hPL: physiologic and pathophysiologic observations.

Serum human placental lactogen (hPL) levels were studied in 806 women in late pregnancy. The hPL levels were positively correlated with birth weight but were unrelated to maternal age, parity, socioeconomic status, or the sex of the newborn. The hPL levels peaked at 37 weeks' gestation and then declined moderately. An individual's hPL levels in late pregnancy are quite constant week to week. Patients with severe chronic hypertension have low hPL values; those carrying twins have high values.

i :1 J f 'I .f! I ,,, ,I , :1 ., I. I hPL: Physiologic and Pathophysiologic Observations FRANK /. ZLA TNIK, MD, MICHAEL W. VARNER, MD, KA THERINE S. HA lISER, MD, AND SYL VIA S. LEE, MS Serum human placental lactogen (hPL) levels were studied in 806 women in late pregnancy. The hPL levels were posi­tively correlated with birth weight but were unrelated to maternal age, parity, socioeconomic status, or the sex of the newborn. The hPL levels peaked at 31 weeks' gestation and then declined moderately. An individual's hPL levels in late pregnancy are quite constant week to week. Patients with severe chronic hypertension have low hPL values; those carrying twins have high values. Despite an extensive literature on serum human pla­cental lactogen (hPL) [also called human chorionic so­matomammotropin (hCS)], several basic clinical ques­tions concerning this placental protein hormone remain unresolved.1 We have recently completed a relatively large study of serial hPL determinations in late pregnancy. The usefulness of hPL as a predictor of perinatal outcome is the subject of a separate report.2 This paper presents data relevant to the following questions: 1) What is the relationship between hPL levels in late pregnancy and the birth weight of the newborn? 2) What is the pattern of hPL values by week in late pregnancy? 3) How consistent are an indi­vidual's weekly hPL values in late pregnancy? and 4) What are the hPL values in certain abnormal condi­tions of pregnancy? Materials and Methods Volunteers from our high-risk clinic and routine ante­partum clinics were enrolled in the study from Sep­tember 1976 through June 1977. The only entry re-from the Departments of Obstetrics and Gynecology and Preventive Medicine and Public Health, University of Iowa College of Medicine, Iowa CIty, Iowa. This investigation was supported in part by Biomedical Research Support Grant RR05372 from the Biomedical Research Support Branch. Division of Research facilities and Resources. National Institute of Health. Submitted for publication february 6, 1979. 314 0029-1844/79/090314-04$01.15 quirements 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 stafldard commercial kit (New England Nuclear). Hospitalized patients had weekly samples taken. The separated serum samples were frozen until analysis. The intraassay coefficient of variation was 3.9%; the interassay coefficient of var­iation was 13.9%. The hPL results were not reported. Therefore, ob­stetric management was not influenced by the serum hPL levels. Following delivery, data sheets were completed us­ing information provided by the hospital records of the mothers and newborns. Demographic data, identi­fied risk factors, maternal course during pregnancy, 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. . Two hundred forty-two babies were delivered to women whose pregnancies were defined as being at risk prior to labor, because 1 or more of the following had occurred: antepartum admission to the hospital, antepartum fetal heart rate (FHR) testing, urinary es­trogen determination, amniocentesis for fetal maturity testing, and/or a t:. optical density (00) qetermination at 450 nm. Factors accounting for the at-risk status are listed in Table 1. Risk patients were placed in group 1 if they spontaneously went into labor and were placed in group 2 if delivery was effected because of concern for the mother and/or fetus. Five hundred seventy-six babies were delivered to women not considered to be at risk prior to labor. They were not admitted to the hospital prior to labor, Obstetrics & Gynecology I I Table 1. Antepartum Factors Accounting for At-Risk Status Factor Group 1 Group 2 Total Postmaturity (known and suspected) 39 20 59 Premature rupture of the membranes 