A standard of fetal growth for the United States of America WILLIAM E, BRENNER, M.D. DAVID A. EDELMAN, Pu.D. HENDRIGKS,“M.D. Chapel Hill and Research Triangle Park, North Carolina CHARLES Ht The appropriate interpretation of monitored fetal growth throughout pregnancy in individual patients and populations is dependent upon the availabilty of adequate sandards. There is no adequate standard of fetal weight throughout pregnancy that i suitable for patients in the U. S. A. To determine sucha standard for infants delivered at about seu level the 10th, 25¢h, 50th, 75h, and 90th percentiles of feat weight for cack menstrual week of gestation were calculated from 130 fetuses at 8 to 20 menstrual tech’ gestation aborted with prostaglandins and from 30,772 lrveborn infants delivered Af pation at 21 t0 44 mmnstrnal works? gestation. Median fetal erown-to-rump lengths and crown-to-hel lengths wvere derived from measurements of 496 aborted fetuses of 8 to 21 weeks gestation. Fetal weight correction factors for parity, race (socioeconomic status), and fetal sex were calculated. The derived fetal grouth curves are useful for clinical, pubic health, and investigational purposes. (AM. J. OssreT. GynecoL. 126: 555, 1976, Wir tHe apveNT of new monitoring techniques, feul and maternal condition can be better evaluated throughout pregnancy. Although assessment of fetal groxth is an important aspect of fetal monitoring, the expected ranges of fetal weight and growth through- out pregnancy have not been adequately determined. Accurate knowledge of fetal growth is clinically impor- tun for at least three reasons: (1) to idemtify the effects of pathologic pregnancies on fetal size, (2) to aid in the ‘valuation of diagnostic measurements of fetal growth, and (3) to identify potentially abnormal children in the antepartum or immediately postpartum periods. From the Department of Obstetrics and Gynecology. University of North Carlin, and the International Feral Research Program. ‘This work was supported in part bythe National Insitutes of Heals, United States Public Health Service (Grant HE, 0191408), the International Fertility Research Program at Chapel Hill (A1D{csd 2979), and by the Pub Health Research Grant K-16 forthe General Clinical Research Center's Branch of the Division of Rearch Resources. Pastaglandin was supplied bs the Upjoin Company, Kalamazee, Michigan. Presid at the Thirtyeighth Annual Mecing ofthe South Wane Association of Obstetricians and Gpreclogts, Hamil, Bermuda, Jonuary 21-29, 176, Reprint quest: Dr. William F. Brenner, Department of Obstris and Gynecology, University of North Carolina, Chapel Hill Noh Carolina 27514. 555 Available data on fetal size and growth throughout pregnancy are not adequate for many clinical pur- poses. Streeter conducted the most comprehensive study of embryo and fetal growth in early pregnancy and reported values for formalin-fixed and embalmed specimens. Although his measurements. of sitting height, head size, and head-to-foot length are similar to those of nonpreserved specimens, fetal weights differ ignificantly, Since the weight change caused by forma- lin is dependent upon several factors, there is no reliable way to determine the weight of the non- preserved from the preserved fetus. Fetal weight and some of the parameters that appear to be associated with differential rates of fetal growth, in the latter part of the second trimester and through- out the third thrimester in specific patient populations have been reported by several investigators. Unfortu- nately, either these studies pertain to selected popula tions who are not representative of all subgroups in the U.S. A. oF the studies are too small to permit accurate determination of fetal growth percentiles. ‘The im: portance of evaluating the relationships between fetal growth 'and such variables as fetal sex. and maternal race, socioeconomic factors, and parity has been previ- ously reported.* * One of the most frequently used standards of fetat growth is the one reported by Lubchienco and assoc556 Brenner, Edelman, and Hendricks Leng in a T clone coe tone fene-r49 +8 manta! | etn. 190 aon 3 S298 935) F 26 0 4 m2 6 ww wm Genaona! Age (ener! Waka Fig 1. Fetal length: lef, come-t-rump right, crown-to-heel Regression of crown-to-rump length(CRL) and crown-to-heel length (CHL) with 90 per cent confidence limits on a future RL and CHL are graphed from data derived from 496 iced abortions. Each dot represents the L for a given gestational age, (S) i the estimated standard deviation about the regression fine, (23) the mulipl correlation coefficient, and the dashed fine is an extrapolation to 0 em. at 2 menstrual weeks’ gestation, Table I. Selected patient characteristies 8-21 weeks gestation 21-44 weeks’ gestation Characteitic [ No. | % Characerivic [No | % “Agctr ec 30 169. 