Efficacy and mechanism of nourishing yin and purging fire therapy for central precocious puberty based on meta-analysis and network pharmacology

Abstract

Background:

Central precocious puberty (CPP) is due to the early activation of the hypothalamus-pituitary-gonadal axis, and its incidence is on the rise. A number of studies have shown that nourishing yin and purging fire (NYPF) therapy can be beneficial for CPP. Therefore, we conducted this review to investigate the efficacy, safety, and mechanism of NYPF therapy for CPP.

Methods:

Electronic databases including PubMed, the Cochrane Library, Web of Science, EMBASE, China National Knowledge Infrastructure, Chinese Biomedical Literature Database, Wan-fang Database, and China Scientific Technical Journals Database and 2 platforms including Clinical Trials and Chinese Clinical Trial Registry were searched for randomized controlled trials of NYPF therapy for CPP. A meta-analysis was conducted using RevMan 5.3 and Stata 17.0 software. The core herb pair of NYPF was identified by data mining using IBM SPSS Modeler 18.0 software. The active ingredients and targets of the core herb pair were obtained through the TCMSP database. The main targets of CPP were acquired form the GeneCards, Disgenet and TTD databases. A protein-protein interaction network was carried out to select the core genes by using STRING platform and Cytoscape 3.7.2 software. Metascape platform was used to conduct gene ontology (GO) and KEGG enrichment analysis. The results were verified utilizing molecular docking.

Results:

A total of 23 studies were included. Meta-analysis shows the NYPF therapy could significantly improve the clinical efficacy rate and secondary sexual indicators (uterine volume, ovarian volume, breast nucleus diameter, follicular diameter), reduce TCM syndrome scores and serum sex hormone (FSH, LH, E2), and slow down bone age maturation compared to GnRHa therapy group. In addition, NYPF therapy was safe and has no obvious adverse events. Data mining revealed that the core herb pair of NYPF was “Anemarrhenae Rhizoma (Zhimu) - Phellodendri Chinensis Cortex (Huangbai).” Network pharmacology predicted that quercetin, kaempferol, beta-sitosterol, etc were the key components of Zhimu-Huangbai for treating CPP. The core targets were TP53, JUN, AKT1, ESR1, TNF, IL6, CCND1, MAPK1, BCL2, EGFR, IL1B, and PTGS2. They played a pivotal role in modulating multiple signaling pathways, such as Endocrine resistance, MAPK signaling pathway, and PI3K-Akt signaling pathway.

Conclusion:

This article revealed that NYPF therapy is effective and safe against CPP. The mechanism of the core herb pair of NYPF therapy for CPP through muti-components, muti-targets and muti-pathways.

1. Introduction

Precocious puberty is a common pediatric endocrine disease characterized by the appearance of secondary sexual characteristics in before age 8 in girls and before age 9 in boys.^[1] With changes in lifestyle, dietary structure and social environment, the rising incidence of CPP has become a focus of social concern.^[2–4] The most common form of precocious puberty is central precocious puberty (CPP). The pathogenesis of CPP is caused by the early activation of the hypothalamus-pituitary-gonadal axis, the increase of pulsatile secretion of gonadotropin releasing hormone (GnRH), and the early secretion of sex hormones, which leads to premature closure of the epiphysis.^[5,6] Children will have short stature as adults, and it also increases the incidence of diabetes, cancer and cardiovascular disease, etc^[7–9] seriously impacting their physical and mental health as well as their long-term social development. Currently, gonadotropin-releasing hormone analogues (GnRHa) is the primary treatment for children with CPP. However, its effect on improving the final adult height of children with advanced bone age is limited.^[10] Meanwhile, the long-term impact of GnRHa on children reproductive function, ovarian reserve function, body weight gain and lipid metabolism remains to be further observed.^[1,11,12]

Traditional Chinese medicine (TCM) is widely used to treat precocious puberty in clinic. According to the theory of TCM, precocious puberty is caused by insufficient kidney yin, which effects liver yin, resulting in an imbalance of kidney yin and yang, excessive phase fire, and the early onset of “Tiangui.” The predominant clinical syndrome is yin deficiency and excessive fire and the treatment principle is nourishing yin and purging fire (NYPF).^[13] In recent years, the number of clinical and animal studies of NYPF therapy for CPP has increased.^[14–17] However, a systematic review and meta-analysis on the efficacy and safety of these studies are still lacking and the molecular mechanism of NYPF therapy for CPP is also unclear.

Therefore, we conducted a systematic analysis and meta-analysis to evaluate the efficacy and safety of NYPF therapy for CPP and explored the mechanism of NYRF therapy for CPP by combining data mining and network pharmacology.

2. Materials and methods

2.1. Meta analysis

This study protocol was registered in PROSPERO (CRD42023433899). It was completed in accordance with the Preferred Reporting Items for Systematic Review and Meta-Analysis (PRISMA) guideline.^[18]

2.1.1. Inclusion and exclusion criteria for this study.

2.1.1.1. Inclusion criteria for this study.

(1) Types of studies

All RCTs with NYRF therapy for CPP published in Chinese and English were included.

