Revision of the MRI stillbirth study protocol

Study of Stillbirth with MRI of the Fetus in the Uterus

Specific Aims:

1. To obtain an MRI of the stillborn infant in the mother’s uterus in order to examine the full umbilical cord in three dimensions to look for its vulnerabilities from fetal position or cord wrapping.

2. To obtain a standard maternal prenatal history that is expanded to include details of events occurring around the presumed time of fetal death.

3. To correlate the umbilical cord findings from the MRI with the clinical history, infant autopsy, and placental examination.

Scientific Rationale/ Previous Work

            Aim 1:

            The percentage of stillbirth due to unknown cause has varied in different studies as the definition of an acceptable cause of death varies, for example 18% in the Stockholm classification1 and 15% in a study of autopsies2.  In a critical and insightful review of classification of stillbirth, the authors point out that many rely on association3. The approach in this study to not classify stillbirth, but to establish a complete chain of causation based on clinical and anatomic evidence. 

            The personal experience of one of us (RB) found that of 336 stillborn autopsies older than 20 weeks of gestation from his personal database, 109 cases (32%) were coded as either “unknown cause” or “acute asphyxia”. The latter code was based on finding intrathoracic petechiae and excluded cases with a retroplacental hematoma covering greater than 50% of the placental surface. Also excluded from these categories were autopsies of previable infants delivered because of preterm labor or preterm premature rupture membranes, as well as infants with fetal hydrops, intrauterine growth restriction or many specific causes such as fetal maternal transfusion or antibody induced hyperthyroidism. In these 109 cases without an anatomic cause of death, an interruption of umbilical cord blood flow was often inferred by exclusion. In some cases, there was reason to suspect an abnormal umbilical cord event, for example a furcate insertion of the cord, impressions in the skin from multiple nuchal cord wrapping or an umbilical cord knot, but no definitive way to prove that the cord lesion caused the death. Rarely there was direct evidence of cessation of umbilical blood flow by color differences on either side of a knot or compressed segment of umbilical cord. Some cases had indirect evidence of flow abnormalities in the fetal circulation, namely, fetal vascular thrombi and other histological evidence of fetal vascular malperfusion4,5. These fetal vascular lesions have been associated with unexplained fetal death and with cord abnormalities, but in most cases without a specific explanation of the cause or of the distribution of the lesions within the placenta. 

            Autopsies of stillborn infants not showing an anatomic cause of death may have lesions that provide clues as to the mechanism of death. The meaning of these lesions can be deduced by comparison with stillbirth of known cause. An example, is the comparison with findings in deaths from premature separation of the placenta/ placental abruption6. If the placental separation is large, the infant typically demonstrates intrathoracic petechiae and may show meconium or excessive vernix aspiration. These findings can reasonably be interpreted as the initiation of gasping respiration in utero. In smaller, but still lethal, placental separations, the typical findings are mild to moderate pleural and pericardial effusions and dilation of the cardiac chambers. These features can be interpreted as acute heart failure from hypoxia and lactic acidosis7. Some infants show both sets of features suggesting an immediate acute, but non-lethal, hypoxia inducing gasping, but then followed by hypoxia/progressive acidosis and death. Infants without a known placental or maternal cause of asphyxia may show this same array of lesions with the implication that there was either acute asphyxia with gasping or heart failure from partial hypoxia leading to death. Both sets of findings can reasonably be attributed to hypoxia from loss of umbilical cord blood flow if there is no other clinical or anatomic explanation of death.

            In addition to autopsy evidence, there is some clinical evidence for a mechanism of acute asphyxia as a cause of fetal death. In a longitudinal study of fetal growth restriction, neither Doppler measurements nor biophysical profile was abnormal prior to a majority of fetal deaths after 34 weeks of gestation8. The findings were suggestive of an acute hypoxic event occurring between fetal screenings. Further evidence for an umbilical cord cause of asphyxia was found in a meta-analysis showing that clinical reports of true knots of the umbilical cord and multiple nuchal cord wrappings had a significant correlation with stillbirth9. Umbilical cord knots can form as a consequence of an infant slipping through a body wrapped umbilical cord during delivery. These studies were based on the cord after delivery, but not all cord wrapping would be evident after delivery of the infant. 

