Reexamination of the Age Limit for Defining When Puberty Is Precocious in Girls in the United States: Implications for Evaluation and Treatment
In 1997 a study from the Pediatric Research in Office Settings network, based on pubertal staging of >17 000 girls between 3 and 12 years of age, indicated that breast and pubic hair development are occurring significantly earlier than suggested by our current guidelines, especially in African-American girls. In response to this article, the Lawson Wilkins Pediatric Endocrine Society undertook a comprehensive review of this topic. The primary conclusions of this review are:
1. The current recommendation that breast development before age 8 is precocious is based on outdated studies. Until 1997, no data were available on pubertal staging in US girls that could have documented a trend to earlier maturation.
2. The 1997 study indicates that stage 2 of breast and pubic hair development is being achieved ∼1 year earlier in white girls and 2 years earlier in African-American girls than previous studies have shown.
3. Concerns that girls with moderately precocious puberty will be significantly short adults are overstated; most have adult height within the normal range.
4. Therapy with gonadotropin-releasing hormone agonists has not been proven to have a substantial effect on adult height in most girls whose puberty starts between 6 and 8 years of age.
5. New guidelines propose that girls with either breast development or pubic hair should be evaluated if this occurs before age 7 in white girls and before age 6 in African-American girls. No changes in the current guidelines for evaluating boys (signs of puberty at younger than 9 years) can be made at this time. normal puberty, breast development, pubic hair.
- BMI =
- body mass index •
- GnRH =
- gonadotropin-releasing hormone •
- SD =
- standard deviation
The 4th edition of Wilkins' textbook The Diagnosis and Treatment of Endocrine Disorders in Childhood and Adolescence states: “By definition isosexual precocity is development of secondary sexual characteristics along female lines in girls who are less than 8 years of age and along male lines in boys who are less than 9 years of age.”1However, a recent article based on examinations of 17 000 healthy 3- to 12-year-old girls seen in the private office setting suggests that puberty is occurring significantly earlier than in the past.2 Since the publication of this article, a letter appeared in the journal Lancet (from a single group of pediatric endocrinologists) proposing revised guidelines for the ages at which precocious puberty should be investigated.3Because any proposed changes in the definition of precocious puberty will affect recommendations regarding treatment of this condition, it seems appropriate to reexamine this issue in greater detail. We undertook a detailed review of the data on which the existing definitions were based, a critical examination of the findings of the new study, and a review of data concerning the benefits of treatment with gonadotropin-releasing hormone (GnRH) analogs in girls with precocious puberty. We propose new guidelines for when a girl with early puberty should be evaluated and considered for medical therapy.
HOW IS THE NORMAL AGE RANGE OF PUBERTY IN GIRLS DEFINED?
It is generally accepted that the appearance of breast tissue is the most reliable physical sign of the activation of the pituitary–gonadal axis (gonadarche) in girls. Appearance of pubic hair, although usually occurring at approximately the same time, is attributable to a distinct hormone process, the onset of androgen secretion by the adrenal glands (adrenarche), and is not a marker for gonadarche. As reviewed by Herman-Giddens et al,2 there have been no recent data that address the age at which either puberty or adrenarche normally occur in a large sample of US or North American girls. The study cited most often as defining the age of normal puberty is the 1969 article by Marshall and Tanner4 that involved 192 white British girls who were seen at 3-month intervals during adolescence (as early as age 8 and as late as age 18). Each subject was photographed nude at each visit, and staging of breast development and pubic hair was done by examination of the photographs (not the patient) using the rating system devised by Tanner. The mean age at which stage 2 of breast development was attained was 11.15 ± 1.1 years, and for stage 2 pubic hair, 11.69 ± 1.21 years. These findings were similar to results of earlier studies by Reynolds and Wines,5 who in 1948 reported a mean onset of breast budding at 10.8 years in 49 girls followed at the Fels Institute in Ohio, and Nicolson and Hanley,6 who reported a mean onset of breast budding of 10.6 years in a sample of 252 California girls born in 1928–1929. However, Tanner and Marshall noted that the girls they studied came primarily from the “lower socioeconomic sector,” were residents of a children's home, and may have received suboptimal care before entering the home (in most cases, between 3 and 6 years of age). They also noted that the girls in their study were “not representative of the British population,” particularly because they reached menarche at a mean age of 13.47 years, or 4 months later than the population of London.4 Nonetheless, their statement that “the first sign of puberty appeared between the ages of 8.5 years and 13 years in 95% of girls” appears to have been widely accepted as a standard for the normal onset of puberty in girls.
