Abstract
Rationale
Knowledge gaps exist regarding health implications of sleep-disordered breathing (SDB) identified in pregnancy and/or after delivery.
Objectives
To determine whether SDB in pregnancy and/or after delivery is associated with hypertension (HTN) and metabolic syndrome (MS).
Methods
nuMoM2b-HHS (Nulliparous Pregnancy Outcomes Study: Monitoring Mothers‐to‐be Heart Health Study) (N = 4,508) followed participants initially recruited during their first pregnancy. Participants returned for a visit 2–7 years after pregnancy. This study examined a subgroup who underwent SDB assessments during their first pregnancy (n = 1,964) and a repeat SDB assessment after delivery (n = 1,222). Two SDB definitions were considered: 1) apnea–hypopnea index (AHI) ⩾ 5 and 2) oxygen desaturation index (ODI) ⩾ 5. Associations between SDB and incident HTN and MS were evaluated with adjusted risk ratios (aRRs).
Measurements and Main Results
The aRR for MS given an AHI ⩾ 5 during pregnancy was 1.44 (95% confidence interval [CI], 1.08–1.93), but no association with HTN was found. ODI ⩾ 5 in pregnancy was associated with both an increased risk for HTN (aRR, 2.02; 95% CI, 1.30–3.14) and MS (aRR, 1.53; 95% CI, 1.19–1.97). Participants with an AHI ⩾ 5 in pregnancy that persisted after delivery were at higher risk for both HTN (aRR, 3.77; 95% CI, 1.84–7.73) and MS (aRR, 2.46; 95% CI, 1.59–3.76). Similar associations were observed for persistent ODI ⩾ 5 after delivery.
Conclusions
An AHI ⩾ 5 in pregnancy was associated with an increased risk of MS. An ODI ⩾ 5 in pregnancy was significantly associated with both HTN and MS. Participants with persistent elevations in AHI and ODI during pregnancy and at 2–7 years after delivery were at the highest risk for HTN and MS.
Clinical trial registered with www.clinicaltrials.gov (NCT 02231398).
Keywords: sleep-disordered breathing, pregnancy, postpartum, cardiometabolic health, hypertension
At a Glance Commentary
Scientific Knowledge on the Subject
We used the nuMoM2b (Nulliparous Pregnancy Outcomes Study: Monitoring Mothers‐to‐be) data to address the knowledge gap that exists regarding the health implications of sleep-disordered breathing (SDB) identified in pregnancy and/or after delivery. Specifically, our objective was to determine whether SDB in pregnancy and/or after delivery is associated with hypertension (HTN) and metabolic syndrome (MS).
What This Study Adds to the Field
In this analysis of objectively assessed SDB in pregnancy and at a 2- to 7-year follow-up visit in a large U.S. cohort, we found that SDB measured during pregnancy and in the 2- to 7-year period after delivery was associated with incident HTN and MS. Associations of SDB in pregnancy were generally stronger when SDB was defined using the oxygen desaturation index, which quantified the frequency of drops in oxygen saturation of ⩾3%, rather than the apnea–hypopnea index, which was defined by a combination of changes in airflow and oxygen saturation. Participants with persistent elevations in apnea–hypopnea index or oxygen desaturation index during pregnancy and the 2- to 7-year follow-up visits were at more than a threefold increased risk for incident HTN and more than a twofold increased risk for MS.
Sleep-disordered breathing (SDB), characterized by recurrent apneas and hypopneas, intermittent hypoxemia, and sleep disruption, is increasingly recognized in pregnancy. Pregnant individuals are at increased risk for SDB, predominantly obstructive sleep apnea, compared with their nonpregnant counterparts, due to physical and hormonal changes that occur during pregnancy (1, 2). A meta-analysis of 33 studies found that the pooled overall prevalence of SDB during pregnancy was 15% (95% confidence interval [CI], 12–18%) (3).
While epidemiologic data from cohorts of middle-aged and older adults indicate that SDB is associated with adverse cardiometabolic outcomes (4–6), less is known about how SDB in pregnancy and the period after delivery impacts maternal health. In pregnancy, increases in inflammation, oxidative stress, and sympathetic nervous system activity, all of which can be exacerbated by SDB, can lead to adverse maternal health events (2, 7). Indeed, SDB in pregnancy has been associated with a two- to threefold increased risk for preeclampsia and gestational diabetes mellitus (GDM) (3, 8). These and other adverse pregnancy outcomes are risk factors for later development of hypertension (HTN) and metabolic disease (9–12). The postpartum period is also of unique relevance regarding SDB epidemiology, given the likely impact of postpartum weight retention on SDB risk (13–15). Thus, there is a critical need to elucidate whether SDB identified in pregnancy and/or in the period after delivery is associated with cardiovascular and metabolic health in young adults.
nuMoM2b-HHS (Nulliparous Pregnancy Outcomes Study: Monitoring Mothers‐to‐be Heart Health Study) was a follow-up study of participants initially recruited during their first pregnancy to the parent nuMoM2b study, completed after delivery. A subset of women in the parent study underwent evaluation for SDB during their first pregnancy (“nuMoM2b SDB substudy”) and was offered a repeat assessment after delivery (16–18). We used the nuMoM2b data to address the knowledge gap that exists regarding the health implications of SDB identified in pregnancy and/or after delivery. Specifically, our objective was to determine whether SDB in pregnancy and/or after delivery is associated with HTN and metabolic syndrome (MS). Some of the results of these studies have been previously reported in the form of an abstract (19).
Methods
Details of the nuMoM2b study and the nuMoM2b-SDB substudy have been previously published (16, 18). Inclusion criteria for the nuMoM2b study were nulliparity (no prior delivery, ⩾20 wk of gestation) and a viable singleton pregnancy at screening (6 wk and 0 d to 13 wk and 6 d of gestation). Participants were excluded from the nuMoM2b-SDB substudy if they were currently using continuous positive airway pressure (CPAP) treatment for SDB, had severe asthma, or required oxygen supplementation.
nuMoM2b-HHS Methods
The complete methods of the follow-up nuMoM2b‐HHS study are described elsewhere (17). Briefly, contacts/interviews began at least 6 months after delivery. An in‐person nuMoM2b‐HHS visit was conducted 2–7 years after the index pregnancy ended. Participants were asked to fast for 8 hours before the visit. Blood pressure (BP), anthropometric measurements, and biological specimens were collected using standardized protocols. Participants in the nuMoM2b-HHS study were eligible to participate in a follow-up assessment of SDB if they had participated in the nuMoM2b-SDB substudy while pregnant.
nuMoM2b-HHS Outcomes
Incident HTN was defined as HTN that developed after the index pregnancy, as determined by BP measurement (systolic BP ⩾ 140 mm Hg or diastolic BP ⩾ 90 mm Hg) or use of an antihypertensive medication for BP control at the 2‐ to 7‐year visit after delivery.
MS was defined as the presence of three of the following five criteria: elevated waist circumference (⩾88 cm [non-Asian] or ⩾80 cm [Asian]); elevated triglycerides (⩾150 mg/dl) or use of a triglyceride-lowering medication; elevated fasting glucose (⩾100 md/dl) or use of a glucose‐lowering medication; elevated BP (systolic BP ⩾ 130 mm Hg or a diastolic BP ⩾ 85 mm Hg) or use of an antihypertensive medication; and reduced HDL levels (<50 mg/dl) or use of medications known to increase HDL (20) (see online supplement for additional details).
SDB Assessments
SDB was assessed at all exams using the same model (Embletta-Gold device Embla) level 3 home sleep apnea test (HSAT) (18). Studies were conducted twice during pregnancy, first between 6 and 15 weeks of pregnancy and then again between 22 and 31 weeks, and then at the 2- to 7-year follow-up visit. Sleep studies were scored by a central reading center by trained polysomnologists blinded to all other data. Event definitions, scoring reliability, and the quality control protocol were previously published (18). Two SDB definitions were considered: 1) apnea–hypopnea index (AHI): the number of apneas and hypopneas per hour of estimated sleep, inclusive of all apneas plus hypopneas accompanied by ⩾3% oxygen desaturation; and 2) oxygen desaturation index (ODI): the number of oxygen desaturations ⩾3% from the before event baseline per hour of estimated sleep (see online supplement for additional details).
