The Use of Morning Urinary Gonadotropins and Sex Hormones in the Management of Early Puberty in Chinese Girls.

CONTEXT: Although gonadotropin-releasing hormone stimulation test (GnRHST) is the gold standard in diagnosing central precocious puberty (CPP), it is invasive, expensive, and time-consuming, requiring multiple blood samples to measure gonadotropin levels.
OBJECTIVE: We evaluated whether urinary hormones could be potential biomarkers for prepuberty or postpuberty, aiming to simplify the current diagnosis and prognosis procedure.
METHODS: We performed a cross-sectional study of a total of 355 girls with CPP in National Clinical Research Center for Child Health in China, including 258 girls with positive and 97 girls with negative results from GnRHST. Twenty patients received GnRH analogue (GnRHa) treatment and completed a 6-month follow up. We measured luteinizing hormone (LH), follicle-stimulating hormone (FSH), estradiol, prolactin, progesterone, testosterone, and human chorionic gonadotropin in the first morning voided urine samples.
RESULTS: Their urinary LH levels and the ratios of LH to FSH increased significantly with the advancement in Tanner stages. uLH levels were positively associated with basal and peak LH levels in the serum after GnRH stimulation. A cutoff value of 1.74 IU/L for uLH reached a sensitivity of 69.4% and a specificity of 75.3% in predicting a positive GnRHST result. For the combined threshold (uLH ≥ 1.74 + uLH-to-uFSH ratio > 0.4), the specificity reached 86.6%. After 3 months of GnRHa therapy, the uLH and uFSH levels decreased accordingly.
CONCLUSION: uLH could be a reliable biomarker for initial CPP diagnosis and screening; uLH could also be an effective marker for evaluating the efficacy of clinical treatment.

Central precocious puberty (CPP) is the result of an early activation of the hypothalamic-pituitary-gonadal axis (HPG axis) (1). Because CPP impairs both physical and psychosocial growth in children (2), testing based especially on hormone measurements is very important for early diagnosis and treatment. The gonadotropin-releasing hormone (GnRH) stimulation test (GnRHST) has been the gold standard for confirming the activation of the HPG axis (3). GnRHST requires outpatient hospitalization and a sequence of blood collections that are not only time-consuming but are also costly and painful. Therefore, an updated proposal and international consortium suggested the use of unstimulated luteinizing hormone (LH) levels in random blood samples instead of GnRHST (4).

Analysis of urinary gonadotropins for differentiating prepubertal and postpubertal individuals was initially carried out almost 50 years ago. This study revealed that sexual maturation is accompanied with more excretion of LH and follicle-stimulating hormone (FSH) in the urine (5). Owing to a marked increase in the night-time amplitude and frequency of gonadotropins (6), gonadotropin concentrations in the first morning voided (FMV) urine samples may reflect the nocturnal secretion and can be used as one of the potential alternative approaches for assessing pubertal development. Even small amounts of intact gonadotropins can now be detected owing to the development of third-generation assays with improved sensitivity (7). These assays have led to the use of urine as a sample of choice, rather than the use of serum.

Various studies have reported consistency between serum and urinary gonadotropins. FMV urinary LH (uLH) correlates with the basal and GnRH-stimulated serum LH (sLH) and gives a sensitivity of 75% and a specificity of 92% according to the GnRHST results (8). Assessment of puberty signs with urinary gonadotropins helps to distinguish the slowly progressive and rapidly progressive precocious puberty (PP) (9). In addition, a recent study reported that FMV uLH provides accurate information about puberty suppression and can be used to monitor the adequacy of GnRH analogue (GnRHa) therapy (10).

Therefore, in the present study we conducted analyses to validate the use of FMV gonadotropins and sex hormones in the diagnosis of early puberty in girls and to establish cutoff values of these biomarkers. We also validated whether FMV sampling is adequate for monitoring the clinical efficacy of GnRHa therapy.

Materials and Methods

Participants

This study was approved by the medical ethics committee of the Children’s Hospital, Zhejiang University School of Medicine (No. 2020-IRB-167). Chinese girls with clinical signs of precocious puberty and concerns were admitted to the Department of Endocrinology, Children’s Hospital, Zhejiang University School of Medicine. Detailed explanations of the study were conveyed to each individual and her parents, and informed consents were obtained.

