The Role of Obstructive Sleep Apnea and CPAP Therapy in the Functional Hypogonadism of Male Patients With Severe Obesity

62% of severely obese male inpatients had low testosterone; severe sleep apnea was an independent predictor, and CPAP raised T after 3 months in 14 men

Journal: The Journal of Clinical Endocrinology and Metabolism | Published: 2026-04-22 | Type: Cross-sectional study with small longitudinal component | PMID: 41284733 Authors: Amodeo A et al. — University of Milan and Istituto Auxologico Italiano-IRCCS, Italy Funding/COI: Italian Ministry of Health–Ricerca Corrente; conflicts of interest not listed

Summary

Among 204 severely obese men hospitalized at a single Italian center, 127 (62%) had total testosterone at or below 10.4 nmol/L. Decompensated obstructive sleep apnea — defined as an apnea/hypopnea or oxygen desaturation index above 30 events/hour — was independently associated with lower testosterone after controlling for BMI, diabetes, and inflammation. In a 14-patient longitudinal substudy, three months of CPAP therapy raised testosterone levels (p=.009), with improvement correlating with the degree of oxygen desaturation independent of weight change.

Claims

• 127 of 204 patients (62.3%) met the threshold for low testosterone (TT ≤ 10.4 nmol/L)
• Of those, 45 (35.4%) had severe hypogonadism (TT < 6.0 nmol/L)
• 76.6% of hypoandrogenic patients with gonadotropin data had the hypogonadotropic form (LH < 9.4 mU/L)
• 112 of 204 patients had severe/decompensated OSAS (AHI or ODI ≥ 30/hour)
• Multiple linear regression identified four independent predictors of lower TT: BMI (p=.039), type 2 diabetes (p=.006), C-reactive protein (p=.003), and decompensated OSAS (p=.014)
• OSAS diagnosis alone (without severity threshold) was not associated with TT difference (p=.951); only severe/decompensated OSAS drove the signal
• In 14 men started on CPAP, TT rose after 3 months (p=.009); oxygen desaturation index independently predicted improvement regardless of BMI change (p=.04)
• SHBG levels were comparable between severe and non-severe OSAS groups in the 51 patients measured (31.6 vs 32.2 nmol/L, p=.79)

Study Quality

The cross-sectional arm enrolled 204 consecutive inpatients, which is a reasonable sample for this heavily comorbid population, and used multiple linear regression to disentangle OSAS from the many overlapping drivers of low testosterone in obesity. Blood draws were standardized to 8–10am. The severity threshold of 30 events/hour rather than any OSAS diagnosis is methodologically sound — the null finding for OSAS diagnosis alone (p=.951) versus the significant finding for severe OSAS (p=.002) is an honest result that strengthens the specificity of the claim.

The longitudinal CPAP arm is where this paper loses credibility. Fourteen patients is not a trial; it is a pilot observation. There is no control group, no randomization, and no blinding. The p-value of .009 for testosterone improvement in n=14, with no comparator arm, means almost nothing statistically. The authors acknowledge this as a limitation, but the framing in the abstract ("CPAP therapy was shown to improve TT independently of BMI changes") overstates what 14 uncontrolled patients can demonstrate.

Red Flags

• The longitudinal CPAP substudy has n=14 with no control group — the headline finding is built on a case series
• Abstract language ("was shown to improve") overclaims what an uncontrolled n=14 observation supports
• SHBG was available in only 51 of 204 patients, so free testosterone — the clinically relevant fraction — could not be assessed for most of the cohort
• Diagnostic method was not uniform: newly diagnosed patients received polysomnography; previously diagnosed patients received overnight pulse oximetry under CPAP, using a different device and metric (ODI vs. AHI). Comparing these groups introduces measurement heterogeneity
• Single center, single country; the inpatient population with "complicated severe obesity" may not generalize to outpatient or moderately obese men
• Conflicts of interest were not disclosed
• Age range of 18–87 years with no age stratification is a substantial confound; testosterone declines markedly with age independent of obesity

Strengths

• Largest reported cross-sectional study of OSAS and hypogonadism specifically in severe/complicated obesity
• Consecutive enrollment limits selection bias
• Regression model controlled for BMI, diabetes, and inflammation simultaneously, isolating OSAS severity as an independent factor
• Null finding for mild/moderate OSAS is reported transparently rather than buried
• Standardized morning blood sampling for testosterone
• Institutional ethics approval and written informed consent documented

Verdict

The cross-sectional finding — that severe, decompensated sleep apnea independently predicts low testosterone in obese men beyond what BMI and diabetes account for — is methodologically credible and comes from the largest sample assembled for this specific question. That part of this paper is worth reading. The CPAP intervention data is a 14-patient observation with no control group and should be treated as hypothesis-generating only, not evidence that CPAP improves testosterone. The authors' abstract language does not adequately reflect that distinction.