Abstract
Background
Women with a history of gestational diabetes (GDM) have an increased risk of cardiovascular disease (CVD) throughout their lifetime. It is still unclear whether adhering to a healthy lifestyle can modify the association of GDM with the risk of CVD and mortality.
Methods
This study included 125,435 parous women from the UK Biobank prospective cohort. The history of GDM was determined by self-reported diagnosis or hospital admission records. A healthy lifestyle score was defined by incorporating self-reported information on five modifiable risk factors, including smoking, alcohol intake, physical activity, diet, and sleep duration. The primary outcome was a composite of major CVD and all-cause mortality.
Results
The mean age was 56.4 ± 7.9 yrs, and 668 participants had a history of GDM. After a median follow-up of 13.6 years, 9371 had major CVD events, and 6750 died. The association between GDM history with the composite of major CVD and all-cause mortality was stronger among women with the least healthy lifestyles (HR, 2.35 [95% CI, 1.72–3.22]) compared to those with moderately healthy (1.31 [0.84–2.06]) or the healthiest lifestyles (1.24 [0.75–2.06]; P = 0.001 for interaction). The relative excess risk due to interaction between GDM history and the least healthy lifestyles was 0.52 (0.27–0.77; P = 0.02). Compared to women with no GDM history and the healthiest lifestyle, those with GDM history and the least healthy lifestyle had a threefold increased risk of the composite outcome (3.00 [2.20–4.10]), while women with GDM history and the healthiest lifestyle did not experience a significantly higher risk (1.19 [0.71–1.97]).
Conclusions
Women with a history of GDM did not experience a higher risk of all-cause mortality and major CVD when adhering to a healthy lifestyle in midlife.
Keywords: Gestational diabetes, Cardiovascular disease, Mortality, Lifestyle
Graphical abstract
1. Introduction
Gestational diabetes mellitus (GDM) is one of the most common complications of pregnancy, which is estimated to influence 14.7 % of pregnant women worldwide [1]. GDM not only affects maternal and perinatal outcomes during pregnancy and delivery [2,3], but is also associated with long-term risks of progressing to cardiometabolic disorders throughout the lifespan for both mothers and offspring [[4], [5], [6], [7], [8], [9], [10]].
However, current evidence on the association of GDM with long-term cardiovascular disease (CVD) and mortality outcomes was mainly from retrospective cohorts, case-control studies, national registries, and administrative databases, with a lack of information on lifestyle risk factors [7]. Thus far, evidence regarding healthy lifestyle factors and long-term CVD risk among women with a history of GDM is limited [8,9]. Notably, whether adherence to a healthy lifestyle attenuates the excess risk of CVD and mortality associated with GDM remains to be elucidated. Moreover, although insufficient sleep has been recognized as an emerging component of the American Heart Association’s “Essential 8″ cardiovascular health metrics [11], little is known about the role of sleep in the association of GDM with incident CVD and mortality.
Using data from the UK Biobank, we evaluated the joint and stratified associations of an overall healthy lifestyle pattern and a history of GDM with major CVD and all-cause mortality to examine whether adhering to a healthy lifestyle could modify the association of a history of GDM with the increased risk.
2. Methods
Study design and participants: The UK Biobank is a prospective cohort study recruiting over 500,000 participants aged 40 to 70 years at 22 assessment centers between 2006 and 2010. The study design and data collection have previously been described in detail [12]. At baseline, participants provided information through touchscreen questionnaires, verbal interviews, physical measurements, and blood samples following standardized procedures. All participants provided written informed consent for the study. The study received ethics approval from the North West Multi-center Research Ethics Committee (reference NO 16/NW/0274).
Among 221,342 women with at least one live birth, we excluded 854 with incomplete GDM history and 91,197 with incomplete lifestyle data (859 for smoking, 40,637 for alcohol consumption, 47,035 for physical activity, 2398 for diet, and 268 for sleep duration). After further excluding participants diagnosed with diabetes before GDM (n = 123), those who developed GDM after baseline (n = 19), and those with prevalent CVD at baseline (n = 3714), the current analysis included 125,435 women (Supplementary Figure S1).
Assessment of history of GDM: At baseline, information on the history of GDM was identified by integrating self-reported history and hospital admission records. Information on the medical history of GDM was self-reported through a touchscreen questionnaire or during a verbal interview with trained staff during the baseline assessment. Additionally, the UK Biobank collected hospital admission data regarding disease diagnoses before enrollment, and the dates and causes were obtained through linkage to Health Episode Statistics (England and Wales) and the Scottish Morbidity Records (Scotland). GDM was identified by the International Classification of Diseases, 10th revision (ICD-10; O24).
Assessment of lifestyle factors: We constructed a healthy lifestyle score that included smoking, alcohol consumption, physical activity, diet, and sleep duration based on recommendations by the World Health Organization (WHO) and the American Heart Association (AHA) [11,13]. All lifestyle factors were collected via a touchscreen questionnaire at baseline (Supplementary Table S1). For smoking, a healthy level was defined as never smoking. Frequency, types, and volume of current alcohol consumption were self-reported, and a healthy level was defined as never/low levels of drinking(<7 U.K. standard drinks [56 g alcohol]/week) according to the guidelines in the UK and the risk threshold reported in prior UK Biobank analyses [14,15]. Regular physical activity was collected using the short-form International Physical Activity Questionnaire, and a healthy level was defined as ≥150 min/week of moderate activity or 75 min/week of vigorous activity or equivalent according to the WHO guideline [16,17]. Dietary information was collected through questionnaires on certain food types in the UK Biobank. According to previous UK Biobank analyses, diet quality was evaluated based on a dietary index that adhered to dietary recommendations for cardiovascular health, considering adequate consumption of fruit, vegetables, whole grains, fish, shellfish, dairy products, and vegetable oils, as well as reduced consumption of refined grains, processed meats, unprocessed meats, and sugar-sweetened beverages [18]. A healthy diet was defined as meeting ≥5 recommendation items. Healthy sleep was defined as 7–9 h/day according to AHA recommendations [11].