A more recent study addressing this issue was reported by Roche et al,7 based on 67 white girls in Ohio, followed longitudinally from ages 9.5 to 16 years of age. The mean age of onset of stage 2 breast development was 11.2 ± 0.7 years, in close agreement with Marshall and Tanner's data. However, because the earliest patients studied were 9.5 years of age, it is not helpful for defining the lower limits of normal for sexual maturation. The usefulness of this study is limited further by the relatively small sample size, the use of only white subjects, and the fact that puberty was staged by self-report rather than by physician examination.
Recognizing the lack of “up-to-date, geographically relevant standards for assessing the onset of pubertal changes in girls,” Herman-Giddens and her collaborators undertook a large cross-sectional study involving 17 000 girls between 3 and 12 years of age being seen in 65 different, primarily suburban practices around the United States who participated in the American Academy of Pediatric's Pediatric Research in Office Settings (PROS) network.2 Patients were eligible if they were being seen for a well-child visit (95%) or for a problem that would require a complete physical examination (5%). All 225 participating clinicians underwent training on assessment of breast and pubic hair by Tanner staging using photographs, and needed to score at least 87.5% on a test to be included in the study. Interrater reliability was found to be 0.86 for breast and 0.93 for pubic hair staging. Because Tanner's original staging was done with photographs and did not involve palpation, participants were instructed to rate breast stage by visual inspection only. If palpation was performed, results were recorded separately.
The mean age at which Tanner 2 breast development was reached was 9.96 ± 1.82 years in white girls (∼1 year earlier than the age cited in most previous studies) and 8.87 ± 1.93 years in African-American girls (∼2 years earlier). The mean age for attaining stage 2 pubic hair, 10.51 ± 1.67 years for white and 8.78 ± 2.0 years for African-American girls, is also significantly earlier than that noted in previous studies. Table 1 shows the data for prevalence of stage 2 breast and pubic hair development in white and African-American girls during the critical ages of 5 to 10 years. Notably, 15.4% of African-American and 5.0% of white girls examined at 7 to 8 years were at Tanner 2 or greater for breast development, and 17.7% of African-American and 2.8% of white girls were at Tanner 2 or greater for pubic hair. The average age of menses was 12.88 years for white and 12.16 years for African-American girls. This indicates no change, at least for white girls in the United States, in the average age at menarche over the past 50 years.
Either the results summarized above are seriously flawed or they should prompt us to reconsider our definition of when puberty is precocious. The authors concede that the practices (mostly suburban) and subjects were not selected randomly, but argue that because of the large sample size, it is likely that the results are applicable to the US population as a whole. There are no data indicating that rural or inner city American girls differ significantly in their growth and puberty from suburban girls. Given the increase in recent years in the prevalence of obesity, which is more pronounced in African-American than in white children,8 it is unlikely that caloric inadequacy is a major problem in inner city areas today as it may have been in the past. It also seems unlikely that preferential enrollment of early-maturing girls being seen for routine well-child visits would have skewed the results. African-American subjects comprised 9.6% of the sample as a whole and 8.3% of the sample of 7- to 8-year-olds (n = 136). However, the 95% confidence interval for the percentage of African-American girls with breast tissue at age 7 to 8 was 10.5% to 22.8%, a high proportion even at the lower end of the interval.
