- Email: [email protected]
The feasibility of using histologic placental sections to predict newborn nucleated red blood cell counts
The Feasibility of Using Histologic Placental Sections to Predict Newborn Nucleated Red Blood Cell Counts William M. Curtin, MD, Bahig M. Shehata, MD, Sadik A. Khuder, MPH, PhD, Haynes B. Robinson, MD, and Brian C. Brost, MD OBJECTIVE: To determine the feasibility of using calculated nucleated red blood cell (RBC) counts from histologic placental slides to predict newborn nucleated RBC counts. METHODS: This retrospective study compared absolute nucleated RBC counts from 24 newborns, diagnosed with fetal distress in labor, with counts calculated from their histologic placental slides. A simple linear regression model was tested with newborn nucleated RBC counts as the dependent variable and calculated placental nucleated RBC counts as the independent variable. RESULTS: The mean ⴞ standard deviation newborn nucleated RBC count was 4.81 ⴛ 109 ⴞ 5.46 ⴛ 109/L compared with 1.37 ⴛ 109 ⴞ 1.78 ⴛ 109/L calculated from placental sections. These data were normalized by logarithmic transformation. A significant linear regression was obtained, r2 ⴝ 0.74, P < .001. The prediction equation obtained was natural logarithm (newborn nucleated RBC count) is equal to 1.002 ⴛ natural logarithm (placental nucleated RBC count) ⴙ 1.173. CONCLUSION: It is feasible to calculate nucleated RBC counts from histologic slides of the placenta that are predictive of newborn nucleated RBC counts. Further work on more homogeneous groups of subjects is necessary to increase the precision of the method. The placenta could serve as a surrogate source for newborn whole blood nucleated RBC counts around the time of birth. (Obstet Gynecol 2002;110:305–10. © 2002 by The American College of Obstetricians and Gynecologists.)
The predominant erythrocyte in the embryo is the nucleated red blood cell (RBC), which declines progressively with advancing gestational age, such that by 12 weeks it comprises less than 2% of circulating red cells.1 Green and Mimouni2 have reported mean absolute nucleated RBC counts for normal term infants of 0.4 ⫻ 109/L and considered values of 1 ⫻ 109/L to be a possible From the Division of Maternal-Fetal Medicine, Department of Obstetrics and Gynecology, University of Rochester Medical Center, Rochester, New York; Department of Pathology, Toledo Children’s Hospital, Toledo, Ohio; and Departments of Medicine and Obstetrics and Gynecology, Medical College of Ohio, Toledo, Ohio.
indicator of intrauterine hypoxia. Values for normal preterm infants are more difficult to interpret and are confounded by the etiology of the preterm delivery. However, fetal blood sampling from normal preterm fetuses demonstrates higher percentages of nucleated RBCs as compared with term fetuses.3 From the data of Forestier et al,3 an estimated mean absolute nucleated RBC count in normal preterm fetuses would be in the range of 0.8 ⫻ 109/L to 1.1 ⫻ 109/L. Increased newborn nucleated RBC counts can be found in maternal diabetes2 and intrauterine growth restriction,4 as well as intrauterine hypoxia.5 Increased numbers of nucleated RBCs in newborns are associated with intrauterine fetal asphyxia and perinatal brain damage.5–7 Buonocore et al7 found significantly higher nucleated RBC counts at birth in three 1-year-old children with abnormal neurodevelopmental status when compared with a normal cohort. Maier et al8 showed a significant correlation between umbilical venous blood erythropoietin concentration at birth and nucleated RBC count in peripheral venous blood. They related the histologic characteristics of meconium phagocytosis in the placenta to the timing of hypoxia and concluded that fetal erythropoietin, produced after a few hours delay in response to a hypoxic stimulus, leads to increased erythropoiesis (manifested by nucleated RBCs). Nucleated RBCs can be readily visualized by microscopic examination of the placenta (Figure 1). The manner in which increased nucleated RBCs are reported is variable among pathologists. The College of American Pathologists recommends descriptive reporting for the presence of increased nucleated RBCs within placental vessels.9 Semiquantitative methods have been used in research studies with nucleated RBCs reported as mildly or moderately increased in number.10 One survey of pathologists on placental examination reported that 38% did not routinely assess for increased nucleated RBCs within placental vessels.11 Theoretically, nucleated RBCs seen within placental vessels reflect the number in
VOL. 110, NO. 2, AUGUST 2002 © 2002 by The American College of Obstetricians and Gynecologists. Published by Elsevier Science Inc.
0029-7844/02/$22.00 PII S0029-7844(02)02041-0
305
Figure 1. A) Nucleated red blood cells within cross-section of umbilical cord blood vessel (hematoxylin-eosin, ⫻ 400, original magnification). B) High-power view of a single nucleated red blood cell (hematoxylin-eosin oil immersion, ⫻ 1000, original magnification). Curtin. Placental Nucleated RBC. Obstet Gynecol 2002.
