Abstract
Objective
To identify if gestational diabetes mellitus (GDM) is a clinically useful marker of future cardiovascular disease (CVD) risk and if GDM combined with other risks (smoking, hypertension or body mass) identifies high-risk groups.
Design
Population-based matched case–control study.
Setting
National Swedish register data from 1991 to 2008.
Population
A total of 2639 women with a cardiovascular event and matched controls.
Methods
Conditional logistic regression examined associations with CVD before and after adjustment for conventional risk factors and confounders. Effect modification for the association of GDM with CVD by body mass index (BMI), smoking and chronic hypertension was assessed by stratification and interaction testing. Adjustment for diabetes post-pregnancy evaluated its mediating role.
Main outcome measures
Inpatient diagnoses or causes of death identifying ischemic heart disease, ischemic stroke, atherosclerosis or peripheral vascular disease.
Results
The adjusted odds ratios (and 95% confidence intervals) for the association of CVD with GDM are 1.51 (1.07–2.14), 2.23 (2.01–2.48) for smoking, 1.98 (1.71–2.29) for obesity and 5.10 (3.18–8.18) for chronic hypertension. In stratified analysis the association of CVD with GDM was only seen among women with BMI ≥25, with an odds ratio of 2.39 (1.39–4.10), but only women with a BMI <30 accounted for this increased risk. Adjustment for post-pregnancy diabetes attenuated it somewhat to 1.99 (1.13–3.52).
Conclusions
In the absence of other recognised cardiovascular risk factors, such as smoking, obesity or chronic hypertension, GDM is a useful marker of raised CVD risk among women with BMI between 25 and 29.
Keywords: Body mass index, cardiovascular disease, gestational diabetes mellitus
Introduction
The immediate adverse maternal and neonatal outcomes of hyperglycaemia during pregnancy are well known,1,2 as well as the long-term complications in the form of raised diabetes risk and the metabolic syndrome later in life.3,4 Future maternal morbidity in terms of cardiovascular disease (CVD) after gestational diabetes mellitus (GDM) has been shown in two studies.5,6 A Finnish population-based cohort study has demonstrated an association with later hypertension, especially among the group of women with GDM who were overweight.7 It is uncertain if this association can be explained by contemporaneous characteristics that are recognised risks for CVD or whether such characteristics may be used to identify a subset of women with GDM who are at particularly raised risk for future CVD. However, not all women with GDM develop subsequent CVD. It would be clinically valuable to have a risk marker at the time of pregnancy so that long-term follow up and possible interventions can be focused on the women at greatest risk in a timely manner.
The aim was therefore to evaluate whether the association of GDM with subsequent severe CVD is independent of conventional CVD risks that are readily identifiable during pregnancy and to identify if a subset of women with GDM is at a particularly raised risk of CVD. The pregnancy characteristics investigated were smoking, overweight and chronic hypertension.
The possible mediating role of diabetes after pregnancy was also evaluated.
Methods
Data were from the Swedish National Board of Health and Welfare and Statistics Sweden, from 1991 to 2008. Linkage between registries was performed using the unique personal identification number issued to Swedish residents.8 The Swedish Medical Birth Register (SMBR) contains data on more than 98% of all births in Sweden from 1973 onwards. Starting with the first visit at the end of the first trimester, information on all hospital births is gathered prospectively and includes demographic data, reproductive history and complications during pregnancy, delivery and the neonatal period. The SMBR was validated in 2002 and the quality is generally high.9 The Patient Register contains data on inpatient discharge diagnosis since 1964 but did not reach full coverage until 1987 and outpatient diagnoses have been registered since 2001. Since 2001 both private and public caregivers have been obliged to report to the register. The Inpatient Register has been validated recently and the validity of the different diagnosis of cardiovascular disease is high.10 The Cause of Death Register was used to identify deaths due to cardiovascular disease. The Prescription Register was used to identify individuals who were prescribed pharmaceutical treatment for diabetes according to ATC codes (A10). This register provides data on prescriptions beginning from 1 July 2005. Statistics Sweden is responsible for the national Register of Population and Population Changes, which was used to select controls and provided data on education.
