Reproductive health is one of the goals for the development humankind adopted by the General Assembly of the UN in 1994. Despite general increase in the population of the world, there has been a steady decline in fertility in the last decades [1, 2] (Figure 1). It must be noted that this problem is linked to the increase in several infertile couples worldwide.

Most research studies indicate that 15% of couples remain infertile after 1 year of attempting to conceive a child [4-6]. This information is based on the WHO report published in 1991 [7], as well as a row of other studies carried out over two past decades [8, 9]. Considering the global tendency to­wards a reduction in fertility, it is possible that cur­rently, the percentage of couples without children is higher than recognized above. Concerning gender, the male factor of infertility accounts for a little over 50% [10]. The ability of a man to impregnate a wom­an is closely related to the qualitative and quantita­tive characteristics of the ejaculate. Currently, the WHO has established a lower level of normal sperm concentration to 15*106/ml [11]. Nonetheless, many authors concede that for normal fertility this parameter should be 40*106/ml or higher [12-14]. The majority of studies in the first half of the 20th Century noted a rather high concentration of sper­matozoa in the ejaculate of healthy men. On the av­erage, this was about 100*106/ml [15, 16]. In 1992, results of a meta-analysis were published for semen analysis of healthy men. The authors included the period from the 1930s to the 1990s of the previ­ous century. From the results of the above, it was established that there is a global tendency towards the worsening of sperm quality (Figure 2) [16]. The initial results appeared doubtful, but after scrutiny and verification and taking into account geographi­cal considerations among others, the results were found to be comparable. The observed reduction in the concentration of sperm of men in North America is about 1.5% per year and in Europe about 3.1% [17, 18]. According to data from contemporary popula­tion series, the sperm concentration in healthy men varies according to a geographical location but on the average ranges between 41-63*106/ml [19-23]. 

It is important to note that the concentration of spermatozoa in the ejaculate is directly related to the size of the testis [24-28]. From the above information, one may propose the presence of a negative dynamic in testicular size. However, information about the trends in change of testicular size in the population is absent. The non-uniformity of methods for measurement of testicular volume [29] and lack of standardization in the latter process leads to a wide variation of current data, thereby making retrospective analysis of the given parameter rather difficult.

Purpose of the study. The objectives of the current study are to evaluate the global secular trend in the change in the size of testes in men.

We searched for studies in English and Russia languages that assessed testicular size. For this purpose, we employed resources from Pubmed and eLibrary. Also, we analyzed citations from the literature obtained, with subsequent inclusion of the same in the selection after satisfying the given criteria. We did not set limits about the date of publication. From this volume of information, we selected data related to testicular size in men aged between 18-60 years. We excluded studies whose subjects were selected based on fertility or the presence of disease conditions such as varicoceles or cryptorchidism, that have the potential to affect testicular volume. From cadaveric studies, we excluded those that assessed volume or weight without removal of the epididymis or those which had no information about the extent of dissection of the specimens. Apart from that, in our review, we did not include studies in which only one linear dimension of the testis was measured, as well as those that employed Hynie orchidometer [30]. The next stage involved the homogenization of the various methods of measurement and calculation of the mean volume, the closest to the actual values.

For this purpose, we used the following formulae for the determination of the volume (V):

(1) V=m/1.038

where m — a mass of the organ; 1.038 — density.

(2) V=a*b*c*0.71

where a, b, c — corresponding length, width and height; 0.71 — Lambert's coefficient.

(3) V=a*b2*0.71

where a, b — corresponding length and width; 0.71 — Lambert's coefficient.

(4) V=V1/0.52*0.71

where V1 — volume calculated using the ellipsoid or prolate spheroid formula; 0.52 — coefficient for ellipsoid or prolate spheroid formula; 0.71 — Lambert's coefficient.

(5) V=a*b2*0.39

where a, b — corresponding length and width, calculated by means of external measurements without taking skin into the account; 0.39 — correction coefficient.

(6) V=VPr*0.816

where VPr — volume obtained with the aid of Prader orchidometer or similar meter; 0.816 — correction coefficient.

(7) V=VTh/0.52*0.39

where VTh — volume obtained with the aid of Takihara orchidometer; 0.52 — coefficient for ellipsoid or prolate spheroid formula; 0.39 — correction coefficient.

