High Rate of Placental Infarcts in Type 2 Compared with Type 1 Diabetes

Abstract

Context:

Timing and cause of pregnancy loss differ between type 1 (T1DM) and type 2 diabetes mellitus (T2DM).

Objective:

The objective of the study was to determine whether placental histology corresponds to differing causes of pregnancy loss in T1DM and T2DM. We hypothesized that placentas from mothers with T2DM would be more likely to demonstrate vascular pathology than those from mothers with T1DM.

Research Design/Setting/Participants:

We reviewed medical histories, pregnancy outcomes, and placental histology of women with pregestational T1DM and T2DM with singleton pregnancies between 2001 and 2009 at a single tertiary care medical center.

Main Outcome Measures:

Placental weight, placental dysmaturity, villous maturation, villitis of unclear etiology, and histological evidence of placental infarction were measured.

Results:

Ninety-eight placentas were available for review, 53 from T1DM mothers (56%) and 45 from T2DM mothers (46%). Mean age and glycemic control each trimester did not differ between diabetes types. T2DM placentas had a higher prevalence of placental infarcts (22 vs. 6%, P = 0.02) and a lower prevalence of placental dysmaturity (12 vs. 29%, P = 0.05) compared with T1DM; rates differed from those reported in the general population. There was no difference in placental weight, villous maturity, or villitis of unclear etiology between diabetes types.

Conclusions:

There were many similarities in placental histological findings between diabetes types. Still, one in five T2DM placentas displayed histological infarcts, consistent with a vascular, rather than glycemic, etiology of pregnancy complications, whereas T1DM placentas showed signs of abnormal development.

Types 1 and type 2 diabetes mellitus (DM) increase the risk of pregnancy, with complications related to hyperglycemia as well as to comorbid conditions. As pregestational type 2 DM has become more common, there has been increasing recognition that risk may differ between diabetes types, irrespective of glycemia (1). Women with type 2 diabetes have similar or poorer pregnancy outcomes, including later pregnancy loss, and high risk of neonatal complications despite better glycemic control, on average, than is seen in type 1 diabetes (2–4). Possible reasons for increased risk in type 2 diabetes relate to older maternal age, higher prevalence of obesity and other metabolic syndrome characteristics, and lower socioeconomic status (5).

In addition to differences in associated medical conditions, placental structure differs between type 1 and type 2 diabetes. Factors other than glycemia, such as hypertension (6) and inflammation (7), affect placental development in pregnancies complicated by pregestational type 2 diabetes (8). The aim of this study was to compare the placental histology of type 1 and type 2 diabetes to provide insight into the differing causes of pregnancy outcomes between diabetes types. We hypothesized that placentas from mothers with type 2 diabetes would have a higher rate of placental histology suggestive of vascular pathology than placentas from mothers with type 1 diabetes.

Materials and Methods

This was a retrospective study performed at Massachusetts General Hospital, an academic medical center in Boston, MA. We searched the pathology database for all placentas delivered at Massachusetts General Hospital between 2001 and 2009 using the terms “diabetes,” “IDDM,” “DM”, “GDM,” and “glucose.” We retained the first placenta for each woman confirmed by chart review to have pregestational type 1 or type 2 diabetes and singleton pregnancies.

Standardized chart review was undertaken to determine diabetes type, maternal age and race (self-reported), weight at first visit and at delivery, mean hemoglobin A1c (HbA1c) during each trimester (mean of available HbA1c in each trimester if more than one was available, with date ranges for each trimester calculated from the gestational age at delivery), preeclampsia or eclampsia as noted in a physician's note in the chart, family history of diabetes, smoking, and hypertension or gestational hypertension. Diabetes type was determined by physician diagnosis, medication prescriptions, medical history, and antibody and C-peptide levels, if available. Pregnancy-induced hypertension was defined as systolic blood pressure of 140 mm Hg or greater and/or diastolic blood pressure of 90 mm Hg or greater for 2 consecutive weeks during pregnancy with normal blood pressure before gestation. Pregestational hypertension was defined as prescription of hypertension drugs before pregnancy and/or having hypertension listed in chart. Obstetric data recorded were gestational age, parity, mode of delivery, Apgar scores at 1 and 5 min, and fetal demise.

The blinded study pathologist re-reviewed the corresponding placenta slides to assure standard classification of placental histology. Parameters scored included placenta weight, villous maturation (9), villitis of unknown etiology [VUE; evidence of inflammatory pathology indicated by chronic villitis (10)], acute chorioamnionitis (11), and evidence of hypertensive damage/ischemia [small placenta by weight, infarcts (12)]. The placental pathological diagnoses were diagnosed by a trained placental pathologist. The placental weight was recorded trimmed of cord and membranes. Villous maturation is based on the presence of well-defined and easily identified vasculosyncytial membranes in a background of mostly tertiary/terminal villi. The diagnosis of VUE or chronic villitis is based on the presence of more than three villi affected by increased mononuclear cells within the villous stroma and often associated with intervillositis in more than three locations. Plasma cells were noted but not required for the diagnosis. Placenta size was based on published percentiles by gestational age (13). Infarcts were defined by the presence of necrotic villi, collapsed maternal space, and grossly identified. We pooled some findings using composite diagnoses. For example, hypertensive damage was the presence of three of the following pathologies: placentas less than the 10th percentile by weight, placental infarct, chronic abruption, large syncytial knots, accelerated villous maturation, decidual vasculopathy, or atherosis. Background population rates of placental characteristics were obtained from previously published reports.