Cases and controls
Cases were defined as women diagnosed with a first cardiovascular event or death in a cardiovascular event (ischaemic heart disease, stroke, atherosclerosis or peripheral vascular disease) using the ninth or tenth revision of the International Classification of Diseases (ICD-9, ICD-10). For each case, five controls were selected. All of the women had to have had a pregnancy before the date that they were defined as cases or controls. Cases and controls were the same age when they gave birth to a child during the same year. The controls had to be alive at the time of matching and resident in Sweden. Neither cases nor controls should have had cardiovascular events before the matched pregnancy. We identified nine individuals with a diagnosis of diabetes mellitus before the index pregnancy, indicating a misdiagnosis of GDM among 4.3%: these women were excluded. After exclusion because of incomplete or inaccurate data, there were 4590 cases and 22 398 controls for analysis. Of these, 2639 cases and 13 310 controls with complete data for all relevant measures were included in the main analysis. Sensitivity analysis was performed for those with missing data to evaluate potential bias. The analysis used information from the pregnancy used for matching cases and controls.
Diabetes after pregnancy was considered as a potential mediating factor. To tackle possible surveillance bias it was necessary for diagnoses of diabetes to have been recorded before the diagnosis of CVD. Cases and controls were assessed for diabetes diagnosed after pregnancy but before the diagnosis of CVD and the equivalent time point among controls. The first diagnosis of diabetes was identified through the Patient Register or through prescribed diabetes treatments recorded in the Prescription Register.
In Sweden, screening strategies for GDM during the study period have been described previously.2,11–13 Diagnosis of GDM is based on the 75-g oral glucose tolerance test (OGTT) and the main diagnostic criteria for GDM were fasting capillary whole blood glucose ≥6.1 mmol/l (fasting plasma glucose ≥7.0 mmol/l) and/or 2-hour blood glucose ≥9.0 mmol/l (plasma glucose ≥10.0 mmol/l).14 Complications during pregnancy are classified according to the Swedish version of the ICD, using the ninth revision (ICD-9) from 1987 until 1996 and the tenth revision (ICD-10) since 1997. Data on smoking are collected in early pregnancy at the first visit in maternal care. In this study, smoking was categorised as smoking/non-smoking in early pregnancy. Parity is categorised as 0, 1 or ≥2 previous births. Ethnicity is defined as Nordic (Sweden, Finland, Denmark, Norway, Iceland) or non-Nordic, according to country of birth.
Body mass index (BMI) was measured in early pregnancy and height and weight are routinely measured in Sweden by midwives, not self-reported. BMI was categorised into following groups <18.5, 18.5–24, 25–29 and >30 kg/m2. Chronic hypertension was defined as hypertension diagnosed before pregnancy, or blood pressure ≥140/90 mmHg before 20 weeks of gestation.
Education level was measured once at the time nearest to the studied pregnancy. Education level was categorised into five classes: I, no education; II, compulsory school (9–10 years of education); III, post compulsory school (2–3 years); IV, further education (3–4 years of education after post compulsory school); and V, higher education (academic studies).
Statistical methods
Conditional logistic regression was used to evaluate the association of GDM with CVD, both before and after adjustment for potential confounding factors; chronic hypertension, smoking, ethnicity, parity, BMI and education level, all modelled as categorical variables. The measure of association is the odds ratio (OR), with 95% confidence intervals (95% CI). Stratification and interaction testing were used to assess effect modification of the association of GDM with CVD by BMI (normal weight, BMI 18.5–24 kg/m2 and overweight, BMI ≥25 kg/m2), smoking and chronic hypertension. To assess the role of post-pregnancy diabetes as a mediating factor for later development of CVD, ORs were calculated with additional adjustment for diabetes mellitus. A P < 0.05 or confidence intervals not including 1.00 were regarded as statistically significant. The chi-squared test was used to compare background characteristics between cases and controls.