In our primary analysis, we assigned 1 point to participants for each healthy category, and the lifestyle score was calculated as the sum of all five factors, with a higher score indicating better adherence to an overall healthy lifestyle. To avoid extreme groups with limited cases, the lifestyle score was categorized into three groups (0–2, 3, 4–5). Although this method has been widely used and is easier to apply in clinical practice [19], the underlying assumption of identical associations between different lifestyle factors and the outcome may not be true. Thus, we also constructed a weighted lifestyle score using the association of each lifestyle factor with the outcome as weights in sensitivity analysis [20].
Covariates: Information on covariates (including age, self-reported race/ethnicity, Townsend deprivation index [21], education, and reproductive history) was obtained by baseline self-reported questionnaires. Hypertension was defined as a self-reported history of diagnosis, use of antihypertensive medication, blood pressure>140/90 mmHg at baseline, or pre-baseline hospital records. Diabetes was defined as a self-reported history of diagnosis, use of glucose-lowering medication, glycated hemoglobin (HbA1c) level ≥6.5 %, random glucose level≥11.1 mmol/L, or pre-baseline hospital records. Body mass index (BMI) and waist-to-hip ratio (WHR) were collected with standardized physical measurements.
Outcome ascertainment: The primary outcome was a composite of major cardiovascular events (including coronary artery disease [I21–I25], stroke [I60–I64, I69], heart failure [I50], and cardiovascular death [I00-I99]) and all-cause mortality, identified by linking hospital records using ICD-10 codes. We used the composite outcome to boost statistical power and repeated analyses with major CVD and all-cause mortality as separate outcomes. We also explored associations for CVD and non-CVD mortality.
Statistical analysis: The hazard ratios (HRs) with 95 % confidence intervals (CIs) for GDM history and lifestyle factors were estimated using Cox proportional hazard models. We first examined the separate associations of GDM history and healthy lifestyles with health outcomes. Secondly, we conducted a stratified analysis by GDM to assess the associations of the healthy lifestyle score with clinical outcomes among women with or without a GDM history. The multiplicative interaction was measured by including cross-product terms, and the additive interaction was estimated by calculating the relative excess risk due to interaction (RERI). Models were adjusted for age, Townsend deprivation index, ethnicity, education, age of menarche, number of live births, age of first birth, history of adverse birth outcomes (stillbirth, spontaneous miscarriage, or termination), menopause status, oral contraceptive use, hormone replacement therapy use, and hysterectomy history. Next, we performed joint analyses by classifying participants into six groups based on their GDM history (yes or no) and healthy lifestyle scores (0–2; 3; 4–5). We estimated the HR of health outcomes in different groups compared with those with no prior history of GDM and the healthiest lifestyle. Moreover, we also repeated the stratified and joint analyses for each individual lifestyle factor.
Several sensitivity analyses were performed. First, we repeated the analyses using a weighted healthy lifestyle score to account for the varied magnitudes of the associations of different lifestyle factors with outcomes [20]. Second, as middle-life obesity status might serve a potential role in the mediation pathway between GDM, lifestyle, and clinical outcomes, healthy BMI was not included as a component of the integrated lifestyle score. Since weight maintenance is also important for the long-term management of GDM, we further reconstructed the lifestyle score by including obesity using two different markers, baseline BMI (18.5–25 kg/m2 as low risk) and WHR (<0.85 as low risk). Third, we excluded events within the first two years of follow-up to minimize potential reverse causation. Fourth, we conducted multiple imputations for missing covariates. Fifth, considering that the GDM history was retrospectively collected, and there is a gap of a mean of 30.0 ± 10.3 years between the index pregnancy and baseline enrollment, we also repeated the analysis stratified by time since index pregnancy (defined as the pregnancy diagnosed with GDM or the first pregnancy in women without GDM history). Sixth, we further adjusted for prevalent cardiovascular risk factors at baseline (diabetes, hypertension, dyslipidemia, and BMI) that may potentially mediate the association of GDM and lifestyle with CVD and outcomes. All analyses were performed utilizing SAS 9.4 (Cary, NC). A p-value<0.05 was considered statistically significant.
3. Results
The mean age at baseline was 56.4 ± 7.9 yrs, and 658 (0.5 %) participants had a GDM history before enrollment (Table 1). Compared with those without GDM, women with a GDM history were, on average, younger, non-white, with higher socioeconomic levels, and reported less smoking and consuming less alcohol, but less physical activity, less adequate sleep duration, and lower diet quality. 541 (82.2 %) women with GDM history have progressed to diabetes at baseline, and those with GDM history also had higher prevalences of hypertension, diabetes, overweight/obesity, and adverse birth outcomes. 50,294 (40.0 %) were in the least healthy lifestyle group (0–1 healthy lifestyle; Supplementary Table S2). Women with healthier lifestyles, regardless of GDM history, generally had higher socioeconomic levels and were less likely to be overweight/obese.
Table 1.
