|
|
| |||||||||||||||||||||||||||||||||||||||
Background Neonates with severe thrombocytopenia can have bleeding leading to death or lifelong residual defects. The predictors, frequency, and consequences of fetal thrombocytopenia are not known, nor is it known if there are maternal clinical features that could predict fetal thrombocytopenia.
Methods We conducted a seven-year cross-sectional study in which platelet counts were determined in newborns' umbilical-cord blood and blood obtained from their mothers at consecutive deliveries in one obstetrical unit. The relations of the umbilical-cord platelet count to maternal risk factors were determined.
Results Platelet counts were determined in blood samples from 15,471 mothers and 15,932 newborn infants. The cord-blood platelet count was less than 50,000 per cubic millimeter in 19 infants (0.12 percent; 95 percent confidence interval, 0.07 to 0.19 percent), whereas the platelet count was less than 150,000 per cubic millimeter in 6.6 percent of the mothers (95 percent confidence interval, 6.2 to 7.0 percent). One infant among those born to 756 mothers with incidental thrombocytopenia, 5 infants among those born to 1414 mothers with hypertension, and 4 infants among those born to 46 mothers with idiopathic thrombocytopenic purpura had cord-blood platelet counts between 20,000 and 50,000 per cubic millimeter. Only 6 infants (0.04 percent; 95 percent confidence interval, 0.01 to 0.08 percent) had cord-blood platelet counts of less than 20,000 per cubic millimeter; all their mothers were among the 18 whose 19 fetuses were at risk for neonatal alloimmune thrombocytopenia. Two of these infants had in utero intracranial hemorrhage. In addition, 3 infants born to these 18 women had cord-blood platelet counts between 20,000 and 50,000 per cubic millimeter; there was 1 stillbirth due to intracranial hemorrhage.
Conclusions Moderate-to-severe fetal thrombocytopenia is a rare event. The only severely affected neonates with morbidity or mortality due to this condition are those born to mothers with antiplatelet alloantibodies.
Methods
We conducted a prospective surveillance study of all women admitted to our obstetrical unit, which is situated in a tertiary care hospital serving a largely white, middle-class population of approximately 1 million, from January 1986 through December 1992. On admission to the labor floor, all women had a complete blood count, including a platelet count. The mother was considered to have a normal hemostatic system if she did not have a history of thrombocytopenia, a bleeding disorder, or an underlying hypertensive or immunologic disorder. Women who had preterm deliveries (defined as those occurring at less than 37 weeks of gestation), multiple fetuses, and gestational and insulin-dependent diabetes mellitus were included in this group if they were otherwise normal. Women who were considered to have a higher-than-normal risk of hemostatic impairment were classified separately into a group with the hypertensive disorders of pregnancy and a group with the immune disorders idiopathic thrombocytopenic purpura and systemic lupus erythematosus. Women were considered to have idiopathic thrombocytopenic purpura if other causes of thrombocytopenia were excluded. All these women had chronic thrombocytopenia, and many had been treated with corticosteroids or intravenous immune globulin. Many also had platelet-survival tests performed to confirm that the thrombocytopenia was destructive in nature. The diagnosis of systemic lupus erythematosus was made according to the standard criteria of the American Rheumatism Association6. Women with hypertension and multiple fetuses or diabetes were included in the group with hypertensive disorders.
All newborn infants had a cord-blood platelet count performed. In some of these infants, a platelet count was not obtained for technical reasons (for example, the cord-blood sample clotted before anticoagulation), in which case one of the investigators reviewed the chart to ensure that the infant did not have a hemostatic defect. Because the study was conducted as an anonymous survey of samples collected for routine hematologic and obstetrical monitoring, the approval of the hospital ethics committee was not required.
Platelet counts were performed in blood samples collected in tubes containing EDTA with use of a Coulter counter (Coulter Electronics, Hialeah, Fla.). Thrombocytopenia was confirmed by examination of the peripheral-blood film to exclude the possibility that platelet clumping had caused an erroneous diagnosis of thrombocytopenia to be made by the electronic counter (pseudothrombocytopenia).
Maternal thrombocytopenia was considered to be present if the platelet count was less than or equal to 150,000 per cubic millimeter. For purposes of risk assessment, only cord-blood platelet counts less than or equal to 50,000 per cubic millimeter were reported. Infants whose platelet counts were below 50,000 per cubic millimeter underwent ultrasonography to identify any intracranial hemorrhages. Confidence intervals were calculated on the basis of a binomial distribution.