            The relationship of cord wrapping to umbilical cord blood flow was investigated with an in-vitro model created by continuous perfusion of the umbilical vein of delivered umbilical cords that were then wrapped around PVC pipes. This model found that twisting the portion of cord distal to the wrapping caused flow in the vein to stop. The degree of twist needed to stop flow was directly related to the length of the cord being twisted. This observation makes physical sense since twisting a tube will produce more twists per length the shorter the tube. At some point, the twisting will result in collapse of the umbilical vein wall, analogous to twisting a paper straw until the lumen collapses. In vivo, the cord that corresponds to the experimental length is the length from the placental insertion to the beginning of cord wrapping around a fetal part10. In vitro, as little as 180 degrees of rotation could collapse venous flow in a 3 to 8 cm segment of umbilical cord. 

            In addition to wrapping that causes a functionally short, vulnerable cord segment, the umbilical cord flow can be compromised by occult prolapse, that is with the cord trapped between the presenting fetal part, usually the vertex, and the pelvis. This is most evident during labor with contractions causing deep variable fetal heart rate decelerations, the fetal response to sudden loss of umbilical cord flow. In complete prolapse of the cord, the cord passes through the cervix before the fetal part and is an obstetrical emergency. The unknown question is whether occult prolapse prior to labor, due to decreased amniotic fluid later in pregnancy or descent of the fetal part into the pelvis, usually in primigravida mothers, can cause umbilical cord compromise. This compromise conceivably may not be just from direct compression but from entrapment and kinking of the cord, similar to a kinked garden hose. Similar to cord wrapping, this entrapment would not be detected after delivery of the infant.

            To make a diagnosis of either a functionally short cord from fetal wrapping or of an occult prolapse or entrapment of the cord requires visualization of the cord prior to delivery of the infant. Recently, MRI of an infant in the uterus have demonstrated that the full length of the umbilical cord can be visualized11. (This was not the purpose of the study, but private communication with the senior author confirmed that this is possible*.) This technology provides a potential to discover previously undetectable umbilical cord anatomy in a stillborn infant prior to delivery. 

            Aim 2

            Recent developments in clinical stillbirth prevention have included more attention to changes in fetal movement pattern perceived by the mother that are more sensitive than traditional kick count. These studies, some still in progress were presented at the last meeting on stillbirth sponsored by Star Legacy. The changes in fetal movement pattern presumably reflect compromise of umbilical blood flow. Sleeping position, respiratory problems, medications and gastrointestinal illness or fasting have been hypothesized to decrease fetal oxygenation and or increase acidosis12-23. Hypothetically, a combination of these factors with or without an added increased umbilical venous resistance could lead to a tipping point with fetal acidosis that decreases fetal cardiac output leading to a progressive increase in fetal hypoxia and lactic acidosis until cardiovascular collapse. 

            Aim 3

            This study will correlate maternal perception of fetal activity, the histological timing of fetal thrombotic vasculopathy and the autopsy findings to estimate the duration of umbilical cord comprise prior to fetal death. The timing of fetal death to delivery will utilize the histological autopsy criteria described by Dr. Genest and colleagues. This timing will allow a more accurate estimate of the time of fetal death after the delivery. The details of the events around the time of fetal death will be compared to the anatomic findings in the placenta and at autopsy and the MRI findings. In addition, the clinical risk factors in the maternal history, and the detailed findings in the placenta and autopsy will be utilized to interpret the chain of causation leading to death. This approach to the mechanism of death as opposed to risk factors to study stillbirth has been pioneered by a study of amniotic fluid markers24. RB has participated in the international conference in Dublin on a consensus in placental pathology and contributed to chapters in the final report of that effort. He also has been posting his experience and published literature review on placental examination and autopsy on a website (www.obstericalpathology.com) that will form the basis for interpreting the pathology findings. The identification of pathological lesions will be confirmed by consensus of at least 2 pathologists, and whenever possible objective, quantitative measures will be made.

            Methods:

Type of study: This is a prospective, observational study applying a new technology, MRI of the stillborn infant in utero, with clinical pathological correlation.

Recruitment: The patients will be recruited from mothers presenting to Magee Hosptial for woman who have an undelivered stillborn infant.

Part 1: 

            Eligibility: A maternal medicine physician will determine that 1) the patient is not in active labor and 2) the fetal death was not inevitable given maternal or fetal disease. An informed consent for the MRI procedure will be obtained. After consent of an eligible patient, every effort will be made to perform an immediate full abdomen/pelvic MRI scan taken with the standard precautions and patient education.

