Patients commencing continuous positive airway pressure therapy (CPAP) unde
rgo overnight airway pressure titration in sleep centres to optimize breath
ing and sleep patterns. We tested the hypothesis that data from formal scor
ing of the sleep and breathing patterns observed at the best achievable pre
ssure during titration can predict CPAP use and effectiveness, as our clini
cal experience suggested otherwise. The relationship between CPAP titration
scores (apnoea/hypopnoea frequency, arousal frequency and sleep staging) a
nd subsequent CPAP use was examined in 150 sleep apnoea/hypopnoea syndrome
patients. One hundred patients were continuing CPAP therapy and 50 were ran
domly selected patients who had discontinued CPAP. Within the CPAP group, t
itration scores were compared with CPAP machine use, subjective daytime sle
epiness and requirements for airway pressure adjustment. Respiratory irregu
larities and arousals during titration did not relate to outcome. Sleep-sta
ge analysis revealed a weak relationship between more wakefulness during ti
tration and CPAP discontinuation (P = 0.02). There was a correlation betwee
n more prolonged Stage 4 sleep during titration and reduced sleepiness on e
stablished therapy (P = 0.002), but this explained less than 12% of the var
iance. The absence of rapid eye movement sleep during titration was not ass
ociated with poorer outcomes. We conclude that routine scoring of breathing
and sleep patterns observed during CPAP titration is of little clinical va
lue, as the results do not predict outcome for individual patients. Satisfa
ctory CPAP therapy may be established even if significant numbers of apnoea
s/hypopnoeas or arousals are observed at the optimal pressure during titrat
ion.