Sleep and the Testis

Abstract

Disordered sleep impairs neurocognitive performance, and is now recognized to cause metabolic ill-health. This review assesses the nascent relationship between insufficient, misaligned, and disrupted sleep with andrological health. High-quality cohort studies show a reduced sperm count in men with sleep disturbances. Well-designed interventional studies show a reduction in testosterone with sleep restriction. Studies of long-term shift workers show no effect of misaligned sleep on mean testosterone concentrations. Men with obstructive sleep apnea (OSA) and more severe hypoxemia have lower testosterone levels, although it is unknown if this relationship is entirely explained by concomitant obesity, or is reversible. Nevertheless, erectile dysfunction, which is common in men with OSA, is clinically improved when OSA is properly treated. Few studies manipulating sleep have been performed in older men, in whom the accumulation of sleep disturbances over decades of life may contribute to age-related illnesses. Improving sleep could ameliorate the development of these disorders.

I. INTRODUCTION

Sleep is an essential part of health maintenance, along with diet and exercise. Unfortunately, millions of Americans have abnormal sleep. More specifically, ~32% of American adults have sleep insufficiency (<7 hours per night) [1, 2], 10% of Americans work evenings and nights, and up to an additional 6% of workers have rotating shifts, split shifts, or an otherwise irregular work schedule, causing misaligned sleep [3]. Obstructive sleep apnea (OSA), which causes disrupted sleep, affects approximately twice as many men compared to women [4]. The prevalence of moderate to severe OSA among men is estimated to be 10% in men 30–49 years of age and 17% in men 50–70 years of age [5].

Insufficient sleep duration, misaligned, and disrupted sleep cause daytime sleepiness and impair cognitive function. Alarmingly, surveys have shown that approximately 1 out of every 25 American adults had fallen asleep while driving during a 30-day period, leading to car accidents and fatalities in some instances [6]. Additionally, abnormal sleep is known to contribute to insulin resistance and cardiometabolic illness [7]. Perhaps less well known, but importantly, studies are beginning to show that poor sleep contributes to disorders of the testicular hormonal axis, including hypogonadism, infertility and erectile dysfunction [8, 9]. Testosterone levels exhibit a diurnal pattern. This diurnal pattern appears to be related to sleep rather than endogenous circadian rhythmicity. Testosterone rises, and peaks with the first episode of rapid-eye-movement (REM) sleep; these elevated levels are maintained until awakening and then decline [10].

This review will discuss important findings between disordered sleep and the testicular axis, with a highlight on high-quality studies where 24-hour testosterone was assessed to determine the diurnal secretion of testosterone. Additionally, an emphasis is placed on recently published findings within the past 5 years. New contributions to the literature include assessment of testosterone pulsatility and the impact of sleep restriction in older men.

II. INSUFFICIENT SLEEP

Sleep duration and reproductive health

Semen analysis is the established, direct measure of male fertility. A notable cohort study examining the relationship between semen analysis and sleep assessed 953 Danish military recruits, and reported that higher and lower sleep disturbances were associated with lower sperm concentrations and lower total sperm counts [11]. Another cohort included 796 Chinese students of a military college, and reported associations between short (<6.5 hour/night) and long (>9 hour/night) sleep with reduced semen volume and total sperm count [12]. A subgroup of 592 of these men provided a second semen sample one year later. Post hoc longitudinal analysis showed men whose sleep duration improved to near an arbitrary ideal sleep duration of 7.0–7.5 hour/night had higher total sperm count [12]. The reason for change in sleep duration was not recorded. Studies have also examined relationships with testicular volume, which is indicative of male reproductive potential. Cross-sectional analyses show that higher sleep disturbances [11] and lower sleep duration [13] are both correlated with smaller testis volumes. In these two studies, these associations persisted after adjusting for body mass index (BMI); however, this does not fully exclude the possibility that the relationship between reproductive health and sleep may still be partly caused by BMI. Prospective studies which alter sleep patterns to assess effects on conception, testicular volume, or semen analysis in humans have not been done. Whether the relationship of sleep duration with reproductive health is direct or mediated by some other factor, such as BMI, is not known.

