Objective
The objective of the study was to examine the rates of gestational diabetes mellitus (GDM) associated with both maternal and paternal race/ethnicity.
Study Design
This was a retrospective cohort study of all women delivered within a managed care network. Rates of GDM were calculated for maternal, paternal, and combined race/ethnicity.
Results
Among the 139,848 women with identified race/ethnicity, Asians had the highest rate ( P < .001) of GDM (6.8%) as compared with whites (3.4%), African Americans (3.2%), and Hispanics (4.9%). When examining race/ethnicity controlling for potential confounders, we found that the rates of GDM were higher among Asian (adjusted odds ratio [aOR], 1.5; 95% confidence interval [CI], 1.4–1.6) and Hispanic (aOR, 1.2; 95% CI, 1.1–1.4) women as well as Asian (aOR, 1.4; 95% CI, 1.3–1.5) and Hispanic (aOR, 1.3; 95% CI, 1.2–1.4) men as compared with their white counterparts.
Conclusion
We found that rates of GDM are affected by both maternal and paternal race/ethnicity. In both Asians and Hispanics, maternal and paternal race are equally associated with an increase in GDM. These differences may inform further investigation of the pathophysiology of GDM.
Interracial couples represent a unique opportunity to study the interplay of genetics and environment on perinatal complications. The majority of research into interracial perinatal outcomes has focused on African American/white interracial couples, with findings of increased rates of stillbirth, low birthweight, and preterm birth seen among interracial African American/white couples as compared with white/white couples. However, the risks of perinatal complications among interracial African American/white couples were lower than African American couples. This finding of a racial/ethnic gradient in these outcomes could be due to either genetic or socioeconomic differences.
There have also been several studies of Asian/white couples. Although it is difficult to estimate the incidence of Asian/white couples, the US Census Bureau’s 2000 report showed that 14.3% of all people reporting Asian race also reported being of mixed Asian/white ancestry. A 2005 retrospective study of 324 Japanese mother/white father couples found no significant difference in the rate of preeclampsia when compared with same-race Japanese couples. However, we previously demonstrated a lower risk of preeclampsia with Asian paternal race and an increased risk of preeclampsia with differing maternal and paternal race. So it may be that these 2 factors were balanced in the study of Japanese couples. In a prior study from a single institution of just Asian/white couples, there was demonstration of greater rates of gestational diabetes from either Asian maternity or Asian paternity.
Aside from these studies, there has been little other work to investigate other permutations of interracial couples and none examining mixed ethnicity studies including Hispanic mothers or fathers. Similar to preeclampsia, gestational diabetes is a pregnancy complication that depends on both the woman’s predilection toward developing the disease and a placental contribution in the form of placental hormones. The genes underlying the placental hormones can be both maternally or paternally derived. Because there is evidence to suggest that gestational diabetes is associated with varying race/ethnicity, this provides a particularly interesting outcome to examine the combined effect of maternal and paternal race/ethnicity.
Given this background, we conducted a retrospective cohort study of both maternal and paternal race/ethnicity and gestational diabetes. We hypothesized that the effect of race/ethnicity should be seen from both parents but that the effect may be stronger from the maternal race/ethnicity as she contributes both to underlying predisposition to type 2 diabetes mellitus as well as to the placental genome.
Materials and Methods
To address this hypothesis, we designed a retrospective cohort study of all women who delivered singleton pregnancies within the Northern California Region of the Kaiser Permanente Medical Care Program (KPMCP) from Jan. 1, 1995, to Dec. 31, 1999. The primary outcome was the incidence of gestational diabetes mellitus (GDM). Maternal and neonatal demographic data were obtained by scanning the KPMCP hospitalization database and linking maternal and neonatal records using methods that have been described elsewhere.
Discharge diagnoses were assigned by the delivering clinician and entered into the KPMCP database by professional medical coders. Of note, all pregnant women cared for within KPMCP during the study time period were screened for gestational diabetes with a 50 g glucose loading test, and women who received a diagnosis of GDM were all entered into a KPMCP-wide perinatal care program within 1-2 weeks of their diagnosis. Maternal and paternal race/ethnicity was obtained from the State of California Birth Certificate Database based on self-reported race/ethnicity and divided into 5 groups: African American, Asian, white, Hispanic, and Native American.
