关键词: 睡眠呼吸暂停, 阻塞性；潜在类别分析；临床亚型；持续气道正压通气

Abstract: Background　Obstructive sleep apnea（OSA）is a highly heterogeneous disorder，and different subtypes suggest different characteristics.In the previous study，we found that there was significant diversity of initial exposure to continuous positive airway pressure（CPAP）therapy among OSA physiologic phenotypes.However，a detailed description for this problem is lacking in current at home and abroad.Objective　To use latent class analysis to explore the clinical subtypes of OSA patients，and to analyze the differences of diagnostic polysomnography（PSG）before and after initial exposure to CPAP therapy among different subtypes.Methods　The study retrospectively analyzed data for 491 OSA patients who underwent initial exposure to CPAP therapy at the Tangdu Hospital Sleep Medicine Center from January 2011 to December 2016.22 kinds of clinical symptoms：difficulty falling asleep，multiple awaking，early awaking，snoring，apnea，sudden awaking with breath holding，nightmare，dreamtalk，enuresis，limb twitching during sleep，morning headache，dizziness，anxiety，depression，cognition dysfunction，lack of concentration，confusion，impaired reaction time，work impairment，palpitation，fatigue and daytime sleepiness were included in the latent category analysis model.Then the symptoms distribution，sleep characteristics of diagnostic PSG and responses to initial CPAP therapy among different subtypes were compared.Results　A model with four phenotype clusters provided the best fit：cluster 1 "apnea with daytime sleepiness group"（143 cases，29.1%），cluster 2 " minimally symptomatic group "（189 cases，38.5%），cluster 3 "severely impaired daytime function group"（64 cases，13.0%）and cluster 4 "combined with insomnia group"（95 cases，19.3%）.The average age of cluster 2 " minimally symptomatic group " was younger and have a higher proportion of men，while the opposite result of cluster 4 "combined insomnia group".Statistical differences were observed among all the OSA symptoms（P<0.05），with the exception of apnea.There were statistically significant differences for sleep latency（SL），N1%，N2%，N3%，REM% of diagnostic PSG among four subtypes（P<?0.05）.These differences for total sleep in time（TST），sleep efficiency（SE），N1%，N2%，N3%，REM%，wake after sleep onset（WASO），oxygen desaturation index（ODI），time spent at pulse oxygen saturation（SpO2）below 90%，the lowest SpO2（LSpO2），average SpO2 after initial exposure to CPAP treatment were also statistically significant（P<?0.05）.Conclusion　The clinical symptoms of OSA can be divided into four clinical subtypes by latent class analysis，and initial exposure to CPAP treatment of four clinical subtypes is statistically different.These findings emphasize the need to identify a series of OSA clinical symptoms and carry out individualized treatment.

Key words: Sleep apnea, obstructive；Obstructive sleep apnea；Latent class analysis；Clinical subtype；CPAP