Background

Intrauterine demise or stillbirth is a despairing situation that raises questions of why pregnancy loss occurred, if it will occur again, and if there is any way to reduce the risk of recurrence. Perinatal autopsy can be helpful in answering these questions and providing guidance for the patients and the family physicians, obstetricians, and other health professionals involved in their care.[1,2] As part of the perinatal autopsy examination, the pathologist may perform ancillary testing, which refers to special testing done in addition to the basic physical and microscopic examinations. Examples of ancillary testing that may be performed as part of a perinatal autopsy include genetic, radiological, ultrastructural (electron microscopy), biochemical, and microbiological studies. Clinicians caring for patients with intrauterine demise or stillbirth may expect or request that certain ancillary tests be done as part of a perinatal autopsy to help them understand the cause of demise and facilitate explaining it to their patients.

Ancillary testing may be ordered by a pathologist at the beginning of the perinatal autopsy (“up-front” testing), regardless of the initially suspected cause of death, or it can be deferred to later in the autopsy process when the potential cause(s) of death may be more clear and allow for more focused testing. The benefits of up-front ancillary testing include reducing the time to report completion and the possibility of identifying abnormalities not detected by physical and microscopic examination. Conducting multiple ancillary tests may also contribute to the sense that everything possible has been done to identify a cause of death, which can be reassuring for the pathologist performing the autopsy, the patient, and the clinicians caring for them. Potential drawbacks of performing up-font ancillary testing include increased laboratory workload, increased cost to the health care system, and the possibility of obtaining false or variant results that could complicate the overall interpretation of the autopsy findings. Presently, there are no robust, evidence-based guidelines on the best use of ancillary testing in the setting of perinatal autopsies.

Amniotic fluid infection/inflammation (AFII) is the placental pathological correlate of clinical chorioamnionitis. AFII is defined by both an acute maternal inflammatory response (i.e., acute chorioamnionitis) and an acute fetal inflammatory response.[3] AFII is a common cause of premature delivery and death, and is a rare cause of intrauterine fetal demise (IUFD);[4] as such, it is frequently encountered in perinatal autopsy practice. The pathogenesis of most cases of AFII is thought to involve invasion of the amniotic cavity by microbes, especially bacteria.[4] Microbes in the amniotic fluid can be inhaled and ingested by the fetus, and demonstration of this at autopsy is an important diagnostic feature of AFII. Ancillary testing in autopsies where AFII is suspected usually focuses on identifying these potential infectious causes and may include microbiological culture, fungal culture, and/or molecular testing for viruses. Culture-based microbiological studies require viable microbes and thus must be performed up front. Additional studies, particularly genetic testing, may also be performed up front to address other potential causes of demise. The value of microbiological and genetic ancillary testing in perinatal autopsies where AFII is the cause of death is unclear.

Review of AFII autopsy-related ancillary testing may identify areas for improving practice by deferring or omitting tests that are unlikely to be of clinical value, which could save workloads and costs, and reduce the risk of obtaining false or variant results without impairing determination of the cause of demise. The use and value of ancillary testing in perinatal AFII autopsies has not been assessed in BC. The purpose of this quality improvement study is to determine the patterns of ancillary testing use in this setting and their value in the autopsy final diagnosis.

Methods

This is a retrospective quality improvement study, and the requirement for research ethics board review was waived by the University of British Columbia Research Ethics Board. Autopsy reports for all noncoronial autopsy cases between 1 August 2014 and 31 December 2019 with “amniotic fluid infection” or “chorioamnionitis” in the body of the report were retrieved from the Anatomical Pathology laboratory information system of BC Children’s Hospital and BC Women’s Hospital and Health Centre (C&W). C&W is the provincial referral centre for perinatal and pediatric autopsies, and the cohort in this study represents the majority of perinatal autopsies performed in the province.

Autopsies were included in the study if the pathologist identified AFII as the cause of death, the favored or most likely cause of death, or the inciting factor leading to death. Autopsies were excluded from the study if AFII was not the cause or favored cause of death, there was termination of pregnancy for developmental and/or genetic anomalies, there was a history of noniatrogenic traumatic rupture of membranes (e.g., motor vehicle accident), there was no internal fetal examination (which precludes microbiological testing of fetal tissue), intrauterine demise occurred more than 1 day prior to delivery based on clinical history and/or macerative change, or the fetus and/or placenta were received in formalin, which precludes microbiological and genetic testing. There was no restriction on gestational age at delivery.

Reports that passed these exclusionary criteria were then assessed for associated clinical information (gestational age at delivery, maternal age, gravidity, prior losses, clinical history), diagnostic pathological features of AFII, ancillary studies performed, and their results. Ancillary testing is defined as a study that is performed outside the Division of Anatomical Pathology and which requires a requisition for testing. Radiology was not included as an ancillary test in this study.