(8) V=V1K/0.71*0.39

where V1K — volume obtained with the aid of

Lambert's formula, using external measurement without consideration of the thickness of skin; 0.71 — Lambert's coefficient; 0.39 — correction coefficient.

Statistical analysis of the data was carried out with the aid of the program STATISTICA for Windows v.10. For the assessment of the size of the gonads in adult males, a time series was constructed. A calcu­lation of mean base and chain-growth rate, as well as a rate of increment, was carried out. The method of enlarging the intervals was utilized (a calculation of mean arithmetic levels of a given period). A calcu­lation of the moving average -mean arithmetic for 3 and 5 years (calculated intervals cross each other), we carried out time series smoothing by 3 and 5 point averages, also we cross-checked the dynamic rows for the presence of any trends with the aid of linear and power functions.

We identified 126 studies following the search carried out, which were published between 1902 and 2018. Further analysis of the publications led to a selection of 33, in which the information contained therein satisfied our set criteria for selection (Table 1). The period of these selected articles ranged from 1902 to 2016. The total number of men involved in the selected studies was 11932. From the selected studies, based on the method of determination of testicular size, we got the following distribution: 8 (24.24%) — assessment was on autopsy specimens by means of weighing or displacement of water, 4 (12.12%) — determination of linear dimensions, also of cadaveric material, 6 (18.18%) — the use of external linear measurements in living subjects, 7 (21.21%) — determination by means of Prader (6) or Takihara (1) orchidometer and 8(24.24%) — by use of ultrasonography (Figure 3).

According to the results of the calculations with the aid of the standardizing formulae, we found the calculated mean volume of each study (Figure 4). After statistical analysis of the data, the weighted arithmetic mean value was 17.43±5.64; 95% CI (17.32, 17.54). The coefficient of variation obtained was 32.34%, and this confirms the uniformity of the sizes within the period under investigation. Therefore, based on the data obtained, there were no statistically significant tendencies observed in the change of volume of testes for the last century.

Testicular volume in adults is mainly determined by the pool of differentiating germ cells as well as Sertoli cells. Many studies have established a direct quantitative relationship between the above­mentioned components of the testes and the outcome of germ cells. Considering the tendency worldwide of a reduction in the concentration of spermatozoa, it would have been logical to expect a similar tendency of reduction in testicular volume. As far as we know, such findings at present are absent.

Giwercman et al. (1991) made such an assumption. However, when they compared results from cadaveric studies from Olesen (1948), which were done at the same facility 40 years earlier, there was no statistically significant reduction in testicular size [42].

Carrying out of retrospective analysis is problematic for a couple of reasons. The main problem lies in the fact that there is no standardization of the methods employed in measuring testicular dimensions. Earlier studies were based often on the assessment of cadaveric material. The authors seldom commented on the extent of dissection of the testes, often weighed the specimens with the epididymis (Spangaro,1902; Stieve,1930 from Lambert,1951[32]). Part of dimensional studies were carried out after fixation of the specimens [53].

As recently as the 1970s, in clinical practice, the testicular size was compared to different objects of similar shape (peas, beans, grapes, hazelnuts), or subjectively characterized by authors as "small- normal" [29].

In 1942 Schonfeld and Beebe were the first to invent an orchidometer, which consisted of 23 ellipsoid models of known volume. Comparison of the testis during palpation with the model of similar size allowed for estimation of testicular volume [31]. The Prader orchidometer, which was proposed in 1966 and became popular, remaining widely in use up to date, was based on the same principle [54]. Its simplicity and low cost popularized the method among researchers and clinicians alike. A series of comparative volumetric studies indicated that a high correlation of volume between the estimated and actual. This notwithstanding, the estimated volumes were found eventually to be higher, besides, there were high reproducibility errors inherent in the process of measurement [55-62].

External measurements with the aid of a ruler or callipers were also actively employed for the evaluation of testicular size [32, 33, 35, 40, 63-70]. In the event of direct measurement of testicular size with the aid of rulers, an unavoidable error arises as a result of the inclusion of the skin and testicular tunics. Therefore, some authors suggested the subtraction of the duplication of the scrotal skin from the results of such measurements [32, 33].

