Abstract
Objective
Evaluate inhibin B and luteinizing hormone (LH) levels in a large, representative cross-sectional sample of U.S. girls and characterize the relationships of these laboratory values with age, clinical signs of puberty, and other correlates.
Design
Cross-sectional analysis of LH and inhibin B in banked serum from 720 girls aged 6-11 years who participated in the Third National Health and Nutrition Examination Survey (NHANES III).
Measurements
Levels of inhibin B and LH, race, ethnicity, and anthropometric measurements were compared for all girls. Visual assessment of pubertal stage was performed on girls 8 years and older. A two-part model was used to establish normative data and tobit regression models were used to evaluate associations with participant characteristics. Receiver operating characteristic (ROC) analysis was performed to identify optimum cut points predictive of puberty onset.
Results
Mean hormone levels progressively increased with age. LH levels progressively increased with pubertal stage. Inhibin B levels increased gradually from breast stage I to II, then more sharply to peak at stage III, followed by a plateau at stages IV and V. ROC curves indicated both hormones were consistent with pubertal onset as indicated by breast stage II.
Conclusions
This study characterizes inhibin B and LH values in a large, representative cross-sectional sample of U.S. girls. Inhibin B can be a useful tool in combination with other clinical and biochemical parameters to evaluate gonadal function as a reflection of pubertal progression in girls.
Keywords: Inhibin B, luteinizing hormone, NHANES, puberty, girls
Introduction
The importance of inhibins in pubertal development has only relatively recently begun to be characterized, sparked by the development of enzyme-linked immunosorbent inhibin assays that distinguish between inhibins A and B.1 Inhibin B is a heterodimeric glycoprotein consisting of an alpha subunit linked to a beta-B subunit.2 It is secreted by granulosa cells in small antral follicles, and plays an important role in the follicle stimulating hormone (FSH) feedback loop.3-5
Previous reports have characterized inhibin B levels in girls at different stages of childhood. During the period in infancy of increased hypothalamic-pituitary-gonadal activation, inhibin B levels are elevated, signifying importance in functional follicle maturation. Levels then fall so that during the prepubertal period, they are much lower and often unmeasurable.6 Evaluation in school-age girls who are clinically prepubertal has revealed a progressive increase in inhibin B levels with age, reflecting increasing follicular activity.7 This finding is consistent with reports of ovarian ultrasonography in girls aged 7-10 years showing an increase in follicle development that can precede physical changes of puberty.8 Evaluation during puberty has shown a slow increase from Tanner breast stage I (B1) to B2, with an increased rate of rise to B3 and variable increases from B4-B5.7,9-11 After the onset of menses, inhibin B rises during the follicular phase, reaching its peak shortly after the LH peak, and then decreases during the luteal phase, reflecting secretion by antral follicles.3
Serum luteinizing hormone (LH) and FSH concentrations in relation to puberty among girls have been known for some time to have a profile of mean levels gradually increasing from slightly before the physical onset of puberty (B2) until menarche.12,13 More recent profiles are similar, albeit with lower values using more recent gonadotropin assays.14 The relationship of gonadotropins to inhibin B has been recently studied suggesting a rise of both throughout puberty, or in a report of a limited number of girls, with a positive correlation with LH before and during early puberty.7,10
The Third National Health and Nutrition Examination Survey (NHANES III), conducted from 1988-1994, involved interviews, physical examination, and collection and banking of serum from U.S. children.15 This cross-sectional sample provides a unique opportunity to evaluate data from a large, nationally representative sample of U.S. boys and girls. In 2007, serum specimens were utilized for assays of inhibin B and LH levels with the original intents of characterizing the relationship between these hormones, with previously reported serum lead levels, age, and the clinical onset of puberty.16 Given limitations in the quantity of stored sera, it was determined that the most useful hormones to study would be inhibin B and LH. These were selected based on information present or lacking in the existing literature, the relatively increased rise of mean levels of LH among girls with puberty compared to FSH, and the potential applicability of further data. The objectives of our study were to examine the LH and inhibin B levels in girls, including relationships with age and pubertal staging, thereby generating normative data, as well as to systematically evaluate associations with other participant characteristics.
Materials and Methods
Approval for this study was obtained by the Centers for Disease Control and Prevention National Center for Health Statistics Research Ethics Review Board. Our target population consisted of 1589 girls who participated in NHANES III from 1988-1994, and were aged 6.0- 11.99 years at the time of an initial household interview. Informed consent was obtained before participation in the survey. Initial sampling was based on a complex, multistage clustered probability sampling design that oversampled non-Hispanic blacks and Mexican-Americans to provide more precise estimates for these groups. Data were subsequently weighted to represent the entire U.S. population of girls for the included ages from 1988-1994. Reweighted sample weights were obtained from the initial sample weights and those calculated for the analytic subsample of girls who had stored residual sera available using propensity score weights. All statistical estimates were adjusted for these complex survey design effects namely, strata, probability sampling units (PSU), and the new reweighted sample weights. Racial/ethnic classifications utilized by NHANES III were also used to characterize our population. Further NHANES III design details are publicly available in the Plan and Operations outline.15
A total of 705 girls with available banked serum were included in this study. The remaining girls in the original cohort did not have adequate residual serum available for testing. Serum was obtained by venipuncture at the time of clinical exam and categorized as morning, afternoon, or evening based on the timing of the blood draw. All subjects had a physical exam by a participating pediatrician and anthropometric assessments. Visual assessment for breast and pubic hair development was performed without palpation of breast tissue in girls aged 8.0-11.9 years. Development was characterized according to apparent Tanner Stage. Girls 8 years or older were asked about menarcheal status.