newborn blood at the time of birth. The purpose of this study was to determine the feasibility of using histologic sections of the placenta to perform quantitative nucleated RBC counts, and to determine if a strong enough correlation existed with newborn nucleated RBC counts such that the placental counts might predict newborn counts. MATERIALS AND METHODS This retrospective study used the medical records database at The Toledo Hospital. A search by International Classification of Diseases, 9th Revision, code 656.31 (fetal distress), during the 2-year period, 1998 and 1999, demonstrated 615 maternal records. Twenty-four cases were found for the study fulfilling the following criteria: record of newborn blood count within 2 hours of birth demonstrating nucleated RBCs and archived histologic slides of the placenta. The maternal data collected included age, parity, medical or obstetric complication, and route of delivery. Data collected on the fetus/new-
306
Curtin et al
Placental Nucleated RBC
born included: gestational age, type of abnormal fetal heart rate pattern, time from onset of abnormal fetal heart rate pattern to delivery, birth weight, time of birth, time from birth to newborn blood count, Apgar scores, and umbilical cord blood gas results. Complete blood counts from the 24 newborns were performed on the Coulter STKS (Beckman Coulter, Inc., Fullerton, CA). Absolute nucleated RBC count expressed in units of ⫻109/L was extracted from the newborn complete blood count report. Many studies report nucleated RBCs as a ratio per 100 white blood cells. However, newborn white blood cell counts are variable; thus, nucleated RBC counts could be misleadingly low or high if reported relative to white cells.2,12 For this reason, as well as to simplify placental counting, we chose to work with absolute numbers of nucleated RBCs. A single board certified pathologist (WMC) reviewed the histologic slides of the placentas without knowledge of the newborn nucleated RBC counts. Umbilical vessels in which the lumen was mostly full of blood were the preferred specimens for counting and volume calculations. In three cases, chorionic plate vessels were used because the umbilical cord vessels provided an insufficient specimen. Internal diameters of all selected blood vessels, measured using a mm ruler, averaged approximately 1.5 mm. The thickness of the histologic section was assumed to be a constant 4 m. The volume of blood contained in the vessels was estimated using the formula for either a cylinder (V ⫽ 兿r2h) or an ellipsoid (V ⫽ 4/3 兿abc), depending on the shape of the blood vessel. Systematic counting of nucleated RBCs was performed within the umbilical or chorionic plate vessels under oil immersion at a magnification of 1000 using an Olympus CH2 microscope (Olympus Optical Co., Ltd., Tokyo, Japan). Placental nucleated RBC counts were converted to #/L to compare with the newborn blood. Nucleated RBC counts, even from large groups of newborns, are characteristically skewed, not normally distributed data.13 Transformation of the nucleated RBC counts to natural logarithms was necessary to normalize the nucleated RBC data to fit a linear regression model. The hypothesis of normality of the natural log-transformed nucleated RBC counts was not rejected using the Kolmogorov-Smirnov and Shapiro-Wilk tests. Further support for normally distributed data was evident by graphic displays of the data: histograms, boxplots, normal, and detrended normal probability plots. Statistical analysis was performed using SPSS 9.0 for Windows (SPSS Inc., Chicago, IL). A simple linear regression model was tested with natural logarithm (newborn nucleated RBC count) as the dependent variable and natural logarithm (placental nucleated RBC
OBSTETRICS & GYNECOLOGY
Table 1. Labor and Birth Characteristics of 24 Newborns With Fetal Distress Gestational age (wk) Mean ⫾ SD Range Delivery by cesarean, n (%) Birth weight (g) Mean ⫾ SD Range AGA, n (%) LGA SGA 1-min Apgar score ⬍4, n (%) 5-min Apgar score ⬍7, n (%) Umbilical artery cord gases Mean ⫾ SD (Range) pH PCO2 (mm Hg) PO2 (mm Hg) Base excess (mmol/L) Start of abnormal FHR pattern to delivery (min) Mean ⫾ SD Range Time interval, birth to newborn blood count (min) Mean ⫾ SD Range
35.0 ⫾ 5.0 24–42 18 (75) 2508 ⫾ 1055 612–3950 20 (83) 3 (13) 1 (4) 12 (50) 9 (37.5) 7.13 ⫾ 0.12 (6.89–7.33) 68.7 ⫾ 13.1 (49–98) 11.5 ⫾ 6.8 (3–26) ⫺8.3 ⫾ 6.2 (⫺28–0.20) 254.7 ⫾ 219.7 35–878 53.5 ⫾ 30.9 8–104
SD ⫽ standard deviation; AGA ⫽ appropriate for gestational age; LGA ⫽ large for gestational age; SGA ⫽ small for gestational age; FHR ⫽ fetal heart rate.