Statistical Analyses
For our primary analyses, we defined SDB using dichotomous AHI and ODI metrics: SDB by AHI (SDB-AHI) was defined as an AHI ⩾ 5, SDB by ODI (SDB-ODI) was defined as an ODI ⩾ 5. Among participants with valid early pregnancy and midpregnancy sleep study data, SDB in pregnancy was defined as an AHI ⩾ 5 in either early pregnancy or midpregnancy (vs. AHI < 5 in both early pregnancy and midpregnancy). A similar categorization was used to define SDB during pregnancy by ODI. In secondary analyses, we separately considered SDB that was present in early pregnancy or midpregnancy. To examine exposure–response relationships, we also grouped participants at early pregnancy and midpregnancy into four categories based on their AHI or ODI values: 0, 0–4.9, 5–14.9, and 15 or more events/h (21). Post hoc tests using orthogonal contrasts were used to assess the exposure–response relationships for linear and quadratic trends in risk on the log scale across the SDB categories.
In a similar fashion, a secondary analysis was performed using the AHI and ODI data obtained at the 2- to 7-year follow-up visit. We also examined the trajectory of SDB (by both definitions) in pregnancy and at the follow-up visit by defining 4 groups: those with no SDB at either time point, those with SDB in pregnancy that persisted at the 2- to 7-year follow-up, those with SDB in pregnancy that resolved at follow-up, and those with new-onset SDB at the 2- to 7-year follow-up visit.
Baseline characteristics of the index pregnancy and cardiovascular characteristics at the 2- to 7-year follow-up visit were summarized according to SDB status during pregnancy. Similarly, cardiovascular characteristics at the 2- to 7-year follow-up visit were also summarized by SDB status at follow-up.
Analyses during pregnancy were restricted to participants who had an early pregnancy or midpregnancy sleep study with adequate data and attended the in-person nuMoM2b-HHS cardiovascular assessment. Participants with baseline chronic HTN in pregnancy were excluded from the HTN analyses; likewise, those with preexisting diabetes were excluded from the MS analyses.
Crude risk ratios (RRs) and adjusted risk ratios (aRRs) and 95% CI were estimated using Poisson regression with robust standard errors to relate SDB to incident HTN and MS (22). P values from likelihood ratio tests were reported. Adjustment covariates chosen a priori included: age and body mass index (BMI) in early pregnancy, self-identified race, and years from delivery of the index pregnancy to the HHS follow-up visit. Similar methods were used for the cross-sectional analyses of SDB in the period after delivery, with age and BMI at the time of the HHS follow-up visit and race used as adjustment covariates. In all our adjusted analyses, BMI was included as a linear and a quadratic term.
We also considered having a hypertensive disorder of pregnancy (HDP) (i.e., gestational hypertension and preeclampsia) during the index pregnancy as a mediator in the associations of SDB during pregnancy with HTN after delivery and HDP or GDM during pregnancy as a mediator in the associations of SDB during pregnancy with MS after delivery.
All tests were performed at a nominal significance level of α = 0.05. All single-degree-of-freedom tests were two-sided. No correction was made for multiple comparisons. Analyses were performed using SAS 9.4 (see online supplement for additional details).
Results
A total of 4,508 participants attended a nuMoM2b-HHS in-person visit between February 4, 2014, and October 9, 2017. Among these, 1,964 had an HSAT from the index pregnancy, with 1,863 having adequate sleep study data available (406 with early pregnancy data only; 92 with midpregnancy data only; 1,365 with both early pregnancy and midpregnancy data) (Figure 1). The median (interquartile range [IQR]) time between delivery of the index pregnancy and the HHS in-person visit was 36 (28–42) months.
Figure 1.
Inclusion in analysis of sleep study results during pregnancy in association with hypertension and metabolic syndrome 2–7 years after delivery. nuMoM2b = Nulliparous Pregnancy Outcomes Study: Monitoring Mothers-to-be.
Among nuMoM2b-HHS participants who had both early pregnancy and midpregnancy sleep study data, 127 (9.3%) had an AHI ⩾ 5 at either the early pregnancy or midpregnancy SDB assessment. Baseline characteristics according to AHI category during pregnancy for the nuMoM2b-HHS participants are presented in Table 1. Participants with an AHI ⩾ 5 in pregnancy were older, had higher BMIs in early pregnancy, and had a lower rate of weight gain in pregnancy.
Table 1.
Characteristics of nuMoM2b Participants with Early Pregnancy or Midpregnancy Sleep Study and Cardiovascular Assessment at 2- to 7-Year Follow-up according to Apnea–Hypopnea Index* Categories
| Characteristics | AHI < 5 in Early Pregnancy and Midpregnancy (n = 1,238) | AHI ⩾ 5 in Early Pregnancy or Midpregnancy (n = 127) | P Value† |
|---|---|---|---|
| Maternal age in early pregnancy | |||
| Mean (SD), yr | 26.6 (5.2) | 29.8 (5.8) | <0.0001 |
| Category, n (%) | <0.0001 | ||
| 13–21 yr | 234 (18.9) | 11 (8.7) | |
| 22–35 yr | 939 (75.8) | 93 (73.2) | |
| >35 yr | 65 (5.3) | 23 (18.1) | |
| Maternal race, n (%) | 0.5663 | ||
| White non-Hispanic | 810 (65.4) | 83 (65.4) | |
| Black non-Hispanic | 133 (10.7) | 18 (14.2) | |
| Hispanic | 199 (16.1) | 16 (12.6) | |
| Asian | 34 (2.7) | 5 (3.9) | |
| Other | 62 (5.0) | 5 (3.9) | |
| BMI in early pregnancy | |||
| Mean, SD, kg/m2 | 25.9 (5.9) | 33.5 (8.2) | <0.0001 |
| Category, n (%) | <0.0001 | ||
| <25 kg/m2 | 680 (55.8) | 18 (14.4) | |
| 25 to <30 kg/m2 | 295 (24.2) | 29 (23.2) | |
| ⩾30 kg/m2 | 244 (20.0) | 78 (62.4) | |
| Smoked during pregnancy, n (%) | 74 (6.0) | 11 (8.7) | 0.2332 |
| Rate of weight gain from early pregnancy to midpregnancy, kg/wk | |||
| Mean, SD | 0.49 (0.22) | 0.42 (0.26) | 0.0011 |
Definition of abbreviations: AHI = apnea–hypopnea index; BMI = body mass index; nuMoM2b = Nulliparous Pregnancy Outcomes Study: Monitoring Mothers‐to‐be.
Including all apneas and hypopneas with ⩾3% oxygen desaturation per hour.
P values are shown for chi-square tests for AHI and the categorical baseline characteristics and from ANOVA F tests for AHI and continuous baseline characteristics.
A repeat sleep study at the 2- to 7-year follow-up visit was conducted on 1,222 of the 1,964 participants who completed a sleep study during pregnancy. The participants who did not complete the follow-up sleep assessment were less likely to be non-Hispanic White and had a longer latency from delivery to the HHS follow-up (see Table E1 in the online supplement). Adequate follow-up sleep study data were available on 1,069 participants (Figure 2). SDB defined as an AHI ⩾ 5 was found in 133 of the 1,069 (12.4%) participants with follow-up sleep data when studied at a median (IQR) of 32 (13) months after delivery. Of the 844 participants with SDB data from all time points (i.e., at both the early pregnancy and midpregnancy visits and at the follow-up visit), 710 (84.1%) did not have SDB-AHI during pregnancy or at follow-up, 49 (5.8%) had SDB-AHI in pregnancy that persisted at follow-up, 22 (2.6%) had SDB-AHI in pregnancy that resolved at follow-up, and 63 (7.5%) had new-onset SDB-AHI detected at the 2- to 7-year follow-up visit.
Figure 2.
Enrollment and inclusion in analysis of nuMoM2b-HHS (Nulliparous Pregnancy Outcomes Study: Monitoring Mothers‐to‐be Heart Health Study) sleep study data. *There were two participants who completed a nuMoM2b-HHS home sleep apnea test (HSAT) who did not have an HSAT during pregnancy.
Cardiovascular and metabolic characteristics and outcome rates at the 2- to 7-year follow-up visit based on AHI status in pregnancy are presented in Table 2. Participants with SDB-AHI in pregnancy had higher BMIs; larger waist circumference; higher systolic and diastolic BP, triglyceride levels, and fasting blood glucose; and lower high-density lipoprotein cholesterol (HDL-C) levels 2–7 years after delivery. At follow-up, incident HTN was more common in participants with an SDB-AHI during pregnancy (15.7% vs. 6.2%, P < 0.0001), as was MS (40.2% vs. 14.2%, P < 0.0001).
Table 2.