            Evaluation of the MRI scan: The complete course of the umbilical cord will be followed. The following points will be identified and measured from the placental insertion: the beginning of a cord wrapping around a fetal part, beginning and end of a cord segment between a fetal part and the bony pelvis (with only the uterus between them), beginning and end of cord segments that are thin or compressed, and the apex of any acute angle made by the cord. Findings of less than a 10 cm segment of the cord between the placental insertion and beginning of fetal wrapping will be considered a positive scan. A focus of the cord between the fetal part and the bony pelvis with either narrowing or an acute angle will also be considered as a positive scan. The fetus, placenta and maternal structures will be conventionally evaluated. 

            Umbilical cord examination: The umbilical cord will be examined according to standard protocols at Magee Hospital. In addition, the umbilical cord insertion will be twisted and its vascular stability noted. The whole cord will be photographed and areas to be sectioned have markers on the photograph. The MRI defined points described above will be measured and marked from the placental insertion. Multiple 2-3 mm sections around each of these points will be identified by location and submitted for histological evaluation along with standard sections from the fetal and placental insertion portions of the cord, and sections of any possible compression of the cord. 

            Microscopic lesions of the cord at MRI identified points will include focal necrosis or thinning of the umbilical vein or artery wall, acute intimal inflammation or mural thrombus, and any morphological evidence of acute torsion in the medial muscle pattern. The presence or absence of such lesions will not influence the designation of a positive or negative scan, but will provide confirmation of likely hemodynamic significance of the umbilical cord configuration. 

Part 2: 

            Eligibility: The same criteria apply as in part 1. The consent will be separate from that of part 1 of the study. In addition, the patient will also be asked to consent to a fetal autopsy. The research will be explained, and the autopsy will be explained. The parent(s) will be asked for separate consents for each.

            Controls: Eligibility will be an uncomplicated labor and delivery of a live born infant of a gestational age within 2 weeks of a matched case. After informed consent, the parents will be asked after delivery the same questions as the cases, but they will be asked about events 3 days prior to delivery.

            The Questions: 

            The parents will be asked to answer questions relating to events within the prior three days with reassurance that these are neutral questions and we do not know of their significance to the fetal death. The questions will be asked, if possible, in the interval, before or after the MRI scan but prior to delivery. If necessary, they will be asked after the delivery and recovery of the mother. When possible, both parents will be present for the questions, but this is not required.

            The baseline time and date of the questions will be recorded. Then a member of the research team will ask the questions and record the answers on a digital data sheet. The presumption is that most deaths will have occurred within 3 days. However, if preliminary evidence from the history, recent ultrasound, or parental questioning suggests that the death may have been even earlier, the questioner can adapt the questions to this longer time span.

The Questions

 1: Perceived Fetal Movement: 

Has there been any change in your perception of fetal movement in the previous days?

            Y/N, if Y:

            A: Have there been an episode or episodes of decreased fetal movement? 

            Y/N, if Y, then record when in hours before interview, more than one period                                 can be recorded.

            B: Have there been an episode or episodes of increased movement?

             Y/N, if Y the intervals can be recorded as above but in addition there is data space for                  duration

            C: Has there been a change in the pattern of fetal movement?

             Y/N if Y, this is an open-ended text entry.

2: Sleeping:

A: How would you describe your sleep pattern? 

            Open ended text data field

B: Do you have a diagnosed history of sleep apnea?

            Y/N

C: Do you take any medications to help you sleep?

            Y/N, if Y, text with name of medication, dose, when taken

D: What is your preferred sleep position during this pregnancy?

            lateral, prone, supine, unknown

3: Respiratory:

A: Over the last few days have you been unusually short of breath?

            Y/N if Y text description including clinical history asthma, etc.

B: Have you used an inhaler?

            Y/N if Y type, dose last used

C: Have you had a cold or cough in the last week?

            Y/N if Y open text

4: Acidosis:

A: Have you had over the last 3 days any gastrointestinal upset that led to vomiting or diarrhea?

            Y/N if Y open text about amount, duration, and effect on oral intake

B: Have you been eating and drinking normally over the last 3 days?

            Y/N if yes text details

5. Medications:

A: Have you taken any medications to relieve pain including aspirin, Tylenol, acetaminophen, Advil, Motrin, ibuprofen?

            Y/N record dose, timings and reason for

B: Are you taking other medications including over the counter, herbal?

            Y/N if Y record dose, timings and reason for

Other:

A: Was there anything else unusual occurring in the last three days?

            Y/N if yes, free text

(This question is being considered)

B: Have you had sexual intercourse in the last three days?

            Y/N if Y when 

Note:   

            In the stillbirth cases, the questions will be reviewed in light of the timing of fetal death based on the Genest criteria. In the parent interview, if the time of death was prior to three days before the interview, the questions will be modified to cover this time period. Parents will be allowed to modify their answers at the autopsy conferences, but both sets of answers will be kept for separate analysis.