Baseline characteristics in the overall population and stratified by history of GDM and lifestyle risk factors in the UK Biobank (N = 125,435).
| Variables | Total (N = 125,435) | No GDM |
GDM |
||||||
|---|---|---|---|---|---|---|---|---|---|
| All (N = 124,777) | Least healthy lifestyle (N = 50,029) | Moderately healthy lifestyle (N = 45,159) | Most healthy lifestyle (N = 29,589) | All (N = 658) | Least healthy lifestyle (N = 265) | Moderately healthy lifestyle (N = 220) | Most healthy lifestyle (N = 173) | ||
| Age | 56.4 ± 7.9 | 56.4 ± 7.9 | 56.3 ± 7.7 | 56.2 ± 7.9 | 56.9 ± 8.0 | 51.3 ± 7.7 | 50.7 ± 7.3 | 51.7 ± 7.8 | 51.9 ± 8.2 |
| Ethnic: White | 119,232 (95.1) | 118,661 (95.1) | 48,429 (96.8) | 42,996 (95.2) | 27,236 (92.0) | 571 (86.8) | 242 (91.3) | 193 (87.7) | 136 (78.6) |
| Highest quintile of Townsend deprivation index | 25,085 (20.0) | 24,921 (20.0) | 11,108 (22.2) | 8544 (18.9) | 5269 (17.8) | 164 (24.9) | 67 (25.3) | 49 (22.3) | 48 (27.7) |
| Education: College or University degree | 42,593 (34.0) | 42,334 (33.9) | 16,002 (32.0) | 15,923 (35.3) | 10,409 (35.2) | 259 (39.4) | 112 (42.3) | 92 (41.8) | 55 (31.8) |
| Healthy lifestyle | |||||||||
| Never smoking | 73,740 (58.8) | 73,322 (58.8) | 12,910 (25.8) | 32,547 (72.1) | 27,865 (94.2) | 418 (63.5) | 85 (32.1) | 168 (76.4) | 165 (95.4) |
| Never or low alcohol use | 50,528 (40.3) | 50,207 (40.2) | 6624 (13.2) | 18,288 (40.5) | 25,295 (85.5) | 321 (48.8) | 56 (21.1) | 105 (47.7) | 160 (92.5) |
| Achieving physical activity recommendation | 103,993 (82.9) | 103,481 (82.9) | 34,077 (68.1) | 40,297 (89.2) | 29,107 (98.4) | 512 (77.8) | 163 (61.5) | 182 (82.7) | 167 (96.5) |
| Sleep duration of 7–9 h/day | 95,032 (75.8) | 94,559 (75.8) | 28,958 (57.9) | 37,404 (82.8) | 28,197 (95.3) | 473 (71.9) | 134 (50.6) | 176 (80.0) | 163 (94.2) |
| Diet score≥5 | 21,094 (16.8) | 20,999 (16.8) | 2330 (4.7) | 6941 (15.4) | 11,728 (39.6) | 95 (14.4) | 6 (2.3) | 29 (13.2) | 60 (34.7) |
| Metabolic factors | |||||||||
| Hypertension | 57,485 (45.8) | 57,176 (45.8) | 23,297 (46.6) | 20,184 (44.7) | 13,695 (46.3) | 309 (47.0) | 119 (44.9) | 99 (45.0) | 91 (52.6) |
| Diabetes | 3963 (3.2) | 3422 (2.7) | 1303 (2.6) | 1147 (2.5) | 972 (3.3) | 541 (82.2) | 212 (80.0) | 185 (84.1) | 144 (83.2) |
| Body mass index, kg/m2 | 26.6 ± 4.7 | 26.6 ± 4.7 | 26.9 ± 4.9 | 26.4 ± 4.6 | 26.2 ± 4.6 | 28.0 ± 5.7 | 28.2 ± 5.7 | 27.9 ± 5.8 | 27.8 ± 5.7 |
| Overweight/obesity (Body mass index≥25) | 71,620 (57.3) | 71,184 (57.3) | 29,990 (60.2) | 25,138 (55.9) | 16,056 (54.5) | 436 (66.6) | 176 (66.4) | 146 (67.0) | 114 (66.3) |
| Abdominal obesity (Waist-to-hip ratio>0.85) | 36,088 (28.9) | 35,824 (28.8) | 15,996 (32.1) | 12,134 (26.9) | 7694 (26.1) | 391 (59.7) | 155 (58.5) | 127 (58.3) | 109 (63.4) |
| Reproductive factors | |||||||||
| Age of menarche | 13.0 ± 1.6 | 13.0 ± 1.6 | 13.0 ± 1.6 | 13.0 ± 1.6 | 13.0 ± 1.6 | 12.9 ± 1.7 | 12.9 ± 1.6 | 13.0 ± 1.7 | 12.8 ± 1.8 |
| Menopause | 76,886 (61.3) | 76,632 (61.4) | 30,885 (61.7) | 27,223 (60.3) | 18,524 (62.6) | 254 (38.6) | 92 (34.7) | 90 (40.9) | 72 (41.6) |
| Number of live births | 2.2 ± 0.8 | 2.2 ± 0.8 | 2.2 ± 0.8 | 2.2 ± 0.8 | 2.2 ± 0.8 | 2.3 ± 0.9 | 2.3 ± 0.9 | 2.3 ± 0.9 | 2.2 ± 1.0 |
| Age at first birth | 26.4 ± 5.1 | 26.5 ± 5.1 | 26.2 ± 5.3 | 26.5 ± 5.1 | 26.5 ± 4.9 | 28.7 ± 6.0 | 28.8 ± 6.1 | 29.0 ± 6.0 | 28.2 ± 5.8 |
| Ever had stillbirth, spontaneous, miscarriage or termination | 44,575 (35.5) | 44,283 (35.5) | 19,009 (38.0) | 15,568 (34.5) | 9706 (32.8) | 292 (44.4) | 121 (45.7) | 101 (45.9) | 70 (40.5) |
| Ever taken oral contraceptive pill | 106,276 (84.7) | 105,715 (84.7) | 44,062 (88.1) | 38,227 (84.6) | 23,426 (79.2) | 561 (85.3) | 240 (90.6) | 187 (85.0) | 134 (77.5) |
| Ever used hormone replacement therapy | 40,389 (32.2) | 40,243 (32.3) | 18,104 (36.2) | 13,852 (30.7) | 8287 (28.0) | 146 (22.2) | 61 (23.0) | 44 (20.0) | 41 (23.7) |
| Ever had hysterectomy | 8764 (7.0) | 8736 (7.0) | 3463 (6.9) | 3092 (6.8) | 2181 (7.4) | 28 (4.3) | 12 (4.5) | 7 (3.2) | 9 (5.2) |
Abbreviations: GDM, gestational diabetes mellitus.