count) as the independent variable. Significance was set at ␣ ⱕ .05. RESULTS The antepartum maternal complications of the 24 newborns with a diagnosis of fetal distress in labor were the following: type 1 diabetes (four), severe preeclampsia (five), both type 1 diabetes and severe preeclampsia (two), choriamnionitis (four), suspected intrauterine growth restriction (one), preterm rupture of membranes (one), placenta previa with bleeding (one), intrapartum bleeding (two), smoking and late prenatal care (one), prior intrauterine fetal demise (one), and none (two). The predominant abnormal fetal heart rate patterns noted were variable decelerations (ten), late decelerations (eight), tachycardia (two), bradycardia (one), only
absent variability (one), and no abnormality (two; both newborns required resuscitation). Selected pregnancy, labor, and birth characteristics are given in Table 1. The mean absolute nucleated red cell counts from newborn blood samples and histologic sections of the placenta are shown in Table 2. The mean and standard deviation nucleated RBC count for the 24 newborns was 4.81 ⫻ 109 ⫾ 5.46 ⫻ 109/L as compared with 1.37 ⫻ 109 ⫾ 1.78 ⫻ 109/L calculated from their placentas. Because the distribution of nucleated red cell concentrations was skewed, transformation of the data to natural logarithms was performed resulting in a normal distribution. The natural logarithmic transformation of the newborn and placental nucleated RBC counts demonstrated a strong correlation (Pearson r ⫽ 0.862, P ⬍ .001). A power calculation demonstrated, with a sample size of 24 and the correlation of 0.862, there was a greater than 90% probability that the correlation was not zero at the .001 level of statistical significance. A simple linear regression was calculated predicting natural logarithm (newborn nucleated RBC count) based on their calculated natural logarithm (placental nucleated RBC count). A significant linear regression was found, P ⬍ .001, with an r2 of 0.743. Seventy-four percent of the variation in newborn nucleated RBC counts was explained by the calculated placental nucleated RBC counts. The prediction equation obtained was natural logarithm (newborn nucleated RBC count) was equal to 1.002 ⫻ natural logarithm (placental nucleated RBC count) ⫹ 1.173 (Figure 2).
DISCUSSION The presence of nucleated RBCs within placental vessels should be assessed and reported upon routinely by the pathologist.9 This subjective impression of nucleated RBCs is dependent on the degree of effort expended; they can be overlooked. Methods used to report the presence of nucleated RBCs in placental examinations are qualitative9 or semiquantitative.10 The method described in this study of calculating a nucleated RBC count from placental slides allows one to quantitate the numbers of nucleated RBCs and express the results in a
Table 2. Correlation of Newborn Nucleated Red Blood Cell Counts (Number of Nucleated Red Blood Cells ⫻ 109/L) With Placental Nucleated Red Blood Cell Counts (n ⫽ 24)
Nucleated RBC*
Newborn
Placenta
Correlation coefficient (r)
P
4.81 ⫾ 5.46 (0.44–20.0)
1.37 ⫾ 1.78 (0.13–8.47)
.862
⬍.001†
RBC ⫽ red blood cell. * Mean ⫾ standard deviation (range). † Pearson correlation coefficient on natural log-transformed data.
VOL. 110, NO. 2, AUGUST 2002
Curtin et al
Placental Nucleated RBC
307
Figure 2. Scattergram with least squares line: natural logarithm (newborn nucleated red blood cell) vs natural logarithm (placental nucleated red blood cell). Curtin. Placental Nucleated RBC. Obstet Gynecol 2002.
manner that allows direct comparison with a neonatal whole blood sample. Calculation of nucleated RBC counts from placental slides would be unnecessary in the case where a newborn complete blood count is obtained in a timely manner after birth. In normal newborns, nucleated RBCs are cleared from the blood such that their numbers at 6 hours after birth are approximately half of the number within the first hour after birth.14 However, most newborns who appear to be doing well after birth do not get routine complete blood counts. Furthermore, most children subsequently found to have cerebral palsy are asymptomatic in the newborn period.15 The indications for placental examination, on the other hand, are fairly broad,9 and about 30% of institutions might be expected to submit 25% or more of all their delivered placentas.11 There would be expected to be many instances in which placental slides would be available, but a newborn complete blood count was not obtained. It is this latter circumstance in which the placental slides could be useful to estimate newborn nucleated RBC counts. The placental counts in this study did substantially underestimate the newborn nucleated RBC counts; however, the strong positive correlation between the two counts fit a linear regression model and thus allowed us to use the relationship in prediction of newborn nucleated red cell counts. The placentas examined in this study were prepared according to a standard institutional protocol.9 It may be possible through selective
308
Curtin et al
Placental Nucleated RBC
preparation of umbilical cord segments to improve the accuracy of the histologic placental counts in relation to newborn or cord blood nucleated RBC counts. This study is limited by the variability inherent in nucleated RBC counts on newborns. The standard deviations in both the newborn and placental counts were high, and this limits the precision of prediction intervals for newborn nucleated red cells from placental counts. For example, a predicted newborn nucleated RBC count of 3.9 ⫻ 109/L would be associated with a wide 95% prediction interval of 1.1 ⫻ 109/L to 14.0 ⫻ 109/L. Current laboratory measurement of whole blood nucleated red cell counts by manual differential cell count may be a source of variability, and increased precision may be possible with direct measurements by newer hematology analyzers.16 This study was retrospective and designed to make use of preexisting placental slides and complete blood counts from newborns. An ideal study would have been prospective comparing nucleated RBC counts from placental slides with nucleated RBC counts on blood obtained immediately after birth either from the umbilical vein or the newborn itself. The mean time from birth to the attainment of a newborn blood sample of 53.5 minutes in this study was not excessively long but could account for some of the variability of the data, as well as underestimation of the placental nucleated RBC counts. The results of this study would be generalizable only to a similar population of acidemic newborns. The equa-
