Human placental lactogen: a predictor of perinatal outcome?

Serial human placental lactogen (hPL) determinations were performed on 806 women with normal and abnormal pregnancies late in the pregnancy. These results were not reported to the clinicians involved. For the study population as a whole, low hPL levels did not effectively predict those adverse perinatal outcome variables evaluated. Further analysis revealed that this was true both for the normal and abnormal pregnancy groups. Our data do not support the routine use of antepartum hPL screening, as advocated by others, as a means of improving perinatal outcome. In certain at-risk patients, there was an association between low hPL values and the presence of 1 or more of the adverse outcome variables. However, these patients had been recognized clinically as having fetuses in jeopardy.

'ertil ; and, Human Placental Lactogen: A Predictor of Perinatal Outcome? FRANK J. ZLATNIK, MD, MICHAEL W. VARNER, MD, AND KA THERINE S. HA USER, MD Serial human placental lactogen (hPL) determinations were performed on 806 women with normal and abnormal preg­nancies late in the pregnancy. These results were not re­ported to the clinicians involved. for the study population as a whole, low hPL levels did not effectively predict those adverse perinatal outcome variables evaluated. further analysis revealed that this was true both for the normal and abnormal pregnancy groups. Our data do not support the routine use of antepartum hPL screening. as advocated by others, as a means of improving perinatal outcome. In cer­tain at-risk patients, there was an association between low hPL values and the presence of 1 or more of the adverse outcome variables. However, these patients had been rec­ognized clinically as having fetuses in jeopardy. The clinical utility of serum human placental lactogen (hPL) determinations in improving perinatal outcome has not been clearly defined. Certain reports suggest that this test may be of value in monitoring abnormal pregnancies,I.2 while others are less hopeful in tHis re­gard. 3 •• Low hPL levels have been reported by some to correlate with adverse fetal/neonatal outcome in clini­cally normal pregnancies as weW'"; others have not found the test to be helpful in the clinically normal pa­tient. ·'? A clinical study was undertaken at the University of Iowa College of Medicine to determine whether hPL determinations are beneficial. The questions asked were: 1) Would routine antepartum hPL testing im­prove perinatal results? 2) In clinically normal preg­nancies would low hPL values effectively predict ad­verse perinatal outcome? and 3) In clinically abnormal pregnancies would hPL testing provide information re­garding perinatal outcome additional to that provided by current clinical and laboratory fetal surveillance? the Department of Obstetrics and Gynecology, University of luwa College of Medicine, Iowa City, Iowa. Supported in part by Biomedical Research Support Grant RR-5372 from, the Biomedical Research Support Branch, Division of Research Factlliles and Resources, National Institutes of Health. Submitted for publication January 15, 1979. VOL 54, NO. 2, AUGUST 1979 Materials and Methods Volunteers from the high-risk and routine antepartum clinics were enrolled in the study from September 1976 through June 1977. The only entry requirements were the presence of a living fetus and the absence of labor. At each antepartum visit from 34 weeks' gesta­tion until delivery, a serum sample was obtained which was analyzed in duplicate for hPL by radioim­munoassay using a standard commercial kit (New England Nuclear). Hospitalized patients had samples obtained weekly. So that obstetric management would not be influ­enced by the serum hPL levels, hPL results were not reported. Following delivery, data sheets were completed us­ing information from the hospital records of the mothers and newborns. Demographic data, identified risk factors, fetal surveillance test results, neonatal outcome data, and hPL values were recorded and sub­sequently transferred to magnetic tape for computer analysis. Eight hundred six patients with at least 1 hPL deter­mination delivered at University Hospitals. They and