A total of 189 women consented to participate in the study (see Fig. 1). Sixty-three women were excluded because of insufficient data as a result of device failure or insufficient sleep time and two women were excluded because of twin pregnancies. A further two women had missing birth data. In total, 122 women were included in the final analyses.
There were no significant differences in age, smoking status, parity, ethnicity, history of asthma, hypertension, diabetes, previous stillbirth, preterm birth, pre-eclampsia or third trimester BMI between participants who were included (n = 122) or excluded (n = 67). There was a trend towards those who were included having a lower first trimester BMI, p = 0.050 (additional files, Table A1). Fewer included participants had gestational hypertension but no other clinical outcomes were different between included and excluded participants (additional files, Table A2).
Of the women included in the study, 16.4% (n = 20) had OSA (Table 1). Women with OSA had higher BMIs, by 6.3 kg/m2 and 7.7 kg/m2 in the first and third trimesters, respectively, and were more likely to report a history of asthma. There was no association between ethnicity and OSA (p value = 0.063). However, OSA prevalence was higher in women of Aboriginal and Torres Strait Islander descent (5/11 = 45.4%) compared with all other ethnicities combined (14/105 = 13.3%).
Table 1
Descriptive and clinical details of women with and without OSA
| OSA (AHI ≥ 5) n = 20 | No OSA (AHI < 5) n = 102 | Test statistic | p value |
Variable | Mean [SD] or Median [IQR} ororMedian [IQR] | Frequency n (%) | Mean [SD] or Median [IQR] | Frequency n (%) | | |
Age (years) | 31.2 [5.9] | | 31.1 [6.0] | | t(1) = 0.019 | 0.985 |
BMI (First trimester) (kg/m2) | 31.2 [10.6] | | 24.9 [6.5] | | U = 370 | p < 0.001*** |
BMI (Third trimester) (kg/m2) a | 36.0 [9.3] | | 28.3 [6.3] | | U = 318 | p < 0.001** |
Gestational age at sleep study (weeks) | 33.0 [6.3] | | 33.4 [6.0] | | U = 991.5 | p = 0.844 |
Smokers a | | 4 (20.0) | | 8 (8.2) | Fishers exact | p = 0.120 |
Parity > 1 | | 14 (70.0) | | 51 (50.0) | χ2(1) = 2.687 | p = 0.101 |
Ethnicity a Caucasian Asian African Indian Middle Eastern Aboriginal Pacific Islander | | 9 (47.4) 1 (5.3) 2 (10.5) 0 (0.0) 0 (0.0) 5 (26.3) 2 (10.5) | | 61(62.9) 11 (11.3) 5 (5.2) 5 (5.2) 5 (5.2) 6 (6.2) 4 (4.1) | χ2(6) = 11.968 | p = 0.063 |
History of asthma a | | 9 (47.4) | | 18 (17.8)) | χ2(1) = 8.006 | p = 0.005* |
Chronic hypertension | | 2 (10.0) | | 8 (7.8) | Fishers exact | p = 0.668 |
Pre-existing diabetes | | 2 (10.0) | | 1 (1.0) | Fishers exact | p = 0.070 |
History of stillbirth b | | 2 (14.3) | | 4 (7.8) | Fishers exact | p = 0.602 |
Previous preterm birth b | | 4 (28.6) | | 15 (29.4) | Fishers exact | p = 1.000 |
Previous preeclampsia a,b | | 4 (30.8) | | 7 (14.3) | Fishers exact | p = 0.221 |
IQR, Interquartile range; SD, Standard deviation; BMI, Body mass index; t, t-statistic students t test, U, U-statistic Mann Whitney U test; X 2 , Chi-squared; OSA, Obstructive sleep apnoea; AHI, Apnoea hypopnoea index
a Missing data for some variables including: third trimester BMI which was available for 119 women (18 with OSA and 101 without OSA); smoking status was available for 118 women (20 with OSA and 98 without OSA); ethnicity was available for 116 women (19 with OSA and 97 without OSA); history of asthma was available for 120 women (19 with OSA and 101 without OSA); pre-existing diabetes was available for 121 women (20 women with OSA and 101 without); and previous preeclampsia was available for 62 multiparous women (13 with OSA and 49 without OSA)
b Data are presented for 65 multiparous women (14 with OSA and 51 without OSA)
Maternal outcomes and obstructive sleep apnoea
No relationship was found between gestational hypertension, pre-eclampsia, gestational diabetes, diabetes requiring insulin or caesarean delivery and OSA (Table 2). The odds of women being diagnosed with maternal postnatal sepsis during their hospital stay were 3.7 times higher in women with OSA. However, this relationship was not maintained after adjusting for BMI and depressive symptoms. Details regarding maternal postnatal sepsis separated by cause in those with and without OSA is available in additional files, Table A3.
