Placental Histomorphometry in Gestational Diabetes Mellitus

Abstract

Objectives

We examined placental histomorphometry in gestational diabetes mellitus (GDM) for factors associated with race/ethnicity and subsequent type 2 diabetes mellitus (T2DM).

Methods

We identified 124 placentas from singleton, full-term live births whose mothers had clinically defined GDM and self-reported race/ethnicity. Clinical and placental diagnoses were abstracted from medical records.

Results

Forty-eight white and 76 nonwhite women were followed for 4.1 years (median, range 0.0-8.9 years). White women developed less T2DM (12.5% vs 35.5%; P = .005) but had higher systolic (mean ± SD, 116 ± 13 vs 109 ± 11 mm Hg; P < .001) and diastolic (71 ± 9 vs 68 ± 7 mm Hg; P = .02) blood pressure, more smoking (35.4% vs 10.5%; P = .004), and more chorangiosis (52.1% vs 30.3%; P = .02) than nonwhite women.

Conclusions

Although more nonwhite women developed T2DM, more white women had chorangiosis, possibly secondary to the higher percentage of smokers among them. Further study is necessary to elucidate the relationship among chorangiosis, subsequent maternal T2DM, and race.

Gestational diabetes mellitus (GDM) affects approximately 7% of all pregnancies in the United States, but prevalence estimates are as high as 18% among women of nonwhite race/ethnicity.¹ GDM increases the risk for type 2 diabetes mellitus (T2DM) and cardiovascular disease subsequent to pregnancy²; therefore, early risk stratification of women with GDM is critical to the development of appropriate primary prevention strategies. Although several risk factors have been identified as associated with progression to T2DM in women with GDM, including higher body mass index, older age, and higher frequency of GDM in past pregnancies,³ these factors incompletely identify the risk. Because the placenta reflects the maternal metabolic milieu during pregnancy, the use of placenta-derived data for subsequent maternal disease is another opportunity to inform risk stratification in this population.

The placenta is the critical organ responsible for fetal growth and development, as well as the transfer of blood, oxygen, nutrients, and waste between the mother and the fetus.⁴ Normal placental anatomy comprises the lacuna, floating and anchoring villi, villous and extravillous cytotrophoblasts, syncytiotrophoblasts, uterine blood vessels, and uterine connective tissue.⁵ Pathologic changes in placental structure and function have been observed in type 1 diabetes mellitus (T1DM)⁶ and T2DM.⁷ In addition, placental abnormalities in GDM have been reported⁸ and have paralleled those associated with pregestational diabetes, including increased fetal and placental weight, diameter, and thickness⁹; cytotrophoblastic hyperplasia; villous edema and fibrin deposits; and chorangiosis.⁸ However, placental abnormalities in GDM have not been associated with subsequent maternal risk for T2DM. In addition, the role of race/ethnicity in GDM placental pathology has not been fully elucidated.

Therefore, we sought to compare placental histomorphometry from pregnancies complicated by GDM and examine factors that distinguished women from racially/ethnically diverse populations compared with white populations. We hypothesized that the placentas of women from racially/ethnically diverse populations would have a higher prevalence of vascular abnormalities given the higher prevalence of GDM and concomitant cardiovascular disease risk among these populations.² Second, because women with GDM have similar placental pathology as women with T1DM and T2DM, we hypothesized that placental pathology would provide insight regarding the risk of subsequent T2DM postpartum.

Materials and Methods

Population Selection

Participants for this study were selected from placental pathology specimens received in the Massachusetts General Hospital (MGH) Pathology Department between January 1, 2001, and December 31, 2009 (n = 765). We then cross-referenced this population with a population of women who presented for prenatal care to the MGH Obstetrical Department between September 1998 and January 2007 and were participants in the MGH Obstetric Maternal Study (MOMS).¹⁰ These women, initially followed during pregnancy for the development of preeclampsia, were subsequently observed for a median of 11.1 years postpartum and had data on the development of T2DM subsequent to pregnancy.