The statistical analysis used spss, version 18 (SPSS Inc., Chicago, IL, USA) and Stata (StataCorp 2007; Stata Statistical Software: Release 10. (College Station, TX, USA).
Results
Of the 2639 cases with complete data on all variables, 1488 (56.4%) had a diagnosis of ischaemic heart disease, 917 (34.7%) had a diagnosis of ischaemic stroke, 119 (4.5%) had a diagnosis of atherosclerosis and 192 (7.3%) had a diagnosis of peripheral ischaemic disease.
A total of 13 310 matched controls were included. Among cases there were 88 (3.3%) CVD-related deaths. Some 80.8% of the pregnancies occurred between 1991 and 1997. The women's mean age at time of diagnosis was 40.7 years (SD 7.3). The mean time period from index pregnancy to cardiovascular event was 9.1 years (SD 4.3). There were significant differences between cases and controls regarding the prevalence of GDM (2.4% versus 1.1%), chronic hypertension during pregnancy (2.1% versus 0.3%), smoking in early pregnancy (35.3% versus 18.1%), overweight (BMI >25, 42.0% versus 30.3%), parity (mean 2.44, SD 1.2 versus 2.62, SD 1.2), lower education level (I/II, 23.4% versus 15.1%) and non-Nordic origin (14.2% versus 11.4%) (Table 1).
Table 1.
Demographic and clinical characteristics for cases and controls
| Characteristics and clinical data | Cases (n = 2639) | Controls (n = 13 310) | Crude |
Adjusted* |
||
|---|---|---|---|---|---|---|
| OR (95% CI) | P value | OR (95% CI) | P value | |||
| Non-GDM | 2577 (97.6) | 13162 (98.9) | Reference | Reference | 0.019 | |
| GDM | 62 (2.4) | 148 (1.1) | 2.19 (1.59–3.01) | <0.001 | 1.51 (1.07–2.14) | |
| No chronic HT | 2584 (97.9) | 13265 (99.7) | Reference | Reference | <0.001 | |
| Chronic HT | 55 (2.1) | 45 (0.3) | 6.24 (3.99–9.78) | <0.001 | 5.10 (3.18–8.18) | |
| No smoking | 1708 (64.7) | 10898 (81.9) | Reference | Reference | <0.001 | |
| Smoking | 931 (35.3) | 2412 (18.1) | 2.48 (2.25–2.74) | <0.001 | 2.23 (2.01–2.48) | |
| Nordic | 2264 (85.8) | 11792 (88.6) | Reference | Reference | 0.125 | |
| Non-Nordic | 375 (14.2) | 1518 (11.4) | 1.28 (1.13–1.46) | <0.001 | 1.12 (0.97–1.29) | |
| Parity | ||||||
| 0 | 565 (21.4) | 1505 (11.3) | Reference | Reference | <0.001 | |
| 1 | 1012 (38.4) | 5680 (42.7) | 0.47 (0.41–0.53) | <0.001 | 0.54 (0.47–0.61) | |
| ≥2 | 1062 (40.2) | 6125 (46.0) | 0.45 (0.40–0.52) | <0.001 | 0.45 (0.39–0.51) | <0.001 |
| BMI | ||||||
| <18.5 | 62 (2.3) | 336 (2.5) | 1.11 (0.83–1.48) | 0.474 | 0.98 (0.73–1.32) | 0.910 |
| 18.5–24 | 1470 (55.7) | 8942 (67.2) | Reference | Reference | ||
| 25–29 | 686 (26.0) | 2981 (22.4) | 1.43 (1.29–1.59) | 0.003 | 1.34 (1.21–1.50) | <0.001 |
| >30 | 421 (16.0) | 1051 (7.9) | 2.37 (2.07–2.71) | <0.001 | 2.01 (1.74–2.33) | <0.001 |
| Education level** | ||||||
| I no education | 47 (1.8) | 136 (1.0) | 1.72 (1.20–2.46) | 0.003 | 1.45 (0.98–2.13) | 0.062 |
| II compulsory school | 569 (21.6) | 1882 (14.1) | 1.46 (1.30–1.65) | <0.001 | 1.23 (1.09–1.40) | 0.001 |
| III post compulsory school | 1389 (52.6) | 6603 (49.6) | Reference | Reference | ||
| IV further education | 334 (12.6) | 2447 (18.4) | 0.63 (0.55–0.72) | <0.001 | 0.76 (0.66–0.87) | <0.001 |
| V higher education ≥3 years | 300 (11.4) | 2242 (16.8) | 0.61 (0.53–0.70) | <0.001 | 0.76 (0.66–0.88) | <0.001 |
HT, hypertension.