Table 2
Summary of Logistic Regression Analysis for Variables Predicting Maternal Complications due to OSA accounting for depressive symptoms.
Dependent | Independent | Observed numbers | Unadjusted | Adjusted |
| | OSA | No OSA | OR [95% C.I.’s] | p | OR [95% C.I.’s] | p |
Hypertension | | | | | | | |
| OSA (AHI ≥ 5) | 4/20 (20.0%) | 17/102 (16.7%) | 1.24 [0.37, 4.16] | 0.733 | 1.85 [0.39, 8.79] | 0.437 |
| BMI (kg/m2) | | | | | 1.03 [0.94, 1.12] | 0.554 |
| EPDS | | | | | 0.85 [0.74, 0.97] | 0.019* |
Gestational hypertension | | | | | | | |
| OSA (AHI ≥ 5) | 2/20 (10.0%) | 9/102 (8.8%) | 1.15 [0.23, 5.77] | 0.867 | 1.97 [0.26, 14.68] | 0.509 |
| BMI (kg/m2) | | | | | 1.01 [0.90, 1.13] | 0.859 |
| EPDS | | | | | 0.83 [0.69, 1.01] | 0.061 |
Pre-eclampsia | | | | | | | |
| OSA (AHI ≥ 5) | 3/20 (15.0%) | 7/102 (6.9%) | 2.40 [0.56, 10.19] | 0.237 | 2.07 [0.32, 13.39] | 0.445 |
| BMI (kg/m2) | | | | | 1.05 [0.95, 1.17] | 0.352 |
| EPDS | | | | | 0.94 [0.81, 1.09] | 0.426 |
Gestational Diabetes | | | | | | | |
| OSA (AHI ≥ 5) | 4/20 (20.0%) | 23/102 (22.5%) | 0.80 [0.26, 2.82] | 0.802 | .32 [0.07, 1.51] | 0.150 |
| BMI (kg/m2) | | | | | 1.09 [1.00, 1.18] | 0.043 |
| EPDS | | | | | 1.07 [0.97, 1.16] | 0.166 |
Diabetes requiring insulin | | | | | | | |
| OSA (AHI ≥ 5) | 3/20 (15.0%) | 8/102 (7.8%) | 2.07 [0.50, 8.61] | 0.315 | 0.99 [0.15, 6.37] | 0.991 |
| BMI (kg/m2) | | | | | 1.06 [0.95, 1.18] | 0.318 |
| EPDS | | | | | 1.07 [0.95, 1.20] | 0.274 |
Maternal postnatal sepsis | | | | | | | |
| OSA (AHI ≥ 5) | 7/19 (36.8%) | 14/101 (13.9%) | 3.65 [1.21, 10.78] | 0.021* | 3.10 [0.80, 11.96] | 0.101 |
| BMI (kg/m2) | | | | | 1.02 [0.93, 1.11] | 0.718 |
| EPDS | | | | | 1.01 [0.92, 1.11] | 0.840 |
Caesarian delivery | | | | | | | |
| OSA (AHI ≥ 5) | 12/20 (60.0%) | 40/100 (40.0%) | 2.25 [0.84, 6.00] | 0.105 | 2.8 [.93, 8.83] | 0.068 |
| BMI (kg/m2) | | | | | .97 [.91, 1.04] | 0.532 |
| EPDS | | | | | 1.04 [0.96, 1.12] | 0.358 |
AHI, Apnoea hypopnoea index; EPDS, Edinburgh Postnatal Depression Scale; BMI, Body mass index; OR, Odds Ratio [95% Confidence Interval]. OSA, hypertension, gestational hypertension, pre-eclampsia, gestational diabetes, diabetes requiring insulin, sepsis and caesarean delivery are each coded as 1 for yes and 0 for no. First trimester BMI and EPDS are continuous, *p < 0 .05.
