Introduction

Puberty is a period of marked change including development of secondary sexual characteristics, acceleration of linear growth, and development of reproductive capability [1,2,3,4,5]. Pubertal delay is a recognized complication in children with chronic kidney disease (CKD) [6]. In addition, puberty itself has been shown to be a risk factor for progression of CKD [7, 8].

Puberty is primarily controlled through the hypothalamic-pituitary–gonadal (HPG) axis. The hypothalamus initially secretes gonadotropin-releasing hormone (GnRH) in a pulsatile manner. In response to GnRH, the pituitary releases the gonadotropins luteinizing hormone (LH) and follicle stimulating hormone (FSH). LH acts on the theca cells of the ovary to form androgenic precursors of estradiol or the Leydig cells of the testes to secrete testosterone. FSH acts on the granulosa cells of the ovary or Sertoli cells of the testes to stimulate gametogenesis and gonadal growth, and in the ovary, FSH stimulates aromatase to form estradiol from thecal androgens [9]. The HPG axis remains quiescent throughout childhood due to suppression by the central nervous system. Puberty ultimately results from an increase in HPG activity with increased pulsatile secretion of GnRH as CNS inhibition declines [9]. Multiple factors likely contribute to the timing of puberty including genetics, environmental factors, and nutrition. Increase in LH is used to signal hormonal onset of puberty [10]. From a clinical standpoint, pubertal onset can be defined by thelarche in girls (Tanner breast stage 2) and testicular enlargement in boys (Tanner genital stage 2).

Previous studies among children with CKD have demonstrated that pubertal onset may be associated with more rapid progression of kidney disease [7, 8]. We previously demonstrated that there was a significant decline in glomerular filtration rate (GFR) after pubertal onset when defined by clinical features including Tanner stage, peak growth velocity, and menarche for girls [7]. However, increase in LH is a more objective marker of pubertal onset, and activation of the HPG-axis via pulsatile increase in various pubertal hormones occurs before certain physical manifestations of puberty such as growth spurt. In addition, children with CKD may have alterations in the HPG axis and impaired kidney metabolism of LH. Therefore, we aimed to (1) describe LH levels in a cohort of children with CKD progressing from pre-puberty to puberty, and (2) evaluate the association between hormonal onset of puberty and GFR decline among children with CKD.

Materials and methods

Study population

We used prospective data from the Chronic Kidney Disease in Children (CKiD) study. The CKiD study’s design and methods have been published previously [11]. A nested sample of participants age 6 through 16 years with at least three consecutive study visits was used in order to follow serial measurements of LH with the goal of detecting change in LH associated with pubertal onset. Baseline clinical and demographic data were obtained at study entry including body mass index (BMI), glomerular versus non-glomerular diagnosis, and urine protein-to-creatinine ratio.

Exposure and definition of pubertal onset

LH levels were measured at study entry and at yearly visits using the ultra-sensitive LH ELISA kit. Limit of detection for LH was 0.009 mIU/mL. The inter-assay coefficient of variance for controls averaged 3.84%. Across the samples, the inter-assay coefficient of variance for the replicates averaged 5.38%. Hormonal onset of puberty was defined as achieving an LH level greater than or equal to 0.3 IU/L based on definitions of puberty in the general population [12, 13].

Outcome

The primary outcome of interest was CKD progression as measured by GFR. GFR was estimated at annual study visits using the CKiD U25 (eGFR) equation based on serum creatinine and cystatin C. We also described changes in GFR measured by iohexol clearance (mGFR) which was conducted every other year. Direct measurement of iohexol is not dependent on serum creatinine which can be affected by muscle mass and pubertal development.

Statistical methods

A linear mixed-effects model with random intercepts and slopes with log-transformed GFR as the outcome was used to compare the average eGFR or mGFR percent change per year before and after hormonal onset of puberty. Models were adjusted for glomerular diagnosis, baseline proteinuria (log-transformed), and BMI category as a time-varying covariate. We fit additional models stratified by glomerular disease status. Analyses were conducted using Stata, version 16.1 [14]. This study was approved by the Institutional Review Board of Penn State College of Medicine.