After a median follow-up of 13.6 years (Interquartile range, 13.0–14.4), 9371 participants had incident major CVD, and 6750 died. In multivariable-adjusted models, participants with GDM history had a higher risk of the composite outcome (HR, 1.67 [95 % CI, 1.33–2.10]; P < 0.001; Supplementary Table S3). Women with a healthier lifestyle score compared with those with a minimum score of 0 had a lower risk of the composite outcome, such that per increment in lifestyle risk score was associated with an HR of 0.89 (0.88–0.91; P < 0.001 for trend) for the composite outcome (Supplementary Table S4).
Table 2 presents the association of GDM history with clinical outcomes, stratified by healthy lifestyle. Compared to no GDM history, a history of GDM was significantly associated with a higher risk of the composite outcome among women with the least healthy lifestyle (HR for GDM, 2.35 [1.72–3.22]). However, the association of GDM with the composite outcome was less pronounced among those with moderately (HR, 1.31 [0.84–2.06]) or most healthy lifestyles (HR, 1.24 [0.75–2.06]; P = 0.001 for multiplicative interaction). The RERI between GDM history and the least healthy lifestyle score was 0.52 (0.27–0.77; P = 0.02) for the composite outcome and 0.59 (0.34–0.84; P = 0.01) for major CVD. Similar patterns were also observed for all-cause and non-CVD mortality, but the interaction was not significant due to a limited number of events (Supplementary Table S5).
Table 2.
Stratified associations of history of GDM and healthy lifestyle scores with the composite of major CVD and mortality (n = 125,435).
| Cases/N (Incidence rate, per 1000 person-year) |
HR (95 % CI) for GDM versus non-GDM | Pmultiplicative | RERI | Padditive | ||
|---|---|---|---|---|---|---|
| No history of GDM | History of GDM | |||||
| Composite outcome | 0.001 | |||||
| Least healthy lifestyle (0–2) | 6313/50,029 (7.90) | 40/265 (17.29) | 2.35 (1.72–3.22) | 0.52 (0.27–0.77) | 0.02 | |
| Moderately healthy lifestyle (3) | 4657/45,159 (6.40) | 19/220 (8.03) | 1.31 (0.84–2.06) | 0.07 (−0.54–0.69) | 0.85 | |
| Most healthy lifestyle (4–5) | 3074/29,589 (6.05) | 15/173 (7.71) | 1.24 (0.75–2.06) | Ref | ||
| Major CVD | 0.001 | |||||
| Least healthy lifestyle (0–2) | 4124/50,029 (5.10) | 32/265 (13.87) | 2.88 (2.03–4.09) | 0.59 (0.34,0.84) | 0.01 | |
| Moderately healthy lifestyle (3) | 3115/45,159 (4.23) | 13/220 (5.48) | 1.36 (0.79–2.35) | 0.05 (−0.70,0.80) | 0.92 | |
| Most healthy lifestyle (4–5) | 2076/29,589 (4.03) | 11/173 (5.63) | 1.31 (0.72–2.37) | Ref | ||
| All-cause mortality | 0.28 | |||||
| Least healthy lifestyle (0–2) | 3164/50,029 (3.65) | 14/265 (5.39) | 1.59 (0.94–2.69) | 0.41 (−0.07,0.90) | 0.28 | |
| Moderately healthy lifestyle (3) | 2120/45,159 (2.70) | 10/220 (4.02) | 1.59 (0.85–2.97) | 0.41 (−0.22,1.03) | 0.34 | |
| Most healthy lifestyle (4–5) | 1437/29,589 (2.60) | 5/173 (2.45) | 0.95 (0.40–2.29) | Ref | ||
Abbreviations: CI, confidence interval; CVD, cardiovascular disease; GDM, gestational diabetes mellitus; HR, hazard ratio; RERI, relative excess risk due to interaction.
Incidence rates were standardized for age and presented as per 1000 person-year.
Models adjusted for age, Townsend deprivation index, ethnic, education, age of menarche, number of live births, age of first birth, history of adverse birth outcomes (ever had stillbirth, spontaneous, miscarriage or termination), menopause status, history of taking oral contraceptive pill, history of using hormone replacement therapy, and history of hysterectomy.
Fig. 1 presents the joint association of GDM history and healthy lifestyles with major CVD and all-cause mortality. Compared with the referent group (healthiest lifestyles and no GDM history), women with GDM history and the least healthy lifestyle had HRs of 3.00 (2.20–4.10; P < 0.001) for the composite outcome, 3.63 (2.56–5.16) for major CVD, and 2.11 (1.24–3.57) for all-cause mortality. By contrast, the risk among women with GDM history but the healthiest lifestyle was not significant, with HRs of 1.19 (0.71–1.97) for the composite outcome, 1.26 (0.69–2.27) for major CVD, and 0.91 (0.38–2.18) for all-cause mortality.
Fig. 1.
