A comprehensive analysis of the association between placental pathology and recurrent preterm birth

Abstract

BACKGROUND:

Histologic examination of the placenta is often performed after preterm birth. Although placental examination cannot change the index pregnancy outcome, it may inform the risk of adverse outcomes in a subsequent pregnancy. Previous research has examined the association between individual histologic lesions and pregnancy outcomes without consistent results.

OBJECTIVE:

This study aimed to determine the independent contributions of the major placental pathology histologic types to recurrent preterm birth.

STUDY DESIGN:

This was a retrospective cohort study of deliveries at a tertiary care center from January 2009 to March 2018. Individuals with ≥2 births, an index birth of <37 weeks of gestation, and a placental pathology report from the index pregnancy were included. The presence of maternal vascular malperfusion, fetal vascular malperfusion, acute inflammation, and chronic inflammation was extracted from the pathology reports for each index placenta and classified as none, low grade, or high grade. A log-binomial model incorporating all 4 placental pathology histologic types, index gestational age, race, and maternal age was used to estimate the associations between each placental histologic type and risk of recurrent preterm birth. Moreover, 2-way interaction terms were studied among placental histologic types. In addition, 2 stratified analyses were completed on the basis of characteristics of the index preterm birth: (1) by late preterm (gestational age of 34–36 weeks) vs early-to-moderate preterm birth (<34 weeks) and (2) a subgroup analysis of those with spontaneous preterm birth.

RESULTS:

A total of 924 pregnancy pairs met the inclusion criteria. Only high-grade chronic inflammation was independently associated with an increased risk of recurrent preterm birth (adjusted risk ratio, 1.37; 95% confidence interval, 1.03–1.81). Stratified analysis by gestational age group demonstrated maternal vascular malperfusion was associated with recurrent preterm birth only among those with early preterm birth (adjusted risk ratio, 1.40; 95% confidence interval, 1.01–1.93). Among participants with spontaneous preterm labor, no association was found between pathology histologic types and risk of preterm birth.

CONCLUSION:

Among index preterm pregnancies, high-grade chronic placental inflammation was associated with recurrent preterm birth. Low-grade maternal vascular malperfusion was associated with recurrent preterm birth only among those with an early or moderate index preterm birth (<34 weeks of gestation). These findings may be useful in determining the risk profile for individual patients and may generate hypotheses as to the pathogenesis of recurrent preterm birth.

Introduction

Preterm birth (PTB) affects 1 in 10 births in the United States and is associated with neonatal morbidity and mortality.¹ The recurrence risk of PTB is estimated to be 30%, compared with a baseline rate of 10% among those with an index term birth.^2–4 This risk may differ based on clinical characteristics, such as gestational age (GA) at index delivery, and type of preterm birth (indicated vs spontaneous).^2–4 Patients who experience PTB in an index pregnancy are counseled on the risk of recurrence; however, few factors independent to a patient’s delivery aid in this counseling.

Histologic examination of the placenta is often completed after PTB and can give insight into the under lying pathophysiology. Contemporary placental pathology using the Amsterdam consensus identifies 4 major patterns of injury, which will be referred to as “histologic types” for this study: acute inflammation (AI), chronic inflammation (CI), fetal vascular malperfusion (FVM), and maternal vascular malperfusion (MVM).⁵ Each histologic type represents a constellation of pathologic lesions, which have been determined to reflect similar pathophysiology. The use of placental pathology histologic types vs individual lesions may be useful in more holistically identifying underlying mechanisms leading to outcomes. Our previous work developed grading systems for CI and MVM, which complements the existing grading systems for AI and FVM.⁶ This comprehensive, systematic grading system and the identification of placental pathology provide more granular information as to the contribution of each histologic type of placental pathology to outcomes of clinical relevance.⁶

Previous studies examining placental pathology in recurrent PTB are conflicting. Moreover, 2 studies assessed placentas from patients experiencing a PTB and examined the association of findings with history of previous PTB.^7,8 In 1 study, there was no association, and in the other, chronic choriodeciduitis, a marker of CI, was associated with previous PTB. Studies looking at index pregnancy pathology and risk in a subsequent pregnancy identify acute inflammatory lesions as markers for risk of recurrent PTB.^9,10 These studies examine individual lesions and the association with PTB. However, placental pathology histologic types frequently cooccur and therefore are important to study in the context of each other.¹¹ The independent contribution of each graded histologic type to the recurrence of PTB is not well studied.

