Diagnosis of Gestational Diabetes in Eldoret, Kenya

1.0 Background

Worldwide, 70 million women in the reproductive age have diabetes or impaired glucose tolerance placing them at risk of complications of hyperglycemia in pregnancy. The global prevalence rates of gestational diabetes mellitus (GDM) vary from between 12% and 14% depending on the population studied.(1,2) The estimated prevalence of GDM has been reported in Western (11.6%) and South Africa (3.8% - 8.8%) but because of the lack of published results and consistency in diagnosis, treatment, and outcomes in Eastern Africa, the extent of the problem of GDM is unknown in Kenya.(2) Estimates from these studies in sub-Saharan Africa (SSA) reveal a widely variable prevalence ranging from 2.4% to 14% with different methodologies and populations of different risk categories. This lack of focused investigation on GDM in Kenya is of concern as preliminary,unpublished data from the Kenyan Global Network Study on birth outcomes suggests that 5.5 % of birth weights in Western Kenya are greater than 4000g, suggesting a substantial potential burden of diabetes mellitus (DM) or GDM in this population.(3) Maternal hyperglycemia during pregancy predisposes the offspring to glucose intolerance in the future by fetal programming. (1,4,5) This vicious cycle can influence and perpetuate the incidence and prevalence of diabetes in any population. Infants born to mothers with diabetes experience double the risk of serious injury at birth, triple the likelihood of caeserean delivery and quadruple the incidence of newborn intensive care unit admissions.(6-8) These maternal, fetal, and neonatal morbidities attributable to diabetes in pregnancy can be prevented with early diagnosis of DM and effective treatment. Within the landscape of care in Western Kenya, pregnancy may be the only time a woman presents for medical care. Therefore, pregnancy is an opportune time to have a receptive audience to screen for possible pregestational diabetes, promote health education, and diagnose glucose intolerance as the woman transitions beyond the postnatal period.

MTRH is the second largest referral hospital in Kenya and serves as a referral center for the whole of Western Kenya. Although access to healthcare is limited in rural areas of Kenya, the Antenatal Clinic at the hospital serves up to 12,000 new mothers annually. Screening for GDM in the antenatal clinic at MTRH is currently limited to a routine urinalysis for glucosuria. Sutherland and colleagues found that 11% of an unselected obstetric population of 1418 women had random glucosuria. However, only 1% of those with glucosuria had an abnormal glucose tolerance test.9 Glucosuria is common in pregnancy and has not been shown to correlate with the diagnosis of GDM.

2.0 Rationale and Specific Aims

The use of a routine urinalysis as a screening strategy for GDM is not consistent with any national or international guidelines.(10-13) There is currently no consensus on the best method to screen for GDM in rural settings lacking infrastructure for venous testing. There are a variety of testing strategies which have been predominantly tested in resource-rich settings and are now used as the standard of care worldwide. Unfortunately, these testing strategies have not been validated or implemented in resource-constrained settings like those existing in Western Kenya.(2) In the United Kingdom, there is a general lack of consensus; most centers employ the 75g glucose tolerance test (Sensitivity 29% and Specificity 96% - these values are for the 75g 2hr postprandial random blood sugar and not the glucose tolerance test), others adopt a two-stage protocol, starting with 50g screen, to be followed, if abnormal, with 75-g glucose tolerance test. Also, according to the WHO criteria and ADA guidelines for GDM diagnosis, a random blood sugar >11mmol/L (198 mg/dL) or a fasting blood sugar >7.0mmol/L (126 mg/dL) meets the threshold for diagnosis of GDM.(12) Because of this uncertainty, the International Association of Diabetes and Pregnancy Study Groups (IADPSG) has recently helped provide greater clarity to this issue by issuing suggested guidelines for diagnosing both gestational diabetes and overt diabetes in pregnancy. They recommend the 75gm oral glucose tolerance test and the cutpoints derived from the Hyperglycemia and Adverse Pregnancy Outcomes (HAPO) study for all patients receiving GDM screening.

As suggested in the IADPSG guidelines and an emerging body of literature, there is growing support for the utilization of the glycosylated hemoglobin (HbA1c) in diagnosing diabetes. (16) Lipska and colleagues report that women and people of African-American ethnicity are more likely to be identified with dysglycemia using HgbA1c than by fasting blood glucose. A notable weakness of this study is that it was performed in adults age 70-79 and thus it is unlcear whether HbAIC can be used as a diagnostic tool in pregnancy. None of the currently available studies evaluating the use of A1c's for GDM have included African women.(17-20) Several studies have also raised concerns about the accuracy of HbA1c in African populations as the HbA1c results tend to be higher than those seen in other populations.(21-23) Because of the various approaches that are currently being proposed by the IADPSG guidelines, there is still considerable debate over the appropriate test to employ as they recommend performing a fasting blood sugar, HbA1c, or random plasma glucose on all women during their first prenatal visit. If a diagnosis of diabetes is not made during this assessment, it is recommended that a 75g oral glucose tolerance test is performed during the 24th to 28th week of the pregnancy to test for GDM. By combining the results of several studies they have also analyzed the additive benefits of combining different strategies. Through this investigation they have concluded that analyzing fasting blood sugars alone enabled them to identify 8.3% of the evaluable population, while the 1 hour glucose measurement identified another 5.7%, and 2 hour glucose measurement identified an additional 2.1%. One major limitation of this investigation, however, was that the cohort did not include patients reflective of the rural SSA population the investigators on this trial serve in Western Kenya. This theme is illustrated throughout this background as there is a deficiency of published results regarding GDM in SSA populations.(10) The IADPSG diagnosis approach also assumes that the vast majority of expectant mothers will have frequent contacts with the healthcare system including early prenatal visits and subsequent visits throughout the pregnancy. This trend is not commonly seen in SSA as a large proportion of expectant mothers deliver at home or have only minimal contact with the healthcare system.(2) Because of these dynamics, there is a clear need to simplify the diagnostic approach for GDM while also assessing the relative utility of each of the testing strategies (ie. random blood sugar, fasting blood sugar, 1 hr/2hr glucose tolerance test, and HbA1c).

Through this proposal, the investigators hope to take the first step towards addressing GDM in this setting by assessing the relative effectiveness of point of care testing strategies, in comparison to the gold standard approach of using the 75gm oral glucose tolerance test in a representative understudied population.