Document o9Qboa0GQvZMRzJY3R57w3N1E

P X 2 & /7 352 THE NEW ENGLAND JOURNAL OF MEDICINE AuR- 5, H DECREASED LEVELS OF HELPER T CELLS A Possible Cause of Immunodeficiency in Pregnancy VIT AYa S r id a m a , M.D., F u r io P a c i n i , M.D., S en-L ian Y a n c , M .D ., A t e f M o a w a d , M.D., M a u r e e n R e i l l y , B.S.N., a n d L e s u e J. D e G r o o t , M.D. Preparation of Peripheral-Blood Mononuclear Celia Heparinized peripheral blood (20 ml) was fractionated by l standard Ficoll-Hypaque gradient technique. Monoclonal Antibod l*s Murine monoclonal antibodies termed OKT3, OKT4, OKT^ffij O K M l, and O K U (Ortho Pharmaceutical, Raritan, N J.) wetefr* used in this study. Functionally, OKT3 defines all peripheral X cells,* OKT4 defines the human hdper/induccr T-cell subset,* OKT8 defines the suppressor/cytotoxic T-ccll subset,* OKM l fines the monocytc/null-ccll population,7 and O K U defines B monocytes, and activated T cells.* 'T'VECREASED maternal immune responsiveness during pregnancy may partly explain the surviv al of the fetus as an allograft. It may also account for changes in disease activity and antibody production in several autoimmune diseases during pregnancy and in the post-partum period. Previous in vitro studies of cellular immunity in pregnant women have demon strated decreased lymphocyte responses to phytohemagglutinin,1 to mixed lymphocyte culture,1 and to soluble antigens.2An increased incidence and severity ofcertain viral infections has also been observed.1,3An imbalance between helper and suppressor T cells has recently been demonstrated in several autoimmune diseases4,5and in immunodeficiency states,4in studies using monoclonal antibodies that react against subsets of T cells. To determine whether abnormalities in T-cclI subsets might help to explain the immunodeficicnt state in pregnancy, we used monoclonal anti bodies to enumerate the proportion and absolute num ber of subsets of T cells and other peripheral-blood mononuclear cells from normal pregnant women, post-partum women, and newborns. We found a sig nificant decrease in relative and absolute numbers of helper T lymphocytes (T4+ cells) throughout preg nancy. .Normalization of T4+ cells occurred dur ing the third to fifth month post partum. Our findings support previous evidence of maternal immunode ficiency during pregnancy and indicate that it may occur because ofa reduction in the population of help er T ceils. M ethods Subjects The study groups included 76 normal pregnant women in various stages of pregnancy (eight in the first trimester, 15 in the second, 10 in the third, and 33 in labor), 13 post-partum women (eight in the fint or second month and five in the third to fifth month post partum), and 15 normal newborns. Thirty-four normal adult volun teers (12 men and 22 women), who had no history of taking hor mones or other medications, were used as controls for studies ofoord blood. Nineteen normal, age-matched, healthy female volunteers from this group were also used as controls for the pregnant women. From the Thyroid Study Unit, Department of Medicine, and the Department of Obstetrics and Gynecology, The University of Chicago. Address reprint requests to Dr. DeGroot it the University of Chicago Department of Medicine. Section of Endocrinology. 950 E. 59th St.. Oifcago. IL 60637. Supported by grams (AM I3J77 and AM 27JM) from the U.S. Public Health Service, Procedures For each subject, 1x10s peripheral-blood mononuclear cells tube were pelleted in five separate tubes, resuspended in 50 pd o f t h i i j appropriate dilution of O K T3, OKT4, OKTB, O K M l, or O K I a ^ l and incubated for 30 mimucs on ice. After two washes with F-10 medium (Microbiological Associates, Walltenville, Md.) oorirJ taining 10 per cent heat-inactivated fetal-calf serum (Grand Biological, Grand Island, N.Y.), they were incubated for 30 im n u ta ^ l on ice with a 1:30 dilution of a fluorescein-conjugated IgG fraction [S -'i goat antimouse IgG (heavy-chain and light-chain) (Cappd, VVcxf-5.j Chester, Pa.) that had previously been adsorbed with human IgG j'^ they were then washed three limes, fixed for 10 minutes in 4 percent1?.'