This study excluded cases of chromosomal abnormalities, twin-to-twin transfusion syndrome, twin anaemia-polycythemia sequence, hydropic foetuses, and pregnant women with previous diabetes. After excluding participants with missing information, 4724 women with twin pregnancies were included in the analysis, with 3053 having a normal prepregnancy BMI, 1123 being overweight/obese, and 548 being underweight. The selection process is shown in Fig. 1.
Table 1 shows that women who were overweight/obese before pregnancy had certain characteristics compared to those with normal prepregnancy weight. They were older, undergone more pregnancies and deliveries, had experienced more nonnatural pregnancies, had more dichorionic diamniotic twin pregnancies, were more likely to develop gestational diabetes and gestational hypertension, and had fewer caesarean sections and less weight gain during pregnancy. Similar results were observed for women with insufficient prepregnancy weight (Table S2).
Table 1
Baseline characteristics of the unmatched sample, PSM sample, and OW sample for comparison between pre-pregnancy normal weight and overweight/obesity.
Characteristics | Unmatched Sample (n = 4176) | p value | PSM (n = 2158) | p value | OW (n = 1576) | p value |
Normal weight (n = 3053) | Overweight/obesity (n = 1123) | Normal weight (n = 1079) | Overweight/obesity (n = 1079) | Normal weight (n = 788) | Overweight/ obesity (n = 788) |
Advanced maternal age | 722 (23.6) | 324 (28.9) | 0.001 | 282 (26.1) | 299 (27.7) | 0.437 | 212.8 (27.0) | 212.8 (27.0) | 1 |
Multiparous | 1138 (37.3) | 467 (41.6) | 0.012 | 461 (42.7) | 440 (40.8) | 0.383 | 318.0 (40.4) | 318.0 (40.4) | 1 |
Parity | | | 0.005 | | | 0.886 | | | 1 |
≤ 1 | 1,357 (44.4) | 440 (39.2) | | 421 (39.0) | 430 (39.9) | | 320.4 (40.7) | 320.4 (40.7) | |
2 | 808 (26.5) | 307 (27.3) | | 294 (27.2) | 295 (27.3) | | 214.4 (27.2) | 214.4 (27.2) | |
≥ 3 | 888 (29.1) | 376 (33.5) | | 364 (33.7) | 354 (32.8) | | 253.1 (32.1) | 253.1 (32.1) | |
Ethnicity | 11 (3.6) | 48 (4.3) | 0.362 | 31 (2.9) | 44 (4.1) | 0.158 | 31.6 (4.0) | 31.6 (4.0) | 1 |
Mode of conception | 1,914 (62.7) | 756 (67.3) | 0.006 | 731 (67.7) | 720 (66.7) | 0.647 | 520.3 (66.0) | 520.3 (66.0) | 1 |
Mode of birth | 2,827 (92.6) | 1,018 (90.7) | 0.045 | 991 (91.8) | 980 (90.8) | 0.444 | 718.3 (91.2) | 718.3 (91.2) | 1 |
Pregnancy weight gain | 1,509 (49.4) | 415 (37.0) | < 0.001 | 411 (38.1) | 409 (37.9) | 0.965 | 317.5 (40.3) | 317.5 (40.3) | 1 |
Adverse pregnancy history | 1,297 (42.5) | 513 (45.7) | 0.067 | 491 (45.5) | 487 (45.1) | 0.897 | 351.5 (44.6) | 351.5 (44.6) | 1 |
Hypertensive disorder of pregnancy | 440 (14.4) | 257(22.9) | < 0.001 | 208 (19.3) | 217 (20.1) | 0.665 | 154.1 (19.6) | 154.1 (19.6) | 1 |
Gestational diabetes | 676 (22.1) | 353 (31.4) | < 0.001 | 317 (29.4) | 316 (29.3) | 1 | 223.2 (28.3) | 223.2 (28.3) | 1 |
Second and third trimester infections | 202 (6.6) | 59 (5.3) | 0.113 | 44 (4.1) | 56 (5.2) | 0.26 | 43.5 (5.5) | 43.5 (5.5) | 1 |
Abnormal thyroid function | 333 (10.9) | 137 (12.2) | 0.246 | 128 (11.9) | 129 (12.0) | 1 | 92.9 (11.8) | 92.9 (11.8) | 1 |
Anemia | 866 (28.4) | 286 (25.5) | 0.066 | 260 (24.1) | 281 (26.0) | 0.321 | 207.7 (26.4) | 207.7 (26.4) | 1 |
Chorionicity | 2,357 (77.2) | 909 (80.9) | 0.010 | 869 (80.5) | 871 (80.7) | 0.957 | 628.9 (79.8) | 628.9 (79.8) | 1 |
PSM, propensity score-matched; OW, overlap weighting. |
1. Primary outcomes
1.1 Normal prepregnancy weight vs. overweight/obesity
1.1.1 Bivariable Analysis
In the unmatched sample of 4176 twins, those born to mothers who were overweight/obese before pregnancy exhibited a higher incidence of 1-min Apgar scores ≤ 7 compared to twins born to mothers with normal weight (7.7% vs. 5.1% for larger twin, and 7.6% vs. 5.2% for smaller twin). The crude ORs (95% CIs) were 1.57 (1.19–2.06) and 1.49 (1.13–1.95), respectively, as determined by the unadjusted logistic regression model (Model 1).