(2) Types of participants

Participants who met the diagnostic criteria for CPP. The diagnostic criteria refers to the diagnostic criteria of expert consensus on the diagnosis and treatment of CPP (2022)^[6] and its past versions. The sex, race, nationality, educational background and medical units were not restricted.

(3) Types of intervention

The intervention group was treated with NYPF therapy alone or combined with GnRHa.

(4) Types of comparison

The control group was treated with GnRHa.

(5) Outcomes

Primary outcomes: Clinical efficacy rate, TCM syndrome scores.

Secondary outcomes: Secondary sexual characteristics [uterine volume (UV), ovarian volume (OV), breast nucleus diameter (BND), follicular diameter (FD)], Serum sex hormone (FSH, LH, E2), Bone age index (BAI), Adverse events (AEs).

2.1.1.2. Exclusion criteria for this study.

• ①Reviews, comments, case reports, animal or cell line experiments.
• ②Studies with incomplete data or duplicate publications.
• ③CPP caused by organic lesions in the central nervous system.
• ④Peripheral precocious puberty including gonadal tumors, adrenal disorders and exogenous intake of sex hormone containing drugs or foods.

2.1.2. Data search strategy.

Electronic databases including PubMed, the Cochrane Library, Web of Science, EMBASE, China National Knowledge Infrastructure (CNKI), Chinese Biomedical Literature Database (CBM), Wan-fang Database, and China Scientific and Technical Journals Database (VIP) were searched from inception to July 2023. In addition, we searched for other sources from Clinical Trials and Chinese Clinical Trial Registry.

The retrieval strategy in PubMed as an example was presented in Table 1. Similar search strategies were applied to other electronic databases.

Table 1.

Retrieval strategy of PubMed.

Number	Search terms
#1	(((((((((((((((((((((((Precocious Puberties[Title/Abstract]) OR (Puberties, Precocious[Title/Abstract])) OR (Pubertas Praecox[Title/Abstract])) OR (Praecox, Pubertas[Title/Abstract])) OR (Precocious Puberty[Title/Abstract])) OR (Precocious Puberty, Central[Title/Abstract])) OR (Central Precocious Puberties[Title/Abstract])) OR (Central Precocious Puberty[Title/Abstract])) OR (Precocious Puberties, Central[Title/Abstract])) OR (Puberties, Central Precocious[Title/Abstract])) OR (Puberty, Central Precocious[Title/Abstract])) OR (Sexual Precocity[Title/Abstract])) OR (Precocities, Sexual[Title/Abstract])) OR (Precocity, Sexual[Title/Abstract])) OR (Sexual Precocities[Title/Abstract])) OR (Idiopathic Sexual Precocity[Title/Abstract])) OR (Idiopathic Sexual Precocities[Title/Abstract])) OR (Precocities, Idiopathic Sexual[Title/Abstract])) OR (Precocity, Idiopathic Sexual[Title/Abstract])) OR (Sexual Precocities, Idiopathic[Title/Abstract])) OR (Sexual Precocity, Idiopathic[Title/Abstract])) OR (Early puberty[Title/Abstract])
#2	(((((((((((((((((Chinese[Title/Abstract]) OR (Chinese drugs[Title/Abstract])) OR (Chinese herbal medicine[Title/Abstract])) OR (Chinese herbal preparations[Title/Abstract])) OR (Chinese herbal remedy[Title/Abstract])) OR (Chinese traditional medicine[Title/Abstract])) OR (traditional[Title/Abstract])) OR (traditional Chinese medicine[Title/Abstract])) OR (traditional medicine[Title/Abstract])) OR (herbal remedy[Title/Abstract])) OR (herbal preparations[Title/Abstract])) OR (TCM[Title/Abstract])) OR (herbal supplements[Title/Abstract])) OR (herbal supplement[Title/Abstract])) OR (Nourishing Yin purging fire[Title/Abstract])) OR (Nourishing Yin[Title/Abstract])) OR (Purging Fire[Title/Abstract])
#3	randomized controlled trial[Publication Type] OR randomized[Title/Abstract] OR placebo[Title/Abstract]
#4	#1AND #2AND #3

2.1.3. Study selection and data extraction.

All the studies searched from the database were input into Endnote X9.3, where they were deduplicated. Two reviewers independently chose related studies after reading the titles, abstracts, and keywords. The remaining documents were screened by reading the full text. All differences between the 2 reviewers were discussed and resolved by a third reviewer.

The data extraction included broad details about the included studies, like the authors and the year of publication, as well as fundamental details of the studies, such as sample size, mean age, intervention measures, intervention time and outcomes.