! paraformaldehyde, mounted on microscopical slides, and examined^.) under ultraviolet light by means of cpfluorescence microscopy. each slide, approximately 200 cells were counted. All analyses w a ity ^ performed without knowledge of the source of the specimen. Mouscm o IgG was used instead of monoclonal antibody to determine badra&a ground staining. Uniformly less than 1 per cent of cells showed)^ fluorescent staining with this serum. Surface-membrane immunoglobulin-positive cells were enum ^ ated with fluoresccin-oonjugated polyvalent goat antihumac iroijj^j! munoglobulin (Bionctics, Kensington, Md.), as previously scribed.9 Before staining, monocytes were labeled by incubation qfgjY- cells in 0.02 per cent uniform latex particles in F-10 medium with lO ^ ; percent fetal-calfserum for 60 minutes at 37*C. Latex-ingesting cells5; were examined by phase-contrast microscopy. Approximately cells were counted on each slide. " ^ The white-cell count and the differential count were used to cal> late the number of positive cells per cubic millimeter. S latitile Student's two-tailed t-test was used to determine significance.^ .,r" R esults As shown in Figure l (top panel) and Table I,'this? percentage and absolute number of T4+ cells wi significantly decreased in normal pregnant womc throughout pregnancy and during labor. The abnoc-j malities were most, obvious among the women in thfc third trimester (29.1 6 .1 per cent and 543169 ceils? per cubic millimeter; P<0.01), as compared with^hc normal female controls (45.1 6.1 per cent 1073441 cells per cubic millimeter). The percentage ofT4+ cells was still decreased in the first and second? months post partum (34-3 --8.4 per cent; P<0.01) `an$ normalized during the third to fifth month post paw turn (39.62.9 per cent), whereas the absolute num' was already norma! in the first and second months poa partum (883 33! cells per cubic millimeter). The ccntagc and absolute number of T cells (T3+ cells (Fig. 2 and Table 1) were significantly decreased in th; second and third trimesters of pregnancy. The lowes 1.307 No. 6 MEDICAL INTELLIGENCE -- SRIDAMA ETAL. 353 3KT4, OK' a". NJ-)>. Uperipheral/ T -cd l tuba ^ a .* OKMl defina B udear odL'p a i:. j din 50/d e f ie d I '.MJ, of OKl&j? ' a with Hanvj?. Wile, Md.) (Grand d lor 30 min(^`. *IrG fracdoo {'< (Cappel, Wc*t^ ih human I r tf lea i*n $/ per c6ent.V. *nd examined'/^ ncrojcopy.For^ I analy*a w ai ecimcn. Monjc. i ctcrrmne back-l>>< !f cells showed^- were enumerf^.. uman ini-- .oujly doAj*' i'cidni*cuumbawtiiothn Ja0**>V';; -IngatingeelL^ -oximitcJy 200 : used toealcu-'SC" ' A ' signifiauK t^#S /iS fable 1, tti cells Wi int wome The abno: >men in wrft 3 169 celjs^ :d with thc^j;" cent ihd^; percentage^ and second#? <0.01) a i $ S J h post par^a--ite numbcCi nonths pas ). The 'T', + cells J in tliej i mc lowest Figure 1. Mean Percentage of T4+ Cells (HelperT Cells) and the T4+/T8+ Ratio in Normal Controls. Pregnant Women and PostPartum Women. Bars represent S.D.; figures in brackets, the numbers of subjects; and ns, no significant difference. values'were observed in the third-trimester group (50.34.9 percent and I052312 cells per cubic milli meter; P<0.01), as compared with the normal female controls (68.28.7 per cent and 1652627 cells per cubic millimeter). When the data were expressed as the T4+/T8+ ratio (Fig. 1, bottom panel), a signifi cant difference was found in the second and third tri mesters, during labor, and one to two months after delivery. The lowest ratio occurred in the third trimes ter (1.1 0.4; PC0.001), as compared with normal con trols (2.00.6). Normalization of the ratio took place during the third to fifth month post partum (1.70.3). The percentage and absolute number of suppressor/ cytotoxic T cells (T8+ cells) (Table 1) were not signifi cantly different from those in normal female controls. There was a significant increase in the percentage of B cells at the time oflabor (9.54.9 per cent; P<0.05), as compared with normal controls (7.0--2.6 per cent) (Fig. 3). No significant change in the absolute number of B cells was found throughout pregnancy (Tabic 2). As shown in Figure 3, the percentage of