Cardiovascular Characteristics at 2–7 Years after nuMoM2b Index Pregnancy, according to Apnea–Hypopnea Index* Categories
| Cardiovascular Characteristic 2–7 Years after nuMoM2b Index Pregnancy | AHI < 5 in Early Pregnancy and Midpregnancy (n = 1,238) | AHI ⩾ 5 in Early Pregnancy or Midpregnancy (n = 127) | P Value† |
|---|---|---|---|
| BMI | |||
| n | 1,225 | 126 | |
| Mean (SD), kg/m2 | 27.0 (7.1) | 35.2 (9.0) | <0.0001 |
| Category, n (%) | <0.0001 | ||
| <25 kg/m2 | 599 (48.9) | 17 (13.5) | |
| 25 to <30 kg/m2 | 314 (25.6) | 19 (15.1) | |
| ⩾30 kg/m2 | 312 (25.5) | 90 (71.4) | |
| Waist circumference over iliac crest | |||
| n | 1,230 | 127 | |
| Mean (SD), cm | 94.6 (15.1) | 111.6 (19.3) | <0.0001 |
| ⩾88 cm (non-Asian) or ⩾80 cm (Asian), n (%) | 768 (62.4) | 115 (90.6) | <0.0001 |
| SBP | |||
| n | 1,235 | 127 | |
| Mean (SD), mm Hg | 110.8 (10.8) | 116.1 (12.2) | <0.0001 |
| DBP | |||
| n | 1,235 | 127 | |
| Mean (SD), mm Hg | 71.7 (9.5) | 77.4 (9.8) | <0.0001 |
| Hypertension | |||
| SBP ⩾ 140, DBP ⩾ 90, or on antihypertensive medication, n (%) | 76 (6.2) | 20 (15.7) | <0.0001 |
| Triglycerides | |||
| n | 1,223 | 127 | |
| Mean (SD), mg/dl | 93.0 (60.6) | 134.5 (83.8) | <0.0001 |
| ⩾150 mg/dl or on lipid-lowering medication, n (%) | 124 (10.1) | 41 (32.3) | <0.0001 |
| HDL-C | |||
| n | 1,223 | 127 | |
| Mean (SD), mg/dl | 55.6 (12.7) | 49.6 (11.5) | <0.0001 |
| <50 mg/dl or on HDL-raising medication, n (%) | 405 (33.1) | 73 (57.5) | <0.0001 |
| Blood glucose (fasting) | |||
| n | 1,220 | 127 | |
| Mean (SD), mg/dl | 89.9 (14.9) | 96.4 (25.2) | <0.0001 |
| ⩾100 mg/dl or on glucose-lowering medication, n (%) | 175 (14.3) | 31 (24.4) | 0.0027 |
| Metabolic syndrome, n (%)‡ | 172 (14.2) | 51 (40.2) | <0.0001 |
| Time from delivery to HHS cardiovascular assessment | |||
| n | 1,236 | 127 | |
| Mean (SD), yr | 3.0 (0.8) | 2.9 (0.8) | 0.1812 |
Definition of abbreviations: AHI = apnea–hypopnea index; BMI = body mass index; DBP = diastolic blood pressure; HDL-C = high-density lipoprotein cholesterol; HHS = Heart Healthy Study; nuMoM2b = Nulliparous Pregnancy Outcomes Study: Monitoring Mothers‐to‐be; SBP = systolic blood pressure.
Including all apneas and hypopneas with ⩾3% oxygen desaturation per hour.
P values are shown for chi-square tests for AHI and the categorical baseline characteristics and from ANOVA F tests for AHI and continuous baseline characteristics.
Metabolic syndrome is defined based on the presence of three of five of the following criteria: elevated waist circumference, elevated triglycerides or associated medication, elevated fasting glucose or associated medication, elevated blood pressure or associated medication, and reduced HDL-C or associated medication.
Table 3 provides crude RRs and aRRs for incident HTN and MS, related to AHI in pregnancy. In adjusted analyses, SDB-AHI in pregnancy (early or mid) was not associated with the risk of incident HTN. Similarly, SDB-AHI in early pregnancy was not associated with a statistically significant increased HTN risk, although the midpregnancy SDB-AHI point estimate was of borderline significance (aRR, 1.73; 95% CI, 0.99–3.01; P = 0.091). AHI grouping by AHI = 0 (referent), 0 < AHI < 5, 5 ⩽ AHI < 15, and AHI ⩾ 15 did not reveal a statistically significant linear or quadratic trend in incident HTN risk with increasing AHI in either early pregnancy or midpregnancy.
Table 3.
Crude and Adjusted Risk Ratios* for Incident Hypertension and Metabolic Syndrome† at 2- to 7-Year Follow-up according to Apnea–Hypopnea Index‡ in Early Pregnancy and Midpregnancy
| AHI Characteristic | n/N (%) | Crude Risk Ratios |
Adjusted Risk Ratios |
||
|---|---|---|---|---|---|
| Estimate (95% CI) | P Value | Estimate (95% CI) | P Value | ||
| Incident hypertension | |||||
| Early pregnancy or midpregnancy | |||||
| AHI < 5 in both early pregnancy and midpregnancy (referent) | 67/1,215 (5.5) | 1.00 | 0.0402 | 1.00 | 0.3344 |
| AHI ⩾ 5 in early pregnancy or midpregnancy | 14/118 (11.9) | 2.15 (1.25–3.71) | 1.39 (0.75–2.56) | ||
| Early pregnancy | |||||
| AHI < 5 (referent) | 96/1,647 (5.8) | 1.00 | 0.0463 | 1.00 | 0.6015 |
| AHI ⩾ 5 | 10/68 (14.7) | 2.52 (1.38–4.62) | 1.21 (0.61–2.41) | ||
| AHI = 0 (referent) | 18/391 (4.6) | 1.00 | 0.1280 | 1.00 | 0.9606 |
| 0 < AHI < 5 | 78/1,256 (6.2) | 1.35 (0.82–2.22) | Trend tests: | 1.03 (0.63–1.68) | Trend tests: |
| 5 ⩽ AHI < 15 | 9/63 (14.3) | 3.10 (1.46–6.60) | 0.3931 linear | 1.26 (0.55–2.88) | 0.8631 linear |
| AHI ⩾ 15 | 1/5 (20.0) | 4.34 (0.71–26.55) | 0.9702 quadratic | 1.11 (0.19–6.61) | 0.8672 quadratic |
| Midpregnancy | |||||
| AHI < 5 (referent) | 72/1,308 (5.5) | 1.00 | 0.0070 | 1.00 | 0.0908 |
| AHI ⩾ 5 | 17/114 (14.9) | 2.71 (1.66–4.43) | 1.73 (0.99–3.01) | ||
| AHI = 0 (referent) | 9/189 (4.8) | 1.00 | 0.0547 | 1.00 | 0.3871 |
| 0 < AHI < 5 | 63/1,119 (5.6) | 1.18 (0.60–2.34) | Trend tests: | 0.86 (0.44–1.69) | Trend tests: |
| 5 ⩽ AHI < 15 | 14/100 (14.0) | 2.94 (1.32–6.55) | 0.1325 linear | 1.43 (0.62–3.30) | 0.3511 linear |
| AHI ⩾ 15 | 3/14 (21.4) | 4.50 (1.37–14.77) | 0.7140 quadratic | 1.92 (0.49–7.60) | 0.5638 quadratic |
| Metabolic syndrome | |||||
| Early pregnancy or midpregnancy | |||||
| AHI < 5 in both early pregnancy and midpregnancy (referent) | 165/1,200 (13.8) | 1.00 | <0.0001 | 1.00 | 0.0211 |
| AHI ⩾ 5 in early pregnancy or midpregnancy | 48/124 (38.7) | 2.82 (2.16–3.66) | 1.44 (1.08–1.93) | ||
| Early pregnancy | |||||
| AHI < 5 (referent) | 260/1,630 (16.0) | 1.00 | <0.0001 | 1.00 | 0.1214 |
| AHI ⩾ 5 | 32/69 (46.4) | 2.91 (2.20–3.84) | 1.28 (0.95–1.73) | ||
| AHI = 0 (referent) | 43/383 (11.2) | 1.00 | <0.0001 | 1.00 | 0.4454 |
| 0 < AHI < 5 | 217/1,247 (17.4) | 1.55 (1.14–2.11) | Trend tests: | 1.07 (0.80–1.43) | Trend tests: |
| 5 ⩽ AHI < 15 | 29/63 (46.0) | 4.10 (2.78–6.05) | 0.1141 linear | 1.38 (0.92–2.07) | 0.5643 linear |
| AHI ⩾ 15 | 3/6 (50.0) | 4.45 (1.91–10.40) | 0.4084 quadratic | 1.25 (0.50–3.13) | 0.7232 quadratic |
| Midpregnancy | |||||
| AHI < 5 (referent) | 178/1,292 (13.8) | 1.00 | <0.0001 | 1.00 | 0.0754 |
| AHI ⩾5 | 45/121 (37.2) | 2.70 (2.06–3.53) | 1.33 (0.99–1.79) | ||
| AHI = 0 (referent) | 13/189 (6.9) | 1.00 | <0.0001 | 1.00 | 0.1497 |
| 0 < AHI < 5 | 165/1,103 (15.0) | 2.17 (1.26–3.74) | Trend tests: | 1.31 (0.78–2.21) | Trend tests: |
| 5 ⩽ AHI < 15 | 37/107 (34.6) | 5.03 (2.80–9.03) | 0.0053 linear | 1.64 (0.91–2.97) | 0.0432 linear |
| AHI ⩾15 | 8/14 (57.1) | 8.31 (4.15–16.62) | 0.4603 quadratic | 2.32 (1.11–4.86) | 0.8503 quadratic |
Definition of abbreviations: AHI = apnea–hypopnea index; BMI = body mass index; CI = confidence interval; HDL-C = high-density lipoprotein cholesterol; HHS = Heart Healthy Study.