their 818 offspring (12 sets of twins) form the study population. Patients were categorized as follows. Groups 1 and 2 consisted of patients whose pregnancies were defined as being at risk prior to labor because 1 or more of the following events had occurred: antepartum admission to the hospital, antepartum fetal heart rate (FHR) test­ihg, urinary estrogen determination, or amniocentesis for fetal maturity testing and/or Il OD (optical den­sity) 450 nm determination. Risk patients were placed in group 1 if they spontaneously went into labor (de­spite their risk status, the clinical opinion apparently was that induction of labor was not required in these patients). Risk patients were assigned to group 2 if de­livery was effected because of concern for the mother and/or fetus. In these patients induction of labor was carried out or cesarean sections were performed be­fore labor had begun. Patients in group 3 were not considered to be at risk prior to labor. They were not 0029-7844/79/080205-06$01.75 20S admitted to the hospital prior to labor, nor did they undergo antepartum FHR testing, urinary estrogen as­says, or third-trimester amniocenteses. Some of these group 3 patients may have been high risk because of maternal age, socioeconomic status, or other factors, but the specific clinical and laboratory measures gen­erally undertaken in pregnancies we consider to be at risk were absent. It should be emphasized that these operational definitions of groups 1-3 were established at the beginning of the study, and the assignment of a particular patient to a particular group was indepen­dent of her serum hPL levels or the perinatal outcome. Adverse outcome variables of interest were peri­natal death, intrauterine growth retardation (IOCR), congenital anomalies, meconium-stained amniotic fluid, FHR abnormalities in labor, fetal acidosis (pH < 7.25) in labor, 5-minute Apgar score < 8, and the need for immediate neonatal resuscitation. Although the significance of meconium-stained amniotic fluid re­mains controversial, it was selected as an adverse out­come variable because it has been included as such in previous hPL studies.5 ,b If there was a discrepancy between the clinical esti­mate of gestational length and that suggested by phys­ical examination of the newborn, the latter was chosen as indicating gestational age. Appropriateness of birth weight for gestational age was determined from the Colorado growth chart.& Low hPLs were defined as those values within the approximate 2~ percentile for each week of gestation (34-43 weeks) in 404 normal pregnancies. These 404 patients were in group 3 (not at risk) and delivered liv­ing singleton fetuses weighing from 2500 to 4000 g who had 5-minute Apgar scores 2:8 and who required no resuscitation. A patient was placed in the low hPL category if any single value was low. The study patients were categorized as having nor­mal or low hPL levels. The outcome variables were compared for patients having normal or low hPL mea­surements for the study population as a whole, indi­vidually for groups 1-3, and for separate risk factors in groups I and 2. Fisher~s exact test was employed to determine the statistical significance of the differences observed. Results This study focuses on perinatal outcome. Details of the study populatibn and certain physiologic and pathophysiologic hPL relationships are the subjects of a separate report: Seven hundred seventy-two babies (94%) were de­livered of mothers with normal serum hPL values; 46 babies (6%) were delivered of mothers with low serum 206 Ziatnik et al hPL Determinations hPL values. Table I compares the low and normal hPL groups in terms of the prt'sence or absence of the sPe­cific adverse outcome variables. The figures in paren­theses represent percentages within each column for each variable. Perinatal Deaths None of the 9 perinatal deaths occurred in patients with low hPL levels. There were 4 fetal deaths (2 of unknown cause in clinically normal patients; I fetus in a twin pregnancy; I growth-retarded fetus in a patient with hypertension). Five babies died neonatally (I fol­lowing severe asphyxia secondary to premature sepa­ration of the placenta; I set of twins because of