2.1.4. Assessment of the risk of bias.

The risk of bias in eligible studies were evaluated based on the Cochrane Risk of Bias Assessment Tool. The items included random sequence generation, allocation concealment, blinding methods, incomplete data, selective outcome reporting, and other biases. Each subject was categorized into 3 levels “low bias,” “unclear bias,” “high bias.” Any disagreements between the 2 researchers were discussed and hold counsel with a third reviewer.

2.1.5. Statistical analysis.

Revman5.3 software and Stata 17.0 software were used for meta-analysis. Relative risk (RR) with 95% confidence interval (95% CI) were used for dichotomous variables, and the standardized mean difference (SMD) were used for continuous data. When the heterogeneity was low (I² ≤ 50%), a fixed-effect model was applied. Otherwise, a random-effect model was used when the heterogeneity was high (I² > 50%). Subgroup analysis or sensitivity analysis was used for high heterogeneity. Egger test was applied to assess reporting bias.

2.2. Data mining

We chose the studies with comprehensive TCM prescriptions. Using Excel 2016, 2 researchers extracted the prescriptions independently. The names of TCM were standardized according to the Pharmacopoeia of the People Republic of China (2020 edition) and Chinese Materia Medica. The core herb pairs were identified using the Apriori algorithm in IBM SPSS Modeler 18.0 software setting up the minimum rule confidence of 80%, the minimum support of 10%, and the maximum number of antecedents of 2. Core herb pair was used for network pharmacology research.

2.3. Network pharmacology research

2.3.1. Screening the targets of core herb group and CPP.

The active ingredients of the herb pair were found through the TCMSP (http://tcmspw.com/tcmsp.php). The parameters were set as oral bioavailability (OB) ≥ 30% and drug-likeness (DL) ≥ 0.18. CPP was selected as the key word to search the GeneCards (https://www.genecards.org/), Disgenet (https://www.disgenet.org/) and TTD databases (http://db.idrblab.net/ttd/) for the disease targets after merging and deleting duplication. Venny 2.1.0 online software (http://www.liuxiaoyuyuan.cn/) was used for screening common targets between herb and disease.

2.3.2. Protein-protein interaction (PPI) network construction and core targets selected.

Common targets were imported into the STRING 11.5 website (https://stringdb.org/) for PPI network construction. The data of this network was imported into Cytoscape 3.7.2 software for topological analysis. The core targets were those that satisfied the medians of degree, betweenness centrality, and closeness centrality of topological eigenvalue nodes. The results were visualized using Cytoscape 3.7.2.

2.3.3. Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analyses.

GO and KEGG enrichment analysis were performed depending on metascape platform (https://metascape.org/gp/index.html). Network visualization was used by bioinformatics online software (https://www.bioinformatics.com.cn/).

2.4. Molecular docking

The core targets and key active ingredients were used for molecular docking. Download the data of core target protein structure from RSCBPDB database (https://www.rcsb.org/), as well as the data of key active ingredient structure from Pubchem database (https://pubchem.ncbi.nlm.nih.gov/) with small resolution and method as the preferred X-ray conditions. AutoDock vina was applied to store protein and ligand in PDBQT format for docking.

3. Results

3.1. Results of meta-analysis

3.1.1. Literature screening results and characteristics of the study included.

A total of 742 studies were retrieved. After removing duplicate studies by Endnote X9.3, 490 studies remained. After reviewing the titles and abstracts, 84 studies were included. At last, 23 studies^[19–41] were collected after reading the full text. The detailed selection process was shown in Figure 1. All the studies were from China. The studies were published from 2001 to 2023. There were a total of 2036 patients, with 1024 assigned to the NYPF therapy group and 1012 to the GnRHa therapy group. The duration of the intervention ranged from 3 to 18 months. The basic characteristics of 23 studies included were shown in Table 2.

Figure 1.

Flow chart of the literature selection process.

Table 2.

Characteristics of the included studies.