Results
During the seven-year study period, there were 15,663 deliveries in our obstetrical unit. Fifty-six women (0.4 percent) gave birth before or immediately after their arrival at the hospital, and a platelet count was therefore not performed antenatally; these women are not included in our analyses. Platelet counts were performed in blood samples obtained before delivery from the remaining 15,607 women, who delivered 16,068 infants. There were no multiple births in which cord-blood platelet counts were below 50,000 per cubic millimeter. There were 136 stillbirths of singleton infants, giving a total of 15,471 pregnancies available for analysis. Cord-blood platelet counts were obtained for 93.8 percent of the infants born to mothers in the "at-risk" categories (those with incidental thrombocytopenia, hypertension of pregnancy, or a history of idiopathic thrombocytopenic purpura). Platelet counts were obtained for cord-blood samples from an estimated 85 percent of the infants of low-risk mothers, who had normal platelet counts, were healthy, and did not have any evidence of hemostatic impairment. We assumed that infants without cord-blood samples whose neonatal course was uneventful had a cord-blood platelet count above 50,000 per cubic millimeter. Overall, cord-blood platelet counts were determined for 15,932 infants.
Nineteen infants had cord-blood platelet counts below 50,000 per cubic millimeter, giving an overall prevalence of 0.12 percent (95 percent confidence interval, 0.07 to 0.19 percent) (Table 1). Six infants (0.4 percent; 95 percent confidence interval, 0.01 to 0.08 percent) had fewer than 20,000 platelets per cubic millimeter. Ten of the 19 infants with thrombocytopenia were born to mothers who had thrombocytopenia, and 9 were born to mothers who had normal platelet counts.
|
A total of 756 women (73.6 percent of those with thrombocytopenia) had incidental thrombocytopenia of pregnancy (Table 1). Cord-blood samples were obtained from 761 of the infants born to these women (94.2 percent of the total), one of whom had thrombocytopenia (platelet count, 44,000 per cubic millimeter). This infant, who was delivered vaginally, had trisomy 21 and was given a diagnosis of congenital marrow dysfunction. Neonatal cranial ultrasonography did not demonstrate an intracranial hemorrhage.
Among the women with thrombocytopenia, 216 (21.0 percent) had a hypertensive disorder of pregnancy (Table 1). Cord-blood samples were obtained from 226 of their infants (93.8 percent), of whom 4 had thrombocytopenia (platelet counts, 34,000, 36,000, 38,000, and 44,000 per cubic millimeter). A total of 1414 women with hypertension delivered 1495 infants. Of the 1198 hypertensive women without thrombocytopenia, 1 delivered an infant who had a cord-blood platelet count of 40,000 per cubic millimeter. All five of the infants with thrombocytopenia born to hypertensive mothers were delivered before term by cesarean section. One infant had a trisomy 13 and died. Three of the remaining infants had birth weights below the 10th percentile for gestational age, and two of these infants had evidence of intracranial hemorrhage on neonatal ultrasonography. One infant's weight was appropriate for gestational age and there were no sequelae of the thrombocytopenia.
Thirty-nine of the women with thrombocytopenia (3.8 percent) had an immunologic disorder complicating pregnancy; 31 women had idiopathic thrombocytopenic purpura, and 8 women had systemic lupus erythematosus (Table 1). None of the eight infants born to mothers with systemic lupus erythematosus had a cord-blood platelet count below 50,000 per cubic millimeter. Two of the infants born to the 31 mothers with a history of idiopathic thrombocytopenic purpura who had thrombocytopenia at delivery had cord-blood platelet counts of 21,000 and 28,000 per cubic millimeter. During the study, 46 women with idiopathic thrombocytopenic purpura delivered their infants in our institution. Two of the 15 infants born to mothers with idiopathic thrombocytopenic purpura who did not have thrombocytopenia had cord-blood platelet counts of 36,000 and 49,000 per cubic millimeter. Of the four infants of mothers with idiopathic thrombocytopenic purpura who had cord-blood platelet counts of 20,000 to 50,000 per cubic millimeter, three were delivered vaginally and one was delivered by cesarean section for an obstetrical indication. No infant had an adverse outcome, and ultrasonographic examination of the infants with moderate or severe thrombocytopenia revealed no evidence of intracranial hemorrhage.
