To review postpartum screening rates and studies designed to prevent progression to type 2 diabetes among women with a history of GDM, we searched published journal articles in English from January 1, 2008, through December 31, 2010. We focused on the most recent literature to augment and update other recent reviews. We searched PubMed by using the search terms 1) "gestational diabetes" and "postpartum" or 2) or "gestational diabetes" and "follow-up." We also included additional articles that we identified by reviewing reference lists of studies and review articles.

Our initial search returned 265 articles. Two authors (A.T. and E.O.) independently reviewed titles and abstracts and obtained all potential articles for additional review; they also conducted all data abstraction and reached consensus through discussion about any disagreement. Articles that were deemed irrelevant and excluded from the review were related to screening for GDM during pregnancy only, screening for metabolic syndrome or cardiovascular disease, or maternal or infant outcomes associated with GDM, type 1 diabetes, obesity during pregnancy, fertility and obstetric complications, interventions during pregnancy to improve GDM or improve pregnancy outcomes for mother or infant, or articles related to other diseases during pregnancy such as preeclampsia or polycystic ovary syndrome. Relevant articles reported on the following topics: 1) recommendations for types of screening tests, 2) rates and results of diabetes screening among women with prior GDM diagnosis at any time period postpartum, 3) results and comparisons of different types of screening tests, 4) predictors of and barriers to screening, and 5) interventions to improve rates of postpartum diabetes screening. We identified 11 published studies that evaluated rates and results of postpartum screening for diabetes among women with GDM (1-3) and 21 studies that discussed predictors of and barriers to screening or interventions to improve rates of postpartum diabetes screening (4,5). We discuss findings for the first postpartum year and the later postpartum period separately.

Both the American College of Obstetricians and Gynecologists (10) and the American Diabetes Association (3,14) recommend that women whose pregnancy was complicated by GDM be screened for persistent glucose abnormality at 6 to 12 weeks postpartum with either a fasting plasma glucose (FPG) test alone or with a fasting 75-g, 2-hour oral glucose tolerance test (OGTT). The OGTT is more sensitive, with reported sensitivities of 100% compared with 67% for FPG. The FPG test may be more acceptable to women because it requires less time, which could help overcome attrition in attendance for repeated follow-up testing in addition to being more easily tolerated (13,15). Criteria for diagnosing diabetes postpartum via a 75-g OGTT are similar to those for nonpregnant adult women (3), and thus results can identify overt diabetes mellitus, impaired fasting glucose (IFG), impaired glucose tolerance (IGT), or normal glycemia; however, the FPG test can identify IFG but not IGT. Both the American College of Obstetricians and Gynecologists and the American Diabetes Association recommend that women with a history of GDM with a normal postpartum screening be rescreened every 3 years, and women with IFG or IGT or both (prediabetes) should be rescreened annually. Recent consensus guidelines recommend follow-up testing with an OGTT at 1 year postpartum, FPG test yearly, and consideration of an OGTT at 3 years (16).

We identified 11 studies published from 2008 through 2010 (17-27) that evaluated rates of postpartum screening for diabetes among women with GDM; of those, 8 were based on review of medical records and 3 were based on mailed questionnaires or interviews (Table). Together, these studies included 32,240 women with pregnancies complicated by GDM from 1999 through 2008. Postpartum diabetes screening rates varied but were generally poor: 34% to 73% (median, 48%) of women overall were screened.

Three of these studies included large, diverse populations of women receiving care at health maintenance organizations (HMOs) in California and Oregon affiliated with Kaiser Permanente. All 3 studies demonstrated substantial increases in screening during recent years. For example, among 14,448 women with GDM who received care at Kaiser Permanente Northern California, rates of postpartum screening increased from 20% in 1995 to 56% in 2006 (18). Even in the most recent years, however, fewer than 60% of women were screened. In 5 studies of smaller populations seen at referral centers in the United States (San Francisco, Texas), Canada, Poland, and Turkey, screening rates were similar (33.7%, 57.0%, 48.2%, 37.2%, and 47.4%, respectively) (17,22,23,26,27). Morrison et al surveyed Australian women whose mean (SD) duration since delivery was 21.2 (8.4) months. Respondents reported a higher overall screening rate of 73.2%, although only 27.3% reported having been screened according to Australian Diabetes in Pregnancy Society guidelines (a 75-g OGTT administered 6 to 8 weeks postdelivery) (24).