Joint associations of history of GDM and healthy lifestyle scores with the composite of major CVD and mortality.
Abbreviations: CI, confidence interval; CVD, cardiovascular disease; GDM, gestational diabetes mellitus; HR, hazard ratio. Models adjusted for age, Townsend deprivation index, ethnicity, education, age of menarche, number of live births, age of first birth, history of adverse birth outcomes (ever had stillbirth, spontaneous, miscarriage or termination), menopause status, history of taking oral contraceptive pill, history of using hormone replacement therapy, and history of hysterectomy.
In further analysis of the lifestyle factors individually, GDM history had significant additive interactions for the composite outcome with smoking, physical activity, and sleep duration, with RERIs of 0.22 (0.18–0.25), 0.58 (0.25–0.90), and 0.21 (0.14–0.29), respectively. When stratified by individual lifestyle factors, the association of GDM history with composite outcome was significantly modified by physical activity and sleep duration. The HRs of GDM for the composite outcome were 2.40 versus 1.44 and 1.33 versus 2.42 among those unhealthy versus healthy groups, respectively (Fig. 2).
Fig. 2.
The association of history of GDM with the composite of major CVD and mortality stratified by each modifiable lifestyle factor.
Abbreviations: CI, confidence interval; CVD, cardiovascular disease; GDM, gestational diabetes mellitus; HR, hazard ratio; RERI, relative excess risk due to interaction. Models adjusted for age, Townsend deprivation index, ethnicity, education, age of menarche, number of live births, age of first birth, history of adverse birth outcomes (ever had stillbirth, spontaneous, miscarriage or termination), menopause status, history of taking oral contraceptive pill, history of using hormone replacement therapy, history of hysterectomy, and mutually adjusted for other modifiable lifestyle factors.
In sensitivity analyses, the results were consistent when using a weighted healthy lifestyle score (Supplementary Table S6) or adding obesity into the lifestyle risk score using either baseline BMI or WHR (Supplementary Table S7). Moreover, results remained stable after removing those with <2 years of follow-up (Supplementary Table S8), repeating the analyses using multiple imputations (Supplementary Table S9), or further stratifying by the stratum of time since the index pregnancy (Supplementary Table S10). After further adjusting for prevalent cardiometabolic risk factors at baseline, the association of GDM with the composite of CVD and mortality was largely attenuated not only in the overall population but also across lifestyle groups. Although the interaction was not significant, there was still some residual GDM-related risk in the least healthy lifestyle group, while the risk was lower in the moderately or most healthy groups (Supplementary Table S11).
4. Discussion
In this large prospective cohort, we found an overall healthy lifestyle significantly modified the association of GDM history with the composite of major CVD and mortality. An unhealthy lifestyle pattern even further exacerbates the excess risk associated with GDM history, especially for smoking, physical inactivity, and insufficient sleep. By contrast, the GDM-related risks of major CVD and mortality were not significant among women with a GDM history but adhering to the healthiest lifestyles, suggesting that an overall healthy lifestyle may prevent much of the elevated risk.
To our knowledge, this is the first study to assess the joint association between adhering to a healthy lifestyle and GDM history with incident CVD and mortality. Our study found that the excess risk associated with GDM was much lower among women with the healthiest lifestyle. Women with a history of GDM did not experience an increased risk of major CVD and all-cause mortality when adhering to the healthiest lifestyle at midlife, compared to those without such a history of GDM. These findings emphasized the importance of maintaining an overall healthy lifestyle to prevent long-term risks among this high-risk population.
Moreover, there were significant additive interactions between GDM history and healthy lifestyles, suggesting that an unhealthy lifestyle pattern even further exacerbates the excess risk associated with GDM history [22,23]. Similarly, although the interaction effects were not significant, the Nurses’ Health Study (NHS) II also found that the association of GDM with incident CVD and all-cause mortality was more pronounced among individuals with several well-established unhealthy lifestyle factors, including diet quality, physical activity, smoking status, alcohol intake, and overweight/obesity [8,9]. Despite current guidelines recognizing GDM history as a risk-enhancing factor for refining cardiovascular risk assessment, there is no specific guideline for cardiovascular risk prevention for women with GDM history after delivery [24]. Given that healthy lifestyles could significantly modify the CVD and mortality risks related to GDM, it is crucial to conduct large clinical trials evaluating the long-term effects of lifestyle changes on CVD risk to develop evidence-based prevention strategies for this high-risk population.
Of note, we found an additive interaction between GDM history and sleep duration, suggesting that sleep duration could be a key target for mitigating long-term risk of CVD and mortality after GDM. In contrast to dietary modification and physical activity, the importance of maintaining healthy sleep has been scarcely addressed in earlier short-term lifestyle modification trials for GDM [24,25]. Although the effects of sleep disturbances on glucose were comparable to traditional risk factors [26], only one small-scale pilot study has tested the effects of sleep extension on glucose metabolism in women with a history of GDM [27]. Recently, there were some attempts to develop feasible, comprehensive, and evidence-based lifestyle modification strategies for postpartum management. However, the clinical efficacy was not observed after short-term interventions, perhaps due to low compliance and retention rates in the post-GDM populations [[28], [29], [30]]. Since engaging postpartum women in lifestyle interventions is particularly challenging, alternate or additional precision prevention approaches are still necessary.
Women with GDM may exhibit different preconception cardiometabolic profiles during pregnancy and could experience more weight gain and obesity risks in their midlife, with metabolic changes occurring shortly after delivery [31,32]. Consistently, we also found that the GDM-related risk was largely explained by postpartum cardiometabolic disorders, and healthy lifestyle may prevent obesity and delay the progression to diabetes and other cardiometabolic disorders in later life [[33], [34], [35], [36], [37]]. Nonetheless, there remains a lack of detailed guidelines on how to manage diet, exercise, and other important lifestyle factors for women with GDM history [38].