Crude and adjusted ORs (95% CI) for the association of different variables with CVD. Values are numbers (%) unless stated otherwise.
Adjustment for GDM, chronic hypertension, smoking, BMI, ethnicity, education level, parity, respectively.
Education level: I, <9 years of education or no education; II, 9–10 years of education; III, 2–3 years of education after compulsory school; IV, 3–4 years of education after compulsory school; V, academic studies.
Associations with cardiovascular disease for GDM and other characteristics are presented in Table 1. Chronic hypertension, smoking, GDM, parity, BMI and lower education were notably associated with CVD events. After adjustment for chronic hypertension, GDM, smoking, BMI, education level, parity and ethnicity, respectively, all factors were statistically significantly associated with an increased risk of CVD events.
The unadjusted odds ratio (and 95% confidence interval) for the association of GDM with cardiovascular disease was 2.19 (95% CI 1.59–3.01) and 1.51 (95% CI 1.07–2.14) after adjustment. Table 1 demonstrates that hypertensive diseases and smoking during pregnancy have higher magnitude associations with cardiovascular disease than GDM.
Stratification by BMI indicated that GDM was statistically significantly associated with CVD among overweight women, but not among normal weight women (Table 2). For women with BMI ≥ 25 the odds ratio for the association of GDM with CVD was 2.23 (95% CI 1.38–3.62) before adjustment and 2.39 (95% CI 1.39–4.10) after adjustment, but only women with a BMI <30 accounted for this increased risk. For women with a BMI ≥30 GDM did not represent an additional risk for CVD, with an odds ratio of 0.76 (95% CI 0.16–3.58). The interaction test in the adjusted model between overweight/normal weight and GDM was statistically significant (P = 0.006). GDM was not associated with CVD in smokers (GDM did not add any risk to the baseline risk for smoking) but a risk with crude odds ratio of 2.78 (95% CI 1.98–3.91) was found in non-smokers (as shown in the Table S1). Analyses were stratified by chronic hypertension, but there were an insufficient number of participants in some cells to produce interpretable estimates.
Table 2.
Odds ratios (95% CI) for GDM as a risk factor for cardiovascular disease in normal weight versus overweight women
| Cases (n = 2639) | Controls (n = 13 310) | Crude |
Adjusted* |
||||
|---|---|---|---|---|---|---|---|
| n (%) | n (%) | OR (95% CI) | P value | OR (95% CI) | P value | ||
| Stratified by BMI | |||||||
| <18.5 (n = 398) | No GDM GDM | 62 (100.0) 0 (0.0) | 335 (99.7) 1 (0.3) | NE | NE | ||
| 18.5–24 (n = 10 412) | No GDM GDM | 1462 (99.5) 8 (0.5) | 8885 (99.4) 57 (0.6) | Reference 0.80 (0.36–1.80) | 0.601 | Reference 0.63 (0.27–1.46) | 0.280 |
| 25–29 (n = 3667) | No GDM GDM | 664 (96.8) 22 (3.2) | 2944 (98.8) 37 (1.2) | Reference 2.39 (1.00–5.72) | 0.050 | Reference 2.56 (0.99–6.63) | 0.053 |
| ≥ 30 (n = 1472) | No GDM GDM | 389 (92.4) 32 (7.6) | 998 (95.0) 53 (5.0) | Reference 1.00 (0.24–4.08) | 1.000 | Reference 0.76 (0.16–3.58) | 0.725 |
| ≥ 25 (n = 5139) | No GDM GDM | 1053 (95.1) 54 (4.9) | 3942 (97.8) 90 (2.2) | Reference 2.23 (1.38–3.62) | 0.001 | Reference 2.39 (1.39–4.10) | 0.002 |
| Interaction P-value | 0.013 | 0.006 | |||||
Adjusted for chronic hypertension, parity, smoking, education level, ethnicity.