Blood samples were obtained from the National Center for Health Statistics/Centers for Disease Control and Prevention. Hormone assays were performed by Rules-Based Medicine, Inc. (http://www.rulesbasedmedicine.com). LH levels were measured using the LH ELISA kit (Bio-Quant BQ049F; Bio-Quant, Inc., San Diego, CA, USA), a solid-phase direct sandwich method. Limit of detection (LOD) for LH was 0.05 mIU/ml. Intra-assay precision ranged from 6.18-10.58 % and inter-assay precision ranged from 8.13 to 11.57%. Inhibin B was measured using the DSL-10-84100 ACTIVE ® Inhibin B ELISA assay (Diagnostic Systems Laboratories, Inc., Webster, TX, USA, Gen I assay), an enzymatically amplified two-site two-step sandwich-type immunoassay. LOD for inhibin B was 7.0 pg/ml. A substitution of LOD/√2 was used in the calculation of descriptive statistics for values that were below the LOD, and thus not accessible to the authors. Intra-assay precision ranged from 3.5-5.6 % and inter-assay precision ranged from 6.2-7.6%. Extreme or potentially outlying values for hormones were reviewed in light of other available data for individuals, and judgments were made concerning inclusion based on clinical feasibility.
Other covariates were obtained from the NHANES III database, including race/ethnicity (non-Hispanic white, non-Hispanic black, Mexican-American, other), census region (Northeast, Midwest, South, West), poverty-income ratio (PIR), health rating, cigarette smoke exposure, and presence of anemia and iron deficiency. PIR was defined as the total household income divided by the poverty threshold for the year of the interview, with higher values indicating increased household income. Iron deficiency was defined as 2 out of 3 measures below age-specific cutoff values for transferrin saturation, serum ferritin, and erythrocyte protoporphyrin, while anemia was defined as a hemoglobin value ≤ the 5th percentile for age.17,18 Health ratings were parental reports of girls’ health status ranging from 1 (excellent) to 5 (poor). Preliminary statistical analysis revealed no significant differences in results based on timing of blood draw so data were pooled for subsequent evaluation.
Several anthropometric indices were derived. Body mass index (BMI) was calculated as weight in kilograms/ height in meters squared. Triceps and subscapular skinfold thickness were measured in millimeters. A two compartment body composition model was built using subcutaneous fatness (from sum of triceps + subscapular skinfolds) to predict body weight (consisting of fatness and lean mass) in regression models. The residuals/error terms of the regression models were defined as a measure of lean mass for each girl in subsequent multivariate analyses.
Statistical Methods
Descriptive statistics for continuous variables were presented as means and standard errors (SE) or medians and standard errors (SE), where appropriate, and categorical variables as percentages and confidence intervals (95% CI) using complex survey procedures. We used non-parametric graphical exploratory data analysis (EDA) 19 techniques to study monotonic trends in hormone levels across age. Subsequently, a penalized beta-spline smoother was used to present the pattern of hormone changes with age.20 This smoother is resistant to outlying values and is able to capture any patterns of hormone levels associated with normal growth.
Smoothed percentiles and Z-scores for normative LH and inhibin B hormone values were calculated for ages 6.00 to 11.99 years in 6-month intervals. A 2-part-approach of percentile estimation was used to accommodate the high proportion of LOD values for LH and inhibin B. For part one, we modeled the continuous part of the distribution above the LOD using LMS method which summarizes hormone measurements using penalized maximum likelihood and cubic splines with 3 smoothed curves L M S.21 They denote: the Box-Cox power transformation (L) term for normalizing the measured covariate, the median (M) and the generalized coefficient of variation (S). The second part involved modeling for the estimated proportion below LOD. And then using a closed-form formula, we calculated the smoothed percentile estimates. All model assumptions were met. We used the worm plots and goodness-of-fit statistics for logistic regressions for continuous and discrete parts of the distribution respectively.22 A full description of the methods for this 2-part-approach has been described elsewhere.23 Additionally, unsmoothed percentiles for the hormone levels were estimated using standard non-parametric ranking methods separately within breast and pubic hair stage categories.
Tobit regression models were used to study associations between hormone and participant descriptors. The use of tobit models facilitates efficient use of all parts of the distribution by applying maximum likelihood regression techniques to model both continuous and discrete parts of the data.24 The use of the tobit regression models thus enabled us to obtain valid associations while accommodating the truncated distributions of the outcome variables, viz., LH and inhibin B, due to values below the LOD. Hormone concentrations were re-expressed to approximate normality to meet regression model assumptions. Consequently, regression estimates are presented in log-natural units as well as in back-transformed units. This approach has been used successfully elsewhere for clinical laboratory data.25,26
Receiver operating characteristic (ROC) curves were used to select optimal hormone cutoffs for correctly identifying those girls with B2 compared with those with B1. These analyses did not include girls with B3 or B4. Because our outcome of interest is not a disease condition, we used minimal misclassification as the criterion for selecting optimal cut-offs, rather than favoring either sensitivity or specificity. To evaluate the performance of the hormone cut-offs in the study population, we used cross-classification tables to assess the diagnostic performance of observed B2 vs. the new optimal cut offs.
All statistical analyses were carried out using PC SAS 9.2 (SAS, Cary, NC, USA). Statistical significance was declared at a 2-tailed p-value of <0.05.
Results
LH levels were completed for 705 girls, with 689 having inhibin B results. Table 1 characterizes sample demographics, with percentages weighted to represent the racial/ethnic makeup of the U.S. population. Mean serum levels tended to increase with age for LH and inhibin B, with the exception of a decrease in mean LH values at age 9 years. Median values for both LH and inhibin B only surpassed the LOD at 10 years of age and older, as the younger age groups had >50% of girls with levels below the limit of detection. Eight outlying LH values were identified; 6 of these in girls < 8 years of age, all with values higher than expected. All outliers were included in the statistical analysis. Only 3 potentially outlying Inhibin B values were identified, all with elevated levels, but with only one girl <8 years of age. Two of the elevated values were in fully pubertal girls. All three values were included in the analysis. Figures 1a and 1b graphically depict the distribution of levels of LH and inhibin B, respectively, based on participant age. When values were smoothed, LH gradually increased from age 8-10 years, with a steeper rate of increase beginning around 10 years of age. Conversely, inhibin B levels showed a less rapid but consistent increase throughout the ages measured.