monocytes (MI + cells or latex-ingesting cells or both) was signifi cantly increased throughout pregnancy, with maxi mally increased M1+ cells at the time of labor (30.3 14.0 per cent and 759415 celts per cubic milli meter), as compared with normal controls (18.48.0 per cent and 469223 cells per cubic millimeter). The absolute number ofcells bearing la antigen (Ia+ cells) did not significantly change, and the percentage Table 1. Absolute Numbers of T3+. T4+. and T8+ Cells In Peripheral-Blood Mononuclear Cells from Normal Controls. Pregnant Women, and Postpartum Women. No. or Subject* Normal coniroli 1st trimester 2d trimester 3d trimester During labor lst-2d month post partum" Jd-Jih month post partum * M m *SD, u compared w ith norm al control*. |P < 0 .0 J u compared ih norm al control** ( P < 0 T0 0 l u com pared w ith n orm a! c o n tro l* . 19 B 15 10 33 E 5 M1AM A O t (R ange) J" 28(20-38) 26 (20-29) 27(I8-X ) 27(21-39) 26(16-38) 27(19-38) J3 (29-36) T3+ C ells * 1652 627 I2093W 1231 309 J I0J23I2 t IJ7567S 1302 321 173l584 T4 + C e lli * C efU fm m i l0734-41 662 203 t 736 232 f 543169$ 737352 t S833l 1 1087 387 T l+ C ells * 568209 546408 498H9 582293 654 421 620S6 72J 321 354 THE NEW ENGLAND JOURNAL OF MEDICINE Aug. 5, i; including the local mechanical barrier,14 block antibodies,14 and soluble suppressor factors from fetus15-- all leading to a successful outcome of p nancy. Several autoimmune diseases have been associate with low numbers of suppressor T cells (T84- cells and a high ratio of helper cells to suppressor cells/1 patients with these disorders become pregnant, a de3 5 40. Figure 2. Mean Percentage of T3+ Cells fTotaf T Cells) In Normal Controls. Pregnant Women, and Post-Partum Women. One asterisk indicates P<0.05, as compared with the mean value in 19 normal women; two asterisks Indicate P<0.001. Bars repre sent S.D. The numbers ol subjects are the same as in Figure 1. changed only in the third trimester (30.55.t per cent), as compared with normal female controls (I8.35.6 per cent) (Fig. 3 and Tabic 2). Studies of cord blood showed significantly de creased percentages of T3+ cells (52.4I6.I per cent; P<0.0!) and T8+ cells (16.77.6 per cent; P<0.01) and significantly increased percentages of M1+ cells (29.915.4 per cent), Ia+ cells (25.98.1 per cent), and B cells (13.64.4 per cent; ali P<0.001), as com pared with normal male and female adult controls (Tabic 3). D is c u s s io n Several studies have demonstrated decreased cellmediated immune responses in normal pregnant women, as measured by in vitro lymphocyte transfor mation,112However, evaluations of the percentage and absolute number of peripheral T lymphocytes in preg nant women have yielded conflicting findings, includ ing no change,10a decrease,11 and an increase.12This discrepancy in total T-lymphocytc enumeration may be due to variability in crythrocyte-rosetting methods. Moreover, recent evidence indicates that not all cells that arc capable of forming erythrocyte rosettes arc thymus-processed cells.13 In this study, we used a monoclonal antibody6 that specifically reacts against almost all T cells (OKT3) and found a significant decrease in the percentage and absolute number ofT lymphocytes (T3+ cells) in the second and third tri mesters of pregnancy. This decrease in total T cells may be explained by a decrease in the proportion and absolute numberofhclpcrTcells (T4+ cells) through out pregnancy. Our finding supports previous evi dence of decreased cell-mediated immune responsive ness during pregnancy, and this mechanism may be important in nonrejcction of the fetal allograft by the mother. Decreased immune responsiveness may oper ate in conjunction with other possible mechanisms, 30 O 20. + 10. 0. 50 ov P\ 40. 30. o + 20. 10. m T i1 0. 40 cnNP 30. UlJ o 20. + 10. 11 0 30 20, 1 M 1 1 r10. 0 " I-- Ti 1 .fid E5 .i . . : KTji o 1-2 3-5 ^ months c postpartum t .. . t r i me s t e r -- Figure 3. Mean Percentage of M1 + Cells (Panel A), Latex-Ingest- -' ing Cells (B). la+ Cells (C). and 8 Cells (D) in Normal Controls, Pregnant Women, and Post-Partum Women. One asterisk indicates PcO.OS, as compared with the mean value^jj^ In 19 normal women; two asterisks indicate P<0.01. Bara repre-^. sent S.D. The numbers of subjects are the same as In Table 2. \ .A . 