Anthropometry, Pubertal Staging, and Imaging

Height, weight, and body mass index (BMI) were recorded at enrollment. BMI was calculated as weight in kilograms divided by height in meters squared. All of these parameters were converted into Z scores using Child Growth Standards from the World Health Organization (11). Pubertal status was staged according to Tanner Breast and Pubic Hair. To improve the validity, bone age was calculated by the Tanner-Whitehouse method.

Pelvic ultrasound (Philips IU Elite) was performed on all patients to assess uterine volume, cervical length, and follicular volume. Pituitary magnetic resonance imaging (Philips Prodiva CX 1.5T) was performed routinely to exclude intracranial causes.

Study Design

The GnRHST was used to determine CPP patients. GnRH (2.5-3 μg/kg, maximum 100 μg, Triptorelin Acetate, Ipsen Pharma Biotech) was administered intravenously, and serum LH and FSH levels were measured at 0-, 30-, 60-, 90-, and 120-minute intervals. Patients were diagnosed as CPP cases when they had peak LH level exceeding 5 IU/L in response to GnRHST and early puberty signs before age 8 years, including progressive breast development accompanied by rapid growth (> 6-7 cm/year) and bone maturation. Patients with an evidence of tumor, adrenal disease, gonadal disease or central nervous system disorder were excluded from this study. Plasma thyroxin and thyrotropin concentrations were measured to exclude hypothyroidism.

The FMV urine specimens were collected in the morning the day of GnRHST and were stored at 4 °C immediately. After centrifuging at 3000 rpm for 10 minutes, urinary supernatants were collected and stored at –80 °C until the hormone assay. Total urine volumes were recorded.

CPP patients were treated with either leuprolide acetate or triptorelin acetate, at a dose of 3.75 mg every 28 days. Stimulated peak serum LH, FSH, and estradiol (E2) concentrations were evaluated 45 minutes after GnRHa injection at 3 and 6 months. Baseline and posttreatment FMV urine samples were collected from 20 girls who were diagnosed with CPP in this cohort.

Hormone Assays

Gonadotropins and other sex hormones levels in urine and serum were measured by immunochemiluminescence assays (ICMAs) (Siemens Healthcare Diagnostics Products Ltd). E2 levels were measured by ICMA (Siemens Healthcare Diagnostics Inc). Urinary creatinine was measured by a biochemical analyzer (Mindray BS-820) and was used as an adjustment factor for urine dilution.

The intra-assay coefficients of variation and the interassay coefficients of variation were less than 3.7% and 7% for LH, 3.4% and 4.8% for FSH, 4% and 4.3% for E2, 3.3% and 4.8% for prolactin (PRL), 9.7% and 12.5% for progesterone (P), 11.7% and 10.3% for testosterone (T), 6.6% and 7.4% for human chorionic gonadotropin (hCG), respectively. The limits of detection (analytical sensitivity) for LH, FSH, E2, PRL, P, T, and hCG were 0.1 IU/L, 0.1 IU/L, 43.6 pmol/L, 10.6 mIU/L, 0.64 nmol/L, 0.69 nmol/L and 1 IU/L, respectively.

Statistical Analyses

For statistical evaluation, concentrations below the detection limits were given a value of the lowest standard. To minimize the possible error in renal function and wide difference in body size, hormone values were multiplied by the corresponding total urine volume (×mL) or were divided by creatinine excretion levels (/uCr).

Statistical analyses were performed using SPSS software (IBM, version 23) and figures were created by GraphPad Prism (version 8) and by the R programming language (version 3.6.0). Normally distributed data were expressed as mean ± SD. Otherwise, the data were represented as median (P25-P75). The t test, Mann-Whitney U test, and chi-square test were performed separately to compare measurements between girls with or without peak LH greater than 5 IU/L (ie, to compare GnRHST positives vs GnRHST negatives). The urinary gonadotropins medians at different pubertal stage groups were compared using a Kruskal-Wallis test. The relationships between urinary and serum hormone levels were further compared using a heat map of correlation coefficients.