Importantly, varying clinical outcomes may result from the same underlying placental pathology seen at the time of delivery. For example, MVM, often associated with hypertensive disorders of pregnancy, was observed in 61.8% of cases of spontaneous PTB.¹² A previous study examining the linkage of placenta pathology with the “phenotype” of PTB identified features of CI, AI, and MVM in multiple different presentations resulting in preterm deliveries, including preterm premature rupture of membranes, abruption, and preeclampsia.¹³

Placental pathology histologic types may reflect an underlying mechanism that relates to the risk of recurrent PTB in a more precise manner than clinical data. We hypothesized that certain histologic placental pathology histologic types are independently associated with recurrent PTB. In addition, we hypothesized that the impact of these placental pathology histologic types may differ based on the qualities of the index PTB, including the type of PTB (spontaneous vs indicated) and timing of PTB (late preterm vs early or moderate PTB).

Materials and Methods

This was a retrospective cohort study of deliveries at Northwestern Memorial Hospital from January 2009 to March 2018. Individuals were included if they experienced an “index” PTB (GA of <37 weeks) for which a placental pathology report was available, followed by a second birth during the study period at the same institution. If an intervening birth took place at an outside hospital, this was not included in the analysis. For patients with >2 deliveries during the study period, only the first 2 consecutive deliveries were included. The study was approved by the Northwestern University Institutional Review Board (STU00209704). Birth outcomes and demographic characteristics were obtained through the electronic data warehouse. Labor onsets were classified as spontaneous or indicated using an automated text search review of medical record data. Specifically, those with “spontaneous” listed in the labor description, “augmentation,” or “active management of labor” and those with rupture of membranes before admission were considered to be spontaneous, whereas those with notes, such as “induction” or “scheduled cesarean delivery,” were considered to be indicated.

During the study period, pathologic examination of placentas was carried out by trained perinatal pathologists. Standardized institutional indications for the examination of placentas by pathology included abnormal placenta and/or cord on gross delivery room examination, intrauterine fetal demise, fetal growth restriction, preterm birth at ≤34 weeks of gestation, severe preeclampsia, clinical concern for maternal infection, clinical suspicion of chorioamnionitis or abruption, history of current or previous gestational trophoblastic disease, compromised clinical condition of the neonate, multiple congenital anomalies with unknown etiology, and/or maternal cancer (active or recent). All placentas were grossed by pathology residents or certified pathologist assistants. Standardized gross description and tissue sampling protocol were used throughout the study period. Standardized diagnostic terminology was used by the pathologists, as appropriate for the period. For the study, placental pathology data points were extracted from pathology reports using an automated text-based search using R (version 4.0).¹⁴ Diagnostic terms used during the study period were relatively consistent, and the text searching algorithms we employed captured all related diagnostic terminology. The presence of MVM, FVM, maternal and/or fetal AI, and CI (based on the definitions provided in the Amsterdam consensus) was extracted from the pathology report for each index placenta. Detailed information on text-based searching protocols and placental lesion definitions and classification is reported elsewhere and summarized in Table 1, which describes lesions contributing to each placental pathology histologic type.⁶ Chronic decidual perivasculitis was included under CI for our analysis given that this is defined as lymphocytes within the wall of decidual vessels.

TABLE 1.

Placental lesions and scoring system

Grading system	Comprising lesions
MVM
Lesions added together on a point system High-grade MVM: score of ≥4 Low-grade MVM: core of 2–3	+1 point: Increased syncytial knots Fibrinoid necrosis or acute atherosis Muscularization of basal plate arterioles Mural hypertrophy of membrane arterioles Basal decidual vascular thrombus Villous agglutination Increased perivillous fibrin deposition Distal villous hypoplasia	+2 points: Multiple Infarcts Retroplacental blood or hematoma with hemosiderin or infarct Small-for-gestational-age placenta (if other MVM indicators present)
FVM
High-grade FVM: avascular villi or VSVK involving >45 villi and/or thrombus in >1 large fetal vessel or >2 stem villous vessels Low-grade FVM: presence of lesions, not high grade	Fetal vascular thrombi or intramural fibrin deposition (umbilical, chorionic, velamentous, stem villous) Avascular villi VSVK
CI
High-grade CI: involvement of > 1 compartment Low-grade CI: involvement of 1 compartment	Lymphocytes, histiocytes, and plasma cells in the following compartments: Chorionic plate or membranes: chronic chorionitis, chronic marginating choriodeciduitis, chronic chorioamnionitis Basal plate: chronic deciduitis with plasma cells, chronic decidual perivasculitis Villous stroma: chronic villitis, chronic basal villitis, intravillous plasma cells Intervillous space: chronic intervillositis Fetal vasculature: eosinophilic T-cell vasculitis, other chronic fetal inflammation not otherwise specified
AI
High-grade AI: presence of acute chorioamnionitis and/or acute umbilical arteritis Low-grade AI: presence of lesions, not high grade	Acute subchorionitis Acute chorionitis Acute chorioamnionitis Acute chorionic vasculitis Acute umbilical phlebitis Acute umbilical arteritis

