Background Obstructive sleep apnea-hypopnea syndrome (OSAHS) can cause secondary polycythemia and elevated hemoglobin, but the prevalence and predictive value of elevated hemoglobin in OSAHS patients remain unclear.
Methods We conducted a retrospective analysis of 1 035 patients with OSAHS diagnosed by polysomnography who were hospitalized at Sleep Medical Center, First People's Hospital of Yunnan Province from 2018 to 2020. Data of polysomnography, clinical parameters, and comorbidity were compared between 145 cases with polycythemia and 145 cases with normal HGB. The propensity score matching (PSM) was used to balance the baseline variables of the age, sex and BMI. Spearman correlation and multiple regression analysis were used to explore the associated factors of HGB level. The receiver operating characteristic (ROC) curve analysis was used to evaluate the predictive value of elevated HGB for pulmonary hypertension and type Ⅱ respiratory failure in OSAHS.
Results The overall prevalence of polycythemia was 16.8% (174/1 035) in OSAHS patients. Males (25.7%, 157/610) had higher prevalence of polycythemia than females (4.0%, 17/425) (P<0.05) . Patients with a high apnea-hypopnea index (AHI) had higher mean HGB level and prevalence of polycythemia than those with a mild or moderate AHI (P<0.05) . In comparison to patients with normal HGB, those with elevated HGB had higher AHI, longer maximum apnea time (maxAT) and sleep time spent at SpO2<90% (TS90%) , higher values in laboratory and clinical parameters including red blood cell count, HGB, hematocrit, mean corpuscular hemcglobin concentration, aspartate aminotransferase, alanine aminotransferase, uric acid (UA) , total cholesterol, triglyceride, T4, T3, partial pressure of carbon dioxide (PaCO2) , tetraiodothyronine and triiodothyronine, but mean oxygen saturation (MSpO2) , the lowest oxygen saturation (LSpO2) and partial pressure of oxygen (PaO2) (P<0.05) . Higher prevalence of hyperuricemia, proteinuria, and hypercapnia and lower prevalence of hypothyroidism were also seen in those with elevated HGB (P<0.05) . Spearman correlation showed that in patients with elevated HGB, HGB level increased with the increase of BMI, AHI, maxAT, TS90%, PaCO2 and UA, but decreased with the increase of MSpO2, LSpO2 and PaO2 (P<0.05) . Age, BMI, maxAT, MSpO2, LSpO2, PaCO2, creatinine and UA were the factors affecting the level of HGB in multiple regression analysis. ROC curve analysis showed that in male patients, the AUC of HBG in predicting the risk of pulmonary hypertension was 0.699〔95%CI (0.504, 0.893) , P=0.033〕with a cutoff value of 169.5 g/L, and its AUC was 0.836〔95%CI (0.682, 0.989) , P=0.005〕in predicting the risk of type Ⅱrespiratory failure with a cutoff value of 181.5 g/L.
Conclusion The prevalence of elevated HGB was high in patients with OSAHS in Kunming, which may be associated with the severity of hypoxemia and lung ventilation during sleep. Patients with elevated HGB had severer conditions and higher prevalence of comorbidities. Elevated HGB may be a predictor of higher risk of pulmonary hypertension and type Ⅱrespiratory failure in male patients.