A diagnosis of diabetes after pregnancy represents a high risk for CVD with an adjusted odds ratio of 4.80 (95% CI 3.23–7.13), and after additional adjustment for diabetes the modest statistical significance of the association of GDM with CVD among all women was eliminated. However, among overweight and obese women, where the association of GDM with CVD is most profound, additional adjustment for post-pregnancy diabetes in analysis stratified by BMI attenuated the association of GDM with CVD only slightly to 1.99 (95% CI 1.13–3.52). Pre-eclampsia was handled as a mediating factor and therefore data are not shown in the analysis. Calculations did, however, show a similar association with CVD as in previous studies and adding pre-eclampsia in the adjusted model did not change the main results.
From the original selected group of cases (4590) and controls (22 398), exclusion because of missing data was mainly due to the BMI variable, which was missing in 40.7% of the cases and 38.8% of controls; and this difference was statistically significant (P = 0.018). In 1991 and 1992 the recording of BMI in SMBR was incomplete, but coverage improved after 1992. In this material, 46.6% of the pregnancies were between 1991 (25.1%) and 1992 (21.5%). Smoking data were missing in 6.3% of the cases and 5.3% of the controls (P = 0.005). Information on education level was lacking for 0.2% in controls and in 0.1% in cases (P = 0.009).
Women among the cases and controls with missing data for any of the variables used in the analysis did not differ by age, prevalence of GDM or ethnicity, compared with women with complete data. Lower education level (classes I and II) was more common in the group with complete data; 20.7% lower education in the group with complete data compared with 16.5% in the group with missing data (P < 0.001). To assess potential selection bias, the unadjusted odds ratio for the association of GDM with CVD was calculated in the entire population (n = 26 988) with an odds ratio of 2.2 (95% CI 1.7–2.8), and this was similar to the odds ratio for the subset with complete data. Excluding the pregnancies from 1991 and 1992 did not change the results. The crude odds ratio for GDM in overweight women was 2.43 (95% CI 1.62–3.63) and in the adjusted model was 2.0 (95% CI 1.3–3.1), indicating that selection bias is unlikely or had minimal impact on the results.
Discussion
Main findings
This general population based case–control study shows that pregnancies complicated by GDM are associated with in increased risk of later CVD. In stratified analysis the association was only seen among women with a BMI ≥25 kg/m2, but only women with a BMI <30 kg/m2 accounted for this increased risk. The association between GDM and CVD in this group was comparable in magnitude to the effects of smoking or obesity and appears to be mediated, at least in part, though diabetes that develops after the pregnancy.
Strengths and weaknesses of the study
Strengths of this study include that it is general population based and that the national health registers used have been validated and have good coverage, the data were collected prospectively and the recording of exposures was not influenced by the outcome. Importantly, it was possible to adjust for maternal characteristics to confirm that the association of GDM with CVD is independent of other CVD risk factors. The matching criteria were used to take into account potential variation for local strategies for screening and diagnosis of GDM or variation in recording of maternal characteristics or disease risk. During the period 1991 to 1997, when most of the pregnancies in this study occurred, the screening and diagnosis of GDM in Sweden was quite uniform (personal communication). Regional variation in screening methods and diagnostic criteria took place during the late 1990s, but this will be addressed by the matched structure of the sample and analysis. The screening strategies used in Sweden do not identify all women with GDM, which means there is undiagnosed GDM among both cases and controls. This could diminish the precision of an association, but probably not overestimate it.