AI, acute inflammation; CI, chronic inflammation; FVM, fetal vascular malperfusion; MVM, maternal vascular malperfusion; VSVK, villous stromal vascular karyorrhexis.

For our primary analysis, we analyzed the association of placental pathology histologic type in an index PTB and risk of recurrent PTB. A log-binomial model (binomial generalized linear model with a log link function) incorporating all 4 placental pathology histologic types was used to estimate the associations (risk ratios [RRs]) between each histologic type and risk of recurrent PTB while controlling for the other placental pathology histologic types and characteristics associated with recurrent PTB, including race and ethnicity, maternal age, and GA at index pregnancy.^15,16 Moreover, initial models included 2-way interactions among the pathology histologic types to evaluate whether the impact of 1 histologic type is modified by the presence of another histologic type. Backward elimination was used to retain statistically significant interaction terms (P < .05). As not all placentas were submitted for examination by pathology, as a sensitivity analysis, we used stabilized inverse probability weighting to account for selection bias among placentas submitted for pathologic review. To calculate the weights, we used the full sample of patients with an index PTB and a subsequent delivery during the study period and modeled the probability of being selected into the pathology sample based on maternal and pregnancy characteristics. Weighting the sample by the inverse probability of selection ensures that the distributions of maternal and pregnancy characteristics in the pathology (included) sample are similar to that of the sample that would have been obtained if all index placentas were submitted for pathologic review.

Placental pathology histologic types may influence each other in ways that are not captured by the primary modeling strategy; including all 4 histologic types in 1 model may result in overadjustment. As a secondary analysis, we modeled each placental histologic type individually to assess whether associations were consistent. Moreover, this analysis facilitated a comparison with other studies that may focus on 1 placental histologic type.

We performed 2 stratified analyses. In the first analysis, we stratified by early or moderate PTB (<34 weeks of gestation) vs late PTB (34–36 weeks of gestation) to determine whether associations between placental pathology and recurrent PTB differed based on the timing of the index delivery. In the second analysis, we stratified by type of index preterm birth (spontaneous vs indicated) to assess whether associations between placental pathology histologic types and recurrent PTB differed based on index PTB type. For this analysis, those with insufficient data to identify the type of PTB were excluded.

An additional exploratory analysis investigated the recurrence of placental pathology in those with placental pathology reports for the subsequent pregnancy. A second exploratory analysis examined each lesion composed of CI and MVM to identify associations with recurrent PTB.

Multiple imputation with 10 imputed datasets was used to account for missing covariate data (race and ethnicity, n=49 missing) and estimates were combined using the Rubin rule.^17,18 Analyses were performed using SAS (version 9.4; SAS Institute, Cary, NC), and a P value of .05 was used to determine statistical significance.

Results

Baseline characteristics

Between January 2009 and March 2018, there were 109,773 deliveries to 84,000 patients, with 21,973 patients having ≥2 deliveries during the period (Figure 1). Of patients with multiple deliveries, 1655 (7.5%) delivered preterm in the index pregnancy. The placenta was submitted for pathologic review for 947 patients (57.2%) of index PTBs. Subsequent pregnancy outcome (GA) was available for 924 patients (97.6%). Among all patients with an index PTB and a subsequent delivery (regardless of pathologic review, n=1655), those included in the sample (ie, patients whose placentas were submitted for pathologic review) were more likely to be nulliparous, have gestational diabetes mellitus and hypertensive disorders of pregnancy, deliver earlier, have small-for-gestational-age neonates, and deliver via cesarean delivery for the index delivery (Supplemental Table 1).

