Placental abruption and neonatal anemia

Placental abruption can cause maternal blood loss and maternal anemia. It is less certain whether abruption can cause fetal blood loss and neonatal anemia. Retrospective multi-hospital 24-month analysis of women with placental abruption and their neonates. Of 55,111 births, 678 (1.2%) had confirmed abruption; 83% of these neonates (564) had one or more hemoglobins recorded in the first day. Four-hundred-seventy (83.3%) had a normal hemoglobin (≥5th% reference interval) while 94 (16.7%) had anemia, relative risk 3.26 (95% CI, 2.66–4.01) vs. >360,000 neonates from previous reference interval reports. The relative risk of severe anemia (<1st% interval) was 4.96 (3.44–7.16). When the obstetrician identified the abruption as “small” or “marginal” the risk of anemia was insignificant. Most abruptions do not cause neonatal anemia but approximately 16% do. If an abruption is not documented as small, it is important to surveille the neonate for anemia.


INTRODUCTION
Placental abruption, separation of the placenta from the uterus before delivery, is reported to occur in 0.4-1.0% of births [1]. In the United States and Canada the incidence is at the higher end of this range and in Scandinavia it is at the lower end [2]. Consequences of abruption on mother's health include blood loss sufficient to cause her hypovolemic shock, disseminated intravascular coagulopathy and may require hysterectomy [3]. Fetal and neonatal sequelae of abruption include hypoxia, ischemia, acidosis, and death [1,2,4,5]. However, whether placental abruption can also cause fetal blood loss has been debated [4][5][6][7][8][9][10][11]. Some researchers have concluded that severe anemia at birth may result from occult fetomaternal hemorrhage coincident with placental abruption [9,10,12]. Indeed, abnormal placentation and placental pathology have been implicated in fetomaternal hemorrhage, possibly as a result of a leaky placental interface [10]. Others contend that placental separation from the uterus in placental abruption is primarily caused by maternal bleeding, leading to compromised oxygen delivery in the fetus but not fetal hemorrhage [1,2,5,6].
Certainly, placental abruption can result in massive maternal blood loss. The World Health Organization lists severe bleeding as the most common cause of postpartum maternal death, and placental abruption as a principal cause of severe postpartum hemorrhage [13]. Many or most abruptions are thought to result from rupture of a maternal decidual artery, causing dissection of maternal blood under arterial pressure at the decidual-placental interface. Thus, the hemorrhage within an abrupting placenta, dissecting and separating it from the uterus, primarily involves maternal blood and the maternal portion of the maternal/fetal circulatory interface.
Likewise, it is certain that abruption can result in fetal morbidity and mortality, but not usually because of massive fetal blood loss.
Abruption-associated fetal injury is typically damage to the brain and other organs from oxygen deprivation [14]. In fact, using the National Inpatient Sample datasets from 2010 to 2018, Acun et al. found that one of the factors most associated with neonatal hypoxic-ischemic encephalopathy was placental abruption (odds ratio, 101; 95% confidence interval 91-112) [14]. Similarly, Parc et al. performed a nested case control study of mothers with placental abruption, and found 29% (44/152) of neonates experienced HIE or death [15]. Therefore, placental abruption typically causes maternal morbidity due to blood loss, but the fetal morbidity it causes is typically from hypoxia/ischemia. However, sometimes a neonate has anemia at birth with no explanation for the anemia other than placental abruption [16]. In such situations it is unclear whether the abruption resulted in fetal blood loss and caused the neonatal anemia.
One method to determine whether abruption can cause neonatal anemia is to identify mothers with a diagnosis of abruption, then examine hemoglobin measurements of their neonates. Such a study was recently reported by Bruinsma et al. [6] who identified placental abruption in 65 mothers (65 neonates) from one hospital over a period of six years. Sixty-eight percent of those women had a partial abruption, 9% had a large abruption, and in 23% the severity was unknown. Fifty-five of the 65 neonates had one or more hemoglobin measurements during their first 24 h, and the mean value was 16.5 g/dL. In 14/65 (26%) fetal blood loss was suspected by the pediatricians based on paleness, tachypnea, and hypotension of the newborn, however anemia was not confirmed because the hemoglobin was normal at birth (not below the 5th percentile lower reference interval). Two had anemia diagnosed by a low hemoglobin, and six (11%, including four with a normal hemoglobin level) received a red blood cell transfusion. The authors concluded that hemoglobin levels of neonates born after abruption are typically within the normal reference interval, and that neonatal anemia is not expected after placental abruption [6].
