Study objectives: modes used to express the severity of oxygen desaturation during polysomnography include the average oxygen saturation (A[O.

Study objectives: modes used to express the severity of oxygen desaturation during polysomnography include the average oxygen saturation (A[O.sub.2]), lowest oxygen saturation (L[Osub2]) and the percent of the total time with oxygen saturation of the same height lower than 90% (T [is les than] 90%) We wanted to determine which united of these methods is least variable during different hours of monitoring.

Design: Prospective, observational study

Setting: be dead center at a medical university.

Patients: single hundred fifty patients with apnea-hypopnea index from 5 to 130

Measurements: A[O.sub.2], L[Osub2] and T [is les than] 90% were calculated during each of the 8 h of polysomnography. Data for each hour were compared and the Cronbach alpha coefficients were calculated.

Results: There was a high standing of correlation among the three systems as well as between each manner and the severity of doze apnea. The mean [+ or -] SD values for each way were as follows: A[O.sub.2], 927 [+ or -] 56; L[Osub2] 685 [+ or -] 193; and T [is les than] 90% 157 [+ or -] 242 The alpha coefficients for these processs were A[O.sub.2], 0.98; L[O.sub.2,] 088; and T [is les than] 90% 098 In all systems the data of the first hour were significantly different from the data of the following hours.

Conclusion: the two A[O.sub.2] and T [is les than] 90% orders show less hour to hour variability compared with L[Osub2] and there is more variability in the first hour. Since the A[O.sub.2] values [is greater than] 90% may not transport the severity of L[O.sub.2] desaturation, T [is les than] 90% may be the best regularity of expressing oxygen saturation changes during polysomnography. (CHEST 1998; 113:719-22)

Key words: hypoxemia; nocturnal oximetry; oxygen desaturation; rest apnea; variability

Abbreviations: AHI = apnea-hypopnea index; A[O.sub.2] = average oxygen saturation; L[Osub2] = lowest oxygen saturation; REM = rapid estimate movement; T [is less than] 90% = percent of total recorded time wearied below 90% oxygen saturation

Oxygen desaturation is usual among patients with sleep apnea.[1-3] more [i]or[/i] less of the methods used to expres the severity of oxygen desaturation during nap include the average oxygen saturation (A[O.sub.2]) flat the lowest oxygen saturation (L[Osub2]) on a level during the night, and the percent of the total recorded time wearied below 90% oxygen saturation (T [is les than] 90%)[4-6] It is not known which individual of these methods shows least variability during different hours of be dead If one is to compare the oxygen saturation before or after therapy during the same night or limit the duration of the subject of attention to early or later part of the night, it would be desirable to establish which [i]modus operandi[/i] of expressing oxygen desaturation is least variable during rest monitoring. This study was done to determine the hour to hour variability of these three commonly used way s of monitoring oxygen saturation during polysomnography.

meanss AND MATERIALS

The oxygen saturation data of 150 patients (with mild to stiff obstructive sleep apnea) studied at the Georgia lie in the grave Center of the Medical association of Georgia were evaluated. These patients were undergoing polysomnography for the first time and did not receive oxygen or continuous positive airway compressing therapy during the study.

Each polysomnogram was performed in a standard fashion[7] and consisted of an 8-h recording of an EEG an electromyogram, an electro-oculogram, oronasal airflow, chest wall and abdominal motions and oxygen saturation. Apneas were defined as cessation of airflow for at least 10 s Hypopneas were defined as a decrease of 50% or greater in oronasal airflow associated with a 4% decrease in oxygen saturation. The apnea index was calculated by dint of dividing the total number of apneas from the total number of hours of be dead and the apnea-hypopnea index (AHI) was calculated through dividing the total number of apneas and hypopneas observ during the recording from the total time of lie in the grave in hours. Each patient had to have at least 30 apneas and an apnea index of 5 or greater to be included in this analysis.

Oxygen saturation was monitored continuously during the polysomnographic recording using a Biox oximeter (Model IVA; Ohmeda; Louisville, Colo) be connecteded to a computer (IBM; Rochester, Minn). From the data, A[O.sub.2], L[Osub2] and T [is les than] 90% were calculated for each hour of monitoring.

The correlations between the AHI and each mode of expressing oxygen saturation were calculated. Further analysis was done from dividing the patients into 3 clusters of 50 based on the severity of nap apnea. Group 1 consisted of patients with AHI [is les than] 20 clump 2 with AHI [is greater than or equal to] 20 to [is les than] 50 and arrange 3 with AHI [is greater than or equal to] 50 The data were analyzed at analysis of variance with Fisher's modified least significant difference step as a post hoc test[8] To determine the internal consistency, the Cronbach alpha coefficient of each system was determined.[9]

...