Table 1.
Participant Characteristics
| Descriptors | Categories | Count | Percent ( 95% CI) |
|---|---|---|---|
| Race | NH-W | 146 | 65.3 (54.8 – 75.6) |
| NH – B | 258 | 17.6 (12.2 – 22.9) | |
| Mexican American | 267 | 7.9 (5.1 – 10.6) | |
| Other | 34 | 9.3(3.6 – 15.0) | |
| Census Region | Northeast | 61 | 17.2 (9.3 – 25.1) |
| Midwest | 153 | 27.4 (17.6 – 37.2) | |
| South | 272 | 32.9 (20.9 – 44.9) | |
| West | 219 | 22.5 (8.5 – 36.5) | |
| Exam Session | Morning | 271 | 30.7 (24.9 – 36.6) |
| Afternoon | 249 | 39.1 (34.2 – 44.1) | |
| Evening | 185 | 30.1 (23.9 – 36.3) | |
| Breast Stage (B) | B 1 | 191 | 53.8 (46.5 – 61.0) |
| B 2 | 127 | 29.8 (21.4 – 38.1) | |
| B 3 | 61 | 10.6 (3.8 – 17.3) | |
| B 4+ | 38 | 5.9 (3.2 – 8.5) | |
| Pubic Hair (PH) | PH 1 | 209 | 55.7 (47.5 – 64.0) |
| PH 2 | 95 | 23.4 (16.5 – 30.4) | |
| PH 3 | 72 | 15.8 (5.4 – 26.3) | |
| PH4+ | 35 | 5.0 (2.3 – 7.7) | |
| Menarcheal Status | Post-menarcheal | 35 | 3.9 (1.8 – 6.0) |
| PIR | <1 | 301 | 53.1 (44.0 – 62.2) |
| 1-2 | 160 | 37.0 (27.0 – 47.0) | |
| 2 + | 50 | 9.9 (4.0 – 15.7) | |
| Anemia and Iron Deficiency* | Anemic (count) | 80 | 7.7 (4.7 – 10.6) |
| Iron Deficient | 32 | 2.3 (0.4 – 4.2) | |
| Health Rating | Excellent | 233 | 43.3 (35.1 – 51.6) |
| Very Good | 188 | 30.1 (24.1 – 36.2) | |
| Good | 213 | 23.3 (18.4 – 28.2) | |
| Fair | 69 | 3.1 (1.7 – 4.5) | |
| Poor | 2 | 0.1 (0.0 – 0.3) | |
Abbreviations: NHW – Non Hispanic white; NHB- Non Hispanic black; PIR- Poverty Income Ratio.
Iron deficiency was defined as 2 out of 3 measures below age-specific cutoff values for transferrin saturation, serum ferritin, and erythrocyte protoporphyrin according to the guidelines provided by the Centers for Disease Control and Prevention. Anemia was defined as hemoglobin value ≤ the 5th percentile for age.
Figure 1a.
Scatter plot describing monotonic trends in smoothed LH values with age. Observed values are plotted as dots, while smoothed values are plotted as triangles. Data were statistically smoothed to generate a pattern of values that is resistant to outliers.
Figure 1b.
Scatter plot describing monotonic trends in smoothed inhibin B values with age. Observed values are plotted as stars, while smoothed values are plotted as squares. Data were statistically smoothed to generate a pattern of values that is resistant to outliers.
Visual assessment of breast development, with assignment to apparent breast stage, was performed in 417 girls, while 411 had Tanner staging for pubic hair available. Thirty-five girls reported being post-menarcheal, with a mean reported age at menarche of 11.48 ± 0.04 years. LH values increased with progression of breast staging and pubic hair staging. Inhibin B values increased with breast and pubic hair staging through stage III, but plateaued at stage IV.
LH and inhibin B values were below the level of detection in 64.9 % and 62.0 % of girls characterized as B1. The percentage of girls below the level of detection for LH decreased with progression through breast staging, and all girls who were B4 or greater had detectable LH levels. Concordant with the mean values, the percentage of participants with detectable inhibin B increased as girls advanced to B3 staging. However, the percentage of undetectable values increased in those who had B4 or greater staging or were post-menarcheal.
Figures 2a and 2b graphically depict the distribution of LH and inhibin B levels in relation to age and breast stage. While illustrating the trends described above, the figures also highlight the inter-individual variation in stages within each age and hormone level. This was especially true for inhibin B levels in girls with B4 or greater staging.
Figure 2a.
Three dimensional scatter plot of measured LH (mIU/ml) vs. age and breast stage.
Figure 2b.
Three dimensional scatter plot of measured inhibin B (pg/ml) vs. age and breast stage.
One aim was to establish normative data for this population based on age and pubertal stage using the two-part model to correct for large percentages of values below the LOD. Results for LH (Table 2) and inhibin B (Table 3) are presented in the form of smoothed percentiles and Z-scores for age, and unsmoothed percentiles for pubertal stage. We would like to emphasize that the coefficients of variation presented in the smoothed percentile tables (2 and 3) only applies to the data above the LOD. As a result, it is not very useful for the ages they represent – even though they provide a partial sense of precision around the point estimates.
