Article Text

Original article
An association between gestational diabetes mellitus and long-term maternal cardiovascular morbidity
  1. Roy Kessous1,
  2. Ilana Shoham-Vardi2,
  3. Gali Pariente1,
  4. Michael Sherf3,
  5. Eyal Sheiner1
  1. 1Department of Obstetrics and Gynecology, Faculty of Health Sciences, Soroka University Medical Center, Ben-Gurion University of the Negev, Beer-Sheva, Israel
  2. 2Epidemiology and Health Services Evaluation, Ben-Gurion University of the Negev, Beer-Sheva, Israel
  3. 3Clalit Health Services (Southern District), Faculty of Health Sciences, Soroka University Medical Center, Ben-Gurion University of the Negev, Beer-Sheva, Israel
  1. Correspondence to Dr Roy Kessous, Department of Obstetrics and Gynecology, Soroka University Medical Center, POB 151, Beer Sheva 84101, Israel; kessousr{at}


Objective To investigate whether a diagnosis of gestational diabetes mellitus (GDM) is a risk factor for subsequent long-term cardiovascular morbidity.

Design A population-based study.

Setting Soroka University Medical Center, a tertiary centre in the southern region of Israel.

Patients A cohort of women with and without a diagnosis of GDM who delivered during the years 1988–1999 with a follow-up period until 2010.

Interventions A comparison of the incidence of cardiovascular morbidity.

Results Of 47 909 deliveries that met the inclusion criteria, 4928 (10.3%) occurred in patients who were diagnosed with GDM. During a follow-up period of more than 10 years, compared with women who gave birth at the same time period, after adjustment for age and ethnicity, patients with GDM had higher rates of cardiovascular morbidity including non-invasive cardiac diagnostic procedures (OR=1.8; 95% CI 1.4 to 2.2), simple cardiovascular events (OR=2.7; 95% CI 2.4 to 3.1) and total cardiovascular hospitalisations (OR=2.3; 95% CI 2.0 to 2.5). In a Cox proportional hazards model, adjusted for comorbidities such as pre-eclampsia and obesity, GDM was independently associated with cardiovascular hospitalisations (adjusted HR 2.6, 95% CI 2.3 to 3).

Conclusions GDM is an independent risk factor for long-term cardiovascular morbidity in a follow-up period of more than a decade.

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.


Gestational diabetes mellitus (GDM) is defined as glucose intolerance that begins or is first recognised during pregnancy.1 The prevalence of GDM differs for different populations and ethnic groups and varies between 1% and 14%,1–5 with increasing rates due to the increased prevalence of obesity and type 2 DM.2 The most common associated risk factors for GDM are advanced maternal age, obesity, excess weight gain during pregnancy, a family history of DM, previous pregnancy with GDM, and hypertensive disorders.2 ,5

The association between GDM and type 2 DM is well recognised, although the exact relative risk and the interval to development of DM are not well established. In a review of 28 studies (performed between the years 1965 and 2001), Kim et al4 found that, in a follow-up of 6 weeks to 28 years after an index pregnancy with GDM, the cumulative incidence of DM was 2.6–70%.4 In a meta-analysis6 of data from 20 cohort studies, women with GDM in their previous pregnancy had a 7.43-fold increased risk of subsequently developing DM.6

Metabolic syndrome encompasses several features such as obesity, glucose intolerance, dyslipidaemia and hypertension2 ,7 ,8. Metabolic syndrome is related to endothelial damage, oxidative stress and inflammatory response, which in turn increase the risk of atherosclerosis and the potential future risk of cardiovascular diseases (CVDs).2 ,7

A higher incidence of metabolic syndrome has been found in women with previous GDM.9 ,10 Di Cianni et al10 studied the prevalence of metabolic syndrome and other cardiovascular risk factors in 166 women with a history of GDM 16 months after the delivery and found a significantly higher rate of metabolic syndrome and higher levels of inflammatory markers (such as C-reactive protein) in women with a previous history of GDM.10 Bo et al11 compared biochemical markers of vascular endothelial dysfunction in 82 patients with GDM and 113 patients without this diagnosis. They found higher levels of markers in women with previous GDM and concluded that these patients have higher rates of endothelial damage and accordingly higher risk of future CVD.11 Vrachnis et al12 discussed the relationship among inflammatory markers, metabolic abnormalities, and vascular dysfunction in women with previous GDM and concluded that these patients should be considered a population at risk of future CVD.

Surprisingly, only a few studies have investigated the association between a previous diagnosis of GDM and subsequent cardiovascular morbidity.13 ,14 Those studies found that patients with GDM are probably more likely to develop long-term CVD. However, in one of these studies the authors concluded that the increased risk is related to the subsequent development of overt DM and in another study the incidence of CVD is based on self-report.

Owing to the relatively scarce data on this subject, the objective of the present population-based study was to investigate whether a diagnosis of GDM is an independent risk factor for subsequent CVD hospitalisations and specific long-term cardiovascular morbidity during a follow-up period of more than a decade.

Materials and methods


The study was conducted at the Soroka University Medical Center, the sole hospital of the Negev, the southern region of Israel, serving the entire population in this region. Thus, the study is based on non-selected population data. The institutional review board (in accordance with the Helsinki declaration) approved the study.

Study population

The study population comprised all patients with or without GDM who delivered between the years 1988 and 1998; these patients were retrospectively followed-up until 2010. Patients with lack of prenatal care, multiple pregnancies, preGDM and known CVD before or during the current pregnancy were excluded from the study.

Study design

A population-based retrospective cohort study was conducted, comparing all deliveries of patients with or without GDM. The primary exposure was GDM. The screening test was the glucose challenge test—that is, glucose measurement 1 h after a 50 g glucose load—which was performed on all pregnant women. Patients with a pathological value (above 140 mg/dL) in this test underwent the 100 g 3 h oral glucose tolerance test performed after an overnight fast. GDM was diagnosed when a patient had at least two pathological values in this test. A retrospective follow-up of hospitalisations due to cardiovascular morbidity 10–20 years after the index birth was performed. Cardiovascular morbidity was defined as hospitalisations for any cardiovascular reason at the first cardiovascular hospitalisation at Soroka University Medical Center. Cardiovascular morbidity was divided into four categories according to severity and type, including simple and complex cardiovascular events (eg, angina pectoris and congestive heart failure, respectively), and invasive and non-invasive cardiac procedures (eg, insertion of a stent and a treadmill stress test, respectively). The exact International Classification of Diseases (ICD) codes for each subtype of cardiovascular morbidity are presented in the online supplementary appendix.

Data were collected from two databases that were cross-linked and merged: the computerised perinatal database and the computerised hospitalisation database of the Soroka University Medical Center. The perinatal database consists of information recorded directly after delivery by an obstetrician. Skilled medical secretaries routinely review the information before entering it into the database. Coding was performed after assessment of prenatal medical care records together with the routine hospital documents. The hospitalisation database includes demographic information and ICD-9 codes for all medical diagnoses made during hospitalisations.

Statistical analysis

Statistical analysis was performed using the SPSS package V.17 edition. Statistical significance was calculated using the χ2 test for differences in qualitative variables, and the Student t test for differences in continuous variables. Multivariate logistic regression models were constructed in order to control for confounders such as ethnicity and maternal age. ORs and their 95% CIs were computed. p<0.05 was considered significant.

Kaplan–Meier survival curves were used to compare cumulative incidence of cardiovascular hospitalisations. Cox proportional hazards models were used to estimate the adjusted HRs and 95% CIs for long-term cardiovascular hospitalisations. p<0.05 was considered significant.


During the study period there were 47 909 women who met the inclusion criteria; 4928 of these (10.3%) were diagnosed with GDM. The first of their deliveries was considered the index birth.

Table 1 summarises characteristics of the index delivery of patients with and without a diagnosis of GDM. Patients in the GDM group were older than those in the comparison group. In addition, a significant difference was noticed in the rate of Jewish versus Bedouin parturients diagnosed as having GDM.

Table 1

Characteristics of patients during the follow-up period divided into those with and without a history of gestational diabetes mellitus (GDM)

Table 2 presents a comparison of cardiovascular morbidity and hospitalisations during the follow-up period of more than 10 years, after ethnicity and age had been controlled for. Before performance of the logistic regression models, patients with GDM had higher rates of cardiovascular morbidity, including invasive and non-invasive cardiac diagnostic procedures, simple and complex cardiovascular events, and hospitalisations due to cardiovascular causes. In multivariate logistic regression models controlling for maternal age and ethnicity, GDM remained an independent risk factor for maternal long-term risk of non-invasive diagnostic procedures, simple cardiovascular events, and cardiovascular hospitalisations.

Table 2

Incidence of cardiovascular morbidity and hospitalisation during the follow-up period in patients with and without a history of gestational diabetes mellitus (GDM) after adjustment for maternal age and ethnicity

Figure 1 presents a Kaplan–Meier survival analysis for the cumulative incidence of cardiovascular hospitalisations after the index birth in both study groups (with and without GDM). Patients with a history of GDM had a significantly higher risk of cardiovascular events during the whole follow-up period.

Figure 1

Kaplan–Meier survival analysis curve (presenting the log survival function) for cardiovascular-associated hospitalisation of patients with and without a history of gestational diabetes mellitus (GDM). p=0.001 (log rank test). This figure is only reproduced in colour in the online version.

Table 3 presents Cox proportional hazards models that were used to estimate the adjusted HRs and 95% CIs for long-term cardiovascular hospitalisations. After other recognised confounders related to the metabolic syndrome, such as pre-eclampsia and obesity, had been controlled for, GDM remained independently associated with an increased risk of subsequent cardiovascular hospitalisations (adjusted HR 2.6, 95% CI 2.3 to 2.9). In addition, we tested interactions between diabetes and pre-eclampsia/toxemia (PET) and between diabetes and obesity; none were found to be statistically significant.

Table 3

Cox multivariate regression model for the risk of cardiovascular hospitalisation


The link between pregnancy complications and future risk of CVD has been previously studied with a specific focus on pre-eclampsia.15–17 Data on the link between GDM and future risk of CVD are relatively scarce. The major finding of the present study is that GDM is an independent risk factor for long-term cardiovascular morbidity and cardiovascular-related hospitalisations.

Shah et al13 compared 8191 patients with previous GDM with 81 262 patients without GDM. In their study, a relative risk of 1.7 (95% CI 1.08 to 2.69) was noted for CVD in patients with GDM. However, after they controlled for subsequent DM, the relative risk decreased to 1.13 (95% CI 0.67 to 1.89) and lost its significance. The authors concluded that these patients have an increased risk of CVD mainly due to the development of DM.13 In our study, we performed several statistical tests in order to adjust for the time of surveillance and control for confounders such as maternal age and confounders related to metabolic syndrome such as obesity and pre-eclampsia. Carr et al14 studied 332 patients with and 663 without a history of GDM in addition to a family history of type 2 DM. They found patients with a history of GDM to be at higher risk of future CVD based on self-reports from the patients. In the present study, data were obtained from the computerised files and not based on self-reports, which eliminates the risk of recall bias.

We performed a further categorisation of the cardiovascular complications according to ICD code for diagnostic procedures and severity. It emphasises the importance of a previous history of GDM as a risk factor not only for the end point of cardiovascular hospitalisations but also for simple cardiovascular events.

A significant difference was noted in the prevalence of GDM between patients of Jewish and Bedouin ethnicity. The higher prevalence of GDM among Jewish patients can possibly be explained by the different accessibility to health services and the fact that Bedouins are a traditional people who tend to underutilise the existing prenatal care services.18 ,19 This probably increases the strength of our results, since we can assume that there were patients with undiagnosed GDM in the control group. Furthermore, the significant association between GDM and long-term cardiovascular morbidity remained significant even after we controlled for ethnicity.

Our population-based study offers the strength of a large sample size that allows a comparison and conclusions regarding this important obstetric complication. Also, we obtained data from computerised files and not based on self-reports, which might lead to recall bias. The main strength of the study lies in the fact that our hospital is the only one serving the entire population of southern Israel. The hospital provides both maternity services and tertiary cardiovascular medical services, thus as long as patients live in the area they would use the hospital. However, cardiovascular events that occurred outside of the hospital could not be ascertained. It is therefore possible that some cardiovascular events were missed. However, the same stands for patients with and without GDM.

In conclusion, according to the results of our study, GDM is an independent risk factor for long-term cardiovascular complications and for hospitalisations due to cardiovascular causes. It is of major importance to obstetricians counselling patients about future risk of cardiovascular complications of GDM for secondary prevention, early detection and perhaps specific screening programmes for this population. It is also important for the primary care physicians and cardiovascular specialists who encounter these patients years after the pregnancy trying to asses risk factors for CVD.



  • Abstract presented in part at the SMFM 2013 Annual Meeting, San Francisco, CA, USA, Control ID: 1473235.

  • Contributors All authors fulfil the criteria of authorship.

  • Competing interests None.

  • Provenance and peer review Not commissioned; externally peer reviewed.