OBSTETRICS & GYNECOLOGY
tion for predicting newborn nucleated RBC counts from placental slides is applicable only at our institution. However, pathologists could use our methods to generate their own prediction equations that would be specific to their institution and populations. The sample size was small and did not allow for statistical analysis of subgroups. Evaluation of residual statistics in our study suggested that the term pregnancy with no underlying maternal disease was associated with the least error in our linear regression model. Further study needs to be performed on more homogeneous groups with stratification based on gestational age, fetal size, and maternal disease states. Intraobserver and interobserver reliability of this technique remains to be determined on a large sample. There has been considerable interest in elevated neonatal nucleated RBCs and their association with perinatal asphyxia.5–7,14 Fox noted that increased nucleated RBCs in blood vessels in histologic sections of the placenta were associated with fetal hypoxia.17 Altshuler and Herman18 confirmed the association of the placental finding of increased nucleated RBCs with fetal hypoxia in an epidemiologic study. In a case control study of 892 infants hospitalized in the neonatal intensive care unit, the placentas of 193 neonates with a diagnosis of perinatal asphyxia were compared with 699 control neonates. The presence of nucleated RBCs, adjusted for ten other placental lesions, gestational age, and birth weight, was a significant independent predictor of perinatal asphyxia (odds ratio 3.00, 95% confidence interval 1.7, 5.4). Placental pathology is often used as evidence in litigation involving neurologically impaired children.19,20 Our method of quantitation of nucleated RBCs from placental slides may be of medicolegal use. Because the placental nucleated RBC counts are reflective of those in the newborn at the time of birth, the placenta could be used as a surrogate for neonatal nucleated RBC counts. This would be particularly valuable in cases where a neonatal blood sample is either not obtained or poorly timed. The quantification of counts expressed as nucleated RBC/L provides more meaningful and objective information than do those done descriptively or on a semiquantitative scale. However, increased nucleated RBCs, either in the placenta or on newborn blood samples, should never be used in isolation to diagnose perinatal asphyxia.12 In summary, we have described a method that directly relates nucleated RBCs identified from placental slides to newborn whole blood nucleated RBCs. This study shows that it is feasible to calculate nucleated RBC counts from placental sections and that these counts could be used to predict newborn nucleated RBC counts from whole blood shortly after birth. Further work needs
VOL. 110, NO. 2, AUGUST 2002
to be done on more homogeneous groups of subjects to increase the precision of the method. Histologic slides of the placenta could be a useful independent data source for quantitation of fetal nucleated RBCs at the time of delivery. REFERENCES 1. Salafia CM, Weigl CA, Foye GJ. Correlation of placental erythrocyte morphology and gestational age. Pediatr Pathol 1988;8:495–502. 2. Green DW, Mimouni F. Nucleated erythrocytes in healthy infants and in infants of diabetic mothers. J Pediatr 1990;116:129 –31. 3. Forestier F, Daffos F, Catherine N, Renard M, Andreux JP. Developmental hematopoiesis in normal human fetal blood. Blood 1991;11:2360 –3. 4. Baschat AA, Gembruch U, Reiss I, Gortner L, Harman CR, Weiner CP. Neonatal nucleated red blood cell counts in growth-restricted fetuses: Relationship to arterial and venous Doppler studies. Am J Obstet Gynecol 1999;181: 190 –5. 5. Korst LM, Phelan JP, Ahn MO, Martin GI. Nucleated red blood cells: An update on the marker for fetal asphyxia. Am J Obstet Gynecol 1996;175:843– 6. 6. Phelan JP, Ahn MO, Korst LM, Martin GI. Nucleated red blood cells: A marker for fetal asphyxia? Am J Obstet Gynecol 1995;173:1380 – 4. 7. Buonocore G, Perrone S, Gioia D, Gatti MG, Massafra C, Agosta R, et al. Nucleated red blood cell count at birth as an index of perinatal brain damage. Am J Obstet Gynecol 1999;181:1500 –5. 8. Maier RF, Gu¨nther A, Vogel M, Dudenhausen JW, Obladen M. Umbilical venous erythropoietin and umbilical arterial pH in relation to morphologic placental abnormalities. Obstet Gynecol 1994;84:81–7. 9. Development Task Force of the College of American Pathologists. Placental Pathology Practice Guidelines. Practice guidelines for examination of the placenta. Arch Pathol Lab Med 1997;121:449 –76. 10. Redline RW, O’Riordan MA. Placental lesions associated with cerebral palsy and neurologic impairment following term birth. Arch Pathol Lab Med 2000;124:1785–91. 11. Gersell DJ. ASCP survey on placental examination. Am J Clin Pathol 1998;109:127– 43. 12. Hermansen MC. Nucleated red blood cells in the fetus and newborn. Arch Dis Child Fetal Neonatal Ed 2001;84: F211–5. 13. Hanlon-Lundberg KM, Kirby RS. Nucleated red blood cells as a marker of acidemia in term neonates. Am J Obstet Gynecol 1999;181:196 –201. 14. Naeye RL, Localio AR. Determining the time before birth when ischemia and hypoxemia initiated cerebral palsy. Obstet Gynecol 1995;86:713–9.
Curtin et al
Placental Nucleated RBC
309
15. Scher MS, Belfar H, Martin J, Painter MJ. Destructive brain lesions of presumed fetal onset: Antepartum causes of cerebral palsy. Pediatrics 1991;88:898 –906. 16. Tsuji T, Hamaguchi Y, Wang F, Houwen B. New rapid flow cytometric method for enumeration of nucleated red blood cells. Cytometry 1999;37:291–301. 17. Fox H. The incidence and significance of nucleated erythrocytes in the foetal vessels of the mature human placenta. J Obstet Gynaecol Br Commonw 1967;74:40 –3. 18. Altshuler G, Herman A. The medicolegal imperative: Placental pathology and epidemiology. In: Stevenson DK, Sunshine P, eds. Fetal and neonatal brain injury: Mechanisms, management, and the risks of practice. Philadelphia: Decker, 1989:250 – 63.