Our ROC analysis indicated that both hormones predicted onset of puberty as measured by B2 breast development, with area under the curve substantially greater than one half. The optimal LH cut-point was found to be 1.04 mIU/ml, (95% confidence interval (CI): 0.71-1.37). This cut-point yielded 77.56% agreement with a 9.24% false negative rate and a 13.20% false positive rate. The optimal cut-point identified for inhibin B was 17.89 pg/ml (95% CI: 11.63- 24.15) with an agreement of 76.57%, a false negative rate of 12.54% and a false positive rate of 10.89%.
Tobit regression models used to evaluate associations of hormone levels with age, race, health rating, anemia-status, PIR, census, region, body mass index, lean mass, skin-fold measurements, and clinical pubertal status/breast staging revealed a significant association between age (in years) and LH levels (Tobit coefficient β (± SE):0. 29 ± 0.11 log natural (Ln)-mIU/ml/year, p = 0.0100; back transformed estimates: 1.34 , 95 % CI :1.07 – 1.67 mIU/ml/year ) when controlled for the other characteristics. The same model showed that girls from the Midwest tended to have higher LH levels (β (± SE): = 1. 02 ± 0.30 Ln-mIU/ml, p = 0.006; back transformed estimates: 2.79, 95 % CI: 1.55 – 5.00 mIU/ml higher on the average) compared with all other regions. Girls classified as non-Hispanic white had a tendency toward lower LH levels (β (± SE): = -0.91 ± 0.32 Ln-mIU/ml, p = 0.004; back transformed estimates: 0.40, 95%CI: 0.22 – 0.75 mIU/ml lower on the average) compared with all other racial groups. No independent relationships were seen between LH levels and PIR, other census regions, skinfold measurements, or lean mass when controlled for the other variables.
Evaluation of associations of inhibin B with the same panel of characteristics revealed a significant association with age in years (β (± SE): = 0.40 ± 0.08 Ln- pg/ml/year, p < 0.0001; back transformed estimates: 1.50, 95 % CI: 1.28 – 1.76 pg/ml/year) when controlled for the other variables. Sum of skinfolds (mm) had a negative association with inhibin B level (β (± SE): = -0.03 ± 0.007 Ln- pg/ml per mm, p < 0.000; back transformed estimates: 0.97, 95% CI: 0.96 – 0.98 pg/ml/ lower on the average per 1 mm increase in skinfolds). No other variables were found to have significant independent relationships.
Discussion
Relative patterns and frequencies for stages of breast and pubic hair development were consistent with expectations for this age range. This study provides a characterization of inhibin B and LH levels in this large, representative population of U.S. girls. Our findings of a progressive LH increase with age and pubertal stage are consistent with previous reports exploring these relationships.14,27 Previous reports regarding inhibin B levels in relation to age and pubertal staging in other populations have been mostly consistent with our findings, showing the positive association with age, as well as the progressive increase within B1, then to B2, followed by a sharper rate of rise with a peak before menarche.6,7,9-11,27 These findings are consistent with a progressive increase in follicular development and recruitment that culminates with a peak in follicular activity during B3, before the initiation of menarche and the subsequent decrease in inhibin B levels seen during the luteal phase of the menstrual cycle.9 Changes in inhibin B levels by staging for pubic hair development mirrored those seen for breast development. This pattern likely reflects progression of central puberty.
ROC analysis yielded an agreement (efficiency) of approximately 77% for both LH and inhibin B as indicators of puberty onset, as measured by B2 development. Using these cutoffs approximately 23% of girls would be misidentified. Although this minimal misclassification rate is relatively high for either test to be useful in isolation for the identification of true puberty, in combination with a more complete clinical picture, each could prove beneficial. Additionally, visual inspection for breast staging (the outcome measure for the ROC analysis) without palpation could lead to misclassification of a prepubertal girl in of itself, especially those with high body fat, as higher BMI has been found to be associated with higher breast area.28 It therefore seems unlikely that these new hormone cut-offs for predicting onset of puberty are underperforming in a nationally diverse and representative sample of girls like those of the current study. Inhibin B can also be useful as a marker of gonadal function in girls with hypothalamic-pituitary axis abnormalities who are at risk for primary gonadal dysfunction.29 For these patients, representative normative data are crucial for appropriate medical decision making and counseling regarding fertility.
Other research has found varying results regarding fat mass and a positive association with pubertal onset and progression.30-33 This study did not find any significant association between hormones and body mass index when controlling statistically for other potential covariates. There was a negative association between skinfold measurements and inhibin B level, which has not, to our knowledge, been previously described. An association with LH and lean body mass was noted originally, but became non-significant once corrected for pubertal staging, and was most likely a reflection of progression through puberty, as has been previously noted.30
The negative association of LH levels and non-Hispanic white race is consistent with previous reports.34,35 Our analysis also showed an independent positive association between LH and Midwest origin. This effect could be the result of a population difference with a greater portion of clinically prepubertal girls having the increase in LH levels which occurs shortly before the transition to B2. It could also reflect a higher LH level across all pubertal stages. Such trends could possibly be due to an environmental effect or to residual confounding. However, these associations must be interpreted cautiously as, although the variables were corrected for each other, there is a possibility that they interact with each other ways that are not obvious.
There are several limitations to our study. Tanner staging was not performed in girls less than 8 years old; however, some of these girls may have had clinical signs of puberty. Staging performed by visual inspection without palpation may have resulted in erroneous assignments. If this occurred, one would expect the described onset of breast stage 2 to be somewhat younger, especially in heavier girls. The immunoradiometric LH assay used in our study is not as sensitive as other immunochemiluminometric LH assays, which may be more useful in younger children who have hormone changes but no clinical signs of puberty.14 Since the serum samples from our study population were analyzed, a more sensitive second-generation inhibin B assay has become available.36 Evaluation with this assay would have been preferable given the high percentage of prepubertal girls with values below the level of detection. Nevertheless, this study comprises the largest representative sample of U.S. girls in which both inhibin B and pubertal staging were evaluated.