22 31 53 Chronic hypertension 12 19 31 Preeclampsia 6 18 24 Third-trimester bleeding 15 5 20 Diabetes mellitus 1 11 12 Suspected IUGR 10 2 12 Twins 10 0 10 Isoimmunization 4 4 8 Previous cesarean section 10 1 11 Renal disease 3 3 6 Premature labor 4 1 5 Bad obstetric history 2 3 5 Autoimmune disease 3 0 3 Cardiac disease 2 1 3 Other 24 5 29 167 124 291" • 242 births---certain patients had multiple risk factors; 576 births occurred in patients who were not at risk antepartum. IUGR = Intrauterine growth retardation. nor did they have antepartum FHR testing, urinary es­trogen assays, or third-trimester amniocenteses per­formed. Some of these patients may have been high risk because of maternal age, socioeconomic status, or other factors, but the specific clinical and laboratory measures generally un4ertaken on our service in preg­nancies we consider to be at risk were absent. If there was a discrepancy between the clinical esti­mate of gestational age and that suggested by physical examination of the newporn, the latter was chosen as indicating gestational age. Appropriateness of birth weight for gestational age was determined from the Colorado growth chart.' Severe hypertension was de­fined as a usual rather ti1an a peak blood pressure of 2:140/90 mmHg. In diabetic pregnancies the average total daily in­sulin requirement in the third trimester was compared with the nonpregnant insulin dosage for each patient. Two patients with an insulin increase of > 100% were compared with 7 patients with an increase of < 100%. The 806 women studied had a total of 1986 serum samples analyzed for hPL. The number of samples per patient ranged from 1 to 8 (Table 2). Seventy-one per­cent of the patients had their fjnal sample obtained within 7 days of delivery and 81% within 10 days of delivery. Statistical methods used were as follows: hPL-birth weight relationships: correlation coefficients and mul- VOL. 54, NO.3, SEPTEMBER 1979 tiple regression analyses4 ; birth weight-gestational age-hPL relationships: Bonferroni t testS; hPL weekly pattern: Duncan's multiple range test6 ; individual pa­tient hPL consistency: intraclass correlation coeffi­cients'; comparison of mean hPL levels in various con­ditions: Student's t test! Results and Discussion Relationship Between hPL and Birth Weight Although hPL levels are generally considered to be Significantly correlated in a positive manner with pla­cental weight, the literature is divided on the question of the relationship between hPL levels in late preg­nancy and birth weight.1 Certain reports have de­scribed a positive relatjonship7,"; others have not.9 - 11 The mean hPL level for each of the 794 women who was delivered' of a single fetus was determined, and these were correlated with the birth weights of the newborns, Mean hPL levels in late pregnancy were significantly correlated with birth weight, r = 0.173 (p < 0.001). Other factors positively correlated with birth weight were: gestational age, r = 0.360 (p < 0.001), male sex of the newborn, r = 0.159 (P < O.OOn and multiparity, r = 0,070 (p < 0.05). Birth weight was un­related to the patient's socioeconomic status. Of the factors studied, gestational age was demonstrated by multiple regression analyses to be the most highly correlated with birth weight. The mean hPL correla­tion was stronger than that of the parity of the mother or of the sex of the newborn. Mean hPf,. levels were unrelated to maternal age, socioeconomic status, par­ity, or sex of the newborn. Figure 1 demonstrates sig­nificant mean hPL differences (p < 0.01) among pa­tients who were delivered of small-, appropriate-, or large-for-gestational-age babies. Since serum hPL levels are related to placental weight, and since newborn and placental weights are positively correlated, it is not surprising that our data Table 2. The Number of Serum Samples per Patient Analyzed for hPL No. of samples No, of patients (%) 1 318 (39) 2 179 (22) 3 112 (14) 4 84 (10) 5 61 (8) 6 36 (4) 7 11 (1) 8 5 (1) 806 (100) hPL '" Human placental lactogen. Ziatnik et al hPL in Late Pregnancy 315 .' 12 10 E 8 "'- O::1J 6 -J .0r..:. 