Study	Sample size	Mean age (yr)		Interventions		Treatment	Outcomes
	(T/C)	T	C	T	C	(mo)
Chen and Wang (2010)[19]	100 (50/50)	6.86 ± 0.78	6.72 ± 0.83	Zaoshu 3 formula	Leuprorelin	6	①②③④⑤⑥
Chou and Wu (2020)[29]	104 (52/52)	7.08 ± 0.42	7.19 ± 0.48	Con + Zhibai dihuang decoction	Leuprorelin	12	①③④⑥
Huang et al (2022)[38]	62 (31/31)	6.85 ± 0.29	6.37 ± 0.28	Con + Zhibai dihuang decoction	Leuprorelin	3	①③④
Shen YX (2018)[26]	88 (44/44)	7.64 ± 2.68	7.51 ± 2.64	Con + Zhibai dihuang formula	Triptorelin	6	①③④⑤⑥
Wu et al (2023)[41]	106 (53/53)	8.18 ± 0.68	8.26 ± 0.57	Con + Zhibai dihuang pill	Triptorelin	6	①②③④⑤⑥
Liu and Wu (2018)[25]	60 (30/30)	5.86 ± 1.64	5.23 ± 1.35	Dabuyin pill + Zhibai dihuang pill	Triptorelin	6	④⑥
Wang and Feng (2020)[31]	85 (43/42)	8.4 ± 1.1	8.5 ± 1.2	Con + Zhibai dihuang pill	Triptorelin	6	④⑤
Zhao et al (2021)[34]	98 (49/49)	7.15 ± 1.08	7.15 ± 1.29	Con + Zhibai dihuang pill	Triptorelin	6	①③④⑤
Huai et al (2021)[32]	100 (50/50)	7.5 ± 1.4	7.2 ± 1.1	Con + Zhibai dihuang decoction	Triptorelin	6	①③④⑤⑥
Shen J (2022)[39]	142 (71/71)	8.21 ± 1.19	8.19 ± 1.22	Con + TCM	Leuprorelin	6	①③④⑥
Lu JW (2023)[40]	85 (43/42)	7.04 ± 1.24	6.84 ± 1.17	Con + Ziyin xiehuo formula	Leuprorelin	3	①③④⑥
Gan DM (2016)[21]	83 (45/38)	8.11 ± 1.97	8.02 ± 2.01	Con + Dabuyin pill	Triptorelin	12	①③④⑤⑥
Zhao et al (2021)[35]	80 (40/40)	6.41 ± 0.40	6.38 ± 0.36	Con + Dabuyin pill	Triptorelin	3	①③④⑥
Liu and Jin (2020)[30]	80 (40/40)	8.12 ± 1.03	8.03 ± 1.12	Con + Dabuyin pill + Zhibai dihuang pill	Triptorelin	3	①③④
Zhang (2016)[23]	70 (35/35)	6.78 ± 0.12	6.79 ± 0.11	Con + Dabuyin pill	Triptorelin	6	①③④⑤⑥
Lin et al (2017)[24]	156 (80/76)	8.28 ± 2.93	8.34 ± 3.16	Jinjia sanjie pill	GnRHa	6	①④
Xu HF (2015)[20]	140 (70/70)	7.32 ± 1.28	7.40 ± 1.31	Con + Zaoshu formula	Triptorelin	12	①③④
Gong HT (2022)[37]	81 (40/41)	6.44 ± 1.29	6.51 ± 1.37	Con + Zhibai jianghuo decoction	Leuprorelin	4	①③④⑥
Li YY (2021)[33]	60 (30/30)	6.31 ± 1.20	6.35 ± 1.22	Zhibai jianghuo decoction	Triptorelin	6	③④
Ying and Wang (2019)[28]	80 (40/40)	7.01 ± 0.48	7.17 ± 0.69	Zhibai jianghuo decoction	Triptorelin	6	①③④⑤⑥
Qi et al (2016)[22]	72 (36/36)	6.80 ± 0.32	7.10 ± 0.45	Con + Dabuyin pill	Triptorelin	18	①③④⑤⑥
Jiang BC (2022)[36]	40 (20/20)	8.05 ± 0.56	8.37 ± 0.39	Zhibai dihuang decoction	Leuprorelin	3	①②③④⑥
Chen YZ (2018)[27]	64 (32/32)	7.13 ± 0.78	7.24 ± 0.66	Ziyin xiehuo formula	Leuprorelin	6	①③④⑤⑥

C = control group, Con = control group, GnRHa = gonadotropin-releasing hormone analogues, T = treatment group, TCM = traditional Chinese medicine; ① clinical efficacy rate; ② TCM syndrome scores; ③ secondary sexual characteristics; ④ serum sex hormone; ⑤ bone age index; ⑥ adverse events.

3.1.2. Risk of bias.

The results of risk of bias were presented in Figure 2. 17 studies^{[22,24–26,28–39,41]} described randomization methods, such as the use of random number tables, which was considered “low risk”; 5 studies^{[19–21,23,27]} only mentioned randomization and did not specify the randomization method, labeled as “unclear risk”; One study^[40] applied parity allocation method, labeled as “high risk.” One study^[27] reported on the shedding of cases, However, the reason and time of the loss were not reported, labeled as “unclear risk.” None of the studies mentioned allocation concealment or blind method. Other biases were not identified in any of the included studies.

Figure 2.

Risk of bias of included studies. (A) Risk of bias graph; (B) Risk of bias summary.

3.1.3. Clinical efficacy rate.

20 studies reported the clinical efficacy rate. There were 919 patients in the experimental group and 910 patients in the control group. After testing, the results demonstrated that the clinical efficacy rate of the NYPF therapy group in treating CPP was superior to the GnRHa therapy group, with a statistically significant difference (RR = 1.15, 95% CI [1.11, 1.20], I² = 39%, P < .00001), as shown in Figure 3.

Figure 3.

Forest plot of clinical efficacy rate.