307 No. 6 MEDICAL INTELLIGENCE -- SR1DAMA ET AL. 355 Table 2. Absolute Numbers of M1 + Cells, Monocytes (Latex-Ingesting Cells). Ia+ Cells, and B Cells in Peripheral-Blood Mononuclear Celts from Normal Controls. Pregnant Women, and Post-Partum Women. No. or Suuicrc Normal contrla In trimci 1er 2d trimester 3d trimester Durinf labor lit-2<J month poil partum Jd--Sth month poit pahum ` Meant S.D, fPCO.OJ u compared th normal controla. IJ 8 15 10 33 8 5 MI+ CeLU * 469 223 660 246 549 236 692255 t 759415t 676 24 5 646247 LaTia-IncaittNG Celli * ctih/mrxt 375*154 491*237 426I86 552* 197 f 68S489 t 476*21 362 90 la* Cells * BCelu * 461 200 J59 143 461 232 620|4B 547* 211 557*262 690408 165 4 6 I88 (40 155*81 159* MS 185*98 189*131 191 81 crease in helper T cells during pregnancy may lead to a lower hclpcr/supprcssor ratio and consequently to lower antibody production. This phenomenon may ex plain the finding of decreased autoantibody produclion during pregnancy in autoimmune thyroid dis eases'6 and systemic lupus erythematosus1' and may also explain the favorable course and clinical improve ment in some cases ofsystemic lupus erythematosus,18 rheumatoid arthritis,19 and Graves' disease20 during pregnancy. However, exacerbation and progression of systemic lupus erythematosus during pregnancy has also been observed.21 This may reflect ^thc natural course ofsystemic lupus erythematosus and a failure of the altered immune status during pregnancy to modify ihat course. A return to prc-prcgnancy levels of helper T tells in the post-parium period may explain the post partum exacerbation and rebound of autoantibody production in autoimmune thyroid diseases16,20 and systemic lupus erythematosus. 8,21 Increased rates of abortion and fetal loss have been observed in pregnant women with systemic lupus erythematosus*8 or Graves' disease.22 These findings may be explained in part by a decrease in T8+ cells in these patients,4*5 which could lead to higher helper/ suppressor ratios than arc present in.normal pregnant women; a low hclpcr/supprcssor ratio is probably important in the maintenance of normal preg nancy. However, other factors may also contribute to ihc high abortion rate among these patients-- for example, uremia, vasculitis, and the hyperthyroid state. A decrease in T4+ cells may be a consequence of the hormonal changes associated with pregnancy. Human chorionic gonadotropin, estrogen, progesterone, corti costeroids, a-fctoprotcin, prolactin, and a-globulin have been shown to,have immunosuppressive effects during in vitro studies.14 Our studies of B cells confirmed the previous finding that there is no change in the absolute numbers of B cells in pregnancy.10The increase in the percentage of monocytes that we noted during pregnancy may have been partly due to the decrease in the proportion of T cells. However, there was also an increase in the abso lute number of monocytes during pregnancy. The in crease in the percentage of Ia+ cells in the third tri mester may have been due to the increase in the percentage of monocytes. Recent studies ofsubpopulations ofT cells by meth ods identifying Fc receptors23 demonstrated increased T y cells and decreased T/i cells (which were thought to be suppressor and helper T cells, respectively) in normal pregnant women. However, several studies have demonstrated that T y and T^t cells arc not homo geneous populations of suppressor and helper T cells.12,24 A high proportion of Ty ceils react with OK.MI, lack T-ccll markers (the T3 antigen),24 and have high natural-killer activity.'3 Our findings sug gest that the immunodcficient stale in pregnancy is probably due primarily to a decrease in helper T cells (T4+ cells). Demonstrations of high suppressive activity of neonatal lymphocytes during in vitro studies led to the hypothesis that suppressor factors from the fetus might cause immunodeficiency in pregnancy. Our finding of decreased T8+ cells in the cord blood is in agreement with previous studies23 but argues against that hy pothesis. Evidence that T4+ and T8+ cells in cord blood arc functionally different from adult T cells25 Table 3. Percentages of Various Cell Types In Mononuclear Cells from Cord Blood of 15 Normal Newborns (Eight Boys and Seven Girls) and from Peripheral Blood of 34 Normal Adult Controls (12 Men and 22 Women).