Of the population of women with pathology specimens who also had longitudinal clinical data from the MOMS data set (n = 178), we selected those categorized as having a clinically confirmed diagnosis of GDM by Carpenter-Coustan criteria.¹¹ Women were diagnosed if they had two or more venous plasma glucose values greater than or equal to the defined threshold levels (fasting, ≥95 mg/dL; 1 hour, ≥180 mg/dL; 2 hours, ≥155 mg/dL; and 3 hours, ≥140 mg/dL) on a 100-g oral glucose tolerance test. Among the women who met the GDM criteria (n = 157), only those who delivered full-term, singleton live births were selected for the study (n = 129).

Race and ethnicity were self-reported, and the women identified themselves as black, white, Asian, Hispanic, or “other.” Those who did not self-identify race/ethnicity were categorized as “unknown” (n = 5) and were not included in the population of women examined by race/ethnicity (n = 124). All participants completed informed written consent to encompass this study, and the study protocol was approved by the Partners Human Research Committee Institutional Review Board.

Placenta Specimen Examination

The pathologic specimens previously had been processed by gross and histologic examinations. A standard gross template was followed, and at least three sections of placenta were sampled for histology (cord, membranes, and parenchyma). One of four perinatal pathology experts rotating on service (including D.J.R.) performed the initial diagnostic examination and reported the findings. All placental specimens selected for our database were reexamined by two authors (D.J.R. and D.L.D.), who were both blinded to clinical history.

The gross parameters examined included placental weight, cord weight, cord length, and cord insertion site. Samples were also coded for villous maturation, categorized as mature, slightly immature, or immature. Mature villi had at least a minimum of one vasculosyncytial membrane and one syncytiotrophoblastic knot per two terminal villi. Slightly immature villi were coded for villi bordering the two other classifications but not meeting criteria for villous maturational arrest.¹² Immature villi were coded when the villi were large without vasculosyncytial membranes and with a prominent cytotrophoblastic layer.

On histologic examination, we scored and diagnosed villous maturational arrest; dysmorphic villi, defined as villi with prominent trophoblastic layers without knots, irregular villous contours, open villous stroma, and hypovascularization; and chorangiosis, defined as 20 or more capillaries in a cluster of villi in three ×40 fields.¹³ We captured maternal inflammation as an aggregate of the measures of villitis of unclear or unknown etiology¹⁴; interface inflammation, confluent basal plate chronic villitis with deciduitis, and/or confluent chronic inflammation on the fetal side of the chorionic plate; and plasma cell deciduitis. Maternal and fetal evidence of acute chorioamnionitis was classified using the Redline nosology.¹⁵ Measures of fetal stress identified included the presence of an increased number of circulating fetal nucleated RBCs and multifocal or diffuse acute villous edema. Measures of placental perfusion pathologies included placental infarcts, distal villous hypoplasia, and fetal thrombotic vasculopathy.¹⁶ Decidual vasculopathy was characterized by fibrinoid necrosis with or without atherosis in the decidua capsularis, decidual necrosis with atherosis in the decidua basalis, or the presence of small muscularized arterioles in the decidua basalis. Meconium, fibrin deposition, intervillous thrombi, and calcification were also evaluated.

Statistical Analysis

The placental pathologic and clinical parameters were compared and used to evaluate the differences in placental pathology and vasculature between the two populations. Continuous variables were summarized using means and standard deviations, while frequency distributions accounting for missing values were used for categorical variables. Summary characteristics of white vs nonwhite subjects were compared using Mann-Whitney tests and χ² tests where appropriate. To complete our analysis of the placenta, clinical and pathologic data were analyzed and compared to examine the differences in placental vasculature in women from the nonwhite population to the white population. The statistical analyses were performed using the SAS for Windows version 9.1 statistical software package (SAS Institute, Cary, NC). A P value less than .05 was considered statistically significant.