pre­sumed coxsackievirus infection; 2 secondary to cardiac anomalies). IUGR In the 10w-hPL group, 7% were growth retarded, ver­sus 2% in the normal-hPL group (Table 1). The appar­ently increased tendency for the 10w-hPL group to be associated with growth-retarded babies is not statisti­cally significant. Furthermore, 93% of the babies deliv­ered of women with low hPL values were not growth­retarded, and the 10w-hPL group accounted for only 17% of the growth-retarded babies in the study. Congenital Anomalies There was no difference between the low- and nor­mal- hPL groups in this category. Meconium Staining of the Amniotic Fluid Again, no difference was noted. Only 6% of the babies with meconium-stained amniotic fluid were in the 10w-hPL group. Abnormal FHR Fetal heart rate abnormalities in labor were recorded in 141 cases. There was no significant difference in dis­tribution between the low- and normal-hPL groups. The 10w-hPL group accounted for 7% of the fetuses with FHR abnormalities. Fetal Addosis Scalp blood pH determinations were performed on 93 fetuses. None of the 17 acidotic fetuses (pH < 7.25) were in the 10w-hPL group. Low 5-Minute Apgar Scores Infants with low 5-minute Apgar scores «8) occurred with equal frequency (7%) in the low- and normal-hPL Obstetrics & Gynecology hPL spe­tren- 1 for ients (2 of us in tient . fol­lepa­pre-rdiac ver­' par­to be ttisti­leliv­wth­only nor-abies 1 the )rded 1 dis­oups. ·tuses )n 93 7.25) urred l-hPL ;%gy Table 1. Correlation of Adverse Perinatal Outcome Variables with hPL Values for Entire Study Population hPL Variable Low Nonnal Total Perinatal death 0 (0) 9 (1) 9 (1) No perinatal dl!ath 46 (100) 763 (99) 809 (99) Total 46 (100) 772 (100) 818 (100) IUGR 3 (7) 15 (2) 18 (2) NolUGR 43 (93) 757 (98) 800 (98) Total 46 (100) 772 (100) 818 (100) Anomaly 1 (2) 16 (2) 17 (2) No anomaly 45 (98) 756 (98) 801 (98) Total 46 (100) 772 (100) 818 (100) Meconium noted 8 (17) 117 (15) 125 (15) Meconium not noted 38 (83) 665 (85) 693 (85) Total 46 (100) 772 (100) 818 (100) Abnonnal FHR 10 (22) 131 (17) 141 (17) FHR nonnal 36 (78) 641 (83) 677 (83) Total 46 (100) 772 (100) 818 (100) pH low « 7.25) 0 (0) 17 (20) 17 (18) pH nonnal (2 7.25) 8 (100) 68 (80) 76 (82) Total 8 (100) 85 (100) 93' (100) 5-minute Apgar score < 8 3 (7) 51 (7) 54 (7) 5-minute Apgar score 2 8 43 (93) 721 (93) 764 (93) Total 46 (100) 772 (100) 818 (100) Resuscitation needed 8 (17) 158 (20) 166 (20) Ressuscitation not needed 38 (83) 614 (80) 652 (80) Total 46 (100) 772 (100) 818 (100) Vascular resuscitation needed 1 (2) 24 (3) 25 (3) Vascular resuscitation not needed 45 (98) 748 (97) 793 (97) Total 46 (100) 772 (100) 818 (100) P> 0.10 for all variables. Figures in parentheses are column percentages for each variable. hPL = Human placental lactogen; IUGR == intrauterine growth retardation; FHR = fetal heart rate . • Ninety-three fetuses had fetal blood sampling. groups. Because I-minute Apgar scores were not en­tered on our data sheets and because with early effec­tive resuscitation a depressed baby at birth might have a high 5-minute score, the need for and type of neo­natal resuscitation were also evaluated. Need for Neonatal Resuscitation Only 5% of the 166 babies requiring resuscitation were in the low-hPL group. Need fQr Vascular Resuscitation Twenty-five babies were depressed enough at birth to require vascular resuscitation (sodium bicarbonate and/or plasma protein) in addition to oxygen. The fre­quency of infants in the low- and normal-hPL groups did not differ. VOL 54, NO.2, AUGUST 1979 For the study population as a whole, assignment to the 10w-hPL group was without predictive value con­cerning the adverse outcome variables that were stud­ied. The frequency of adverse outcome was the same in the low- and nOrn1al-hPL groups; a large majority in the 10w-hPL group was without each of the adverse outcome variables evaluated. Separate evaluation of groups 1-3 gave similar neg­ative results. Group 3 (not at risk antepartum) con­sisted of 576 babies. Twenty-eight (5%) infants were delivered of mothers with low hPL levels. The low­hPL group had either no association or a lower fre­quency of association, with 5 of the adverse outcome variables (perinatal