Exploring Sleep Quality and Chronic Rhinosinusitis Risk Using Wearable Technology

Chronic rhinosinusitis (CRS) is a prevalent sinonasal inflammatory disorder with substantial impacts on patient quality of life [1]. CRS pathogenesis involves interactions between immune responses, environmental factors, and behavioral interventions [1]. Sleep has emerged as a possible contributor, as it regulates mucosal immunity, epithelial barrier integrity, and inflammation [2]. Disrupted sleep can heighten cytokine production and weaken mucosal defenses, suggesting sleep disturbance may play a role in CRS onset [3]. An epidemiologic study linked insomnia symptoms to higher subsequent CRS risk, though it relied on subjective sleep measures [4]. This study leverages longitudinal, wearable-derived metrics to evaluate sleep as a potentially modifiable risk factor in CRS pathogenesis. We used Fitbit-derived sleep data from participants in the All of Us (AoU) Research Program. Sleep metrics included nightly sleep duration and stages (light, deep, REM). These stages are estimated by Fitbit's algorithms using heart rate and movement [5]. Three quality metrics were derived, including sleep irregularity (standard deviation of monthly average sleep duration), sleep efficiency (time asleep relative to time in bed) and REM-to-non-REM ratio (REM relative to light and deep sleep). Patient inclusion criteria are detailed in Figure 1. The outcome was an incident diagnosis of CRS using ICD-10 codes (J32.X) in the electronic health record (EHR). Participants with CRS diagnoses within the first 180 days were excluded to reduce reverse causation. Sleep patterns were aggregated monthly to capture habitual sleep behaviors. Monthly averages were then categorized into tertiles. Time-varying Cox hazard regression models estimated hazard ratios (HRs) and 95% confidence intervals (CIs) for incident CRS codes in the EHR by comparing sleep periods in the higher tertiles of each sleep metric with the lowest tertile. The model was adjusted for demographic, socioeconomic, and clinical covariates (Table 1), with missing data handled by predictive mean matching. Analyses were conducted in Python on the AoU Researcher Workbench. A final cohort of 9831 patients with 10,131,340 person-nights was included in our study (Figure 1). About 332 (3.20%) participants had a diagnosis of CRS at the beginning of the study period. Demographic and comorbidity variables associated with CRS were incorporated into the adjusted Cox regression model (Table 1). Sleep quality metrics indicated several significant findings. Greater sleep irregularity was associated with an increased risk of receiving a CRS diagnosis. Participants in the highest (1.5–7 h) and moderate (1–1.5 h) irregularity tertile had a significantly higher risk compared with those with < 1 h (HR 1.79, 95% CI 1.73–1.86, and HR 1.33, 95% CI 1.29–1.38, p < 0.005, respectively). Higher sleep efficiency was associated with a lower risk of receiving a CRS diagnosis. Participants with high (> 89.4%) and moderate (86.7%–89.4%) sleep efficiency had a lower risk compared with those < 86.7% (0.81, 0.78–0.84, and 0.97, 0.94–0.99, p = 0.03, respectively). Higher REM-to-non-REM ratios were associated with a reduced risk of CRS diagnosis. Participants with high (29%–85%) and intermediate (23%–29%) ratios had a lower risk compared with those with < 23% REM proportion (0.74, 0.72–0.77, and 0.78, 0.76–0.81, p < 0.005, respectively) (Table 1). This longitudinal population‑based study suggests that poor sleep quality, characterized by greater sleep irregularity, reduced sleep efficiency, and lower REM‑to–non‑REM proportions, may be associated with an increased risk of a diagnosis of CRS. Published literature links impaired sleep with chronic inflammatory conditions, for example, obesity, cardiovascular disease, and diabetes [6]. Although few studies have examined sleep disturbances and CRS incidence [4], inflammatory mechanisms associated with poor sleep are consistent with known CRS pathophysiology. We observed a graded increase in CRS risk across groups with higher sleep irregularity, lower efficiency, and reduced REM sleep. These findings align with evidence that circadian rhythm disruption promotes chronic inflammation [2]. Sleep irregularity alters circadian timing, and circadian dysregulation can influence pathways relevant to CRS [7]. Circadian‑regulated genes (e.g., NFIL3) modulate cytokine suppression, and disrupted nocturnal sleep may enhance Th2 activity [8]. Animal models demonstrate that circadian misalignment amplifies Th2 and Th17 responses in allergic airway disease, implicating a mechanistic link for CRS [3]. Clinical studies also report higher rates of inflammatory airway disorders among individuals with chronic sleep conditions [9]. Moreover, circadian regulation affects epithelial barriers, leukocyte trafficking, and cytokine expression [2, 3]; persistent disruption of these cues may shift acute inflammation toward the chronic, tissue‑remodeling phenotype characteristic of CRS. This study has limitations. To ensure sufficient sample size for analysis, given the limited number of CRS cases in the AoU database, we included individuals with a single ICD‑10-coded CRS diagnosis. Coding without imaging confirmation thus introduces potential misclassification. We adjusted for demographic and clinical confounders in the Cox model; however, utilization‑driven coding and diagnostic overlap cannot be excluded. The cohort, limited to individuals who use wearable devices, may not represent the broader population. Device‑related noise (e.g., motion artifact, placement, and fit) may affect sleep‑metric accuracy. Although we excluded participants diagnosed with CRS within the first 180 days of Fitbit monitoring, undiagnosed or subclinical disease may have influenced sleep patterns; thus, residual reverse causation cannot be ruled out. To our knowledge, this is the first study to examine the association between wearable‑derived sleep quality and CRS. Greater sleep irregularity, lower efficiency, and reduced REM proportion were associated with increased risk of CRS diagnosis. Basic science studies and prospective trials should clarify causal pathways, refine risk prediction, and evaluate sleep‑related interventions in CRS. The Houston Methodist Department of Otolaryngology Head and Neck Surgery expresses its sincere gratitude to Willis and Patricia Johnson for their generous private donation in support of research within the department. The authors have nothing to report. Michael T. Yim is a consultant for Aerin Medical and Chitogel. Masayoshi Takashima is a consultant for Neurent Medical, LivaNova and Medtronic ENT. Omar G. Ahmed is a consultant for Aerin Medical, LivaNova and Medtronic ENT. The other authors declare no conflicts of interest.