FIGURE 1. Flow diagram of inclusion.

GA, gestational age.

The risk of recurrent PTB was 27% in our sample. Using PTB classification as defined above, 609 patients (66%) were identified as having a spontaneous PTB, and 210 patients (23%) had indicated PTB. Moreover, 105 patients (11%) were not able to be classified. Those with a subsequent preterm birth (recurrent PTB, n=250) were less likely to be White and nulliparous and more likely to have had an index spontaneous PTB than those with a subsequent term delivery (Table 2).

TABLE 2.

Maternal and pregnancy characteristics for the full sample stratified by subsequent birth outcome

Variable	Index PTB (n=924)	Subsequent term birth (n=674)	Recurrent PTB (n=250)	P valuea
Age, index delivery	30.6 (5.0)	30.8 (4.9)	30.1 (5.3)	.09
Tobacco use	10 (1.1)	8 (1.2)	2 (0.8)	.99
Race				<.01
White	549 (59.4)	416 (61.7)	133 (52.2)
Black	115 (12.5)	67 (9.9)	48 (19.2)
Hispanic	149 (16.1)	109 (16.2)	40 (16.0)
Asian	65 (7.0)	50 (7.4)	15 (6.0)
Other	46 (5.0)	32 (4.8)	14 (5.6)
GA, index delivery	33.5 (3.2)	33.9 (3.0)	32.7 (3.6)	<.01
SGA, index delivery	75 (8.3)	58 (8.7)	17 (7.0)	.41
Mode of index delivery				.18
Cesarean	312 (34.1)	236 (35.4)	76 (30.6)
Vaginal	603 (65.9)	431 (64.6)	172 (69.4)
Nulliparous, index delivery	785 (85.6)	587 (88.0)	198 (79.2)	<.01
Interpregnancy interval (y)	1.8 (1.2)	1.8 (1.1)	1.9 (1.3)	.15
Type of PTB, index delivery				<.01
Spontaneous	609 (74.4)	430 (71.8)	179 (81.4)
Indicated	210 (25.6)	169 (28.2)	41 (18.6)
Unknown	105	75	30
Maternal medical condition, index delivery
Pregestational diabetes mellitus	20 (2.2)	11 (1.6)	9 (3.6)	.07
Gestational diabetes mellitus	57 (6.2)	34 (5.0)	23 (9.2)	.02
Hypertensive disorder of pregnancyb	174 (18.8)	133 (19.7)	41 (16.4)	.25
Clinical chorioamnionitis	8 (0.9)	8 (1.2)	0 (0.0)	.08
Placenta previa	43 (4.7)	32 (4.8)	11 (4.4)	.82
Clinical abruption	29 (3.1)	23 (3.4)	6 (2.4)	.43
Cervical insufficiency	14 (1.5)	8 (1.2)	6 (2.4)	.18

Data are presented as number (percentage), unless otherwise indicated.

GA, gestational age; PTB, preterm birth; SGA, small for gestational age.

^aP values of those with a subsequent term birth (n=674) were compared with those with a subsequent PTB (n=250) and were calculated using chi-square tests for categorical variables and t tests for continuous variables;

^bIncludes chronic hypertension, gestational hypertension, preeclampsia, eclampsia, and chronic hypertension with superimposed preeclampsia.

Placental histologic types and recurrent preterm birth

In the primary analysis investigating associations between placental pathology histologic types and recurrent PTB, only high-grade CI was associated with an increased risk of recurrent PTB (adjusted RR [aRR], 1.37; 95% confidence interval [CI], 1.03–1.81) (Figure 2). None of the other placental pathology histologic types were significantly associated with increased risk of recurrent PTB, although high-grade AI was associated with decreased risk of recurrent PTB (aRR, 0.70; 95% CI, 0.49–0.99) (Figure 2). Moreover, we tested all 2-way interaction terms among placental pathology histologic types, although none were statistically significant (P>.05) and all were removed from the model. Results were largely consistent in a sensitivity analysis, where inverse probability weighting was used to account for selection bias among those included in the pathology sample, except for high-grade AI, which was no longer statistically significantly associated with recurrent PTB (differences described in Supplemental Table 1 and weighted results are shown in Supplemental Table 2). Furthermore, results were consistent when each placental pathology histologic type was modeled separately (Supplemental Table 3). We found no association between chronic villitis and basal villitis with recurrent PTB, but we did find an association between chronic deciduitis with plasma cells and recurrent PTB (aRR, 1.31; 95% CI, 1.05–1.62; P=.02) (Supplemental Table 4).