Another way to investigate whether abruption can cause neonatal anemia is to study neonates who have severe anemia diagnosed within a few hours after birth and determine how many of those had a placental abruption identified as the only recognized cause of that anemia. We recently took that approach [16], reporting that during a 10-year period in our multi-hospital healthcare system, 344 neonates were identified with severe anemia in the first 6 h after birth. A plausible cause for the anemia was identified in only 176/344 (51%) neonates. Of these, 129 (37%) were born to mothers with placental abruption and no other cause for the neonatal anemia was identified. We concluded that severe anemia at birth is rare, but when found, placental abruption might be one of the more common causes [16].
Given the conflicting results of these two studies, we performed a new retrospective analysis, using the approach of the Bruinsma study [6]. We obtained a large and recent sample, using the multihospital Intermountain Healthcare databases during the most recent 24 months. We identified every case of placental abruption diagnosed during the period July 1, 2020 through June 30, 2022. Our aims were, in each documented case, to; (1) record the size of the abruption, if available, by whatever metric each delivering obstetrician employed, (2) record all hemoglobin measurements made on the neonates born to these women during their first 24 h after birth and compare these values with gestational-age matched controls.

METHODS
The Institutional Review Board (IRB) of Intermountain Healthcare reviewed this proposal and approved it as exempt from the need for individual informed consent. We did not collect patient names, photographs, Social Security numbers, addresses, phone numbers, or zip codes as part of the process.
Our study was a retrospective records analysis of neonates in the Legacy Intermountain Healthcare databases. Intermountain Healthcare is a not-forprofit organization that owns and manages hospitals in the Intermountain West of the USA. The diagnosis of placental abruption was initially screened from electronic data marts; Case Mix (the billing, coding, and financial data mart used by Intermountain Healthcare), EVOX (the extended Vermont-Oxford database), Storkbytes (the labor and delivery database), Fetal Link (a replacement for Storkbytes), provider problem lists, and ICD-10 coding.
The electronic medical record of every delivery that was identified by our screening ascertainment methods was individually reviewed by a member of the research team, not relying on coded information or on data tables. This was done to verify that a diagnosis of placental abruption was indeed made, either from notes by the delivering obstetrician or from reports by a pathologist examining the placenta. Patients were included as having a confirmed placental abruption if the obstetrician documented abruption in their procedure note, or if the pathologist confirmed abruption in their report. We characterized each confirmed abruption using whatever descriptive metric we found in the records. Other information we recorded on all women with a confirmed diagnosis of placental abruption included; age, gravity, parity, gestational age at delivery, and mode of delivery.
The blood hemoglobin concentration was measured at birth either from the umbilical cord blood, generally venous, or the first neonatal blood draw, which may have been capillary, arterial, or venous. If there was only an umbilical arterial cord blood hemoglobin value, this was recorded as the initial hemoglobin. Anemia was defined by a hemoglobin concentration plotting below the 5th percentile lower reference interval for gestational age on the day of birth [16,17]. Moderately severe anemia was defined as a hemoglobin below the 5th but above the 1st percentile lower reference interval, and severe anemia by a hemoglobin below the 1st percentile.
The dataset was collected and managed using an Intermountain Healthcare REDCap (Research Electronic Data Capture) electronic data capture tool. REDCap is a secure cloud-based application designed to support data capture and provide an intuitive interface for validated data entry, audit trails for tracking data manipulation and export procedures, automated export procedures for seamless data downloads to common statistical packages, and procedures for importing data from external sources. Summary statistics (means, counts, and proportions) were the primary quantitative tools used for analysis. Differences in continuous variables by group were assessed using one-way ANOVA. Differences in categorical variables were assessed using either chi-square tests or Fisher's exact test. Data management and statistical analysis were done in the R language and environment for statistical computing (R Foundation).

RESULTS
During the 24-month study period 55,111 births were recorded at the Intermountain Healthcare hospitals. In 1050 of these we identified "abruption" somewhere in the medical record using our various ascertainment methods. Individual chart reviews confirmed a placental abruption preceding the birth of 678 neonates (Fig. 1). The remaining 372 neonates generally had "possible abruption" or "suspected abruption" listed prenatally with vaginal spotting or bleeding, but at delivery no signs of abruption were identified. Thus, the incidence of confirmed placental abruption in this series was 1.2% (678 neonates born after abruption/55 111 births).
