Can pregnant women accurately report snoring? A pilot study

Sleep-disordered breathing (SDB) is a spectrum of breathing abnormalities that occur during sleep (1), ranging from primary snoring to obstructive sleep apnoea (OSA) (2). OSA is the most serious form of SDB, with repeated occurrences of partial (hypopnoeas) or complete (apnoeas) upper airway collapse (3). OSA also impacts on daytime performance including experiencing daytime sleepiness and fatigue (4). When snoring occurs, but is not associated with apnoeas and hypopneas during sleep and the individual has no daytime complaints, then this is known as ‘simple snoring’ (5). Snoring is quite common, it is estimated that approximately 40% of adult men and 20% of adult women are habitual snorers. Snoring is an audible indicator of increased upper airway resistance and as such is a sign of the presence of some degree of SDB (6).

Snoring in pregnant women can go unnoticed and this may also mean that SDB may go undiagnosed (7). This is problematic, because SDB is thought to be associated with poor pregnancy outcomes, such as foetal growth restriction, gestational diabetes and hypertension (8), especially if the SDB begins during pregnancy (7). SDB can also cause sleep disturbance, which may, in turn, increase negative outcomes such as preterm birth (9).

Accurate screening for SDB in pregnancy is important, because screening leading to treatment could assist in reduction of negative outcomes (10). That said, at present there is little evidence to support universal screening for SDB in pregnancy (11). This is because it has not yet been determined what degree of SDB may be clinically relevant. An Apnoea-Hypopnoea Index (AHI) of greater than 5 is used in many adult populations as the starting point for concern (12). However, it is not yet known if a pregnant woman should have a lower or higher alarm threshold (7, 11). Furthermore, screening indicators for SDB that are used in non-pregnant populations, such as body mass index and neck circumference (13) may not be appropriate in pregnancy due to normal physiological weight gain during this time.

Finally, several of the self-reported screening tools used in the general population, such as the Epworth Sleepiness Scale (ESS), ask about level of day time sleepiness which may be resulting from SDB (4). However, pregnant women are often sleepy during the day due to hormonal changes and the physical demands of the pregnancy, meaning that this type of screening can also be unreliable in pregnancy especially if used in isolation.

Logic would suggest that a combination of screening tools for SDB may be useful in pregnancy. Work by Facco and others in 2012 (11) as well as Wilson and colleagues in 2013 (10) indicate that combining a range of objective (overnight sleep study, BMI) and self-reports of snoring may be the most useful way of screening for SDB in pregnancy, however, little is known regarding which combination of screening measures used in the general adult population may be the most useful in pregnancy, hence, further exploration and validation is necessary.

This pilot study aimed to investigate whether simply asking the woman if she snores as a basic screening tool for the presence of SDB is valid. Questions about snoring and other symptoms of SDB from the Basic Nordic Sleep Questionnaire (BNSQ) (14), and the Multivariable Apnea Risk Index (MAP Index) (15) were used. Additionally body mass index (BMI) and neck circumference were measured. Prospective studies show an association between increasing BMI and/or neck circumference and with an increased risk of SDB in both the general (13, 16) and pregnant populations (17, 18). We therefore included these measures in our analysis to provide further exploration of their value for identifying women at risk of SDB in pregnancy.

Ten participants (71%) were identified as snorers by video-scoring. BNSQ and MAP reports of snoring were concordant with video-determined snoring in ≤ 50% of participants (n = 7). The Watch-PAT 200 dB measure of snoring was concordant with video-determined snoring in 35% (n = 5) participants (Figure 1, left). Specificity was 86% for MAP > 2 and 76% for BNSQ > 2 compared to video-determined snoring. Sensitivity, specificity, PPV and NPV were ≤ 55% for all other comparisons to video-determined snoring (Figure 1, right).

Figure 1.

Four participants (27%) had an RDI greater than 10 and/or AHI greater than 5. While the correlation between RDI and AHI was r = 0.83, one person had an RDI of 12.8 and an AHI of 4.8, and another had an RDI of 9.6 and an AHI of 7.0. Highest concordance between self-reports and RDI/AHI measures was for BNSQ > 2 (n = 8, 53%) (Figure 1, left). While sensitivity and NPV were ≥ 67% for self-reports compared to RDI > 10 and AHI > 5, specificity and PPV were ≤ 45% (Figure 1, right).

There were no significant differences in BMI or neck circumference in snorers compared to non-snorers, or those with indicators for SDB compared to those without (Table 2).

This study aimed to determine if self-report tools in late pregnancy may be used as a screen for the presence of SDB. More than half of our participants (n = 10) were identified as snorers while much fewer (n = 4) met the current cut off for diagnosis of mild SDB (AHI ≥ 5). Investigation of concordance and predictive value of self-report measures compared to standard video-scoring and Watch-PAT 200 determined SDB indicators suggests that self-reports may not provide an accurate assessment of SBD symptoms in late pregnancy. This is consistent with previous work finding low-moderate predictive capacity of MAP index measures compared to PSG for predicting SDB in pregnancy (10). Self-report in this study, resulted in an underestimation of the number of participants who experienced SDB symptoms.

Combining self-report with objective measures such as BMI and neck circumference when screening for SDB has been validated within general adult populations (25, 26). However, there were no differences in BMI or neck circumference in snoring or potential SDB groups in our study. It may be that these measures are not as reflective of SDB risk during pregnancy as it is normal for pregnant women to gain weight, which may also result in increase in neck circumference (27, 28). Interestingly, there has been recent work towards development of a model using three factors, obesity (BMI ≥ 32), snoring volume and tiredness upon wakening, to accurately predict the presence of SDB in pregnancy (10). While our study suggests that self-reports of snoring and SDB symptoms may not be accurate in this population, perhaps combining them with other measures such as tiredness upon waking, may be more effective.

The snore sensor of the Watch-PAT 200 was sensitive to ambient noise other than snoring (for example: television, fans in the room and outside traffic) and this issue may have meant all ‘snore’ data recorded by the Watch-PAT 200 was unreliable. Indeed, there was a relatively poor concordance between the Watch-PAT 200 measure of snoring and video-determined snoring.

This was a pilot study, with a small sample size. However, our study lends weight to other studies which found poor predictive value of the scales used in the general population to detect SDB in pregnancy. Future study is therefore needed to validate screening tools, which may need to be a combination of measures. Such a tool would need to take account of the normal physiological changes of pregnancy but also should be administered in pregnancy early enough to allow diagnosis of SDB with enough time for any intervention (such as CPAP) to be effective in improving maternal and foetal outcom