Navigating the Recent Articles on Girls’ Puberty in <i>Pediatrics</i>: What Do We Know and Where Do We Go from Here?

After the publication of the Pediatric Research in Office Settings (PROS) study on the age of onset of pubertal characteristics and menses in US girls in 1997,1 a spate of related articles have appeared on emerging questions and controversies over recent pubertal data and the implications of these findings for clinical practice. The purpose of this commentary is to 1) summarize the consistencies and contradictions among some of these newer communications, 2) address misconceptions and misinterpretations of the PROS data, and 3) identify legitimate points of disagreement and areas for additional investigation.A survey of just some of the recent articles demonstrates the scope of additional research both in our country and abroad.2–17 The 1997 PROS study, a convenience sample of 17 077 white and black girls seen in pediatric practices across the United States and Puerto Rico used the Tanner method18 to describe the ages of onset of breast development, pubic hair growth, and menarche. It found that the mean ages for these characteristics varied significantly between white and black girls (with black girls being at younger ages), the median age of menarche for black girls had dropped over the past several decades, and the ages for the onset of development seemed to be earlier than previous US studies as well as Marshall and Tanner’s classic 1969 study.18 The PROS study pointed out that the prevalence of secondary sexual characteristics in girls <8 years old was substantially higher than what had been believed previously and “that more appropriate standards for defining delayed and precocious puberty may need to be developed, that the timing of sex education in the schools may need revision, and that the etiology and effects require further study.” The authors stated, “The findings of this study need to be confirmed in other research including a nationally representative sample such as HANES [Health and Nutrition Examination Survey].”1 After the PROS study, Kaplowitz et al2, using its data, provided additional analyses and new recommendations calling for the age for referral for precocious puberty to be lowered.Between October 2002 and April 2003, Pediatrics alone has published 10 articles on puberty markers or issues.19–28 Several of these articles beg for comment, in particular the articles that propose changes in practice or present interpretations of findings that contradict those of other recent articles. Six of the articles have been based wholly or in part on the most recent National Health and Nutrition Examination Survey (NHANES) data, and some present overlapping results or conflicting conclusions.20,21,23–25,28The October 2002 article by Wu et al20 analyzed data from the NHANES to report on ethnic differences in secondary sexual characteristics and menarche. The authors presented mean ages of onset for breast and pubic hair growth and for menses by race and ethnicity as well as odds ratios of having attained pubertal milestones among the 3 racial/ethnic groups studied in the NHANES. Tables 1 and 2 compare these results with those of the PROS study1 and the analyses of the NHANES data for average ages of onset of breast and pubic hair growth and menses by Sun et al24 and Chumlea et al25, respectively. Age at menarche was estimated by Wu et al by both the status quo method as well as an estimate based on the self-reported age using a failure time model, both under the assumption of a normal distribution of the event (Table 2). Their mean ages for menarche differ slightly from those of the Chumlea et al analysis (see below) of the NHANES data published in January 2003 because of different statistical methods. Wu et al concluded that black girls enter puberty earliest, followed by Hispanic and then white girls. Numerous studies, including the 1997 PROS study, have found earlier puberty among black girls. The Wu et al analysis provides the important additional information that racial and ethnic differences among the NHANES populations are independent of select social and economic factors.In the same issue of Pediatrics, the article by Freedman et al22 looked at the relation of age at menarche to race, time period, and anthropometric dimensions by using the Louisiana population followed in the Bogalusa Heart Study. Their assessment of secular trends in menarcheal age between 1973 and 1994 found that the mean menarcheal age decreased by 9.5 months for black girls and 2 months for white girls over the 20-year time period. As in other studies, they also found that black girls matured earlier than white girls.The November 2002 article by Sun et al, “National Estimates of the Timing of Sexual Maturation and Racial Differences Among US Children,”24 used NHANES data to look at ages at entering a sexual maturity stage as well as being in the stage for both boys and girls and by race and ethnic groups, whereas the Wu et al20 article reported only on mean ages at entering a stage (Tables 1 and 3). As would be expected, the results of these 2 articles are very similar, and the authors stressed that the degree of racial differences requires separate normative reference data, a recommendation with which we agree.The NHANES data support the findings of the PROS study and, for the most part, show close or similar results, although the PROS population was a convenience sample of girls being seen in pediatric practices across the country, as compared with the nationally representative population sample from the NHANES. Examination of Table 3 shows confidence intervals from the PROS data overlapping proportions from the NHANES data in a number of cells. Where results are not overlapping, in some cells the NHANES reports a higher proportion by race and age of girls with a particular characteristic than in the PROS study but in others reports a lower proportion (selected and weighted to be nationally representative, sample sizes for breast and pubic hair examination [ages 8–13] for each race and age ranged from 41 to 77 for the NHANES; similarly, the cell ranges for the convenience sample for the PROS study were 91–1334).Sun et al24 proposed that the results of their NHANES analysis provide “normative reference data.” Although we are focusing our comments on girls, this recommendation with regard to boys especially should be questioned, given the very young ages at which 25% of all US boys were found to have begun