Part 3 Collection of all information on the stillbirth 

            1. Medical chart review: This will be entered into the data collection forms of the NIH collaborate stillbirth network25.

            2. Autopsy:

            After permission for an autopsy is granted, the infant will undergo the standard protocol used at Magee Woman’s Hospital. This protocol includes all the external measures and organ weights and anatomic features such as intrathoracic petechiae, islet cell cytomegaly, etc. that will be used in the study analysis. The following procedures will be specifically added or evaluated for this study:

Gross measures:

  1. Volume measurements of pleural and pericardial effusions
  2. Circumferential ventricle measurements at the midpoint between atrioventricular valve and the cardiac apex. 

Specific samples for histology:

  1. A skin sample from the thoracic incision to evaluate edema, thickness of adipose and skn maturation
  2. A longitudinal section of costo-chondral junction to evaluate for endochondral growth columns and transverse arrest lines
  3. Heart samples: Longitudinal samples that include papillary muscle and cross sections midway from valve to apex from both ventricles.
  4. Cross section of ductus arteriosus
  5. Cross section of the pancreas at the head, mid-section and tail
  6. Cross section of the adrenal frozen for fat stain

Additional histologic evaluation:

  1. Measure of blood space as a percentage area  in the adrenal, lung, liver, spleen, kidney
  2. Measure of percentage area of cortex in the thymus
  3. Measure of percentage area of lymphoid tissue in the spleen
  4. Measure of lipid-stained area in the adrenal
  5. Measure of percentage area of insulin and somatostatin cells in pancreas
  6. Measure of percentage volume of colloid in the thyroid
  7. Measure of thickness of the ductal intima
  8. Measure of percentage area of erythropoiesis in liver

Special immun0-staining

  1. GFAP of frontal and occipital cortical white matter
  2. BNF of heart
  3. Somatostatin and insulin stain of pancreas

Other: The frozen adrenal sample will be saved as a potential DNA source.

Analysis

            The histological percentage areas will be multiplied by the appropriate organ weight, and if determinable, by the volume from the MRI. The immuno-staining for GFAP and BNF will be graded on a semiquantitative scale. The rationale and background for these additional studies are in the autopsy manual on www.obstetricalpathology.com.

            Given the variation in gestational age and of intrauterine postmortem retention time superimposed on the variable duration, severity, and variety of mechanism of death, it is anticipated that strong correlations will be found in this preliminary study. The goal is to improve techniques. For example, percentage areas are routinely done by point counting on a microscopic grid, but automated methods on scanned slides will be explored for this study. The value of immune-stains in stillbirth with variable antigen preservation postmortem will critically evaluated, and some not yet included immune-stains will be explored.

            The long-term goal from a larger study is to test hypothesis about mechanism of death based on correlation of clustered autopsy features that reflect specific mechanisms and in turn these clusters will correlate with clinical events, risk factors, placental findings, and MRI umbilical cord findings. Correlations in this pilot will be tested using cluster analysis.

Hypotheses to be tested in this small study include:

  1. That a fetal gasping cluster will correlate intrathoracic petechiae with increased lung airspace volume, increased capillary blood, volume, and evidence of meconium/vernix aspiration.
  2. That a subacute hypoxia cluster will correlate with increased pericardial and pleural effusion volume, increased mid ventricular diameter: heart weight ratio, and decreased brain natriuretic peptide staining.
  3. That a chronic hypoxia cluster will correlate with increased erythropoiesis, thymic cortical atrophy, increased somatostatin and increased thickness of lipid laden cells in the adrenal
  4. That an acute hypoxic cluster with natural resuscitation (such as temporary cord torsion or compression) will correlate neuronal necrosis and increased GFAP white matter staining.
  5. That a subacute asphyxia injury will be associated with increased circulating nucleated red cells and decreased liver erythropoiesis and that chronic hypoxia will correlate with increased circulating nucleated red cells and increased liver erythropoiesis.
  6. That a fetal shock cluster will correlate liver sinusoidal thrombi, adrenal hemorrhage, and papillary muscle necrosis
  7. That a ductal constriction cluster will correlate ductal intimal thickness, increased blood volume in the lung, increased blood volume in the liver, colloid loss in the thyroid and dilated right ventricle.
  8. That an immaturity cluster will correlate increased insulin cells, increased subcutaneous adipose, increased erythropoiesis, and increased cardiac wall thickness.

