Proposed age-stratified reference intervals of FSH derived from normozoospermic men

Emily-Jane Waller, Jason Conceicao, Phillip Matson?, John L Yovich,3

1School of Medical and Health Sciences, Edith Cowan University, Joondalup, Perth, Australia

2PIVET Medical Centre, Leederville, Perth, Australia 3School of Pharmacy and Biomedical Sciences, Faculty of Health Sciences, Curtin University, Bentley, Perth, Australia

ABSTRACT

KEYWORDS: Male infertility; FSH; Reference intervals; Age;Obstructive azoospermia; Non-obstructive azoospermia

1. Introduction

Infertility is a global health issue affecting an estimated 15% of couples[1], and males are found to be solely responsible for 20%-30% of infertility cases and contribute to 50% of cases overall[2].Semen analysis remains the main screening test for male fertility on the basis that it is the quality of the ejaculated semen that appears most relevant[3] but serum follicle stimulating hormone (FSH)has long been known to be an important biochemical marker of testicular function[4-6].

Different assay platforms measuring FSH can give significantly different results when analysing the same sample[7-9] but this is not always the case[10], and a far more pressing issue is the selection of an appropriate reference interval for clinical interpretation.Reference intervals commonly use the central 95% of a healthy population[11] but should take into account normal physiological changes such as age[12]. A number of studies have demonstrated that serum FSH in men rises with age[13-15] and yet many FSH reference intervals applied do not have age-related partitioning for men of reproductive age[8,9,16-19].

This study analysed retrospectively the serum FSH concentrations and semen quality at the initial consultation of men attending a clinic for the investigation of their fertility, and aimed to: 1) determine the effect of age upon serum FSH in men producing normozoospermic semen samples, 2) derive an age-based reference interval for these men, 3) apply these reference intervals to men producing oligozoospermic and azoospermic semen samples, and 4) validate further the age-based reference intervals on men with obstructive azoospermia and non-obstructive azoospermia.

2. Materials and methods

2.1. Study population

A total of 1 190 men were enrolled who were sexually mature and aged ＞20 years. Hence, changes around puberty were excluded as a pre-analytical issue. They attended for initial fertility assessment between January 2010 - July 2018 where semen collection and analysis were performed on the same date as blood collection.Semen samples were excluded if they 1) were not fresh semen (i.e.epididymal, testicular, and cryopreserved samples); 2) had a semen volume of ＜1.5 mL; 3) had recorded days of abstinence outside of 2-7 days; 4) were samples that were not initial assessments; 5) were not collected within the study centre (i.e. home collections); and 6)were produced by men taking exogenous hormone medication, or had HIV or any form of cancer. Men producing azoospermic semen samples were identified as having non-obstructive azoospermia by clinical diagnostic criteria including genetic analysis and surgical exploration.

2.2. Semen analyses

Methods for the analysis of semen[20] and lower reference limits for the interpretation of results[21] were those currently recommended by the World Health Organisation (WHO). Semen samples included in the study were classed as either normozoospermic with the three parameters of sperm concentration, motility and morphology all above the lower reference limits, or oligozoospermic (reduced sperm concentration but normal motility and morphology), and azoospermic (an absence of sperm in the semen). The laboratory participated in an external quality assurance scheme to ensure acceptable performance in terms of accuracy and precision for sperm concentration, sperm motility and sperm morphology (EQASRM,PO Box 162, Northlands, Western Australia 6905).

2.3. Blood collection and FSH measurement

All blood collections were done within the PIVET Medical Centre on the same day as semen sample collection and analysis. Bloods were non-fasting samples collected into plain tubes, allowed to clot and the serum were aspirated and frozen prior to analysis.The Siemens ADVIA CentaurXP immunoassay system was the platform used to analyse all blood samples in this study. Internal quality controls at three levels (Liquichek Immunoassay Plus Control, Trilevel, Bio-Rad) were run on each day the FSH analyses were made, and clinical samples were analysed only if the quality control samples were within defined ranges. Participation in an external quality assurance scheme (RIQAS-Randox Immunoassay Quality Assurance Scheme) also confirmed acceptable performance of the FSH assay. The assay laboratory is accredited on an annual basis by the National Australian Testing Authority.

2.4. FSH reference intervals

The FSH reference intervals were stratified according to age[12],namely 21-30, 31-40 and 41-50 years, and each age bracket had more than 120 samples[22]. The serum FSH concentrations for each age bracket were tested with the Kolmogorov-Smirnov test to determine if it matched the characteristics of a normal distribution and, if not, concentrations were transformed to logvalues and retested for normality. The reference intervals in the present study were determined by calculating the 95% centile range after logarithmic transformation as the upper and lower limits of two standard deviations for men producing normozoospermic semen samples and thereby having normal spermatogenesis.