A concern raised about this study is that breast staging was done by visual inspection alone, without the benefit of palpation. In overweight girls, fat can be confused with breast tissue on visual inspection, but with palpation, one can usually distinguish true glandular breast tissue from adipose tissue. Is it possible that the large increase in the percentage of girls between 6 and 10 years of age with Tanner 2 breast development could be explained by an increase in the prevalence of obesity and thus fat in the breast area? To address this issue, Herman-Giddens provided the authors with data comparing the recorded Tanner staging by both visual inspection and by palpation, which was done for 39% of the patients (personal communication, 1997). For the entire sample in which breast development was scored by both methods, overrating by inspection was found in 4% and underrating was found in 1.7%. In 15% of girls rated Tanner 2 by visual inspection in whom results of palpation were recorded, no breast tissue was found. However, for the most obese girls (1st quartile body mass index [BMI]), the percentage who were Tanner 2 by inspection but who had no palpable tissue was 15%, versus 13% for the thinnest girls (4th quartile BMI). In other words, the occasional misclassification of a Tanner 1 girl as a Tanner 2 girl was as likely to occur in the thinner as in the fatter girls. In addition, the percentage of obese girls in the study, defined as a BMI above the 90th percentile for age, did not increase greatly between 6 and 10 years of age for white girls (14%–15%) or for African-American girls (21%–28%) (E. Slora, R. Wasserman, personal communication, 1997). Thus, there is no evidence that the high percentage of girls found to have stage 2 breast development before 8 years of age can be explained by observer error in scoring fat tissue as breast tissue. We believe that it reflects a trend toward earlier onset of pubertal maturation.
WHAT IS THE EVIDENCE THAT PRECOCIOUS PUBERTY LEADS TO BEHAVIOR PROBLEMS?
The data presented above suggest that the onset of breast development between 7 and 8 years of age in white girls and between 6 and 8 years in African-American girls may be part of the normal broad variation in the timing of puberty, and does not, in most cases, represent a pathologic state. It is these age groups, however, that comprise the majority of girls referred to pediatric endocrinologists for evaluation of sexual precocity. Puberty in most such patients is idiopathic and central (because of early activation of an otherwise normal hypothalamic–pituitary–gonadal axis). What then are the arguments that support aggressively evaluating and subsequently suppressing puberty in these girls? Behavioral problems have been noted in girls with precocious puberty, although there are few well-designed studies of this issue. Sonis et al9 administered the Child Behavior Checklist to 33 early-maturing girls 6 to 11 years of age and an age-matched control group; 27% of the patients scored >2 standard deviations (SD) above the mean on the Total Behavior Problem scale. As a group, these girls appeared to be more depressed, socially withdrawn, aggressive, and moody than those in the control group. However, no attempt was made to relate these problems to the age of the children at diagnosis or at the time of testing, and it is not clear which, if any, of these behavior problems would have responded to hormonal therapy. A follow-up study of girls with idiopathic precocious puberty at 17½ years of age found no significant lasting psychological effects, except a tendency for excessive psychosomatic complaints.10 A related concern is that early-maturing girls will have difficulty coping with early menarche if no treatment is offered. The 7-year-old who has Tanner 2 breasts will in most cases not have menarche for another 2 to 3 years, allowing the pediatrician and parents ample time to prepare the child for this event. However, individual variation in the child's coping mechanisms and readiness for adolescent development, and in the family's level of concern, needs to be taken into consideration when making recommendations regarding treatment.
WHAT IS THE EVIDENCE THAT PRECOCIOUS PUBERTY COMPROMISES ADULT HEIGHT?