3.1.4. TCM syndrome scores.

3 studies reported the TCM syndrome scores. There were 123 patients in the experimental group and 123 patients in the control group. Meta-analysis showed that compared with GnRHa therapy group, NYPF therapy group could significantly improve TCM syndrome scores (SMD = −1.38, 95% CI [−1.96, −0.80], I² = 74%, P < .00001), as shown in Figure 4.

Figure 4.

Forest plot of TCM syndrome scores. TCM = traditional Chinese medicine.

3.1.5. Secondary sexual indicators.

20 studies reported the secondary sexual indicators. There were 869 patients in the experimental group and 864 patients in the control group. Meta-analysis showed that compared with control group, experimental group could significantly improve UV (SMD = −0.80, 95% CI [−1.21, −0.38], I² = 94%, P = .0002), OV (SMD = −0.63, 95% CI [−0.97, −0.29], I² = 91%, P = .0003), BND (SMD = −1.46, 95% CI [−1.93, −0.99], I² = 85%, P < .00001), FD (SMD = −0.92, 95% CI [−1.56, −0.28], I² = 94%, P = .005). Due to the high heterogeneity, a subgroup analysis and sensitivity analysis were conducted. Subgroup analysis was conducted according to the presence or absence of joint GnRHa in experimental group. The subgroup analysis revealed that NYRH combined with GnRHa could significantly reduce UV, OV and FD compared with GnRHa alone (UV: SMD = −1.12, 95% CI [−1.53, −0.72], I² = 92%, P < .00001; OV: SMD = −0.88, 95% CI [−1.25, −0.51], I² = 91%, P < .00001; FD: SMD = −1.24, 95% CI [−1.94, −0.54], I² = 94%, P = .0005). There was no statistical difference between NYPF alone group and GnRHa group in OV (SMD = 0.14, 95% CI [−0.19, 0.47], I² = 57%, P = .41) and FD (SMD = 0.21, 95% CI [−0.13, 0.54], I² = 0%, P = .23). Meanwhile, GnRHa group could significantly reduce UV compared with NYPF alone group (SMD = 0.45, 95% CI [0.10, 0.81], I² = 58%, P = .01), as shown in Figure 5. Sensitivity analysis was performed using the method of excluding studies one by one and the result did not affect the results, indicating that the results of the meta-analysis were reliable, as shown in Figure 6.

Figure 5.

Forest plot of secondary sexual indicators. (A) Uterine volume; (B) Ovarian volume; (C) Breast nucleus diameter; (D) Follicular diameter.

Figure 6.

Sensitivity analysis of secondary sexual indicators. (A) Uterine volume; (B) Ovarian volume; (C) Breast nucleus diameter; (D) Follicular diameter.

3.1.6. Serum sex hormone.

22 trails reported serum sex hormone. There were 978 patients in the experimental group and 971 patients in the control group. Meta-analysis showed that compared with control group, experimental group could significantly improve FSH (SMD = −1.56, 95% CI [−2.06, −1.06], I² = 96%, P < .00001), LH (SMD = −1.52, 95% CI [−2.07, −0.98], I² = 96%, P < .00001) and E₂ (SMD = −1.09, 95% CI [−1.46, −0.72], I² = 93%, P < .00001). A subgroup analysis and sensitivity analysis were conducted because of the high heterogeneity. The subgroup analysis revealed that NYPF combined with GnRHa could significantly reduce FSH, LH and E₂ compared with GnRHa alone (FSH: SMD = −2.00, 95% CI [−2.50, −1.50], I² = 93%, P < .00001; LH: SMD = −2.14, 95% CI [−2.67, −1.62], I² = 94%, P < .00001; E₂: SMD = −1.44, 95% CI [−1.87, −1.01], I² = 92%, P < .00001). There was no statistical difference between NYPF alone and GnRHa group (FSH: SMD = −0.43, 95% CI [−1.45, 0.58], I² = 96%, P = .40; LH: SMD = −0.17, 95% CI [−0.90, 0.57], I² = 94%, P = .66; E₂: SMD = −0.38, 95% CI [−0.97, 0.22], I² = 91%, P = .21), as shown in Figure 7. Figure 8 demonstrates that a sensitivity analysis disclosed the results of the meta-analysis to be reliable.

Figure 7.

Forest plot of serum sex hormone. (A) FSH; (B) LH; (C) E2.

Figure 8.

Sensitivity analysis of serum sex hormone. (A) FSH; (B) LH; (C) E2.

3.1.7. Bone age index.

11 studies reported the BAI. There were 502 patients in the experimental group and 497 patients in the control group. Meta-analysis showed that compared with control group, experimental group could significantly improve BAI (SMD = −0.63, 95% CI [−1.13, −0.14], I² = 93%, P = .01). The subgroup analysis revealed that NYPF combined with GnRHa could significantly reduce BAI (SMD = −0.85, 95% CI [−1.50, −0.21], I² = 94%, P = .010). As shown in Figure 9A, there was no statistical difference between NYPF alone and GnRHa group (SMD = −0.08, 95% CI [−0.34, 0.17], I² = 0%, P = .52). Sensitivity analysis revealed that the results of the meta-analysis were reliable, as shown in Figure 9B.