FIGURE 2. Placental histology and risk of recurrent preterm birth.

Risk ratio adjusted for presence of remaining placental lesions, race, gestational age at index pregnancy, and maternal age.

AI, acute inflammation; aRR, adjusted risk ratio; CI, chronic inflammation; FVM, fetal vascular malperfusion; MVM, maternal vascular malperfusion.

We examined the test performance of high-grade CI as a predictor of recurrent PTB. Our specificity was 88%, sensitivity was 16.4%, positive predictive value was 33.3%, and negative predictive value was 73.9% using the presence of high-grade CI as test “positive” and absence or low grade as test “negative.”

Placental histologic types and recurrent preterm birth by delivery timing

High-grade AI and high-grade MVM were more common in early or moderate PTB (<34 weeks of gestation) compared with late PTB (34–36 weeks of gestation) (Supplemental Table 5). In models stratified by timing of the index PTB, the association between MVM and recurrent PTB differed based on the timing of the index PTB (Table 3). Among those with an early or moderate PTB (<34 weeks of gestation), low-grade MVM was associated with recurrent PTB (aRR, 1.40; 95% CI, 1.01–1.93). High-grade MVM was not associated with recurrent PTB (aRR, 0.87; 95% CI, 0.63–1.21). Among those with a late PTB (34–36 weeks of gestation), there was no association between MVM and recurrent PTB (aRR of low-grade MVM, 0.84 [95% CI, 0.55–1.26]; aRR of high-grade MVM, 0.79 [95% CI, 0.52–1.19]). Point estimates for the association between high-grade CI and recurrent PTB were consistent across the timing of the index PTB and similar to the primary results; however, CIs were wide because of sample size. In our exploratory analysis investigating individual MVM lesions, including lesions representing histologic evidence of abruption, none were associated with recurrent PTB.

TABLE 3.

Placental pathology histologic types and recurrent preterm birth, stratified by timing of delivery

	Early preterm (n=376)				Late preterm (n=548)
	Unadjusted		Adjusteda		Unadjusted		Adjusteda
Variable	RR	95% CI	RR	95% CI	RR	95% CI	RR	95% CI
Acute inflammation
None	Reference		Reference		Reference		Reference
Low	0.80	0.54–1.18	0.74	0.50–1.08	0.89	0.60–1.32	0.98	0.66–1.45
High	0.88	0.59–1.32	0.77	0.52–1.14	0.71	0.34–1.50	0.76	0.36–1.60
Chronic inflammation
None	Reference		Reference		Reference		Reference
Low	0.96	0.67–1.40	0.99	0.70–1.41	0.97	0.65–1.44	0.85	0.57–1.26
High	1.42	0.98–2.06	1.33	0.91–1.94	1.37	0.91–2.08	1.26	0.84–1.90
Fetal vascular malperfusion
None	Reference		Reference		Reference		Reference
Low	0.97	0.67–1.40	0.94	0.66–1.33	0.73	0.49–1.10	0.76	0.51–1.14
High	0.56	0.16–1.98	0.62	0.20–1.89	0.29	0.04–1.95	0.38	0.06–2.50
Maternal vascular malperfusion
None	Reference		Reference		Reference		Reference
Low	1.42	1.01–1.99	1.40	1.01–1.93	0.91	0.60–1.37	0.84	0.55–1.26
High	0.86	0.60–1.24	0.87	0.63–1.21	0.90	0.61–1.34	0.79	0.52–1.19

CI, confidence interval; RR, risk ratio.

^aModels include all 4 placental pathology histologic types, race and ethnicity, and maternal age.

Placental histologic types and recurrent preterm birth by spontaneous vs indicated preterm birth

The only placental pathology histologic type that differed in prevalence between spontaneous and indicated births was MVM, which was present in 42.5% of spontaneous index deliveries and 63.3% of indicated index deliveries (P<.01) (Supplemental Table 6). We had a limited sample of indicated deliveries and therefore only performed a subgroup analysis examining the association of pathologic histologic types and recurrent PTB among those with spontaneous PTB. Although our point estimate was similar for the association of high-grade CI and recurrent PTB, it was not statistically significant (Supplemental Table 7).