The reason cited most often for treating central precocious puberty is concern that adult height will be compromised if GnRH agonists are not used. However, before GnRH analogs were available, many girls with precocious puberty achieved adult heights well within the normal range.11 A more recent analysis of 20 untreated patients with idiopathic precocious puberty for whom adult height was available found that it was normal (>3rd percentile) in 90% and that the mean adult height for this group was 161.4 ± 7.7 cm, ∼2 cm below the population mean.12 Correlation between initial height prediction based on bone age using the Bayley-Pinneau tables13 and adult height was good (r = 0.85). A similar conclusion was reached in a study that looked retrospectively at the relationship between age at menarche and adult height of women attending a gynecology practice.14 For women who reported menarche at age 9 (in whom breast development probably started 2 to 3 years earlier), mean height was 159.5 ± 6.5 cm (n = 13), compared with 160.8 ± 5.1 cm (n = 23) for those reporting menarche at age 10, and ≈163 cm (n = 260) for those reporting menarche at age 11, 12, or 13. Thus, for otherwise normal women who have menarche 3 to 4 years earlier than average, the average impact on adult height is <4 cm. Insight into why early puberty has a relatively small influence on adult height is provided by a recent study that examined the interval between the appearance of breast development and menarche as a function of age at onset of puberty.15 The younger the age at onset of puberty, the longer the duration of puberty (mean of 2.77 years for onset of puberty at age 9, decreasing steadily to 1.44 years at age 12). The longer duration of puberty in early-maturing girls thus results in a more prolonged pubertal growth spurt that in part offsets the loss of adult height expected because of earlier skeletal maturation. (This factor is incorporated into the separate Bayley-Pinneau height prediction table for “accelerated” girls.)13 Roche16 reported that the mean increase in stature for all girls after menarche is 7.4 cm, but this figure is higher (≈10 cm) for girls with early menarche and lower (≈5 cm) for those with delayed menarche. Thus, the concern that girls with borderline early onset of puberty (ie, between 6 and 8 years of age) often will become very short adults without intervention is overstated, because adult height within the normal range is reached in the majority of cases.
WHAT IS THE EVIDENCE THAT TREATMENT OF GIRLS WITH PRECOCIOUS PUBERTY STARTING BETWEEN 6 AND 8 YEARS OF AGE IS BENEFICIAL?
Potent, synthetic long-acting GnRH analogs that inhibit gonadotropin and sex steroid secretion were first used to treat children with sexual precocity in 1979. Many studies have shown increased predicted adult height (based on bone age) after 1 to 3 years of therapy, and a smaller number have shown increased adult height over that predicted before the onset of therapy.17–21 It is not possible, however, to conclude from these studies that GnRH analogs are beneficial in all early-maturing girls. Some girls with precocious puberty (particularly those with bone age ≤20% above chronologic age) appear to have a slowly progressing variant and do not show loss of predicted height if followed without treatment.22 ,23Furthermore, in most studies, patients of all ages are pooled for analysis, and the outcomes for the girls who were older at the onset of puberty are not reviewed separately. One study that does separate younger from older girls is the review of Kletter and Kelch,20 which included retrospective data pooled from 10 groups of investigators on 131 treated and 66 untreated girls with idiopathic central precocious puberty. For patients younger than 6 years at diagnosis, treatment was associated with a greater adult height (160.4 cm) than no treatment (153.9 cm), but for the larger group of girls who were older than age 6 at diagnosis, mean adult height was 157.5 cm in GnRH-treated patients and 157.0 cm in control subjects. In a study in which height outcomes of treated patients were compared with a historical untreated group,21the difference in mean adult height was quite substantial for patients started on therapy before age 5 (164.3 vs 150.2 cm) but was relatively small (157.6 vs 153.4 cm) for girls started on therapy after age 5. Another study of adult height in girls with idiopathic central precocious puberty compared two groups of girls: those with a height prediction of ≤155 cm who were treated with GnRH analogs, and those with a predicted height of >155 cm who were not treated.21 Treated girls (4–9 years of age at start of therapy, with a mean bone age advancement of 3 years) had a mean predicted height of 152.1 cm and an adult height of 159.0 cm, a significant improvement, whereas the untreated girls (all 6–8 years of age at onset of puberty) had a similar mean predicted height (162.5 cm) and adult height (162 cm). These results confirm that there is a subset of early-maturing girls who, because of significant bone age acceleration, may have compromised adult height and therefore benefit from GnRH therapy. However, for the 6- to 8-year-old girls with a more slowly progressive form of puberty and less bone age advancement, withholding therapy does not appear to compromise adult height.24 In summary, the studies available to date support the view that treatment of girls with early puberty is often unnecessary, and that in the girls with pubertal onset at 6 to 8 years, little if any beneficial effect on adult height is obtained in the majority of cases.