Figure 9.

Forest plot and Sensitivity analysis of Bone age index. (A) Forest plot of Bone age index. (B) Sensitivity analysis of Bone age index.

3.1.8. Adverse events.

A total of 16 studies mentioned adverse events and 10 studies reported the AEs. The results indicated that the AEs in the experimental group was no statistical difference than control group (RR = 0.62, 95% CI [0.36, 1.08], I² = 0%, P = .09). Nausea and vomiting were the most prevalent adverse effects reported. The incidence of nausea and vomiting in the experimental group did not differ significantly from the control group (RR = 1.14, 95% CI [0.43, 2.99], I² = 0%, P = .79), according to a meta-analysis of 5 studies. 3 studies mentioned the adverse event of pain at the injection site. There was no statistical difference between the 2 groups (RR = 0.99, 95% CI [0.34, 2.87], I² = 0%, P = .99). 2 studies reported the adverse events of rash, pubic hair growth, and transient vaginal bleeding. NYRF intervention can reduce the occurrence of transient vaginal bleeding (RR = 0.11, 95% CI [0.01, 0.86], I² = 0%, P = .04), but no obvious effect on decreasing the incidence of rash and public hair development (Rash: RR = 0.50, 95% CI [0.09, 2.65], I² = 0%, P = .42; Public hair development: RR = 0.34, 95% CI [0.05, 2.08], I² = 0%, P = .24), as shown in Figure 10.

Figure 10.

Forest plot of adverse events.

3.1.9. Publication bias.

Publication bias was applied to the clinical efficacy rate, FSH, LH, E2, UV, OV, and BAI. The Egger test suggested the possibility of publication bias in FSH, LH, E₂ (FSH: Egger test: P = .014; LH: Egger test: P = .009; E2: Egger test: P = .025). According to Table 3, the clinical efficacy rate, UV, OV, and BAI did not exhibit publication bias (Clinical efficacy rate: Egger test: P = .994; UV: Egger test: P = .051, OV: Egger test: P = .237; BAI: Egger test: P = .367).

Table 3.

The results of Egger test.

Outcomes	Number of trials	P value
Clinical efficacy rate	20	.994
Secondary sexual indicators
UV	19	.051
OV	20	.237
Serum sex hormone levels
FSH	21	.014
LH	22	.009
E2	21	.025
Bone age index	11	.367

OV = ovarian volume, UV = uterine volume.

3.2. Data mining results

A total of 14 prescriptions were included. There were 44 kinds of herbs in total. The most frequently herbs were Phellodendri Chinensis Cortex (Huangbai) and Anemarrhenae Rhizoma (Zhimu). SPSS Modeler 18 was utilized to analyze the association rules of prescriptions. Combined with the ranking of support degree and confidence, core herb pairs were Anemarrhenae Rhizoma (Zhimu) - Phellodendri Chinensis Cortex (Huangbai), Anemarrhenae Rhizoma (Zhimu) - Moutan Cortex (Mudanpi), Phellodendri Chinensis Cortex (Huangbai)- Moutan Cortex (Mudanpi), as shown in Table 4. “ Zhimu- Huangbai “ with the highest support and confidence was chosen as the core herb pair for network pharmacology research.

Table 4.

Core herbal pairs of NYPR therapy for CPP.

Core herbal pairs	Support (%)	Confidence (%)
Anemarrhenae Rhizoma (Zhimu) - Phellodendri Chinensis Cortex (Huangbai)	85.714	100
Anemarrhenae Rhizoma (Zhimu) - Moutan Cortex (Mudanpi)	78.571	100
Phellodendri Chinensis Cortex (Huangbai)- Moutan Cortex (Mudanpi)	78.571	100

CPP = central precocious puberty.

3.3. Network pharmacology

3.3.1. Active targets of “Zhimu- Huangbai” and CPP.

A total of 36 active ingredients and 235 targets of “Zhimu- Huangbai” were collected. 2160 targets of CPP were obtained. 110 overlapping targets were obtained by venny2.1.0, as shown in Figure 11A. Cytoscape3.7.2 was used to establish the “herb-component-targets-disease” network, as shown in Figure 11B. The network contained 235 nodes and 932 edges. The key active ingredients were screened by network analysis function and were listed according to the degree, as shown in Table 5.

Figure 11.

Network pharmacological analysis. (A) Venny of herb and disease. (B) Network of core herbal pair -active ingredients-targets-disease. CPP = central precocious puberty, HB = Phellodendri Chinensis Cortex (Huangbai), ZM = Anemarrhenae Rhizoma (Zhimu).

Table 5.

Key active ingredients of “Zhimu-Huangbai” against CPP.

