Background

Early puberty is one of the prevalent conditions witnessed nowadays in pediatric endocrinology clinics. It is defined as the appearance of secondary sexual characteristics before age 8 years in girls and 9 years in boys, with a clear female predominance [1]. According to the published guidelines by Lawson Wilkins Pediatric Endocrine Society, precocious puberty should be classified according to race as breast development occurring before the age of 7 in Caucasian girls and 6 in African-American girls [2].

Precocious puberty may be associated with psychosocial issues and has detrimental effects on long-term health, such as an increased predisposition to type 2 diabetes [3], weight gain, obesity, cardiovascular disease [4], depressive disorders [5], and early mortality [6], in addition to higher risk of breast cancer [7]. Hence, it has been quite concerning to recognize the earlier timing of puberty as well as the increase in the number of referrals of children with suspected precocious puberty [3].

There are three types of precocious puberty (PP): benign pubertal variants, peripheral precocious puberty (PPP), and central precocious puberty (CPP). PP can be caused by central nervous system tumors, or ovarian tumors or the cause may be unknown (idiopathic) [8]. Central precocious puberty has diverse and heterogeneous etiologies, including congenital and acquired causes leading to structural or functional changes in the brain [9]. Reactivation of the hypothalamic-pituitary–gonadal (HPG) axis prematurely results in earlier pulsatile production of hypothalamic GnRH, which causes CPP [10].

The diagnosis and management of PP constitute a major challenge. To distinguish between different categories of PP, the GnRH stimulation test (GnRHST), which evaluates peak luteinizing hormone (LH) after GnRH agonist stimulation, is still recognized as the gold standard test [1]. However, it necessitates repeated sampling which makes it expensive and time-consuming [11]. Moreover, the inconsistency in age-specific cut-off values, and various timings for LH sampling after stimulation, different gonadotrophin (especially LH) assay techniques, and the unavailability of pharmacological agents for gonadotrophin stimulation render it challenging [12].

Therefore, there is an unmet clinical need to search for simple affordable biomarkers such as serum LH and AMH that could overcome the disadvantages of GnRHST and evaluate their performance in diagnosing CPP in girls.

Methods

This was a retrospective case–control study carried out on 26 Egyptian girls aged less than 8 years who were evaluated for early signs of puberty in the Pediatric Endocrine Clinic at Alexandria University Children’s Hospital in Alexandria, Egypt, in comparison to 26 age-matched healthy girls as controls. Girls with peripheral precocious puberty, those with precocity due to exogenous drugs, and normal variants were excluded from the study.

Demographic data was documented including chronological age, age of onset of breast development, family history, and anthropometric measurements including weight standard deviation score (SD), height SD, and body mass index SD. Tanner pubertal staging of breast, pubic, and axillary hair development were assessed in all girls [13]. The bone age (BA) was measured using the Greulich and Pyle method, and it was considered advanced if the difference between bone age and chronological age was 2 years or more [14]. Ultrasonography of the abdomen and pelvis to assess the dimensions of the uterus was done.

Baseline serum LH, follicle-stimulating hormone (FSH), and estradiol levels were withdrawn to all girls between 9 and 10 am. The GnRHST was performed in patients with prepubertal basal LH levels < 0.5 IU/L. Initially, 100 ug of GnRH analog (GnRHa) (triptorelin 0.1 mg vial) was administered intramuscularly, and then blood samples were collected again at 30, and 45 min. Peak LH concentration after GnRHST ≥ 5 IU/L was defined as being representative of the pubertal pattern [15]. LH and FSH were measured by highly sensitive electrochemiluminescence immunoassay (cobas e801) with a lower detection limit of 0.3 IU/L for both FSH and LH. Basal LH/FSH ratio and LH/FSH ratio following GnRH administration was calculated, and CPP was diagnosed with peak LH/FSH ratio ≥ 0.6. Basal Anti-Mullerian hormone (AMH) was measured by electrochemiluminescence immunoassay (Cobas e801) with a lower detection limit of 0.01 ng/mL and a limit of quantification of 0.03 ng/mL.