Poisson regression models are used for the analyses with robust error covariance used to compute CIs and to test for trends. P values are based on likelihood ratio tests. Analyses are adjusted for age and BMI (linear and quadratic) in early pregnancy, race, and time from delivery to HHS cardiovascular assessment.
Metabolic syndrome is defined based on the presence of three of five of the following criteria: elevated waist circumference, elevated triglycerides or associated medication, elevated fasting glucose or associated medication, elevated blood pressure or associated medication, and reduced HDL-C or associated medication. Analyses are adjusted for age and BMI in early pregnancy, race, and time from delivery to HHS cardiovascular assessment.
Including all apneas and hypopneas with ⩾3% oxygen desaturation per hour.
The adjusted risk for MS given SDB-AHI in pregnancy was 1.44 (95% CI, 1.08–1.93). Separate analyses of AHI ⩾ 5 in early pregnancy or midpregnancy with MS were not statistically significant. There was, however, a significant linear trend suggesting an exposure–response relationship between increasing AHI in midpregnancy and the risk of MS (P = 0.043). Participants in the highest AHI group (⩾15) had an aRR for MS of 2.32 (95% CI, 1.11–4.86).
Table 4 gives crude RRs and aRRs relating ODI in pregnancy to incident HTN and MS diagnosed at the 2- to 7-year follow-up. In adjusted analyses, SDB-ODI in pregnancy (early or middle) was associated with an increased risk for incident HTN (aRR, 2.02; 95% CI, 1.30–3.14). A similar statistically significant risk was observed separately for ODI ⩾ 5 at midpregnancy but not early pregnancy.
Table 4.
Crude and Adjusted Risk Ratios* for Incident Hypertension and Metabolic Syndrome† at 2- to 7-Year Follow-up according to Oxygen Desaturation Index‡ in Early Pregnancy and Midpregnancy
| ODI Characteristic | n/N (%) | Crude Risk Ratios |
Adjusted Risk Ratios |
||
|---|---|---|---|---|---|
| Estimate (95% CI) | P Value | Estimate (95% CI) | P Value | ||
| Incident hypertension | |||||
| Early pregnancy or midpregnancy | |||||
| ODI < 5 in both early pregnancy and midpregnancy (referent) | 41/984 (4.2) | 1.00 | <0.0001 | 1.00 | 0.0041 |
| ODI ⩾ 5 in early pregnancy or midpregnancy | 40/349 (11.5) | 2.75 (1.81–4.18) | 2.02 (1.30–3.14) | ||
| Early pregnancy | |||||
| ODI < 5 (referent) | 82/1,521 (5.4) | 1.00 | 0.0048 | 1.00 | 0.2918 |
| ODI ⩾ 5 | 24/194 (12.4) | 2.29 (1.49–3.53) | 1.32 (0.81–2.13) | ||
| ODI = 0 (referent) | 5/141 (3.5) | 1.00 | 0.0220 | 1.00 | 0.6595 |
| 0 < ODI < 5 | 77/1,380 (5.6) | 1.57 (0.65–3.82) | Trend tests: | 1.30 (0.53–3.19) | Trend tests: |
| 5 ⩽ ODI < 15 | 21/174 (12.1) | 3.40 (1.32–8.80) | 0.1462 linear | 1.73 (0.63–4.75) | 0.5806 linear |
| ODI ⩾ 15 | 3/20 (15.0) | 4.23 (1.09–16.36) | 0.7345 quadratic | 1.41 (0.33–6.11) | 0.5015 quadratic |
| Midpregnancy | |||||
| ODI < 5 (referent) | 47/1,094 (4.3) | 1.00 | <0.0001 | 1.00 | 0.0013 |
| ODI ⩾ 5 | 42/328 (12.8) | 2.98 (2.00–4.44) | 2.16 (1.41–3.30) | ||
| ODI = 0 (referent) | 4/57 (7.0) | 1.00 | 0.0002 | 1.00 | 0.0092 |
| 0 < ODI < 5 | 43/1,037 (4.1) | 0.59 (0.22–1.59) | Trend tests: | 0.50 (0.18–1.39) | Trend tests: |
| 5 ⩽ ODI < 15 | 35/287 (12.2) | 1.74 (0.64–4.70) | 0.0605 linear | 1.09 (0.39–3.08) | 0.3663 linear |
| ODI ⩾ 15 | 7/41 (17.1) | 2.43 (0.76–7.77) | 0.2666 quadratic | 1.37 (0.40–4.74) | 0.2533 quadratic |
| Metabolic syndrome | |||||
| Early pregnancy or midpregnancy | |||||
| ODI < 5 in both early pregnancy and midpregnancy (referent) | 106/972 (10.9) | 1.00 | <0.0001 | 1.00 | 0.0014 |
| ODI ⩾ 5 in early pregnancy or midpregnancy | 107/352 (30.4) | 2.79 (2.19–3.54) | 1.53 (1.19–1.97) | ||
| Early pregnancy | |||||
| ODI < 5 (referent) | 219/1,505 (14.6) | 1.00 | <0.0001 | 1.00 | 0.0188 |
| ODI ⩾ 5 | 73/194 (37.6) | 2.59 (2.08–3.22) | 1.35 (1.06–1.70) | ||
| ODI = 0 (referent) | 10/137 (7.3) | 1.00 | <0.0001 | 1.00 | 0.0456 |
| 0 < ODI < 5 | 209/1,368 (15.3) | 2.09 (1.14–3.85) | Trend tests: | 1.49 (0.82–2.71) | Trend tests: |
| 5 ⩽ ODI < 15 | 63/173 (36.4) | 4.99 (2.66–9.35) | 0.0021 linear | 2.02 (1.07–3.81) | 0.0917 linear |
| ODI ⩾ 15 | 10/21 (47.6) | 6.52 (3.09–13.76) | 0.1995 quadratic | 1.77 (0.82–3.78) | 0.1362 quadratic |
| Midpregnancy | |||||
| ODI < 5 (referent) | 121/1,083 (11.2) | 1.00 | <0.0001 | 1.00 | 0.0031 |
| ODI ⩾ 5 | 102/330 (30.9) | 2.77 (2.19–3.49) | 1.48 (1.15–1.90) | ||
| ODI = 0 (referent) | 3/54 (5.6) | 1.00 | <0.0001 | 1.00 | 0.0148 |
| 0 < ODI < 5 | 118/1,029 (11.5) | 2.06 (0.68–6.28) | Trend tests: | 1.38 (0.47–4.04) | Trend tests: |
| 5 ⩽ ODI < 15 | 82/286 (28.7) | 5.16 (1.69–15.74) | <0.0001 linear | 1.94 (0.65–5.75) | 0.0214 linear |
| ODI ⩾ 15 | 20/44 (45.5) | 8.18 (2.60–25.75) | 0.6269 quadratic | 2.57 (0.82–8.01) | 0.9481 quadratic |
Definition of abbreviations: BMI = body mass index; CI = confidence interval; HDL-C = high-density lipoprotein cholesterol; HHS = Heart Healthy Study; ODI = oxygen desaturation index.
