A recent study found that older men with higher body mass index are more likely to receive a diagnosis of obstructive sleep apnea, but women are accounting for a growing number of new diagnoses.
Older men with high body mass index (BMI) are the demographic most likely to receive a diagnosis of obstructive sleep apnea (OSA), according to a recent study published in Sleep Science, but the number of women with diagnosed OSA has increased over the course of 2 decades. The aim of this study was to analyze long-term trends in characteristics of patients undergoing diagnostic polysomnography (PSG) and subsequently receiving a diagnosis of OSA.
For this study, researchers used data on patients attending the West Australian Sleep Disorders Research Institute (WASDRI) as part of an ongoing large-scale study, the Western Australian Sleep Health Study. The study used data collected from the inception of the study in January 1989 to the most recent available analysis from June 2013.
The participants in the study who underwent diagnostic PSG were all aged 17 years and older. The participants who underwent PSG and received a diagnosis of OSA were split into the OSA subgroup and those who did not have OSA were split into the “no-OSA” group. The current published findings focused on the OSA subgroup.
Different PSG scoring rules were used to define apnea-hypopnea index (AHI), which determines the diagnosis of OSA. For the first period, from January 1989 to July 2002, the American Sleep Disorders Association (ASDA) definition was used. For the second period, from August 2002 to June 2013, the American Academy of Sleep Medicine (“Chicago”) rules were applied. Respiratory monitoring systems also changed during this time, with airflow measured by thermistor prior to September 2002 and more sensitive nasal pressure cannulae and oronasal thermocouple used subsequently.
The researchers also analyzed PSG data of sleep efficiency index (SEI), arousal index (ARI), and percentage of study time with peripheral arterial oxygen saturation below 90% (T90%).
When participants attended their PSG test, their age, sex, height, and weight were recorded, and BMI was calculated. Socioeconomic indexes for area (SEIFA) scores were calculated by the Australian Bureau of Statistics. Comparisons between groups were performed with χ2 tests (categorical), Student t tests (age), or Mann-Whitney U tests for other continuous variables with skewed distributions. Analyses for each testing period were done separately.
Linear models were used for analyses of patient age and SEIFA score and generalized linear models were used for analyses of BMI and sex. Multivariable analyses were used to investigate the OSA risk variables with the diagnosis and severity of OSA for each time period.
Between January 1989 and June 2013, 47,054 participants resided in Western Australia and underwent PSG at WASDRI. Of those, 24,510 received an initial diagnostic PSG and had complete PSG data. The number of patients who underwent a PSG increased from 205 in 1989 to 1104 in 1996. The number fluctuated from 1996 to 2013, with a high of 1399 and a low of 875.
The study participants were primarily older, male, obese, and of relatively high socioeconomic status by Australian standards. Compared with the no-OSA subgroup, participants in the OSA subgroup were older, more obese, and more likely to be male, and had lower SEIFA scores.
The patients undergoing PSG during the Chicago period were more likely to be female compared with the patients of the ASDA period. Median AHI values were 125% higher in the Chicago period compared with the ASDA period. Other PSG results during the Chicago period showed increased OSA severity, with higher ARI and T90% and lower SEI.
The Chicago period had a 32% higher median AHI and participants with a diagnosis of OSA were more likely to have severe cases. Mild OSA decreased (ASDA vs Chicago: 38.4% vs 25.4%), moderate OSA slightly increased (ASDA vs Chicago: 25.0% vs 28.1%), and severe OSA rose significantly (ASDA vs Chicago: 36.6% vs 46.5%). The Chicago period also saw sleep quality worsen but oxygenation improve.
During the study period, 81% of all participants were diagnosed with OSA. There was an increase of proportion of participants with a diagnosis of OSA from the ASDA period (68%) to the Chicago period (91%). Mean age varied from year to year with a range of 48.9 to 51.6 years. There was a significant increase in mean age over time, with a model-predicted change from 1989 to 2013 of an additional 1.91 years. There was also a significant increase of 3.19 kg/m2 in mean BMI over time.
The mean SEIFA of the participants decreased during the ASDA period from 7.4 in 1989 to 6.7 in 2001. There was also an increase in percentage of women who received a diagnosis of OSA, with an increase in percentage of 30.3% from 1989 to 2013. The odds ratio of a woman receiving a diagnosis of OSA in 2013 relative to 1989 was 4.57 (95% CI, 4.10-5.08).
The strongest predictor of OSA diagnosis was obesity, followed by male sex and older age. Year of PSG and SEIFA had little and no effect, respectively.
There were several limitations to this study. The researchers did not have data on referral patterns for OSA diagnosis and could not provide direct information on this trend. The increased PSG capacity provided by sleep services that may have opened during the time of the study is not known. The investigators also did not have data for reason for referral from physicians, nor did they exclude participants with central sleep apnea in their data set.
The researchers concluded that, although the study shows an increase in the number of women getting referred for a PSG test, they have lower odds of receiving an OSA diagnosis overall compared with their male counterparts. BMI was the strongest predictor of AHI in this study, with increasing obesity and age representing other trends in patients with a diagnosis of OSA.
The authors noted that the differences in diagnosing OSA from the ASDA period to the Chicago period could have contributed to an increase in diagnoses. “There are important resource implications for health service delivery when changes in diagnostic methods lead to higher diagnosis rates,” they wrote. “There is a need to carefully evaluate whether associated higher diagnosis rates lead to improved health outcomes.”
Marriott RJ, McArdle N, Singh B, et al. The changing profile of obstructive sleep apnea: long term trends in characteristics of patients presenting for diagnostic polysomnography. Sleep Sci. 2022;15(suppl 1):28-40. doi:10.5935/1984-0063.20210005