Hypoxic burden linked to sleep apnea severity, inflammation and weaker breathing muscles
In 137 adults evaluated for sleep apnea, total overnight oxygen deprivation was strongly linked to disease severity, higher inflammation and weaker breathing muscles.
Summary
Sleep apnea is usually graded by counting breathing interruptions per hour, but that number says little about how low oxygen drops or for how long. Researchers in Lahore, Pakistan studied 137 adults with suspected sleep apnea using overnight sleep tests and blood work. They measured hypoxic burden, the total oxygen deprivation across the night. It rose sharply from mild to severe disease and showed the strongest association with severity of the measures examined. Inflammation markers such as C-reactive protein rose in step, and breathing muscle strength fell. Weaker inspiratory muscles were independently linked to both greater oxygen deprivation and higher inflammation, even after accounting for body weight, age and sleepiness. Because this was a snapshot in time, it cannot show cause and effect. It does point to oxygen exposure and breathing muscle strength as potentially useful, complementary markers.
Detailed Summary
Obstructive sleep apnea (OSA) is graded by the Apnea-Hypopnea Index (AHI), which counts breathing interruptions per hour. The AHI ignores how deep oxygen levels fall and how long they stay low. Hypoxic burden, the cumulative area under the oxygen desaturation curve, captures both and has been tied more closely to cardiovascular risk. Few studies had examined how it relates to inflammatory biomarkers and to breathing muscle strength, which is why this study was done.
The researchers enrolled 137 adults with suspected OSA (from 156 recruited) across five tertiary sleep centers in Lahore, Pakistan. Everyone had attended overnight polysomnography scored to AASM criteria. Hypoxic burden was computed offline in R: the desaturation area after each apnea or hypopnea was summed and normalized to total sleep time (min·%). Fasting morning blood gave high-sensitivity CRP, IL-6, TNF-α, neutrophil-to-lymphocyte ratio and platelet-to-lymphocyte ratio. Maximal inspiratory and expiratory pressures (MIP, MEP) were measured with a handheld manometer. People with lung disease, autoimmune disease, recent infection or prior OSA treatment were excluded.
Across mild (n=45), moderate (n=46) and severe (n=46) OSA, hypoxic burden rose from 22.0 to 54.0 to 110.5 min·%. It had the largest effect size of any sleep-study variable (η²=0.73, p=0.001), ahead of AHI (0.48), oxygen desaturation index (0.44) and minimum oxygen saturation (0.41). CRP climbed from 3.4 to 8.1 mg/L, IL-6 from 2.6 to 6.2 pg/mL and TNF-α from 3.2 to 6.8 pg/mL. MIP fell from 86 to 65 cmH₂O and MEP from 102 to 78 cmH₂O. Hypertension prevalence rose from 22.2% to 43.5%.
In the prespecified nine-predictor model (adjusted R²=0.47), higher BMI (β=1.35, p=0.002), older age (β=0.25, p=0.024), higher Epworth Sleepiness Scale score (β=0.50, p=0.015) and lower MIP (β=−0.68, p=0.007) were independently associated with greater hypoxic burden. MEP (p=0.026) narrowly missed the Bonferroni-adjusted threshold of 0.025. Physical activity, diet quality, socioeconomic status and medication use were not significant. In the CRP model (adjusted R²=0.58), BMI (β=0.28, p=0.003), hypoxic burden (β=0.045, p=0.004) and lower MIP (β=−0.10, p=0.009) were independent predictors. BMI×MIP interactions were not significant.
The findings support hypoxic burden as a complement to AHI. They also suggest weak inspiratory muscles may be part of the picture linking nocturnal hypoxia and low-grade inflammation, which matters for cardiovascular and metabolic healthspan. Inspiratory muscle strength is potentially modifiable, but this study did not test any intervention.
Several caveats apply. The cross-sectional design cannot show whether weak muscles worsen hypoxia, hypoxia weakens muscles, or both. Participants came from one city, and there was no non-OSA control group. Sex and smoking were not in the primary models, and hypoxic burden was computed with a custom R pipeline on two different polysomnography platforms. IL-6 and TNF-α were not modeled separately. The text supplied cut off before the authors' full discussion, so any further stated limitations are not captured here.
Key Findings
- Hypoxic burden rose from 22.0 to 54.0 to 110.5 min·% across mild, moderate and severe OSA (n=137), with the largest effect size of any sleep-study metric (η²=0.73, p=0.001).
- AHI (η²=0.48), oxygen desaturation index (η²=0.44) and minimum SpO₂ (η²=0.41) showed smaller severity effects than hypoxic burden.
- CRP increased from 3.4 to 8.1 mg/L, IL-6 from 2.6 to 6.2 pg/mL and TNF-α from 3.2 to 6.8 pg/mL from mild to severe OSA.
- MIP fell from 86 to 65 cmH₂O and MEP from 102 to 78 cmH₂O from mild to severe OSA (both p<0.05).
- Lower MIP was independently associated with greater hypoxic burden (β=−0.68, p=0.007), alongside higher BMI (β=1.35), age (β=0.25) and sleepiness score (β=0.50); adjusted R²=0.47.
- CRP was independently predicted by hypoxic burden (β=0.045, p=0.004), BMI (β=0.28, p=0.003) and lower MIP (β=−0.10, p=0.009); adjusted R²=0.58.
- Hypertension prevalence rose from 22.2% in mild OSA to 43.5% in severe OSA (p=0.031).
Methodology
Multicentre cross-sectional study of 137 adults with suspected OSA, recruited consecutively from five tertiary sleep centers in Lahore, Pakistan, with a sample size of at least 120 calculated for a nine-predictor regression. Attended overnight polysomnography was scored to AASM criteria. Hypoxic burden was derived from SpO₂ using an event-based area-under-the-desaturation-curve method, and inflammatory markers and MIP/MEP were measured the next morning. Groups were compared with ANOVA or Kruskal-Wallis tests, with multivariable linear regression for hypoxic burden and CRP. A Bonferroni threshold of p<0.025 was applied across the two primary models.
Study Limitations
The cross-sectional design prevents conclusions about causality or direction between muscle weakness, hypoxia and inflammation. The sample came from one city, had no non-OSA control group, and left sex and smoking out of the primary models. The supplied text ended before the authors' full discussion and conflict-of-interest statement, so those are not reported here.
Enjoyed this summary?
Get the latest longevity research delivered to your inbox every week.
Enter your email to subscribe: