Objective
We examined the risk of adverse pregnancy outcomes, including low birthweight (LBW), preterm birth, small for gestational age (SGA), cesarean section (CS), low Apgar score (at 5 minutes after delivery), and preeclampsia in pregnant women with and without obstructive sleep apnea (OSA).
Study Design
Our subjects included 791 women with OSA and 3955 randomly selected women without OSA. We performed conditional logistic regression analyses to examine the risks of adverse pregnancy outcomes between women with and without OSA.
Results
Compared with women without OSA, adjusted odds ratios for LBW, preterm birth, SGA infants, CS, and preeclampsia in women with OSA were 1.76 (95% confidence interval [CI], 1.28–2.40), 2.31 (95% CI, 1.77–3.01), 1.34 (95% CI, 1.09–1.66), 1.74 (95% CI, 1.48–2.04), and 1.60 (95% CI, 2.16–11.26), respectively.
Conclusion
Pregnant women with OSA are at increased risk for having LBW, preterm, and SGA infants, CS, and preeclampsia, compared with pregnant women without OSA.
Obstructive sleep apnea (OSA), a common sleep-related breathing disorder, is characterized by recurrent collapse or blockage of the pharynx during sleep that causes intermittent cessation of airflow and a hallmark snoring-gasping pattern. The prevalence of OSA among women ranges from 0.3% to 5 %. A study by Loube et al, based on self-reports, found that frequent snoring is reported more often in pregnant women than in nonpregnant women. However, the incidence of OSA in pregnant women is unknown.
See Journal Club, page 176
Patients with OSA commonly had decreased quality of life. In particular, OSA affects sleep quality and duration of sleep in pregnant women. Pregnancy causes anatomic, physiologic, and endocrinologic changes, including narrowing of the upper respiratory tract, which may increase the risk for OSA or worsen preexisting sleep apneas. Studies have associated OSA in pregnant women with low birthweight (LBW), preterm birth, small for gestational age (SGA), cesarean section (CS), lower Apgar scores at birth, and preeclampsia.
On the other hand, an empirical study by Loube et al reported no association between mothers with frequent snoring and LBW infants (mean birthweights were 3534 ± 474 g and 3450 ± 652 g for women with and without OSA, respectively). Furthermore, previous studies on OSA and pregnancy outcomes were limited to case reports and selective data or small sample sizes ; furthermore, all the studies reported on subjects from Western countries. Therefore, whether there was an association between OSA and adverse pregnancy outcomes remains unanswered.
Using 2 large-scale, nationwide, population-based datasets, this study aimed to examine the risk of adverse pregnancy outcomes, including LBW, preterm birth, SGA, CS, lower Apgar score (at 5 minutes after delivery), and preeclampsia/eclampsia, between pregnant women with and without OSA in Taiwan. The large dataset available from Taiwan presents an exceptional opportunity to examine this issue among Asian women.
Materials and Methods
Database
We linked 2 nationwide, population-based datasets and used the resulting compilation for our analyses. In March 1995, Taiwan initiated its National Health Insurance (NHI) program to finance health care for all its citizens. Taiwan’s NHI has a unique combination of characteristics: universal coverage, a single-payer payment system with the government as the sole insurer, unrestricted access to any medical institution of the patient’s choice, and a wide variety of providers including primary care physicians.
The Taiwan National Health Insurance Research Dataset (NHIRD) is derived from the NHI program and includes all the original claims data as well as registry files of contracted medical facilities, board-certified specialists, other medical service providers, and prescriptions covered by the program for the 25.68 million enrollees in Taiwan (the coverage rate was greater than 98.5% in 2007). Therefore, the NHIRD includes comprehensive information on the medical utilization of virtually all the pregnant women in Taiwan and thus offers an excellent opportunity to examine the relationship between OSA and pregnancy outcomes.
The national birth certificate registry is maintained and publicly released by the Taiwanese Ministry of the Interior. According to law, all births in Taiwan must be registered within 10 days following the birth. This dataset contains both infants’ and parents’ demographic, reproductive, and socioeconomic characteristics and infants’ birth characteristics, including birthweight, gestational age, birth order, and sex. A previous study has verified the completeness and showed high levels of validity in Taiwan’s birth registry.