2.5. Statistical analysis

The association between age and serum FSH concentration was analysed by using Spearman’sρ(rank correlation coefficient). The effect of age was assessed by using the Kruskal-Wallis test with post-hoc analysis by the Dunn method adjusted by the Benjamini-Hochberg procedure, whilst proportions were compared by using Chi square (χ) analysis with post-hoc analyses using the Bonferroni correction. Associations and differences were considered significant if P＜0.05.

2.6. Ethics approval, Assisted Reproductive Technology (ART)Unit accreditation, and patient consent

Ethics approval had been obtained to use data contained in the clinical database for retrospective research, and all patient information was de-identified to protect the identity of all subjects.Specifically, the reporting of this data was approved under Curtin University Human Ethics Committee approval No. RD_25-10 general approval for retrospective data analysis in 2010, updated in 2015, and again further updated in August 2020. The ART Unit was accredited by both the Reproductive Technology Accreditation Committee, a committee within the Fertility Society of Australia, as well as the Reproductive Technology Council of Western Australia.Consent forms permitting the retrospective use of data received approval under both regulatory bodies.

3. Results

3.1. Study population

Table 1 shows the age and semen parameters for the normozoospermic (n=1 037), oligozoospermic (n=94) and azoospermic (n=59) men in the study. There was a significant difference in age between normozoospermic, oligozoospermic and azoospermic groups (Kruskal-Wallis test; P＜0.05), with post-hoc testing confirming the oligozoospermic men were younger than their normozoospermic counterparts (P＜0.05). The azoospermic men were excluded from the comparisons of semen quality on the basis that their samples were, by definition, devoid of sperm. The normozoospermic men had significantly higher sperm concentrations in their semen than the oligozoospermic men (Kruskal-Wallis test;P＜0.01) although this was to be expected by definition according to the lower reference limit for sperm concentration set by the WHO[20].Whilst all oligozoospermic men had normal sperm morphology and progressive motility above the lower reference limit of the WHO 5th edition manual[20], their proportion of normal forms and sperm with progressive motility were nevertheless both lower than for the normozoospermic men (Kruskal-Wallis test; both P＜0.01).

3.2. Men producing normozoospermic samples

Figure 1 shows the scatterplot of serum FSH concentration for normozoospermic men against their age, and there was a significant positive association (Spearman’sρ= 0.17, P＜0.01). The men were then arranged into the three age groups as shown in Table 2, namely 21-30, 31-40, and 41-50 years. The FSH concentrations for each of these three groups were not normally distributed (Kolmogorov-Smirnov test; all P＜0.05 thus rejecting the null hypothesis that the values were sampled from a population that follows a Gaussian distribution). The FSH values were then transformed to logvalues and these were found to be normalized with each null hypothesis then being accepted. Further analysis on the normalized data (Table 2) showed that the serum FSH concentrations for men aged 21-30,31-40 and 41-50 years were different (Kruskal-Wallis test; P＜0.005),with post-hoc analysis confirming there were progressive and significant rises from men aged 21-30 years to 31-40 years to 41-50 years (all P＜0.05). Calculation of the upper and lower limits of 2 standard deviations of the geometric transformation gave the 95th centile range and enabled stratified reference intervals to be obtained for these three age brackets, namely 21-30 years [(1.21-9.26) IU/L], 31-40 years [(1.37-10.05) IU/L], and 41-50 years[(1.57-10.89) IU/L].

3.3. Qualitative measures of serum FSH according to semen quality

The stratified reference intervals and the proportion of normozoospermic, oligozoospermic and azoospermic men that were above or below these intervals are shown in Table 3. Just under 5% ofmen producing normozoospermic samples had FSH concentrations outside of the reference intervals, consistent with the intervals being derived from 95% of the men. The proportion of men with normal FSH concentrations in the three age groups was different overall(χtest; P＜0.05) and post hoc testing showed that the proportion of oligozoospermic men with normal FSH concentrations was less than the normozoospermic men, which in turn was less in the azoospermic men (χtest; both P＜0.01). The azoospermic men were further broken down according to the nature of the azoospermia as either obstructive azoospermia or non-obstructive azoospermia.A total of 86.4% (38/44) men with non-obstructive azoospermia had elevated serum FSH concentrations whereas only 6.7% (1/15)men with obstructive azoospermia had high FSH levels showing significant difference (χtest; P＜0.01).

Table 1. The population and semen quality of the 1 190 men in the study.

Table 2. The serum concentrations of FSH according to age in 10-year brackets for men producing a normozoospermic semen sample at their first assessment (IU/L).

Table 3. The identification of men with abnormal serum FSH concentrations according to their semen quality [n(%)].

Figure 1. The scattergram of serum FSH concentrations against age for the 1 037 normozoospermic men. A trendline shows the overall change in pattern,and there is a significant positive association between FSH and age (Spearman’sρ= 0.170, P＜0.01). FSH: follicle-stimulating hormone.