310
Curtin et al
Placental Nucleated RBC
19. Benirschke K. The placenta in the litigation process. Am J Obstet Gynecol 1990;162:1445–50. 20. Altshuler G. Placenta within the medicolegal imperative. Arch Pathol Lab Med 1991;115:688 –95.
Address reprint requests to: William M. Curtin, MD, University of Rochester Medical Center, Department of Obstetrics and Gynecology, 601 Elmwood Avenue, Box 668, Rochester, NY 14642; E-mail: [email protected]. Received December 7, 2001. Received in revised form February 27, 2002. Accepted March 14, 2002.
OBSTETRICS & GYNECOLOGY
The predominant erythrocyte in the embryo is the nucleated red blood cell (RBC), which declines progressively with advancing gestational age, such that by 12 weeks it comprises less than 2% of circulating red cells.1 Green and Mimouni2 have reported mean absolute nucleated RBC counts for normal term infants of 0.4 ⫻ 109/L and considered values of 1 ⫻ 109/L to be a possible From the Division of Maternal-Fetal Medicine, Department of Obstetrics and Gynecology, University of Rochester Medical Center, Rochester, New York; Department of Pathology, Toledo Children’s Hospital, Toledo, Ohio; and Departments of Medicine and Obstetrics and Gynecology, Medical College of Ohio, Toledo, Ohio.
indicator of intrauterine hypoxia. Values for normal preterm infants are more difficult to interpret and are confounded by the etiology of the preterm delivery. However, fetal blood sampling from normal preterm fetuses demonstrates higher percentages of nucleated RBCs as compared with term fetuses.3 From the data of Forestier et al,3 an estimated mean absolute nucleated RBC count in normal preterm fetuses would be in the range of 0.8 ⫻ 109/L to 1.1 ⫻ 109/L. Increased newborn nucleated RBC counts can be found in maternal diabetes2 and intrauterine growth restriction,4 as well as intrauterine hypoxia.5 Increased numbers of nucleated RBCs in newborns are associated with intrauterine fetal asphyxia and perinatal brain damage.5–7 Buonocore et al7 found significantly higher nucleated RBC counts at birth in three 1-year-old children with abnormal neurodevelopmental status when compared with a normal cohort. Maier et al8 showed a significant correlation between umbilical venous blood erythropoietin concentration at birth and nucleated RBC count in peripheral venous blood. They related the histologic characteristics of meconium phagocytosis in the placenta to the timing of hypoxia and concluded that fetal erythropoietin, produced after a few hours delay in response to a hypoxic stimulus, leads to increased erythropoiesis (manifested by nucleated RBCs). Nucleated RBCs can be readily visualized by microscopic examination of the placenta (Figure 1). The manner in which increased nucleated RBCs are reported is variable among pathologists. The College of American Pathologists recommends descriptive reporting for the presence of increased nucleated RBCs within placental vessels.9 Semiquantitative methods have been used in research studies with nucleated RBCs reported as mildly or moderately increased in number.10 One survey of pathologists on placental examination reported that 38% did not routinely assess for increased nucleated RBCs within placental vessels.11 Theoretically, nucleated RBCs seen within placental vessels reflect the number in
VOL. 110, NO. 2, AUGUST 2002 © 2002 by The American College of Obstetricians and Gynecologists. Published by Elsevier Science Inc.
0029-7844/02/$22.00 PII S0029-7844(02)02041-0
305
Figure 1. A) Nucleated red blood cells within cross-section of umbilical cord blood vessel (hematoxylin-eosin, ⫻ 400, original magnification). B) High-power view of a single nucleated red blood cell (hematoxylin-eosin oil immersion, ⫻ 1000, original magnification). Curtin. Placental Nucleated RBC. Obstet Gynecol 2002.