34330 52168 309 2029207599 B00) 7785. 377 3039 34 3039 6401 208 ioe 308 40 00 “13 Pang Pas 0 280 5G ie 10,032 1g 168 389-2 7836 H a 83 ha 393 3 7 1a be ‘09 ae Rae: Whie 952° 508. Whie 16401 58.3 Prive 15077 49.0 Saft 133443 Nowwhite 24492. Nomwhite 13371 467 Primate 179336 Saf 12648 411 ates,t derived from deliveries at 24 to 42 menstrual weeks’ gestation. Their study was based on relatively few (9,639) selected infants born to Caucasian mothers of low socioeconomic status at a high elevation (5,280 feet). ‘Their derived fetal growth rates may be inap- propriate for other areas of the United States, since fetal growth’ may be affected by altitude, and these other areas may have populations with different races, socioeconomic, and parity compositions. Fetal growth 2 November gs ‘Am. J Obstet yet st dards for the past generations, for other cam: tries," and for other specific U.S. A. groups? may not be representative of present-day fetal growth yan dards in the U. S. A. To determine the “normal” (expected) range of feud weight throughout pregnancy at sea level for mar U S.A. patient types, the association of selected variabhs with fetal weight and Fetal length in early pregnana, pertinent data from 31,268, prostaglandin indued abortions and deliveries after spontaneous labor ata 44 menstrual weeks’ gestation were analyzed, Materials and methods To determine fetal growth from 8 to 21 mensral weeks’ gestation, 641 fetuses aborted with prostigh- dins (PG's) were weighed and measured. All aborios were legally approved and performed in the Cll Research Unit of the Memorial Hospital, Universiy North Carolina at Chapel Hill (elevation 513 fe, during the period of 1972 to 1975. Abortions ver induced with PG's according to investigational po tocols; PGFra by the intra-amniotic and vaginal routs, PGE, by the vaginal route, 15(S)-15-methyl PG by the intraamniotic and intramuscular. souies, and 15(S)-15-methyl PGE: by the intramuscular rote Neither the systemic nor intrauterine administationol any of these prostaglandins is known to cause efecsin the fetus that would change either its weight ots. Fetal weights and crown-to-rump and crowntobed Iengths from intact fetuses: were obtained within 9 minutes of delivery with a Mettler electronic sale and ruler. Gestational age for the’ abo: by a combination of the physician's estimate bad on uterine size and the patient's stated last sorta menstrual period (LNMP). Neither of the two e¢- mates of gestional age alone was satisfactory. There fore, only data from 496 of the 641 pregnancies whet the LNMP estimate and the physician's estimate gestional age differed by no more than 2 weeks we analyzed for fetal length. Fetal weights were deied from 430 fetuses at 8 to 20 weeks’ gestation, Fetal weight from 21 to 44. menstrual weeks go tion was determined with the use of data from 30772 deliveries during the period from 1962 to 1969 MacDonald House, University Hospitals of Cleveland Ohio (elevation 660 feet). Data. were analyzed fra single births with the infant living at the onset of ab among pregancies not complicated by pre-elampia breech, erythroblastosis, diabetes mellitus, or conges+ tal abnormalities. Subsets of some of these data hit been previously reported.®-*: 1” Gestational age calculated to the nearest week of gestation from | |Nedume 125, Nanber5 Fecal weg (Grama Fetal growth standard for United States 557 B 2s 6810 “ew me [Gestaonal Age (Mensa Weeks) Fig. 2. Fetal weight during early weight in grams during the: ini wregnancy. The 50th (median), 25th, and 75th percentiles of fetal ‘24 menstrual weeks of pregnancy as calculated from 430 prostaglandin-induced abortions at €20 weeks’ and 30,772 “spontaneous” deliveries from 21 to 44 ‘weeks gestation are graphed. Dashed lines are extrapolations between the data sts and to 0 Gm. at 2 ‘menstrual weeks gestation, patien’s LNMP. Infants were weighed on a balance sale immediately after delivery. Fetal weight percentiles were caleulated from the daa at each week of gestation. The resulting. 