Poisson regression models are used for the analyses with robust error covariance used to compute CIs and to test for trends. P values are based on likelihood ratio tests. Analyses are adjusted for age and BMI (linear and quadratic) in early pregnancy, race, and time from delivery to HHS cardiovascular assessment.
Metabolic syndrome is defined based on the presence of three of five of the following criteria: elevated waist circumference, elevated triglycerides or associated medication, elevated fasting glucose or associated medication, elevated blood pressure or associated medication, and reduced HDL-C or associated medication. Analyses are adjusted for age and BMI in early pregnancy, race, and time from delivery to HHS cardiovascular assessment.
Number of desaturations with ⩾3% per hour of sleep.
In adjusted analyses, SDB-ODI in pregnancy was associated with an increased risk of MS (aRR, 1.53; 95% CI, 1.19–1.97). This risk was also separately observed for both early pregnancy and midpregnancy, and the risk increased with increasing categories of ODI severity. Notably, participants with an ODI ⩾ 15 in midpregnancy had an aRR of 2.57 (95% CI, 0.82–8.01) for MS (P = 0.02 for linear trend).
To better understand the relationships between SDB in pregnancy and MS, we examined the individual components of MS in relation to the presence of SDB-AHI or SDB-ODI in pregnancy. In adjusted analyses, an SDB-AHI in pregnancy was associated with elevated triglycerides and reduced HDL. SDB-ODI in pregnancy was associated with elevated triglycerides and elevated BP (data shown in Tables E2 and E3).
Given the associations between AHI and ODI in pregnancy and hypertensive disorders of pregnancy and GDM, and the relationship between these pregnancy complications and later life cardiometabolic disease, we performed a mediation analysis aimed at assessing the controlled direct effect of SDB in pregnancy on HTN and MS at the 2- to 7-year follow-up. In these analyses, we did not observe a statistically significant controlled direct effect (AHI only, without mediation or interaction) between SDB-AHI and the outcomes. Similarly, no direct effect was observed for SDB-ODI on HTN risk at the 2- to 7-year follow-up. However, the controlled direct effect of SDB-ODI in pregnancy on MS was statistically significant. A total excess aRR of 0.60 (95% CI, 0.10–1.09) was observed with SDB-ODI in pregnancy and the proportion explained by mediation and/or interaction was estimated at ⩾11% with 95% confidence (data in Tables E4 and E5).
Cardiovascular and metabolic characteristics and outcome rates at the 2- to 7-year follow-up by current SDB status by AHI are detailed in Table E6. In summary, participants with an SDB-AHI at the 2- to 7-year follow-up had higher BMIs; larger waist circumferences; higher systolic and diastolic BP, triglyceride levels, and fasting blood glucose levels; and lower HDL-C levels. Incident HTN was more common in participants with an SDB-AHI after delivery (19.5% vs. 5.3%, P < 0.0001), as was MS (44.4% vs. 14.0%, P < 0.0001).
Table 5 gives crude RRs and aRRs relating SDB by AHI and ODI status at the 2- to 7-year follow-up to incident HTN and MS. In adjusted analyses, an SDB-AHI after delivery was associated with an increased likelihood of HTN (aRR, 2.01; 95% CI, 1.17–3.46) and MS (aRR, 1.51; 95% CI, 1.14–2.00). Similarly, SDB-ODI was associated with increased risks for incident HTN (aRR, 1.75; 95% CI, 1.05–2.92) and MS (aRR, 1.60; 95% CI, 1.21–2.12). There were statistically significant linear trends suggesting exposure–response relationships between increasing values of AHI and risk of incident HTN and MS, and between increasing values of ODI and risk of incident HTN.
Table 5.
Crude and Adjusted Risk Ratios* for Incident Hypertension† and Metabolic Sydrome‡ at 2- to 7-Year Follow-up according to Apnea–Hypopnea Index and Oxygen Desaturation Index§ at the Follow-up Assessment
| Sleep Characteristic | n/N (%) | Crude Risk Ratios |
Adjusted Risk Ratios |
||
|---|---|---|---|---|---|
| Estimate (95% CI) | P Value | Estimate (95% CI) | P Value | ||
| AHI | |||||
| Incident hypertension | |||||
| AHI < 5 (referent) | 45/918 (4.9) | 1.00 | 0.0008 | 1.00 | 0.0271 |
| AHI ⩾ 5 | 21/124 (16.9) | 3.45 (2.13–5.60) | 2.01 (1.17–3.46) | ||
| AHI = 0 (referent) | 4/118 (3.4) | 1.00 | 0.0048 | 1.00 | 0.0580 |
| 0 < AHI < 5 | 41/800 (5.1) | 1.51 (0.55–4.14) | Trend tests: | 1.11 (0.39–3.15) | Trend tests: |
| 5 ⩽ AHI < 15 | 11/94 (11.7) | 3.45 (1.14–10.49) | 0.0015 linear | 1.67 (0.50–5.57) | 0.0131 linear |
| AHI ⩾ 15 | 10/30 (33.3) | 9.83 (3.31–29.19) | 0.3869 quadratic | 4.18 (1.19–14.76) | 0.2783 quadratic |
| Metabolic syndrome | |||||
| AHI < 5 (referent) | 122/911 (13.4) | 1.00 | <0.0001 | 1.00 | 0.0073 |
| AHI ⩾ 5 | 54/127 (42.5) | 3.18 (2.45–4.12) | 1.51 (1.14–2.00) | ||
| AHI = 0 (referent) | 11/116 (9.5) | 1.00 | <0.0001 | 1.00 | 0.0073 |
| 0 < AHI < 5 | 111/795 (14.0) | 1.47 (0.82–2.65) | Trend tests: | 1.00 (0.57–1.75) | Trend tests: |
| 5 ⩽ AHI < 15 | 34/96 (35.4) | 3.73 (2.00–6.97) | <0.0001 linear | 1.32 (0.71–2.46) | 0.0079 linear |
| AHI ⩾ 15 | 20/31 (64.5) | 6.80 (3.66–12.65) | 0.5670 quadratic | 2.14 (1.11–4.15) | 0.1936 quadratic |
| ODI | |||||
| Incident hypertension | |||||
| ODI < 5 (referent) | 28/695 (4.0) | 1.00 | 0.0002 | 1.00 | 0.0349 |
| ODI ⩾ 5 | 38/347 (11.0) | 2.72 (1.70–4.35) | 1.75 (1.05–2.92) | ||
| ODI < 5 (referent)‖ | 28/695 (4.0) | 1.00 | 0.0004 | 1.00 | 0.0244 |
| 5 ⩽ ODI < 15 | 23/279 (8.2) | 2.05 (1.20–3.49) | Trend tests: | 1.52 (0.86–2.68) | Trend tests: |
| ODI ⩾ 15 | 15/68 (22.1) | 5.48 (3.08–9.74) | 0.0010 linear 0.5825 quadratic | 2.89 (1.50–5.57) | 0.0080 linear 0.6403 quadratic |
| Metabolic syndrome | |||||
| ODI < 5 (referent) | 71/689 (10.3) | 1.00 | <0.0001 | 1.00 | 0.0009 |
| ODI ⩾ 5 | 105/349 (30.1) | 2.92 (2.22–3.83) | 1.60 (1.21–2.12) | ||
| ODI = 0 (referent) | 2/22 (9.1) | 1.00 | <0.0001 | 1.00 | 0.0014 |
| 0 < ODI < 5 | 69/667 (10.3) | 1.14 (0.30–4.35) | Trend tests: | 0.74 (0.20–2.79) | Trend tests: |
| 5 ⩽ ODI < 15 | 68/280 (24.3) | 2.67 (0.70–10.18) | <0.0001 linear | 1.10 (0.29–4.16) | 0.3061 linear |
| ODI ⩾ 15 | 37/69 (53.6) | 5.90 (1.55–22.52) | 0.4812 quadratic | 1.57 (0.41–6.01) | 0.4813 quadratic |
Definition of abbreviations: AHI = apnea–hypopnea index; BMI = body mass index; CI = confidence interval; HDL-C = high-density lipoprotein cholesterol; HHS = Heart Healthy Study; ODI = oxygen desaturation index.
Poisson regression models are used for the analyses with robust error covariance used to compute CIs and to test for trends. P values are based on likelihood ratio tests. Analyses are adjusted for age and BMI (linear and quadratic) at the time of the HHS cardiovascular assessment, and race.
Women with baseline hypertension are excluded from the hypertension analyses.