newborn blood at the time of birth. The purpose of this study was to determine the feasibility of using histologic sections of the placenta to perform quantitative nucleated RBC counts, and to determine if a strong enough correlation existed with newborn nucleated RBC counts such that the placental counts might predict newborn counts. MATERIALS AND METHODS This retrospective study used the medical records database at The Toledo Hospital. A search by International Classification of Diseases, 9th Revision, code 656.31 (fetal distress), during the 2-year period, 1998 and 1999, demonstrated 615 maternal records. Twenty-four cases were found for the study fulfilling the following criteria: record of newborn blood count within 2 hours of birth demonstrating nucleated RBCs and archived histologic slides of the placenta. The maternal data collected included age, parity, medical or obstetric complication, and route of delivery. Data collected on the fetus/new-
306
Curtin et al
Placental Nucleated RBC
born included: gestational age, type of abnormal fetal heart rate pattern, time from onset of abnormal fetal heart rate pattern to delivery, birth weight, time of birth, time from birth to newborn blood count, Apgar scores, and umbilical cord blood gas results. Complete blood counts from the 24 newborns were performed on the Coulter STKS (Beckman Coulter, Inc., Fullerton, CA). Absolute nucleated RBC count expressed in units of ⫻109/L was extracted from the newborn complete blood count report. Many studies report nucleated RBCs as a ratio per 100 white blood cells. However, newborn white blood cell counts are variable; thus, nucleated RBC counts could be misleadingly low or high if reported relative to white cells.2,12 For this reason, as well as to simplify placental counting, we chose to work with absolute numbers of nucleated RBCs. A single board certified pathologist (WMC) reviewed the histologic slides of the placentas without knowledge of the newborn nucleated RBC counts. Umbilical vessels in which the lumen was mostly full of blood were the preferred specimens for counting and volume calculations. In three cases, chorionic plate vessels were used because the umbilical cord vessels provided an insufficient specimen. Internal diameters of all selected blood vessels, measured using a mm ruler, averaged approximately 1.5 mm. The thickness of the histologic section was assumed to be a constant 4 m. The volume of blood contained in the vessels was estimated using the formula for either a cylinder (V ⫽ 兿r2h) or an ellipsoid (V ⫽ 4/3 兿abc), depending on the shape of the blood vessel. Systematic counting of nucleated RBCs was performed within the umbilical or chorionic plate vessels under oil immersion at a magnification of 1000 using an Olympus CH2 microscope (Olympus Optical Co., Ltd., Tokyo, Japan). Placental nucleated RBC counts were converted to #/L to compare with the newborn blood. Nucleated RBC counts, even from large groups of newborns, are characteristically skewed, not normally distributed data.13 Transformation of the nucleated RBC counts to natural logarithms was necessary to normalize the nucleated RBC data to fit a linear regression model. The hypothesis of normality of the natural log-transformed nucleated RBC counts was not rejected using the Kolmogorov-Smirnov and Shapiro-Wilk tests. Further support for normally distributed data was evident by graphic displays of the data: histograms, boxplots, normal, and detrended normal probability plots. Statistical analysis was performed using SPSS 9.0 for Windows (SPSS Inc., Chicago, IL). A simple linear regression model was tested with natural logarithm (newborn nucleated RBC count) as the dependent variable and natural logarithm (placental nucleated RBC
OBSTETRICS & GYNECOLOGY
Table 1. Labor and Birth Characteristics of 24 Newborns With Fetal Distress Gestational age (wk) Mean ⫾ SD Range Delivery by cesarean, n (%) Birth weight (g) Mean ⫾ SD Range AGA, n (%) LGA SGA 1-min Apgar score ⬍4, n (%) 5-min Apgar score ⬍7, n (%) Umbilical artery cord gases Mean ⫾ SD (Range) pH PCO2 (mm Hg) PO2 (mm Hg) Base excess (mmol/L) Start of abnormal FHR pattern to delivery (min) Mean ⫾ SD Range Time interval, birth to newborn blood count (min) Mean ⫾ SD Range
35.0 ⫾ 5.0 24–42 18 (75) 2508 ⫾ 1055 612–3950 20 (83) 3 (13) 1 (4) 12 (50) 9 (37.5) 7.13 ⫾ 0.12 (6.89–7.33) 68.7 ⫾ 13.1 (49–98) 11.5 ⫾ 6.8 (3–26) ⫺8.3 ⫾ 6.2 (⫺28–0.20) 254.7 ⫾ 219.7 35–878 53.5 ⫾ 30.9 8–104
SD ⫽ standard deviation; AGA ⫽ appropriate for gestational age; LGA ⫽ large for gestational age; SGA ⫽ small for gestational age; FHR ⫽ fetal heart rate.
count) as the independent variable. Significance was set at ␣ ⱕ .05. RESULTS The antepartum maternal complications of the 24 newborns with a diagnosis of fetal distress in labor were the following: type 1 diabetes (four), severe preeclampsia (five), both type 1 diabetes and severe preeclampsia (two), choriamnionitis (four), suspected intrauterine growth restriction (one), preterm rupture of membranes (one), placenta previa with bleeding (one), intrapartum bleeding (two), smoking and late prenatal care (one), prior intrauterine fetal demise (one), and none (two). The predominant abnormal fetal heart rate patterns noted were variable decelerations (ten), late decelerations (eight), tachycardia (two), bradycardia (one), only
absent variability (one), and no abnormality (two; both newborns required resuscitation). Selected pregnancy, labor, and birth characteristics are given in Table 1. The mean absolute nucleated red cell counts from newborn blood samples and histologic sections of the placenta are shown in Table 2. The mean and standard deviation nucleated RBC count for the 24 newborns was 4.81 ⫻ 109 ⫾ 5.46 ⫻ 109/L as compared with 1.37 ⫻ 109 ⫾ 1.78 ⫻ 109/L calculated from their placentas. Because the distribution of nucleated red cell concentrations was skewed, transformation of the data to natural logarithms was performed resulting in a normal distribution. The natural logarithmic transformation of the newborn and placental nucleated RBC counts demonstrated a strong correlation (Pearson r ⫽ 0.862, P ⬍ .001). A power calculation demonstrated, with a sample size of 24 and the correlation of 0.862, there was a greater than 90% probability that the correlation was not zero at the .001 level of statistical significance. A simple linear regression was calculated predicting natural logarithm (newborn nucleated RBC count) based on their calculated natural logarithm (placental nucleated RBC count). A significant linear regression was found, P ⬍ .001, with an r2 of 0.743. Seventy-four percent of the variation in newborn nucleated RBC counts was explained by the calculated placental nucleated RBC counts. The prediction equation obtained was natural logarithm (newborn nucleated RBC count) was equal to 1.002 ⫻ natural logarithm (placental nucleated RBC count) ⫹ 1.173 (Figure 2).