The UK Biobank is a large prospective cohort with comprehensive information on healthy lifestyle and health status collected using standardized procedures, and the long-term clinical endpoints were rigorously collected based on national registries and hospital admission records. However, several limitations should be addressed. First, the UK Biobank used a convenience sample, and the participants are less likely to be obese, to smoke and drink, and more likely to live in areas with higher socioeconomic status compared with the general population in the UK [39]. This may also explain why the prevalence of GDM was relatively lower than the estimates during the same period in the UK (0.97–2.69 %) [40]. Therefore, the generalization into other populations warrants further investigation. Second, the information on GDM was retrospectively collected during study enrollment (2006–2010). Considering the mean duration since the first live birth was 30.0 ± 10.3 years, most women had been pregnant before that when universal GDM screening was adopted into routine obstetrics practice in the late 1990s [41]. Thus, many GDM cases might have been undiagnosed or misdiagnosed. The small number of GDM cases also limits our ability to further evaluate the association for individual components of the composite outcome. Third, most women with a history of GDM have already progressed to diabetes at baseline, which limits our ability to prospectively assess the role of progression to diabetes in the association of GDM and healthy lifestyle with CVD and mortality. Fourth, data on the severity and recurrence of GDM were not collected, and whether a healthier lifestyle could modify related CVD risk could not be assessed. Fifth, data on lifestyle factors were self-reported, and misclassification is likely. In particular, diet quality was simply assessed based on frequencies of limited food groups without considering portion sizes, although prior studies have shown that the dietary index was significantly linked to mortality and CVD risk [18,42,43]. However, given the prospective nature of this study, such misclassification would be nondifferential, which could lead to an underestimation of the true association. Sixth, survivor bias might also exist since women with a GDM history have a higher risk of adverse outcomes and are therefore less likely to participate in the study at an older age, which has slightly underestimated the observed associations of GDM. Seventh, individuals’ lifestyles can change over time during midlife, potentially weakening the contrast between different healthy lifestyle groups and diluting the modified effects on GDM. Further investigation is needed to understand how lifestyle changes (i.e., trajectories) throughout midlife affect the association between GDM and long-term risk. Lastly, although a comprehensive set of covariates, including reproductive characteristics, has been carefully adjusted in the current analysis, some residual confounding effects cannot be controlled, such as underlying cardiometabolic risk before pregnancy and the recurrence of GDM across multiple pregnancies.
In conclusion, unhealthy lifestyles may exacerbate the increased risk of major CVD and all-cause mortality among women with a GDM history, especially for smoking, physical inactivity, and insufficient sleep. By contrast, women with a history of GDM did not experience a higher risk of major CVD and all-cause mortality when adhering to a healthy lifestyle at midlife. Our findings reinforce the importance of lifestyle modification in long-term prevention in this high-risk population, and further clinical trials are necessary to assess the long-term effects of lifestyle modifications to inform evidence-based precision prevention strategies.
Funding
This study was supported by the National Natural Science Foundation of China (grant No 82473722; 82204059) and the USTC Research Funds of the Double First-Class Initiative (grant No YD9100002029; YD9100002060).
CRediT authorship contribution statement
Sidong Li: Writing – review & editing, Writing – original draft, Validation, Supervision, Software, Resources, Project administration, Methodology, Investigation, Formal analysis, Conceptualization. Yuxiao Wu: Writing – review & editing, Writing – original draft, Methodology, Formal analysis. Yuxiang Yan: Writing – original draft, Visualization, Validation, Methodology, Investigation, Formal analysis. Yang Du: Writing – review & editing, Methodology. Shuhan Chen: Writing – review & editing, Methodology, Data curation. Deirdre K. Tobias: Writing – review & editing. Cuilin Zhang: Writing – review & editing. Wei Bao: Writing – review & editing, Supervision, Project administration, Methodology, Investigation, Funding acquisition, Conceptualization.
Declaration of competing interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Acknowledgments
UK Biobank data are available for researchers after acceptance of a research proposal to the UK Biobank, and the present study was conducted under application number 117331 of the UK Biobank resource. Dr. Wei Bao and Dr. Sidong Li had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Footnotes
Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.ajpc.2025.101406.