Sleep duration and testosterone

Large cross-sectional studies in young [12], older [14, 15] and young and older [16] men have not consistently shown a correlation between sleep duration and testosterone levels. There are no longitudinal reports. Nevertheless, sleep restriction in controlled laboratory settings has generally been shown to reduce blood testosterone concentrations: Table 1 [17–28].

Table 1:

Insufficient Sleep

Study	Subjects (n)	Age (yr) Mean ± SD	BMI (kg/m2) Mean ± SD	Study Design	Control Sleep Conditions	Interventional Sleep Conditions	Time of testosterone measurement	Results
24 Hour Assessment of Testosterone
Liu 2020 [17]	17 M (“young men”)	24.1 ± 2.9	Median 25.0 (IQR 22.9 – 27.5),	Randomized crossover	1 night × 8 hr	1 night × 0 hr	Q10 min × 24 hr, 1800 – 1800	Mean 24 hr T decreased, 6 AM - 9 AM T decreased, 3 PM - 6 PM T decreased in young and older men (from main effect analyses)
	18 M (“older men”)	63.9 ± 4.0	Median 29.5 (IQR 26.4–31.7)
Leproult 2011 [18]	10 M	24.3 ± 4.3	23.5 ± 2.4	Fixed order	3 nights × 10 hr	8 nights × 5 hr	Q15–30 min × 24hr, 1400–1400	T decreased during waking hours (0800–2200), particularly so 1400–2200
Dattilo 2019 [19]	10 M Undergoing strenuous exercise	24.5 ± 2.9	22.7 ± 2.3	Randomized crossover	3 nights × 8 hr	2 nights × 0 hr	Q2 hr × 24 hr 1900–1900	No change in 24h AUC
<24 Hour Assessment of Testosterone
Schmid 2012 [20]	15 M	27.1 ± 5.0	22.9 ± 1.2	Crossover, Balanced order	2 nights × 8 hr (allowed to leave lab 0700 – 2000)	2 nights × 4 hr (allowed to leave lab 0700 – 2000)	0740, then Q1 hr, 0800–2300	No change
Schmid 2012 [20]	8 M	24.5 ± 3.1	23.7 ± 1.7	Randomized balanced order, 3-period	1 night × 7 hr	1 night × 0 hr	0700	AM T decreased
					1 night × 7 hr	1 night × 4.5 hr	0700	AM T decreased
Reynolds 2012 [21]	14 M	27.4 ± 3.8	23.5 ± 2.9	Fixed order	2 nights × 10 hr	5 nights × 4 hr	0900, then Q2 hr, 1000–2000	No change
Cote 2013 [22]	Intervention: 11 M Control: 13 M	Intervention: 20.55 ± 2.21 Control: 19.23 ± 1.48	NR	Randomized parallel groups	1 night × 8 hr	1 night × 0 hr	0715, 1600 (salivary)	AM T decreased, No change PM T
Arnal 2016 [23]	14 M	31.4 ± 3.9	24.0 ± 2.0	Randomized fixed order, crossover	6 nights × 8–8.5 hr (last night in-lab)	1 night × 0 hr (In-lab)	0700, 1700	AM T decreased, No change PM T
					6 nights × 10 hr (last night in-lab)	1 night × 0 hr (In-lab)	0700, 1700	AM T decreased, PM T decreased
Sauvet 2017 [24]	16 M	27.3 ± 5.4	23.6 ± 0.6	Fixed order	1 night × 8 hr	1 night × 0 hr	0700	AM T decreased
Carter 2012 [25]	14 M	22 ± 1	25.5 ± NR	Randomized order	1 night (at home) Sleep duration 7.3 ± 0.2 hr in preceding nights (home actigraphy)	1 night × 0 hr (In-lab)	0730	AM T decreased
Akerstedt 1980 [26]	12 M	NR (Range 19 – 30)	NR	Fixed order	1 night × “normal” sleep	2 nights × 0 hr	0800	AM T decreased
Smith 2019 [27]	11 M	36.6 ± 5.6	24.1 ± 1.1	Randomized crossover	5 nights × 9 hr	3–5 nights × 4 hr	0730	No change AM T
Jauch-Chara 2013 [28]	10 M	25.3 ± 4.4	NR (Range 20.7–25.0)	Balanced order crossover	1 night × ~7 hr	1 night × 0 hr	0700	AM T decreased