Recurrent pathologic findings among those with index preterm birth

Of 924 patients, 304 (32.9%) had placentas from the subsequent pregnancy submitted for pathologic review. Of these patients, 175 (57.6%) had a recurrent PTB, and 129 (42.4%) had a subsequent term delivery. CI was the only pathology with a significant recurrence risk, identified in the subsequent delivery for 68.9% of those with CI in the index delivery (RR, 1.99 [95% CI, 1.58–2.51] with referent as those without CI in index delivery who developed CI in subsequent pregnancy) (Table 4). MVM showed a similar trend, with recurrence in 61.5% (RR, 1.22; 95% CI, 0.99–1.50).

TABLE 4.

Recurrence of placental pathology in subsequent pregnancy among those with subsequent placenta submitted for histologic review (n=304)

Variable	Recurrence risk	Recurrence RRa
Acute inflammation	42.9%	1.30 (0.96–1.77)
Chronic inflammation	68.9%	1.99 (1.58–2.51)
Fetal vascular malperfusion	39.0%	1.02 (0.73–1.41)
Maternal vascular malperfusion	61.5%	1.22 (0.99–1.50)

RR, risk ratio.

^aRR compared with risk among those without placental pathology lesion in index pregnancy.

Comment

Principal findings

Placental pathology with high-grade CI in the index PTB was associated with an increased risk of recurrent PTB, whereas low-grade MVM was associated with an increased risk of recurrent PTB only among those patients who delivered at <34 weeks of gestation in an index pregnancy. We did not find a significant interaction among the placental pathology histologic types. In an analysis restricted to those with an index spontaneous PTB, no placental pathology histology type was statistically significantly associated with increased risk of recurrent PTB, although these analyses were limited by sample size.

Results in the context of what is known

Our findings suggested that acute chorioamnionitis is a response to nonrecurring factors that may not increase the a priori risk of subsequent PTB, whereas chronic placental inflammation may reflect immune-mediated processes that might predispose to the risk of recurrent PTB. AI is correlated with acute chorioamnionitis in the clinical setting, and predominantly, neutrophils are noted in this pathology.¹⁹ This may be due to infection or sterile inflammation secondary to labor processes. In contrast, CI is represented by lymphocytic and histiocytic processes that may be secondary to an intolerance of the maternal immune system to the fetal antigen or potentially a response to a viral or other chronic infection causing a maternal immune response.^19,20 CI has been implicated particularly in late PTB.²⁰ We found that the recurrence of CI was high, at 70% among those with index PTB, supporting our hypothesis that this may represent underlying pathology that predisposes a patient to PTB. Previous reports have demonstrated the role of antifetal alloantibodies in the pathogenesis of chronic intervillositis, which may be an example of a mechanism underlying recurrence.²¹

Our findings have added to the ambiguity from previous studies, in which acute inflammatory lesions are sometimes identified as markers of recurrent PTB and, in other studies, do not mark any predisposition to recurrence.^9,10,22 In our weighted analyses to account for differences among patients with placentas sent to pathology, AI was not associated with either decreased or increased risk of recurrent PTB. We hypothesized that this discrepancy may be due to the inclusion of factors, such as GA and race in our model, although it may also be secondary to a selection bias in our pathology-based sample.

The association between MVM and recurrent PTB was dependent on the GA at index delivery, with increased risk of recurrent PTB only observed for those with low-grade MVM who delivered in the index pregnancy at <34 weeks of gestation. MVM is thought to represent abnormal remodeling of the spiral arteries and resultant abnormal oxygenation and flow in the intervillous space, and an early development of MVM may represent an underlying predisposition to abnormal placental development.²³ A previous study did not find an association between MVM and recurrent PTB among those with spontaneous PTB.¹² Further research is needed to elucidate the significance of MVM in predicting recurrent PTB. In the sample of those with 2 placentas submitted for pathologic review, MVM trended toward an elevated recurrence risk (P=.051). Our prevalence of recurrent pathologies was highly biased, based on sample selection of patients with 2 placentas sent to pathology.

