ORIGINAL CONTRIBUTION Exposure to Soy-Based Formula in Infancy and Endocrinological and Reproductive Outcomes in Young Adulthood Brian L. Strom, MD Context A large body of evidence documents the role of phytoestrogens in influ- Rita Schinnar, MPA encing hormone-dependent states. Infants fed soy formula receive high levels of phy- Ekhard E. Ziegler, MD toestrogens, in the form of soy isoflavones, during a stage of development at which permanent effects are theoretically possible. However, a paucity of data exists on the Kurt T. Barnhart, MD long-term effects of infant soy formulas. Mary D. Sammel, ScD Objective To examine the association between infant exposure to soy formula and George A. Macones, MD health in young adulthood, with an emphasis on reproductive health. Design, Setting, and Participants Retrospective cohort study conducted from March Virginia A. Stallings, MD to August 1999 among adults aged 20 to 34 years who, as infants, participated dur- Jean M. Drulis, BA ing 1965-1978 in controlled feeding studies conducted at the University of Iowa, Iowa Steven E. Nelson, BA City (248 were fed soy formula and 563 were fed cow milk formula during infancy). Main Outcome Measures Self-reported pubertal maturation, menstrual and re- Sandra A. Hanson, BA productive history, height and usual weight, and current health, compared based on OY ISOFLAVONES, A SUBCAT- type of formula exposure during infancy. egory of phytoestrogens, are Results No statistically significant differences were observed between groups in ei- naturally occurring plant com- ther women or men for more than 30 outcomes. However, women who had been fed pounds. Phytoestrogens have soy formula reported slightly longer duration of menstrual bleeding (adjusted mean dif- Sbeen shown to bind to estrogen recep- ference, 0.37 days; 95% confidence interval [CI], 0.06-0.68), with no difference in se- tors in the adult1-3 and to act either as verity of menstrual flow. They also reported greater discomfort with menstruation (un- estrogens4 or as antiestrogens.5,6 For ex- adjusted relative risk for extreme discomfort vs no or mild pain, 1.77; 95% CI, 1.04-3.00). ample, phytoestrogens can block the ac- Conclusions Exposure to soy formula does not appear to lead to different general tion of endogenous estrogens on the health or reproductive outcomes than exposure to cow milk formula. Although the uterus.5-7 few positive findings should be explored in future studies, our findings are reassuring about the safety of infant soy formula. A large body of evidence docu- ments the role of phytoestrogens in JAMA. 2001;286:807-814 www.jama.com influencing hormone-dependent 4,8-10 women with differences in menstrual recent work has not confirmed this states. Dietary phytoestrogens dur- 27 28 ing adulthood have been suggested in cycle length. Lowered sperm counts pattern. numerous epidemiological studies to be have also been suggested as possibly as- Infants fed soy formula receive rela- protective against cancer of the pros- sociated with phytoestrogens, but more tively high doses (per unit of body tate, colon, rectum, stomach, breast, Author Affiliations: Center for Clinical Epidemiol- Mr Nelson, and Mss Drulis and Hanson). and lung and to exert similarly protec- ogy and Biostatistics and Department of Biostatistics Financial Disclosure: The Fomon Infant Nutrition Unit tive effects against chronic conditions and Epidemiology (Drs Strom, Barnhart, Sammel, and receives research grants from and conducts clinical tri- Macones and Ms Schinnar), Division of General In- als funded by Ross Products Division, Nestle, and Mead such as atherosclerosis and osteoporo- ternal Medicine, Department of Medicine (Dr Strom), Johnson Nutritionals. Dr Ziegler has received hono- sis.11-15 Evidence in some animal Department of Obstetrics and Gynecology (Drs Barn- raria for speaking from the same companies, mostly hart and Macones), and Division of Gastroenterol- in continuing education settings. species of a contraceptive or sexual de- ogy and Nutrition, Children’s Hospital, and Depart- Corresponding Author and Reprints: Brian L. Strom, velopment effect from dietary phytoes- ment of Pediatrics (Dr Stallings), University of MD, University of Pennsylvania, Center for Clinical Epi- 16-26 Pennsylvania School of Medicine, Philadelphia; demiology and Biostatistics, 824 Blockley Hall, 423 trogens has led some to suggest that and Fomon Infant Nutrition Unit, Department of Guardian Dr, Philadelphia, PA 19104-6021 (e-mail: dietary habits should be investigated in Pediatrics, University of Iowa, Iowa City (Dr Ziegler, [email protected]). ©2001 American Medical Association. All rights reserved. (Reprinted) JAMA, August 15, 2001—Vol 286, No. 7 807 SOY-BASED FORMULA AND HEALTH OUTCOMES weight) of phytoestrogens during a pleted, all available infants were as- phone interview that took 30 to 60 developmental stage at which perma- signed to the next study formula. Infant minutes to complete. Reliability checks nent changes are theoretically pos- characteristics or parental preferences were done on 5% of