The diagnostic value of the urine hormones’ concentrations was evaluated using receiver operating characteristic (ROC) analysis. Optimal cutoff values for biomarkers were computed using sensitivity and specificity. Sensitivity is defined as the probability of detecting positive cases in GnRHST-positive individuals, and specificity is defined as the probability of detecting negative cases in GnRHST-negative individuals. The Youden index, a summary measure of diagnostic accuracy, is computed by subtracting one from the sum of sensitivity and specificity. The optimal cutoff value for a single biomarker or for the linear combination of biomarkers in the ROC model (a logistic regression model) was computed using the probability of detecting positive cases that maximizes the Youden index.

A Wilcoxon signed rank test was performed to compare hormone concentrations between pretreatment and posttreatment groups. Differences were considered statistically significant if P was less than .05.

Results

General Characteristics

A total of 355 girls who presented early signs of puberty before age 8 years were recruited, and they were aged between 3.92 and 9.83 years during the first hospitalization. The 258 patients with a peak LH of greater than 5 IU/L were considered as the GnRHST-positive (+) group, and the rest were considered as the GnRHST-negative (–) group. Baseline clinical and radiological characteristics of the participants are shown in Table 1. The average age of the GnRHST (+) group was 8.23 years, whereas that of the GnRHST (–) group was 7.82 years. The GnRHST (+) group had advanced Tanner stages, larger uterine volume, and longer cervical length. Significant increases in the basal sLH and serum FSH levels were observed in the GnRHST (+) group.

Table 1.

Comparison of auxological and laboratory data between gonadotropin-releasing hormone stimulation test positive and negative groups

Characteristics	Total	GnRHST (+)	GnRHST (–)	P
Patient No.	355	258	97	–
CA, y	8.12 ± 0.77	8.23 ± 0.67	7.82 ± 0.94	< .001a
BA-CA, y	1.61 ± 0.93	1.65 ± 0.97	1.50 ± 0.82	.233
Wt Z score	0.67 ± 0.92	0.67 ± 0.88	0.67 ± 1.03	.933
Ht Z score	0.77 ± 0.93	0.81 ± 0.90	0.65 ± 1.00	.143
BMI Z score	0.34 ± 0.96	0.31 ± 0.93	0.41 ± 1.04	.388
Tanner B, 1/2/3/4	0/217/133/5	0/144/110/4	0/73/23/1	.004a
Tanner PH, 1/2/3	331/23/1	244/13/1	87/10/0	.167
Uterine volume, cm3	3.91 (2.66-5.61)	4.57(3.15-6.22)	2.72 (2.02-3.44)	< .001a
Uterine + cervical length, cm	4.20 (3.80-4.6)	4.3(3.9-4.7)	4.0 (3.7-4.2)	< .001a
Fundocervical ratio	1.41 (1.28-1.56)	1.43 (1.31-1.58)	1.36 (1.22-1.53)	.01a
Basal LH, IU/L	0.28 (0.13-0.89)	0.50 (0.19-1.19)	0.13 (0.10-0.22)	< .001a
Basal FSH, IU/L	2.70 (1.66-4.18)	3.12 (2.00-4.57)	1.75 (1.34-2.68)	< .001a
Basal E2, pmol/L	171.0 (121.0-229.9)	181.5 (123.5-241.8)	151.5 (111.0-184.0)	.055
Basal PRL, mIU/L	145.5(106.0-237.0)	146.5 (108.5-237.8)	139.0 (104.23-236.0)	.089

Values are presented as mean ± SD for normal data which were compared by t test and the median (25%-75%) for nonnormal data, which were compared by Mann-Whitney U test.

Abbreviations: BA, bone age; BMI, body mass index; CA, chronological age; E2, estradiol; FSH, follicle-stimulating hormone; GnRHST, gonadotropin-releasing hormone stimulation test; LH, luteinizing hormone; PRL, prolactin; Tanner B, Tanner breast stage; Tanner PH, Tanner pubic hair.

aP < .05.