Compound name	Degree
quercetin	155
kaempferol	64
beta-sitosterol	39
Anhydroicaritin	38
Isocorypalmine	37
Stigmasterol	33
Cavidine	29
Dehydrotanshinone II A	22
rutaecarpine	19
diosgenin	17
Hippeastrine	12
asperglaucide	6
Anemarsaponin F_qt	2

CPP = central precocious puberty.

3.3.2. PPI network construction.

Using the String database and selecting the highest confidence (0.900), a PPI network was constructed for the 110 herb-disease targets. There were 101 nodes and 306 edges in the network. CytoNCA was utilized to identify important targets. As shown in Figure 12, the primary targets were TP53, AKT1, JUN, ESR1, TNF, IL6, MAPK1, CCND1, BCL2, IL1B, EGFR, and PTGS2.

Figure 12.

PPI network and hub targets. PPI = protein-protein interaction.

3.3.3. Gene enrichment analysis.

1796 GO enrichment items were displayed. Among them, 1586 items in biological process (BP) were identified, such as response to hormone, response to steroid hormone, response to nutrient levels and response to growth factor. Cellular component (CC) was made up of 68 items, such as plasma membrane raft, transcription regulator complex, protein kinase complex and so on. Molecular function was consisted of 142 items, such as transcription factor binding, nuclear receptor activity, ligand-activated transcription factor activity, etc. As depicted in Figure 13A to C, the first 20 items in BP and 10 items in CC, molecular function were selected for visualization.

Figure 13.

Gene enrichment analysis. (A) Go biological process; (B) Go cellular components; (C) Go molecular function; (D) KEGG analysis. GO = gene ontology, KEGG = Kyoto encyclopedia of genes and genomes.

A total of 188 signal pathways were enriched after KEGG analysis and the first 20 items were selected for visualization, as shown in Figure 13D. Among them, the main items including Endocrine resistance, MAPK signaling pathway and PI3K-Akt signaling pathway, as shown in Figure 14A to C.

Figure 14.

Signaling pathway diagram for this study. (A) Endocrine resistance; (B) MAPK signaling pathway; (C) PI3K-Akt signaling pathway.

3.3.4. Molecular docking.

To validate the results of network pharmacology, molecular docking was performed between key ingredients (quercetin, kaempferol, beta-sitosterol) and hub targets (AKT1, ESR1, EGFR). The results showed that the docking scores were all ≤ −6 kcal/mol, showing good binding activity, as shown in Table 6.

Table 6.

Results of molecular docking.

Ligand	Receptor	Free energy (Kcal/mol)
quercetin	AKT1(1UNQ)	−6.3
quercetin	ESR1(7BAA)	−7.8
quercetin	EGFR(8A27)	−9.2
kaempferol	AKT1(1UNQ)	−6.0
kaempferol	ESR1(7BAA)	−7.9
kaempferol	EGFR(8A27)	−9.4
beta-sitosterol	AKT1(1UNQ)	−6.9
beta-sitosterol	ESR1(7BAA)	−8.2
beta-sitosterol	EGFR(8A27)	−9.5

4. Discussion

With changes in lifestyle, dietary structure and social environment, the rising incidence of CPP has become a focus of social concern. In China, TCM is widely used in the treatment of CPP. There is no “precocious puberty” or corresponding disease name in TCM. The fundamental pathogenesis of CPP in TCM is insufficient kidney yin, imbalance of yin and yang, deficiency of yin and excessive fire and reckless movement of phase fire, which results in the early arrival of “Tiangui.” The NYRF method is the fundamental treatment principle for CPP. Zhibai Dihuang Pill and Dabuyin Pill are commonly prescribed medications for treating CPP.^[42,43]

This is the first study to combine meta-analysis, data mining, and network pharmacology to investigate the efficacy and potential mechanisms of NYPF therapy for CPP. This study included 23 studies with 2036 patients in total. Compared to the GnRHa therapy group, NYPF therapy group could significantly improve the clinical efficacy rate and secondary sexual indicators (UV, OV, BND, FD), reduce TCM syndrome scores and serum sex hormone (FSH, LH, E2) and slow down bone age maturation. However, subgroup analysis found that there was no significant difference between the the NYPF alone group and GnRHa therapy group. Nausea and vomiting, pain at the injection site, rash, pubic hair growth, and transient vaginal bleeding were the main adverse events. The NYPF therapy group had a lower incidence of transient vaginal bleeding compared to the GnRHa therapy group. But there were no significant differences between the 2 groups in terms of nausea and vomiting, pain at the injection site, rash, and pubic hair development. The study revealed that NYPF therapy could reduce transient vaginal bleeding caused by GnRHa.