Results

The cohort included 124 children (48 girls and 76 boys) (Table 1) contributing 585 observations. The median duration of follow-up for girls was 4 (IQR 3, 6) years and for boys was 5 (IQR 3, 6) years. Eight children entered following hormonal onset of puberty (i.e., prevalent puberty). There were 89 children with incident hormonal onset of puberty during study follow-up (31 girls and 58 boys) (Table 2). The median age at pubertal LH level for girls was 9.9 (8.8, 11.9) years and for boys was 10.2 (9.2, 11.0) years. Figure 1 displays the individual trajectories of LH; for those with incident hormonal onset of puberty, the enhanced boxplot depicts the distribution of ages at which onset was first observed. As expected, there is a general upward trend in absolute LH level over time after pubertal onset. The majority of children had a normal BMI and a non-glomerular disease diagnosis. The median eGFR and mGFR for those with incident hormonal onset of puberty was 47 ml/min|1.73 m2 and 43 ml/min|1.73 m2, respectively, at the time of pubertal onset.

Table 1 Clinical characteristics of 48 girls and 76 boys with measured luteinizing hormone at study entry. Median (IQR) or percent (n), or N
Full size table
Table 2 Clinical characteristics at the time of hormonal onset of puberty for those with incident puberty. Median (IQR) or percent (n), or N
Full size table
Fig. 1
figure 1

Luteinizing hormone levels for each subject with incident hormonal onset of puberty and enhanced boxplots describing age at hormonal onset of puberty (girls, n = 31—contributing 162 samples; boys, n = 58—contributing 314 samples). Percentile box plots depict the distributions of age at first luteinizing hormone ≥ 0.3 IU/L

Full size image

Table 3 presents results from mixed effects linear models. Faster decline in eGFR and mGFR after onset of hormonal puberty was observed for boys, and a similar but attenuated accelerated eGFR decline was observed for girls and no difference for mGFR. For boys, the percent change in eGFR was − 1.1% prior to hormonal pubertal onset and − 5.7% following hormonal pubertal onset (p < 0.001). Similarly, the percent change in mGFR was − 0.9% prior to hormonal pubertal onset and − 6.6% following onset (p < 0.001). For girls, the percent change in eGFR was − 3.5% prior to hormonal pubertal onset and − 5.0% following onset (p = 0.092). Using mGFR, the percent change in GFR was − 4.8% prior to hormonal pubertal onset and − 4.4% following hormonal pubertal onset (p = 0.84). Figure 2 displays the model-estimated change in eGFR for a child with non-glomerular disease, urine protein-creatinine ratio of 0.5 and normal BMI. In a secondary analysis, results stratified by glomerular disease status showed consistent results as the main analysis for the non-glomerular group. The glomerular diagnosis group did not have enough data for reliable inference (Table 4).

Table 3 Average glomerular filtration rate (GFR) percent change before and after hormonal onset of puberty for girls and boys from a mixed effects model adjusted for age, body mass index (BMI), glomerular disease diagnosis, and baseline proteinuria on a log scale
Full size table
Fig. 2
figure 2

Example of change in estimated glomerular filtration rate (eGFR) before and after hormonal onset of puberty from a mixed effects model for a hypothetical profile experiencing LH ≥ 0.3 IU/L at 9.9 years for girls and 10.2 year for boys with a normal BMI, non-glomerular disease diagnosis, and urine protein-to-creatinine ratio of 0.5

Full size image
Table 4 Average glomerular filtration rate (GFR) percent change before and after hormonal onset of puberty for girls and boys with non-glomerular disease from a mixed effects model adjusted for age, BMI, and baseline proteinuria on a log scale
Full size table

Discussion

In this longitudinal cohort study of children with CKD, we found that pubertal levels of LH occurred at a median age of 9.9 years among girls and 10.2 years among boys. Although we did not directly study children with CKD compared to healthy controls, the age at hormonal onset of puberty was generally similar as compared to previously reported literature from the general population at approximately 10 years of age for both boys and girls [10, 13, 14]. We found that hormonal onset of puberty was associated with a faster decline in GFR particularly among boys with CKD.