Urinary Gonadotrophins in Relation to Pubertal Status

All girls in this study were of breast Tanner stage 2, 3, or 4. Overall, the levels of urinary gonadotrophins increased with pubertal advancement (Fig. 1). Significant differences in levels of uLH, uLH-to-uCr, and uLH-to-uFSH were observed in girls with advanced Tanner stages. However, overlapping hormone levels were observed among the Tanner stages both in the original and creatinine-normalized hormones.

Figure 1.

Distribution of unadjusted and adjusted hormone levels in different Tanner breast stages. Changes of A, uLH; B, uLH-to-uCr; C, uFSH; D, uFSH-to-uCr; E, uE2; F, uE2-to-uCr; G, uP; H, uP-to-uCr; I, uPRL; J, uPRL-to-uCr; K, uHCG; L, uHCG-to-uCr; M, uT; N, uT-to-uCr; and O, uLH-to-uFSH ratio in advanced Tanner Breast stages, respectively. uE2, urinary estradiol; uE2-to-uCr, adjusted uE2 by urinary creatinine; uFSH, urinary follicle-stimulating hormone; uFSH-to-uCr, adjusted uFSH by urinary creatinine uHCG, urinary human chorionic gonadotropin; uHCG-to-uCr, adjusted uHCG by urinary creatinine; uLH, urinary luteinizing hormone; uLH-to-uCr, adjusted uLH by urinary creatinine; uLH-to-uFSH ratio, unadjusted uLH divided by uFSH; uP, urinary progesterone; uP-to-uCr, adjusted uP by urinary creatinine; uPRL, urinary prolactin; uPRL-to-uCr, adjusted uPRL by urinary creatinine; uT, urinary testosterone; uT-to-uCr, adjusted uT by urinary creatinine;. The lines in the boxes represent the median values, the lower and upper limits of the boxes correspond to the 25th and 75th percentiles, and the whiskers represent the 95% CI limits. Note the logarithmic scale of the Y axes. (*P < .05. **P < .01. ***P < .001).

Correlation Between Urinary and Serum Hormone Levels

The Pearson correlation coefficient was calculated to evaluate the correlation between urinary hormone levels and blood indexes or ovary maturation (Fig. 2). Positive correlations were found between uLH and sLH both in the basal (r = 0.451, P < .001) and peak levels (r = 0.326, P < .001). These correlations also were confirmed using normalized uLH levels adjusted by the total volume of urine or uCr. Positive correlations were found between uLH-to-uFSH ratio and basal LH, sum LH, peak LH, insulin-like growth factor-1, uterine volume, and uterine length.

Figure 2.

Correlations between urinary hormone concentrations and blood hormone concentrations, uterine indexes. Sum, the summation of corresponding gonadotropin concentrations measured at measured at 0-, 30-, 60-, 90-, and 120-minute intervals in gonadotropin-releasing hormone stimulation test.

Receiver Operating Characteristic Analysis

Using the 355 patients who received GnRHST, we performed a ROC analysis to estimate the optimal cutoff value for predicting a positive GnRHST result (peak LH > 5 IU/L) (Fig. 3). uLH or uCr-adjusted uLH had the best performance of any individual urinary hormone as indicated by having the largest area under the ROC curve (AUC). However, it was interesting that the AUC of the uLH-to-uFSH ratio was even larger and reached a value of 0.853.

Figure 3.

Receiver operating characteristic (ROC) analysis to calculate the optimal cutoff point for a single biomarker. ROC curve of unadjusted hormone concentrations and urinary luteinizing hormone (uLH) to urinary follicle-stimulating hormone (uFSH) ratio. A, ROC curve of adjusted urinary hormone concentrations.

With a sensitivity of 95%, which could possibly detect the great majority of positive cases, the optimal uLH cutoff value for screening was 0.55 IU/L (Table 2). The cutoff value for diagnosis was 1.74 IU/L with a sensitivity of 69.4% and a specificity of 75.3%. Based on AUC for a single marker, a uLH cutoff value of 1.74 IU/L combined with a uLH-to-uFSH ratio of greater than 0.4 had a higher specificity of 86.6%.

Table 2.