The clusters significant clusters will be correlated with non-autopsy data. For example, cluster 8 with placental villous dysmaturity, and maternal glucose intolerance, or clusters 1 and 4 with umbilical cord wrapping creating a functional short cord. In addition to statistical analysis, all correlations/associations will be compared to known clinical and experimental physiology and known risk factor for stillbirth, to look for plausible causation. Using both physiology and clinical and experimental timing of pathologic lesions, to compare onset with maternal events prior to fetal death, and identify window of potential diagnosis of risk and of potential intervention. It is not anticipated that this type of analysis can be completed on this pilot study. The goal is to test the analytic structure for plausibility.

                        3. Placenta examination: A routine protocol for such examination at Magee Hospital will be utilized. In addition, an estimate will be made of functional placental reserve: 

1) the volume of placental tissue will be estimated from the disc outline and average thickness and morphometrically from a photograph of the sliced placenta when appropriate 

2) the volume of placenta is not compromised by perivillous fibrinoid deposition, fetal vascular malperfusion including thrombi, villitis, infarctions and retroplacental hematoma, based on gross images, and on microscopic samples taken until a reasonable distribution function of the involved volume percentage of lesion is obtained. 

3)A novel approach to mapping fetal vascular malperfusion lesions, fetal vascular thrombi and vascular resistance (as measured by diameter and wall thickness) will be attempted using 4 orthogonal sections around the cord insertion to look at major vessel branches, and sections parallel to the surface to obtain larger samples of stem vessels, as well a following visible thrombi to establish extent of compromised circulation.

4) the total volume of compromised placenta and an estimate of the timing of the lesion based on histological criteria of tissue repair. 

5) the linear extent and average density of capillary syncytial membranes within a volume of functional placenta derived from a sampling of functional microscopic areas.

6) A count of syncytial knots in the same areas as a measure of decreased uteroplacental flow.

            A portion of clean placental villi will be preserved for possible genomic studies.

            In the process, the study will search for automated techniques to make measurements, and of novel ways to find information such as tissue ultrasound and MRI. 

The team conference:

            .

            After all the studies are complete, including the MRI, autopsy, placental examination, medical chart review and any clinical care of the patient during the current admission, the research team will meet to review the case. The process will be to deduce from the evidence a biologically probable timeline of pathogenesis that led to the fetal death, and an assessment of potential points of prevention. This chain of causation will utilize all the information available: the clinical risk factors including prenatal Doppler studies, maternal medical history, events described by the parents, and the anatomic findings in the placenta and autopsy. This process will reveal gaps in understanding in forming pathogenetic connections. 

            From the proceedings of these conferences, an exploratory post hoc analysis of correlations of specific anatomic and clinical features will be performed. The goal is to develop hypotheses to test in a larger multi-institutional study.

            The conclusions from our follow-up conference will be conveyed to the parents during the obstetrical conference to review the autopsy with parents with the goal of a time frame of 6 weeks to complete analysis of each case. If the parental questions need to be re-asked because the timing of death was beyond 3 days, done at the autopsy mortem conference with the parents, then the case may be reassessed at an additional follow up conference (see below) after that meeting if there is additional information

Anticipated outcomes: 

Part 1

            Based on 224 records of autopsies that RB performed in Louisville, KY from 2000 to 2014, approximately 31% have no specific cause of death after a complete autopsy and placental examination. This group is composed of 58 cases classified as “unknown” and 12 classified as “unexplained acute asphyxia” based on the finding of intrathoracic petechiae. Eleven of the “unknown” cases were also coded as having fetal thrombotic vasculopathy which is considered a possible indicator of umbilical cord blood flow occlusion, making 33 patients with highest probability of compromised umbilical cord blood flow (12 with acute asphyxia and 11 “unknown” with fetal thrombi). Magee Hospital had 35 autopsies performed last year. We would anticipate that at least half of those patients would have been eligible for the study, and that anyone consenting to an autopsy would also consent to an MRI. This would provide 35 patients for the study over 2 years of which 10 (31%) would be expected to have an unknown cause of death, and of those 5 would have a high probability of a cord abnormality. Therefore, we would anticipate between 5 to 10 cases with an umbilical cord compromise as a cause of death.

            The null hypothesis is that there will be zero abnormal cord configurations on MRI from 35 patients. Any number found would be significant. Until we undertake a control group with normal pregnancy and delivery there is no valid control group. For this pilot study, finding even one predicted cord configuration on MRI would demonstrate that such a mechanism of fetal death exists. There is experimental and indirect clinical evidence that such a mechanism would be biologically plausible. A positive event in the pilot would merit a larger, multi-institutional study. 