In conclusion, this study provides normative data for inhibin B and LH levels in a large population of U.S. girls aged 6-11 years. Levels showed a clear increase with age as well as with progression through puberty to B3, with high inter-individual variation thereafter. LH levels also increased with age, with a steeper rate of increase after the onset of puberty. Both hormones can be useful instruments in adjunct with other clinical and biochemical markers in assessing onset of puberty.
Table 2a.
Smoothed Percentiles for Luteinizing Hormone for Peri-pubertal US Girls Ages 6.00-11.99 (N=705) Using the 2-part Model
| Age, y |
Luteinizing hormone (mIU/ml)
|
||||||
|---|---|---|---|---|---|---|---|
| Smoothed Prob. of Exceeding LOD | Coefficient of Variation | Smoothed Percentiles |
|||||
| 10th | 25th | 50th | 75th | 90th | |||
| 6.00-6.49 | 0.09 | 1.43 | <0.05 | <0.05 | <0.05 | <0.05 | <0.05 |
| 6.50-6.99 | 0.33 | 1.43 | <0.05 | <0.05 | <0.05 | 0.12 | 0.51 |
| 7.00-7.49 | 0.41 | 1.43 | <0.05 | <0.05 | <0.05 | 0.22 | 0.78 |
| 7.50-7.99 | 0.42 | 1.42 | <0.05 | <0.05 | <0.05 | 0.31 | 1.06 |
| 8.00-8.49 | 0.42 | 1.42 | <0.05 | <0.05 | <0.05 | 0.38 | 1.34 |
| 8.50-8.99 | 0.4 | 1.39 | <0.05 | <0.05 | <0.05 | 0.37 | 1.32 |
| 9.00-9.49 | 0.38 | 1.34 | <0.05 | <0.05 | <0.05 | 0.30 | 1.10 |
| 9.50-9.99 | 0.38 | 1.27 | <0.05 | <0.05 | <0.05 | 0.32 | 1.08 |
| 10.00-10.49 | 0.44 | 1.21 | <0.05 | <0.05 | <0.05 | 0.56 | 1.54 |
| 10.50-10.99 | 0.59 | 1.15 | <0.05 | <0.05 | 0.30 | 1.32 | 2.82 |
| 11.00-11.49 | 0.79 | 1.11 | <0.05 | 0.16 | 1.07 | 2.56 | 4.59 |
| 11.50-11.99 | 0.90 | 1.08 | <0.05 | 0.60 | 1.80 | 3.70 | 6.15 |
The coefficients of variation only apply to the data above the LOD.
Table 2b.
Smoothed Z Scores for Luteinizing Hormone for Peri-pubertal US Girls Ages 6.00-11.99 (N=705) Using the 2-part-Model
| Age, y | Smoothed Prob. of Exceeding LOD | Coefficient of Variation | Estimated Smoothed Z scores for Luteinizing hormone, mIU/ml |
||||||
|---|---|---|---|---|---|---|---|---|---|
| -1.5 | -1 | -0.5 | 0 | 0.5 | 1 | 1.5 | |||
| 6.00 -6.49 | 0.09 | 1.43 | <0.05 | <0.05 | <0.05 | <0.05 | <0.05 | <0.05 | 0.12 |
| 6.50 -6.99 | 0.33 | 1.43 | <0.05 | <0.05 | <0.05 | <0.05 | 0.06 | 0.28 | 0.77 |
| 7.00 -7.49 | 0.41 | 1.43 | <0.05 | <0.05 | <0.05 | <0.05 | 0.14 | 0.46 | 1.13 |
| 7.50 -7.99 | 0.42 | 1.42 | <0.05 | <0.05 | <0.05 | <0.05 | 0.19 | 0.63 | 1.53 |
| 8.00 - 8.49 | 0.42 | 1.42 | <0.05 | <0.05 | <0.05 | <0.05 | 0.23 | 0.80 | 1.92 |
| 8.50 - 8.99 | 0.4 | 1.39 | <0.05 | <0.05 | <0.05 | <0.05 | 0.21 | 0.78 | 1.88 |
| 9.00 - 9.49 | 0.38 | 1.34 | <0.05 | <0.05 | <0.05 | <0.05 | 0.17 | 0.66 | 1.55 |
| 9.50 - 9.99 | 0.38 | 1.27 | <0.05 | <0.05 | <0.05 | <0.05 | 0.18 | 0.66 | 1.49 |
| 10.00 - 10.49 | 0.44 | 1.21 | <0.05 | <0.05 | <0.05 | <0.05 | 0.37 | 1.02 | 2.05 |
| 10.50 - 10.99 | 0.59 | 1.15 | <0.05 | <0.05 | <0.05 | 0.30 | 1.00 | 2.04 | 3.54 |
| 11.00 - 11.49 | 0.79 | 1.11 | <0.05 | <0.05 | 0.36 | 1.07 | 2.10 | 3.55 | 5.53 |
| 11.50 - 11.99 | 0.90 | 1.08 | <0.05 | 0.22 | 0.85 | 1.80 | 3.13 | 4.92 | 7.22 |
The coefficients of variation only apply to the data above the LOD.
Table 2c.