Different cutoff values of urinary luteinizing hormone, urinary follicle-stimulating hormone, their ratio, and combination algorithm on gonadotropin-releasing hormone stimulation test results

	Cutoff value	Sensitivity, %	Specificity, %	Youden index
uLH, IU/L	> 0.55	95.0	25.8	0.207
	> 1.74	69.4	75.3	0.446
	> 20.1	7.4	99	0.06
uFSH, IU/L	> 1.23	95.0	5.2	0.02
	> 4.09	56.6	66.0	0.226
	> 21.05	4.3	99.0	0.032
uLH-to-uFSH	> 0.416	78.68	80.41	0.591
uLH ≥ 1.74 IU/L + uLH-to-uFSH > 0.4		65.50	86.60	0.521
0.399*uLH-0.17*uFSH > –0.087		78.8	80.4	0.591

Abbreviations: uFSH, urinary follicle-stimulating hormone; uLH, urinary luteinizing hormone.

We then created a scatterplot in which uLH and uFSH were used for a single data series. The GnRHST-positive cases were marked in blue and the negative cases were marked in red. Logistic regression was applied to calculate the predictive probability based on using both uLH and uFSH for the diagnosis of CPP (Fig. 4). An ROC curve for the combination of biomarkers was constructed based on logistic regression, and the resulting AUC was higher than the AUCs obtained using single biomarkers. Table 2 lists the logistic regression analysis parameters for CPP diagnosis compared with single markers. The results showed that when 0.399*uLH-0.17*uFSH > –0.087, the patient was predicted to have GnRHST-positive results. The results showed that the sensitivity of CPP diagnosis was 78.8% and specificity was 80.4% when calculated based on the maximum Youden index.

Figure 4.

Receiver operating characteristic (ROC) analysis to calculate the optimal cutoff point for the linear combination of urinary luteinizing hormone (uLH) and urinary follicle-stimulating hormone (uFSH). A, Scatter plot of uLH and uFSH. B, ROC curve of combination of uLH and uFSH.

Follow-up Study

Twenty-three girls received GnRHa therapy, and 20 of them were enrolled in our study. Before each injection, FMV urine samples were collected from 20 participants at 3 months and at 6 months of treatment. After the subcutaneous administration of GnRHa, blood samples were collected 45 minutes later and serum LH, FSH, and E2 levels were checked.

All postpeak LH levels throughout this follow-up study were under 1.7 IU/L, which indicates that HPG axis was well suppressed by the GnRHa treatment. The mean uLH was 0.37 ± 0.20 with a range of 0.16 to 0.84 IU/L at 3 months, and at 6 months the mean was 0.50 ± 0.25 with a range of 0.24 to 1.07 IU/L. Stimulated peak LH was highest at baseline and declined significantly after 3 months of GnRHa therapy (Fig. 5). The uLH and uFSH levels decreased at 8 weeks, indicating good efficacy of the therapy; however, the levels increased slightly at 6 months (Table 3).

Figure 5.

The changing tendency of serum and urinary hormone levels during the follow-up assessment.

Table 3.

Comparison of serum and urinary gonadotropins between pretreatment and posttreatment levels

3 mo (compared to initial treatment)	Serum peak LH	uLH	uFSH	uE2
z score	–3.920a,e	–3.92a,e	–3.92a,e	–2.464a,c
P	< .001	< .001	< .001	.014
6 mo (compared to 8 wk)	Serum peak LH	uLH	uFSH	uE2
z score	–0.543a	–2.375b,c	–2.875b,d	–1.269b
P	.587	.018	.004	.204

Abbreviations: E2, estradiol; LH, luteinizing hormone; uE2, urinary estradiol; uFSH, urinary follicle-stimulating hormone; uLH, urinary luteinizing hormone.

Based on positive ranks.

Based on negative ranks.

P less than .05.

P less than .01.

P less than .001.

Discussion

In this study we evaluated the diagnostic value of urinary hormones in CPP girls. We found that uLH and uLH-to-uFSH ratio were associated with sLH levels and Tanner stages, which could predict puberty status. We concluded that uLH could be a convenient and noninvasive biomarker as compared to the serum peak LH of GnRHST. The optimal cutoff values of uLH were 0.55 IU/L for screening and 1.74 IU/L for diagnosis, respectively.