Demographic features of the 678 neonates born after confirmed abruption are listed in Table 1. These are shown along-side two comparison groups, (1) the 372 where abruption was excluded by individual chart reviews and (2) the remaining neonates born at Intermountain Healthcare hospitals during this 24-month period. The control group (neonates where abruption was neither identified electronically nor confirmed by chart review) had higher birthweights, older gestational ages at birth, a lower proportion born to primiparous mothers, and a lower rate of delivery by cesarean section. Those neonates where abruption was electronically identified but excluded by chart review also had slightly higher birthweights, older gestational ages at birth, a lower proportion born to primiparous mothers, and a lower rate of delivery by cesarean section; but these differences were not statistically significant.
Of the 678 neonates born after abruption, 83.1% (564) had one or more blood hemoglobin values recorded during the first 24 h after birth. Among the 114 who did not have a hemoglobin recorded, 45 were fetal or very early neonatal deaths with no laboratory tests obtained. The other 69 neonates were apparently well, because all were cared for in a newborn nursery with no hemoglobin or hematocrit values obtained. These neonates had an average gestational age of 37.8 weeks (SD = 2.03), average birth weight of 2981 g (SD = 618 g) and 63% had the abruption Fig. 1 Consort diagram depicting neonates who were included in this analysis.
classified as small, marginal or <30%. In the infants who had a hemoglobin reported we found no difference in birth weight or gestational age between those with a normal hemoglobin, vs. those with anemia, vs. those with severe anemia. Table 2 shows the demographics for infants who were identified with moderate and severe anemia, including those who were transfused and those who were not transfused. In moderately anemic patients, those who received a transfusion had lower gestational ages, lower birth weights, lower average hemoglobin values and lower Apgar scores compared to neonates who were not transfused. Figure 2A shows the lowest blood hemoglobin level recorded during the first 24 h after birth in each of these 564 neonates. The majority of these neonates (n = 470, 83.3%) did not have anemia. However, 94 (16.7%) did. Of those with anemia 65 (69.1%) had a hemoglobin below the 5th percentile but above the 1st percentile, which we labeled "moderately severe anemia", and 29 (30.9%) had a hemoglobin below the 1st percentile, which we labeled "severe anemia". Eighteen of the 29 with severe anemia received a red blood cell transfusion within the first 30 h after birth. Thus, the relative risk of neonatal anemia (<5th percentile) after placental abruption was 3.33 (95% CI, 2.77-4.02) when compared with over 360,000 neonates from our previous reference interval report [17] and the relative risk of severe anemia (<1st percentile) was 5.14 (3.60-7.34).
Two hundred eighty-one of the neonates born after confirmed abruption (41.4%) had the abruption size described as "small", "marginal", or a percentage ≤30% of the maternal surface covered with an adherent thrombus. When we plotted the lowest blood hemoglobin level recorded during their first 24 h from these neonates (Fig. 2B) we found the relative risk of anemia, and of severe anemia, were not greater than in the reference population. A "large" or "complete" abruption was described in 23 neonates born after abruption (3.4%). Most of these (n = 17) died without ever having a hemoglobin drawn; six of these were immediate neonatal deaths and 11 were fetal deaths. Only six with a large abruption lived; all six had a hemoglobin below the 50th percentile for gestational age, two had a hemoglobin below the 5th percentile and were transfused and four had a hemoglobin between the 5th and 50th percentile. One of the six survivors after a large abruption was diagnosed with hypoxic ischemic encephalopathy (HIE). This compares with HIE diagnosed in one of the 269 who survived after a small abruption, and two of 328 who survived after no abruption. The relative risk of anemia in the group with a large abruption was 5.7 (95% CI, 1. 8-18.4). There were 374 (55.2%) neonates who had no metric listed relative to the size of the abruption. The relative risk of anemia in this group was 4.9 (4.0-5.9).
Repeat hemoglobin checks, over the 10 h following birth, are shown in Fig. 3. These repeat hemoglobin values were not obtained by protocol but were likely ordered as a result of concerns of the clinicians caring for the neonate. We identified serial hemoglobin checks among 78 neonates, who did not have a confirmed abruption, and 58 among those who did. Regressions lines tended to increase in those who did not have an abruption and to decrease in those who did.