4 2 0 SGA AGA LGA N=12 N=628 N=154 Figure 1. Mean hPL levels as related to birth weight-gestational age categories. SGA = Small for gestational age; AGA = appropriate for gestational age; LGA = large for gestational age. Lines above bars represent 1 standard deviation. demonstrate a relationship between mean hPL levels in late pregnancy and birth weight. As can be seen in Figure 1, however, considerable overlap in hPL levels exists among the different birth weight-gestational age categories. This helps explain the limited value of low hPL levels in identifying small-for-gestational-age fe­tuses in a general obstetric population! Weekly hPL Pattern in Late Pregnancy Several patterns of serum hPL levels in late pregnancy have been described .•. 10.12-1. Peak values have been re­ported at 35," 37-38,15 39,13 and 40 weeks' gestation.'6 Certain authors describe a decline in late preg­nancy, 13.1S.16 while others describe a plateau after 34'2 or 36 weeks' gestation. 18 Our data, based on 1983 determinations, are pre­sented in Figure 2 (3 determinations for week 43 have been excluded). Peak mean hPL levels (8.8 pg/ml) oc- 10 ~ 9 (122;(128; (231; (276; (354) (421} (309; (107) (35) Weeks' Gestation Figure 2. Mean hPL levels by weeks' gestation. The number of samples tested for each week are in parentheses. 316 Zlatnik et al hPL in Late Pregnancy curred at 37 weeks' gestation. This level is signifi­cantly higher than that at 34 weeks' gestation (7.8 p.g/ ml; P < 0.05). After 37 weeks' gestation there was a moderate decline in mean hPL levels to 7.1 p.g/ml at 42 weeks' gestation. The 42-week value is significantly lower than all of the others (P < 0.05). As has pre­viously been reported,'9 the distribution of hPL values in late pregnancy is skewed with a greater spread of values above than below the mean. Individual Consistency of hPL Values Although an individual's hPL levels in late pregnancy may vary within a day by ± 15%,19 there is little infor­mation regarding the consistency of an individual pa­tient's hPL values week to week in late pregnancy. To estimate this consistency in our study population, in­traclass correlation or reliability coefficients were com­puted separately for those patients tested 2, 3, 4, or more times in late pregnancy. Due to computational limitations, the coefficients summarized in Table 3 were computed separately for the number of samples tested in late pregnancy. The reliability coefficient quantifies the consistency of hPL determinations from week to week within a subject. If the hPL readings show little variation rela­tive to the variation among patients, the reliability co­efficient will be significantly larger than 0, although by definition it will never be greater than 1. The levels of significance (Table 3) indicate that the reliability of hPL readings is greater than 0, regardless of the num­ber of times a patient is tested in late pregnancy. This suggests that multiple weekly samples from an indi­vidual in late pregnancy may provide little more infor­mation than that provided by 1 or 2 samples. hPL Levels in Various Conditions Serum hPL levels reflect placental mass. Therefore one might predict that certain conditions would be associ­ated with high or low hPL levels. The 12 women with twins in our study population had a mean hPL level of 12.7 pg/ml, which was significantly higher (P < 0.01) Table 3. Reliability Coefficients Related to the Number of Samples per Patient No. of No. of Reliability samples patients coefficient Significance 2 179 0.71 P<0.OO1 3 112 0.64 P<O.(XH 4 84 0.66 P<O.OOI 5 61 0.74 P<O.OOI 6 36 0.64 P<O.OOI 7 11 0.51 P<O.OOI 8 5 0.33 P<O.Ol Obstetrics & Gynecology than the mean level of patients with singleton preg­nancies (8.3 ",giml). Patients with severe hypertension in our study population (N = 9) had a mean hPL level of 5.9 ",gimi, which was significantly lower (P < 0.01) than that of the remainder of the study population (8.3 ",glml). No other statistically significant differences in mean hPLs for the conditions listed in Table 1 were found. This may reflect in part small numbers in some of the categories and also the heterogeneity of the pa­tients in each category, with some having mild or no disease (eg, Rh-sensitized patients carrying Rh-nega­tive fetuses, and patients with suspected intrauterine growth retardation delivering appropriate-for-gesta­tional- age babies). Because hPL affects carbohydrate and lipid metabo­lism and is thought to be in part