DISCUSSION The presence of nucleated RBCs within placental vessels should be assessed and reported upon routinely by the pathologist.9 This subjective impression of nucleated RBCs is dependent on the degree of effort expended; they can be overlooked. Methods used to report the presence of nucleated RBCs in placental examinations are qualitative9 or semiquantitative.10 The method described in this study of calculating a nucleated RBC count from placental slides allows one to quantitate the numbers of nucleated RBCs and express the results in a
Table 2. Correlation of Newborn Nucleated Red Blood Cell Counts (Number of Nucleated Red Blood Cells ⫻ 109/L) With Placental Nucleated Red Blood Cell Counts (n ⫽ 24)
Nucleated RBC*
Newborn
Placenta
Correlation coefficient (r)
P
4.81 ⫾ 5.46 (0.44–20.0)
1.37 ⫾ 1.78 (0.13–8.47)
.862
⬍.001†
RBC ⫽ red blood cell. * Mean ⫾ standard deviation (range). † Pearson correlation coefficient on natural log-transformed data.
VOL. 110, NO. 2, AUGUST 2002
Curtin et al
Placental Nucleated RBC
307
Figure 2. Scattergram with least squares line: natural logarithm (newborn nucleated red blood cell) vs natural logarithm (placental nucleated red blood cell). Curtin. Placental Nucleated RBC. Obstet Gynecol 2002.
manner that allows direct comparison with a neonatal whole blood sample. Calculation of nucleated RBC counts from placental slides would be unnecessary in the case where a newborn complete blood count is obtained in a timely manner after birth. In normal newborns, nucleated RBCs are cleared from the blood such that their numbers at 6 hours after birth are approximately half of the number within the first hour after birth.14 However, most newborns who appear to be doing well after birth do not get routine complete blood counts. Furthermore, most children subsequently found to have cerebral palsy are asymptomatic in the newborn period.15 The indications for placental examination, on the other hand, are fairly broad,9 and about 30% of institutions might be expected to submit 25% or more of all their delivered placentas.11 There would be expected to be many instances in which placental slides would be available, but a newborn complete blood count was not obtained. It is this latter circumstance in which the placental slides could be useful to estimate newborn nucleated RBC counts. The placental counts in this study did substantially underestimate the newborn nucleated RBC counts; however, the strong positive correlation between the two counts fit a linear regression model and thus allowed us to use the relationship in prediction of newborn nucleated red cell counts. The placentas examined in this study were prepared according to a standard institutional protocol.9 It may be possible through selective
308
Curtin et al
Placental Nucleated RBC
preparation of umbilical cord segments to improve the accuracy of the histologic placental counts in relation to newborn or cord blood nucleated RBC counts. This study is limited by the variability inherent in nucleated RBC counts on newborns. The standard deviations in both the newborn and placental counts were high, and this limits the precision of prediction intervals for newborn nucleated red cells from placental counts. For example, a predicted newborn nucleated RBC count of 3.9 ⫻ 109/L would be associated with a wide 95% prediction interval of 1.1 ⫻ 109/L to 14.0 ⫻ 109/L. Current laboratory measurement of whole blood nucleated red cell counts by manual differential cell count may be a source of variability, and increased precision may be possible with direct measurements by newer hematology analyzers.16 This study was retrospective and designed to make use of preexisting placental slides and complete blood counts from newborns. An ideal study would have been prospective comparing nucleated RBC counts from placental slides with nucleated RBC counts on blood obtained immediately after birth either from the umbilical vein or the newborn itself. The mean time from birth to the attainment of a newborn blood sample of 53.5 minutes in this study was not excessively long but could account for some of the variability of the data, as well as underestimation of the placental nucleated RBC counts. The results of this study would be generalizable only to a similar population of acidemic newborns. The equa-
OBSTETRICS & GYNECOLOGY
tion for predicting newborn nucleated RBC counts from placental slides is applicable only at our institution. However, pathologists could use our methods to generate their own prediction equations that would be specific to their institution and populations. The sample size was small and did not allow for statistical analysis of subgroups. Evaluation of residual statistics in our study suggested that the term pregnancy with no underlying maternal disease was associated with the least error in our linear regression model. Further study needs to be performed on more homogeneous groups with stratification based on gestational age, fetal size, and maternal disease states. Intraobserver and interobserver reliability of this technique remains to be determined on a large sample. There has been considerable interest in elevated neonatal nucleated RBCs and their association with perinatal asphyxia.5–7,14 Fox noted that increased nucleated RBCs in blood vessels in histologic sections of the placenta were associated with fetal hypoxia.17 Altshuler and Herman18 confirmed the association of the placental finding of