Results

In total, 382 autopsies were identified where “amniotic fluid infection” and/or “chorioamnionitis” was included in the autopsy report. Seventy-eight autopsies were excluded, primarily because AFII was not identified as the cause or favored cause of death. Lack of fetal internal examination resulting from autopsy restrictions was the second most common exclusionary criterion, and the remainder were excluded because IUFD occurred more than 1 day prior to delivery, tissues were submitted in formalin, or pregnancy was terminated due to fetal anomalies. In total, 304 autopsy reports were included in the study, which represents approximately 15% of all pediatric and perinatal noncoronial autopsies performed during the study period. This proportion of AFII cases to the overall autopsy workload was similar to that reported elsewhere.[5]

The mean gestational age at delivery was 20.1 weeks (range 6 to 42). The mean maternal age at the time of delivery was 32 years (range 15 to 47), mean gravida was 2.6 (range 1 to 10), and mean number of prior pregnancy losses was 0.7 (range 0 to 9). The most common clinical presentation was spontaneous labor/spontaneous vaginal delivery (51%), followed by premature preterm rupture of membranes ([PPROM] 29%), IUFD (9%), and incompetent cervix (8%). All IUFD cases were delivered spontaneously or induced because of clinical concern about chorioamnionitis.

Slightly more than 90% of the autopsy reports had complete information on the stage and grade of both the maternal and fetal inflammatory responses. Staging and grading of AFII is an attempt to provide a semiquantitative measure of the severity of inflammation.[3] Stage reflects the duration of the inflammatory response and is divided into stage 1 (early), stage 2 (intermediate), and stage 3 (advanced), while grade reflects intensity and is divided into grade 1 (not severe) and grade 2 (severe). Generally, AFII associated with preterm delivery is intermediate or advanced stage. This pattern was apparent in the study cohort, which had a mean maternal inflammatory stage of 2.4/3.0 and grade of 1.2/2.0. The mean fetal inflammatory response stage was 1.6/3.0, and grade was 0.8/2.0. This pattern of lower stage and/or grade for the fetal response is typical of early preterm AFII and reflects the relatively immature fetal immune system in premature deliveries.

In total, 310 bacterial cultures were performed in the study cohort. The proportion of autopsy cases with at least one tissue site cultured and the proportion of autopsies with at least one positive or negative culture result are presented in Figure 1A. The site most frequently cultured was gastric contents (59% of cultured autopsies), lung (31%), subamniotic space of the placental disc fetal surface (6%), and spleen (4%).

The bacterial culture results are presented in Figure 2, and highlight bacteria that are typically isolated in AFII.[4] Bacteria were seen on microscopic examination in 19% of the total autopsy cases. For autopsies where culture was performed, culture and histology (the presence of visible bacteria) were concordant in 14%, culture was positive and histology negative in 73%, and culture was negative and histology positive in 13%.

Fungal culture was performed in 8% of autopsies: one culture was positive for Candida albicans (4% of fungal cultures submitted). Fungal elements were seen microscopically in 2% of the autopsy cases. There was only one autopsy where both fungal culture was performed and fungal elements were identified microscopically. There were no autopsies where fungal culture was positive and histology was negative.

The genetic testing protocol employed for perinatal autopsies over the time frame of this study involved rapid aneuploidy testing, followed by either karyotype (2014 to mid-2015) or chromosomal microarray (mid-2015 to 2019) testing if initial rapid aneuploidy testing was negative. In total, 159 autopsies were submitted for genetic testing. The proportion of autopsies tested and the results are shown in Figure 1B. The abnormal genetic results included a previously known inherited translocation associated with fetal anomalies and three incidental findings. One of the incidental findings was associated with increased risk of neurodevelopmental abnormalities; two were of unclear clinical significance. No fetal developmental anomalies were associated with the three incidental findings.

Polymerase chain reaction for cytomegalovirus was performed in 14% of cases, and all were negative. Other studies, such as electron microscopy, were not performed.

Conclusions

As might be expected from the pathogenic mechanism of AFII, microbiological testing and particularly bacterial culture is the most common ancillary test. Bacterial culture appears to be a valuable ancillary test in defining an underlying infectious etiology in the setting of AFII. A significant proportion of cases had cultures that were positive, while bacteria were not detected by microscopic examination, which indicates that culture is a more sensitive test for bacterial involvement in AFII. In a smaller set of autopsies, histology was positive and culture was negative, which could indicate bacterial inactivation by storage conditions prior to autopsy (i.e., refrigeration) or fastidious bacteria that are difficult to culture. These results indicate that culture and histology should be used concurrently to maximize the possibility of detecting bacteria in AFII. The main drawback of bacterial culture is that presently it has to be ordered up front; however, the imminent introduction of molecular and mass spectrometry-based bacterial testing will improve bacterial identification in AFII-related autopsies and allow for a more selective approach.

Why bacterial culture was not employed in all cases of suspected AFII is unclear, although this finding is similar to prior data that indicate that bacterial culture is performed in only up to 70% of cases.[5] Our study was not designed to assess the practice decisions made by autopsy pathologists at the initiation of the autopsy and what might have influenced their decision to order cultures up front; however, not recognizing AFII as a possibility at the initiation of the autopsy, when cultures need to be performed, is a likely explanation. In this study, most of the clinical presentations were preterm spontaneous delivery, PPROM, or incompetent cervix, which should prompt consideration of AFII and bacterial culture. Concerns that contaminants (i.e., false positives) might confuse the autopsy results and evidence that not all cases of AFII are associated with a detectable infectious etiology may also play a role in not ordering bacterial cultures.