In addition to analysing OSA at the clinical cut-off (AHI ≥ 5), increasing AHI (as a continuous variable) was investigated in a separate analysis as a predictor of gestational hypertension, pre-eclampsia, gestational diabetes, diabetes requiring insulin, maternal postnatal sepsis and caesarean delivery, both independently and after adjusting for first trimester BMI and depressive symptoms measured on the EPDS (additional files, Table 4). Increasing AHI was related to maternal postnatal sepsis although the relationship was not maintained after adjusting for BMI and depressive symptoms.
In a separate analysis (additional files, Table A5), increasing ODI3% was investigated as a predictor of maternal complications. A model accounting for depressive symptoms showed gestational diabetes was associated with increasing BMI and decreasing ODI3%. Additionally, in a multivariate model accounting for ODI3% and EPDS, increased BMI was associated with pre-eclampsia. No other associations were found in this cohort.
Newborn outcomes and obstructive sleep apnoea
After adjusting for BMI and depressive symptoms, maternal OSA was associated with 18 times greater odds of having an infant with an Apgar score of < 7 at 5 minutes after birth, but not at 1 minute after birth. A greater proportion of mothers with OSA also had infants with transient tachypnoea of the newborn, although this relationship was not maintained after accounting for BMI, depressive symptoms and caesarean delivery (Table 3).
Table 3
Summary of logistic regression analysis for variables predicting newborn complications due to OSA.
Dependent | Independent | Observed numbers | Unadjusted | Adjusted |
Predictor | | OSA | No OSA | OR [95% C.I.’s] | p | OR [95% C.I.’s] | p |
Preterm birth | | | | | | | |
| OSA (AHI ≥ 5) | 6/20 (30.0%) | 34/102 (33.3) | 0.86 [0.30, 2.43] | 0.772 | 0.40 [0.10, 1.58] | 0.191 |
| BMI (kg/m2) | | | | | 1.01 [0.93, 1.09] | 0.911 |
| EPDS | | | | | 1.17 [1.06, 1.28] | 0.001** |
Special care nursery admission | | | | | | | |
| OSA (AHI ≥ 5) | 13/20 (65.0%) | 51/101 (50.4%) | 1.82 [0.67, 4.94] | 0.239 | 1.10 [0.33, 3.62] | 0.877 |
| BMI (kg/m2) | | | | | 1.06 [0.98, 1.15] | 0.165 |
| EPDS | | | | | 1.04 [0.97, 1.12] | 0.298 |
Transient tachypnoea | | | | | | | |
| OSA (AHI ≥ 5) | 4/20 (20.0%) | 1/100 (1.00%) | 24.8 [2.60, 235.77] | 0.005** | 6.48 [0.40, 106.30] | 0.190 |
| BMI (kg/m2) | | | | | 1.16 [0.98, 1.36] | 0.076 |
| Caesarean | | | | | 2.28 [0.16, 31.68] | 0.540 |
| EPDS | | | | | 1.01 [0.84, 1.21] | 0.890 |
Apgar < 7 at 1 minute | | | | | | | |
| OSA (AHI ≥ 5) | 6/20 (30.0%) | 18/101 (17.8%) | 1.87 [0.64, 5.51] | 0.254 | 1.29 [0.31, 4.64] | 0.787 |
| BMI (kg/m2) | | | | | 1.03 [0.95, 1.12] | 0.425 |
| EPDS | | | | | 1.04 [0.96, 1.14] | 0.334 |
Apgar < 7 at 5 minutes | | | | | | | |
| OSA (AHI ≥ 5) | 3/20 (15.0%) | 1/101 (0.99%) | 17.82 [1.75, 181.51] | 0.015* | 18.41 [1.17, 289.29] | 0.038* |