Appendix. Supplementary materials
References
- 1.Saeedi M., Cao Y., Fadl H., Gustafson H., Simmons D. Increasing prevalence of gestational diabetes mellitus when implementing the IADPSG criteria: a systematic review and meta-analysis. Diabetes Res Clin Pr. 2021;172 doi: 10.1016/j.diabres.2020.108642. [DOI] [PubMed] [Google Scholar]
- 2.McIntyre H.D., Catalano P., Zhang C., Desoye G., Mathiesen E.R., Damm P. Gestational diabetes mellitus. Nat Rev Dis Primers. 2019;5(1):47. doi: 10.1038/s41572-019-0098-8. [DOI] [PubMed] [Google Scholar]
- 3.Wu Y., Liu B., Sun Y., et al. Association of maternal prepregnancy diabetes and gestational diabetes mellitus with congenital anomalies of the newborn. Diabetes Care. 2020;43(12):2983–2990. doi: 10.2337/dc20-0261. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Ying Q., Xu Y., Zhang Z., Cai L., Zhao Y., Jin L. Gestational diabetes mellitus and risk of long-term all-cause and cardiac mortality: a prospective cohort study. Cardiovasc Diabetol. 2024;23(1):47. doi: 10.1186/s12933-024-02131-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Vounzoulaki E., Khunti K., Abner S.C., Tan B.K., Davies M.J., Gillies C.L. Progression to type 2 diabetes in women with a known history of gestational diabetes: systematic review and meta-analysis. Bmj. 2020;369:m1361. doi: 10.1136/bmj.m1361. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Tobias D.K., Hu F.B., Forman J.P., Chavarro J., Zhang C. Increased risk of hypertension after gestational diabetes mellitus: findings from a large prospective cohort study. Diabetes Care. 2011;34(7):1582–1584. doi: 10.2337/dc11-0268. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Kramer C.K., Campbell S., Retnakaran R. Gestational diabetes and the risk of cardiovascular disease in women: a systematic review and meta-analysis. Diabetologia. 2019;62(6):905–914. doi: 10.1007/s00125-019-4840-2. [DOI] [PubMed] [Google Scholar]
- 8.Tobias D.K., Stuart J.J., Li S., et al. Association of history of gestational diabetes with long-term cardiovascular disease risk in a large prospective cohort of US women. JAMA Intern Med. 2017;177(12):1735–1742. doi: 10.1001/jamainternmed.2017.2790. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Wang Y.X., Mitsunami M., Manson J.E., et al. Association of gestational diabetes with subsequent long-term risk of mortality. JAMA Intern Med. 2023 doi: 10.1001/jamainternmed.2023.4401. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Khandelwal R., Meshram R.J., Patel A., Reddy S., Manek Y.B. Maternal weight and gestational diabetes impacts on child health: a narrative review. Cureus. 2024;16(9) doi: 10.7759/cureus.70192. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Lloyd-Jones D.M., Allen N.B., Anderson C.A.M., et al. Life's essential 8: updating and enhancing the American heart association's construct of cardiovascular health: a presidential advisory from the American heart association. Circulation. 2022;146(5):e18–e43. doi: 10.1161/CIR.0000000000001078. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Sudlow C., Gallacher J., Allen N., et al. UK biobank: an open access resource for identifying the causes of a wide range of complex diseases of middle and old age. PLoS Med. 2015;12(3) doi: 10.1371/journal.pmed.1001779. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.World Health Organization . 2017. Tackling NCDs: ‘Best Buys’ And Other Recommended Interventions For The Prevention And Control Of Noncommunicable Diseases.https://www.who.int/publications/i/item/WHO-NMH-NVI-17.9 Accessed March 26, 2024. [Google Scholar]
- 14.Tu S.J., Gallagher C., Elliott A.D., et al. Risk thresholds for total and beverage-specific alcohol consumption and incident atrial fibrillation. Clin Electrophysiol. 2021;7(12):1561–1569. doi: 10.1016/j.jacep.2021.05.013. [DOI] [PubMed] [Google Scholar]
- 15.Kalinowski A., Humphreys K. Governmental standard drink definitions and low-risk alcohol consumption guidelines in 37 countries. Addiction. 2016;111(7):1293–1298. doi: 10.1111/add.13341. [DOI] [PubMed] [Google Scholar]
- 16.Craig C.L., Marshall A.L., Sjostrom M., et al. International physical activity questionnaire: 12-country reliability and validity. Med Sci Sports Exerc. 2003;35(8):1381–1395. doi: 10.1249/01.MSS.0000078924.61453.FB. [DOI] [PubMed] [Google Scholar]
- 17.Bull F.C., SS Al-Ansari, Biddle S., et al. World Health Organization 2020 guidelines on physical activity and sedentary behaviour. Br J Sports Med. 2020;54(24):1451–1462. doi: 10.1136/bjsports-2020-102955. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Said M.A., Verweij N., van der Harst P. Associations of combined genetic and lifestyle risks with incident cardiovascular disease and diabetes in the UK biobank study. JAMA Cardiol. 2018;3(8):693–702. doi: 10.1001/jamacardio.2018.1717. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Zhang Y.B., Pan X.F., Chen J., et al. Combined lifestyle factors, all-cause mortality and cardiovascular disease: a systematic review and meta-analysis of prospective cohort studies. J Epidemiol Comm Health. 2021;75(1):92–99. doi: 10.1136/jech-2020-214050. [DOI] [PubMed] [Google Scholar]
- 20.Warren Andersen S., Zheng W., Sonderman J., et al. Combined impact of health behaviors on mortality in low-income Americans. Am J Prev Med. 2016;51(3):344–355. doi: 10.1016/j.amepre.2016.03.018. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Foster H.M.E., Celis-Morales C.A., Nicholl B.I., et al. The effect of socioeconomic deprivation on the association between an extended measurement of unhealthy lifestyle factors and health outcomes: a prospective analysis of the UK Biobank cohort. Lancet Public Health. 2018;3(12):e576–e585. doi: 10.1016/S2468-2667(18)30200-7. [DOI] [PubMed] [Google Scholar]