In 19 pregnancies in 18 women there was a potential for alloimmune neonatal thrombocytopenia. All had maternal-paternal platelet alloantigen discrepancy according to radioimmunoprecipitation analysis, 15 had previously had affected infants, and 3 were sisters of women with alloimmune thrombocytopenia. Three infants with cord-blood platelet counts from 20,000 to 50,000 per cubic millimeter and six infants with counts below 20,000 per cubic millimeter were born to these 18 women (platelet counts, 6,000, 8,000, 9,000, 11,000, 11,000, 13,000, 30,000, 34,000 and 40,000 per cubic millimeter) (Table 1). A 10th infant was stillborn at 35 weeks of gestation, having died of an intracranial hemorrhage confirmed at autopsy. Two of the severely affected live-born infants were first identified on routine ultrasonography as having had in utero intracranial hemorrhage. There were no new neonatal intracranial events identified by neonatal ultrasonography in the nine infants with severe thrombocytopenia due to alloimmunization. Three mothers carrying infants with potential alloimmune thrombocytopenia themselves had thrombocytopenia (platelet counts at delivery, 86,000, 120,000, and 139,000 per cubic millimeter) (Table 1). All three of their infants had cord-blood platelet counts below 20,000 per cubic millimeter (8,000, 9,000, and 11,000 per cubic millimeter). These mothers had both incidental thrombocytopenia of pregnancy and alloimmune thrombocytopenia.
Discussion
Thrombocytopenia in newborn infants, like that in adults, can be caused by underproduction, increased destruction, and sequestration of platelets within an enlarged spleen. Most infants in whom thrombocytopenia develops are ill, are premature, and have other disorders that could contribute to the thrombocytopenia, including disseminated intravascular coagulation and bacteremia. In these sick infants, the frequency of thrombocytopenia is as high as 15 percent, and it is most severe several days after delivery4,5. Thrombocytopenia can also be present at birth. Usually, the thrombocytopenia is caused by maternal factors such as antiplatelet alloantibodies or autoantibodies that have crossed the placenta1,2,3. Concern about the optimal care of these infants has prompted a variety of maneuvers, but there has been little attention to the frequency of thrombocytopenia among otherwise healthy infants and to whether these infants who will have thrombocytopenia can be identified by a maternal characteristic or associated illnesses.
During a seven-year period, we prospectively studied more than 15,000 women and their infants. We previously described the frequency of maternal thrombocytopenia in some of these women7,8,9,10; the main causes of thrombocytopenia were incidental thrombocytopenia of pregnancy (74 percent), hypertensive diseases of pregnancy (21 percent), and immune thrombocytopenic disorders of pregnancy (idiopathic thrombocytopenic purpura and systemic lupus erythematosus) (4 percent). In the current study, our focus was on fetal thrombocytopenia identified by cord-blood sampling. We found that moderate thrombocytopenia (cord-blood platelet count, 20,000 to 50,000 per cubic millimeter) or severe thrombocytopenia (<20,000 per cubic millimeter) is a distinctly uncommon event, occurring in fewer than 0.2 percent of all infants delivered in our institution.
The disorders that cause thrombocytopenia in mothers -- such as incidental thrombocytopenia, idiopathic thrombocytopenic purpura, systemic lupus erythematosus, and the hypertensive disorders of pregnancy -- usually did not cause moderate-to-severe fetal or neonatal thrombocytopenia in their infants. The one group of infants at very high risk were those whose mothers had alloimmune antiplatelet antibodies. These mothers had the infants with the most severe thrombocytopenia; all infants whose cord-blood platelet count was below 20,000 per cubic millimeter were born to the women in this group. Indeed, it was this group that accounted for all fetal morbidity and mortality.
There were three adverse outcomes, all of which were related to bleeding. All three infants had serologically confirmed alloimmune thrombocytopenia. One death occurred in 1987 in utero before aggressive intravenous immune globulin therapy was routinely used. Two infants had intracranial hemorrhages unexpectedly discovered on routine ultrasonography during apparently uncomplicated pregnancies (an occurrence that may reflect the high rate of use of ultrasonography in low-risk obstetrics in Canada). Others have also described antenatal morbidity and mortality in alloimmunized infants11. These two infants were the only ones with alloimmune thrombocytopenia in this study who were not identified by family history (alloimmune thrombocytopenia in either the mother or the mother's sister).
