Sleep disordered breathing a very common disorder with prevalence rates of up to 49% in large epidemiological studies on subjects older than 40 years. A recent study showed that applying CPAP treatment to patients with sleep disordered breathing recruited by their number of apnea and hypopnea events alone, does improve sleepiness but does not improve overall cardiovascular mortality. Based on older large studies however it is knownthat sleep disordered breathing is a cardiovascular risk and that treatment lowers mortality and morbidity. These results appear to be contradictory. However, they might be explained if patient population investigated are carefully reviewed further, and if sleep apnea severity metrics are reconsidered. According to this, it appears that studies speak of different populations. Whereas epidemiological studies use sampled subjects willing to participate, earlier studies used patients contacting a sleep center with complaints and symptoms. In this paper two studies are presented with an assessment of anatomical metrics for upper airway morphology in order to derive parameters for better prediction. Different phenotypes can explain why some people benefit from treatment and others do not benefit equally. Therefore more than just counting apnea and hypopnea events is needed in order to identify patients at risk and patients who have a lower risk when treated. This will require large data set evaluations with hard outcome data.