We previously demonstrated that GFR declined faster after clinical onset of puberty [7]. There are several proposed mechanisms to explain the accelerated GFR decline associated with puberty. One proposed mechanism for the loss of kidney function around the time of puberty includes hyperfiltration of remaining glomeruli as a result of acceleration in growth, body mass, and blood pressure. In the ItalKid study of children with CKD, a faster decline in eGFR was demonstrated after pubertal growth spurt. However, the authors concluded that increase in body mass was likely not the sole explanation for decline in GFR after pubertal onset given that eGFR continued to rapidly decline despite slowing or cessation of growth [8]. We previously showed that the decline in GFR after clinical onset of puberty remained significant after adjustment for BMI, and in this study, we show a faster decline in GFR after hormonal onset of puberty in boys which occurs prior to the pubertal growth spurt. For reference, peak growth velocity was shown to occur at age 11.9 years in boys and 11.1 years in girls which is later than hormonal onset of puberty (Fig. 2) [7]. Interestingly, the difference before and after hormonal onset of puberty in girls is attenuated as compared to our previous reports of more subjective clinical markers (Tanner stage, peak growth velocity, and menarche). This could be a reflection of fewer girls included in the study, but there is also literature that suggests there are sex differences in progression of CKD with males having been described as having faster disease progression [15, 16]. It is still possible the physical development that occurs after initial hormonal onset plays a larger role in disease progression rather than upregulation of LH levels directly or that other hormones upregulated secondary to LH and FSH such as testosterone and estradiol to higher levels later in puberty play a stronger role than LH itself.

Numerous animal models have shown a negative effect of androgens on kidney disease progression. Androgens have been implicated in the development of proteinuria, glomerulosclerosis, and tubulointerstitial fibrosis among males [17,18,19,20,21]. Estrogen may also contribute to CKD progression among females. In a study of over 4000 women, pre-menopausal women on oral contraceptives and post-menopausal women on hormone replacement therapy had an increased risk of microalbuminuria [22]. In a study of over 5000 post-menopausal women, hormone replacement therapy was associated with a significant decrease in eGFR. A higher cumulative dose of estrogen was associated with a greater decline in eGFR [23]. The association of sex hormones and kidney disease progression has not been studied in the pediatric CKD population. We hypothesize that sex hormones could contribute to the acceleration in GFR decline after pubertal onset, but do not minimize the role that physical growth itself may play. This needs to be evaluated further in future studies. In addition, changes in medication compliance and self-care during adolescence may contribute to the decline in kidney function with puberty [24].

This study has several limitations. Although we utilized data from the large CKiD study, our sample size was small as we restricted the cohort to children who were likely to be in the pubertal period and had three consecutive visits. The models indicated accelerated GFR decline after puberty, but it is not known whether the acceleration persists or the GFR decline stabilizes after completion of puberty. Further work could assess long-term GFR changes, but would need to address the competing risk problem of dialysis or transplant. Second, LH is released in a pulsatile fashion, and the level can vary throughout the day. As our samples were obtained from stored serum that were obtained at various times of day there may be variability in LH levels, and these results may not be peak levels. It is also possible that the LH level used to define hormonal onset of puberty in children with CKD may need to be different than the definition used for the general population. LH is thought to be primarily cleared by the kidney where it is endocytosed by tubular cells and catabolized. Therefore, patients with CKD are thought to be at increased risk for elevated LH levels as a result of reduced clearance by the kidney. When synthetic LH-releasing hormone was administered to adult patients on hemodialysis, prolonged plasma elevations of both LH and FSH were seen, and children with CKD and on dialysis have been shown to have longer half-life of several isoforms of LH [25,26,27,28]. Further study is needed to optimize the definition of LH level indicative of hormonal onset of puberty in CKD. Finally, the absence of testosterone and estradiol levels limits our ability to more definitively comment on the impact of sex hormones on kidney disease progression.

In conclusion, hormonal onset of puberty was associated with faster decline in GFR, particularly among boys with CKD. Clinicians should be aware that puberty is a high-risk period for kidney disease progression. Future studies should explore the role of sex hormones on CKD progression in children and interventions to mitigate the risk of progressive disease.