Clinical and research implications

Strategies to prevent spontaneous recurrent PTB include 17-alphahydroxyprogesterone caproate and cervical cerclage.²⁴ Options to prevent indicated PTB are limited, with aspirin recommended in the setting of previous preeclampsia.²⁵ The prevention of recurrent PTB is primarily based on a clinical scenario in an index pregnancy. For example, progesterone-based treatments are used in index spontaneous PTB but are not offered with previously indicated PTB. Previous studies have revealed that the same placental pathology, such as MVM, may result in different clinical presentations, such as PTB, preeclampsia, or placental abruption.¹³ In this manner, the placental pathology as an objective data element may offer additional insights into determining the appropriate treatment.

The most appropriate treatment to prevent recurrent spontaneous PTB is not entirely clear.²⁶ For example, a recent study has called into question the purported efficacy of intramuscular progesterone.²⁷ It is not known whether certain subcategories of patients with a history of PTB might benefit from supplemental progesterone. Of note, 1 mechanism of action proposed for intramuscular progesterone is through its anti-inflammatory properties.^28–31 Our identification of placental CI as being correlated with increased risk of recurrent PTB might be investigated for use to target therapy for the prevention in a subsequent pregnancy.

Finally, previous authors have questioned the use of placental pathologic examination.³² Here, we described manners in which placental pathology can be used to change risk assessment and, in the future, potentially further personalize clinical care. Importantly, the negative predictive values of high-grade CI and specificity were high at 73.9% and 88%, respectively, indicating the use of placental pathology as a potential “rule-out” test.

Strengths and limitations

The strengths of this study included the large number of participants and placental pathology reports performed by trained perinatal pathologists. The limitations included its retrospective nature and the use of automatically extracted electronic medical record information to separate indicated and spontaneous PTB, which may result in some inaccurate classifications. Clinical characteristics were derived from electronic medical records, which may have limitations for data points, such as smoking and clinical chorioamnionitis, which both seem to be underreported based on the data points available to us. Placentas from index PTBs are often sent for pathologic review. However, there is still some bias in the portion that is sent for pathology. We attempted to account for this with inverse probability weighting using demographic and clinical characteristics, although selection bias may still impact our results.

Regarding placental lesion classification, the Amsterdam consensus classification forms the basis of most of our analysis except for chronic decidual perivasculitis, which is classified as CI as opposed to MVM. Chronic decidual perivasculitis, defined by the presence of lymphocytes within the wall of decidual vessels, is, in our opinion, intuitively representative of a chronic inflammatory lesion involving decidual vessels. The prevalence of this lesion in our cohort is very low, recorded only 6 times in 924 placentas included in the analysis (0.6%). Our primary outcome was unchanged with the reclassification of this lesion (Supplemental Table 8). In addition, we have acknowledged that the classification of avascular villi can be difficult within the 4 main histologic types and, in some cases, can be subjective. Generally, our pathology practice over the years of the study followed the following guidelines for consistency:

1. Chronically inflamed villi (ie, with lymphocytes within their stroma) were diagnosed as chronic villitis; however, although these villi may have lacked fetal vessels, which are often destroyed by the inflammation, these were not considered to be FVM.

2. Cases of chronic villitis with foci of uniformly avascular villi away from the main inflammatory focus, which were uninflamed, were diagnosed as avascular villi and, in our opinion likely, represented upstream fetal vascular obstruction related to chronic villitis.

These kinds of cases were coded with lesions in both CI and FVM. Continued descriptions of classifications within the literature will continue to improve the generalizability of these findings.

Lastly, because of the retrospective design and review of placentas following delivery, we were only able to estimate the associations between pathology and recurrent PTB and cannot establish causality or mechanistic pathways. Further work is needed to elucidate underlying mechanisms.

Conclusions

High-grade CI in the placenta was associated with an increased risk of recurrent PTB, and low-grade MVM was associated with an increased risk of recurrent PTB among those with an early GA at delivery (<34 weeks of gestation). Further work is needed to develop risk stratification models using placental histology and clinical factors.

AJOG at a Glance.

Why was this study conducted?

This study aimed to determine the independent contribution of placental histologic types to the risk of recurrent preterm birth (PTB).

Key findings

High-grade chronic inflammation in an index placenta was associated with recurrent PTB. Low-grade maternal vascular malperfusion was associated with recurrent PTB only among those with an early (<34 weeks of gestation) index PTB.

What does this add to what is known?

This study presented placental histologic types that may identify those at increased risk of recurrent PTB and suggested underlying mechanisms behind recurrent PTB.