Statistical analysis

Data was analyzed using IBM SPSS software package version 20.0. (Armonk, NY: IBM Corp). Qualitative data were described using numbers and percentages. The Shapiro–Wilk test was used to verify normality of distribution. Quantitative data were described using mean ± SD for normally distributed data, median (minimum and maximum), and interquartile range (IQR) for data that are not normally distributed. Chi-square test was used to compare different categorical variables. For non-normally distributed quantitative variables, Mann–Whitney test was used to compare between twoindependent groups and Friedman test to compare between more than two periods. Post Hoc Test (Dunn’s) was used for pairwise comparisons. A level of significance of ≤ 0.05 was considered statistically significant. Receiver operating characteristics (ROC) curve analysis was performed using MedCalc software version 15.8 (MedCalc Software, Ostend, Belgium) to evaluate the diagnostic test accuracy of variables that showed statistical significance by univariate analysis. Based on ROC curves, the best cut-off values of the studied CPP biomarkers were deduced from Youden's index.

Results

This retrospective study enrolled 26 girls diagnosed with CPP at the Pediatric Endocrinology clinic at Alexandria University Children’s Hospital. The demographic data of this cohort are included in Table 1. The mean age of the girls at diagnosis was 5 ± 2.8 years. Half of the girls presented with precocious puberty at age ≥ 4 years, while the remainder were diagnosed before 4 years. The youngest patient in our cohort presented at age 1.3 years.

Table 1 Distribution of the studied cases according to clinical parameters (n = 26)
Full size table

It is noteworthy that the most common presentation was isolated breast enlargement (53.8%), while 9 patients (34.6%) presented with both thelarche and adrenarche, and 3 patients (11.5%) presented after develo** menarche. At presentation, the majority of patients (76.9%) had breast Tanner 3. Almost 54% of the patients had average height for age, while 12 girls (46.2%) had tall stature. Twenty patients (76.9%) had average BMI for age, while the prevalence of obesity in this cohort was 7.7%, while overweight was 15.4%.

As a routine investigation for CPP patients, bone age (BA) was assessed for all CPP patients showing advanced BA in 17 girls (65.4%). Half of the patients (13/26) had pubertal uterine dimensions by ultrasonography. All patients with CPP had magnetic resonance imaging (MRI) of the brain focusing on sella turcica and pituitary gland done, to detect any underlying pathology as an etiology of pubertal precocity. Only 7 patients (26.9%) were diagnosed with organic CPP and had abnormal MRI findings such as tuber cinereum hamartoma (3/42.9%) and arachnoid cyst (2/28.6%), while 19 patients (73.1%) were classified as idiopathic CPP, which was the most common cause of CPP in our cohort. The results of those investigations are shown in Table 2.

Table 2 Distribution of the studied cases according to different investigations (n = 26)
Full size table

Hormonal profile of these patients was assessed and compared to an age-matched control group (n = 26) as shown in Table 3. GnRHST was done in 22 patients (84.6%) showing pubertal response in all of them (peak LH ≥ 5 IU/L). Girls who did not undergo the test had shown evident pubertal basal levels of both LH and FSH. Thus, there was no need to undergo further GnRHST. There was a statistically significant difference in basal LH and FSH levels between CPP and control groups (p < 0.001*, p < 0.001* respectively). However, basal AMH and basal LH/FSH failed to show any statistically significant difference between both groups (p = 0.07, p = 0.913 respectively). Basal estradiol was assessed in only 19 patients according to availability; 12 of them (63.2%) had pubertal levels ≥ 10 pg/ml.

Table 3 Comparison between the studied groups according to basal hormonal profile
Full size table

Table 4 showed a statistically significant difference between basal LH in comparison to 30 min (min) and 45 min LH after stimulation (p1 < 0.001*, p2 < 0.001*). Moreover, LH levels after 30 and 45 min of GnRHST revealed a statistically significant difference (p3 = 0.016*).

Table 4 Comparison between the three studied periods in GnRH stimulation test in cases group
Full size table

In the scope of studying the factors that could affect basal LH and AMH levels in CPP girls, basal LH had a statistically significant correlation with the presence of adrenarche in addition to thelarche (p < 0.001*), height (p = 0.005*), bone age at diagnosis (p < 0.001*), and pubertal uterine dimensions (p = 0.007*). On the other hand, there was no statistically significant correlation between basal AMH and any of the parameters involved in assessment of CPP patients as shown in Tables 5 and 6.