* C a u .T m T 3+ (%) T 4+(% ) T 8+(% ) T 4+/T S+ ratio M I+(%) Latex-ingesting U+{%) B exHi (%) A dult Co w tio u 65.48.6 43.07.5 24.:5,4 l.90.6 19.5*8.1 14.64.0 18.5* 6.8 7.3 4 .0 C o ld- B lood Samples 52.4 16.1 t 43.5 15.9 16.7*7.6 t 3.22.2 % 29.9 15.4 j I8.7 18.5 25.9*8.1 f 13 .6 * 4.4 } ` Muni iS.D. tP < 0 .O t c o m p lie d w ith adu lt c o n tfo li. J P < 0 0 i at com p ile d w ith id u lt control. ( P < 0.001 u c o m p ile d - i l h a d u lt co n If o il. 356 THE NEW ENGLAND JOURNAL OF MEDICINE Aug. 5, IS makes a full interpretation of the results of the cordblood study impossible. In conclusion, our finding of decreased helper T cells in normal pregnancy supports the evidence of immunodeficiency in pregnant women and may indi cate one mechanism allowing survival of the fetal allo graft during normal pregnancy. INTRAMUSCULAR OR INTRALIPOMATO INJECTIONS? W. P e t e r C o c k s h o t t , M.D., G e r a r d T. T h o m p s o n , M.B., LoisJ. H o w l e t t , a n d E l iz a b e t h T. S e e l e y , B.Sc.N., M.Sc.N. R eferences 1. Tbonf YH, Steele RW, Vincent MM, Hensen SA. Bcllanti JA. Impaired ta vitro edi-mediated immunity to rubella virus during pregnancy. N Engl J Mod. 1973; 289:604-6. 2. Smith iK, Caspar? EA. Field EJ. Lymphocyte reactivity lo antigen in preg nancy. Arts J Obstet Gynecol. 1972; 113:602-6. 3. D'Crui LA, Balanj SG. Iyer LS. Infectious Scpali(it and pregnancy. Obstet Gynecol. 1968; 31:449-33. 4. Bach M-A. Bach 1-F. The use of monoclonal anii-T cell antibodies lo study T cell imbalances in human diseases. G in Eap Immunol. 1981; 43: - 449-56. 3. Sridima V, Pacint F. DeGroot U . Decreased suppressor T-lyrophocyiei in autoimmune thyroid diseases detected by monoclonal antibodies, i Gin Endocrinol Meub. 1982: 54:316-9. 6. Reinherz EL, Kung PC. Goldstein G. Schlossman SF. Separation of func tional subsets of human T-cells by monoclonal antibody. Proc Nad Acad Set USA. 1979; 76:4061-5. 7. Breard /. Reinherz F i., Kung PC, Goldstein G, Schlossman SF. A mooocloaal antibody reactive with human peripheral blood monocytes. J Im munol. 1980; 124:1943-8. 8. Retnhcrz EL. Kung PC, Pesando fM, Ritz f, Goldstein G. Schlossman SF. la determinants on human T-ceJI sublets defined by monoclonal anti body: activation stimuli required for eapression, i Exp Med. 1979; 130: 472-82. 9. Alexander EL, Sanders SK. F (ab')j reagents ait not required if goat, nther than rabbit, antibodies arc used to detect surface immunoglobulin. J Im munol. 1977; 119:1084-8. 10. Dodson MG, Kerman RH, Lange CF. Sefant SS, O'Leary JA. T and B cells in pregnancy. Obstet Gynecol. 1977; 49:299-302. 11. Bulmer R. Hancock KW, Depletion of circulating T lymphocytes in preg nancy. Gin Exp Immunol. 1977; 28:302-5. 12. Clements PJ. Yu DTY, Levy J. Pearson CM. Human lymphocyte subpOfMlations: the effect of pregnancy. Proc Soc Exp Biol Med. 1976: 152: 664-6. 13. Beverley PCL, Calltrd RE. Distinctive functional characteristics of human T " lymphocyte* defined by E rescuing or a monoclonal anti-T cell anti body. Eur I Immunol. 1981; 11:329-34. 14. Lawrence R. Church JA. Richards W. Borzy M. Immunological mech anisms in the maintenance of pregnancy. Ann Allergy. 1980; 44;]66* 73. 15. Roclieh Cf. Goodwin 7$, Bankhum AD, Williams RC. Pregnancy, a tem porary fetal graft of suppressor cells in autoimmune disease? Am J Mod. __ 1980; 69-J29-3I. 16. Amino N. Kuro R. Tanizawa O. a ml. Changes o f serum anti-thyroid anti bodies during and after pregnancy in autoimmune thyroid diseases. Clin Exp Immunol. 1978; 31:30-7. 17. Bresnihxn B, Grigor RR, Oliver M. A al. Immunological mechanism for fpontanerxg abortion in systemic lupus erythematosus. Lancet. 1977; 2:1205-7. 18. Hanson GC, Ghosh 5. Systemic lupus oythematosus and pregnancy. Br Med J. 1965; 2:1227-8. 19. fcrsellin RH. The effects of pregnancy on rheumatoid arthritis. Bull Rheum DU. 1976; 27.-922-7. 