With advanced techniques of ICMA and electrochemiluminescence immunoassay, several studies measured urinary gonadotropins in children going through puberty and reported that urinary LH performed well in predicting the pubertal process (Table 4). Sex hormones in the urine, including LH, T, and E2, correlated cross-sectionally and longitudinally with age, anthropometry, and Tanner stages (12). Urinary gonadotrophins reflected the clinical/physical status using the ratio of LH to FSH (13). Furthermore, FMV uLH levels were strongly correlated with stimulated LH levels (r = 0.91) and with basal LH levels in the blood (r = 0.65) (10). FMV uLH and uLH-to-uFSH performed equally well as GnRHST in distinguishing early puberty and prepuberty status (14). Similarly, another study found that uFSH-to-uCr ratio and uLH-to-uCr ratio correlated with serum FSH and sLH levels, respectively; the authors recommended that uLH-to-uCr might play a role in identifying CPP and that uFSH-to-uCr assessment could be used to recognize germ cell failure (15). Based on Zung’s analysis, an FMV uLH cutoff value of 1.16 IU/L had higher sensitivity (83%) and predictive values than that of serum basal LH (9). Likewise, it was reported that a cutoff SD score of 2 for uLH had a sensitivity of 75% and a specificity of 92% in the COPENHAGEN puberty study with a large population (8). In another important study random urinary gonadotropins were recommended for the initial assessment of CPP in girls because there was no difference between FMV and random urinary gonadotropins levels (16). We consider that the main reason for various cutoff values was difference in sample size and hormone assay.

Table 4.

Urinary luteinizing hormone cutoff values used to diagnose precocious puberty

	Sensitivity, %	Specifity, %	Patient No., sex	Publication y	Assay	Study
FMV uLH > 1.01 IU/L to predict RP-PP	83	72	47, F	2014	Two-site immunochemiluminescence assay	Zung et al (9)
FMV uLH > 1.75 IU/L	91.5	82.7	138, M	2016	Sandwich fluoroimmunoassays	Demir et al (14)
FMV uLH > 1.2 IU/L	80	74	52, F
Nontimed uLH-to-uCr > 0.05 IU/mmol	86	71	41, M	2016	Chemiluminescent microparticle immunoassays	Lucaccioni et al (15)
Random uLH > +2 SD	75	92	12 of 479, F	2017	Time-resolved immunofluorometric assays	Kolby et al (8)
FMV uLH > 0.58 IU/L	91.9	63.2	100, F	2019	Electrochemiluminescence immunoassays	Shim et al (16)
Random uLH > 0.2 IU/L	77.4	73.7
FMV uLH > 1.01 mIU/mL	92.3	100	68, F	2020	Electrochemiluminescence assays	Yüce et al (10)

Abbreviations: F, female; FMV, first morning voided; M, male; RP-PP, rapidly progressive precocious puberty; uCr, urinary creatinine; uLH, urinary luteinizing hormone.

In general, our results are consistent with previous findings, and we recommend uLH as a superior alternative to GnRHST. A significant increase in levels of uLH, uLH-to-uCr, and uLH-to-uFSH with the advanced Tanner stages was observed in girls (12). A cutoff value of 1.74 IU/L for FMV uLH reached a sensitivity of 69.4% and a specificity of 75.3% in predicting a positive GnRHST; however, when 0.399*uLH-0.17*uFSH > –0.087, the patient was predicted to have GnRHST-positive results with a sensitivity of 78.8% and a specificity of 80.4%. For the combined threshold of the urinary gonadotropins (uLH ≥ 1.74 + uLH-to-uFSH ratio > 0.4), the specificity was 86.6%. A cutoff value of 0.55 IU/L for uLH showed a sensitivity of 95%, indicating the potential value of uLH for screening in large-scale population and diagnosis in clinical use. We suggest that girls who reach the cutoff level should receive further physical examination to confirm Tanner stage. Compared to uLH, urinary E2, T, HCG, and PRL levels overlapped among Tanner stages and therefore had poor correlation with puberty advancement.