death, congenital anomalies, fetal acidosis, the need for resuscitation, and the need for vascular resuscitation) compared with the normal-hPL group (N = 518). Data for the other 4 variables are Ziatnik et al hPL Determinations 207 Table 2. Correlation of Adverse Perinatal Outcome Variables with hPL Values in Group 3* hPL Variable Low Normal Total HJGR 1 (4) 8 (1) 9 (2) NolUGR 27 (96) 540 (99) 567 (98) Total 28 (100) 548 (100) 576 (100) Meconium noted 6 (21) 90 (16) 96 (17) Meconium not noted 22 (79) 458 (84) 4SO (83) Total 28 (100) 548 (100) 576 (100) Abnonnal FHR 5 (18) 74 (14) 79 (14) FHRnonnal 23 (82) 474 (87) 497 (86) Total 28 (100) 548 (100) 576 (100) 5-minute Apgar score < 8 3 (11) 32 (6) 35 (6) 5-minute Apgar score <!: 8 25 (89) 516 (94) 541 (94) Total 28 (100) 548 (100) 576 (100) P > 0.10 for all variables. Figures in parentheses are column percentages for each variable. hPL = Human placental lactogen; IUGR = intrauterine growth retardation; FHR = fetal heart rate . .. Group 3 = Patients not considered to be at risk prior to labor. shown in Table 2. Again, no statistically significant differences were observed between the low- and nor­mal- hPL groups. Two hundred forty-two babies were delivered of mothers in the at-risk groups 1 and 2. Mothers were assigned to the risk groups based on medical (eg, dia­betes mellitus and chronic hypertension) and obstetric [eg, premature rupture of the membranes (PROM) without labor, third-trimester bleeding, and pre­eclampsia] factors. One hundred forty-four babies were delivered of mothers in group 1. Although increased risk explained certain management steps (eg, antepartum FHR test­ing) not taken in group 3 patients, no indicated effec­ting of delivery was undertaken. Ten of these babies (7%) were delivered of mothers with low hPL levels. This low.,.hPL group had no association or a lower fre­quency of association with 8 of the 9 adverse oqtcome variables than did the normal-hPL group (N = 134). Data for the remaining variable (meconium-stained amniotic fluid) showed no real differences (Table 3). There were 98 babies delivered to mothers in group 2. These at-risk patients had delivery effected by in­dicated induction of labor or by cesarean section per­formed before labor had begun. Eight of these babies (8%) were delivered of mothers with low hPL levels. This low-hPL group had no association or a lower fre­quency of association with 4 of the 9 adverse outcome variables (perinatal death, meconium-stained amniotic fluid, fetal acidosis, and low 5-minute Apgar score) than did the normal-hPL group (N == 90). The 5 re­maining variables are considered in Table 4. In group 2 patients with low hPL levels there is a suggestion of an increased rate of IUGR (statistically significant) and FHR abnormalities (not statistically significant) when compared with group 2 patients with normal hPL lev­els, although the small numbers limit interpretation. The 11 low-hPL adverse variable entries in Table 4 represent 6 patients (3 with hypertension, 1 with renal disease and suspected IUGR, and 2 with PROM with­out labor). Five of the 6 patients had had amniotic fluid maturity testing and/or antepartum fetal well- Table 3. Meconium-Stained Amniotic Fluid as Related to hPL Values (Group 1) Meconium noted Meconium not noted Total P = O.SO. 2 8 10 Figures in parentheses are column percentages. hPL = Human placental lactogen. 2.08 Ziatnik et al hPL Determinations Low ------------------------------~------------------ (20) (SO) (100) hPL 21 113 134 Normal (16) (84) (100) 23 121 144 Total (16) (84) (100) Obstetrics & Gynecolog T • being testing. All 6 patients were admitted to the hos­pital prior to labor and had indicated inductions of la­bor. Their doctors made their management decisions without knowledge of the 10w-hPL test results. Analysis of outcome by specific risk factors for the low- and normal-hPL groups did not demonstrate hPL testing to be of predictive value. Discussion In a recent review, Spellacy summarized the literature regarding control factors for hPL and clinical studies of hPL as a placental function test.'o It had been noted previously that hypertensive women who sustained intrauterine fetal deaths tended