Study subjects are healthy. Studies are in-laboratory unless otherwise stated. All testosterone measurements are of total testosterone from serum, unless otherwise stated

Abbreviations:

AUC – area under the curve

BMI – body mass index

Hr – hour

IQR – interquartile range

M – males

Min - minutes

NR – not reported

Q – every

SD – standard deviation

T – testosterone

A new contribution to the study of sleep restriction and the testicular axis is a prospective, randomized, crossover study of total sleep deprivation (complete nighttime wakefulness) and 8 hours of nighttime sleep in 17 young male adults (average age 24 ± 2.9 years) and 18 older men (average age 63.9 ± 4.0) conducted by Liu and colleagues [17]. This is the first assessment of sleep manipulation on the testicular axis in older men, a population at higher risk for andrological disorders [29], and in whom sleep debt accumulated over decades of life could contribute to diseases of aging. Testosterone secretion was measured at 10-minute intervals over a 24-hour period. With total sleep deprivation, mean testosterone concentration as well as basal and total secretion were significantly decreased in both young and older men, whereas pulsatile secretion and pulse number were decreased only in older men, by main effect analyses. This is the first time that the effects of manipulating sleep in older men have ever been examined. Time-of-day analyses showed that testosterone was significantly reduced in the morning and afternoon in both age groups after sleep deprivation. This finding is consistent with another study with frequent blood sampling every 15–30 minutes over a 24-hour period in 10 men (average age 24.3 ± 4.3) which showed a decrease in testosterone during waking hours of 8 AM to 10 PM [18].

The high fidelity blood sampling (every 10 minutes) allowed for the assessment of pulsatile secretion and pulse frequency of testosterone with sleep restriction by mathematical deconvolution for the first time by Liu et al [17]. The current deconvolution method has been validated mathematically [30] and empirically in 3 mammalian species [31]. It makes an unbiased estimate of the timing of pulses; the mass per pulse; the basal, total and pulsatile secretion; and its biexponential elimination [32, 33]. These findings are important because the pulsatile secretion, and not solely the hormone levels alone, is important for hormonal signaling in endocrine networks [32–34]. The episodic secretion of hormones creates pulses, which cause sudden variations in circulating hormone concentrations that allows for rapid responses to the environment. The frequency, regularity, and size (mass per pulse) of the pulses, as well the basal (non-pulsatile) secretory component, contain information that is crucial for signaling between endocrine glands, and for the formation of these pulses [32–34].

This study also assessed changes in LH (luteinizing hormone), thereby allowing appraisal of the entire testicular axis [17]. Sleep restriction had no effect on 24-hour LH parameters, but reduced LH in the morning for both age groups. LH parameters were not affected by sleep restriction in the afternoon. This suggests potential time-of-day differences in the regulation of morning and afternoon testosterone levels with sleep restriction: the morning decrease in LH is appropriately coupled with a decrease in morning testosterone levels; whereas an uncoupling of LH and testosterone in the afternoon is indicated by a recovery of LH levels despite testosterone levels remained low at this time [17].

Another recently published study of 10 young men (average age 24.5 ± 2.9) who underwent total sleep deprivation and regular nighttime sleep in a crossover, randomized fashion, showed no difference in total or free testosterone levels when measured at 2-hour intervals for a total of 24 hours. However, figures showing the hourly data indicate that both total and free testosterone levels were decreased (although perhaps not significantly) at 9 AM in the sleep restricted condition [19]. All subjects underwent strenuous exercise during both sleep conditions; this may have contributed to the lack of significant difference in 24-hour testosterone concentrations as exercise is known to acutely increase serum testosterone levels [35]. As previously discussed, the two studies which measured testosterone for 24 hours at more frequent intervals of 10 minutes [17] or 15–30 minutes [18] did show a reduction in testosterone with sleep restriction. The majority of the remaining studies, but not all individual studies, collectively provide strong evidence that sleep restriction reduces testosterone. Table 1 shows these studies and is ordered by frequency of blood sampling, and then by sample size.