4. Discussion

Reference intervals are usually derived from a group of healthy individuals against which others can be deemed unhealthy if the analyte under investigation is outside of the reference interval[11].In this study dealing with the measurement of serum FSH as part of the investigation of male infertility, the reference group were sexually mature men producing normozoospermic semen samples,

with men who had possible semen abnormalities through disease or medication being excluded. The identification of such a reference group of men based upon the quality of their semen is not without difficulty as the performance of semen analysis itself has long been known through external quality assurance schemes to potentially have poor accuracy and precision[23], and more recent findings have shown some counting chambers to give erroneous results when counting human sperm[24]. Furthermore, the reference limits for semen analysis have changed over the years[25-27], confounding the definition of men whose semen can be categorized within normal limits. The laboratory in this study therefore used the methods[20]and lower reference limits[21] most recently recommended by the WHO for semen analysis and the interpretation of results. The laboratory also participated in an external quality assurance scheme to demonstrate acceptable performance in the measurement of sperm concentration, sperm motility and sperm morphology.

Age-related increases in serum FSH during reproductive life have been reported previously[13-15], although some reports only show an increase in serum FSH from the 5th[28] or 6th[29] decade of life. Nevertheless, the reality is that the vast majority of reference intervals available do not take age into account[8,16-19]. This should be seen as a major limitation in the use of FSH as a biomarker of testicular function, with a possible loss of sensitivity in the detection of men with testicular deficiency.

The measurement of FSH in men is a fundamental investigation since it is considered to be an indicator of germinal function[30], and hence elevated serum FSH concentrations can be seen in some men with seminal abnormalities such as azoospermia, oligozoospermia,and the presence of varicocoeles[31]. More specifically, a raised concentration of FSH is considered a reliable indicator of germinal epithelial damage and is usually associated with azoospermia or severe oligozoospermia[5,32,33]. The present study reported 13.8%of oligozoospermic men to have elevated FSH levels. This is particularly important as the oligozoospermic group will inevitably have a range of underlying pathologies and the measurement of FSH helps to identify the sub-group of men with a degree of germinal epithelial damage. In further examining the group of azoospermic men, non-obstructive azoospermia was an important sub-category as these men had no obstruction within the male reproductive tract and would therefore contain a high proportion of men with extremely low levels or absence of spermatogenesis[34]. The present study showed a high proportion of men with non-obstructive azoospermia to have high FSH concentrations outside of the age-related reference intervals when compared with men with obstructive azoospermia.Low serum FSH can be a good predictor of the surgical recovery of sperm from men with non-obstructive azoospermia[35] such that motile sperm, even in extremely low numbers, can be used for intracytoplasmic sperm injection and generate pregnancies[36].Nevertheless, others were able to collect surgically sperm even in the presence of high FSH concentrations[37]. Whether age-stratified reference intervals can help further improve the reliability of the predictive capacity of sperm retrieval linked to circulating FSH values remains to be seen.

The use of a set of age-related reference intervals for FSH from just one laboratory based upon a group of reference subjects from one location is technically only valid for that laboratory. An argument therefore exists for the use of the development of common reference intervals by laboratories using the same analytical platform[38], and choosing a number of suitable laboratories to collaborate in the production of such common age-based reference intervals could prove beneficial. The choice of laboratories is important as there are a number of pre-requisites for the establishment of such a common reference interval at the pre-analytical and analytical stages[39]. In the case of serum FSH interval ranges, the contributing laboratories should at least use 1) current WHO methodology to analyse semen and interpret results for the identification of normozoospermic men,and 2) the Siemens Centaur assay system for measuring serum FSH.The development of these reference intervals would be extremely valuable to users of the Siemens Centaur automated analyser and an improved alternative to the expected results offered by the company Information for Users document[40] and the numbers should be large enough, and then continuous reference intervals could be determined as found useful in other disciplines[41].

The limitations of the study were that patients were from only one geographical location, and the FSH analyses were made on one assay platform.

In conclusion, the present study has shown that male serum FSH concentrations increase with age making the use of age-stratified reference intervals important in the detection of men with abnormal testicular function, and the development of common reference intervals for users of the Siemens ADVIA Centaur XP automated analyser world-wide should be considered. FSH above method and age-specific reference intervals should alert the clinician to increased risk of underlying semen abnormalities and causes of male infertility.

Conflict of interest statement

The authors declare that they have no conflict of interest

Acknowledgements

The authors would like to thank the staff at PIVET Medical Centre for their help and support with this project.

Authors’ contributions

All authors have contributed positively to this manuscript. Emily-Jane Waller and Phillip Matson conceived and designed this study, undertook statistical analysis and wrote the initial draft of the manuscript. John Yovich and Jason Conceicao facilitated data collection and patient management. All authors contributed significantly to the proofreading, revision and final approval of the manuscript.