Women with preexisting diabetes are excluded from the metabolic syndrome analyses. Metabolic syndrome is defined based on the presence of three of five of the following criteria: elevated waist circumference, elevated triglycerides or associated medication, elevated fasting glucose or associated medication, elevated blood pressure or associated medication, and reduced HDL-C or associated medication.
AHI is all apneas and hypopneas with ⩾3% oxygen desaturation per hour; ODI is the number of desaturations ⩾3% per hour of sleep.
There were no women in the ODI = 0 with hypertension, so the ODI = 0 group was combined with the 0 < ODI < 5 group for this analysis.
Among participants with data across all time points (826 participants), Table 6 examines the relationship between SDB trajectory by AHI and ODI criteria from pregnancy to the 2- to 7-year follow-up visit in relation to risk of incident HTN and MS. Rates of both outcomes differed by SDB trajectory. Participants with persistent SDB, defined using either the AHI or ODI, had the highest risk for incident HTN and MS (aRRs ranging from 2.3 to 3.8).
Table 6.
Crude and Adjusted Risk Ratios* for Incident Hypertension† and Metabolic Sydrome‡ at 2- to 7-Year Follow-up according to Apnea–Hypopnea Index and Oxygen Desaturation Index§ during the nuMoM2b Pregnancy and at the Follow-up Assessment
| Sleep Characteristic | n/N (%) | Crude Risk Ratios |
Adjusted Risk Ratios |
||
|---|---|---|---|---|---|
| Estimate (95% CI) | P Value | Estimate (95% CI) | P Value | ||
| AHI | |||||
| Incident hypertension | |||||
| Never: AHI < 5 in early pregnancy and midpregnancy and after delivery (referent) | 31/702 (4.4) | 1.00 | 0.0130‖ | 1.00 | 0.0327‖ |
| New-onset: AHI < 5 in early pregnancy and midpregnancy; AHI ⩾ 5 after delivery | 8/59 (13.6) | 3.07 (1.48–6.37) | 2.18 (1.07–4.48) | ||
| Resolved: AHI ⩾ 5 in early pregnancy or midpregnancy; AHI < 5 after delivery | 0/20 (0.0) | Not applicable | Not applicable | ||
| Persistent: AHI ⩾ 5 in early pregnancy or midpregnancy; AHI ⩾ 5 after delivery | 10/45 (22.2) | 5.03 (2.64–9.60) | 3.77 (1.84–7.73) | ||
| Metabolic syndrome | |||||
| Never: AHI < 5 in early pregnancy and midpregnancy and after delivery (referent) | 81/697 (11.6) | 1.00 | <0.0001 | 1.00 | 0.0033 |
| New-onset: AHI < 5 in early pregnancy and midpregnancy; AHI ⩾ 5 after delivery | 20/60 (33.3) | 2.87 (1.90–4.33) | 1.81 (1.18–2.80) | ||
| Resolved: AHI ⩾ 5 in early pregnancy or midpregnancy; AHI < 5 after delivery | 4/22 (18.2) | 1.56 (0.63–3.89) | 1.12 (0.49–2.54) | ||
| Persistent: AHI ⩾ 5 in early pregnancy or midpregnancy; AHI ⩾ 5 after delivery | 24/47 (51.1) | 4.39 (3.11–6.22) | 2.46 (1.59–3.79) | ||
| ODI | |||||
| Incident hypertension | |||||
| Never: ODI < 5 in early pregnancy and midpregnancy and after delivery (referent) | 15/468 (3.2) | 1.00 | 0.0029 | 1.00 | 0.0401 |
| New-onset: ODI < 5 in early pregnancy and midpregnancy; ODI ⩾ 5 after delivery | 9/144 (6.3) | 1.95 (0.87–4.36) | 1.69 (0.76–3.74) | ||
| Resolved: ODI ⩾ 5 in early pregnancy or midpregnancy; ODI < 5 after delivery | 5/68 (7.4) | 2.29 (0.86–6.11) | 1.81 (0.71–4.60) | ||
| Persistent: ODI ⩾ 5 in early pregnancy or midpregnancy; ODI ⩾ 5 after delivery | 20/146 (13.7) | 4.27 (2.25–8.13) | 2.94 (1.50–5.77) | ||
| Metabolic syndrome | |||||
| Never: ODI < 5 in early pregnancy and midpregnancy and after delivery (referent) | 38/468 (8.1) | 1.00 | <0.0001 | 1.00 | 0.0004 |
| New-onset: ODI < 5 in early pregnancy and midpregnancy; ODI ⩾ 5 after delivery | 28/143 (19.6) | 2.41 (1.54–3.78) | 1.93 (1.24–3.01) | ||
| Resolved: ODI ⩾ 5 in early pregnancy or midpregnancy; ODI < 5 after delivery | 9/65 (13.8) | 1.71 (0.87–3.36) | 1.17 (0.62–2.18) | ||
| Persistent: ODI ⩾ 5 in early pregnancy or midpregnancy; ODI ⩾ 5 after delivery | 54/150 (36.0) | 4.43 (3.06–6.43) | 2.42 (1.59–3.67) | ||
Definition of abbreviations: AHI = apnea–hypopnea index; BMI = body mass index; CI = confidence interval; HDL-C = high-density lipoprotein cholesterol; HHS = Heart Healthy Study; nuMoM2b = Nulliparous Pregnancy Outcomes Study: Monitoring Mothers-to-be; ODI = oxygen desaturation index.
Poisson regression models are used for the analyses with robust error covariance used to compute CIs and to test for trends. P values are based on likelihood ratio tests. Analyses are adjusted for age and BMI (linear and quadratic) in early pregnancy, race, and time from delivery to HHS cardiovascular assessment.
Women with baseline hypertension are excluded from the hypertension analyses.
Women with preexisting diabetes are excluded from the metabolic syndrome analyses. Metabolic syndrome is defined based on the presence of three of five of the following criteria: elevated waist circumference, elevated triglycerides or associated medication, elevated fasting glucose or associated medication, elevated blood pressure or associated medication, and reduced HDL-C or associated medication.
AHI is all apneas and hypopneas with ⩾3% oxygen desaturation per hour; ODI is the number of desaturations ⩾3% per hour of sleep.
None of the 20 participants with AHI ⩾5 during the nuMoM2b pregnancy study and AHI <5 at the 2- to 7-year assessment had hypertension at the follow-up assessment. To compute relative risk estimates and an upper bound on the P value for differences between the AHI categories, one observation among the 20 was categorized as having hypertension, such that the observation selected resulted in the smallest estimate of adjusted relative risk for this group compared to the referent group.
Given that the relationships between our outcomes were stronger for ODI in pregnancy compared with AHI, we sought to understand how AHI and ODI related to each other in the larger pregnancy dataset. As expected, AHI and ODI were strongly correlated (Spearman correlation coefficients at early pregnancy and midpregnancy visits were 0.79 and 0.84, respectively). The median (IQR) for the difference (delta) between ODI and AHI in early pregnancy and midpregnancy were 0.65 (1.47) and 1.16 (2.11), respectively. A larger ODI-AHI delta was seen with higher BMI in pregnancy (additional data on ODI-AHI delta can be found in the online supplement).
Discussion
In this analysis of objectively-assessed SDB in pregnancy and at a 2- to 7-year follow-up visit in a large U.S. cohort, we found that SDB measured during pregnancy and in the 2- to 7-year period after delivery was associated with HTN and MS. Associations of SDB in pregnancy were generally stronger when SDB was defined using the ODI, which quantified the frequency of drops in oxygen saturation of ⩾3%, than the AHI, which was defined by a combination of changes in airflow and oxygen saturation. Participants with persistent elevations in AHI or ODI during pregnancy and 2- to 7-year follow-up visits were at more than threefold increased risk for incident HTN and more than twofold increased risk for MS. Our data suggest that the simpler measure of ODI can identify those at increased risk just as well, if not better, than AHI.
In nonpregnant populations that have been followed after an SDB diagnosis, such as the Wisconsin Sleep Cohort, increasing AHI has been linked to higher rates of incident HTN (23). In our prospective analysis of SDB in pregnancy, we did not find an association between AHI in pregnancy and incident HTN. However, there was an association between the ODI in pregnancy and incident HTN. AHI and ODI are very tightly correlated, but we found that the ODI tended to be slightly higher than AHI, meaning that there were oxygen desaturation events that were not associated with clearly evident respiratory events (apneas and hypopneas). In our analysis, the difference between ODI and AHI during pregnancy could not be explained by an asthma diagnosis or restless legs syndrome symptoms (see online supplement). However, it was correlated with BMI, as was AHI. It may be that there were subtle respiratory flow events that did not reach the amplitude criteria for hypopnea but impaired ventilation, leading to oxygen desaturations. Other researchers found that the difference between AHI and ODI increases progressively with obesity level and similarly postulated that higher BMI may result in desaturation events even in the absence of notable changes in breathing amplitude as measured during routine sleep studies (24).