Weaknesses of the study include that a proportion of the women could not be included in the analysis due to missing data, particularly for BMI. Registration of BMI in the SMBR started in the early 1990s and during the first years, when many of our pregnancies were identified, registration was poorer. As the associations of GDM with CVD were not notably different to the main results in the group with missing values, the exclusion of those with missing data should not introduce important selection bias. Weight and height are routinely recorded by a midwife, reducing the inaccuracies associated with self-reported height and weight. There were also few diagnoses of chronic hypertension, which made stratification by hypertension not possible.
The interval between diagnosis of GDM and first CVD event was relatively short and a longer interval would have increased the rate of women with CVD and strengthened the associations.
Identification of post-pregnancy diagnoses of diabetes is likely to be incomplete, as the registers available to us did not include diagnoses from primary care. As well as inpatient and outpatient diagnoses, the registers covered prescriptions for diabetes during the later calendar period of the study. Hemce, while more severe disease is more likely to have been identified, type 2 diabetes diagnosed in primary care and managed without prescribed pharmaceutical treatment is less likely to have been identified. It is therefore possible that the mediating role of post-pregnancy onset diabetes has been somewhat underestimated in the association of GDM with CVD.
Interpretation
The main evidence before this research that GDM is a risk for subsequent CVD comes from two studies. Shah et al.6 conducted a retrospective cohort study based on administrative data from Ontario Canada, and reported an unadjusted hazard ratio for GDM of 1.71 (95% CI 1.08–2.69). Further clinical information on the women during pregnancy was not reported. Compared with their study, our research has the strength that it was possible to adjust for multiple potential confounding factors during pregnancy, so demonstrating that GDM is a useful independent marker of subsequent CVD risk. We have also shown in this study that the combination of GDM with overweight or obesity represents a particularly raised risk for later CVD and this association remains significant even after adjustment for DM that developed after pregnancy. We could also identify characteristics that potentially modify the risk of GDM for CVD. Carr et al.5 showed that in women with a family history of type 2 diabetes, previous GDM was associated with an increased prevalence of CVD. Their study, unlike ours, was limited by self-reported measures of GDM and CVD, potentially introducing recall and selection bias. In addition, the earlier study may have been rendered less precise as screening during pregnancy was not uniform. Which specific levels of hyperglycaemia are predictive of future CVD is not known. In this study we show an association for the specific diagnostic criteria used in Sweden. Whether similar associations exist for women with lower levels of hyperglycaemia, remains to be studied. The two earlier studies evaluating GDM as a risk for CVD5,6 have provided valuable information, but we believe that this study adds further clinically relevant information. Our finding that women with GDM who are overweight are at particular risk for later severe CVD has not been shown previously, as earlier work identified an association with the less severe outcome of hypertension. GDM was rarer among the normal weight women reflecting the fact that the aetiology of GDM differs. The aetiology of GDM in overweight women may be more closely related to components of the metabolic syndrome,15 while among normal weight women autoimmune and monogenic diseases are more frequently the causes.16,17 Women with GDM have characteristics of hyperinsulinaemia and hyperlipidaemia with the metabolic syndrome, which is also a risk for future CVD. Hence the association between GDM and later CVD, especially among overweight women, is reasonable. GDM is predictive of CVD in the overweight group but not the obese group. One explanation for the apparently counterintuitive result among obese women is that the risk for CVD is already so much increased in obese women that GDM does not add to this baseline risk. A similar explanation may be applicable to the association of GDM with CVD among women who were smokers. This study and the study by Pirkola et al.7 basically identified similar associations with BMI: the combination of GDM with overweight identifies a subgroup of women with underlying pathology or a greater predisposition to cardiovascular disease.