Results

Clinical Findings

From the initial 765 specimens available, we identified the 129 women with biochemically defined GDM who had full-term, singleton live births. Among these 129 women, 5 women did not identify their race/ethnicity and were excluded from the analysis population (n = 124). Maternal and neonatal baseline characteristics for the study population are listed in Table 1. Of the 124 women who provided information regarding their race, 48 were white, four were black, 14 were Asian, 51 were Hispanic, one was Native American, and six were “other” nonwhite but unspecified race/ethnicity. The women were followed for a median of 4.1 (range, 0.0-8.9) years from delivery to last MGH encounter, and we observed that the development of T2DM after pregnancy among the nonwhite population was nearly three times more frequent compared with the white population (35.5% vs 12.5%; P = .005). White women also had higher systolic (mean ± SD, 116 ± 13 vs 109 ± 11 mm Hg; P < .001) and diastolic (71 ± 9 vs 68 ± 7 mm Hg; P = .02) blood pressure compared with nonwhite women. Notably, the number of past or current smokers was significantly higher among the white compared with the nonwhite women (35.4% vs 10.5%; P = .004). Otherwise, both the white and nonwhite populations had similar clinical, prenatal, and perinatal parameters.

Table 1.

Clinical Characteristics of the Study Population With Gestational Diabetes Mellitus^a

Characteristic	All	White	Nonwhiteb	P Value
No. of patients	129	48	76
Age, y	32.4 ± 5.8	32.7 ± 5.6	32.0 ± 6.0	.71
Body mass index, kg/m2	30.0 ± 6.7	30.5 ± 7.2	29.8 ± 6.4	.71
Systolic blood pressure, mm Hg	112 ± 12	116 ± 13	109 ± 11	.001c
Diastolic blood pressure, mm Hg	69 ± 8	71 ± 9	68 ± 7	.02c
Total cholesterol, mg/dL	195 ± 38	189 ± 41	199 ± 37	.52
High-density lipoprotein, mg/dL	53 ± 13	55 ± 14	51 ± 13	.44
Low-density lipoprotein, mg/dL	114 ± 28	109 ± 28	116 ± 29	.49
Triglycerides, mg/dL	172 ±128	188 ± 174	163 ± 90	.75
Gestational age at prenatal visit, wk	12.1 ± 6.0	11.5 ± 5.2	12.2 ± 6.3	.88
No. of total pregnancies	2.7 ± 1.7	2.4 ± 1.4	2.7 ± 1.6	.44
No. of live births	1.0 ± 1.3	0.7 ± 0.8	1.1 ± 1.3	.19
Weight gain, lb	21.3 ± 12.7	21.9 ± 13.8	20.7 ± 12.1	.85
Preeclampsia	2 (1.6)	1 (2.1)	1 (1.3)	.74
Gestational age at delivery, wk	39.5 ± 1.1	39.4 ± 1.1	39.4 ± 3.3	.53
Baby weight, g	3,670 ± 618	3,667 ± 505	3,676 ± 699	.66
Birth weight for gestational age percentile, %	62.9 ± 30.0	64.2 ± 29.0	61.7 ± 31.3	.64
Time from MGH delivery to follow-up, median (range), y	4.1 (0-8.9)	3.7 (0-8.5)	4.1 (0-8.9)	.96
Essential hypertensiond	17 (13.2)	6 (12.5)	11 (14.5)	.53
Diabetes mellitusd	34 (26.4)	6 (12.5)	27 (35.5)	.005c
Smoking status				.004c
Never	56 (43.4)	18 (37.5)	36 (47.4)
Past	18 (14.0)	17 (35.4)e	8 (10.5)e
Current	8 (6.2)
Race
White	48 (37.2)
Black	4 (3.1)
Asian	14 (10.9)
Hispanic	51 (39.5)
Native American	1 (0.8)
Other nonwhite	6 (4.7)
Unknown	5 (3.9)

MGH, Massachusetts General Hospital,

^aValues are presented as mean ± SD or number (%) unless otherwise indicated, Percentages that do not add up to 100% denote missing or unknown data,

^bNonwhite group includes all categories except white and “unknown,”

^cSignificant at P < .05,

^dPercentage denotes development of disease following delivery,

^eNumbers include past and current smokers.

Pathologic Findings

Gross and histomorphometric examinations were performed on all 129 placentas in the study cohort, and the findings are listed in Table 2. Placental weight (mean ± SD, 532 ± 108 vs 540 ± 148 g; P = .88) and gestational age at delivery (39.4 ± 1.1 vs 39.4 ± 3.3 weeks; P = .53) were similar between the white and nonwhite populations. We observed a larger number of slightly immature placentas in the white population compared with the nonwhite population (91.7% vs 89.5%), but this difference did not achieve statistical significance in the overall maturation category. Chorangiosis was significantly higher in placentas from the white population compared with the nonwhite population (52.1% vs 30.3%, P = .02) Image 1. Excluding chorangiosis, the placental pathology parameters were not statistically significantly different between the white and nonwhite populations (Table 2).