We compared continuous variables using Student's t tests and categorical variables using the Fisher exact test for variables with few outcomes. We ran multivariable logistic regression models that were exploratory, given the limited number of outcome events. All analyses were performed using SAS software, version 9.2 (SAS Institute, Cary, NC). A P ≤ 0.05 was considered statistically significant. The study was approved by the Partners HealthCare Institutional Review Board.

Results

We identified 765 placentas, most from pregnancies complicated by gestational diabetes mellitus (GDM). After excluding GDM, retaining the first placenta for each subject and excluding one twin pregnancy and one with lost placenta slides, 98 subjects remained in the sample, 53 with type 1 (56%) and 45 with type 2 diabetes (46%).

There was no significant difference in age or glycemic control (HbA1c) between mothers with type 1 and type 2 diabetes. Other demographic and clinical characteristics are shown in Table 1. Hypertension was nearly twice as common in women with type 2 diabetes (33 vs. 17% in type 1 DM, P = 0.056), and women with type 2 diabetes had higher weight at first visit and before delivery than women with type 1 diabetes. Preeclampsia and eclampsia rates were not statistically significantly lower in type 2 than in type 1 diabetes (11 vs. 25%, P = 0.09). Obstetric data, including gestational age at delivery, delivery by cesarean section, birth weight, Apgar scores, and rate of fetal demise did not differ between diabetes types, but women with type 2 diabetes had higher parity.

Table 1.

Baseline and pregnancy characteristics in women with pregestational type 1 and type 2 diabetes

	Type 1 (n = 53)	Type 2 (n = 45)	P value
Demographic data
Age at delivery (yr), mean (sd)	31 (5.7)	32 (6.3)	0.37
Race/ethnicity
Black, n (%)	3 (6)	8 (18)	<0.0001a
White, n (%)	41 (77)	13 (29)
Hispanic, n (%)	7 (13)	20 (44)
Other, n (%)	2 (4)	4 (9)
Clinical data
Weight at first visit (lb), mean (sd)	160.0 (32.5)	200.1 (43.3)	<0.0001
Weight prior to delivery (lb), mean (sd)	190.2 (34.1)	218.7 (46.1)	0.002
First-trimester HbA1c (%), mean (sd)	7.4 (1.4)	7.5 (1.6)	0.9
Second-trimester HbA1c (%), mean (sd)	6.4 (1.0)	6.3 (1.0)	0.6
Third-trimester HbA1c (%), mean (sd)	6.5 (1.0)	6.4 (0.9)	0.6
Hypertension, n (%)	9 (17)	15 (33)	0.056
Preeclampsia or eclampsia, n (%)	13 (25)	5 (11)	0.09
Family history of diabetes mellitus, n (%)	31 (58)	28 (62)	0.5
Smoking, n (%)	2 (4)	5 (11)	0.2
Obstetric data
Gestational age at delivery (wk), mean (sd)	37.0 (3.74)	37.5 (3.8)	0.6
Cesarean deliveries, n (%)	35 (67)	24 (53)	0.2
Birth weight (g), mean (sd)	3,309 (962)	3,372 (842)	0.7
One-minute Apgar score, mean (sd)	7.6 (2.2)	7.7 (1.9)	0.8
Five-minute Apgar score, mean (sd)	8.3 (2.1)	8.7 (0.6)	0.2
Parity, mean (sd)	1.9 (1.2)	2.8 (1.9)	0.004
Fetal demise	2 (4)	2 (4)	0.9

^aP value for χ² test across all race/ethnicity groups.

Some, but not all, features of placental histology differed between diabetes types and appeared to differ from published population rates of placental pathology (Table 2). Placental infarcts were significantly more common in type 2 diabetes [odds ratio (OR) 4.7, 95% confidence interval (CI) 1.22–18.6] than in type 1 diabetes and appeared to be twice as high as the rate seen in the general population. This relationship was attenuated after adjusting for hypertension (OR for type 2 DM relative to type 1 DM: 3.4, 95% CI 0.8–13.9; OR for hypertension: 4.4, 95% CI 1.2–16.2, P = 0.03, data not shown). Villitis of unclear etiology appeared to be more prevalent in type 2 DM (20%) than in type 1 DM (13%) placentas, but this was not statistically significant.

Table 2.

