Background Evidence linking OSA with hypertension in population studies is conflicting. We examined longitudinal and cross-sectional associations of previously unrecognized OSA, including OSA occurring in rapid eye movement (REM) sleep, with hypertension. Methods The Men Androgens Inflammation Lifestyle Environment and Stress (MAILES) study is a longitudinal study of community-dwelling men in Adelaide, South Australia. Biomedical assessments at baseline (2002-2006) and follow-up (2007-2010) identified hypertension (systolic ≥ 140 mm Hg and/or diastolic ≥ 90 mm Hg, or medication) and risk factors. In 2010 to 2011, 837 men without a prior diagnosis of OSA underwent full in-home unattended polysomnography of whom 739 recorded ≥ 30 min of REM sleep. Hypertension at follow-up (concomitant with OSA status) was defined as prevalent hypertension. Recent-onset hypertension was defined as hypertension at biomedical follow-up (56 months mean follow-up range, 48-74) in men free of hypertension at baseline. Results Severe REM OSA (apnea hypopnea index ≥30/h) showed independent adjusted associations with prevalent (OR, 2.40, 95% CI, 1.42-4.06), and recent-onset hypertension (2.24 1.04-4.81). Significant associations with non-REM AHI were not seen. In men with AHI < 10, REM OSA (apnea hypopnea index) ≥ 20/h was significantly associated with prevalent hypertension (2.67 1.33-5.38) and the relationship with recent-onset hypertension was positive but not statistically significant (2.32 0.79-6.84). Similar results were seen when analyses were confined to men with non-REM AHI < 10. Conclusions In men not considered to have OSA (AHI < 10), hypertension was associated with OSA during REM sleep. REM OSA may need consideration as an important clinical entity requiring treatment but further systematic assessment and evidence is needed.