Weighted Unsmoothed Percentiles for Luteinizing Hormone by Pubertal Stages with Non-parametric Ranking for Peri-pubertal US Girls Ages 6.00-11.99(N=705)
| Pubertal stage | Unweighted N | Proportion Exceeding LOD | 10th | 25th | 50th | 75th | 90th |
|---|---|---|---|---|---|---|---|
| B 1 | 191 | 0.35 | <0.05 | <0.05 | <0.05 | 0.2 | 1.02 |
| B 2 | 127 | 0.72 | <0.05 | <0.05 | 0.40 | 1.27 | 4.09 |
| B 3 | 61 | 0.87 | 0.18 | 1.10 | 1.10 | 2.93 | 3.81 |
| B 4+ | 38 | 1.00 | 0.19 | 1.92 | 3.02 | 6.33 | 11.65 |
| PH 1 | 209 | 0.43 | <0.05 | 0.04 | 0.04 | 0.2 | 0.67 |
| PH 2 | 95 | 0.63 | <0.05 | 0.04 | 0.36 | 2.34 | 4.35 |
| PH 3 | 72 | 0.86 | <0.05 | 0.35 | 1.1 | 2.91 | 4.23 |
| PH 4+ | 35 | 1.00 | 0.19 | 1.59 | 2.68 | 6.1 | 11.65 |
Table 3a.
Smoothed Percentiles for Inhibin B for Peri-Dubertal US Girls Ages 6.00-11.99 (N=705) Using the 2-Dart Model
| Age, y |
Inhibin B, pg/ml
|
||||||
|---|---|---|---|---|---|---|---|
| Smoothed Prob. of Exceeding LOD | Coefficient of Variation | Smoothed Percentiles |
|||||
| 10th | 25th | 50th | 75th | 90th | |||
| 6.00-6.49 | 0.09 | 0.39 | <7.0 | <7.0 | <7.0 | <7.0 | <7.0 |
| 6.50-6.99 | 0.09 | 0.44 | <7.0 | <7.0 | <7.0 | <7.0 | <7.0 |
| 7.00-7.49 | 0.10 | 0.50 | <7.0 | <7.0 | <7.0 | <7.0 | <7.0 |
| 7.50-7.99 | 0.13 | 0.55 | <7.0 | <7.0 | <7.0 | <7.0 | 7.78 |
| 8.00-8.49 | 0.18 | 0.60 | <7.0 | <7.0 | <7.0 | <7.0 | 11.65 |
| 8.50-8.99 | 0.25 | 0.63 | <7.0 | <7.0 | <7.0 | <7.0 | 17.29 |
| 9.00-9.49 | 0.33 | 0.64 | <7.0 | <7.0 | <7.0 | 11.31 | 24.21 |
| 9.50-9.99 | 0.43 | 0.66 | <7.0 | <7.0 | <7.0 | 16.67 | 31.85 |
| 10.00-10.49 | 0.52 | 0.67 | <7.0 | <7.0 | 7.64 | 22.60 | 40.13 |
| 10.50-10.99 | 0.59 | 0.69 | <7.0 | <7.0 | 12.47 | 28.47 | 48.39 |
| 11.00-11.49 | 0.63 | 0.71 | <7.0 | <7.0 | 15.81 | 34.22 | 56.71 |
| 11.50-11.99 | 0.67 | 0.73 | <7.0 | <7.0 | 19.51 | 40.58 | 65.87 |
The coefficients of variation only apply to the data above the LOD.
Table 3b.
Smoothed Z Scores for Inhibin B for Peri-pubertal US Girls Ages 6.00-11.99 (N=705) Using the 2-part Model
| Age, y | Smoothed Prob. of Exceeding LOD | Coefficient of Variation | Estimated Smoothed Z scores for Inhibin B, pg/ml |
||||||
|---|---|---|---|---|---|---|---|---|---|
| -1.5 | -1 | -0.5 | 0 | 0.5 | 1 | 1.5 | |||
| 6.00 -6.49 | 0.09 | 0.39 | <7.0 | <7.0 | <7.0 | <7.0 | <7.0 | <7.0 | 6.32 |
| 6.50 -6.99 | 0.09 | 0.44 | <7.0 | <7.0 | <7.0 | <7.0 | <7.0 | <7.0 | 6.89 |
| 7.00 -7.49 | 0.10 | 0.50 | <7.0 | <7.0 | <7.0 | <7.0 | <7.0 | <7.0 | 8.12 |
| 7.50 -7.99 | 0.13 | 0.55 | <7.0 | <7.0 | <7.0 | <7.0 | <7.0 | <7.0 | 10.83 |
| 8.00 - 8.49 | 0.18 | 0.60 | <7.0 | <7.0 | <7.0 | <7.0 | <7.0 | 7.15 | 15.61 |
| 8.50 - 8.99 | 0.25 | 0.63 | <7.0 | <7.0 | <7.0 | <7.0 | <7.0 | 11.94 | 22.77 |
| 9.00 - 9.49 | 0.33 | 0.64 | <7.0 | <7.0 | <7.0 | <7.0 | 7.72 | 17.49 | 31.11 |
| 9.50 - 9.99 | 0.43 | 0.66 | <7.0 | <7.0 | <7.0 | <7.0 | 13.28 | 23.87 | 39.75 |
| 10.00 - 10.49 | 0.52 | 0.67 | <7.0 | <7.0 | <7.0 | 7.64 | 18.74 | 31.01 | 48.80 |
| 10.50 - 10.99 | 0.59 | 0.69 | <7.0 | <7.0 | <7.0 | 12.47 | 23.98 | 38.18 | 57.79 |
| 11.00 - 11.49 | 0.63 | 0.71 | <7.0 | <7.0 | <7.0 | 15.81 | 29.01 | 45.32 | 66.92 |
| 11.50 - 11.99 | 0.67 | 0.73 | <7.0 | <7.0 | <7.0 | 19.51 | 34.60 | 53.18 | 77.06 |
The coefficients of variation only apply to the data above the LOD.
Table 3c.