Table 5 Correlation between basal LH and basal AMH according to different parameters in cases group (n = 26)
Full size table
Table 6 Comparison between both basal LH and basal AMH with uterine dimensions by ultrasonography in cases group (n = 26)
Full size table

ROC curve was performed to assess the diagnostic accuracy of basal LH as the initial screen in patients with early pubertal signs, while considering GnRHST as the gold standard test. Basal LH showed higher sensitivity 84.62% and specificity 76.92% and an area under the curve (AUC) of 0.840 (95% CI: 0.719–0.962, p < 0.001*). The best cut-off was chosen at a value of 0.29 IU/L based on the best Youden index (Table 7 and Fig. 1).

Table 7 Validity (AUC, sensitivity, specificity) for basal LH to discriminate CPP patients (n = 26) from control (n = 26)
Full size table
Fig. 1
figure 1

ROC curve for Basal LH to discriminate CPP patients (n = 26) from control (n = 26)

Full size image

Discussion

Recently, the incidence of CPP has increased worldwide. The diagnosis of CPP is based on clinical features, which are confirmed by hormonal and radiographic investigations [16]. However, this can be misleading, which makes it more challenging to diagnose [17]. CPP can result in various adverse effects including early menarche, premature epiphyseal closure leading to shorter target adult height, and psychological issues. Hence, it is essential to detect CPP early and develop simple screening methods to make easier decisions about starting treatment and improve the quality of life of patients [18].

The gold standard test for diagnosing CPP is intravenous GnRHST which confirms the premature activation of HPG axis [19]. A pubertal response is defined by an increase in LH concentration ≥ 5 IU/L which indicates CPP [15]. However, this test is time-consuming and expensive as it requires repeated blood sampling. Moreover, the clinical utility of the test in certain circumstances can be limited [11].

The availability of commercial intravenous synthetic GnRH is limited in some countries, including Egypt. Therefore, GnRH analogs such as triptorelin 0.1 mg injection are sometimes used [20]. The half-life of triptorelin is longer, and it has a stronger affinity for the GnRH receptor, raising uncertainties about the ideal sampling time of gonadotrophins after triptorelin stimulation [21]. There are also concerns regarding the lack of agreement on basal and stimulated LH and FSH cut-off values used to diagnose CPP [22]. Hence, a conclusive single diagnostic biomarker is needed to eliminate the need to perform GnRHST.

Previous studies have suggested basal serum LH as the most sensitive biomarker to differentiate between CPP and premature thelarche, rendering it suitable as a substitute for GnRHST [23, 24]. This is similar to our study that shows that there was a statistically significant difference in basal LH levels between CPP and control groups (p < 0.001*).

Moreover, there was a statistically significant positive correlation between basal LH and different clinical and radiological parameters such as Tanner staging (p < 0.001*), height SDS (p = 0.005*), bone age advancement (p < 0.001*), and uterine dimensions (p = 0.007*). Therefore, combining those parameters with high basal LH levels could aid in the diagnosis of CPP. Similarly, You et al. [20] developed a clinical score model for CPP among girls which consisted of four variables: age at start of puberty, basal LH level, largest ovarian volume, and uterine volume. This score may predict CPP and has been proposed to help in assessing the need for further GnRHST.

However, until now, there are no consistent cut-off values of basal LH as a diagnostic predictor of CPP. In different centers, it has varied from 0.1 to 1.5 IU/L with a sensitivity and specificity ranging between 60 and 100% [21, 23,24,25]. In the present study, the best basal LH cut-off was chosen at a value of 0.29 IU/L, with sensitivity 84.62% and specificity 76.92% (AUC 0.840). Therefore, we propose that basal LH > 0.29 IU/L could be used as a better discriminator for children with CPP.