Smoking and hypertension are known risk factors for CVD18–20 and the first recommendation for prevention of CVD is smoking cessation. The present study also confirms that smoking is a strong risk factor for CVD and that GDM and smoking do not have a multiplicative effect on total risk of CVD. The incidence of hypertension in pregnancy is increasing, probably due to increasing overweight among women,21 and evaluation of the risk for CVD associated with the combination GDM and overweight might have given clinically valuable results. Unfortunately, there were too few women with chronic hypertension in our material for this analysis to produce meaningful results.
Conclusion
If recognised risks for CVD such as smoking and obesity are not present, then GDM is a potentially useful marker of future cardiovascular disease risk among women with a BMI between 25 and 29 kg/m2. The risk increase in this group is comparable in magnitude to the effects of smoking and obesity but lower than for chronic hypertension. Further studies are needed regarding the association of later CVD and degree of hyperglycaemia during pregnancy.
Disclosure of interests
None to declare.
Contribution to authorship
All authors (HF, AM, IÖ, SM, UH, ES) contributed to the design and conduct of the paper, critically reviewed manuscript details, and approved the final version.
Details of ethics approval
The study was approved by the regional Ethics Committee in Uppsala, Sweden, Number 2009/027, date of approval 2009-03-11.
Funding
The study was supported by grants from the research Committee of Örebro County Council, Sweden.
Supporting Information
Additional Supporting Information may be found in the online version of this article:
Table S1. Odds ratios (95% confidence interval) for GDM as a risk factor for cardiovascular disease in smoking versus non-smoking women
References
- 1.Metzger BE, Lowe LP, Dyer AR, Trimble ER, Chaovarindr U, Coustan DR, et al. Hyperglycemia and adverse pregnancy outcomes. N Engl J Med. 2008;358:1991–2002. doi: 10.1056/NEJMoa0707943. [DOI] [PubMed] [Google Scholar]
- 2.Fadl HE, Ostlund IK, Magnuson AF, Hanson US. Maternal and neonatal outcomes and time trends of gestational diabetes mellitus in Sweden from 1991 to 2003. Diabet Med. 2010;27:436–41. doi: 10.1111/j.1464-5491.2010.02978.x. [DOI] [PubMed] [Google Scholar]
- 3.Bellamy L, Casas JP, Hingorani AD, Williams D. Type 2 diabetes mellitus after gestational diabetes: a systematic review and meta-analysis. Lancet. 2009;373:1773–9. doi: 10.1016/S0140-6736(09)60731-5. [DOI] [PubMed] [Google Scholar]
- 4.Lauenborg J, Mathiesen E, Hansen T, Glumer C, Jorgensen T, Borch-Johnsen K, et al. The prevalence of the metabolic syndrome in a danish population of women with previous gestational diabetes mellitus is three-fold higher than in the general population. J Clin Endocrinol Metab. 2005;90:4004–10. doi: 10.1210/jc.2004-1713. [DOI] [PubMed] [Google Scholar]
- 5.Carr DB, Utzschneider KM, Hull RL, Tong J, Wallace TM, Kodama K, et al. Gestational diabetes mellitus increases the risk of cardiovascular disease in women with a family history of type 2 diabetes. Diabetes Care. 2006;29:2078–83. doi: 10.2337/dc05-2482. [DOI] [PubMed] [Google Scholar]
- 6.Shah BR, Retnakaran R, Booth GL. Increased risk of cardiovascular disease in young women following gestational diabetes mellitus. Diabetes Care. 2008;31:1668–9. doi: 10.2337/dc08-0706. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Pirkola J, Pouta A, Bloigu A, Miettola S, Hartikainen AL, Jarvelin MR, et al. Prepregnancy overweight and gestational diabetes as determinants of subsequent diabetes and hypertension after 20-year follow-up. J Clin Endocrinol Metab. 2010;95:772–8. doi: 10.1210/jc.2009-1075. [DOI] [PubMed] [Google Scholar]
