In this cross-sectional study of an occupational Japanese population, 29% of subjects were suspected to have SRBD, and increased 3%ODI was significantly associated with an increased risk for having hypertension. However, the association was attenuated after adjustment for obesity status. Noteworthy, the influence of increased 3%ODI on having hypertension was more evident for individuals without obesity compared with those of obesity (p for interaction = 0.03). These findings suggest that the presence of SRBD should be kept in mind even in non-obese individuals.

One previous retrospective study investigated the association between 3%ODI and blood pressure in Japanese truck drivers [9]. In that study, the odds ratio for having hypertension was significantly higher only when 3%ODI was 15 or higher. When stratified by obesity status (BMI 25 kg/m2 or higher), the similar association was observed only in the obese group, without significant interaction. Although the results differ from those of our study, the different definition of hypertension could be one reason for the difference in results. That is, because hypertension was defined as 160/95 mmHg or higher in the previous study, moderate or severe SRBD could be associated with moderate or severe hypertension in that analysis.

Other previous studies that defined hypertension as 140/90 mmHg or higher reported that the odds ratio for hypertension significantly increased according to increased 3%ODI, similar to the present study [2, 3, 5,6,7,8]. In contrast, to the best of our knowledge, there has been no stratified analysis by obesity status, with the exception of the study mentioned above [9]. While obesity has been shown to be associated with hypertension and this association was greater in the younger generation [10], in the present study, significant association was observed between higher 3%ODI and higher blood pressure after adjusting for age. Therefore, SRBD may be significantly associated with hypertension regardless of age.

The results of this study could be clinically important because they suggest that SRBD may be present even in non-obese individuals, and that SRBD may be associated with hypertension with a dose-response relationship. Japanese people have small jaws, suggesting the need to watch for the presence of SRBD even in non-obese Japanese individuals. Epidemiological studies have reported that not only hypertension but also high-normal blood pressure (130–139/85–89 mmHg) is associated with risk of developing cardiovascular diseases, especially stroke [22], so better blood pressure control is important even in mild cases of hypertension. Clinically, for example, we should keep in mind that nocturnal hypertension is common in SRBD [23], which would emphasize the importance of home blood pressure monitoring in order to detect clinical findings suggestive of nocturnal hypertension (early morning hypertension in home blood pressure monitoring).

In contrast, the presence and severity of SRBD was not significantly associated with hypertension in obese individuals in the present study. Many previous studies have reported that obesity itself is a risk factor for developing hypertension, which may mask the influence of higher 3%ODI on hypertension. And the effect of obesity on hypertension has been reported to be greater in younger people [10], suggesting the importance of obesity prevention in young generation. In addition, the mechanisms by which SRBD could be associated with hypertension include not only sympathetic nervous system hyperactivity [24], but also increased insulin resistance [25], activation of the Renin-Angiotensin-Aldosterone (RAA) system [26,27,28], activation of mineralocorticoid receptors [29], oxidative stress [30], and chronic inflammation [31, 32]. These mechanisms overlap with those by which obesity is associated with hypertension, and thus the impact of SRBD on hypertension may not be significant in obese individuals.

The study had several strengths. First, the study enrolled a large occupational population. Second, unlike studies of general populations with a relatively large percentage of elderly, this study included young and middle-aged participants, most of whom were male, and thus was able to examine the association between SRBD and hypertension in young and middle-aged males, which is a problem in Japan. Third, our results were based on reliable data with few missing values, since the participants were required to undergo regular health examinations and simplified PSG.

However, several limitations of this study should be noted. First, selection bias may have occurred because we recruited study participants from among train and bus drivers of a company based in Japan. In addition, the generalizability of the present findings to populations with different genetic backgrounds and lifestyles may have been limited. Second, simplified PSG was used to evaluate SRBD. Full PSG comprehensively assesses biological activities through a variety of measures over the course of a night: electroencephalography, electrocardiography, electromyography, and respiratory curve analysis, along with measurements of snoring, eye movements and oxygen saturation. In contrast, the simplified PSG calculates the 3% oxygen desaturation index (3%ODI: events/h), which is the number of the times when SpO2 decreases by 3% or more per hour. 3%ODI is well known to correlate with AHI and is often used as a screening test [16,17,18,19]. However, it does not record sleep. Therefore, the apnea/hypopnea per hour is calculated using the total recording time as the denominator, not the total sleep time. As a result, the AHI may have been underestimated. The simplified PSG has been reported to have an overall sensitivity of 73% when compared to that of the full PSG, but a sensitivity of 52% for AHI < 30, suggesting that the frequency of mild/moderate SRBD may have been underestimated [18]. On the other hand, this does not affect the results of the dose-response relationship for the association between elevated 3%ODI and hypertension. Third, clinical symptoms were not investigated. For occupational safety reasons, it is unlikely that any of the participants would have daytime symptoms (excessive daytime sleepiness, headache, fatigue, and impaired concentration), but it is possible that some would have nighttime symptoms (subjective symptoms: mid-afternoon awakening and nocturia; other symptoms: apnea and snoring). In the present study, however, we defined SRBD based on objective indices regardless of clinical symptoms, following the ICSD-3 definition. Fourth, the impact of the class of antihypertensive agents on the severity of SRBD was not investigated. RAA system inhibitors and diuretics have been reported to improve AHI on SRBD patients [33]; since lack of data prevented that analysis here, those antihypertensive agents might have affected 3%ODI in the present study. Finally, since this was only a cross-sectional study, the causal relationship between 3%ODI elevation and blood pressure elevation remains unclear.

Asians have been reported to have smaller jaws than other ethnic groups [34]. Further studies, including non-obese Asian individuals are needed to confirm the causal relationship between 3%ODI elevation and blood pressure elevation.