to have low hPL levels, and it was suggested that serum hPL might serve as a test of placental function. It A prospective study of a high-risk clinic population (both hypertensive and normotensive) resulted in fewer perinatal deaths if low hPL results were reported than if they were not.' Yli­korkala found that low hPL levels were associated with fetal distress and/or IUGR in a variety of preg­nancy complications.2 Following these earlier observations, the potential usefulness of hPL testing as an adjunct in obstetric management was extended from use in hypertensive patients to use in general high-risk patients, and fi­nally, with the studies of Letchworth and Chard~ and England et at 6 to use as a general screening test in nor­mal pregnancies. Other workers have been less enthusiastic. They have found the test to be less helpful in predicting perinatal outcome or altering clinical management in high-risk or normal patientsY In a separate report based on these 806 women with normal and abnormal pregnancies, we have demon­strated a significant correlation between hPL levels and birth weight and have noted mean hPL differences in conditions where large or small placentas are antici­pated. 9 In the present report, however, we are unable to demonstrate a relationship between Jow hPL values and adverse perinatal outcome. Why? Placental weights are correlated in a positive man­ner with serum hPL levels. The placenta in patients with hypertension or with growth-retarded fetuses tends to be small. Patients with these conditions are predisposed to adverse perinatal outcome. Therefore, it is not surprising that some patients with these en­tities and low hPL levels have problems. This associa­tion does not, however, establish that serum hPL can serve as a clinically useful marker of placental func­tion or as an effective predictor of perinatal outcome. The production of this hormone is related to placental size, but considerable variation exists. Although the mean hPL level of patients with growth-retarded fe­tuses is significantly lower than that of patients whose Table 4. Correlation of Adverse Perinatal Outcome Variables with hPL Values (Group 2) hPL Variable Low Nonnal Total IUGR 2 (25) 1 (1) 3 (3) NolUGR 6 (75) 89 (99) 95 (97) Total 8 (100) 90 (100) 98 (100) P<0.05 A.nomaly 1 (13) 0 (0) 1 (1) No anomaly 7 (88) 90 (100) 97 (99) Total 8 (100) 90 (100) 98 (100) P=0.08 A.bnonnal FHR 5 (63) 30 (33) 35 (36) FHRnonnal 3 (38) 60 (67) 63 (64) Total 8 (100) 90 (100) 98 (100) p .. 0.10 Resuscitation needed 2 (25) 17 (19) 19 (19) Resuscitation not needed 6 (75) 73 (81) 79 (81) Total 8 (100) 90 (100) 98 (100) P=0.82 Vascular resuscitation needed 1 (13) 3 (3) 4 (4) Vascular resuscitation not needed 7 (88) 87 (97) 94 (96) Total 8 (100) 90 (100) 98 (100) P=O.97 Figures in parentheses are column percentages for each yariable. hPL '" Human placental lactogen; IUGR '" intrauterine growth retardation; FHR - fetal heart rate. VOL 54, NO.2, AUGUST 1979 Ziatnik et al hPL Determinations Z09 Vi Ii I' ,i' Ii !" fetuses are not growth retarded/ the overlap between the 2 groups limits the predictive value of the test. Ninety-three percent of our patients with low hPL lev­els gave birth to babies who were not growth retarded. Furthermore, the important transport functions of the placenta may be independent of the placenta's ability to produce hPL. Blood levels of another placental hor­mone, progesterone, have not been helpful in clinical obstetric management." It should be emphasized that those group 2 patients in this study with low hPL levels and an apparently in­creased frequency of certain adverse outcome vari­ables had been identified clinically without knowledge of the hPL test results. Therefore, in the one subset of the study population where there may be a relation­ship between low hPL levels and adverse outcome (group 2), the test apparently would not have provided additional information. The decision to deliver these patients had already been made. In the hands of the authors routine hPL screening would result in many false positives, ie, low hPL levels in patients who did not have adverse perinatal out­come. The increased concern for patients with low hPL levels might result in additional expense in fetal sur­veillance and possibly in iatrogenic prematurity as well. 