Part 2

            Using a P <.05, we will statistically test for the differences between answers to our questions between control and mothers of stillborn infants. However, this test will not prove that the answer in the stillborn mother is not related to the death of her infant even if the question is not significantly different between the groups. Our hypothesis is that multiple factors may combine to initiate a tipping point of progressive fetal acidosis. If an important risk factor, such a multiple wrapping of the umbilical cord, is not present in the control mother, then the added factor, for example, taking Advil for a headache the day before delivery, may not have had a sufficient effect on her infant to cause death, but would have if the infant also had multiple cord wrappings. The main anticipated benefit of the answers to the questions in mothers of live born infants is to know the frequency of positive answers in uncomplicated pregnancies.

            The outcome for this part of the study is to pilot ideas about possible correlations of conventional risk factors, immediate premortem events, autopsy, placental findings and MRI findings. Some hypotheses may be testable using the approximately 70 stillborn autopsies that we would anticipate over the two years: Example anticipated correlations: 1) Maternal perception of a period of sudden intense fetal movement with anatomic evidence of fetal gasping. 2) Prolonged fetal calm prior to the time of death with anatomic evidence of subacute/fetal hypoxia, 3) Sleeping in the supine position or respiratory illness with anatomic evidence of chronic hypoxia. 4) Fasting with fetal heart failure due to hypoglycemia or fetal acidosis in diabetic patients. In practice, few correlations could be statistically tested given the small numbers of cases. However, the observational aspects of the study would create a detailed chain of causation which would aid the development of future studies. The decisions about which correlations to test post hoc would be developed by the research team from the follow up case conferences.

             As a hypothetical example of the process: A specific umbilical cord configuration is associated with an increased risk of fetal death, with fetal vascular malperfusion lesions in the placenta of more than 7 days duration and with days of perceived decreased fetal movement prior to fetal death. Some stillborn infants with this cord configuration were also associated with other risk factors such as obesity with supine sleeping or with NSAID taken after 36 weeks of gestation, and these events occurred on the day of fetal death. Finally, the cord configuration is also correlated with a specific prenatal Doppler pattern in the fetus. This set of correlations would imply a more complex hypothesis that the cord configuration decreases umbilical blood flow and a second event triggers an additional hypoxic stress that leads to fetal death. In future studies, this hypothesis could be tested by screening for the Doppler pattern, documentation of the cord configuration with an MRI, and then in this positive sample, comparing, for example, education to avoid secondary risks of stillbirth compared to in hospital continuous FHR monitoring to reduce stillbirth. This is a purely hypothetical construct.

            The anticipated success for the pilot study is to discover/develop novel hypotheses about stillbirth, having a basis in a chain of causation, that can be tested either observationally or by intervention studies. This will lead to larger multi-institutional studies with enough power to test these more specific hypotheses. The ultimate goal to understand and prevent what are now unexplained stillbirths. 

            Part 3: There will not be enough cases in this pilot to make definitive correlations, but hypothesis should become clearer during the process. For example, a key question is one of timing, how long was the fetus compromised before death. Preventative strategies would be influenced by this timing. 

            If this interval was short, the autopsy would demonstrate intrathoracic petechiae, no thymic atrophy or adrenal lipid accumulation. The maternal history may include a period of rapid fetal activity before decreased fetal activity. This group of features might occur with cord wrapping with a very short free cord length from insertion to wrapping. The hypothesis to be tested in a larger study is that given this cord configuration correlates with a high risk of sudden fetal death. If the correlation is strong, a risk benefit analysis might show that the cord configuration would require intervention on the basis of the cord lesion prior to any other indication of fetal distress.

            If the anatomic evidence demonstrated a longer duration of fetal compromise, such as evidence of heart failure including dilatation and loss of natriuretic peptide staining and modest pleural and pericardial effusions with normal lung weights, and evidence of placental changes of fetal vascular malperfusion. If such cases also showed abnormal cord configurations such as occult prolapse or a furcate insertion, then the hypothesis to be tested in a larger group of infants is that the cord lesion is correlated with potentially detectable fetal changes that cost benefit analysis might indicate therapeutic intervention when features of decompensation were beginning. 

            In growth restricted infants, a loss of gas exchange volume (capillary syncytial membranes in the whole placenta might be correlated with a greater tendency to fetal loss if additional components of fetal hypoxia were superimposed including maternal respiratory or supine sleep position. These infants might also show smaller thymus, lipid in the adrenal detected by MRI, or changes in doppler studies of placental/fetal perfusion. The hypothesis to be tested in a larger study is that there is a correlation with the extent of compromised placenta to detectable cord abnormalities and fetal death.  Such a correlation may explain the failure of weekly or biweekly monitoring of growth restricted infants to prevent sudden fetal death, and allow a different approach to infants with cord abnormalities.