These 2 nationwide, population-based datasets were linked with assistance from the Bureau of Health Promotion, Department of Health, Taiwan. Because the NHIRD consists of deidentified secondary data released to the public for research purposes, this study was granted approval via summary review by the institutional review board.
Study sample
This cross-sectional design includes a study group and a comparison group. To form the study group, we first identified 218,776 women in Taiwan who had live singleton births between Jan. 1, 2005, and Dec. 31, 2005. If the selected women had more than 1 singleton birth during the study period, we included only the first in the study sample and designated it an index delivery. Of the 218,776 women, 791 had been diagnosed with OSA ( International Classification of Diseases , Ninth Revision , Clinical Modification [ICD-9-CM] codes 780.51, 780.53, 780.57, or 327.23) after receiving polysomnograms during ambulatory care visits within 1 year prior to their index deliveries.
When a physician suspects that a patient has OSA, the physician may give the patient a tentative diagnosis of OSA during their first visit to perform the related clinical or laboratory tests to confirm the OSA diagnosis and avoid any possible fines for performing unnecessary or inappropriate procedures. Therefore, we selected only women who had been given at least 2 consensus OSA diagnoses after undergoing polysomnographic studies to increase coding reliability and validity from this administrative database.
To form the comparison group, we randomly extracted 3955 women (5 women for every woman with OSA) matched with the study group in terms of age group (<20, 20-24, 25-29, 30-34, and ≥35 years) using the SAS surveyselect procedure (SAS System for Windows, version 8.2; SAS Institute Inc, Cary, NC). We also assured that selected women in the comparison group had never received a diagnosis of OSA since the initiation of the NHI program in 1995.
Variables of interest
The independent variable for this study was whether each woman was diagnosed with OSA within 1 year prior to her index delivery. The outcome variables selected for this study were all dichotomous. They included LBW (<2500 g); preterm gestation (<37 completed weeks of gestation); SGA babies (SGA has been defined as a birthweight of less than the 10th percentile for gestational age by Lubchenco et al and Battaglia and Lubchenco in the 1960s); Apgar score at 5 minutes less than 7; CS; preeclampsia (eclampsia); gestational diabetes; and gestational hypertension.
We also adjusted for several maternal characteristics (highest educational level, marital status, geographic region, coronary heart disease [ICD-9-CM codes 410-414 or 429.2], anemia [ICD-9-CM codes 280-285], hyperlipidemia [ICD-9-CM codes 272 and 272.0-272.9], and obesity [ICD-9-CM codes 278, 278.0, 278.00, and 278.01]); infant sex and parity; and father’s age in the regression modeling to assess the independent effect of OSA on the specified pregnancy outcomes.
Statistical analysis
We performed all analyses in this study using the SAS package (SAS Institute). Pearson χ 2 tests were used to compare differences between women with and without OSA in terms of the characteristics of mother, infant, and father identified in the above-mentioned text. We also used conditional logistic regression analyses that were conditioned on maternal age to examine the risk of adverse pregnancy outcomes between women with and without OSA. A 2-sided P < .05 was considered statistically significant for this study.
Results
The mean age of the 4746 sampled women was 30.3 ± 4.4 years (SD; range, 14–45 years). The mean birthweight for women with OSA and women without OSA were 3063 ± 584 g (SD; range, 361–4650 g) and 3147 ± 418 g (SD; range, 1426–4760 g), respectively. Moreover, the mean gestational age for women with OSA and women without OSA were 38 ± 2.28 weeks (SD; range, 24–41 weeks) and 38 ± 1.45 weeks (SD; range, 29–43 weeks), respectively.
Table 1 reports the distribution of characteristics of mothers, infants, and fathers across the study and comparison groups. After matching for maternal age, we found no significant differences between women with and without OSA in infant sex ( P = .216), maternal education level ( P = .156), anemia ( P = .989), and hyperlipidemia ( P = .998). However, there were significant differences in infant parity ( P < .001), maternal marital status ( P < .001), coronary heart disease ( P < .001), obesity ( P < .001), geographic region ( P < .001), and paternal age ( P = .003) between women with and without OSA.