puberty (≤8 years old) and the lack of any discussion about data quality.Several statements in their article are unclear: In January 2003, Chumlea et al published “Age at Menarche and Racial Comparisons in US Girls,”25 which also used data from the NHANES study and shares many of the same authors as the previous article. This article repeats the analysis on menarcheal age by the Wu et al20 article, with some differences in technique. The former calculated the average age of onset of menses both using probit and failure time methods, and the latter used the percentage of girls having begun menses during 3-month age groupings in a probit model. Chumlea et al also present their results in a format differing from Wu et al, including probit plots and the ages at which 10%, 25%, 50%, 75%, and 90% of girls by race had reached menarche. As reported by Wu et al, they found that black girls began menses earlier than white and Mexican American girls. Statistically significant racial differences were reported by P values and 95% confidence intervals. The authors calculated a median overall age for onset of menses of 12.43 years of age for NHANES girls and compared it with the overall of age of 12.76 years from the MacMahon29 analysis of the National Health Examination Survey (NHES) conducted ∼25 years ago, a difference of 4.0 months. (For the NHANES, “overall” included white, black, and Mexican American girls; for the NHES, “overall” included white and black girls.) The authors concluded that, “overall, US girls are not gaining reproductive potential earlier than in the past.” In contrast, the April 2003 article in Pediatrics by Anderson et al28 presented their own analyses of NHES and NHANES menarcheal data, focusing on secular trends and the influence of weight and race on the age of onset of menses. These authors concluded that the average age of menarche had dropped by 0.21 years (∼2.5 months) during this 25-year time period and that it represented a statistically significant secular trend (Table 2).Several points in the Chumlea et al article deserve additional comment.Also in the January 2003 Pediatrics issue was the article by Midyett et al, “Are Pubertal Changes in Girls Before Age 8 Benign?”26 This article presented data on 223 patients referred for precocious puberty to Children’s Mercy Hospital in Kansas City, Kansas, during a 5-year period. The patients were picked to include white girls 7 to 8 years old and black girls 6 to 8 years old, because this represents the group that might not have received an endocrine evaluation if the age limits for evaluation proposed in 1999 were followed.2 They found that 105 girls (47%) had both breast and pubic hair and defined this group as having “true precocious puberty.” After finding that 12.3% of the patients had other endocrine conditions, most commonly acanthosis nigricans/hyperinsulinism (n = 15) or hypothyroidism (n = 4), the authors recommended that “all girls with any secondary sexual development before 8 years of age deserve at the minimum a bone age assessment and close longitudinal follow-up.”Although the generalizability of the results to all girls in the United States, as the authors advocate, can be questioned as well as certain other aspects of the study, we restrict our comments to the points below.We do not take issue with the fact that some girls with pubertal signs between 6 and 8 years of age will have other endocrine pathology, although acanthosis nigricans may be seen in very obese children at any age unrelated to the presence or absence of signs of puberty. No doubt, some children >8 years old have endocrine problems such as the ones found in the Midyett et al study, and they may not be identified at that age because they have fallen out of the recommended age for evaluation for many years. No system of screening is perfect. Primary care physicians should not blindly use the age of a child as a rote guide for referral. A careful history, examination, and follow-up looking for evidence of rapid progression of pubertal changes and growth acceleration should always be used in conjunction with age guidelines.The age of pubertal events is important individually, socially, culturally, and as a public health indicator. The rise and fall of the age of onset of secondary sexual characteristics, the attainment of gonadal maturation, and the age of complete sexual maturity may serve as the “canary in the mine” for environmental problems just as height and weight indices already do in individual cultures and countries. Growth data have been shown to be sensitive to times of stress, war, and famine, to cultural changes affecting diet and lifestyle, and to times of prosperity. In the same way, given the remarkable changes over the last few decades in the environment with regard to factors suspected of affecting puberty such as endocrine disrupters7,8,15,17 and the epidemic of overweight and obesity,3,11,14,23,27,28 pubertal events may be a sensitive indicator of conditions that are not healthful. We need to pay attention to these markers, and therefore, we find the recent pubertal studies and emerging questions and issues gratifying, especially as they relate to the quality and accuracy of pubertal measurements and clinical practice.It is important to recognize that there are differences in the pubertal markers “age of onset of secondary sexual characteristics” and “age of menarche.” The timing of these events may be regulated differently, making it important to study both. Some data have indicated that the earlier girls begin the onset of secondary sexual characteristics, the longer the time period until menarche is reached.32–34 Related to that, it is interesting to note that, in the recent study of pubertal characteristics in children with cerebral palsy by Worley et al,21 white girls began puberty earlier than the general population (25% of girls had Tanner stage 2 or greater breast development by 8.1 year of age) but did not reach menarche until later than average, at 14.0 years.We feel it is inaccurate to posit that “girls are not gaining reproductive potential earlier than in the past,”25 given the US studies suggesting otherwise.1,2,3,22,28 The cost of evaluations alone indicates the need for additional examination of guidelines for referral for early onset of puberty in boys and girls, how factors in our environment may be affecting pubertal development, and additional studies, especially on boys, to delineate the age at which pubertal markers are occurring.