After controlling for confounding variables by multivariate logistic regression (Model 2), the association between maternal prepregnancy overweight/obesity and an increased risk of low 1-min Apgar scores for twin infants remained significant, with ORs (95% CIs) of 1.60 (1.20–2.13) and 1.45 (1.09–1.92) for larger and smaller twins, respectively (Table 2).
Table 2
The results of OR (95% CIs) for comparisons between pre-pregnancy normal weight and overweight/obesity in each model.
| Normal weight (n = 3053) | Overweight/ Obesity (n = 1123) | OR (95%CIs) |
| Crude | Adjusted | PSM | IPTW | OW |
Total (n/%) | | | | | | | |
GA < 37 weeks | 1909 (62.5) | 723 (64.4) | 1.08 (0.94–1.25) | 0.99 (0.84–1.15) | 0.99 (0.82–1.20) | 1.00 (0.87–1.13) | 0.99 (0.80–1.23) |
GA < 34 weeks | 483 (15.8) | 215 (19.1) | 1.25 (1.05–1.50) | 1.21 (0.99–1.47) | 1.19 (0.93–1.52) | 1.20 (1.02–1.42) | 1.21 (0.91–1.61) |
BWDT ≥ 20% | 520 (17.0) | 191 (17.0) | 1.00 (0.83–1.20) | 0.96 (0.80–1.16) | 0.92 (0.74–1.15) | 0.95 (0.81–1.11) | 0.96 (0.74–1.25) |
Larger twin | | | | | | | |
1 min Apgar ≤ 7 | 155 (5.1) | 87 (7.7) | 1.57 (1.19–2.06) | 1.60 (1.20–2.13) | 1.60 (1.11–2.33) | 1.67 (1.31–2.12) | 1.65 (1.08–2.57) |
5 min Apgar ≤ 7 | 24 (0.8) | 16 (1.4) | 0.93 (0.27–2.41) | 0.82 (0.24–2.17) | 0.62 (0.15–2.26) | 0.80 (0.26–1.96) | 0.82 (0.17–3.71) |
NICU admission | 1426 (46.7) | 531 (47.3) | 1.02 (0.89–1.17) | 0.961 (0.83–1.11) | 0.95 (0.80–1.14) | 0.97 (0.86–1.10) | 0.96 (0.78–1.19) |
Low Birth weight | 1365 (44.7) | 468 (41.7) | 0.88 (0.77–1.01) | 0.80 (0.68–0.92) | 0.76 (0.63–0.91) | 0.80 (0.71–0.91) | 0.95 (0.81–1.11) |
Smaller twin | | | | | | | |
1 min Apgar ≤ 7 | 159 (5.2) | 85 (7.6) | 1.49 (1.13–1.95) | 1.45 (1.09–1.92) | 1.55 (1.07–2.25) | 1.48 (1.17–1.87) | 1.47 (0.97–2.25) |
5 min Apgar ≤ 7 | 24 (0.8) | 15 (1.3) | 1.71 (0.87–3.24) | 1.40 (0.70–2.73) | 0.96 (0.43–2.14) | 1.46 (0.84–2.52) | 1.42 (0.53–4.02) |
NICU admission | 1536 (50.3) | 576 (51.3) | 1.04 (0.91–1.19) | 0.96 (0.83–1.11) | 0.96 (0.81–1.15) | 0.96 (0.85–1.09) | 0.96 (0.78–1.18) |
Low Birth weight | 2188 (71.7) | 770(68.6) | 0.86 (0.74-1.00) | 0.78 (0.67–0.91) | 0.75 (0.62–0.91) | 0.78 (0.68–0.89) | 0.78 (0.62–0.98) |
Complications of pregnancy | | | | | | | |
Hypertensive disorder of pregnancy | 440 (14.4) | 257 (22.9) | 1.71 (1.43–2.03) | 1.85 (1.55–2.21) | 1.73 (1.38–2.16) | 1.87 (1.61–2.18) | 1.85 (1.43–2.42) |
Gestational diabetes | 676 (22.1) | 353 (31.4) | 1.61 (1.38–1.88) | 1.49 (1.27–1.74) | 1.47 (1.21–1.78) | 1.48 (1.30–1.68) | 1.49 (1.19–1.86) |
Second and third trimester infections | 202 (6.6) | 59 (5.3) | 0.78 (0.58–1.05) | 0.76 (0.56–1.02) | 0.68 (0.48–0.96) | 0.75 (0.58–0.97) | 0.76 (0.50–1.15) |
Abnormal thyroid function | 333 (10.9) | 137 (12.2) | 1.13 (0.92–1.40) | 1.14 (0.92–1.42) | 1.04 (0.81–1.35) | 1.15 (0.96–1.38) | 1.14 (0.84–1.56) |
Anemia | 866 (28.4) | 286 (25.5) | 0.86 (0.74–1.01) | 0.83 (0.71–0.97) | 0.84 (0.69–1.01) | 0.84 (0.73–0.95) | 0.83 (0.66–1.03) |