DISCUSSION
In the past five editions of the textbook Wintrobe's Clinical Hematology, we included the following statement about neonatal anemia following placental abruption, "Although the majority of the blood loss is maternal, loss of fetal blood can also occur, thus in neonates born after abruption it is important to monitor blood pressure, hemoglobin/hematocrit, and tissue perfusion" [18].
Using our multihospital data, we found that placental abruption usually (84% of cases) did not result in neonatal anemia. This is concordant with the report of Bruinsma et al. [6]. In fact, when our delivering physicians recorded that the abruption was small, the risk of neonatal anemia was nil; meaning it was the same risk as in the general population for that gestational age. However, we found that 16% of neonates born after abruption had anemia, defined by a hemoglobin below the 5th percentile lower reference interval for gestational age [17]. Moreover 1/3rd of neonates with anemia after abruption had severe anemia, defined by a hemoglobin below the 1st percentile lower reference interval, and the majority of those neonates received emergent transfusions. We identified some neonates born after abruption who initially had a normal hemoglobin level, which fell during the hours immediately after birth. In healthy neonates the hemoglobin typically rises slightly in the hours following birth, as a result of placental transfusion and fluid shifts out of the vasculature concentrating the hemoglobin level [7]. One explanation for a falling hemoglobin in a neonate after abruption is fetal hemorrhage immediately before birth. Such a hemorrhage can result in pallor, tachypnea, tachycardia and hypotension but with a normal hemoglobin level initially, which falls over the subsequent hours as fluid is drawn into the vasculature compensating for hypovolemia and diluting the hemoglobin level [7]. We speculate that many or all of the neonates we identified with a falling hemoglobin in the hours after birth illustrate this dilutional drop in hemoglobin associated with fetal blood loss just before birth. This phenomenon might also explain why six of the neonates in the Bruinsma study who had a normal hemoglobin value at birth were transfused for clinical signs of anemia [6]. representing the lowest hemoglobin level recorded in the first day for each neonate (n = 286) born after an abruption that was described as "small", "marginal", or "<30% abruption". It may seem curious that only 15 of the 28 neonates with severe anemia at birth received a red blood cell transfusion in the first day after birth. We defined severe anemia as a hemoglobin value plotting below the 1st percentile lower reference interval for gestational age [16]. This proportion receiving a transfusion (15/ 28) seems low, however it is consistent with a recent report by our group that only 56% of neonates with severe anemia at birth had anemia recognized by caregivers and recorded in the medical record [16]. Thus, it is likely that many cases of severe anemia at birth in our present study were not recognized by the clinical team as being anemic. We are still working on reporting to clinicians the exact percentile at which each hemoglobin plots, so they will better recognize severe anemia at birth and consider transfusion if it is severe and falls within our transfusion guidelines [17].
There are several limitations of our study. First, it is a retrospective analysis relying on documentation and manual chart review to collect important information. To confirm an abruption, we required the obstetrician to document abruption in a procedure note or a pathologist to state abruption in their report. The accuracy of determining whether an abruption is present can be poor before delivery or before the pathologist's examination. Additionally, we based the size of abruption on whatever metric was documented (descriptive, percent estimation), which is subjective. Moreover, no size was recorded in 55% (374/678) of these deliveries.
Another limitation involves retrospective recording of hemoglobin values. Most often these were from the umbilical cord venous blood because this was the most consistently available measurement. However, if this was not available, we recorded the umbilical arterial, value or the neonatal arterial, venous, or capillary hemoglobin measurement. We did not collect information on Kleihauer Betke testing as this was not routinely done for patients with abruption but might have been useful information. Finally, we did not collect information on whether mothers were diagnosed with anemia or required oral or IV iron, which might have provided information on the neonate's iron status.
In conclusion, our multi-hospital dataset analysis suggests to us that the report by Bruinsma et al. is correct, in that anemia is generally not seen among neonates following a placental abruption [6]. However, we caution that anemia, indeed severe anemia, can occur after abruption and might not be identified by the initial hemoglobin level. The relative risk of anemia in neonates born after placental abruption was 3.76 and was 5.7 if the abruption was described as large or complete. We advise that unless the delivering clinician documents that the abruption is small, and unless the neonate appears completely well with no signs of hypovolemia, surveillance for the possibility of significant fetal blood loss should be routinely instituted during the hours after birth.