10th, Buh, 50th, 75th, and OOth percentile curves for fetal ‘eight were then smoothed with two point weighted means, Simple linear regression techniques were used for the analyses of crown to heel and crow-to-rump lengths. Whenever statistical tests were used, only the P ‘ahie—the significance level of the test—is presented. Subjects, Although the race distribution for patients inthe 8to 21 week group was similar to that of those in the 21 to 44 week group, the distributions of age and Parity differed (Table 1). In both groups 98 per cent of. the nonwhite patients were blacks. Patients in the 21.t0 H weeks’ group were classified as staff or private, depending on theie ability to. pay private physic (0st, whereas patients in the 8 to 21 weeks’ group were ‘were nonwhite and 89.7 per cent of the pri were white, the effects of these two variahles, race socioeconomic status, on fetal growth could not be adequately separated. Results Fetal length (8 to 21 weeks’ gestation). The relation- ship of both crown-to-rump (CRL) and crown-to-heel (CHL) lengths to gestational age appears to be linear during this period of gestation (Fig. 1). The calculated regressions of the RL and CHL on gestational age are given by: GRL = ~7.42 + 1.19 (gestational age). s CHL = 11.72 + 1,80 (gestational age) s = 36 cm. 3.87 cm, In the range of 8 to 21 menstrual weeks’ gestation the expected CRL and CHL are calculated from the‘558. Brenner, Edelman, and Hendhicks Nowemer 5 ‘Am. J Obie Cyead Table 11, Fetal weight percentiles throughout pregnancy Simoothed percentiles Gestational age = (menstrual weeks) No. of women 10 23 50 75 30 8 6 — = om B 9 7 S = 73" eS 10 5 s = cay S nL 8 = = 119" = 19 8 ma a = 13 a = 225 95.3 - 4 61 = 345 Bia = 15 63 = 510 16.7 = 16 39 = 798 u7 =: "7 36 =e 125 166 = 18 28 = 172 220 - 19 a = 27 283, 20 a1 = 255, 335. = 21 280. 330 410 0 a 320 410 480 x 29 m 370 460 350 0 24 ™ 420 530 640 1,080 25 48 490 630 740 1180 26 86 570 730 860 130 2 6 660 a0 990 tam 28 at 710 980 1,150 1660 20 88 890 1,100 11910 1890 30 138 1.030 11360 1460 2100 31 us 1,180 110 1,630 220 32 210 1310 15570, 1,810 2300 3 22+ 1,480 1,720 2,010 260 34 373 1.670 1910 2,220 2880 35 499 raza 2130 2430 son 36 1,085, 2190 2470 2,650 320 31 1,798, 2310 2580 2870 samo 38 3,908 2510 2770 3,030 3610 39 Sais 2080 2910 3,170 30 40 10,585 2750 3,010 3.280 3370 4a 3,399 2/800 3,070 3.360 3980 2 1,735 2830 3,110 3,10 4060 8 507 2840 3,10 3,420 4,10 “a 147 2790 3,030. 3,390 40 weeks gestation. above equations. The 95 per cent confidence limits ona future CRL or CHL for gestations over 10 weeks are approximated by the expected CRL *3.9 cin. and by the expected CHL 6.0 cm. The relationship of fetal length to age is curvilinear in the very early pregnancy. Fetal weight. The 25th, 50th, and 75th percentiles of fetal weight during the initial 24 weeks’ gestation are portrayed in Fig. 2. The continuity of the percentiles of. the two data'sets [(1) abortions at Chapel Hill and (2) neonates at Cleveland] is apparent. The 10th, 25th, 50th, 75th, and 90th percentiles of fetal weight throughout pregnancy are'portrayed in Fig. 3 and detailed in Table I]. These percentiles are not adjusted for fetal sex or maternal parity and race. ian fetal weights may be overestimated. They were derived from only a small proportion of the fetuses delivered at thot Growth rate expressed as the median Fetal wight gain per week increased progressively until 34 m 38 weeks’ gestation, after which the weekly fetal wet increment decreased (Fig. 4). In contrast, when growth Tate (weight gain) is expressed as the percentage ir crease in weight over the previous week. the maximun percentage increase in weight isin early pregnancyand progressively decreases throughout pregnancy (Fig 5. Variables associated with sdifferences: in. fetl weight. Maternal parity and race (socioeconomi) ant fetal sex are associated with significant (p <0.10 dit ferences in median fetal weights in the latter part pregnancy (Fig. 6).,After $4 to 38 weeks’ gestation median fetal weights were’ significantly differena Fetal growth standard for United States. 559 FETAL WEIGHT (GRAMS) FETAL WEIGHT (GRAMS) 8 eae DEVIATION FROM MEDUN (GRANS) B23 36 BO — GESTATIONAL. AGE (MENSTRUAL, WEEKS)... Fi. 3: el weigh The 30 aia), 1 28s and perenne gh ins throughout pregnancy and correction factors for parity. rare (secinermomic), and sex Alexived from 31.202 prostaglandin-induced aborions and "spontaneous" deliveries are graphed.560. Brenner, Edelman, and Hendricks Weignt Gain per Week (Grams) 250) 159] 109] 04 8 12 16 20.26 ma 2 40 44 Gestational Age (Menstrual Weeks) Fig. 4. Weekly fetal weight gain throughout pregnancy. The median weight gain in grams each week throughout preg. rhancy among 31,202. prostaglandin-induced abortions and “spomtancous" deliveries are graphed. ~0.10) for primiparas and multiparas, for whites (pri- fs (saff), and for male and fernale ibutions of fetal weights about die media weights at each week after the Sth week for each of the above variables were similar, Before 34 weeks’ fetal weights wer different (p >0-10) for these variables. 