Table 2.

Placental Pathology Characteristics in the Study Population With Gestational Diabetes Mellitus^a

Characteristic	All	White	Nonwhiteb	P Value
Placental weight, g	536 ±132	532 ±108	540 ± 148	.88
Placental weight  percentile			.95
<10	24 (18.6)	9 (18.8)	14 (18.4)
10-90	72 (55.8)	27 (56.3)	41 (54.0)
>90	33 (25.6)	12 (25.0)	21 (27.6)
Cord length, cm	36.1 ± 13.8	33.7 ±14.8	37.2 ± 13.2	.19
Cord insertion	5.0 ± 2.4	4.7 ± 2.5	5.2 ± 2.3	.25
Cord insertion location				.26
Membranous	2 (1.6)	0	2 (2.6)
Maturation				.92
Mature	9 (7.0)	3 (6.3)	6 (7.9)
Slightly immature	117 (90.7)	44 (91.7)	68 (89.5)
Immature	3 (2.3)	1 (2.1)	2 (2.6)
Chorangiosis	49 (38.0)	25 (52.1)	23 (30.3)	.02c
Chorangiosis—  focal diffuse	7 (5.4)	4 (8.3)	3 (4.0)	.30
Dysmorphic	16 (12.4)	5 (10.4)	11 (14.5)	.51
Maturation arrest	0	0	0	NA
Inflammation	51 (39.5)	19 (39.6)	30 (39.5)	.99
Acute chorioamnionitis	22 (17.1)	6 (12.5)	14 (18.4)	.38
Villitis of unclear/  unknown etiology	26 (20.2)	11 (22.9)	14 (18.4)	.54
Interface inflammation	4 (3.1)	2 (4.2)	2 (2.6)	.64
Deciduitis	2 (1.6)	2 (4.2)	0	.07
Fetal acute  chorioamnionitis	2 (1.6)	1 (2.1)	1 (1.3)	.74
Vasculitis	6 (4.7)	2 (4.2)	3 (4.0)	.74
Nucleated RBCs	6 (4.7)	3 (6.3)	3 (4.0)	.54
Acute villous edema	14 (10.9)	5 (10.4)	9 (11.8)	.81
Placental infarcts	20 (15.5)	8 (16.7)	12 (15.8)	.90
Intervillous thrombi	26 (20.2)	9 (18.8)	16 (21.1)	.76
Fetal thrombotic  vasculopathy	9 (7.0)	1 (2.1)	8 (10.5)	.07
Avascular villi	5 (3.9)	1 (2.1)	4 (5.3)	.38
Edema	1 (0.8)	1 (2.1)	0	.21
Meconium	39 (30.2)	11 (22.9)	27 (35.5)	.14
Extensive meconium	2 (1.6)	0	2 (2.6)	.26
Decidual vasculopathy	2 (1.6)	1 (2.1)	1 (1.3)	NA
Calcification	14 (10.9)	4 (8.3)	9 (11.8)	.53
Fibrin	11 (8.5)	2 (4.2)	8 (10.5)	.21
Atherosis	1 (0.8)	0	1 (1.3)	.42

NA, not applicable.

^aValues are presented as mean ± SD or number (%). Percentages that do not add up to 100% denote missing or unknown data.

^bNonwhite group includes all categories except white and “unknown.”

^cSignificant at P < .05.

Image 1.

Mature placentas at ×20. A, Chorangiosis. Multiple capillaries are in all villi with villous expansion (H&E). B, Normal nonchorangiotic placenta showing normal number of capillaries in villi (H&E).

Discussion

In our study, we analyzed the placental histomorphometry of GDM among women from white and nonwhite populations. To our knowledge, this is the first study that has examined racial/ethnic differences in the placental pathology of women with GDM with a consideration of the clinical implications during and subsequent to pregnancy. Clinically, we observed that white women had higher blood pressure, were more likely to be smokers, and were less likely to develop subsequent T2DM than the nonwhite women. From a pathology perspective, we observed that placental pathology was largely similar between the white and nonwhite populations. However, we did observe that white women with GDM had a greater percentage of chorangiosis than was evident among nonwhite women with GDM, a finding worthy of further consideration.