questionnaires. Out- sible.29-31 Delayed effects of soy-based were not taken into account. The same come variables were selected for inves- infant formula on subsequent child or formula was fed for the entire study pe- tigation a priori based on 3 criteria. They adult health have thereby been postu- riod. Selected solid foods were received were expected to be (1) potentially re- lated,30,31 generating substantial con- by some infants. The study protocols in- lated to estrogenic effects, (2) clinically troversy in the lay and medical press. volved periodic measurements of weight important, and (3) likely to have a suf- However, effects on human pubertal and and length, measurement of formula con- ficient sample size to detect a clinically reproductive development of phytoes- sumption, and determination of se- relevant effect, based on a priori stan- trogen exposure in infancy have not lected serum chemical indexes. The ac- dard type I and type II error levels. Be- been systematically investigated.13 tual phytoestrogen content of the infant cause potential estrogenic effects may be Given the paucity of data on the long- formulas was not measured. different in men and women, and many term effects of soy formulas, their wide- Criteria for inclusion in the present reproductive and sexual outcomes are spread use, and expressed concerns follow-up study were that the formula sex-specific, these outcomes were se- about their safety,30,31 we undertook a fed be classifiable as “milk based” or “soy lected and analyzed separately by sex. study to examine the association based,” ie, containing protein derived The outcomes chosen for primary between exposure to soy formula in from cow milk or isolated soy protein analysis in women were adult height, infancy and any subsequent possible or, in the case of one formula (12 sub- usual weight since the age of 18 years, effects on adult health, focusing on out- jects in follow-up), from soy flour. In- usual body mass index, pubertal matu- comes that could be estrogen related. dividuals who received formulas with ration (age at menarche, age when both soy and cow milk protein were ex- breasts developed enough to start wear- METHODS cluded. In addition, subjects had to have ing a bra), number of days between pe- Study Design completed the original feeding study as riods (during times when not using birth This was a retrospective cohort study of planned (about 85% of enrolled infants). control pills, shots, or implants), num- young adults, 20 to 34 years of age, who Those who were adopted were ex- ber of days requiring pads or tampons, as infants had participated in multiple cluded from follow-up. Also excluded regularity of menstrual period, men- controlled but nonrandomized feeding were individuals who were profoundly strual flow, pain with menstrual period studies conducted at the University of disabled who would not have been able (none, mild, severe), physical symp- Iowa during the years 1965 to 1978. A to be interviewed, deceased individu- toms of pain (eg, dysmenorrhea, head- telephone interview was conducted be- als, and individuals from countries other aches), breast tenderness during men- tween March and August 1999 with eli- than the United States, because of dif- strual cycle, premenstrual symptoms, gible subjects who could be located and ficulty in tracking and differences in life- breast size (bra cup size), reproductive who agreed to be interviewed. The study style and language barriers. outcomes (number of pregnancies, de- was approved by the University of Penn- liveries, abortions, and other complica- sylvania Committee on Studies Involv- Subject Follow-up tions), and education level attained as ing Human Subjects. A search was conducted to locate the a proxy measure for intelligence. 952 subjects of the original cohort. The In men, in addition to adult height, The Cohort search used a variety of approaches, in- usual weight, and education level, out- Methods and procedures used in the cluding the use of national telephone comes investigated in the primary original feeding studies have been pre- and address directories. Records of the analyses included pubertal matura- viously described.32,33 Briefly, partici- parents’ and often grandparents’ names tion (age at first ejaculation, age when pants were healthy term infants with and addresses were also available from voice changed, age when hair began to birth weights of more than 2500 g whose the original study. Efforts to locate sub- grow on chest, face, or pubic area) and mothers elected not to breastfeed. With jects were aided by the fact that many pregnancy outcomes in sexual part- few exceptions, they were white, reflect- of the individuals born from 1966 ners impregnated by the male study ing the population in and around Iowa through 1971 had been located when subjects. Other outcomes, such as con- City. Infants were enrolled before the age they were 8 years old as part of a pre- genital malformations in the offspring of 9 days and were studied through 16 vious follow-up study.34 of study subjects, hormonal disor- weeks of age.