- 22.Thirunavukarasu S., Ansari F., Cubbon R., et al. Maternal cardiac changes in women with obesity and gestational diabetes mellitus. Diabetes Care. 2022;45(12):3007–3015. doi: 10.2337/dc22-0401. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.He Y., Wu N. Research progress on gestational diabetes mellitus and endothelial dysfunction markers. Diabetes Metab Syndr Obes. 2021;14:983–990. doi: 10.2147/DMSO.S295737. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Jowell A.R., Sarma A.A., Gulati M., et al. Interventions to mitigate risk of cardiovascular disease after adverse pregnancy outcomes: a review. JAMA Cardiol. 2022;7(3):346–355. doi: 10.1001/jamacardio.2021.4391. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Shepherd E., Gomersall J.C., Tieu J., Han S., Crowther C.A., Middleton P. Combined diet and exercise interventions for preventing gestational diabetes mellitus. Cochrane Database Syst Rev. 2017;11(11) doi: 10.1002/14651858.CD010443.pub3. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Anothaisintawee T., Reutrakul S., Van Cauter E., Thakkinstian A. Sleep disturbances compared to traditional risk factors for diabetes development: systematic review and meta-analysis. Sleep Med Rev. 2016;30:11–24. doi: 10.1016/j.smrv.2015.10.002. [DOI] [PubMed] [Google Scholar]
- 27.Reutrakul S., Martyn-Nemeth P., Quinn L., et al. Effects of sleep-extend on glucose metabolism in women with a history of gestational diabetes: a pilot randomized trial. Pilot Feasibility Stud. 2022;8(1):119. doi: 10.1186/s40814-022-01076-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Potzel A.L., Gar C., Seissler J., Lechner A. A smartphone app (TRIANGLE) to change cardiometabolic risk behaviors in women following gestational diabetes mellitus: intervention mapping approach. JMIR Mhealth Uhealth. 2021;9(5) doi: 10.2196/26163. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Potzel A.L., Gar C., Banning F., et al. A novel smartphone app to change risk behaviors of women after gestational diabetes: a randomized controlled trial. PLoS One. 2022;17(4) doi: 10.1371/journal.pone.0267258. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Tandon N., Gupta Y., Kapoor D., et al. Effects of a lifestyle intervention to prevent deterioration in glycemic status among South Asian women with recent gestational diabetes: a randomized clinical trial. JAMA Netw Open. 2022;5(3) doi: 10.1001/jamanetworkopen.2022.0773. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Bahador E., Saber M., FadakarDavarani M.M., Khanjani N., Gohari B.H., Safinejad H. The relationship between lifestyle and metabolic evaluation in women with a history of gestational diabetes. J Educ Health Promot. 2021;10:403. doi: 10.4103/jehp.jehp_12_21. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Chen L-W, Soh S.E., Tint M.T., et al. Combined analysis of gestational diabetes and maternal weight status from pre-pregnancy through post-delivery in future development of type 2 diabetes. Sci Rep. 2021;11(1):5021. doi: 10.1038/s41598-021-82789-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Borkoles E., Carroll S., Clough P., Polman R.C. Effect of a non-dieting lifestyle randomised control trial on psychological well-being and weight management in morbidly obese pre-menopausal women. Maturitas. 2016;83:51–58. doi: 10.1016/j.maturitas.2015.09.010. [DOI] [PubMed] [Google Scholar]
- 34.Tobias D.K., Hu F.B., Chavarro J., Rosner B., Mozaffarian D., Zhang C. Healthful dietary patterns and type 2 diabetes mellitus risk among women with a history of gestational diabetes mellitus. Arch Intern Med. 2012;172(20):1566–1572. doi: 10.1001/archinternmed.2012.3747. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Bao W., Tobias D.K., Bowers K., et al. Physical activity and sedentary behaviors associated with risk of progression from gestational diabetes mellitus to type 2 diabetes mellitus: a prospective cohort study. JAMA Intern Med. 2014;174(7):1047–1055. doi: 10.1001/jamainternmed.2014.1795. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Semnani-Azad Z., Gaillard R., Hughes A.E., Boyle K.E., Tobias D.K., Perng W. Precision stratification of prognostic risk factors associated with outcomes in gestational diabetes mellitus: a systematic review. Commun Med. 2024;4(1):9. doi: 10.1038/s43856-023-00427-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Yang J., Qian F., Chavarro J.E., et al. Modifiable risk factors and long term risk of type 2 diabetes among individuals with a history of gestational diabetes mellitus: prospective cohort study. Bmj. 2022;378 doi: 10.1136/bmj-2022-070312. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Huang J., Wu Y., Li H., et al. Weight management during pregnancy and the postpartum period in women with gestational diabetes mellitus: a systematic review and summary of current evidence and recommendations. Nutrients. 2023;15(24) doi: 10.3390/nu15245022. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Fry A., Littlejohns T.J., Sudlow C., et al. Comparison of sociodemographic and health-related characteristics of UK biobank participants with those of the general population. Am J Epidemiol. 2017;186(9):1026–1034. doi: 10.1093/aje/kwx246. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40.Farrar D., Simmonds M., Griffin S., et al. The identification and treatment of women with hyperglycaemia in pregnancy: an analysis of individual participant data, systematic reviews, meta-analyses and an economic evaluation. Southampton (UK) NIHR J Libr. 2016 doi: 10.3310/hta20860. https://www.ncbi.nlm.nih.gov/books/NBK401113/ (Health Technology Assessment, No. 20.86.) Chapter 4, Prevalence of gestational diabetes in the UK and Republic of Ireland: a systematic review. Available from. Nov. Accessd March 26, 2024. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41.Danilenko-Dixon D.R., Van Winter J.T., Nelson R.L., Ogburn P.L., Jr Universal versus selective gestational diabetes screening: application of 1997 American Diabetes Association recommendations. Am J Obs Gynecol. 1999;181(4):798–802. doi: 10.1016/s0002-9378(99)70304-2. [DOI] [PubMed] [Google Scholar]
- 42.Zhang Y.B., Chen C., Pan X.F., et al. Associations of healthy lifestyle and socioeconomic status with mortality and incident cardiovascular disease: two prospective cohort studies. Bmj. 2021;373:n604. doi: 10.1136/bmj.n604. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 43.Han H., Cao Y., Feng C., et al. Association of a healthy lifestyle with all-cause and cause-specific mortality among individuals with type 2 diabetes: a prospective study in UK Biobank. Diabetes Care. 2022;45(2):319–329. doi: 10.2337/dc21-1512. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.