increased nucleated RBCs with fetal hypoxia in an epidemiologic study. In a case control study of 892 infants hospitalized in the neonatal intensive care unit, the placentas of 193 neonates with a diagnosis of perinatal asphyxia were compared with 699 control neonates. The presence of nucleated RBCs, adjusted for ten other placental lesions, gestational age, and birth weight, was a significant independent predictor of perinatal asphyxia (odds ratio 3.00, 95% confidence interval 1.7, 5.4). Placental pathology is often used as evidence in litigation involving neurologically impaired children.19,20 Our method of quantitation of nucleated RBCs from placental slides may be of medicolegal use. Because the placental nucleated RBC counts are reflective of those in the newborn at the time of birth, the placenta could be used as a surrogate for neonatal nucleated RBC counts. This would be particularly valuable in cases where a neonatal blood sample is either not obtained or poorly timed. The quantification of counts expressed as nucleated RBC/L provides more meaningful and objective information than do those done descriptively or on a semiquantitative scale. However, increased nucleated RBCs, either in the placenta or on newborn blood samples, should never be used in isolation to diagnose perinatal asphyxia.12 In summary, we have described a method that directly relates nucleated RBCs identified from placental slides to newborn whole blood nucleated RBCs. This study shows that it is feasible to calculate nucleated RBC counts from placental sections and that these counts could be used to predict newborn nucleated RBC counts from whole blood shortly after birth. Further work needs
VOL. 110, NO. 2, AUGUST 2002
to be done on more homogeneous groups of subjects to increase the precision of the method. Histologic slides of the placenta could be a useful independent data source for quantitation of fetal nucleated RBCs at the time of delivery. REFERENCES 1. Salafia CM, Weigl CA, Foye GJ. Correlation of placental erythrocyte morphology and gestational age. Pediatr Pathol 1988;8:495–502. 2. Green DW, Mimouni F. Nucleated erythrocytes in healthy infants and in infants of diabetic mothers. J Pediatr 1990;116:129 –31. 3. Forestier F, Daffos F, Catherine N, Renard M, Andreux JP. Developmental hematopoiesis in normal human fetal blood. Blood 1991;11:2360 –3. 4. Baschat AA, Gembruch U, Reiss I, Gortner L, Harman CR, Weiner CP. Neonatal nucleated red blood cell counts in growth-restricted fetuses: Relationship to arterial and venous Doppler studies. Am J Obstet Gynecol 1999;181: 190 –5. 5. Korst LM, Phelan JP, Ahn MO, Martin GI. Nucleated red blood cells: An update on the marker for fetal asphyxia. Am J Obstet Gynecol 1996;175:843– 6. 6. Phelan JP, Ahn MO, Korst LM, Martin GI. Nucleated red blood cells: A marker for fetal asphyxia? Am J Obstet Gynecol 1995;173:1380 – 4. 7. Buonocore G, Perrone S, Gioia D, Gatti MG, Massafra C, Agosta R, et al. Nucleated red blood cell count at birth as an index of perinatal brain damage. Am J Obstet Gynecol 1999;181:1500 –5. 8. Maier RF, Gu¨nther A, Vogel M, Dudenhausen JW, Obladen M. Umbilical venous erythropoietin and umbilical arterial pH in relation to morphologic placental abnormalities. Obstet Gynecol 1994;84:81–7. 9. Development Task Force of the College of American Pathologists. Placental Pathology Practice Guidelines. Practice guidelines for examination of the placenta. Arch Pathol Lab Med 1997;121:449 –76. 10. Redline RW, O’Riordan MA. Placental lesions associated with cerebral palsy and neurologic impairment following term birth. Arch Pathol Lab Med 2000;124:1785–91. 11. Gersell DJ. ASCP survey on placental examination. Am J Clin Pathol 1998;109:127– 43. 12. Hermansen MC. Nucleated red blood cells in the fetus and newborn. Arch Dis Child Fetal Neonatal Ed 2001;84: F211–5. 13. Hanlon-Lundberg KM, Kirby RS. Nucleated red blood cells as a marker of acidemia in term neonates. Am J Obstet Gynecol 1999;181:196 –201. 14. Naeye RL, Localio AR. Determining the time before birth when ischemia and hypoxemia initiated cerebral palsy. Obstet Gynecol 1995;86:713–9.
Curtin et al
Placental Nucleated RBC
309
15. Scher MS, Belfar H, Martin J, Painter MJ. Destructive brain lesions of presumed fetal onset: Antepartum causes of cerebral palsy. Pediatrics 1991;88:898 –906. 16. Tsuji T, Hamaguchi Y, Wang F, Houwen B. New rapid flow cytometric method for enumeration of nucleated red blood cells. Cytometry 1999;37:291–301. 17. Fox H. The incidence and significance of nucleated erythrocytes in the foetal vessels of the mature human placenta. J Obstet Gynaecol Br Commonw 1967;74:40 –3. 18. Altshuler G, Herman A. The medicolegal imperative: Placental pathology and epidemiology. In: Stevenson DK, Sunshine P, eds. Fetal and neonatal brain injury: Mechanisms, management, and the risks of practice. Philadelphia: Decker, 1989:250 – 63.
310
Curtin et al
Placental Nucleated RBC
19. Benirschke K. The placenta in the litigation process. Am J Obstet Gynecol 1990;162:1445–50. 20. Altshuler G. Placenta within the medicolegal imperative. Arch Pathol Lab Med 1991;115:688 –95.
Address reprint requests to: William M. Curtin, MD, University of Rochester Medical Center, Department of Obstetrics and Gynecology, 601 Elmwood Avenue, Box 668, Rochester, NY 14642; E-mail: [email protected]. Received December 7, 2001. Received in revised form February 27, 2002. Accepted March 14, 2002.
OBSTETRICS & GYNECOLOGY