“We found a dose-response association between sleep-disordered breathing at base line and the presence of hypertension four years later that was independent of known confounding factors.”