Weighted Unsmoothed Percentiles for Inhibin B by Pubertal Stages with Non-parametric Ranking for Peri-pubertal US Girls Ages 6.00-11.99(N=705)
| Pubertal Stage | Unweighted N | Proportion Exceeding LOD | 10th | 25th | 50th | 75th | 90th |
|---|---|---|---|---|---|---|---|
| B 1 | 187 | 0.38 | <7.0 | <7.0 | <7.0 | 20.6 | 33.2 |
| B 2 | 123 | 0.60 | <7.0 | <7.0 | 14.2 | 29.4 | 67.95 |
| B 3 | 61 | 0.74 | <7.0 | 14.85 | 40.65 | 71.35 | 71.35 |
| B 4+ | 37 | 0.62 | <7.0 | <7.0 | 35.7 | 55.75 | 83.75 |
| PH 1 | 203 | 0.39 | <7.0 | <7.0 | <7.0 | 20.4 | 33.2 |
| PH 2 | 95 | 0.62 | <7.0 | <7.0 | <7.0 | 36.8 | 75.45 |
| PH 3 | 72 | 0.67 | <7.0 | <7.0 | 28.5 | 55.75 | 71.35 |
| PH 4+ | 34 | 0.68 | <7.0 | <7.0 | 28.6 | 41.5 | 62.55 |
Acknowledgments
This research was supported in part by the Intramural Research Program of the National Institutes of Health, Eunice Kennedy Shriver National Institute of Child Health and Human Development. E.K.S. is supported by NIH grant T32DK065549.
Footnotes
Disclosures: Nothing to declare
References
- 1.Groome NP, Tsigou A, Cranfield M, et al. Enzyme immunoassays for inhibins, activins, and follistatins. Molecular and Cellular Endocrinology. 2001;180:73–77. doi: 10.1016/s0303-7207(01)00523-8. [DOI] [PubMed] [Google Scholar]
- 2.Ravio T, Dunkel L. Inhibins in childhood and puberty. Best Practice and Research Clinical Endocrinology and Metabolism. 2002;16:43–52. doi: 10.1053/beem.2001.0179. [DOI] [PubMed] [Google Scholar]
- 3.Groome NP, Illingworth PJ, O'Brien M, et al. Measurement of dimeric inhibin B throughout the human menstrual cycle. Journal of Clinical Endocrinology and Metabolism. 1996;81:1401–1405. doi: 10.1210/jcem.81.4.8636341. [DOI] [PubMed] [Google Scholar]
- 4.Lahlou N, Roger M. Inhibin B in pubertal development and pubertal disorders. Seminars in Reproductive Medicine. 2004;22:165–75. doi: 10.1055/s-2004-831892. [DOI] [PubMed] [Google Scholar]
- 5.Elsholz D, Padmanabhan V, Rosenfield R, et al. GnRH agonist stimulation of the pituitary-gonadal axis in children: age and sex differences in circulating inhibin-B and activin-A. Human Reproduction. 2004;19:2748–58. doi: 10.1093/humrep/deh519. [DOI] [PubMed] [Google Scholar]
- 6.Bergada I, Rojas G, Ropelato G, et al. Sexual dimorphism in circulating monomeric and dimeric inhibins in normal boys and girls from birth to puberty. Clinical Endocrinology. 1999;51:455–460. doi: 10.1046/j.1365-2265.1999.00814.x. [DOI] [PubMed] [Google Scholar]
- 7.Sehested A, Juul A, Andersson AM, et al. Serum inhibin A and inhibin B in healthy prepubertal, pubertal, and adolescent girls and adult women: relation to age, stage of puberty, menstrual cycle, follicle-stimulating hormone, luteinizing hormone, and estradiol levels. Journal of Clinical Endocrinology and Metabolism. 2000;85:1634–1640. doi: 10.1210/jcem.85.4.6512. [DOI] [PubMed] [Google Scholar]
- 8.Cacciatore B, Apter D, Alfthan H, et al. Ultrasonic characteristics of the uterus and ovaries in relation to pubertal development and serum LH, FSH and estradiol concentrations. Adolescent and Pediatric Gynecology. 1991;4:15–20. [Google Scholar]
- 9.Crofton PM, Evans AE, Groome NP, et al. Dimeric inhibins in girls from birth to adulthood: relationship with age, pubertal stage, FSH and oestradiol. Clinical Endocrinology (Oxf) 2002;56:223–30. doi: 10.1046/j.0300-0664.2001.01449.x. [DOI] [PubMed] [Google Scholar]
- 10.Chada M, Prusa R, Bronsky J, et al. Inhibin B, follicle stimulating hormone, luteinizing hormone, and estradiol and their relationship to the regulation of follicle development in girls during childhood and puberty. Physiological Research. 2003;52:341–6. [PubMed] [Google Scholar]
- 11.Foster CM, Phillips DJ, Wyman T, et al. Changes in serum inhibin, activin and follistatin concentrations during puberty in girls. Human Reproduction. 2000;15:1052–7. doi: 10.1093/humrep/15.5.1052. [DOI] [PubMed] [Google Scholar]
- 12.Lee PA, Xenakis T, Winer J, et al. Puberty in girls: correlation of serum levels of gonadotropins, prolactin, androgens, estrogens, and progestins with physical changes. Journal of Clinical Endocrinology and Metabolism. 1976;43:775–84. doi: 10.1210/jcem-43-4-775. [DOI] [PubMed] [Google Scholar]
- 13.Apter D. Serum steroids and pituitary hormones in female puberty: a partly longitudinal study. Clinical Endocrinology. 1980;12:107–20. doi: 10.1111/j.1365-2265.1980.tb02125.x. [DOI] [PubMed] [Google Scholar]
- 14.Neely EK, Hintz RL, Wilson DM, et al. Normal ranges for immunochemiluminometric gonadotropin assays. Journal of Pediatrics. 1995;127:40–6. doi: 10.1016/s0022-3476(95)70254-7. [DOI] [PubMed] [Google Scholar]
- 15.Plan and Operation of the Third National Health and Nutrition Examination Survey, 1988-1994. Series 1: programs and collection procedures. Vital and Health Statistics 1. 1994;32:1–407. [PubMed] [Google Scholar]