RECOMMENDATIONS CONCERNING CHILDREN NEEDING EVALUATION FOR EARLY PUBERTY
1) Girls With Breast and/or Pubic Hair Development Before Age 7 (White Girls) or Age 6 (African-American Girls)a
The 1-year difference between white and African-American girls is justified by both the data shown in Table 1 and the fact that the interracial difference in mean age for transition to Tanner stage 2 was 1.1 years for breast development and 1.7 years for pubic hair. To estimate the percentage of girls who are maturing earlier than the proposed age limits, one should note that the data in Table 1 represent the prevalence of Tanner stage 2 breast and pubic hair in groups of girls whose mean ages are at the midpoint of the intervals noted. For example, 2.9% of white girls in the 6.0- to 6.99-year group, (mean age of ∼6.5 years) had reached stage 2 breast development. Thus, the percent prevalence of stage 2 breast development by age 7 in white girls is approximately midway between the prevalence for the 6.0- to 6.99-group and the 7.0- to 7.99-year group ([2.9% + 5.0%]/2 ≈ 4.0%). Similarly, the percent of African-American girls reaching the same stage by age 6 is ∼4.4%. For early pubic hair development, the estimated percents needing evaluation according to the proposed guidelines are 2.1% for white girls and 6.4% for African-American girls. Girls with pubic hair but no breast development appearing before age 8 usually are diagnosed with premature adrenarche, a benign normal variant. However, those with other evidence of increased androgen production such as linear growth acceleration, clitoral enlargement, or acne require evaluation to exclude a more serious virilizing disorder.
2) Girls With Breast Development Beginning After Age 7 (White) or After Age 6 (African-American) Who Have the Following Additional Conditions
a) unusually rapid progression of puberty resulting in rapid skeletal advancement (bone age >2 years ahead of chronologic age)and a predicted height (based on bone age) of either 2 SD (10 cm or 4 inches) or more below their genetic target heightor <150 cm (59 inches);
b) new central nervous system-related findings including headaches, seizures, or focal neurologic deficits, or an underlying neurologic problem such as hydrocephalus; or
c) behavior-based factors suggesting that their emotional state (or the family's emotional state) is affected adversely by the progression of puberty and the potential for early onset of menses.
The majority of 6- to 8-year-old girls with breast development will have evidence of growth acceleration as well as an advanced bone age, yet few of these girls are at risk of short stature. For example, a 7-year-old girl with stage 2 breast development, a height of 50 inches (at the 90th percentile) and a bone age of 9 years has, according to the Bayley-Pinneau tables,13 a height prediction of 50 inches/0.79 = 63.3 inches. If her father is 70 inches and her mother is 65 inches, her genetic target height (defined for girls as [father's height + mother's height − 5 inches)/2], is 70 inches + 65 inches − 5 inches/2 = 65 inches. Because her predicted height based on bone age falls well within 2 SD (4 inches) of her genetic target height, there is no indication for detailed evaluation and therapy with GnRH analogs. However, a 7-year-old girl with stage 2 breast development whose height is only 46 inches (25th percentile) has, if the bone age is 9 years, a predicted height of only 58 inches, which is more than 4 inches below her target height of 65 inches; she might be an appropriate candidate for GnRH agonist therapy. It should not be necessary to obtain a bone age on all girls with breast development appearing between 7 and 8 years of age. However, for the subset of girls whose tempo of puberty seems to be more rapid than normal (eg, Tanner 3 or greater breast development before age 8) or whose height is not well above the 50th percentile, a bone age helps to identify those girls who may end up significantly shorter than their target height. In some cases, reexamination in 3 to 6 months will aid in deciding whether puberty is progressing too rapidly and whether an evaluation should be performed.