Our data demonstrated a relationship between pregnancy ODI and MS and a weaker association (demonstrated only in the analysis of AHI in early pregnancy or midpregnancy) between pregnancy AHI and MS. This association has been reported in prior cohorts, but this is the first large prospective study in young women to confirm this association (25). It is important to note that, unlike the HTN analyses, we cannot assert that the MS was incident in the follow-up period as we do not have assessments of fasting blood glucose and lipids before pregnancy for comparison. Furthermore, how best to interpret this relationship between SDB and MS is complicated by the very tight relationship of both SDB and MS with BMI. Yet, even after adjustment for BMI (linear and quadratic), the associations between SDB measures and MS remained statistically significant. Also, notably, in our mediation analysis, we considered pregnancy complications and found that ODI ⩾ 5 in pregnancy had a significant controlled direct effect on MS risk unrelated to the impact of HDP and GDM.
In our cross-sectional analyses of AHI and ODI at the 2- to 7-year follow-up period, we demonstrated strong associations between AHI and ODI with both HTN and MS. We found significant exposure–response relationships between increasing AHI values and higher rates of both outcomes and a significant exposure–response relationship between increasing ODI and incident HTN. When we examined these associations in relation to prior sleep data from pregnancy, we found that participants who had persistent SDB (defined either by AHI or ODI) were at the highest risk.
A major strength of this study is the combined prospective and cross-sectional design from pregnancy through 2 to 7 years of follow-up in which the pregnancy AHI and ODI results were blinded to the care providers, investigators, and participants (unless an urgent alert was identified [see online supplement for details]). This limited the possibility of ascertainment bias. Our SDB ascertainment was optimized using an independent and blinded central reading center. We were able to control for important confounding factors, including BMI (BMI in early pregnancy for prospective analysis, BMI at HHS visit for cross-sectional analysis), and we had objective assessments of cardiovascular and metabolic outcomes at the 2- to 7-year follow-up visit. However, given the observational and voluntary nature of the study and moderate aRRs, the possibility of residual confounding due to selection bias and unmeasured confounders cannot be definitively excluded. Additionally, given sample size limitations, we were unable to perform a meaningful analysis considering the impact of intercurrent pregnancies on the HTN and MS risk. Also, we do not have objective SDB data from all of our participants at the 2- to 7-year follow-up. While it was offered to most nuMoM2b-SDB substudy participants, only 62.2% repeated sleep testing at the 2- to 7-year follow-up assessment. We also recognize the limitations of utilizing a level 3 HSAT for measuring AHI. Unattended home sleep apnea testing may modestly underestimate the AHI and ODI values due to overestimation of sleep time, leading to some degree of misclassification bias (26). Unlike full polysomnography, our home sleep testing procedures did not employ an electroencephalogram so that events associated with arousals (without desaturations) were not ascertained. Finally, because many analyses were done and adjustment for multiple statistical testing was not performed, chance alone might be responsible for some statistically significant results.
Although we found associations with AHI, ODI, and incident HTN and MS, we cannot conclude that universal screening for and treatment of SDB in pregnancy and/or in the period after delivery would reduce the risks of these adverse outcomes. The most widely prescribed treatment for SDB is CPAP. The benefit of treatment with CPAP has been reliably demonstrated when excessive daytime sleepiness and sleep quality are used as endpoints (27, 28). CPAP has also been demonstrated in randomized controlled trials to reduce BP (29, 30). However, data conflict regarding whether treatment of SDB can reduce the risk of developing HTN, other cardiovascular disease, or diabetes (31–34). This is especially true for milder forms of SDB (AHI < 30), which our study confirms represent the vast majority of SDB cases in young pregnant persons.
In summary, in this prospective analysis of objectively assessed SDB in pregnancy, ODI but not AHI ⩾ 5 in pregnancy was associated with incident HTN 2–7 years after the index pregnancy. Both AHI and ODI were associated with MS, though associations were stronger for ODI in pregnancy. AHI and ODI ⩾ 5 at 2–7 years after delivery were strongly associated with both incident HTN and MS, and participants who had persistent AHI and ODI elevations (present at pregnancy and after-delivery visit) were at the highest risk for these adverse outcomes. Further longitudinal studies are needed to examine the temporal or causal relationships between SDB and cardiometabolic risk, to determine if simple oximetry monitoring can reliably be used to identify individuals at risk, and to study if treatment with CPAP during or after pregnancy can modify these risks.
Footnotes
Supported by cooperative agreement funding from the NHLBI and the Eunice Kennedy Shriver National Institute of Child Health and Human Development: U10-HL119991, U10-HL119989, U10-HL120034, U10-HL119990, U10-HL120006, U10-HL119992, U10-HL120019, U10-HL119993, and U10-HL120018. Support was also provided by the NIH: Office of Research on Women’s Health through U10-HL-119991; Office of Behavioral and Social Sciences Research through U10-HL119991 and U10-HL119992; and the National Center for Advancing Translational Sciences through UL-1-TR000124, UL-1-TR000153, UL-1-TR000439, and UL-1-TR001108; by the Barbra Streisand Women’s Cardiovascular Research and Education Program; and by the Erika J. Glazer Women’s Heart Research Initiative, Cedars-Sinai Medical Center, Los Angeles. The content of this article is solely the responsibility of the authors and does not necessarily represent the official views of the NIH.
Author Contributions: All authors made substantial contributions to the conception or design of the work or the acquisition, analysis, or interpretation of data for the work; drafting the work or revising it critically for important intellectual content; final approval of the version to be published; and agreement to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.
This article has an online supplement, which is accessible from this issue’s table of contents at www.atsjournals.org.
Originally Published in Press as DOI: 10.1164/rccm.202104-0971OC on February 11, 2022
Author disclosures are available with the text of this article at www.atsjournals.org.