96 weeks’ the median weights were plotted in Fig. Gand deviations from the ot significantly At each gestational age afi median were plotted in Tig. 3 for patients of different paritics and. races, (socioeconomic) and. fetal: sexes. Analyses indicated that the effects of these variables on median fetal weight, could: be considered additive. From Fig. 3 the expected fetal weight (EFW) adjusted for sex, and/or parity, and/or race (socioeconomic) can be determined within £50 grams. For a selected ma- ternal gestational age the adjustment is made by adding to the median fetal weight the grams deviation for the effects of sex, parity, and race (socioeconomic). For FW) for a single ton delivery to a white private primigravid 39 weeks! gest gto the uncor- rected median weight for 39 weeks’ gestation (3,170 example, the expected fetal weight (1 ‘November 1 ‘Am. J. Obset Ghat grams) the correction factors at $9 weeks’ for prin Bravidas (~65 grams) and for white patients (pi vate (+75. grams) [EFW = 3,170 ~ 65 + 75 = 3,10 grams}. If the fetal sex is known, the EFW can be fr ther adjusted by adding the fetal sex correction fax Comment A standard of fetal weight throughout Pregnancy for singleton uncomplicated. pregrancs lefivered at about sea level has been determined fr male and feimale fetuses, nulliparas, and mulipars and for whites and nonwhites. Akhough median weight of specific U. S. A. populations may vary fom these medians, with appropriate corrections for Ke, the derived median fetal weighs ‘age will probably be accurate eno il investigational purposes. Ober variables such as maternal smoking habits, dsc, abnormal presentations, nutrition, and addicis which may affect fetal weight were not evahuatinth It was the objective of this study to evaluate fe) growth for “normal” pregnancies rather thao f pathologic and nutritional conditions and pero! hhabits that may be altered and/or may depend ut local conditions. “The validity of using fetal weights from delves after spontaneous labor ("spontancous” delivers) uring the late second and early third trimester Int been questioned, since’ a large proportion of the Pregnancies may have been abnor and reste it fetal weights that are not representative of “orm growth, Because there was good continuity inthe fal weight percentile curves derived from prostaghas induced abortions and from fetuses born after“ taneous” labor, it appears that the fetal weight ccemtiles derived from “spontaneous” deliveries ia fate second and third trimesters are valid Ldn ing “normal” (usual fetal growth. : Mean crown-trramp lengih RL) (tn He and crown-to-heel length (CHL) at each week of £6 tion reported in this study were similar to those Pe? ously reported.': From 8 to 21 weeks’ sy relationship ty fetal hength as linear in all of! studies. In contrast, Haase's rule, a commonly calculation for determining the expected ct es curvilinear growth rate during the inital estat Save! a Hingsr rinbenip thereafter: Hise ers for the yx determined weeks oun” fifth Io rule: The expected CHIL in centimete five lunar months of pregnane four and squaring the quotient; after the AU month the expected CHL is determined by 1° the number of menstrual weeks gestation PY,Nidame 196 unl Fetal growth standard for United States 561 Percent Canoe 2 » 6 2 40) 20 2 10 ° 10 12 1S 18 wz 4 we 78 3 a HW SD (Geraional Age Mermruat Weeks) Fig. 8. Percentage increase in. weight over: previous week throughout pregnancy. Throughout Pregnancy, the median percentage increase in fetal weight during the week prior tothe plot among 1,202 prostaglandin-induced abortions and “spontaneous” deliveries is graphed. APARITY BRACE ‘Socmecmonscd sex = ‘soa 204 é Wree Mae 4 i g oe : 3000) cuit 00 & 1 1. 1. 4 3a ae «oe xe WEEKS GESTATION Fig. 6. Median fetal weight by panty, re, (Socioéconomic) nd sex. ‘The median feral weight in fram during 36 ty 42 menstrual wees gestation derived foun 31, 202 prostaglandin induced abortions and “spontaneous” deliveries (central unmarked line) are graphed with fetuses of different sexes and born to mothers of different parities and races (socioeconomic). Deviations prior to 36 ‘eels gestation were not significant (p 70.10).562. Brenner, Edelman, and Hendricks amt en, 0] aco) 1.000 Goraiond Ae Wats): 3M OE Mont tom Bt ois aoa es Fig, 7. Median perinatal weight. Graphed are the median weight in grams throughout the later part of pregnancy de- rived from 30,772 “spontaneous” deliveries, and the media ‘weight throughout the intial 3 months of ife