III. MISALIGNED SLEEP

Circadian misalignment and reproductive health

Two population-based studies of couples randomly selected from distinct geographic locations indicate that shift work does not impair male fertility and may not cause hypogonadism without actual loss of sleep. The first study was a cohort from multiple European countries of 6630 couples planning pregnancy and 4035 couples with a pregnant female partner. The non-standard work hours of the male partner were not related to fecundity (probability of a live birth per menstrual cycle) [36]. The second study did not show a difference in semen parameters, including DNA fragmentation, between standard and non-standard male shift workers in 501 couples in 2 states of the United States [37].

Additional cross-sectional studies are not population-representative and are conflicting and inconclusive [38–41]. Prospective, randomized controlled studies of circadian misalignment induced by experimental shift work or by forced desynchrony that assess effects on semen parameters or other measures of male fertility are not available.

Circadian misalignment and testosterone

Only one study has examined normal healthy volunteers in a controlled laboratory environment [42], whereas the remainder have all been performed in actual long-term shift workers in a naturalistic setting [43–48]. Only a single blood sample was used as a surrogate for the 24-hour testosterone rhythm in 3 studies [46–48], and 2 of these did not actually assess the effect of circadian misalignment by shift work on testosterone [47, 48]. The studies that assessed shift work effects on testosterone exposure are summarized in Table 2, and are ordered according to the frequency of sampling. A major flaw in all available studies is that the effect of sleep itself has not been entirely accounted for by way of a constant routine that fully controls all environmental factors, including food intake, sleep and physical activity [49]. This is critical because sleep, particularly the evolution of REM sleep, is believed to trigger testosterone secretion [10]. Furthermore, night shift workers typically sleep less overall compared to day shift workers [50], and sleep insufficiency decreases testosterone levels (see Table 1).

Table 2:

Misaligned Sleep

Study	Subjects (n)	Age (yr) Mean ± SD	BMI (kg/m2) Mean ± SD	Study Design	Control Sleep Conditions	Interventional Sleep Conditions	Time of testosterone measurement	Results
24 Hour Assessment of Testosterone
Axelsson 2005 [42]	7 M	25 ± 3	Range 21–25	Balanced order In lab	Nocturnal sleep (2300–0700)	Diurnal sleep (0700–1500)	Q1 hr ×24 hr, 2100 – 2100	No difference in 24 hr mean. T higher during night sleep than night wake (2300–0700). T significantly increased over sleep period for both conditions
Papantoniou 2015 [43]	Day shift: 21 M Night Shift: 39 M Hospital, car factory, or railroad employees	Day: 38.4 ± 9.2 Night: 39.9 ± 9.8	Day: 26.0 ± 3.4 Night: 26.6 ± 4.0	Case-control	8-hr shifts starting between 0545 and 0700 ×5 days/week	8-hr shifts 2200 – 0600 ×3–5 consecutive nights (n=35), or 10-hr shifts 2100 – 0700 ×2– 3 consecutive nights (n=4)	24 hr urine, each void collected as separate sample (measured by GC- MS)	24 hr T: no change T peak: no change
<24 Hour Assessment of Testosterone
Jensen 2016 [44]	73 M Police officers with rotating shifts	38 ± 10	NR	Crossover	2+2, 4+4, 7+7 (night shifts + recovery days)	2+2, 4+4, 7+7 (night shifts + recovery days)	Awakening, bedtime and Q4 hours 0700 – 0300 while awake on last night shift or recovery day (salivary)	No difference in rhythm when comparing time since awakening and clock time
Touitou 1990 [45]	Control: 6 M “Healthy” volunteers Night shift: 4 M Oil refinery operators with rotating shifts	Control: 2431 Night shift: 25–34	NR	Case-control	Lights on 0700 ± 1 hr, lights off 2300 ± 1 hr	Rotating shift system every 3–4 days	Q2 hr, 0000 – 0800	Mean T decreased. Lower T at peak and trough times Sleep was not controlled for – blood samples collected at night, while night shift workers were awake/working and while control group was asleep
Smith 2006 [46]	26 M Physicians with rotating shifts	Range 24–38	NR	Crossover	7 vacation days (off work)	7 days working night shift (shift hours and shift duration NR)	0800–0930 ×1 measurement	No change T No change bioavailable T

Study subjects are healthy. Studies are naturalist, unless otherwise stated. All testosterone measurements are of total testosterone from serum, unless otherwise stated.