The apparent coincidence that three of the infants with the most severe thrombocytopenia were born to mothers who also had incidental thrombocytopenia of pregnancy is noteworthy. Before delivery, the infants of all three women were known to be at risk for neonatal alloimmune thrombocytopenia caused by platelet-specific alloantibody PLA1 (HPA-1a). Had these women's true status not been known, they might inappropriately have been considered to have idiopathic thrombocytopenic purpura in pregnancy, and a much more severe outcome would have been attributed to the idiopathic thrombocytopenic purpura than actually occurred. Perhaps the association of alloimmune neonatal thrombocytopenia and incidental thrombocytopenia of pregnancy explains some of the reports of severely affected infants born to mothers with mild idiopathic thrombocytopenic purpura.
The incidence of alloimmunization in our infants at term was approximately 1 per 5000 live births. This is in agreement with the results of several retrospective studies12,13 in which the reported incidence ranged from 1 per 5000 to 1 per 10,000 and the one prospective study14 in which a range of values from 1 per 2000 to 1 per 3000 live births was estimated. The incidence of alloimmunization in our study could be an underestimate, since we included only women identified by their infants' cord-blood platelet count of less than 50,000 per cubic millimeter. A more inclusive and comprehensive evaluation of all cord-blood platelet counts below 150,000 per cubic millimeter might have identified more cases of lesser severity.
In most recent studies of idiopathic thrombocytopenic purpura in pregnancy,15,16,17,18 severe fetal thrombocytopenia (in utero) was distinctly uncommon. These studies also indicated that although neonatal morbidity and mortality can occur, some bleeding problems can be prevented with early recognition and therapeutic intervention. Our report provides the obstetrician and internist with population-based estimates of the risk of severe thrombocytopenia in the fetuses and neonates of women with various thrombocytopenic disorders. So long as an obstetrical unit does not serve a population in which there is a high prevalence of unique conditions associated with maternal and possibly infant thrombocytopenia -- such as human immunodeficiency virus infection -- our results should be generalizable to most populations.
The importance of maternal anti-HLA class I antibodies against paternal platelets as a cause of neonatal thrombocytopenia has been debated. In some studies, these antibodies were associated with severe thrombocytopenia, but other investigators have suggested that the optimal techniques to detect platelet-specific alloantibodies were not always used in those studies. We did not address this issue specifically. However, on the basis of one study,19 it can be estimated that 4 to 25 percent of women (depending on parity) form anti-HLA alloantibodies to paternal platelets. Only one of the women we studied had only an anti-HLA alloantibody detected, and her infant did not have a platelet count below 50,000 per cubic millimeter. These observations suggest that anti-HLA alloantibodies are not a common cause of severe neonatal thrombocytopenia.
Moderate-to-severe neonatal thrombocytopenia does occur, but it is uncommon. In our study severe thrombocytopenia was seldom associated with idiopathic thrombocytopenic purpura, incidental thrombocytopenia of pregnancy, or the hypertensive disorders of pregnancy, but it was associated with platelet alloimmunization. These results should not be interpreted to indicate that severe thrombocytopenia can never occur in the infants of mothers who have idiopathic thrombocytopenic purpura or systemic lupus erythematosus. Moreover, the platelet count typically falls after delivery,20 and consequently, infants born to mothers with idiopathic thrombocytopenic purpura and other conditions that place the infant at risk for thrombocytopenia in the peripartum period should be followed carefully to ensure that the platelet count does not fall to dangerous levels.
Supported by the Medical Research Council of Canada.
Source Information
From the Departments of Obstetrics and Gynecology, Medicine, and Pathology, McMaster University Medical Centre, and the Canadian Red Cross Blood Transfusion Service Hamilton Centre, Hamilton, Ont.
Address reprint requests to Dr. Burrows at the Department of Obstetrics and Gynecology, McMaster University Medical Centre, 1200 Main St. W., Hamilton, ON L8N 3Z5, Canada.
References
| |||||||||||||||||||||||||||||||||||||||
Related Letters:
Fetal and Maternal Thrombocytopenia
Fischer J., Dietl J., Goelz R., Burrows R. F., Kelton J. G.
Extract |
Full Text
N Engl J Med 1994;
330:940-941, Mar 31, 1994.
Correspondence
This article has been cited by other articles:
HOME | SUBSCRIBE | SEARCH | CURRENT ISSUE | PAST ISSUES | COLLECTIONS | PRIVACY | TERMS OF USE | HELP | beta.nejm.org Comments and questions? Please contact us. The New England Journal of Medicine is owned, published, and copyrighted © 2009 Massachusetts Medical Society. All rights reserved. |
Previous