Although the recommended timing of screening is 6 to 12 weeks postpartum, all but 1 of the studies included longer follow-up (18). However, screening rates were not markedly improved when the time window was broader. For example in 2006, screening rates were 50.3% in a study within 3 months of postpartum follow-up (20), 48.2% to 51.0% in 2 studies within 6 months of postpartum follow-up (19,22), and 55.9% in a study within 12 months of postpartum follow-up (18). Among women seen at Kaiser Southern California who were screened within 6 months, 41% were screened from 7 days through 6 weeks postpartum, 46% from 6 through 12 weeks postpartum, and 12% from 12 weeks through 6 months postpartum (19).

The type of screening tests included and results varied across the studies. Ogonowski and Miazgowski and Swan et al included only OGTT (23,25). The other 9 studies included either FPG or OGTT (17-22,24,26,27) although, of these, 2 did not report percentages of women screened with each test (17,21). In 4 studies, 79% to 100% of women were screened with FPG (18-20,27), whereas in the remaining 3 studies, most women (56%-95%) were screened with OGTT (22,24,26) (Table). A number of women screened postpartum were found to have type 2 diabetes (1.2%-4.5%) or prediabetes (12.2%-36%) (Table). In Turkey, Kerimoglu et al reported that 50% (5 of 10) of the patients who received an OGTT and 7.4% (2 of 27) of the patients who received FPG were diagnosed with type 2 diabetes (27). Other studies have found that up to one-third of women with GDM will have abnormal postpartum screening (diabetes, IFG, or IGT) (28,29). According to Kwong et al, using only FPG in postpartum screening would cause up to 75% of women with IGT or type 2 diabetes to be missed (22). Similarly, Hunt et al found that of the women with postpartum OGTT with IGT, approximately one-third had outside range values on the 2-hour value alone (26). These findings are consistent with a substantial literature that a single FPG has a sensitivity of 16% to 89% when compared with an OGTT (13,15,30).

Three studies of postpartum screening rates used different methods and did not include test results. Almario et al identified women with GDM by using laboratory data from a university teaching hospital and found that only one-third of medical records included plans to perform postpartum screening (21). Morrison et al mailed questionnaires to Australian postpartum women with a history of GDM and found a higher rate of self-reported screening at any time (73.2%) compared with the 38%-45% elsewhere at the time of the study, although many fewer women (27.3%) received an OGTT within 6 to 8 weeks postpartum as per Australian recommendations (24). Another survey of 77 Australian women with a GDM pregnancy found that 61% self-reported postpartum diabetes screening (25).

Rates of diabetes screening among postpartum women with a history of GDM differed by race and ethnicity, although findings varied (Table). For example, Stasenko et al found that Hispanic women had the lowest follow-up frequency (18% compared with 28% of white women and 29% of African American women) (17), whereas Lawrence et al found Hispanic women in another population were more likely to be screened (51% vs 48% of white women and 27% of African American women) (19). Asian women were most likely to be screened compared with women of other racial and ethnic backgrounds (odds ratio, 1.4; 95% confidence interval, 1.2-1.5) (18,19). Other patient characteristics associated with higher rates of postpartum screening included older age, nulliparity, and higher income or education. Women who received prenatal care, were treated with insulin during pregnancy, or completed a 6-week postpartum visit were also more likely to receive a postpartum diabetes screening (Table).

Of the 21 additional studies identified, 3 discussed barriers to postpartum screening (31-33). Barriers to postpartum screening included provider, patient, and system-level issues. In surveys of providers in Massachusetts, North Carolina, and Canada, providers commonly identified lack of patient attendance at a postpartum visit as the primary reason for not completing postpartum screening (31,32). However, in a chart review of data from 197 Massachusetts women with pregnancies complicated by GDM from 2000 through 2001, Smirnakis et al found that 94% of the women completed postpartum Papanicolaou screening, whereas 67% received any postpartum glucose testing and only 37% had a FPG or OGTT (34). Other barriers to postpartum screening identified by clinicians included 1) clinicians’ perception that screening guidelines were “inconsistent,” 2) patient cost or inconvenience (31), 3) lack of documentation of GDM on problem lists, and 4) poor communication between obstetricians and primary care providers (33). Patients identified time pressure as a primary barrier (32).

Postpartum reminders may help to improve diabetes screening rates. In a randomized trial of 223 women, mailed reminders resulted in improved screening rates within 12 months postpartum when sent to physicians only (52%), patients only (55%), or both (61%), compared with no reminders (14%) (35). However, only 48% of women seen at a Canadian diabetes referral clinic completed postpartum glycemic screening, even though women were given laboratory requisitions before delivery and received a telephone reminder if testing was not completed by 6 months postpartum (22). Ferrara et al reported that the proportion of women receiving an OGTT increased from 17% in 2005 to 72% in 2006, when Kaiser Permanente Northern California instituted a nurse-managed care program that included greater attention to postpartum screening guidelines (18).

Women who have a family history of diabetes, are from a high-risk racial/ethnic group (eg, South Asian), have a higher prepregnancy or postpartum body mass index (BMI) (>25 kg/m²), have a higher glucose or greater number of abnormal glucose values on prenatal glycemic screening, who are diagnosed with GDM earlier in pregnancy, or who require insulin during pregnancy are more likely to have prediabetes or type 2 diabetes on postpartum screening (7,13,28,36-41). Fewer studies have examined potentially modifiable characteristics such as maternal diet (total energy intake, fat intake), physical activity, and breastfeeding (37), but these studies suggest that certain dietary patterns may play a role in increasing risk for developing type 2 diabetes.

Risks for abnormal glycemia persist even after the early postpartum period. In a study by Retnakaran et al, among 70 women with GDM but normal OGTT at 3 months postpartum, 17% had prediabetes (IFG, IGT, or both) when reevaluated with OGTT at 12 months postpartum, compared with 3% of 73 women with normal glucose tolerance both during pregnancy and at 3 months postpartum (4). Even women with lesser degrees of abnormal glucose during pregnancy were at elevated risk for abnormal 12-month OGTT despite a normal 3-month OGTT (39). Risk factors for abnormal 12-month screening included higher 3-month postpartum glucose, BMI, low density lipoprotein cholesterol, triglycerides, and leptin, and lower high density lipoprotein cholesterol and adiponectin (4).

Risks for prediabetes, type 2 diabetes, and metabolic syndrome continue to increase over time, since pregnancy. Recent studies found that 16% to 30% of women with GDM develop type 2 diabetes by 5 to 10 years postpartum (42-44), and more cases accumulate even later. In a 2009 meta-analysis of 20 studies that included 675,455 women and 10,859 type 2 diabetes events, women with a GDM pregnancy had a nearly 7.5-fold increased risk of developing diabetes (8). Within 5 years of a pregnancy complicated by GDM, women had a relative risk of 4.7, which increased to 9.3 in those who were examined more than 5 years postpartum. In a 2010 population-based cohort study of more than 6,000 women from northern Finland followed since their pregnancies in 1986, the cumulative incidence of diabetes continued to increase even at 20 years postpartum. Rates were particularly high among women with a history of GDM who were overweight (26%), compared with controls who had no risk factors for GDM (0.7%) (45). Risk factors for late progression to type 2 diabetes are similar to risk factors for early progression and include low insulin sensitivity, insulin resistance and progressive insulin secretory defect, and gain in weight and body fat.

Although we identified several recent screening studies completed within the first postpartum year, we found no recent studies that evaluated rates of diabetes screening in clinical settings after this time period. We found only 1 published intervention to promote longer-term follow-up of women with a history of GDM; in South Australia beginning in 2002, women with GDM were invited to provide contact information to a GDM registry, which then sent women annual mailed reminders for glycemic screening and requests for diabetes screening updates beginning at 15 months postpartum (46). Findings suggest that these reminders helped promote screening. While 56% of women reported glycemic screening in the first postpartum year (before the intervention began), 75% reported having been screened during the second postpartum year. However, fewer than half of eligible women responded to the update requests, and thus reporting bias was likely.