Abbreviations:

BMI – body mass index

GS-MS – gas chromatography-mass spectrometry

Hr – hour

IQR – interquartile range

M – males

NR – not reported

Q – every

SD – standard deviation

T – testosterone

Nevertheless, in the most thoroughly controlled and the only in-laboratory study, testosterone was measured every hour for 24 hours, and there was no difference in mean 24-hour testosterone observed in the same subjects when sleeping at night versus when awake at night under acute circadian misalignment [42]. As expected, a significant increase in testosterone was observed with sleep, irrespective of whether sleep occurred during the day or night.

The potential for sleep to confound findings explains the only study to report a lower mean testosterone concentrations in night-shift compared with day-shift workers [45]. Importantly, blood was sampled for a discrete 8-hour period (from midnight to 8 AM), and not for a full 24 hours, in 4 day shift workers and 6 other night shift workers of similar ages. Furthermore, sleep was not controlled for, meaning testosterone was measured in the night shift workers while awake, and measured in the day shift workers while asleep. Hence the reported reduction in testosterone in night shift workers was entirely due to the fact that they slept, whereas the day shift workers did not, during the time of blood sampling.

The remaining studies show no effect of shift work on testosterone. The first was a comprehensive study of 21 day shift and 39 night shift workers who had been in that shift pattern for over a year, were of the same occupation, and similar in age and BMI. No change in 24-hour urinary testosterone measured by mass spectrometry, with each void measured as a separate sample, was reported. Furthermore, peak testosterone did not change, although urine collection was deferred until awakening while the subjects were asleep [43]. An additional study in rotating shift workers measured salivary testosterone at multiple time points while awake only on the last day of a day or night shift in a rotating system of 2 days followed by 2 nights, 4 days followed by 4 nights, or 7 days followed by 7 nights. There was no difference in testosterone rhythm when comparing both time since awakening and clock time [44]. These findings are limited as testosterone was not measured while the subjects were asleep, and it was not measured over a 24-hour period. The final study of 16 physicians showed no change in testosterone from a single morning blood collection after 7 days of night shift work compared to 7 days of vacation [46].

While most of the above case-control and crossover studies in chronic shift workers or rotating shift workers did not measure testosterone during both wake and sleep over a 24-hour period, they generally show no difference in testosterone levels with misaligned sleep: Table 2. No study has fully excluded confounding by behavioral or environmental factors by way of a constant routine [49].

IV. DISRUPTED SLEEP

Disrupted sleep and reproductive health

Relationships between OSA and fecundity or semen parameters have not been studied in large cross-sectional studies. It is, however, known that andrological conditions are prevalent in men with OSA [51]. For example, low libido is a frequent complaint by obese males with OSA. Furthermore, 50% of men with OSA have erectile dysfunction [9]. A randomized, sham controlled (and therefore double-blinded) study was conducted on OSA subjects with erectile dysfunction. It is the only randomized, sham controlled trial available. Erectile function was increased by 6 units on the International Index of Erectile Function (IIEF), a clinically relevant improvement in the gold standard measure, in 61 men with moderate-severe erectile dysfunction after 3 months of adherent continuous positive airway pressure (CPAP) use [52].

Findings from many studies show a clinically significant improvement in erectile function with CPAP. These findings are seen in a large majority of studies utilizing the IIEF as well as studies assessing erectile function assessed by other methods [9].

Disrupted sleep and testosterone

Low testosterone concentrations, OSA and obesity are known to frequently occur together, but the mechanism by which they are connected is not well established [51]. While cohort studies show that more severe OSA, as indicated by degree of hypoxemia, is associated with decreased testosterone concentrations [14, 53–55], they differ in whether adiposity does [14] or does not [53–55] explain this relationship.

Three studies with frequent blood sampling showed that OSA is associated with low testosterone or alterations in its diurnal pattern [56–58]. These findings are not explained by obesity in the 2 studies that controlled for it [57, 58]. Conclusions regarding the relationship with adiposity are limited, however, due to the small number of study subjects and lack of well-matched obese controls.