Chorangiosis refers to the presence of excess blood vessels in the placental villi. Ogino and Redline¹³ observed that chorangiosis is significantly associated with having enlarged placenta, immature villi, and maternal diabetes. In addition, they observed in the literature that chorangiosis is also associated with maternal anemia, smoking, twin gestations, and delivery at high altitudes. In our study, all women were healthy, and none were reported to have anemia or cardiovascular disease. In addition, all pregnancies in our study cohort were singleton and delivered at MGH, so ambient altitude was not a factor. However, a significantly greater number, 35.4%, of the white women identified as past or current smokers compared with 10.5% of the nonwhite population (P = .004). Therefore, it is possible that the relatively high proportion of past or current smokers in the white population contributed to the greater degree of chorangiosis in that population, although both populations’ insulin dependence during pregnancy should also be considered.

Several studies have considered glycemic control during pregnancy when investigating placental pathology associated with diabetes. Calderon et al¹⁷ found that abnormal glycemic levels could contribute to morphometric abnormalities observed in the GDM placenta. Gauster et al⁸ supported these findings, concluding that poorly controlled GDM could result in villous edema and increased fibrin in the placenta. While GDM has numerous implications for the vasculature and gross characteristics of the placenta, it has also been shown to contribute to impaired placental function, including reduced glucose metabolism and the ability to metabolize other substrates.^18,19

Alternatively, good glycemic control during pregnancy reduces adverse clinical pregnancy outcomes associated with GDM,²⁰ but normoglycemia does not mediate all the pathologic characteristics associated with GDM.^21-23 In our population, the only pathologic difference was that chorangiosis was significantly more prevalent among the white compared with the nonwhite women with GDM. Because we do not know the degree of insulin use or glycemic control of these women, we do not know if the greater incidence of chorangiosis among white women resulted from greater intrapartum insulin use, presumably leading to improved intrapartum glycemic control and consequently less T2DM development subsequent to pregnancy. Conversely, more intrapartum insulin use among nonwhite women with GDM would be consistent with data suggesting an increased rate of subsequent development of T2DM in this setting.²⁴ Nonetheless, we require the data regarding insulin use and glycemic control to clarify this relationship.

Accordingly, one limitation of our study is the lack of information on glucose control or insulin use during pregnancy. These data would facilitate the elucidation of the etiology of chorangiosis in our population. In addition, to elucidate the observed racial difference in the degree of chorangiosis, it would be beneficial to examine a population of white women without GDM to distinguish this association with chorangiosis as race mediated or GDM associated. Furthermore, we had limited racial/ethnic diversity within our nonwhite population. This could have limited the differences between the white and nonwhite populations.

In summary, we identified that chorangiosis was more prevalent among white compared with nonwhite women with GDM. Excluding chorangiosis, we found that the placental pathology of the white and nonwhite women did not differ significantly. This finding is clinically relevant for at least two reasons. First, the placentas of women with GDM may not undergo routine pathologic examination depending on institutional practice. However, our findings suggest that placental pathologic examination may demonstrate changes with implications for maternal postpartum cardiometabolic disease risk surveillance. Second, the identification of placental pathology may serve as an indicator of underlying maternal disease. For example, in our study, we identified a racial/ethnic difference in the presence of chorangiosis, which has been directly linked to clinical parameters such as maternal dysglycemia and anemia.¹³ Consequently, this may be correlated with the prenatal measurements of these parameters and potentially lead to the consideration of additional prenatal intervention in certain populations.

Further study is warranted to determine the cause of greater chorangiosis in our white population and examine whether this observation is related to smoking, insulin use, glycemic control, or specifically race. We also found that more nonwhite women developed T2DM after pregnancy than white women in our GDM population. However, further research is necessary to determine if the higher rate of subsequent T2DM in nonwhite populations is associated with differences in placental pathology. This may provide another opportunity for risk stratification in GDM populations to inform evidence-based interventions to alleviate racial/ethnic disparities in GDM and T2DM.