            The study cannot anticipate a prior what correlation may be found. The approach is to gather quantitative data from the autopsy that can be related to clinical and experimental fetal physiology to create biologically plausible chains of causation leading to death. Our approach is not to just document placental lesions but to try to see their likely effect on physiology, to give meaning to the idea of placental reserve, and to try to find measures of timing. The problem is always why did this particular growth restricted infant or this infant of a diabetic or obese mother die, not just to correlate fetal death with risk factors. 

            Brain lesions will also be correlated with the above cases. The main goal of stillbirth is to prevent death, a secondary goal is to act therapeutically in a way that also prevents neurologic injury to the fetus. One potential correlation is that a particular cord lesion on MRI or decreased functional placental volume correlates with evidence of a given brain lesion such as white matter gliosis or neuronal apoptosis in brain nuclei. 

            Even if the pilot study does not find correlations of MRI detected cord lesions with maternal history, and placental and autopsy pathology, the informal correlating of quantitative anatomic abnormalities with fetal death may provide insight for further studies. Many previous studies of stillbirth treated the anatomic findings in a more subjective manner. This study will improve the value of anatomic studies. Even with the anticipated 35 autopsies, the variation of events, the differences in gestational age and of the interval of intrauterine postmortem retention, the study will only capture a small percentage of the possibilities. We anticipate finding some basic commonalities based on the mode (timing) and basic mechanisms of death to provide provisional insights to show the value of the study’s approach to the pathologic examination and to the analysis of individual autopsies.

Innovation Statement:

            A systematic study of postmortem MRI scan of a stillborn infant to determine an anatomic mechanism for inferred umbilical cord accidents has not been previously published.

Resources and environment:

            The major resources are those of the Magee Women’s Hospital system with approximately 10,000 deliveries per year. The institution has a strong track record of perinatal research including participation in the NICH MFM network. The MRI facility is capable of performing an abdominal study of a gravid patient. Certified pediatric pathologists oversee the perinatal autopsies and placental examinations. 

            The pathology co-PI (RB) has experience with collaborative clinical pathological studies as PI for the NICHD histology component of the PROM study26,27.  The MFM co-PI

Plans for investigator interactions:

The interactions are inherent in the case conferences that are part of the study protocol. 

Budget:

            Personnel:

                        Radiologist time

                        Research assistant to aid in consent, in maternal questions and in chart                               review

                        Statistical support

            Procedures:

                        MRI technical fees X 10

1.         Varli IH, Petersson K, Bottinga R, et al. The Stockholm classification of stillbirth. Acta Obstet Gynecol Scand 2008;87:1202-12.

2.         Horn LC, Langner A, Stiehl P, Wittekind C, Faber R. Identification of the causes of intrauterine death during 310 consecutive autopsies. Eur J Obstet Gynecol Reprod Biol 2004;113:134-8.

3.         Silver RM, Varner MW, Reddy U, et al. Work-up of stillbirth: a review of the evidence. Am J Obstet Gynecol 2007;196:433-44.

4.         Parast MM, Crum CP, Boyd TK. Placental histologic criteria for umbilical blood flow restriction in unexplained stillbirth. Hum Pathol 2008;39:948-53.

5.         Ryan WD, Trivedi N, Benirschke K, Lacoursiere DY, Parast MM. Placental histologic criteria for diagnosis of cord accident: sensitivity and specificity. Pediatr Dev Pathol 2012;15:275-80.

6.         Bendon RW. Review of autopsies of stillborn infants with retroplacental hematoma or hemorrhage. Pediatr Dev Pathol 2011;14:10-5.

7.         Bendon RW. Review of some causes of stillbirth. Pediatr Dev Pathol 2001;4:517-31.

8.         Crimmins S, Desai A, Block-Abraham D, Berg C, Gembruch U, Baschat AA. A comparison of Doppler and biophysical findings between liveborn and stillborn growth-restricted fetuses. Am J Obstet Gynecol 2014;211:669 e1-10.

9.         Hayes DJL, Warland J, Parast MM, et al. Umbilical cord characteristics and their association with adverse pregnancy outcomes: A systematic review and meta-analysis. PLoS One 2020;15:e0239630.