Placental weight and histology comparing type 1 with type 2 diabetes and background population

	Type 1 (n = 53)	Type 2 (n = 45)	P value	Population prevalence
Placental weight (g), mean (sd)	468 (161)	471 (143)	0.9	626 (133)
Villous maturity			0.3
Hypermature, n (%)	6 (11)	6 (13)
Immature, n (%)	14 (26)	5 (11)
Mature, n (%)	32 (60)	33 (73)
Other, n (%)	1 (2)	1 (2)
VUE, n (%)	7 (13.2)	9 (20.0)	0.4	5–15%b
Placental dysmaturity, n (%)	15 (29)	5 (12)	0.05	7.7%c
Infarction, n (%)	3 (5.7)	10 (22.2)	0.02	12%d
Maternal choriamnionitis, n (%)	7 (13)	6 (13)	0.9	1.5%c

^aSee reference 19. Based on Norwegian 50th percentile for placental weight at 37 wk gestation, the mean gestational age in this study.

^bSee Ref. 10.

^cSee Ref. 15.

^dSee Ref. 20.

In contrast, placental dysmaturity (29 vs. 12%, P = 0.05) was more common in type 1 compared with type 2 diabetes and with published population rates (Table 2). Villous immaturity also appeared more common in type 1 DM (26%) than type 2 DM (5%) placentas, but this did not reach statistical significance. In underpowered exploratory models adjusting for hypertension and glycemia, the increased rate of villous immaturity in type 1 DM became significant (P = 0.04). Placental weight and maternal choriamnionitis did not differ by diabetes type.

Discussion

Although there were many similar placental characteristics in type 1 and type 2 diabetes, it is notable that despite equivalent glycemic control, gestational age at delivery, and cesarean section rates, we observed significant differences in placental pathology. Consistent with previous findings (14, 15), type 1 diabetes placentas had a higher prevalence of placental dysmaturity, indicating problems in placental development. In contrast, type 2 diabetes placentas had a 4-fold higher rate of placental infarction than type 1 diabetes, with infarcts found in one of five placentas from mothers with type 2 diabetes (and in one of the two placentas associated with fetal demise); this relationship was attenuated after controlling for hypertension. Other differences, such as those in villitis of unclear etiology and villous maturity, did not reach statistical significance. Placental weight did not differ between diabetes types but did appear to be lower than population reference ranges in both diabetes groups.

These histological differences suggest that differing pathophysiology associated with type 1 or type 2 diabetes affects placentation despite similar medical care and glycemic characteristics; other nonsignificant differences, in this small sample, favor this interpretation as well. The high rates of placental dysmaturity in type 1 diabetes and placental infarcts in type 2 diabetes provide histological correlates of the prior observation that the cause and timing of pregnancy loss differ between diabetes types. Due to the retrospective study design, this study failed to capture first-trimester pregnancy losses and placental tissue. Consequently, our findings are derived from surviving pregnancies that presumably had less severe abnormalities. Extreme hyperglycemia in the first trimester, which is more common in type 1 than in type 2 diabetes, may have led to increased first-trimester losses in that group, yielding a somewhat healthier subset of type 1 diabetes pregnancies in this study. Prior studies have shown that congenital anomaly and first-trimester losses are more common in type 1 and stillbirth more common in type 2 diabetes (3).

Diabetes is associated with a 2.5 increased odds of stillbirth (16). Even well-controlled diabetes (with periconception HbA1c < 6.9%) is associated with perinatal mortality of 2.1% compared with 0.75% in an unaffected population (17). Similarly, it appears that many of the rates of pathological characteristics of type 1 and type 2 diabetes pregnancies differ from those in background populations.

There are few prior reports comparing placental histology in type 1 with type 2 diabetes. Higgins compared placental structure by stereology in 10 nondiabetic, eight type 2 DM, and 10 type 1 DM women and found an association between maternal diabetes, glycemia, and terminal villous volume and type 1 DM and increased capillary length (variables we were not able to assess in this study) (8). In that study, there was no association between glycemia and stromal development, suggesting that nonglycemic factors also influence placental development. In a Kuwaiti study of placental pathology in association with White's classification of diabetes, placental infarcts did not differ by White class, but dysmaturity appeared to be more prevalent in White classes C/D (mostly type 1 DM, n = 7) compared with normal controls and white classes A (GDM) or B (mostly type 2 DM) (15).

This present report extends these prior studies by including a larger number of subjects with both diabetes types and documenting a 4-fold rate of placental infarcts in type 2 compared with type 1 diabetes. An earlier, larger study that searched for placental vascular disease in association with pregestational diabetes did not find a higher risk of infarction between all pregestational diabetes pregnancies compared with normoglycemic pregnancies, but this study used International Classification of Diseases, ninth revision, codes rather than histological review, which, as the authors acknowledged, was likely insensitive for this finding (18).

Although our single-center findings are additionally limited by the retrospective study design, failure to capture early pregnancy losses, and the small sample size, the detailed review provides histological correlates for the differing causes and timing of pregnancy loss in pregestational type 1 and type 2 diabetes. In addition, these findings support the hypothesis that the complications associated with type 2 diabetes in pregnancy are related to vascular pathology in addition to hyperglycemia, analogous to the relationship between type 2 diabetes and risk of cardiovascular disease later in life. Future studies of placental pathology should follow patients with diabetes matched to controls without diabetes prospectively from the beginning of pregnancy to capture all pregnancy losses and events and to allow comparison with normal placental pathology to determine whether these findings can elucidate the etiology of pregnancy complications in pregestational diabetes.