- 16.Gollenberg AL, Hediger ML, Lee PA, et al. Association between lead and cadmium and reproductive hormones in peripubertal U.S. girls. Environmental Health Perspectives. 2010;118:1782–7. doi: 10.1289/ehp.1001943. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Centers for Disease Control and Prevention Blood lead levels- United States, 1999-2002. Morbidity and Mortality Weekly Report. 2005;54:513–16. [PubMed] [Google Scholar]
- 18.Looker AC, Dallman PR, Carroll MD, et al. Prevalence of iron deficiency in the United States. Journal of the American Medical Association. 1997;277:973–6. doi: 10.1001/jama.1997.03540360041028. [DOI] [PubMed] [Google Scholar]
- 19.Hoaglin DC, Mosteller F, Tukey JW. Understanding Robust and Exploratory Data Analysis. John Wiley & Sons, Inc.; New York: 1983. [Google Scholar]
- 20.Smith AR. [October 3,2011];Spline Tutorial Notes. 1983 77 Available at http://www.intelligence.tuc.gr/~petrakis/courses/computervision/splines.pdf. [Google Scholar]
- 21.Cole TJ, Green PJ. Smoothing reference centile curves: The LMS method and penalized likelihood. Stat Med. 1992;11:1305–19. doi: 10.1002/sim.4780111005. [DOI] [PubMed] [Google Scholar]
- 22.van Buuren S, Fredriks M. Worm plot: A simple diagnostic device for modeling growth reference curves. Stat Med. 2001;20:1259–77. doi: 10.1002/sim.746. [DOI] [PubMed] [Google Scholar]
- 23.Zhang Z, Addo OY, Himes JH, et al. A two-part model for reference curve estimation subject to a limit of detection. Statistics in medicine. 2011;30:1455–1465. doi: 10.1002/sim.4189. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Tobin J. Estimation of Relationships for Limited Dependent Variables. Econometrica. 1958;26:24–36. [Google Scholar]
- 25.Mehta NN, Krishnamoorthy P, Martin SS, et al. Usefulness of insulin resistance estimation and the metabolic syndrome in predicting coronary atherosclerosis in type 2 diabetes mellitus. American Journal of Cardiology. 2011;107:406–411. doi: 10.1016/j.amjcard.2010.09.035. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Reilly MP, Wolfe ML, Localio AR, et al. Coronary artery calcification and cardiovascular risk factors: impact of the analytic approach. Atherosclerosis. 2004;173:69–78. doi: 10.1016/j.atherosclerosis.2003.10.010. [DOI] [PubMed] [Google Scholar]
- 27.Mogensen SS, Aksglaede L, Mouritsen A, et al. Diagnostic work-up of 449 consecutive girls who were referred to be evaluated for precocious puberty. Journal of Clinical Endocrinology and Metabolism. 2011;96:1393–1401. doi: 10.1210/jc.2010-2745. [DOI] [PubMed] [Google Scholar]
- 28.Novotny R, Daida Y, Morimoto Y, et al. Puberty, body fat, and breast density in girls of several ethnic groups. American Journal of Human Biology. 2011;23:359–65. doi: 10.1002/ajhb.21145. [DOI] [PubMed] [Google Scholar]
- 29.Cuny A, Trivin C, Brailly-Tabard S, et al. Inhibin B and antimullerian hormone as markers of gonadal function after treatment for medulloblastoma or posterior fossa ependymoma during childhood. Journal of Pediatrics. 2011;158:1016–22. doi: 10.1016/j.jpeds.2010.11.019. [DOI] [PubMed] [Google Scholar]
- 30.Buyken AE, Karaolis-Danckert N, Remer T. Association of prepubertal body composition in healthy girls and boys with the timing of early and late pubertal markers. American Journal of Clinical Nutrition. 2009;89:221–30. doi: 10.3945/ajcn.2008.26733. [DOI] [PubMed] [Google Scholar]
- 31.Boyne MS, Thame M, Osmond C, et al. Growth, body composition, and the onset of puberty: longitudinal observations in Afro-Caribbean children. Journal of Clinical Endocrinology and Metabolism. 2010;95:3194–200. doi: 10.1210/jc.2010-0080. [DOI] [PubMed] [Google Scholar]
- 32.Lee JM, Appugliese D, Kaciroti N, Corwyn RF, Bradley RH, Lumeng JC. Weight status in young girls and the onset of puberty. Pediatrics. 2007;119:e624–30. doi: 10.1542/peds.2006-2188. [DOI] [PubMed] [Google Scholar]
- 33.Demerath EW, Li J, Sun SS, Chumlea WC, et al. Fifty-year trends in serial body mass index during adolescence in girls: the Fels Longitudinal Study. American Journal of Clinical Nutrition. 2004;80:441–6. doi: 10.1093/ajcn/80.2.441. [DOI] [PubMed] [Google Scholar]
- 34.Kaplowitz PB, Slora EJ, Wasserman RC, et al. Earlier onset of puberty in girls: relation to increased body mass index and race. Pediatrics. 2001;108:347–53. doi: 10.1542/peds.108.2.347. [DOI] [PubMed] [Google Scholar]
- 35.Herman-Giddens ME, Slora EJ, Wasserman RC, et al. Secondary sexual characteristics and menses in young girls seen in office practice: a study from the Pediatric Research in Office Settings network. Pediatrics. 1997;99:505–12. doi: 10.1542/peds.99.4.505. [DOI] [PubMed] [Google Scholar]
- 36.Kalra B, Kumar A, Patel K, et al. Development of a second generation Inhibin B ELISA. Journal of Immunological Methods. 2010;362:22–31. doi: 10.1016/j.jim.2010.08.002. [DOI] [PubMed] [Google Scholar]