Global estimates suggest 15–20 million polio survivors worldwide, with 2.8 million in China (1). Obstructive sleep apnea (OSA) prevalence among this population ranges from 7.3 to 65% (2), highlighting their substantial demographic presence and the growing public health burden of post-polio sequelae. Multiple mechanistic hypotheses have been proposed to explain OSA pathogenesis in polio survivors, including progressive respiratory muscle dysfunction, neuromuscular scoliosis-induced restrictive lung disease, pharyngeal dilator muscle impairment, upper airway narrowing, or combinatorial effects of these factors (3, 4). However, longitudinal evidence regarding OSA progression in this population remains scarce (5, 6).

Most prior studies on OSA in polio survivors have been cross-sectional and clinically based (7–9). This 5-year longitudinal study aimed to characterize OSA progression in a community-recruited cohort of polio survivors. By comparing outcomes with an age-matched general control group, we sought to determine the role of polio sequelae in OSA progression. To our knowledge, this represents the first longitudinal investigation of OSA in this population. We hypothesized that polio sequelae would be associated with accelerated OSA severity changes over the 5-year follow-up period.

This prospective 5-year longitudinal follow-up study was conducted from 2014 to 2019 in Dongyang City, Zhejiang Province, China. The study protocol was approved by the Institutional Review Board of Binzhou Medical University, and written informed consent was obtained from all participants.

Table 1 presents baseline and follow-up descriptive statistics for all participants and those who completed follow-up. ODI₄ increased significantly from 8.11 ± 9.13 to 10.35 ± 11.63 events/h (p = 0.01) (Figure 2a), with 58% of polio survivors showing ODI₄ elevation. Baseline ODI₄ severity distribution (normal: 50%, mild: 38%, moderate–severe: 13%) shifted to 46, 31, and 22% at follow-up (p = 0.027) (Figure 2b), reflecting a statistically significant increase in moderate–severe cases.

Change scores (Δ) were calculated as follow-up minus baseline values. To identify predictors of ODI₄ change, correlation analysis evaluated ΔODI₄ against age, BMI, years since polio onset, scoliosis, and baseline ODI₄ (Table 2). Significant positive correlations emerged between ΔODI₄ and scoliosis (r = 0.27, p = 0.005) and ΔBMI (r = 0.25, p = 0.008).

In linear multivariate regression analyses, the change in ODI₄ was significantly associated with baseline BMI, BMI changes, baseline ODI₄ and scoliosis (Table 3). While other conditions remain constant, compared with patients without scoliosis, the average change in ODI₄ in patients with scoliosis is 5.53 n/h higher.

The 112 survivors who completed follow-up did not differ in age, sex, or BMI from the 33 who underwent PSG follow-up. However, the PSG subgroup had higher baseline ODI₄ (16.62 ± 13.1 vs. 10.35 ± 11.63 events/h, p = 0.009). Table 4 compares baseline and follow-up demographics, anthropometrics, and PSG metrics between polio survivors with PSG follow-up and controls. At baseline, groups were matched for age, sex, BMI, and AHI. The polio group exhibited higher baseline hypopnea index (p < 0.001) and greater proportion of respiratory events during REM sleep (p = 0.005).

In polio survivors, REM sleep latency increased significantly (p = 0.002), with a trend toward reduced REM duration. AHI rose significantly in both groups: polio survivors (26.57 ± 21.25 to 33.86 ± 22.43 n/h, p = 0.02) and controls (27.14 ± 21.91 to 37.24 ± 24.55 n/h, p = 0.004) (Figure 3). Elevated AHI occurred in 73% of polio survivors and 67% of controls. Both groups showed significant increases in maximum apnea duration and total apnea proportion, with slight trends toward reduced SpO₂. Besides, at baseline, the arousal index of polio survivors was already higher than that of the control group. After follow-up, the arousal index in both groups increased significantly.

As shown in Table 5, ΔAHI did not differ between groups (p = 0.89). However, polio survivors exhibited a significant increase in Δ mixed sleep apnea (MSA, p = 0.02) index and ΔREM latency (p = 0.009) compared to controls.

This longitudinal study prospectively evaluated the progression of OSA in patients with polio sequelae using a community-based cohort with a mean follow-up of 5 years. Results demonstrated that ODI₄, AHI, and maximum apnea duration significantly increased over time in this patient population. Notably, changes in ODI₄ were positively correlated with changes in BMI and scoliosis.

Several studies have demonstrated that AHI increases progressively over time and that the progression of OSA is correlated with age, male gender, and BMI in the general population (11–13). Mechanistically, older OSA patients have a higher propensity for upper airway collapse than younger ones due to genioglossus neurogenic degeneration and diminished ventilatory sensitivity to oxygen/CO₂ with aging (14–16). Obesity remains the most critical risk factor for OSA progression, as adipose tissue accumulation in the tongue and pharyngeal regions reduces upper airway caliber and enhances collapsibility during sleep (12, 17). Peri-pharyngeal fat deposition additionally contributes to anatomical narrowing of the upper airway (18, 19). However, in aging polio patients, BMI alone may not adequately predict OSA risk due to muscle atrophy-induced weight loss. A shift from muscle to fat mass in these individuals might instead facilitate OSA development (2), highlighting the need for age- and disease-specific risk assessment metrics.