Fungal culture was performed much more sporadically and appears to be much less valuable as an up-front ancillary test. Concordance between fungal culture and histology could not be reasonably assessed because there was only one autopsy where both fungal culture was performed and fungal elements were seen microscopically. Nevertheless, because fungal AFII is relatively rare and fungal elements are typically histologically evident, it seems reasonable that fungal culture should be done only if there is a strong suspicion of fungal involvement at the outset of the autopsy. Retrospective molecular testing can be done in cases where fungal AFII was not initially expected but subsequently discovered on microscopic examination if further characterization is necessary.

Ancillary viral testing was also used sporadically, and in this cohort was confined to cytomegalovirus testing. Viruses are not typically associated with AFII and are a rare cause of infection-related preterm birth and IUFD.[5] Accordingly, no cases of cytomegalovirus or other virus-related AFII were identified. Since viral infection can be identified by microscopy and immunohistochemistry, and retrospective molecular testing may be used to identify many of the typical perinatal viruses, viral testing should be deferred unless there is strong clinical concern.

Genetic testing was performed frequently in this cohort despite the lack of a history suggesting developmental abnormalities or clinical concern about a genetic abnormality in all but one case. Most genetic testing was noninformative, and this includes a small subset of testing failures, which are typically related to the degenerative changes and/or maternal cell contamination associated with the maternal inflammatory response. Incidental findings, although rare, have the potential to confuse the autopsy interpretation and could lead to additional unwarranted genetic testing, surveillance, and patient concern in future pregnancies. These data argue strongly against genetic testing in the setting of AFII unless there is pre-existing clinical concern about a specific genetic abnormality. This is perhaps the most important finding for clinicians who care for patients with intrauterine demise or stillbirth, as requests for genetic testing in the setting of AFII are very common, regardless of whether there is specific clinical concern about a genetic developmental abnormality. These data demonstrate that genetic testing in this setting is a waste of resources and is more likely to produce a spurious result than diagnostically useful information.

Although not the primary intent of this quality improvement study, the clinical data and autopsy findings also provide information about the characteristics of AFII-related death in BC. Almost all cases of AFII-related death involved some form of spontaneous premature delivery. IUFD is rare in AFII, and the mechanism of most deaths appears to involve the deleterious effects of preterm birth rather than the direct inflammatory effects of AFII on the fetus. The bacteria identified by culture are consistent with those previously associated with AFII,[4,5] although Group B Streptococcus was the predominant identifiable cause in our cohort. This may be of clinical relevance because Group B Streptococcus has been identified as a potential cause of recurrent AFII-related pregnancy loss.[6]

This review of ancillary testing in AFII autopsies in BC identified a potential algorithm [Figure 3] for reducing associated workloads and costs for the health care system and avoiding rare but potentially confusing test results without impairing diagnostic accuracy. To illustrate this algorithm, consider a hypothetical example of a 34-year-old primigravida woman who presents to hospital at 21 weeks gestational age reporting a gush of clear vaginal fluid the previous day followed by the onset of lower abdominal cramping. On examination, there are findings consistent with PPROM, and she goes on to spontaneously deliver a stillborn infant the same day. She reports that her prenatal genetic screening was low risk and that she had a normal detailed ultrasound the week prior. An autopsy is requested. Given this history and presentation, AFII would be a likely explanation for premature delivery and death, and bacterial cultures at autopsy would be warranted; however, up-front studies for viral and fungal etiologies and genetic studies would not. When reimagining the same scenario with an abnormal detailed ultrasound suggestive of trisomy 21, it would be reasonable to do up-front genetic testing in addition to bacterial cultures.

The findings of this study can also inform patients and their family physicians, obstetricians, and other health care professionals about the rationale behind why certain ancillary tests are done or not done in perinatal autopsies where AFII is the cause or likely cause of death. For example, a family physician is discussing the autopsy findings with a patient after she had a spontaneous delivery and stillbirth at 19 weeks gestational age. The autopsy report identifies AFII as the cause of premature delivery and death, and indicates that the fetus was normally developed. The autopsy does not mention genetic testing, and the patient is concerned that a genetic abnormality might have been missed because genetic testing was not performed. Based on the data in this study, the family physician can reassure their patient that genetic testing in the setting of AFII and normal fetal development does not provide useful diagnostic information and is unnecessary.

Summary

This quality improvement study focused on AFII-related autopsies, which represent approximately 15% of perinatal autopsies performed in BC each year; however, some of the remaining 85%, such as acute abruption, may also be amenable to algorithms to improve the use of ancillary testing. Additional study in these areas is warranted.

Competing interests

None declared.

This article has been peer reviewed.

Dr Terry is a pediatric and perinatal pathologist at the BC Children’s Hospital and BC Women’s Hospital and Health Centre, a clinical associate professor in the Faculty of Medicine at the University of British Columbia, and an investigator at the Women’s Health Research Institute and BC Children’s Research Institute.