Our results are consistent with the findings of Heo et al. [23] who studied the cut-off values in CPP and showed that levels > 0.245 IU/L had 88% sensitivity in confirming the diagnosis (p = 0.049*, AUC = 0.764). Durá-Travé et al. [26] results also showed higher sensitivity and specificity for basal LH levels (89 and 82% respectively) in 241 girls with breast enlargement between 6 and 8 years. Moreover, Lee et al. [27] studied 21 girls with CPP demonstrating a strong correlation between single basal unstimulated LH and GnRHST (r = 0.532, p < 0.001*) and proposed that basal LH levels ≥ 0.22 IU/L indicate CPP diagnosis, with a lower sensitivity compared to our results (69.4%) [28].

Additionally, Lee et al. [27] demonstrated that basal LH/FSH ratio was a significant predictor of positive response during GnRHST with significant differences between the 2 studied groups (p < 0.001*). Basal LH/FSH ratio has also been suggested to be more efficient in diagnosing children with pubertal changes with a cut-off value of 0.07 (low sensitivity and specificity of 73.2% and 89.5%) [28]. This was not observed in the present study as there was no statistically significant difference in basal LH/FSH ratio between girls with CPP and controls (p = 0.913). Perhaps, this could be attributed to the smaller sample size in our study (26 patients) in comparison to 121 CPP patients in the former study. Furthermore, combining basal LH level (cut-off 0.2) and basal LH/FSH ratio (cut-off 0.1) was proposed to be an easier method to diagnose CPP in 86 Thai girls between 6 and 8 years, with 71.4% sensitivity and 100% specificity [29].

In females, AMH is primarily produced by the ovarian granulosa cells and plays an important role in regulating sex hormone production in relation to onset of puberty. During childhood, AMH is at its lowest level. It increases gradually within 3 years before puberty, reaching its highest levels after puberty [30]. Despite demonstrating that basal LH can aid in the diagnosis of CPP in the present study, basal AMH failed to show a statistically significant difference between CPP girls and control group (median 1.3 vs 1.1 ng/ml, p = 0.07). This indicates that basal AMH could not be used as a marker in the initial screening of patients with early puberty.

Conversely, Muratoğlu Şahin et al. [30] investigated AMH in infants with premature thelarche (PT) and reported that AMH levels in PT were significantly lower than control group (1.66 vs 1.96 ng/ml, p = 0.025*), even though both were within normal range. These results implied that AMH may play an inhibitory role in breast development during infancy. Furthermore, Chen et al. [31] proposed that girls with progressive CPP had lower serum AMH in comparison to non-progressive CPP in a study enrolling 148 girls with early puberty. However, in the present study, the rate of progression of CPP was not assessed.

CPP is characterized by increased height velocity, advancement of bone age, and premature epiphyseal closure due to the effect of estradiol on bone growth. Therefore, it has been suggested that when BA is advanced, more than 2 SD in comparison to CA, premature thelarche is unlikely [26]. Similarly, BA was advanced in 17 CPP girls (65.4%) in our cohort indicating that BA could be frequently used as an additional clue, raising clinical suspicion of CPP. Nonetheless, we cannot rely on BA as a sole investigation to diagnose CPP, because only 63.2% of patients (12/19) had pubertal basal estradiol levels ≥ 10 pg/ml. These results are in accordance with the findings of Xu et al. [32].

Finally, the role of pelvic ultrasonography in CPP has been described in recent studies [33,34,35]. Normally, elevated gonadotrophin levels lead to enlargement of the ovaries and uterus in the later stages of puberty [35]. Pubertal uterine dimensions were observed in only half of the patients (13/26) in the present study. However, in the literature, there are contradictory results regarding the effectiveness of uterine dimensions in predicting CPP [33,34,35].

The present study has some limitations, including the small sample size which might have hindered finding a statistically significant difference between CPP and control groups. Lack of intravenous GnRHa due to unavailability made it difficult to compare peak LH, FSH, and LH/FSH ratio after GnRH stimulation with similar studies using intravenous route. Another limitation was the lack of a follow-up period to detect the course of CPP.

Conclusions

From the present study, we can conclude that basal LH levels at a cut-off point of ≥ 0.29 were superior to FSH and AMH as a suitable screening test to detect CPP, which could save time and expenses by potentially limiting the need to perform GnRHST. Furthermore, AMH cannot be used as screening test in patients with CPP. Yet, multi-center studies with larger sample size are needed to further evaluate the adequacy of basal LH in addition to other parameters in diagnosing CPP.