\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAlbekery MA, Alhomoud IS, Alabdulathim LS, et al. Underutilization of albuminuria screening in adults with diabetes mellitus or hypertension: a systematic review and meta-analysis. BMC Nephrol. 2025;27(1). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1186/S12882-025-04672-5\u003c/span\u003e\u003cspan address=\"10.1186/S12882-025-04672-5\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eQuestionnaires NHANES. Datasets, and Related Documentation. National Center for Health Statistics. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://wwwn.cdc.gov/nchs/nhanes/\u003c/span\u003e\u003cspan address=\"https://wwwn.cdc.gov/nchs/nhanes/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e Accessed 12 Apr 2026.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNational Death Index | Data Linkage | CDC. Centers for Disease Control and Prevention. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.cdc.gov/nchs/linked-data/mortality-files/index.html\u003c/span\u003e\u003cspan address=\"https://www.cdc.gov/nchs/linked-data/mortality-files/index.html\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e Accessed 12 Apr 2026.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSalmon AH, Satchell SC. Endothelial glycocalyx dysfunction in disease: albuminuria and increased microvascular permeability. J Pathol. 2012;226(4):562\u0026ndash;74. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1002/path.3964\u003c/span\u003e\u003cspan address=\"10.1002/path.3964\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMolitoris BA, Sandoval RM, Yadav SPS, Wagner MC. Albumin uptake and processing by the proximal tubule: physiological, pathological, and therapeutic implications. Physiol Rev. 2022;102(4):1625. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1152/physrev.00014.2021\u003c/span\u003e\u003cspan address=\"10.1152/physrev.00014.2021\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMatsushita K, Coresh J, Sang Y, et al. Estimated glomerular filtration rate and albuminuria for prediction of cardiovascular outcomes: a collaborative meta-analysis of individual participant data. Lancet Diabetes Endocrinol. 2015;3(7):514\u0026ndash;25. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/S2213-8587(15)00040-6\u003c/span\u003e\u003cspan address=\"10.1016/S2213-8587(15)00040-6\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWright EE, Cebrian A, Ngui D. A primary care guide to the screening and pharmacologic management of chronic kidney disease in people living with type 2 diabetes. Clin Diabetes. 2025;43(4):531\u0026ndash;44. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.2337/cd25-0013\u003c/span\u003e\u003cspan address=\"10.2337/cd25-0013\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSimon N, Friedman J, Hastie T, Tibshirani R. Regularization paths for Cox\u0026rsquo;s proportional hazards model via coordinate descent. J Stat Softw. 2011;39(5):1. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.18637/jss.v039.i05\u003c/span\u003e\u003cspan address=\"10.18637/jss.v039.i05\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhang Y, Wong G, Mann G, Muller S, Yang JYH. SurvBenchmark: comprehensive benchmarking study of survival analysis methods using both omics data and clinical data. Gigascience. 2022;11:1\u0026ndash;13. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1093/gigascience/giac071\u003c/span\u003e\u003cspan address=\"10.1093/gigascience/giac071\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003evan Belle V, Pelckmans K, van Huffel S, Suykens JAK. Improved performance on high-dimensional survival data by application of Survival-SVM. Bioinformatics. 2011;27(1):87\u0026ndash;94. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1093/bioinformatics/btq617\u003c/span\u003e\u003cspan address=\"10.1093/bioinformatics/btq617\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHerrmann M, Probst P, Hornung R, Jurinovic V, Boulesteix AL. Large-scale benchmark study of survival prediction methods using multi-omics data. Brief Bioinform. 2021;22(3). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1093/bib/bbaa167\u003c/span\u003e\u003cspan address=\"10.1093/bib/bbaa167\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWaikar SS, Rebholz CM, Zheng Z, et al. Biological variability of estimated GFR and albuminuria in CKD. Am J Kidney Dis. 2018;72(4):538. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1053/j.ajkd.2018.04.023\u003c/span\u003e\u003cspan address=\"10.1053/j.ajkd.2018.04.023\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRiley RD, Ensor J, Snell KIE, et al. External validation of clinical prediction models using big datasets from e-health records or IPD meta-analysis: opportunities and challenges. BMJ. 2016;353. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1136/bmj.i3140\u003c/span\u003e\u003cspan address=\"10.1136/bmj.i3140\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCollins GS, Moons KGM, Dhiman P et al. TRIPOD\u0026thinsp;+\u0026thinsp;AI statement: updated guidance for reporting clinical prediction models that use regression or machine learning methods. BMJ. 2024;385. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1136/bmj-2023-078378\u003c/span\u003e\u003cspan address=\"10.1136/bmj-2023-078378\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"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":"sn-comprehensive-clinical-medicine","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"sncm","sideBox":"Learn more about [SN Comprehensive Clinical Medicine](https://www.springer.com/journal/42399)","snPcode":"42399","submissionUrl":"https://submission.nature.com/new-submission/42399/3","title":"SN Comprehensive Clinical Medicine","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Springer Hybrid","inReviewEnabled":true,"inReviewRevisionsEnabled":false},"keywords":"Albuminuria, Estimated glomerular filtration rate, Mortality, NHANES, Machine learning","lastPublishedDoi":"10.21203/rs.3.rs-9420411/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-9420411/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003ch2\u003eBackground\u003c/h2\u003e \u003cp\u003eAlbuminuria is an established marker of kidney damage and cardiovascular risk, but its prognostic significance in individuals with preserved estimated glomerular filtration rate (eGFR) remains incompletely characterized. We evaluated the association of albuminuria with preserved eGFR and all-cause mortality in pooled National Health and Nutrition Examination Survey (NHANES) 2007\u0026ndash;2018 data and compared conventional and machine learning survival models.\u003c/p\u003e\u003ch2\u003eMethods\u003c/h2\u003e \u003cp\u003eWe performed a retrospective cohort study of adults in the pooled NHANES 2007\u0026ndash;2018 with linked mortality follow-up. Albuminuria was defined as a urinary albumin-creatinine ratio at least 30 mg/g, and preserved kidney function as eGFR\u0026thinsp;\u0026ge;\u0026thinsp;60 mL/min/1.73 m\u0026sup2;. Cox proportional hazards models evaluated the association between albuminuria with preserved eGFR and all-cause mortality. Sensitivity analyses included age-group and survey-weighted Cox models. Secondary analyses compared standard Cox, least absolute shrinkage and selection operator (LASSO) Cox, Elastic Net Cox, and Random Survival Forest models.\u003c/p\u003e\u003ch2\u003eResults\u003c/h2\u003e \u003cp\u003eThe final cohort included 32,290 participants, including 3,090 with albuminuria and preserved eGFR; 2,768 deaths occurred during follow-up. In the fully adjusted pooled model, albuminuria with preserved eGFR was independently associated with higher all-cause mortality (hazard ratio, 1.57; 95% CI, 1.42\u0026ndash;1.73; P \u0026lt; .001). Findings were consistent in age-group and survey-weighted sensitivity analyses. In secondary analyses, Random Survival Forest showed the best predictive performance, but gains over standard and penalized Cox models were modest.\u003c/p\u003e\u003ch2\u003eConclusions\u003c/h2\u003e \u003cp\u003eAlbuminuria, despite preserved eGFR, was independently associated with higher all-cause mortality in pooled NHANES 2007\u0026ndash;2018. Although Random Survival Forest performed best, conventional Cox-based models remained informative for mortality risk stratification.\u003c/p\u003e","manuscriptTitle":"Association of Albuminuria with Preserved eGFR and Mortality in Pooled NHANES 2007–2018: Conventional and Machine Learning Survival Analyses","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-05-04 06:01:07","doi":"10.21203/rs.3.rs-9420411/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"editorInvitedReview","content":"","date":"2026-05-12T03:07:54+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"144018739652755042214888856468376516397","date":"2026-05-08T01:03:15+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2026-04-21T08:53:13+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2026-04-21T01:34:23+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2026-04-20T23:41:15+00:00","index":"","fulltext":""},{"type":"submitted","content":"SN Comprehensive Clinical Medicine","date":"2026-04-15T01:35:11+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"sn-comprehensive-clinical-medicine","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"sncm","sideBox":"Learn more about [SN Comprehensive Clinical Medicine](https://www.springer.com/journal/42399)","snPcode":"42399","submissionUrl":"https://submission.nature.com/new-submission/42399/3","title":"SN Comprehensive Clinical Medicine","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Springer Hybrid","inReviewEnabled":true,"inReviewRevisionsEnabled":false}}],"origin":"","ownerIdentity":"d782bd9e-4593-424f-aae0-cc5d2c974d0e","owner":[],"postedDate":"May 4th, 2026","published":true,"recentEditorialEvents":[{"type":"editorInvitedReview","content":"","date":"2026-05-12T03:07:54+00:00","index":46,"fulltext":""},{"type":"reviewerAgreed","content":"144018739652755042214888856468376516397","date":"2026-05-08T01:03:15+00:00","index":45,"fulltext":""}],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[],"tags":[],"updatedAt":"2026-05-08T19:53:14+00:00","versionOfRecord":[],"versionCreatedAt":"2026-05-04 06:01:07","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-9420411","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-9420411","identity":"rs-9420411","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}