Current Age of Onset of Puberty

To the Editor. The appearance of breast and pubic hair development in girls has long been considered to be premature if before 8 and 9 years of age, respectively.1 Recently the Drug and Therapeutics and Executive Committees of the Lawson Wilkins Pediatric Endocrine Society reviewed the data on which to base recommendations for the diagnostic work-up and therapy of precocious puberty.2 They were concerned that current recommendations may be based on outdated standards for the definition of the normal age of puberty. They concluded that “recent data demonstrate that in the United States, the onset of puberty in girls is occurring earlier than previous studies have documented, with breast and pubic hair development appearing on average 1 year earlier in white girls and 2 years earlier in African-American girls.” Taken together with recent data on the impact of sexual precocity on psychosocial development and on height potential and data on the limitations of treating children with gonadotropin-releasing hormone agonists after the age of 6 years, they concluded that aggressive evaluation and treatment are unlikely to be beneficial in girls with the onset of puberty after 6 (African-Americans) or 7 (white) years of age.However, many of us in the field of pediatric endocrinology believe that it is premature to conclude that the normal age of puberty is occurring earlier. The conclusions cited above that “demonstrate” an earlier age of puberty were based solely on the paper by Herman-Giddens et al, which reported the age at which these pubertal milestones were present on visual inspection in the practice of a large number of pediatricians' offices.3 We are of the opinion that the study has a fundamental flaw: the prevalence of the early signs of puberty was not ascertained in a sample of children drawn at random from the general population. Thus, the study is prone to ascertain bias. How might this have arisen? One can postulate that a number of girls are brought to their pediatrician with the concern about early onset of breast buds or pubic hair as a hidden agenda, not a stated complaint. The committee report has proved in depth the reasons why the study's conclusions may or may not be valid. However, in support of the possibility that the study may be flawed was the notable finding in the same practices that age of menarche averaged 12.16 years in African-Americans and 12.88 years in whites; this is not clearly different from that reported over 25 years ago, at which time menarche occurred at 12.7 ± 1.0, SD, years, with African-Americans approximately .1 years earlier and whites approximately 1.0 years later than the average.4 Thus, the Herman-Giddens report carries the connotation that puberty is beginning earlier, but that its tempo is slower, for which there is no explanation. Furthermore, liberalizing the definition of normal carries the risk of overlooking pathology. Therefore, we are of the opinion that a well-designed study is necessary before a conclusion can be drawn about the normal age of puberty.We are in agreement, however, with the general thrust of the committees' statement. That is, we agree that pubertal development occurring after the age of 6 years often is slowly progressive and generally does not have a serious cause or require treatment. However, this does not mean that puberty at this age is normal. Indeed, evidence is emerging that premature thelarche and premature pubarche may on occasion be risk factors for subsequent reproductive endocrine system dysfunction.5 When breasts or pubic hair appear before 8 or 9 years of age, respectively, regardless of race, we are of the opinion that a diagnostic evaluation should be initiated with evaluation of bone age and height prediction. Further work-up and management should then be based on these results, as well as the other findings in the history and physical examination, which have been outlined in the committees' report, on an individualized basis.In Reply. Now that the paper by Drs Kaplowitz and Oberfield and the Drug and Therapeutics Executive Committees of the Lawson Wilkins Pediatric Endocrine Society recommending new age limits for defining precocious puberty has been published,1 Drs Rosenfield et al have written concerning the issue of “ascertainment bias” in our 1997 study on the age of onset of puberty in 17 000 young girls.2 We addressed this issue in the “Discussion” section of our paper. Although the subjects were not randomly selected and, therefore, might not be representative of the population at large, we went on to note that the large numbers of girls studied would make it unlikely that they were different from the population as a whole. Furthermore, we stated that “the findings of this study need to be confirmed in other research including a nationally representative sample such as HANES. Until such studies are done, our data offer the best information about young girls in the United States.”2 Thus, we feel our data provide a far better estimate of the age distribution of the onset of puberty in US girls than the Marshall and Tanner study of 192 institutionalized white British girls published in 1969 upon which the 8-year cutoff has been based. We also pointed out that the girls could have been brought in selectively because of a problem with puberty, one of the mechanisms for selection bias discussed by Drs Rosenfield et al. We further stated that if such a bias had occurred we should have expected to see the same bias operating among parents of the older girls with no development, leading to a decrease in the prevalence of secondary sexual characteristics in that age group, a finding which did not occur.2 For such a “hidden agenda” to affect the findings, the majority of parents would have had to be reluctant to bring up a concern about puberty even though that was the reason for the visit. It seems unlikely that so many would be reticent to discuss their reason for bringing a child to their pediatrician.We noted that the age of menses had not dropped for white girls over the past 45 years and had dropped by several months for African-American girls since MacMahon's analysis of HANES data from the 1960s.3 Our colleagues correctly noted that the implications of our study are that the tempo of puberty is slower. The purpose of our study was not to offer explanations for any of the findings, but simply to note the proportions at a given age with secondary sexual characteristics and menses from a large population of girls. Earlier puberty has been noted to be associated with a longer duration until menses.4 It is interesting to note that the lengthening tempo has also been noted among Hong Kong girls. With 10% of their study population now with stage 2 breast developmentbefore the age of 8, the mean duration from breast budding to menses for Hong Kong girls is 6 months longer than it was in the early 1960s.5We agree that puberty at an early age may not be “normal” even though a large proportion of girls are experiencing it because factors that may be contributing are not yet understood. We do not dispute that some girls with early signs of puberty may be at risk for subsequent reproductive dysfunction. However, we question the recommendation that all girls with the sole factor of breast development or pubic hair growth before 8 and 9 years of age, respectively, have a diagnostic evaluation. If this recommendation (which reflects standard practice before the recent revised guidelines) was followed, about 8% of all white girls in this country and 34% of all African-American girls would need such an evaluation. We believe the new recommendations for the evaluation of early puberty1 are sound. They were approved by the leadership of the Lawson Wilkins Pediatric Endocrine Society in large part because 7- to 8-year-old white and 6- to 8-year-old African-American girls are commonly referred to endocrinologists for pubertal changes and are rarely found to have a pathologic cause.