The gold-standard biochemical diagnosis of CPP is mainly based on the assessment of peak gonadotropin levels in GnRHST. The disadvantages of GnRHST are the high cost and invasiveness. Other alternatives to GnRHST are serum basal gonadotropin levels, imaging, and gene expression (for familial PP). Basal serum LH levels were suggested by Heo et al (2019) as the most sensitive parameter for screening CPP and early puberty (17). Other investigators recommended basal LH levels instead of GnRHST, but the cutoff values of basal LH varied from 0.1 to 0.83 IU/L (ICMA) with 64% to 93% diagnostic sensitivity (18, 19). Similarly, pelvic ultrasonography or pituitary magnetic resonance imaging could not distinguish prepubertal and pubertal stages (20, 21). As a major inhibitor of hypothalamic GnRH secretion, serum MKRN3 concentration declines before puberty onset in girls and might be useful for the assessment of pubertal development (22); however, it is also invasive. In comparison, urinary gonadotropins can provide a simple, inexpensive, noninvasive, and well-accepted alternative to assess pubertal stages and the efficacy of CCP therapy; thus, urinary gonadotropins may even help to build a monitoring/screening system for pediatric growth in a large population.

During the follow-up period of our study, pituitary suppression during GnRHa therapy was optimally assessed by GnRH stimulation, and suppressed gonadotropin levels indicated the effectiveness of GnRHa therapy in CPP patients. In clinical practice the common method to monitor the efficacy of CPP therapy is to inject GnRHa monthly and to examine the plasma LH levels every 3 to 4 months. A single serum sample obtained 30 to 60 minutes after leuprolide injection is an accurate way to evaluate treatment efficacy (23). Peak plasma LH levels of less than 2 IU/L are considered indicative of biochemical suppression (24). Efforts were made to simplify this test by using preinfection serum basal LH, but they failed (25); so researchers tried to replace GnRHST by evaluating FMV urinary hormone levels before the administration of GnRHa. Zung et al (26) found that 8% of uLH levels in 36 CPP patients with adequate gonadotropin suppression were above the prepubertal threshold. Their results suggested that uLH could not fully replace the sLH in monitoring CPP. Another study conducted by Yüce et al (10) showed a significant difference in FMV uLH levels between suppressed and unsuppressed participants. In the COPENHAGEN puberty study, urinary concentrations of LH decreased after 3 months of GnRH treatment to levels below +2 SD. In our study the pronounced difference in uLH levels between pretreatment and posttreatment time points indicated the good efficacy. Several cases showed elevated uLH levels at 20 weeks; however, their uLH levels were not higher than baseline. Pubertal development and body growth should be examined every 3 to 6 months, and repeated blood tests are not necessary if the uLH or uLH-to-uFSH ratio is not elevated (27). By no means should clinical manifestations of pubertal regression or arrest be given priority over hormonal tests because the probability of a positive GnRHST response after GnRHST increases with the increasing difference between bone age and chronological age (28).

To summarize, the relatively large number of 355 participants enrolled in our study provided convincing cutoff values of uLH for screening and diagnosis of CPP in clinical practice. A few studies found that uLH degrades with time when stored at –20 °C (9, 29); therefore, we stored our samples at –80 °C after centrifugation and extraction. We performed comprehensive assays that cover all sorts of sex hormones in urine. Furthermore, we also recorded the total urine volume and uCr levels to check whether urine volumes or uCr levels affected normalized hormone values. However, the performance of the tests used to recognize precocious puberty differs in younger age groups, especially at ages younger than 3 years (30) because of mini-puberty. The present study does not have this younger age group so further studies are needed to explore the association of urinary hormones and sexual development. Although 20 participants were followed during the 6 months of treatment, none showed inadequate suppression of HPG function. The value and reference of uLH in well-suppressed patients should be confirmed in large-scale and multicenter studies. In addition, randomized urinary gonadotropins were promising in clinical practice (16), so further work is required to clarify whether elevated randomized uLH is sufficiently sensitive to replace FMV uLH.

Conclusion

Based on our results, we recommend uLH as a reliable biomarker for initial diagnosis and effective screening for CPP in suspected populations. In well-suppressed CPP patients, uLH and uFSH decreased significantly as indicators of effective therapy.