as reported by H.C. Swart, MLD, and associates The dashed line is an ‘extrapolation of the best iting line to indicate simlla rates of growth. ‘multiplying the quotient by five.” Although Haase’s ule may be uselul for rough determinations of ex- pected fetal lengths, it should not be used asa standard for fetal lengths or for determining rates of fetal growth, Fetal weights for use as clinical standards must be derived from. recently. delivered, nonmacerated, nonfixed specimens with an accurately determined igestational age. To obtain estimates of fetal size at 8 to 28 weeks’ gestation for ise by embryologists, Streeter? ‘used 704 formalin-fixed or embalmed specimens in different conditions and states’ of preservation. ob- tained from many sources over a period of several years. The gestational ages were derived by. several methods. He derived the estimates of the expected formalin-fixed fetal weight at 28 to 40 weeks’ gestation by adding 5 per cent (an estimate of the amount added 1g) 1 the mean. weights reported by and embalm- ing do not add a comsistent amount or percentage of weight and the amour of weight added depends upon several factors" In our study dhe median weights at each week's gestation were from 0 to 40 grams higher igher at 27 {0 30 seeks’ gestation than those reported by Strecter. November 16 ‘Am. J Obtet Goma Unfortunately, itis not always made clear in texbot presentations that Streeter's data are for the en bryologist’s use to estimate gestational age from te weight of arfixed specimen and not for clinica dealing with nonfixed specimens. Many of the values reported for fetal weight and growth during the latter part of pregnancy are nt satisfactory for use in today’s U.S. A. population Feal weights derived many years ago* were lower and there fore are unsatisfactory for today’s standards. Nedin weights in India* and Singapore’ were consent lower, and those in Sweden,* Holland, and Scotand* were higher, whereas those reported from Cana" parts of the U. S. A..3~! and Austria® were siibrio subgroups in this study. National standards change and periodically need revision. Some methods, suds tuse of birth certificate data and the inclusion of weighs of dead fetuses or only fetuses that live throughout te neonatal period, result in inaccurate fetal weight sit dards. Standards for fetal weight derived from very pcic populations are usually satisfactory for only that pop lation. Among the indigents in Atlanta, Negroes hd lower male and female birth weights at each gestaioral age after 37 weeks’ gestation than did Caucasians ‘The median birth weights of Negroes were sii only the nonwhite subpopulation in our study. Poke bly the most frequent standard of fetal weight seit the U.S. A. is the one reported by Lubchenco ant associates." Using various “corrections,” they repoted combined weights of infants born to 5,635 indget Spanish American patients in Denver, Colorado (ck tion 5,280 feet). Their 90th, 50th, and 10th percente weights after 30 weeks’ gestation were about 100 grams less than those reported ini our study. Growth curves are potentially tiseful for seven! purposes in. the clinical practice of obstetrics 2nd gynecology and pediatrics as well asin public heath and the social sciences. Adequate fetal growth suit dards may be used as a basis for more informed cial judgments and for better interpretation of the resuts Of fetal monitoring. For example, in complicated preg ‘es, when evaluating the best time for delve, ination of -the expected ‘fetal weight 2 te Present gestational age and the expected weight git ‘with each: duration of delay ‘may be helpful. Seid determinations of fetal: growth by ultrasound axl other methods, if correlated with fetal weight, couldte used to assess growth retardation and the effects of i- ferent therapies. Although neutrates weighing les that 3.5 pounds and/or with “severe intrauterine grom reurdation” have higher rates of iorcaliy and sterYolume 126 Kanter rated somatic, psychological, and intellectual growth as compared to their peers,” * better definition of thehigh- and low-risk neonate groups may be possible with the availability of accurate growth standards. Growth curves may: be useful for selecting groups that might benefit from specific public health and ‘educational programs and for partially evaluating the effecs of such programs. Nutritional and special cducational and psychological adjustment programs forthe very premature and severely growth-retarded infants may decrease their mortality rate, improve their psychological adjustment, and promote their ‘maximum somatic and intellectual