Women without information regarding race/ethnicity on themselves or the father of the baby were excluded as were those with a preexisting diagnosis of diabetes mellitus. Finally, if the mothers or the fathers were of racial/ethnic groups that did not clearly fit into 1 of the 5 racial/ethnic groups listed above (eg, mixed ethnicity), they were also excluded. Of note, Asian race/ethnicity includes the following designations: Japanese, Chinese, Korean, Vietnamese, Indian, Laotian, Thai, Cambodian, and Filipino. Because Tongan and Samoan individuals have specific phenotypic differences and different perinatal outcomes than other Asian women, these individuals were excluded from this study. Hispanic ethnicity includes the general Spanish/Hispanic category as well as those from Mexico, Central or South America, Cuba, and Puerto Rico.
Rates of GDM were compared among each of the 4 major racial/ethnic groups overall and by paternal race/ethnicity utilizing the χ 2 test. To control for potential confounders, multivariable analyses that included maternal parity, age, educational attainment, year of delivery, and obesity (body mass index [BMI] ≥30kg/m 2 ) were used and compared Asians, African Americans, and Latinas to whites. Interaction terms, created as cross-product variables, between ethnicity, education, maternal age, and parity were created and investigated in the multivariate analyses to examine for effect modification. Using backward, stepwise logistic regression, variables that did not have a P value ≤ .20 were removed from the full model and the restricted vs full models compared using the likelihood ratio χ 2 test.
A cluster analysis was utilized to control for women with multiple pregnancies. Several variables were left in the model for their clinical relevance despite lack of statistical significance, including parity and age, as long as they did not distort the results of the remaining variables.
To assess goodness of fit for the multivariable models, we utilized the Hosmer-Lemeshow χ 2 test. For this test, if the P value is large, the model is well calibrated; conversely, if the P value is small, the model may need to be further modified. We used a threshold of P > .05 as acceptable. We evaluated the discrimination of the final model using the c-statistic. A model that accurately discriminates perfectly would have a c-statistic of 1.0, whereas with completely random predictions, as in the flip of a coin, the c-statistic would be 0.5.
For a clinical prediction rule, it is generally considered that a c-statistic of less than 0.6 has minimal clinical value, greater than 0.8 has discrimination that could be utilized clinically, and values between 0.6 to 0.8 perhaps having utility in various situations. Finally, the relative contribution of each of variables included in the final model was determined using a log-likelihood ratio comparison between the full and individually restricted models using a technique described.
All analyses were conducted using STATA (Stata Corp, College Station, TX), and P ≤ .05 were considered statistically significant. This study was approved by the KPMCP Institutional Review Board for the Protection of Human Subjects, which has jurisdiction over all the hospitals included in this study as well as the Committee on Human Research at University of California, San Francisco.
Results
Overall, there were 139,848 couples with singleton gestations in which both parents belonged to 1 of the 5 racial/ethnic groups defined and the mother did not have preexisting diabetes mellitus. The women of differing race/ethnicity varied in terms of maternal age, parity, education, and other obstetric factors ( Table 1 ). The incidence of GDM differed by race/ethnicity as well ( Table 2 ), with Asians having the highest rate (6.8%) then, in decreasing order, Native Americans (5.6%), Latinas (4.9%), whites (3.4%), and African Americans (3.2%; P < .001). These trends appeared to hold when examined by paternal racial/ethnic groups.
| Characteristic | White | African American | Latina | Asian | Native American | P value a |
|---|---|---|---|---|---|---|
| Total number of women | 62,823 | 14,069 | 35,786 | 26,370 | 800 | |
| Maternal age, y (mean ± SD) | 29.1 ± 6.0 | 26.7 ± 6.5 | 27.0 ± 6.0 | 30.0 ± 5.7 | 26.2 ± 8.1 | < .001 |
| Maternal age ≥35 y, % | 19.1 | 13.6 | 11.6 | 22.1 | 11.1 | < .001 |
| Nulliparous, % | 43.7 | 43.5 | 37.8 | 46.4 | 36.5 | < .001 |
| High school graduate, % | 92.8 | 89.6 | 69.7 | 93.5 | 86.7 | < .001 |
| College graduate, % | 32.3 | 18.3 | 11.6 | 40.0 | 19.6 | < .001 |
| Birthweight, g (mean ± SD) | 3569 ± 484 | 3374 ± 486 | 3507 ± 478 | 3359 ± 461 | 3447 ± 569 | < .001 |
| ≥4000 g, % | 18.2 | 10.1 | 14.7 | 8.6 | 12.3 | < .001 |
| ≥4500 g, % | 3.3 | 1.7 | 2.6 | 1.3 | 1.8 | < .001 |
| ≥41 wks, % | 19.5 | 17.5 | 17.1 | 12.5 | 16.9 | < .001 |
| ≥42 wks, % | 3.0 | 2.7 | 2.3 | 1.7 | 2.34 | < .001 |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