| BMI (kg/m2) | | | | | 1.00 [.86, 1.17] | 0.861 |
| EPDS | | | | | 0.98 [0.81, 1,19] | 0.945 |
Neonatal problems | | | | | | | |
| OSA (AHI ≥ 5) | 16/20 (80.0%) | 57/99 (57.6%) | 2.06 [0.68, 6.07] | 0.206 | 1.26 [0.34, 4.72] | 0.734 |
| BMI (kg/m2) | | | | | 1.03 [0.95, 1.10] | 0.522 |
| EPDS | | | | | 1.07 [0.98, 1.17] | 0.118 |
AHI, Apnoea hypopnoea index; EPDS, Edinburgh Postnatal Depression Scale; BMI, Body mass index; OR, Odds Ratio [95% Confidence Interval]; OSA, preterm birth, special care nursery admission, transient tachypnoea, Apgar < 7 at 1 minute, Apgar < 7 at 5 minutes and neonatal problems are each coded as 1 for yes and 0 for no. First trimester BMI and EPDS are continuous, *p < 0 .05, **p < 0 .01.
Considering AHI as a continuous predictor, univariate analysis identified that increasing maternal AHI was associated with having an infant requiring admission to special care nursery, suffering from transient tachypnoea of the newborn, having an Apgar score of < 7 at 5 minutes after birth and having neonatal problems (see Table 4). After adjusting for BMI and depressive symptoms, these relationships were no longer significant.
Table 4
Summary of logistic regression analysis for newborn complications predicted by apnoea hypopnoea index.
Dependent | Independent | Unadjusted | Adjusted |
| | OR [95% C.I.’s] | p | OR [95% C.I.’s] | p |
Preterm birth | | | | | |
| AHI | 1.53 [0.56, 4.19] | 0.407 | 1.09 [0.31, 3.83] | 0.899 |
| BMI (kg/m2) | | | 0.98 [0.90, 1.05] | 0.543 |
| EPDS | | | 1.14 [1.04, 1.24] | 0.004** |
Special care nursery admission | | | | | |
| AHI | 2.98 [1.02, 8.78] | 0.047* | 2.16 [0.63, 7.38] | 0.218 |
| BMI (kg/m2) | | | 1.02 [0.95, 1.10] | 0.552 |
| EPDS | | | 1.04 [0.96, 1.13] | 0.301 |
Transient tachypnoea | | | | | |
| AHI | 14.36 [2.30, 89.67] | 0.004** | 3.05 [0.25, 37.02] | 0.382 |
| BMI (kg/m2) | | | 1.17 [1.01, 1.35] | 0.041* |
| Caesarian | | | 1.00 [1.00, 1.01] | 0.336 |
| EPDS | | | 1.11 [0.94, 1.31] | 0.226 |
Apgar < 7 at 1 minute | | | | | |
| AHI | 2.76 [0.91, 8.42] | 0.074 | 2.16 [0.57, 8.17] | 0.257 |
| BMI (kg/m2) | | | 1.02 [0.94, 1.10] | 0.682 |
| EPDS | | | 1.03 [0.95, 1.13] | 0.468 |
Apgar < 7 at 5 minutes | | | | | |
| AHI | 9.15 [1.32, 63.35] | 0.025* | 6.68 [0.59, 75.19] | 0.124 |
| BMI (kg/m2) | | | 1.03 [0.89, 1.19] | 0.707 |
| EPDS | | | 1.03 [0.86, 1.24] | 0.727 |
Neonatal problems | | | | | |
| AHI | 3.63 [1.09, 12.09] | 0.036* | 2.78 [0.70, 11.11] | 0.148 |
| BMI (kg/m2) | | | 1.00 [0.93, 1.08] | 0.908 |
| EPDS | | | 1.06 [0.97, 1.15] | 0.231 |
AHI, Apnoea hypopnoea index; EPDS, Edinburgh Postnatal Depression Scale; BMI, Body mass index; OR, Odds Ratio [95% Confidence Interval]; Preterm birth, special care nursery admission, transient tachypnoea, Apgar < 7 at 1 minute, Apgar < 7 at 5 minutes and neonatal problems are each coded as 1 for yes and 0 for no. First trimester BMI, AHI and EPDS are continuous, *p < 0 .05, **p < 0 .01.