BWDT, birth weight discordance in twins; CIs, confidence intervals; GA, gestational age; IPTW, inverse probability treatment weighting; OR, odds ratios; OW, overlap weighting; PSM, propensity score-matched. |
1.1.2 Propensity Score-Matched Analysis
To investigate the impact of maternal prepregnancy BMI on the 1-min Apgar score of twin infants, it was crucial to eliminate confounding factors such as preeclampsia and gestational diabetes, which have been indicated to affect Apgar scores.14, 15 We performed a 1:1 PSM analysis (Model 3). As shown in Fig. 2, all baseline variables had SMD values of less than 0.1, indicating a balanced state between the two groups. After matching, the analysis showed that twin infants born to women with prepregnancy overweight/obesity had a higher risk of 1-min Apgar score ≤ 7 than those born to women with normal prepregnancy weight, with ORs (95% CIs) of 1.60 (1.11–2.33) and 1.55 (1.07–2.25), respectively (Table 2).
1.1.3 Inverse Probability Treatment Weighting or Overlap Weighting Analysis
After IPTW or OW, weighted multivariable logistic regression analysis showed that maternal overweight/obesity before pregnancy could increase the risk of Apgar scores ≤ 7 at 1 minute in large twin, with ORs (95% CI) of 1.67 (1.31–2.12) and 1.65 (1.08–2.57), respectively (Table 2, Table S3). Similar results were observed for smaller twin in the analysis with IPTW, while the analysis with OW almost reached significance, with ORs (95% CI, p value) of 1.48 (1.17–1.87, 0.001) and 1.46 (0.97–2.25, 0.07), respectively (Table 2, Table S3). The weighted baseline characteristics were poorly balanced for IPTW and well-balanced for OW (shown in Figure. 2).
1.2 Normal prepregnancy weight vs. underweight
Similar findings were observed when comparing prepregnancy underweight and normal prepregnancy weight. Specifically, maternal prepregnancy underweight was associated with a decreased risk of 1-min Apgar score ≤ 7 in twins. In larger twin, the ORs (95% CIs) values were 0.56 (0.32–0.92) (Model 1), 0.50 (0.28–0.83) (Model 2), 0.48 (0.24–0.89) (Model 3), 0.49 (0.29–0.78) (Model 4), and 0.50 (0.24–0.98) (Model 5) (Table S4). In smaller twin, the ORs (95% CIs) values in Models 1–5 were 0.58 (0.34–0.94), 0.56 (0.32–0.91), 0.49 (0.25–0.91), 0.55 (0.33–0.86), and 0.55 (0.27–1.07), respectively (Table S4). The balance of the baseline characteristics between the groups before and after matching or weighting is presented in Figure S1.
2. Other outcomes
Furthermore, the study also examined the impact of different prepregnancy BMI values on various outcomes, including 5-min Apgar scores ≤ 7, NICU admission, preterm birth (GA < 37 weeks), early preterm birth (GA < 34 weeks), LBW, and BWDT. Compared to normal prepregnancy BMI, prepregnancy overweight/obesity had a protective effect on smaller twin, reducing the likelihood of them having LBW. However, maternal normal weight before pregnancy did not have a significant effect on other outcomes, such as 5-min Apgar scores ≤ 7, NICU admission, preterm birth, early preterm birth, and BWDT, when compared to maternal prepregnancy overweight/obesity or underweight (see Table 2 and Table S4).