20. Burrow GN. Maternal-fetal considrai ions in hyperthyroidism. Gin Endo crinol Meub. 1978; 7:115-23. 21. Zulman JI, Txlal N. Hoffman GS, Epstein WV. Problems associated with the management of pregnancies in patients with systemic lupus erythemato sus. J Rheumatol. 1980; 7:37-49. 22. Hawc P, Francis HH. Pregnancy and thyrotoxicosis. Br Med J. 1962:2:81722. 23. Hirahara F, Gormi 1. Tanaka K. Mauuzakt Y, Sumiyoshi Y, Shlojimi Y, Cellular immunity in pregnancy: subpopulalionx of T lymphocytes bearing Fc receptors for IgG and IgM in pregnant women. Clin Exp Immunol. 1980; 41:333-7. 24. Reinherz EL, Morrtta L, Roper M, A al. Human T lymphocyte subpopulationi defined by Fc receptors and monoclonal antibodies: a comparison. J Exp Med. 1980; 151:969-74. 23, Yachic A. MiyswakiT. NagaokiT, a al. Regulation of B cell differentiation by T ceil subsets defined with monoclonal OKT4 and OKT8 andbodies in human cord blood. J Immunol. 1981; 127:1314-7. VIEWING routine computerized-tomography (i scans of the pelvis suggests that in a North Arner^ ican population the majority of injections intended ttfyi be intramuscular arc actually delivered into fat. To confirm or refute this impression we studiedthe* thickness of gluteal fat in a hospital population. found that few female patients and under 15 per cent ofcr male patients do n fact receive an intramuscular mjec lion when a needle of the usual size is inserted into thL -- buttocks. The implications of this finding are discussMt below. M ethods To learn where (he needle is introduced in intramuscular injects' tions, we observed both student and graduate nurses. The d u n e area of the upper outer quadrant of the bullock was used in`a llj': instances by both groups. In simulated injections in which a b*U-*^ point pen was used to demarcate the injection site, 12 student nurses scattered their injections on the upper quadrant over a +.5-cm cirdefe when the patient lay prone and over a 5.6-cm circle when th c ^ patient was in an oblique decubitus position. Although the epicenteiji was the same distance from the iliac crest when the patient was mini cither of these positions, it lay 3 cm posteriorly when the patient waitg; in a scmiobliquc position. In adult patients who were about to have'di CT scanning or the pelvis, graduate senior nurses taped a lead shot^L to the skin to mark the position where they would give the injection^'. These shots were placed in the same demarcated area at Icvds^>varying from the lower border of the fifth lumbar vertebra to thej'v; lower part of the second sacral segment. It was found that the^v thickness of the prcmuscular fat did not vary significantly cither, vertically or transversely over the likely injection sites selected by both student and graduate nurses. We measured (he distance tween the skin and musdc at these levels and found negligible diiTer-'ij enccs in the thickness of the layer of fat, although the musdc bulk, diminished as the level selected approached the level of the iliac--.-- _, *i>4i The gluteal-fat thickness was then measured on the left tide oT.^ 213 adults (both inpatients and outpatients) scheduled for a C T"S examination that included the pelvis. We assumed that this sam ple'^ hospital population could be considered representative of all p-;S tients receiving intramuscular injections. The total of premuscular & fat -- both superficial and deep -- to the deep fascia at the SI level~U~ was measured in the 213 adults, and body weight was recorded' i in 123. ' iig R esults At any given weight women had a mean gluteal-fat' thickness nearly 2.5 cm greater than men, confirming'^ ihc obvious fact that women arc more stcaiopygous {Fig. 1). Tlie depth offal increased until the fourth-1*"'" decade of life and leveled off or diminished [hcreaficr\^i in the overall series of 213 adults. The needle generally used for intramuscular injec- tions in our hospital is 3.5 cm long. Observation haj'j^ .J From McMxxter University McdicxJ Centre xnd McMxJier University. Hamil- Ion. On!. Address reprint request! io Dr, Cockshott si the Department of Radiol- 'A ogy, McMaiter University Medical Centre, 1200 Msin Si. W,, Hamilton, ON U N 3Z3. Canada. ^