3) Boys With Evidence of Increased Androgen Production Other Than Pubic Hair (eg, Penile and Scrotal Enlargement, Acne, Growth Acceleration)
With or Without Testicular Enlargement Before Age 9
Boys are not discussed in this review, because the Herman-Giddens study only examined puberty in girls, and there is no new data on which to base revised recommendations in boys. Two recent studies,9 ,25 both published in 1995, did not show a notable trend toward earlier maturation in boys, and one of them,25 which examined 515 boys (278 white and 237 African-American) between 10 and 15 years of age, did not find a significant difference in the timing of puberty based on race. With the current age limit of 9 years, far fewer boys than girls are referred to pediatric endocrinologists for early puberty.26Furthermore, whereas the majority of girls with central precocious puberty are idiopathic, a large proportion of boys with this problem will prove to have a central nervous system disorder.17 ,26Thus, no change in the lower limit for normal puberty in boys is proposed at this time.
1. 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.
2. In most cases, evaluation of girls with early breast and/or pubic hair development to look for a pathologic etiology of precocious puberty need not be performed for white girls older than 7 years of age or for African-American girls older than 6 years of age.
3. No change in the guidelines for similar assessment in boys can be made at this time. Investigation for pathologic etiologies should occur in boys who have onset of pubertal changes before 9 years of age.
4. Use of GnRH therapy to suppress puberty in 6- to 8-year-old girls with slowly progressive puberty and/or an acceptable predicted adult height based on bone age has not been proven to have a significant effect in improving adult height.
We thank Dr Melvin Grumbach for his critical review of this manuscript. Drs Grumbach and Styne reviewed this topic for the 9th edition of Williams Textbook of Endocrinologyand independently arrived at similar recommendations.27
- Received February 8, 1999.
- Accepted March 29, 1999.
Reprint requests to (P.B.K.) Box 980140, Richmond, VA 23298-0140. E-mail:
FNa Although not reviewed here, it should be noted that girls who are less than 2 years of age with a small amount of unilateral or bilateral breast tissue, no enlargement of the areola, normal linear growth, and no other signs of puberty usually have premature thelarche, a benign normal variant.
- ↵Rogol A, Blizzard RM. Variations and disorders of pubertal development. In: Kappy MS, Blizzard RM, Migeon CJ, eds. Wilkins' The Diagnosis and Treatment of Endocrine Disorders in Childhood and Adolescence. Springfield, IL: Charles C Thomas; 1994;857–917
- Herman-Giddens ME,
- Slora EJ,
- Wasserman RC,
- et al.
- Marshall WA,
- Tanner JM
- Bayley N,
- Pinneau SR
- Roche AF
- Paul D,
- Conte F,
- Grumbach MM,
- Kaplan SL
- Fontoura M,
- Brauner R,
- Prevot C,
- Rappoport R
- ↵Kaplan SL, Grumbach MM. Pathogenesis of sexual precocity. In: Grumbach MM, Sizonenko PC, Aubert ML, eds. Control of the Onset of Puberty. Baltimore. MD: Williams and Wilkins; 1990:620–660
- ↵Grumbach MM, Styne DM. Puberty: ontogeny, neuroendocrinology, physiology, and disorders. In: Wilson JD, Foster DW, eds. Williams Textbook of Endocrinology. 9th ed. Philadelphia, PA: WB Saunders; 1998
- Copyright © 1999 American Academy of Pediatrics