3. Prepregnancy BMI and pregnancy complications.
Table 1 shows that different BMI values are associated with maternal pregnancy-related diseases. According to the analysis of Models 1–5, compared with normal prepregnancy weight, maternal overweight/obesity before pregnancy not only increased the risk of developing gestational diabetes during pregnancy, with ORs (95% CIs) of 1.61 (1.38–1.88), 1.49 (1.27–1.74), 1.47 (1.21–1.78), 1.48 (1.30–1.68), and 1.49 (1.19–1.86), respectively, but also increased the likelihood of developing gestational hypertension, with ORs (95% CIs) of 1.71 (1.43–2.03), 1.85 (1.55–2.21), 1.73 (1.38–2.16), 1.87 (1.61–2.18), and 1.85 (1.43–2.42) (Table 2 and Table S5). Conversely, prepregnancy underweight did not pose such risks (Tables S4 and S6). Furthermore, different prepregnancy BMI values did not exhibit significant effects on the occurrence of thyroid function abnormalities, anaemia, and infections during mid-late pregnancy.
4. Sensitivity analysis
To ensure the reliability of the findings on the impact of different prepregnancy BMI on the 1-min Apgar score, sensitivity analysis was conducted by stratifying twins based on GA or BWDT status. The results showed that compared to normal weight, maternal prepregnancy overweight/obesity significantly increased the risk of having 1-min Apgar score ≤ 7 in twins with a GA less than 34 weeks (Table 3). After stratified for BWDT, the risk of prepregnancy weight failed to show consistent significance for 1-min Apgar scores ≤ 7 in twins (Table 3). Similar results were obtained when comparing maternal prepregnancy underweight with normal weight (Table S7). In addition, sensitivity analyses were also performed by adjusting the matching algorithm using different calliper values (0.02, 0.1, 0.2), ratios (1, 2, 3), or by changing to a probit model. However, the statistical significance of the results remained unchanged, indicating the robustness of our effect estimates (data not shown).
Table 3
Summary of Sensitivity Analysis between pre-pregnancy normal weight and overweight/obesity.
| Normal weight (n = 3053) | Overweight/ Obesity (n = 1123) | OR (95%CIs) |
| Crude | Adjusted | PSM* | IPTW# | OW& |
Larger twin (n/%) | | | | | | | |
GA>37 weeks | 11 (1.0) | 3 (0.8) | 0.78 (0.18–2.51) | 0.73 (0.14–2.62) | 3.82 (0.49–2.16) | 7.85 (0.21–2.37) | 0.92 (0.12–6.64) |
34–36 weeks | 61 (4.3) | 20 (3.9) | 0.92 (0.53–1.51) | 1.09 (0.63–1.83) | 0.72 (0.37–1.38) | 1.12 (0.70–1.74) | 1.12 (0.51–2.45) |
GA < 34 weeks | 83 (17.2) | 64 (29.8) | 2.04 (1.40–2.97) | 1.94 (1.30–2.89) | 1.90 (1.17–3.12) | 1.98 (1.43–2.73) | 1.94 (1.08–3.53) |
BWDT ≥ 20% | 128 (5.1) | 76 (8.2) | 1.12 (0.52–2.24) | 1.16 (0.52–2.42) | 0.90 (0.33–2.40) | 1.14 (0.59–2.11) | 1.22 (0.39–3.86) |
BWDT<20% | 27 (5.2) | 11 (5.8) | 1.67 (1.24–2.23) | 1.73 (1.26–2.35) | 1.87 (1.23–2.87) | 1.83 (1.41–2.37) | 1.79 (1.12–2.91) |
Smaller twin | | | | | | | |
GA>37 weeks | 16 (1.4) | 4 (1.0) | 0.71 (0.20–1.96) | 0.88 (0.24–2.49) | 0.46 (0.11–1.63) | 0.85 (0.28–2.17) | 0.86 (0.16–4.51) |
34–36 weeks | 63 (4.4) | 18 (3.5) | 0.79 (0.45–1.33) | 0.89 (0.50–1.51) | 0.97 (0.48–1.97) | 0.88 (0.55–1.39) | 0.89 (0.40–1.94) |
GA < 34 weeks | 80 (16.6) | 63 (29.3) | 2.09 (1.43–3.05) | 1.97 (1.32–2.93) | 1.71 (1.07–2.78) | 1.94 (1.41–2.67) | 1.97 (1.11–3.57) |
BWDT ≥ 20% | 133 (5.3) | 76 (8.2) | 2.10 (1.12–3.87) | 2.34 (1.20–4.47) | 2.20 (0.89–5.86) | 2.35 (1.35–4.09) | 2.34 (0.88–6.84) |
BWDT<20% | 26 (5.0) | 66 (34.6) | 1.38 (1.01–1.86) | 1.30 (0.94–1.78) | 1.15 (0.77–1.71) | 1.34 (1.02–1.74) | 1.32 (0.82–2.12) |
BWDT, birth weight discordance in twins; CIs, confidence intervals; GA, gestational age; IPTW, inverse probability treatment weighting; OR, odds ratios; OW, overlap weighting; PSM, propensity score-matched. |