In patients with polio sequelae, residual neuromuscular impairments may promote OSA through multiple mechanisms. Bulbar involvement contributes to hypopharyngeal collapse via hypotonia of upper airway dilator muscles, predisposing patients to obstructive events during sleep (20). Neuromuscular scoliosis and chest wall deformities alter diaphragmatic mechanics, reducing its efficiency in counteracting upper airway collapse (8). Consistent with our findings, results show that changes in ODI₄ are associated with scoliosis.

Based on the predisposing factors of polio sequelae, we hypothesized that OSA progression in these patients might be more severe than in the general population. However, no statistically significant difference in AHI progression was observed between the polio sequelae group and controls. This may first be attributed to stabilized neurological damage from polio, which does not accelerate its progression. Additionally, our sample comprised community-dwelling patients with relatively mild polio and preserved pulmonary function. We cannot exclude that the follow-up duration was insufficient to detect differences.

Among respiratory events, the dominant types were obstructive apnea and hypopnea, consistent with prior data from Dahan et al. (9) and Hsu et al. (8). At follow-up, the MSA index increased significantly in the polio sequelae group compared to controls. MSA is known to correlate with respiratory control system instability, OSA severity, age, male sex, and obesity (21). Long-term intermittent hypoxia—particularly in severe OSA—may impair respiratory center function through repeated hypoxic stress (22). Thus, the elevated MSA index in this cohort likely reflects worsening OSA. Besides, the high variability in respiratory muscle paresis distribution and bulbar involvement could explain the broad spectrum of SDB pattern observed among polio survivors (4). Further investigation is warranted to clarify these mechanisms.

In the post-polio cohort, we observed a significant increase in REM latency and a trend decrease in REM duration compared to controls, consistent with prior reports of delayed REM sleep onset and reduced REM duration in polio sequelae patients (23). It’s hypothesized that bulbar involvement may indicate additional brainstem poliovirus tropism (particularly in the reticular formation), potentially disrupting REM sleep initiation mechanisms (23). The arousal index was higher in post-polio cohort at baseline. Sliva et al. (24) reported that post-polio patients presented lower sleep efficiency, higher arousal index, higher REM latency and a larger number of stage changes. The increase in the arousal index may be related to factors such as abnormal limb sensations, OSA, and periodic limb movements (PLM). Unfortunately, due to problems like muscle atrophy in this post-polio group, PLM signals could not be accurately identified. Although no significant group difference was found in AHI progression, the elevated proportion of MSA and prolonged REM latency in post-polio patients highlight the characteristics of sleep-disordered breathing compared to controls.

This study has several limitations, including a small sample size, loss to follow-up, and a relatively short follow-up period. Especially a small sample size leads to insufficient statistical power, large sampling errors, poor stability and representativeness of results, making it difficult to detect true differences and limiting extrapolation; it also increases sensitivity to missing data, raises bias risks, and reduces research reliability. Owing to mobility constraints and limited PSG accessibility, the oxygen desaturation index ODI₄ was employed as a surrogate metric. As reported, when incorporating AHI data from PSG, ODI₄ demonstrates the highest accuracy in diagnosing and classifying the severity of OSA during nocturnal SpO₂ analysis (25). Although ODI₄ is less sensitive and specific than the AHI, it reliably tracks longitudinal disease progression. To validate this, we conducted targeted PSG follow-ups where feasible, confirming concordant trends between AHI and ODI₄ trajectories. Besides, although the sample undergoing PSG follow-up did not differ in age, sex, or BMI from the oximetry group, the PSG subgroup had higher baseline ODI₄ (16.62 ± 13.1 vs. 10.35 ± 11.63 events/h, p = 0.009). There may be selection bias in the sample undergoing PSG follow-up. Because patients with more severe symptoms are more willing to complete in-lab PSG monitoring, the PSG subgroup may fail to accurately reflect the overall changing trends due to more severe baseline OSA. While BMI and scoliosis alone may not fully characterize OSA risk in this population, data limitations—such as the absence of neck circumference measurements and electromyography (EMG) assessments—may have affected the study’s validity. And due to the cross-sectional of the control data, comorbidities such as asthma and diabetes, which can affect the progression of OSA, were not matched. These may also lead to bias in the results.

This study demonstrates that OSA-related parameters, particularly mixed apnea and REM alterations, progress differently in polio survivors. Changes in ODI₄ were positively correlated with BMI fluctuations and scoliosis. Despite declining poliomyelitis incidence over recent decades, polio sequelae remain prevalent in aging post-polio populations, posing an escalating challenge for clinical care. Prospective longitudinal follow-up studies with larger samples are needed to inform patient management and optimize treatment strategies.