- 8.Ludvigsson JF, Otterblad-Olausson P, Pettersson BU, Ekbom A. The Swedish personal identity number: possibilities and pitfalls in healthcare and medical research. Eur J Epidemiol. 2009;24:659–67. doi: 10.1007/s10654-009-9350-y. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Centre for epidemiology, National board of health and welfare. The Swedish Medical Birth Register—A summary of content and quality. Stockholm, Sweden: National Board of Health and Welfare; 2003. [Google Scholar]
- 10.Ludvigsson JF, Andersson E, Ekbom A, Feychting M, Kim JL, Reuterwall C, et al. External review and validation of the Swedish national inpatient register. BMC Public Health. 2011;11:450. doi: 10.1186/1471-2458-11-450. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Fadl HE, Ostlund IK, Hanson US. Outcomes of gestational diabetes in Sweden depending on country of birth. Acta Obstet Gynecol Scand. 2012;91:1326–30. doi: 10.1111/j.1600-0412.2012.01518.x. [DOI] [PubMed] [Google Scholar]
- 12.Ostlund I, Hanson U. Repeated random blood glucose measurements as universal screening test for gestational diabetes mellitus. Acta Obstet Gynecol Scand. 2004;83:46–51. doi: 10.1111/j.1600-0412.2004.00267.x. [DOI] [PubMed] [Google Scholar]
- 13.Anderberg E, Kallen K, Berntorp K, Frid A, Aberg A. A simplified oral glucose tolerance test in pregnancy: compliance and results. Acta Obstet Gynecol Scand. 2007;86:1432–6. doi: 10.1080/00016340701532444. [DOI] [PubMed] [Google Scholar]
- 14.Lind T, Phillips PR. Influence of pregnancy on the 75-g OGTT. A prospective multicenter study. The Diabetic Pregnancy Study Group of the European Association for the Study of Diabetes. Diabetes. 1991;40(Suppl 2):8–13. doi: 10.2337/diab.40.2.s8. [DOI] [PubMed] [Google Scholar]
- 15.Carpenter MW. Gestational diabetes, pregnancy hypertension, and late vascular disease. Diabetes Care. 2007;30(Suppl 2):S246–50. doi: 10.2337/dc07-s224. [DOI] [PubMed] [Google Scholar]
- 16.Mauricio D, de Leiva A. Autoimmune gestational diabetes mellitus: a distinct clinical entity? Diabetes Metab Res Rev. 2001;17:422–8. doi: 10.1002/dmrr.237. [DOI] [PubMed] [Google Scholar]
- 17.Weng J, Ekelund M, Lehto M, Li H, Ekberg G, Frid A, et al. Screening for MODY mutations, GAD antibodies, and type 1 diabetes-associated HLA genotypes in women with gestational diabetes mellitus. Diabetes Care. 2002;25:68–71. doi: 10.2337/diacare.25.1.68. [DOI] [PubMed] [Google Scholar]
- 18.Mosca L, Benjamin EJ, Berra K, Bezanson JL, Dolor RJ, Lloyd-Jones DM, et al. Effectiveness-based guidelines for the prevention of cardiovascular disease in women — 2011 update: a guideline from the American Heart Association. J Am Coll Cardiol. 2011;57:1404–23. doi: 10.1016/j.jacc.2011.02.005. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Garovic VD, Hayman SR. Hypertension in pregnancy: an emerging risk factor for cardiovascular disease. Nat Clin Pract Nephrol. 2007;3:613–22. doi: 10.1038/ncpneph0623. [DOI] [PubMed] [Google Scholar]
- 20.Davis MB, Duvernoy CS. How to stay heart healthy in 2011: considerations for the primary prevention of cardiovascular disease in women. Womens Health (Lond Engl) 2011;7:433–51. doi: 10.2217/whe.11.34. [DOI] [PubMed] [Google Scholar]
- 21.Seely EW, Ecker J. Clinical practice. Chronic hypertension in pregnancy. N Engl J Med. 2011;365:439–46. doi: 10.1056/NEJMcp0804872. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Table S1. Odds ratios (95% confidence interval) for GDM as a risk factor for cardiovascular disease in smoking versus non-smoking women