References
- 1. Ayyar L, Shaib F, Guntupalli K. Sleep-disordered breathing in pregnancy. Sleep Med Clin . 2018;13:349–357. doi: 10.1016/j.jsmc.2018.04.005. [DOI] [PubMed] [Google Scholar]
- 2. Izci-Balserak B, Pien GW. Sleep-disordered breathing and pregnancy: potential mechanisms and evidence for maternal and fetal morbidity. Curr Opin Pulm Med . 2010;16:574–582. doi: 10.1097/MCP.0b013e32833f0d55. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Liu L, Su G, Wang S, Zhu B. The prevalence of obstructive sleep apnea and its association with pregnancy-related health outcomes: a systematic review and meta-analysis. Sleep Breath . 2019;23:399–412. doi: 10.1007/s11325-018-1714-7. [DOI] [PubMed] [Google Scholar]
- 4. Bradley TD, Floras JS. Obstructive sleep apnoea and its cardiovascular consequences. Lancet . 2009;373:82–93. doi: 10.1016/S0140-6736(08)61622-0. [DOI] [PubMed] [Google Scholar]
- 5. Reichmuth KJ, Austin D, Skatrud JB, Young T. Association of sleep apnea and type II diabetes: a population-based study. Am J Respir Crit Care Med . 2005;172:1590–1595. doi: 10.1164/rccm.200504-637OC. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Marshall NS, Wong KK, Phillips CL, Liu PY, Knuiman MW, Grunstein RR. Is sleep apnea an independent risk factor for prevalent and incident diabetes in the Busselton Health Study? J Clin Sleep Med . 2009;5:15–20. [PMC free article] [PubMed] [Google Scholar]
- 7. Izci Balserak B. Sleep disordered breathing in pregnancy. Breathe (Sheff) . 2015;11:268–277. doi: 10.1183/20734735.009215. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Facco FL, Parker CB, Reddy UM, Silver RM, Koch MA, Louis JM, et al. Association between sleep-disordered breathing and hypertensive disorders of pregnancy and gestational diabetes mellitus. Obstet Gynecol . 2017;129:31–41. doi: 10.1097/AOG.0000000000001805. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Rich-Edwards JW, McElrath TF, Karumanchi SA, Seely EW. Breathing life into the lifecourse approach: pregnancy history and cardiovascular disease in women. Hypertension . 2010;56:331–334. doi: 10.1161/HYPERTENSIONAHA.110.156810. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Sattar N, Greer IA. Pregnancy complications and maternal cardiovascular risk: opportunities for intervention and screening? BMJ . 2002;325:157–160. doi: 10.1136/bmj.325.7356.157. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Smith GN, Walker MC, Liu A, Wen SW, Swansburg M, Ramshaw H, et al. A history of preeclampsia identifies women who have underlying cardiovascular risk factors. Am J Obstet Gynecol . 2009;200:58.e1–8. doi: 10.1016/j.ajog.2008.06.035. [DOI] [PubMed] [Google Scholar]
- 12. Bellamy L, Casas JP, Hingorani AD, Williams DJ. Pre-eclampsia and risk of cardiovascular disease and cancer in later life: systematic review and meta-analysis. BMJ . 2007;335:974. doi: 10.1136/bmj.39335.385301.BE. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Street LM, Aschenbrenner CA, Houle TT, Pinyan CW, Eisenach JC. Gestational obstructive sleep apnea: biomarker screening models and lack of postpartum resolution. J Clin Sleep Med . 2018;14:549–555. doi: 10.5664/jcsm.7042. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Sharkey KM, Boni GM, Quattrucci JA, Blatch S, Carr SN. Women with postpartum weight retention have delayed wake times and decreased sleep efficiency during the perinatal period: a brief report. Sleep Health . 2016;2:225–228. doi: 10.1016/j.sleh.2016.05.002. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Kapsimalis F, Kryger M. Sleep breathing disorders in the U.S. female population. J Womens Health (Larchmt) . 2009;18:1211–1219. doi: 10.1089/jwh.2008.1054. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Haas DM, Parker CB, Wing DA, Parry S, Grobman WA, Mercer BM, et al. NuMo Mbs A description of the methods of the Nulliparous Pregnancy Outcomes Study: monitoring mothers-to-be (nuMoM2b) Am J Obstet Gynecol . 2015;212:539.e1–539.e24. doi: 10.1016/j.ajog.2015.01.019. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Haas DM, Ehrenthal DB, Koch MA, Catov JM, Barnes SE, Facco F, et al. National Heart, Lung, and Blood Institute nuMoM2b Heart Health Study Network Pregnancy as a window to future cardiovascular health: design and implementation of the nuMoM2b heart health study. Am J Epidemiol . 2016;183:519–530. doi: 10.1093/aje/kwv309. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Facco FL, Parker CB, Reddy UM, Silver RM, Louis JM, Basner RC. NuMoM2b sleep-disordered breathing study: objectives and methods. Am J Obstet Gynecol . 2015;212:542.e1–127. doi: 10.1016/j.ajog.2015.01.021. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Facco F.Sleep disordered breathing in pregnancy and 2-7 years post-delivery: associations with hypertension and metabolic syndrome [abstract] Society for Maternal Fetal Medicine 202130194–198.. [Google Scholar]
- 20. Grundy SM, Cleeman JI, Daniels SR, Donato KA, Eckel RH, Franklin BA, et al. American Heart Association National Heart, Lung, and Blood Institute. Diagnosis and management of the metabolic syndrome: an American Heart Association/National Heart, Lung, and Blood Institute Scientific Statement. Circulation . 2005;112:2735–2752. doi: 10.1161/CIRCULATIONAHA.105.169404. [DOI] [PubMed] [Google Scholar]
- 21. Berry RB, Budhiraja R, Gottlieb DJ, Gozal D, Iber C, Kapur VK, et al. American Academy of Sleep Medicine Deliberations of the Sleep Apnea Definitions Task Force of the American Academy of Sleep Medicine. Rules for scoring respiratory events in sleep: update of the 2007 AASM Manual for the Scoring of Sleep and Associated Events. J Clin Sleep Med . 2012;8:597–619. doi: 10.5664/jcsm.2172. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. Zou G. A modified poisson regression approach to prospective studies with binary data. Am J Epidemiol . 2004;159:702–706. doi: 10.1093/aje/kwh090. [DOI] [PubMed] [Google Scholar]
- 23. Peppard PE, Young T, Palta M, Skatrud J. Prospective study of the association between sleep-disordered breathing and hypertension. N Engl J Med . 2000;342:1378–1384. doi: 10.1056/NEJM200005113421901. [DOI] [PubMed] [Google Scholar]
- 24. Ernst G, Bosio M, Salvado A, Dibur E, Nigro C, Borsini E. Difference between apnea-hypopnea index (AHI) and oxygen desaturation index (ODI): proportional increase associated with degree of obesity. Sleep Breath . 2016;20:1175–1183. doi: 10.1007/s11325-016-1330-3. [DOI] [PubMed] [Google Scholar]
- 25. Xu S, Wan Y, Xu M, Ming J, Xing Y, An F, et al. The association between obstructive sleep apnea and metabolic syndrome: a systematic review and meta-analysis. BMC Pulm Med . 2015;15:105. doi: 10.1186/s12890-015-0102-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26. Chai-Coetzer CL, Antic NA, Rowland LS, Catcheside PG, Esterman A, Reed RL, et al. A simplified model of screening questionnaire and home monitoring for obstructive sleep apnoea in primary care. Thorax . 2011;66:213–219. doi: 10.1136/thx.2010.152801. [DOI] [PubMed] [Google Scholar]
- 27. Weaver TE, Mancini C, Maislin G, Cater J, Staley B, Landis JR, et al. Continuous positive airway pressure treatment of sleepy patients with milder obstructive sleep apnea: results of the CPAP Apnea Trial North American Program (CATNAP) randomized clinical trial. Am J Respir Crit Care Med . 2012;186:677–683. doi: 10.1164/rccm.201202-0200OC. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28. Jenkinson C, Davies RJ, Mullins R, Stradling JR. Comparison of therapeutic and subtherapeutic nasal continuous positive airway pressure for obstructive sleep apnoea: a randomised prospective parallel trial. Lancet . 1999;353:2100–2105. doi: 10.1016/S0140-6736(98)10532-9. [DOI] [PubMed] [Google Scholar]
- 29. Faccenda JF, Mackay TW, Boon NA, Douglas NJ. Randomized placebo-controlled trial of continuous positive airway pressure on blood pressure in the sleep apnea-hypopnea syndrome. Am J Respir Crit Care Med . 2001;163:344–348. doi: 10.1164/ajrccm.163.2.2005037. [DOI] [PubMed] [Google Scholar]
- 30. Pepperell JC, Ramdassingh-Dow S, Crosthwaite N, Mullins R, Jenkinson C, Stradling JR, et al. Ambulatory blood pressure after therapeutic and subtherapeutic nasal continuous positive airway pressure for obstructive sleep apnoea: a randomised parallel trial. Lancet . 2002;359:204–210. doi: 10.1016/S0140-6736(02)07445-7. [DOI] [PubMed] [Google Scholar]
- 31. Haentjens P, Van Meerhaeghe A, Moscariello A, De Weerdt S, Poppe K, Dupont A, et al. The impact of continuous positive airway pressure on blood pressure in patients with obstructive sleep apnea syndrome: evidence from a meta-analysis of placebo-controlled randomized trials. Arch Intern Med . 2007;167:757–764. doi: 10.1001/archinte.167.8.757. [DOI] [PubMed] [Google Scholar]
- 32. Bratton DJ, Stradling JR, Barbé F, Kohler M. Effect of CPAP on blood pressure in patients with minimally symptomatic obstructive sleep apnoea: a meta-analysis using individual patient data from four randomised controlled trials. Thorax . 2014;69:1128–1135. doi: 10.1136/thoraxjnl-2013-204993. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33. Barbé F, Durán-Cantolla J, Sánchez-de-la-Torre M, Martínez-Alonso M, Carmona C, Barceló A, et al. Spanish Sleep And Breathing Network Effect of continuous positive airway pressure on the incidence of hypertension and cardiovascular events in nonsleepy patients with obstructive sleep apnea: a randomized controlled trial. JAMA . 2012;307:2161–2168. doi: 10.1001/jama.2012.4366. [DOI] [PubMed] [Google Scholar]
- 34. McEvoy RD, Antic NA, Heeley E, Luo Y, Ou Q, Zhang X, et al. SAVE Investigators and Coordinators CPAP for prevention of cardiovascular events in obstructive sleep apnea. N Engl J Med . 2016;375:919–931. doi: 10.1056/NEJMoa1606599. [DOI] [PubMed] [Google Scholar]