growth. More ap- propriate and successful treatment and education pro- sams may be devised if concentrated follow-up pro- rams of high-risk children are created to detect neu- rologica, speech, hearing, eye, and behavior problems. Theeffects on fetal growth of genetic factors, illnesses, and personal habits such as smoking, diet, and addic- son cat be evaluated so that potential parents cat be counneled and effective public health programs eau be inated. Serial fetal growth rate curves uver several years wuld be used (0 partially evaluate specific nutri- tonal, medical cane, and educational programs. large, conuolled, long-term comparative studies of various methods of managing pregnancies with i trauterine growth retardation from different causes andat various stages of gestation need to be conducted. ‘The-most frequently used method of managing the gronth-retarded fetus is cither by the continued obser- vation of the pregnancy until spontaneous labor oc- ‘urs, or by effecting early delivery if there is presump- live exidence that continuation of the pregnancy will Permanently injure the mother or there is a high probability that fetal death is imminent. However, there are reasons for believing that this approach rather than earlier delivery and adequate neonatal hutrition may result in less than optimal development in some fetuses. Reduced intrauterine growth rate is probably assaci- ated with suboptimal fetal development. After a tran- sient decrease in fetal growth while adapting to. the extrauterine environment. the neonate gains™ weight ta rate similar to the maximum rate noted during inrauterine development (Fig 7). Intrauterine growth retardation appears frequently to stem froma deficient intrauterine environment rather than an abnormality in the fetus since (1) after birth, low-birth-weight Aconates often gain weight at a more rapid rate than their normal or high birth. weight peers; (2) the inrauterine growth rate of-each fetus in multiple Pregnancies is negatively related to the number of Fetal growth standard for United States 563 fetuses", (8) fetuses of malnourished mothers * ** and/or pre-eclamptic mothers! have attenuated ‘growth; and (4) in some populations the maximum rate of fetal growth continues into the latter part of preg- nancy? ‘The anatomical and functional effects of human intrauterine growth retardation are unknown. How- ever, the decreased fetal brain growth that has been reported in animals when there are maternal nutri- tional deficiencies in the latter part of pregnancy may also occur in some growth-retarded human fetuses.**** Defects resulting from undernutrition during devel- ‘opment may not be repaired even if adequate nutrition is later resumed. Fetuses of undernourished bitches and mice have smaller brains and a decreased number of central nervous system neurons than the offsprin of well-nourished animals." This decrease neuroris persists into later life eveu hough there has been good neonatal nut ion * Some investigators believe that this decrease in anatomical growth docs ‘not imply a functional deficiency. Although intrauterine growth retardation results potentially in less than the optimal neurologic and somatic development, the functional deficit may not be demonstrable by the use of standardized tests jormal” populations. Although the mean somatic for size and intellectual and behavioral ratings of in- trauterine growth-retarded infants are within “nor- mal” ranges by 5 years of age, they are consistently less than those of their “peers."! Neurological deficiencies stemming from intrauterine growth. retardation may not be adequately defined by 1Q scores. OF 96 children, born with “severe” growth retardation afier 38 weeks! gestation, the mean 1Q was 95 for males and 101 for females, but 50 per cent of males and 36 pér cent of females had “poor” school performance: 59 per cent of males and 69 per cent of females had speech defects: 1 per cent had cerebral palsy: 6 per cent had convul- sions: and 25 per cent had evidence of cerebral dys- function.= In consideration of such- findings it appears that delivery of the growth-retarded fetus at a time when there isa high probability of extrauterine survival may result in less permanent damage to the infant than does the present practice of continued observation until there is evidence of maternal danger or impend- ing fetal death. However; appropriate studies will be eee to determine the best method of managing ts at different gestational ages with different Hlegrecs of intrauterine growth tetardation from dif ferent causes and for managing the. growth-retarded neonate.