In a separate analysis (also presented in additional files, Table A6), increasing ODI3% was investigated as a predictor of newborn complications. ODI3% was also related to transient tachypnoea of the newborn in univariable analysis, although the relationship was not maintained after accounting for BMI and EPDS.
Interestingly, preterm birth (birth at < 37 weeks of gestation) was not associated with OSA, AHI or ODI3% but was associated with increasing depressive symptoms (additional files, Tables A3, A4 and A6), and this relationship was maintained after controlling for BMI. To investigate this further, analysis of the relationship between preterm birth and the clinically defined dichotomous EPDS variable (EPDS ≥ 13) was also undertaken. The odds of preterm birth in women with EPDS ≥ 13 was found to be 6.58 [1.64, 26.41], p = 0.008. After adjusting for BMI, the odds of preterm birth in women with EPDS ≥ 13 were 6.85 [1.57, 29.95], p = 0.011.
Birthweight
Birthweight percentiles were positively related to AHI and ODI3% (see Table 5) in a separate univariable analysis. After accounting for BMI and EPDS, the model using ODI3% remained moderately (r = 0.31) positively associated with increased birthweight (p < 0.01).
Table 5
Summary of regression analyses for birthweight percentile predicted by apnoea hypopnoea index and ODI3% in separate analysis.
| | Birthweight (Z score) |
Model | | Unadjusted | Adjusted |
| Variable | B[95%CI] | SE B | p | B | SE B | p |
1 | AHI | 0.70 [0.14, 1.25] | 0.28 | 0.015* | 0.61 [-0.04, 1.27] | 0.33 | 0.066 |
EPDS | | | | -0.01 [-0.02, 0.06] | 0.02 | 0.548 |
BMI (kg/m2) | | | | 0.18 [-0.02, 0.06] | 0.02 | 0.339 |
R2 | 0.05 | 0.06 2.52 |
F (114) | 6.16 |
2 | ODI3% | 0.79 [0.33, 1.26] | 0.24 | 0.001** | 0.75 [0.20, 1,29] | 0.27 | 0.008** |
EPDS | | | | -0.01 [-0.05, 0.03] | 0.02 | 0.624 |
BMI (kg/m2) | | | | 0.01 [-0.03, 0.05] | 0.02 | 0.622 |
R2 | 0.09 11.29 | 0.10 3.89 |
AHI, Apnoea hypopnoea index; EPDS, Edinburgh Postnatal Depression Scale; BMI, Body mass index; ODI, Oxygen desaturation index; F, F test statistic (degrees freedom);
*p < 0 .05, **p < 0 .01
Cord blood gases
Cord blood gas results were available for 74 of the 122 women included in the study. No significant relationships were found between either having an umbilical cord blood pH ≤ 7.2 (see additional files, Table A7) or increasing umbilical cord blood lactate levels (continuous variable) (see additional files, Table A8) and either AHI or ODI3% in univariable models, or following adjustment for caesarean delivery, gestational age, BMI and depressive symptoms.