responsible for the diabetogenic effects of pregnancy, the relationship of hPL levels to the insulin requirements of pregnant dia­betics has received previous attention. Although Soler et a120 could not relate the increased insulin require­ment during pregnancy to hPL levels, Ursell and co­workers" found significantly higher hPL levels in pa­tients whose insulin requirements rose> 100% during pregnancy. The few patients in this study who required insulin prior to pregnancy (N = 9) did not demonstrate a rela­tionship between the incremental insulin required and hPL levels. Conclusions The hPL levels in late pregnancy are positively corre­lated with birth weight. The hPL levels peak at about 37 weeks' gestation and then decline. This decline continues until at least 42 weeks' gestation. In late pregnancy, individuals generally show little variation in hPL levels from week to week. The hPL levels re­flect placental mass. They are high in twin pregnancy and low in patients with severe chronic hypertension. References 1. 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 2. Ziatnik Fl, Varner MW, Hauser KS: hPL: A predictor of peri­natal outcome? Obstet GynecoI54:205, 1979 3. Lubchenco LO, Searls DT, Brazie JV; Neonatal mortality rate: Relationship to birth weight and gestational age. J Pediatr 81:814, 1972 4. Snedecor GW, Cochran WG: Statistical Methods. Sixth edition. Ames, Iowa State University Press, 1967 5. Miller RG; Simultaneous Statistical Inference. New York, McGraw-Hill Book Co, 1966, pp 15-16 VOL 54, NO.3, SEPTEMBER 1979 6. Bancroft TA: Topics in Intermediate Statistical Methods, Vol I. Ames, Iowa State University Press, 1968, pp 103-109 7. Seppala M, Ruoslahti E: Serum concentration of human placen­tal lactogenic hormone (hPL) in pregnancy complications. Acta Obstet Gynecol Scand 49:143,1970 8. Cramer DW, Beck P, Makowski EL: Correlation of gestational age with maternal human chorionic somatomammotropin and maternal and fetal growth hormone plasma concentrations dur­ing labor. Am J Obstet Gynecoll09:649, 1971 9. Sciarra JJ, Sherwood LM, Varma AA, et al: Human placental lac­togen (HPL) and placental weight. Am J Obstet Gynecol 101:413,1968 10. Singer W, Desjardins P, Friesen HG: Human placental lactogen, an index of placental function. Obstet Gynecol36:222, 1970 11. Spellacy WN, Usategui-Gomez M, Fernandez-deCastro A: Plasma human placental lactogen, oxytocinase, and placental phosphatase in normal and toxemic pregnancies. Am J Obstet Gynecol 127:10, 1977 12. Saxena BN, Refetoff S, Emerson K, et al: A rapid radioimmuno­assay for human placental lactogen. Am J Obstet Gynecol 101:874,1968 13. Ylikorkala 0; Maternal serum HPL levels in normal and com­plicated pregnancy as an index of placental function. Acta Ob­stet Gynecol Scand (Suppl) 26:1,1973 14. Persson B, Lunell NO, Aubert ML, et al: Determination of plasma human chorionic somatomammotrophin and urinary estriol in diabetic pregnancies. Acta Obstet Gynecol Scand 52:63,1973 15. Genazzani AR, Cocola F, Casoli M, et al: Human chorionic so­matomammotrophin radioimmunoassay in evaluation of pla­cental function. 1 Obstet Gynaecol Br Commonw 78:577, 1971 16. Spencer TS: Human chorionic somatomammotropin in the third trimester of pregnancy. J Obstet Gynaecol Br Commonw 78;232, 1971 17. Josimovich JB, Kosor B, Boccella L, et al: Placental lactogen in maternal serum as an index of fetal health. Obstet Gynecol 36:244,1970 18. Lebech PE, Borggaard B: Serum levels of human chorionic so­matomammotropin (HCS) in normal and abnormal preg­nancies. Acta Endocrinol 75 (SuppI182): 35, 1974 19. Teoh ES, Spellacy WN, Buhi WC: Human chorionic somato­mammotrophin (HCS); A new index of placental function. J Obstet Gynaecol BrCommonw 78:673, 1971 20. Soler NG, Nicholson HO, Malins JM: Serial determination of human placental lactogen in the management of diabetic preg­nancy. Lancet 2;54, 1975 21. Ursell W, Brudenell M, Chard T: Placental lactogen levels in diabetic pregnancy. 8r Med J 2:80, 1973 Address reprint requests to; Frank J. Zlatnik, MD Department of Obstetrics & Gynecology University of Iowa College of Medicine Iowa City, IA 52242 Accepted for publication February 27, 1979. Copyright © 1979 by The American College of Obstetricians and Gynecologists. Zlatnik et al hPL in Late Pregnancy 317