'In this study, low hPL levels were defined as those within the approximate 2~ percentile of carefully characterized normal patients. This is consistent with levels chosen by other authors .of large series.l • U Al­though a less rigidly selected 10w-hPL group would in­clude more patients with adverse perinatal outcome, the problem with false positives would likewise be ex­pected to be greater. Possible explanations for our failure to find low hPL levels to be of predictive value may relate to differ­ences in our patient population or in our clinical man­agement. For example, patients with preeclampsia who do not respond rapidly to in-hospital treatment on our service are generally delivered. This study was designed to evaluate the clinical util­ity of hPL determinations in late pregnancy. Our data do not support the use of hPL testing as a routine screening procedure to improve perinatal outcome. Low hPL values did not effectively predict adverse perinatal outcome in clinically normal patients. In clinically abnormal pregnancies our tentative con- . clusion is that hPL determinations did not provide in­formation regarding perinatal outcome over and above that provided by our current clinical and laboratory fetal surveillance. A definite statement cannot be made, however, because the numbers of patients in specific risk categories (eg, chronic hypertension) are not large enough to permit meaningful evaluation. Others are encouraged to study hPL testing in their 210 Zlatnik et al hPL Determinations high-risk patient populations on a blind or prospec_ tively controlled basis. The value of the test, if any ex­ists, must be in providing additional information to that obtained by the careful· clinician using the cur­rently available diagnostic tools. References 1. Spellacy WN. Buhi WC, Birk SA: The effectiveness of human placental lactogen measurements as an adjunct in decreaSing perinatal mortality: Results of a retrospective and randomized controlled prospective study. Am J Obstet Gynecol 121:835, 1975 2. Ylikorkala 0: Maternal serum HPL levels in normal and com­plicated pregnancy as an index of placental function. Acta Ob­stet Gynec Scand (Suppl) 26:1,1973 3. Josimovich JB, Koser B, Boccella L, et al: Placental lactogen in maternal serum as an index of fetal health. Obstet Gynecol 36:244,1970 4. Hollingsworth DR. Moser RJ. Carlson JW, et al: Abnormal ado­lescent primiparous pregnancy: Association of race. human chorionic somatomammotropin production, and smoking. Am J Obstet GynecoI126:230, 1976 5. Letchworth AT, Chard T: Placental lactogen levels as a screen­ing test for fetal distress and neonatal asphyxia. Lancet 1:704, 1972 6. England P, Lorrimer D, Fergusson IC, et al: Human placental lactogen: The watchdog of fetal distress. lancet 1:5, 1974 7. Granat M, Sharf M, Diengott D, et al: Further investigation on the predictive value of human placental lactogen in high-risk pregnancies. Am I Obstet Gynecol129:647, 1977 8. Lubchenco LO, Searles DT, Brazie JV: Neonatal mortality rate: Relationship to birth weight and gestational age. I Pediatr 81:814, 1972 9. Zlatnik FI, Varner MW, Hauser KS, et al: hPL: Physiologic and pathophysiologic observations. (In preparation) 10. Spellacy WN: Monitoring of high-risk pregnancies with human placental lactogen, Management of the High-Risk Pregnancy. Edited by WN Spellacy. Baltimore, University Park Press, 1976 11. Spellacy WN, Tech ES, Buhi We: Human chorionic somato­mammotropin (HCS) levels prior to fetal death in high-risk pregnancies. Obstet GynecoI35:68S, 1970 12. Dawood MY: Circulating maternal serum progesterone in high­risk pregnancies. Am J Obstet Gynecoll2S:832, 1976 13. Teoh ES, Spellacy WN, Buhi WC: Human chorionic somato­mammotrophin (HCS): A new index of placental function. J Obstet Gynaecol Br Commonw 78:673, 1971 Address reprint requests to; Frank ,_ Zlatnik, MD Department of Obstetrics & Gynecology The University of Iowa Hospitals and Clinics Iowa City, lA 52242 Accepted far publicatioll February 1, 1979. Copyright © 1979 by The American College of Obstetricians and Gynecologists. Obstetrics & Gynecology