10.       Bendon RW, Brown SP, Ross MG. In vitro umbilical cord wrapping and torsion: possible cause of umbilical blood flow occlusion. J Matern Fetal Neonatal Med 2014;27:1462-4.

11.       Coblentz AC, Teixeira SR, Mirsky DM, Johnson AM, Feygin T, Victoria T. How to read a fetal magnetic resonance image 101. Pediatr Radiol 2020;50:1810-29.

12.       Cronin RS, Wilson J, Gordon A, et al. Associations between symptoms of sleep-disordered breathing and maternal sleep patterns with late stillbirth: Findings from an individual participant data meta-analysis. PLoS One 2020;15:e0230861.

13.       Cronin RS, Li M, Thompson JMD, et al. An Individual Participant Data Meta-analysis of Maternal Going-to-Sleep Position, Interactions with Fetal Vulnerability, and the Risk of Late Stillbirth. EClinicalMedicine 2019;10:49-57.

14.       O’Brien LM, Warland J, Stacey T, Heazell AEP, Mitchell EA, Consortium S. Maternal sleep practices and stillbirth: Findings from an international case-control study. Birth 2019;46:344-54.

15.       Li M, Thompson JMD, Cronin RS, et al. The Collaborative IPD of Sleep and Stillbirth (Cribss): is maternal going-to-sleep position a risk factor for late stillbirth and does maternal sleep position interact with fetal vulnerability? An individual participant data meta-analysis study protocol. BMJ Open 2018;8:e020323.

16.       Heazell A, Li M, Budd J, et al. Association between maternal sleep practices and late stillbirth – findings from a stillbirth case-control study. BJOG 2018;125:254-62.

17.       McCowan LME, Thompson JMD, Cronin RS, et al. Going to sleep in the supine position is a modifiable risk factor for late pregnancy stillbirth; Findings from the New Zealand multicentre stillbirth case-control study. PLoS One 2017;12:e0179396.

18.       Stacey T, Thompson JM, Mitchell EA, Ekeroma AJ, Zuccollo JM, McCowan LM. Association between maternal sleep practices and risk of late stillbirth: a case-control study. BMJ 2011;342:d3403.

19.       Heazell AEP, Budd J, Li M, et al. Alterations in maternally perceived fetal movement and their association with late stillbirth: findings from the Midland and North of England stillbirth case-control study. BMJ Open 2018;8:e020031.

20.       Olagbuji BN, Igbarumah S, Akintayo AA, Olofinbiyi BA, Aduloju PO, Alao OO. Maternal understanding of fetal movement in third trimester: a means for fetal monitoring and reducing stillbirth. Niger J Clin Pract 2014;17:489-94.

21.       Tveit JV, Saastad E, Stray-Pedersen B, et al. Reduction of late stillbirth with the introduction of fetal movement information and guidelines – a clinical quality improvement. BMC Pregnancy Childbirth 2009;9:32.

22.       Miodovnik M, Skillman CA, Hertzberg V, Harrington DJ, Clark KE. Effect of maternal hyperketonemia in hyperglycemic pregnant ewes and their fetuses. Am J Obstet Gynecol 1986;154:394-401.

23.       Miodovnik M, Lavin JP, Harrington DJ, Leung LS, Seeds AE, Clark KE. Effect of maternal ketoacidemia on the pregnant ewe and the fetus. Am J Obstet Gynecol 1982;144:585-93.

24.       Pacora P, Romero R, Jaiman S, et al. Mechanisms of death in structurally normal stillbirths. J Perinat Med 2019;47:222-40.

25.       Dudley DJ, Goldenberg R, Conway D, et al. A new system for determining the causes of stillbirth. Obstet Gynecol 2010;116:254-60.

26.       Bendon RW, Faye-Petersen O, Pavlova Z, et al. Fetal membrane histology in preterm premature rupture of membranes: comparison to controls, and between antibiotic and placebo treatment. Pediatr Dev Pathol 1999;2:552-0558.

27.       Bendon RW, Faye-Petersen O, Pavlova Z, et al. Histologic features of chorioamnion membrane rupture: development of methodology. Pediatr Pathol Lab Med 1997;17:27-42.

Footnote:

*Email communication from Dr. Teresa Victoria: “OK so I think I understand the project a bit better. You want to measure the length of the cord unwrapped around the fetus and its length to the placenta.

GIven that the fetus is dead. you should have no motion. In this sense you should be able to easily do a 3DS TruFISP sequence that you could then put in Terarecon  and evaluate in the console in 3D. Does this make sense? This should be able to answer the question I think you are asking.”

%d bloggers like this: