key: cord-0883265-pq5f82gf
authors: PATBERG, Elizabeth T.; ADAMS, Tracy; REKAWEK, Patricia; VAHANIAN, Sevan A.; AKERMAN, Meredith; HERNANDEZ, Andrea; RAPKIEWICZ, Amy V.; RAGOLIA, Louis; SICURANZA, Genevieve; CHAVEZ, Martin R.; VINTZILEOS, Anthony M.; KHULLAR, Poonam
title: COVID-19 Infection and Placental Histopathology in Women Delivering at Term
date: 2020-10-19
journal: Am J Obstet Gynecol
DOI: 10.1016/j.ajog.2020.10.020
sha: 39c04db59eb6761d483e6c2a62d9af8627b463fe
doc_id: 883265
cord_uid: pq5f82gf

Background – There is a paucity of data describing the effects of COVID-19, especially in asymptomatic patients, on placental pathology. Although the pathophysiology of COVID-19 is not completely understood, there is emerging evidence that it causes a severe systemic inflammatory response and results in a hypercoagulable state with widespread microthrombi. We hypothesized that it is plausible that a similar disease process may occur in the fetal-maternal unit. Objective – The aim of this study was to determine whether COVID-19 in term patients admitted to Labor and Delivery, including women without COVID-19 symptomatology, is associated with increased placental injury compared to a cohort of COVID-19 negative controls. Study Design – This was a retrospective cohort study performed at NYU Winthrop Hospital between 3/31/2020 and 6/17/2020. During the study period all women admitted to Labor and Delivery were routinely tested for SARS-CoV-2 regardless of symptomatology. The placental histopathological findings of COVID-19 patients (n=77) who delivered a singleton gestation at term were compared to a control group of term patients without COVID-19 (n=56). Controls were excluded if they had obstetric or medical complications including fetal growth restriction, oligohydramnios, hypertension, diabetes, coagulopathy or thrombophilia. Multivariable logistic regression models were performed for variables that were significant in univariable analyses. A subgroup analysis was also performed comparing asymptomatic COVID-19 cases to negative controls. Results – In univariable analyses, COVID-19 cases were more likely to have evidence of fetal vascular malperfusion, i.e. presence of avascular villi and/or mural fibrin deposition (32.5% (25/77) vs. 3.6% (2/56), p<0.0001) and villitis of unknown etiology (20.8% (16/77) vs. 7.1% (4/56), p=0.030). These findings persisted in a subgroup analysis of asymptomatic COVID-19 cases compared to COVID-19 negative controls. In a multivariable model adjusting for maternal age, race/ethnicity, mode of delivery, preeclampsia, fetal growth restriction and oligohydramnios, the frequency of fetal vascular malperfusion abnormalities remained significantly higher in the COVID-19 group (OR= 12.63, 95% CI [2.40, 66.40]). While the frequency of villitis of unknown etiology was more than double in COVID-19 cases compared to controls, this did not reach statistical significance in a similar multivariable model (OR=2.11, 95% CI [0.50, 8.97]). All neonates of mothers with COVID-19 tested negative for SARS-CoV-2 by PCR. Conclusions – Despite the fact that all neonates born to mothers with COVID-19 were negative for SARS-CoV-2 by PCR, we found that COVID-19 in term patients admitted to Labor and Delivery is associated with increased rates of placental histopathologic abnormalities, particularly fetal vascular malperfusion and villitis of unknown etiology. These findings appear to occur even among asymptomatic term patients.

Fisher's exact test, as deemed appropriate, for categorical variables and the two sample t-test or 169

Mann-Whitney test for continuous data. 170

The inter-rater reliability for five classes of placental histopathologic lesions was 171 calculated using Gwet's "first-order agreement coefficient," (AC 1 ) 21 . The AC 1 variables that were significant in the univariate analyses and Model 2 excluded those factors 180 which could be in the causative pathway, such as pre-eclampsia, fetal growth restriction, and 181 oligohydramnios (as feasible based on sample size). Pre-existing and gestational diabetes were 182 not adjusted for in these models due to inadequate numbers. A numerical measure of the 183 accuracy of the model was obtained from the area under the curve (AUC), where an area of 1.0 184 signifies near perfect accuracy, while an area of less than 0.5 indicates that the model is worse 185 than just flipping a coin. The following was used as a guide for AUC: 0.9-1.0 Excellent, 0.8-0.9 186 Very good, 0.7-0.8 Good, 0.6-0.7 Average, and 0.5-0.6 Poor. The Hosmer and Lemeshow 187

Goodness-of-Fit test was also used to test how well each model fits the data. Results from the 188 multivariable logistic regression models are reported as odds ratios with their corresponding 95% 189 confidence intervals. 190 J o u r n a l P r e -p r o o f For the secondary objective, controls were compared to asymptomatic Chi-square test or Fisher's exact test, as deemed appropriate, was used to compare these two 192 groups for pathological outcomes. A result was considered statistically significant at the p<0.05 193 level of significance. All analyses were performed using SAS version 9.4 (SAS Institute Inc. 194 Cary, NC). Table 1 . COVID-19 cases were significantly more likely to be 201 younger (29.9 years versus 32.2 years, p=0.02), Hispanic (55.8% (43/77) versus 17.6% (10/56), 202 p=0.001) and to deliver vaginally (74% (57/77) versus 25% (14/56), p<0.0001) compared to 203 controls. The majority of study patients 67/77 (87%) were asymptomatic; the remaining 10 204 patients had mild COVID symptomatology including fever, cough, URI symptoms and shortness 205 of breath. None required supplemental oxygen administration or admission to the ICU. 206

Laboratory testing such as inflammatory markers and coagulation studies were not performed on 207 these patients as they were not clinically indicated. Obstetric and medical complications were 208 noted in less than 10% of the COVID-19 cohort including: oligohydramnios (6/77, 7.8%), fetal 209 growth restriction, defined as ultrasound estimated fetal weight less than the 10 th percentile for 210 gestational age(4/77, 5.2%), pre-eclampsia (5/77, 6.5%) and pre-existing or gestational diabetes 211 (7/77, 9.1%). Neonatal outcomes including birthweight, APGAR scores and NICU admissions 212

were similar between groups. 213 J o u r n a l P r e -p r o o f possible colliders (Model 2) this difference remained statistically significant (OR=12. 26 COVID-19 cases were more likely to have evidence of meconium on placental evaluation 246 compared to controls (20.8% (16/77) versus 7.1% (4/56), p=0.030; Table 2 ). While 14.3% of 247 COVID cases had no significant histopathologic abnormalities, this was significantly less 248 common than for the controls (48.2% (27/56) vs. 14.3% (11/77, p<0.0001). Other placental 249 histopathologic findings were similar between groups (Table 2) . 250

For the secondary objective, when comparing asymptomatic COVID-19 patients (n=67) 251 to the controls, a significant difference was again observed in the frequency of fetal vascular 252 malperfusion and villitis of unknown etiology (31.3% (21/67) vs 3.6% (2/56), p<0.0001; and 253 22.4% (15/67) versus 7.1% (4/56), p=0.020, respectively; Table 5 ). In another subgroup analysis 254 of COVID-19 patients comparing those with symptoms (n=10) to those without (n=67), there 255 was no difference in any of the placental histopathologic findings between groups (data not 256 shown). We found that placental histopathological abnormalities were more frequent among 261 COVID-19 cases compared to controls. Both fetal vascular malperfusion and villitis of unknown 262 etiology were significantly more common in the COVID-19 cases than the controls (32.5% 263 (25/77) versus 3.6% (2/56), p<0.001; and 20.8% (16/77) vs. 7.1% (4/56), p=0.030). The 264 difference in fetal vascular malperfusion lesions persisted in multivariable analysis, after 265 adjusting for potential confounders; however, this difference did not persist in a similar 266 multivariable model for villitis of unknown etiology. The statistically significant increase in 267 incidence of fetal vascular malperfusion and villitis of unknown etiology persisted even among 268 asymptomatic COVID-19 cases compared to controls (31.3% (21/67) vs 3.6 (2/56)%, p<0.0001; 269 and 22.4% (15/67) versus 7.1% (4/56), p=0.020, respectively). We did not observe an increase in 270 the other placental histopathologic lesions, including those related to maternal vascular 271 malperfusion, except for presence of meconium, which was likely attributable to higher rate of 272 vaginal delivery in the COVID-19 group and of little clinical significance. 273

In a recent structured review including twenty studies with histopathology findings in 275 third trimester placentas following maternal SARS-CoV-2 infection, evidence of fetal vascular 276 malperfusion was reported in 35% of cases, which is similar to the rate observed in our cohort 277 (32.5%) 24 . In this review, the rate of maternal vascular malperfusion was reported to be higher 278 than that observed in our cohort, while villitis of unknown etiology was less frequent than in our 279 study. The authors note that limited conclusions can be drawn regarding the effects of SARS-280

CoV-2 on the placenta because most of the studies included few patients and lacked control 281 282 identified articles, which we also observed. The exact mechanism by which SARS-CoV-2 may 283 cross the placental barrier is not fully elucidated. However, it is hypothesized that the observed 284 low rate of vertical transmission is due, in part, to the fact that two key receptors which must be 285 present for viral entry into host cells (ACE-2 and TMPRSS2) are rarely co-expressed in third 286 trimester placentas 25 . 287

We identified six studies which reported third trimester placental histopathology findings 288 in greater than five patients with COVID-19 and included a control group 26-31 ( between COVID-19 and controls groups, despite the significantly higher rate of placental 307 abnormalities in the COVID-19 cohort. This suggests that the placenta may act as a defensive 308 biological "filter", offering a degree of protection to the neonate, may modify its immune status 309 or may actively produce factors which may protect the fetus in utero. However, our findings 310 suggest that there is 'injury" to the placenta with possible adverse effects in the neonates, 311 warranting follow up for possible long-term side effects. 312

The etiology and significance of the increased rate of lesions related to fetal vascular 313 malperfusion in our COVID-19 cohort is unclear. In our COVID-19 placentas, most fetal 314 vascular malperfusion lesions had a global distribution, suggesting a partial or intermittent 315 obstruction in umbilical blood flow, rather than segmental, as is seen in the case of localized 316 thrombosis/occlusion of chorionic plate or stem villous vessels. In this global pattern, endothelial 317 damage may reflect altered flow rather than thrombosis per se 37 . Usually, but not always, this 318 pattern is seen in cases with predisposing umbilical cord lesions, none of which were observed in 319 our cohort. If or how SARS-CoV-2 and/or COVID-19 disease may lead to such alterations in 320 fetal umbilical blood flow and ensuing endothelial damage should be a focus of future research. 321

We suspect that indirect immune-and/or viral-mediated systemic damage, rather than local 322 pathophysiologic changes (i.e. hypoxemia) of COVID-19 may be responsible, since our findings 323 were similar when we included only asymptomatic cases in the analysis. Interestingly, we found 324 only two cases of fetal vascular thrombi (both in chorionic plate vessels), in comparison to others 325 who found that many of the cases of fetal vascular malperfusion had evidence of fetal vascular 326 thrombosis 26, 38, 39 , which they hypothesized may be attributable to the increased hypercoagulable 327 state induced by COVID-19. 328

We did not observe an increase in the incidence of placental lesions associated with 329 maternal vascular malperfusion -including, placenta weight <10 th percentile, intraparenchymal 330 or retroplacental hematomas, infarcts or decidual vasculopathy -compared to controls. Maternal 331 vascular malperfusion lesions in the placenta are associated with adverse pregnancy outcomes 332 including stillbirth, preterm delivery and placental abruption and have been observed with 333

increased frequency compared to controls in one study on placental pathology in COVID-19 334 patients 27 . Similarly, we did not find an increase in the incidence of pregnancy outcomes related 335 to ischemic placental disease (fetal growth restriction, oligohydramnios and pre-eclampsia).This 336 may be related to the severity of the disease, with more severe COVID-19 infection resulting in 337 greater damage on the maternal unit of the placenta. Alternatively, the timing and duration of 338 infection may play an important role, with immature placentas being more vulnerable to certain 339 types of viral-mediated damage than term placentas. Importantly, our study included only term 340 placentas, while others have included a significant proportion of preterm deliveries, in which 341 lesions of maternal vascular malperfusion are known to be more common. 342

Our study is the first to observe an increase in the incidence of villitis of unknown 343 etiology in placentas of COVID-19 patients (20.8% (16/77) vs. 7.1% (4/56), p=0.030). While 344 this statistical significance did not persist in multivariable modelling, this was likely secondary 345 Type II error from small sample size. Nevertheless, the rate of villitis of unknown etiology in our 346 COVID-19 cohort is three times higher than in controls and is higher than expected in term 347 placentas (5-15%) 37 . This may be due to the study time period extending several months past the disease, such as maternal hypoxia, influenced these findings because our results were similar in a 360 subgroup analysis including only asymptomatic cases. 361

Chronic intervillositis was observed in placentas from three (4.5%) COVID-19 patients 362 (Image 7 and 8), all of whom were asymptomatic, and none in the control group. This difference 363 was not statistically significant, although the comparison is limited by small sample size. Two 364 previous case reports on placental histopathology in COVID patients have described chronic 365 intervillositis as a significant finding 16, 32 ; however, both cases were preterm gestations with 366 severe COVID-19 symptomatology. The fact that our study included only term gestations, the 367 majority of whom were asymptomatic or with mild symptoms, may contribute to the observation 368 that chronic intervillositis as an uncommon abnormality in our COVID positive cohort. Another strength is that our sample of women with COVID-19 consists of consecutive 387 patients who were routinely tested on admission to Labor and Delivery, regardless of 388 symptomatology, allowing us to compare findings among women with and without symptoms, 389 which few others have done 30, 31 . This study design captured all eligible patients who met the 390 inclusion criteria. 391

One of our study's limitations is that we were not able to assess timing of the COVID-19 392 infection in our study patients; however, in a limited (one-month) audit we identified that for 393 75% of these patients, the finding of positive COVID-19 status was new and the remaining 25% 394 patients were known to be COVID-19 positive by a previous PCR test before admission. Another 395 limitation is to identify the "ideal" control group since placentas are sent to pathology 396 predominantly in patients with high risk maternal or fetal conditions in our institution. Therefore, 397 we had to include many placentas from patients whose only indication for placental evaluation 398 was delivery via cesarean section, which decreased our sample size and ability to match for 399 certain clinical characteristics. This also accounts for the significantly higher rate of C-section in 400 the control group. However, the majority of patients in the control group who delivered via 401 cesarean section underwent labor prior to delivery and there was no statistically significant 402 difference between groups in pre-delivery labor (7/20, 35% versus 20/42, 47.6%; p=0.350), nor 403 would we expect mode of delivery to affect placental lesions related to fetal vascular 404 malperfusion or villitis of unknown etiology. 405

Control placentas had fewer sections examined per placenta than COVID-19 cases, which 406 may contribute to the higher frequency of certain lesions in the latter group. However, if this 407 were the case we would expect all lesions to be more frequent in the COVID-19 group, which 408 was not our observation. It is unlikely that oversampling alone could account for the three-fold 409 increase in lesions of villitis of unknown etiology or the nearly ten-fold increase in lesions 410 associated with fetal vascular malperfusion in cases compared to controls. Although differences 411 in Pathologists' interpretation could influence differences in findings between groups, we found 412 excellent inter-observer reliability (AC 1 >0.9; Table 7 ) upon re-review of approximately one 413 third of all placentas in our study (26 cases and 21 controls). Finally, Pathologists were not 414 blinded to COVID-19 status, as this study was performed in a pragmatic fashion. However, 415 during the reliability assessment, Pathologists were blinded to COVID-19 status and we found 416 excellent interclass correlation. What are the key findings? 31

• Our findings suggest that there are more histopathological abnormalities in term placentas 32 from women with COVID-19, even among asymptomatic patients. 33

• The placental histopathologic findings of fetal vascular malperfusion (32.5% vs. 3.6%) 35 and villitis of unknown etiology (20.8% vs. 7.1%) were significantly more common 36 among women with COVID-19 at term, even in asymptomatic patients. and microangiopathic disease in almost every organ system, including lungs, kidney, heart, and 95 brain, has been identified both clinically and by autopsy examination 17, 18 . Thus, it is reasonable 96 to investigate if a similar disease process occurs in the fetal-maternal unit. 97

We hypothesize that the placentas of pregnant women may have histologic evidence of 98 injury since they could have harbored the virus for weeks prior to delivery, even if the women 99 were asymptomatic. Therefore, we undertook this retrospective cohort study to compare 100 placental pathology findings between women with COVID-19 at term versus a control group of 101 between asymptomatic COVID-19 patients and placental pathology. 103 104

This was a retrospective cohort study involving women who delivered at term at NYU for each group are presented in Table 1 . COVID-19 cases were significantly more likely to be 201 younger (29.9 years versus 32.2 years, p=0.02), Hispanic (55.8% (43/77) versus 17.6% (10/56), 202 p=0.001) and to deliver vaginally (74% (57/77) versus 25% (14/56), p<0.0001) compared to 203 controls. The majority of study patients 67/77 (87%) were asymptomatic; the remaining 10 204 patients had mild COVID symptomatology including fever, cough, URI symptoms and shortness 205 of breath. None required supplemental oxygen administration or admission to the ICU. 206

Laboratory testing such as inflammatory markers and coagulation studies were not performed on 207 these patients as they were not clinically indicated. Obstetric and medical complications were 208 noted in less than 10% of the COVID-19 cohort including: oligohydramnios (6/77, 7.8%), fetal 209 growth restriction, defined as ultrasound estimated fetal weight less than the 10 th percentile for 210 gestational age(4/77, 5.2%), pre-eclampsia (5/77, 6.5%) and pre-existing or gestational diabetes 211 (7/77, 9.1%). Neonatal outcomes including birthweight, APGAR scores and NICU admissions 212 were similar between groups. Table 2 . Placental weights were similar between groups. COVID-19 cases 229

were more likely to have findings associated with fetal vascular malperfusion including 230 avascular villi (Image 1 and 2) and/or mural fibrin deposition (Image 3) (32.5% (25/77) vs. 3.6% 231

(2/56), p<0.0001; Table 2 ). While 14.3% of 247 COVID cases had no significant histopathologic abnormalities, this was significantly less 248 common than for the controls (48.2% (27/56) vs. 14.3% (11/77, p<0.0001). Other placental 249 histopathologic findings were similar between groups (Table 2) . 250

For the secondary objective, when comparing asymptomatic COVID-19 patients (n=67) 251 to the controls, a significant difference was again observed in the frequency of fetal vascular 252 malperfusion and villitis of unknown etiology (31.3% (21/67) vs 3.6% (2/56), p<0.0001; and 253 22.4% (15/67) versus 7.1% (4/56), p=0.020, respectively; Table 5 ). In another subgroup analysis 254 of COVID-19 patients comparing those with symptoms (n=10) to those without (n=67), there 255 was no difference in any of the placental histopathologic findings between groups (data not 256 shown). 257 258 J o u r n a l P r e -p r o o f identified articles, which we also observed. The exact mechanism by which SARS-CoV-2 may 283 cross the placental barrier is not fully elucidated. However, it is hypothesized that the observed 284 low rate of vertical transmission is due, in part, to the fact that two key receptors which must be 285 present for viral entry into host cells (ACE-2 and TMPRSS2) are rarely co-expressed in third 286 trimester placentas 25 . 287

We identified six studies which reported third trimester placental histopathology findings 288 in greater than five patients with COVID-19 and included a control group 26-31 (Table 6) 

Pregnant women with COVID-19 can present with COVID-19 related symptoms, 300 obstetrical complications such as pregnancy loss, fetal death, pre-eclampsia and preterm birth 301 (among others), or can be asymptomatic at term [32] [33] [34] [35] . Despite the fact that nearly all neonates 302 born to COVID-19 infected mothers are SARS-CoV-2 PCR negative 35, 36 , the virus could impact 303 perinatal outcomes through placental injury. Indeed, placental injury has been reported in 304 305 outcomes in our study, including birthweight, NICU admissions and APGAR scores were similar 306 between COVID-19 and controls groups, despite the significantly higher rate of placental 307 abnormalities in the COVID-19 cohort. This suggests that the placenta may act as a defensive 308 biological "filter", offering a degree of protection to the neonate, may modify its immune status 309 or may actively produce factors which may protect the fetus in utero. However, our findings 310 suggest that there is 'injury" to the placenta with possible adverse effects in the neonates, 311 warranting follow up for possible long-term side effects. 312

The etiology and significance of the increased rate of lesions related to fetal vascular 313 malperfusion in our COVID-19 cohort is unclear. In our COVID-19 placentas, most fetal 314 vascular malperfusion lesions had a global distribution, suggesting a partial or intermittent 315 obstruction in umbilical blood flow, rather than segmental, as is seen in the case of localized 316 thrombosis/occlusion of chorionic plate or stem villous vessels. In this global pattern, endothelial 317 damage may reflect altered flow rather than thrombosis per se 37 . Usually, but not always, this 318 pattern is seen in cases with predisposing umbilical cord lesions, none of which were observed in 319 our cohort. If or how SARS-CoV-2 and/or COVID-19 disease may lead to such alterations in 320 fetal umbilical blood flow and ensuing endothelial damage should be a focus of future research. 321

We suspect that indirect immune-and/or viral-mediated systemic damage, rather than local 322 pathophysiologic changes (i.e. hypoxemia) of COVID-19 may be responsible, since our findings 323 were similar when we included only asymptomatic cases in the analysis. Interestingly, we found 324 only two cases of fetal vascular thrombi (both in chorionic plate vessels), in comparison to others 325 who found that many of the cases of fetal vascular malperfusion had evidence of fetal vascular 326 327 state induced by COVID-19. 328

We did not observe an increase in the incidence of placental lesions associated with 329 maternal vascular malperfusion -including, placenta weight <10 th percentile, intraparenchymal 330 or retroplacental hematomas, infarcts or decidual vasculopathy -compared to controls. Maternal 331 vascular malperfusion lesions in the placenta are associated with adverse pregnancy outcomes 332 including stillbirth, preterm delivery and placental abruption and have been observed with 333

increased frequency compared to controls in one study on placental pathology in COVID-19 334 patients 27 . Similarly, we did not find an increase in the incidence of pregnancy outcomes related 335 to ischemic placental disease (fetal growth restriction, oligohydramnios and pre-eclampsia).This 336 may be related to the severity of the disease, with more severe COVID-19 infection resulting in 337 greater damage on the maternal unit of the placenta. Alternatively, the timing and duration of 338 infection may play an important role, with immature placentas being more vulnerable to certain 339 types of viral-mediated damage than term placentas. Importantly, our study included only term 340 placentas, while others have included a significant proportion of preterm deliveries, in which 341 lesions of maternal vascular malperfusion are known to be more common. 342

Our study is the first to observe an increase in the incidence of villitis of unknown 343 etiology in placentas of COVID-19 patients (20.8% (16/77) vs. 7.1% (4/56), p=0.030). While 344 this statistical significance did not persist in multivariable modelling, this was likely secondary 345 Type II error from small sample size. Nevertheless, the rate of villitis of unknown etiology in our 346 COVID-19 cohort is three times higher than in controls and is higher than expected in term 347 placentas (5-15%) 37 . This may be due to the study time period extending several months past the 348 peak of COVID-19 infections in our region, with this lesion developing more often in women 349 who have a longer interval from infection to delivery. In all cases of villitis of unknown etiology 350 in our COVID-19 positive cohort, there was no clinical history that could explain this finding 351 and placental weights were normal (between 10 th to 90 th percentile) in these patients. The 352 presence of villitis of unknown etiology in otherwise healthy patients with normal placental 353 weights suggest that exposure to SARS-CoV-2 may have occurred earlier in the third trimester, 354 with villitis of unknown etiology developing as a sequelae, after weeks of harboring the virus. 355

The etiology of the villitis of unknown etiology seen in these placentas is uncertain; it may be a 356 direct or indirect consequence of viral infection resulting in an exaggerated systemic immune 357 response, including increased levels of cytokines, as has been noted following other respiratory 358 viral infections 40 . It appears unlikely that pathophysiologic changes from severe COVID-19 359 disease, such as maternal hypoxia, influenced these findings because our results were similar in a 360 subgroup analysis including only asymptomatic cases. 361

Chronic intervillositis was observed in placentas from three (4.5%) COVID-19 patients 362 (Image 7 and 8), all of whom were asymptomatic, and none in the control group. This difference 363 was not statistically significant, although the comparison is limited by small sample size. Two 364 previous case reports on placental histopathology in COVID patients have described chronic 365 intervillositis as a significant finding 16, 32 ; however, both cases were preterm gestations with 366 severe COVID-19 symptomatology. The fact that our study included only term gestations, the 367 majority of whom were asymptomatic or with mild symptoms, may contribute to the observation 368 that chronic intervillositis as an uncommon abnormality in our COVID positive cohort. 369 J o u r n a l P r e -p r o o f 

Clinical course of severe and critical COVID-433 19 in hospitalized pregnancies: a US cohort study

Detection of SARS-COV-2 in Placental

Severe acute respiratory syndrome 439

coronavirus 2 in pregnancy: symptomatic pregnant women are only the tip of the iceberg

Evidence for and 442 against vertical transmission for severe acute respiratory syndrome coronavirus 2

Visualization of severe acute respiratory 445 syndrome coronavirus 2 invading the human placenta using electron microscopy

Coronavirus disease 2019 (COVID-19) and pregnancy: a 448 systematic review

Coronavirus disease 2019 in pregnant women: a report based on 451 116 cases

Coronavirus disease 2019 during pregnancy: 453 a systematic review of reported cases

COVID-19 and acute 455 coagulopathy in pregnancy

COVID-19 infection among 457 asymptomatic and symptomatic pregnant women: Two weeks of confirmed presentations to 458 an affiliated pair of New York City hospitals

Outcome of Coronavirus spectrum infections 461 (SARS, MERS, COVID 1 -19) during pregnancy: a systematic review and meta-analysis 462

Maternal and perinatal outcomes with COVID-19: A systematic 465 review of 108 pregnancies

Coronavirus disease 2019 (COVID-19) pandemic 467 and pregnancy

Severe respiratory SARS-CoV2 infection

Does ACE2 receptor matter?

Factors preventing materno-fetal transmission of SARS-471

Transplacental transmission of SARS-CoV-2 473 infection

COVID-19 and its implications for thrombosis and anticoagulation

Pregnancy and postpartum outcomes in a 502 universally tested population for SARS-CoV-2 in New York City: a prospective cohort 503 study

Placental Pathology in 505 COVID-19

SARS-CoV-2 can infect the placenta and is not 507 associated with specific placental histopathology: a series of 19 placentas from COVID-19-508 positive mothers

Detection of SARS-CoV-2 in placentas with pathology 510 and vertical transmission

Third Trimester Placentas of SARS-CoV-513 2-Positive Women: Histomorphology, including Viral Immunohistochemistry

Histopathologic evaluation of placentas after 517 diagnosis of maternal severe acute respiratory syndrome coronavirus 2 infection

SARS-CoV-2 infection of the placenta

Second-Trimester Miscarriage in a Pregnant Woman With 523

SARS-CoV-2 Infection

Pre-eclampsia-like syndrome induced by severe 526

COVID-19: a prospective observational study

COVID-19 in Pregnant Women and Neonates: A 529

Systematic Review of the Literature with Quality Assessment of the Studies

Maternal and neonatal response to COVID-19

Placental and Gestational Pathology 534 (Diagnostic Pediatric Pathology)

Analysis of complement deposition 536 and viral RNA in placentas of COVID-19 patients

Placental Pathology in Covid-19 Positive Mothers

High rate of chronic villitis in placentas 540 of pregnancies complicated by influenza A/H1N1 infection

Maternal vascular malperfusion (12/15; 80%)* Fetal vascular malperfusion -formal diagnosis (1/15; 7%) Findings suggestive of FVM (any feature): Clustered avascular villi

Low-grade chronic lymphocytic villitis (2/15; 7%) Chronic deciduitis (2/15; 7%)

Normal" controls -historical uncomplicated term deliveries with GBS+ as the only clinical indication for pathology exam

Pathologic" controlshistorical late preterm and term deliveries of infants with HIE (n=130) Maternal vascular malperfusion

29 74 Consecutive term and preterm SARS-CoV-2 negative placentas submitted based on "institutional criteria of maternal and fetal conditions" and matched for delivery time period (n=290) Maternal vascular malperfusion "Vasculopathy

Fetal vascular malperfusion "Thrombosis" (18/74; 24%) Avascular villi

subgroup analysis of symptomatic versus asymptomatic infections performed Maternal vascular malperfusion Decidual vasculopathy: 3/51 (6%) Accelerated villous maturity: 10/51 (20%) Villous agglutination: 21/51 (41%)* Infarcts: 7/51 (14%) Intervillous thrombus: 8/51 (16%) Subchorionic thrombus: 9/51 (18%)* Fetal vascular malperfusion Avascular villi (segmental): 5/51 (10%) Fetal thrombotic vasculopathy: 4/51 (8%) Chorangiosis: 8/51 (16%) Chronic villitis

31 50 Term and preterm singleton placentas from historical controls submitted to pathology at Maternal vascular malperfusion Accellerated villous maturation

Distal villous hypoplasia 2/50 (4%) Excessive infarction: 4/50 (8%) Old hemorrhage in membranes: 1/50 (2%)

Villitis of unknown etiology: Terminal villi with a T cell lymphocytic inflammatory 617 infiltrate in the villous stroma. CD3 immunohistochemical stain x 400

Chronic intervillositis: Aggregates of histiocytes between terminal villi. H&E stain x

Chronic intervillositis: Terminal villi with intervillous histiocytes seen on CD 68

Acknowledgements: We would like to thank the Pathology Physician Assistants and Resident 427Physicians who were integral to our study, and responsible for the processing of placentas for 428 this study. We would also like to thank Rosanne Vertichio for her assistance in data collection. 429 430 J o u r n a l P r e -p r o o f

number 2GV6 Rabbit MAB) were obtained from Ventana Medical Systems (Tucson, AZ). The 147 slides were counterstained with hematoxylin. 148Microscopic examination was performed by a single Pathologist for all placentas from 149 COVID-19 cases (PK), and by one of three Pathologists for the control placentas (PK, VD, AH). 150The Pathologist were not blinded to clinical information, including COVID-19 status. The 151 findings recorded for the study were identical to that provided to the Ob/Gyn for clinical 152 purposes. Histopathologic lesions were categorized according to the 2016 Amsterdam criteria 20 . 153Seventy-one neonates born to mothers with COVID-19 had test results available in the EMR for 154 SARS-CoV-2 via PCR of nasal swabs using Cepheid Xpert Xpress. 155Subsequently, we performed a second review on a subset of 26 COVID-19 placentas and 156 21 control placentas in order to assess inter-observer reliability. Representative slides from each 157 placenta were independently reviewed by the above three Pathologists (PK, VD, AH) to assess 158 for the presence or absence of the following: fetal vascular malperfusion with avascular villi, 159 high grade; fetal vascular malperfusion with avascular villi, low grade; fetal vascular 160 malperfusion with mural fibrin deposition; villitis of unknown etiology, high grade and villitis of 161 unknown etiology, low grade. The Pathologists were blinded to COVID-19 status during the re-162 review and the previously collected data were not changed by the results of this inter-observer 163 reliability assessment. 164

Descriptive statistics (mean ± standard deviation or median (25 th , 75 th percentiles) for 166 continuous variables; frequencies and percentages for categorical variables) were calculated 167 separately by group (cases vs. controls). Groups were compared using the chi-square test or 168

Principal findings 260We found that placental histopathological abnormalities were more frequent among 261 COVID-19 cases compared to controls. Both fetal vascular malperfusion and villitis of unknown 262 etiology were significantly more common in the COVID-19 cases than the controls (32.5% 263 (25/77) versus 3.6% (2/56), p<0.001; and 20.8% (16/77) vs. 7.1% (4/56), p=0.030). The 264 difference in fetal vascular malperfusion lesions persisted in multivariable analysis, after 265 adjusting for potential confounders; however, this difference did not persist in a similar 266 multivariable model for villitis of unknown etiology. The statistically significant increase in 267 incidence of fetal vascular malperfusion and villitis of unknown etiology persisted even among 268 asymptomatic COVID-19 cases compared to controls (31.3% (21/67) vs 3.6 (2/56)%, p<0.0001; 269 and 22.4% (15/67) versus 7.1% (4/56), p=0.020, respectively). We did not observe an increase in 270 the other placental histopathologic lesions, including those related to maternal vascular 271 malperfusion, except for presence of meconium, which was likely attributable to higher rate of 272 vaginal delivery in the COVID-19 group and of little clinical significance. 273

In a recent structured review including twenty studies with histopathology findings in 275 third trimester placentas following maternal SARS-CoV-2 infection, evidence of fetal vascular 276 malperfusion was reported in 35% of cases, which is similar to the rate observed in our cohort 277 (32.5%) 24 . In this review, the rate of maternal vascular malperfusion was reported to be higher 278 than that observed in our cohort, while villitis of unknown etiology was less frequent than in our 279 study. The authors note that limited conclusions can be drawn regarding the effects of SARS-280CoV-2 on the placenta because most of the studies included few patients and lacked control 281 limitation is to identify the "ideal" control group since placentas are sent to pathology 396 predominantly in patients with high risk maternal or fetal conditions in our institution. Therefore, 397 we had to include many placentas from patients whose only indication for placental evaluation 398 was delivery via cesarean section, which decreased our sample size and ability to match for 399 certain clinical characteristics. This also accounts for the significantly higher rate of C-section in 400 the control group. However, the majority of patients in the control group who delivered via 401 cesarean section underwent labor prior to delivery and there was no statistically significant 402 difference between groups in pre-delivery labor (7/20, 35% versus 20/42, 47.6%; p=0.350), nor 403 would we expect mode of delivery to affect placental lesions related to fetal vascular 404 malperfusion or villitis of unknown etiology. 405Control placentas had fewer sections examined per placenta than COVID-19 cases, which 406 may contribute to the higher frequency of certain lesions in the latter group. However, if this 407 were the case we would expect all lesions to be more frequent in the COVID-19 group, which 408 was not our observation. It is unlikely that oversampling alone could account for the three-fold 409 increase in lesions of villitis of unknown etiology or the nearly ten-fold increase in lesions 410 associated with fetal vascular malperfusion in cases compared to controls. Although differences 411 in Pathologists' interpretation could influence differences in findings between groups, we found 412 excellent inter-observer reliability (AC 1 >0.9;