The Hynie orchidometer [30] was quite popular and widely used among clinicians. The basis for calculation for this device lies in the measurement of the length of the testis (it's longest dimension-d) and applying this in the formula V=d3/4 to determine the volume of the testis [29]. It was quite predictable that comparative volumetric investigations will reveal wide margins of error using this device [55]. Another type of orchidometer was introduced by Takihara in 1983. This device consists of stencil ellipsoid rings of different sizes. The corresponding size to the testis was slipped over the gonad with the scrotal skin stretched. The calculation of the volume using this device is based on the formula of a prolate spheroid: V=0.52*a*b2 [71].

In contemporary research as well as in current clinical practise, we often employ ultrasonography for the assessment of testicular volume. This method allows us to determine more accurate dimensions due to clear visualization. However, just as with the external linear measurements, there was no consensus on the most appropriate formula for calculating the volume. Most often we employ already installed formulae on the ultrasound machines for calculating the volume of an ellipsoid: (V=0.52*a*b*c) or that of a prolate spheroid: (V=0.52*a*b2) [48]. In recently published recommendations, the working group of the European Society of Urogenital Radiology agreed to use Lambert's formula (V=0.71*a*b*c), in computing the volume of the testis during ultrasound investigations, as the closest in correlation to the actual testicular volume [72].

Another important factor that makes retrospective analysis difficult to carry out is the fact that testicular size varies greatly among different age groups. Most contemporary studies focus on the measurement of testicular volume in children and adolescents as a factor of sexual development [46, 61, 62, 66, 67, 69, 73-91]. The size of gonads after puberty alters only insignificantly. Testicular volume will tend to reduce only after 60 years of age [48, 53, 92-95].

Most often researches focus their investigations on certain disease states of the male reproductive system which have the potential to affect the size of the gonads. This may be a selection based on fertility [28, 96, -100], the presence or absence of a varicocele [41, 81, 99, 101, 102], cryptorchidism in the past medical history [68, 103] or indeed from andrological patients [26, 60].

Finally, some researches point to the fact that there are racial and ethnic differences in the size of gonads in adult males [71, 79, 104].

In the current study, we analyzed the volume of the testes in healthy men aged 18-60 years using available publications in the literature. After the selection of data that satisfied the required criteria, we homogenized the given volumes. The measurement of testicular volume based on the Archimedes principle of displacement of water was considered the most accurate. The density of the testis without the epididymis is 1.038 g/cm3 [94]. This parameter was used to calculate the volume in the studies that were carried out through a measurement of the mass of the testis (Formula 1).

Even though the shape of the testis is like a prolate spheroid, many studies indicate that more accurate results can be obtained using Lambert's formula with a coefficient of 0.71[32, 55, 59, 105, 106]. For this reason, we used Lambert's formula for calculation of testicular volume, whose linear dimensions are known (Formula 2 and 3).

In his original article, Lambert pointed out the importance of subtracting the skin thickness when calculating testicular volume with the aid of external linear measurements [32]. In studies where the authors subtracted the skin thickness but then calculated the testicular volume using the formula of an ellipsoid, we recalculated by employing Lambert's coefficient (Formula 4).

V. Dornberger and G. Dornberger [55] in their studies based on comparative volumetric investigations suggested a correction coefficient of 0.39 in the ellipsoid formula during external measurements which were carried out without taking into account the thickness of the skin. We utilized this in formula 5. In the case where the authors calculated the volume of the testis using Lambert's formula without subtracting the skin thickness, we employed formula 8. The principle behind the Takihara orchidometer is based on the ellipsoid formula. The idea behind this method is similar to that of external linear measurements without correction for skin thickness. For the determination of testicular volume in such cases we employed formula 7.

The margin of error with the use of the Prader orchidometer varies widely depending on the age and volume of the testis [61, 85]. For volume calculation, we used the adjusted coefficient of 0.816 which was suggested by Sakamoto [59] (formula 6).

Ultrasound results were recalculated using Lambert's formula.

Analysis of the data we obtained confirms the absence of any kind of significant global secular trends in the testicular size of adult males. This does not correlate with the concept of a global tendency towards a reduction of sperm concentration in the ejaculate.

There is a possibility that the factors which have an inhibitory action on spermatogenesis, do not necessarily lead to a reduction in the number of Sertoli and germ cells, i.e. the weight and volume of the testis, but influences the differentiation of the latter.

In the light of global trends towards reduction infertility, it is necessary to carry out further wideranging investigations in order to elucidate the nature of the process. Standardization of methods of assessment of the results will aid in reducing the margins of error and this will assist in the analysis of the existing tendencies.