Using data mining, the core herb pairs in the NYPF cohort were analyzed. Finally, 14 prescriptions containing 44 botanicals were retrieved. Huangbai and Zhimu were the most frequently used herbs. In the meantime, association rule analysis revealed that “Zhimu- Huangbai” was the pair of essential herbs with the highest level of confidence and support. This combination of “Zhimu- Huangbai” originated from Zishen Pill in Li Dongyuan “Lanshi Secret Collection.” Zhimu possesses a bitter and sweet flavor, a cold nature, and is associated with the lung, stomach, and kidney meridians. It can clear heat, extinguish fire, nourish yin and moisten dryness. Huangbai has a bitter taste, a cold nature, and belongs to the kidney, bladder, and large intestine meridians. It can remove heat, dry dampness, and eliminate fire to strengthen yin. In TCM, the herb pair of “Zhimu - Huangbai” is commonly used to treat qi deficiency and excess fire-related diseases.^[44] Classic formulas such as Zhibai Dihuang pill and Dabuyin pill contain the herb pair of “Zhimu- Huangbai.” The mutual compatibility of “Zhimu- Huangbai” is a “Xiangxu” compatibility of TCM and enhances the efficacy of NYPF. Relevant studies have found that the herb pair of “Zhimu- Huangbai” has potential therapeutic effects on diabetes, hyperlipidemia, and inflammatory diseases.^[45–47]

The pathogenesis of CPP is still unclear. According to the majority of researchers, it is associated with the early activation of the hypothalamus-pituitary-gonadal axis and the release of sex hormones. Network pharmacology research found that quercetin, kaempferol, β-sitosterol, etc are the main ingredients of “Zhimu- Huangbai” to interfere with CPP. Quercetin, exhibiting various antioxidant and anti-inflammatory activities, is a flavonoid that is abundant in vegetables, fruits, herbs and teas.^[48] β-Sitosterol, a natural phytosterol from many vegetable oils, nuts, and plant medicines, exhibits anti-oxidative, anti-inflammatory and anticancer effects.^[49] Quercetin and β-sitosterol are the active ingredients of Huangbai. Related studies have found that quercetin can inhibit ovarian cell proliferation and the release of ovarian steroids and peptide hormones. Kaempferol is a nature compound widely found in many vegetables and herbs and displays anti-inflammatory, antioxidant, antitumor, cardioprotective, neuroprotective, and antidiabetic activities.^[50] Zhimu active constituent, kaempferol, has the effect of inhibiting E₂ secretion when E₂ levels rise in the body.^[51]

PPI network analysis found that TP53, JUN, AKT1, ESR1, TNF, IL6, CCND1, MAPK1, BCL2, EGFR, IL1B, and PTGS2 are the core targets of “Zhimu- Huangbai” to interfere with CPP. Estrogen plays an essential role in the development and function of reproductive physiology. The action of estrogen is mediated by estrogen receptors (ERs). ESR1 plays an important role in precocious puberty.^[52] Several studies have documented the growing significance of EGFR in bone development.^[53,54] GO enrichment analysis revealed that the core pair “Zhimu- Huangbai” might produce a therapeutic effect by response to hormone, response to steroid hormone, response to nutrient levels, response to growth factor and other BPs. KEGG enrichment analysis suggested that Endocrine resistance, MAPK signaling pathway, and PI3K-Akt signaling pathway were the main pathways for the core herb pair against CPP. Finally, the results of molecular docking also preliminarily verified that these key ingredients have good binding activities with hub targets.

This review has limitations: First, the quality of the included literatures was not high and the random methods used in some studies were not clear. The blinding and allocation concealment of most studies were unclear. The heterogeneity of some secondary outcome indicators in this study might be related to significant differences in baseline values of parameters between different studies, as well as differences in specific treatment plans, prescription composition, and dosage of interventions. Second, it lacked representativeness because the included literatures were all in Chinese and there was no English literature. It needs to be further verified by a randomized double-blind trial using a unified treatment plan, multi-center, large sample, and more rigorous research design. Thirdly, the majority of the research findings regarding the mechanism by which NYPF treats CPP were derived from network pharmacology and still required experimental confirmation.

5. Conclusion

In conclusion, this study showed that NYPF therapy is more effective and safer for CPP compared to GnRHa therapy. Future data mining and network pharmacology revealed that the core herb pair “Zhimu- Huangbai” exerted therapeutic effects through muti-components, muti-targets and muti-pathways. However, due to lack of high-quality studies, rigorously designed RCTs are required in the future. Further biological experiments in vivo and vitro are needed to validate the active ingredients, core targets and key pathways in order to clarify the mechanisms of these findings.

Author contributions

Conceptualization: Yuan Ma, Fengping Sun.

Data curation: Yuan Ma, Erbing Zhang.

Funding acquisition: Fengping Sun.

Investigation: Jing Li, Shangsai Yue.

Methodology: Yuan Ma, Yunyun Fu.

Supervision: Yunyun Fu, Suling Zhang.

Validation: Yuan Ma, Suling Zhang.

Writing – original draft: Yuan Ma.

Writing – review & editing: Yuan Ma, Fengping Sun.