Table 3 summarizes studies showing the effect of CPAP on testosterone concentrations in men with OSA, ordered by the frequency of sampling, and then by sample size. There are no studies examining the effect of CPAP on 24-hour testosterone. Only one study assessed testosterone during a 12-hour period from 7 PM to 7 AM [59], and the remaining studies examined the effect of CPAP on morning testosterone, often assessed only once [53, 60–68]. Accordingly, the effects of CPAP on circulating testosterone levels in men with OSA are likely inconclusive due to significant study limitations. Nevertheless, a recent meta-analysis of 232 men that included 7 of these studies reported no significant change in testosterone levels with CPAP therapy for OSA [69], but the veracity of that finding ultimately depends on the quality of the individual studies included.

Table 3:

Disrupted Sleep in men with obstructive sleep apnea

Study	Subjects (n)	Age (yr) Mean ± SD	BMI (kg/m2) Mean ± SD	Study Design	Control Sleep Conditions	Interventional Sleep Conditions	Time of testosterone measurement	Results
<24 Hour Assessment of Testosterone
Luboshitzky 2003 [59]	5 M Severe OSA	49.5 ± 5.2	31.7 ± 4.7	Fixed order (untreated followed by nasal CPAP)	Pre-CPAP RDI 56.6 ± 22.4	CPAP duration 9 months	Q20 min, 1900 – 0700	Mean T increased. No change in AM T
Vlkova 2014 [60]	28 M OSA	54 (range 26 – 78)	34.9 (range 21.7 – 55.6)	Fixed order (untreated followed by nasal CPAP)	Pre-CPAP RDI 62.0 (range 28.8 – 107.4)	CPAP duration 1 night, adherence NR	Evening before sleep & morning within 1 hr of waking (plasma & salivary)	No change
Meston 2003 [61]	101 M Moderate- Severe OSA 49 M placebo 52 M CPAP	Placebo median: 48 CPAP median 50	Placebo median 35.0 CPAP median 35.1	Parallel randomized, sham-placebo controlled vs nasal CPAP	Pre-Placebo median ODI 28.5 Pre-CPAP median ODI 32.9	Duration: 4 weeks Sham adherence 4.6 ±2.4 hr/day CPAP adherence 5.4 ±1.6 hr/day	“Mid-morning”	AM T increased with CPAP compared to sham CPAP. No change in T with CPAP compared to baseline (pre-treatment)
Celec 2014 [62]	67 M Severe OSA	56.2 ± 9.4	33.7 ± 4.3	Fixed order (untreated followed by nasal CPAP)	Pre-CPAP RDI 54.4 ± 16.6	CPAP duration 1 month & 6 months, adherence NR	AM fasting	No change
Zhang 2016 [63]	53 M Severe OSA and erectile dysfunction	43.87 ± 9.17	28.71 ± 3.25	Fixed order (untreated followed by CPAP)	Pre-CPAP AHI 63.45 ± 12.03	CPAP duration 3 months, adherence 6.8 hr/day	“Morning”	No change
Grunstein 1989 [53]	43 M Severe OSA	51.8 ± 11.1	33.0 ± 6.9 **	Fixed order (untreated followed by nasal CPAP)	Pre-CPAP MOS <70%	CPAP duration 3 months, adherence NR	0600–0630 ×1 measurement	T increased Free T – no change
Li 2016 [64]	32 M Severe OSA and erectile dysfunction	47.4 ±11.0**	31.3 ± 4.1**	Fixed order (untreated followed by CPAP)	Pre-CPAP AHI 51.6 ± 16.3	CPAP duration 1 month, adherence 5.8 hr/day, 5.2 days/week	“Morning”	AM T increased
Hoekema 2007 [65]	27 M OSA	49 ± 9	31 ± 4	Fixed order (untreated followed by CPAP)	Pre-CPAP AHI 46.7	CPAP duration 8–12 weeks (additional 4 weeks for total of 12 weeks if AHI >5 & CPAP adjustment needed)	0800–1000 ×1 measurement	No change in total, free and bioavailable T
Knapp 2014 [66]	27 M OSA and diabetes mellitus type 2	65.4 ± 9.6	32.0 ± 5.2	Fixed order (untreated followed by CPAP)	Pre-CPAP AHI >15, median 43 (IQR 28–62)	CPAP duration 3 months, adherence 5.1 ±1.9 hr/day	AM Fasting	No change
Bratel 1999 [67]	11 M Moderate- Severe OSA	51.3 ±14.8**	32.0 ±6.4**	Fixed order (untreated followed by nasal CPAP, n=10, UPPP, n=1)	Pre-CPAP AHI 43.3 ± 18.8 **	CPAP duration 7 months (range 6–10 months), adherence NR	0800	No change
Macrea 2010 [68]	10 M OSA	57 ± 9	32 ± 4	Fixed order (untreated followed by CPAP)	Pre-CPAP AHI >5 in 9/10 subjects	CPAP duration 11–39 months, adherence 6.2 ± 0.9 hr/day	0700	No change

All testosterone measurements are of total testosterone from serum, unless otherwise stated.

^**Values based on larger population screened

Abbreviations:

AUC – area under the curve

AHI – apnea hypopnea index (events per hour)

BMI – body mass index

CPAP – continuous positive airway pressure

Hr – hour

IQR – interquartile range

M – males

Min - minutes

MOS – mean minimum SaO2, an index of hypoxemia

NR – not reported

ODI – oxygen desaturation index (events per hour)

OSA – obstructive sleep apnea

Q – every

RDI – respiratory disturbance index (events per hour)

SD – standard deviation

T – testosterone

UPPP – uvulopalatopharyngoplasty

Only one study has examined the effect of CPAP on the testicular axis by sampling blood for testosterone levels every 20 minutes for a 12-hour period from 7 PM to 7 AM [59]. The study subjects included 5 males with severe OSA with average age of 49.5 ± 5.3 and average BMI of 31.7 ± 4.7. CPAP improved mean and area under the curve (AUC) testosterone concentrations and marginally increased LH AUC (P=0.06) but not mean LH. Subjects were compliant with CPAP (average use 5.2 hours/night) for a duration of at least 9 months [59]. Furthermore, this study assessed pulse characteristics using a model-free method that relied on assay precision. CPAP improved increment (a measure of pulsatile secretion) of LH, but not of testosterone.

Overall, the literature is lacking frequently sampled studies assessing changes in testosterone concentrations before and after CPAP use in men with OSA. This may explain the inconclusive findings on whether or not CPAP improves testosterone levels. While studies have not assessed testosterone levels with CPAP use in the treatment of OSA over a 24-hour period, the study that measured testosterone in frequent intervals for a 12-hour period did show improvement in mean and AUC testosterone. The majority of the remaining studies show no significant change in testosterone with CPAP use for the treatment of OSA: Table 3.

V. SUMMARY AND CONCLUSIONS

The effects of disordered sleep on male reproductive health have been investigated in both epidemiological and interventional studies. Cohort studies have shown reduced sperm concentration in men with insufficient sleep. Cross-sectional data do not show impaired fertility in male shift workers. The literature is lacking large epidemiological studies of fertility in men with disrupted sleep (OSA). As cross-sectional studies are often confounded by multiple variables, future research should be in longitudinal studies in larger cohorts, which would broaden our knowledge on the effects of poor sleep.

Few studies have examined the effect of sleep restriction, circadian misalignment and disrupted sleep on 24-hour testosterone profiles: see Tables 1, 2 and 3. Furthermore, studies that sample testosterone more frequently throughout the day would elucidate changes in the diurnal variation of testosterone, including peak, nadir, and rate of decline. For example, studies attempting to assess circadian rhythmicity may require sampling every hour, whereas those attempting to determine pulsatility may need to measure testosterone as frequently as every 10 minutes. Of the studies measuring testosterone over 24 hours, only one has examined an older population [17]. Considerable opportunities therefore exist to assess the impact of accumulated disordered sleep on illnesses that are more prevalent with aging, including andrological disorders. One such disorder is erectile dysfunction in association with OSA, in which reversal of OSA with adherent CPAP improves erectile function [52]. The effect of improving sleep on hypogonadism has not been directly assessed; however, it does appear that restricted sleep decreases testosterone, whereas the quality of studies examining the effects of misaligned and disrupted sleep on circulating testosterone are insufficient to draw definitive conclusions. Nevertheless, it does appear prudent for clinicians to provide education on proper sleep hygiene and to correct any disordered sleep in the hope that the health of their patients will ultimately be improved.

Highlights.

• The hypothalamo-pituitary testicular axis is a tightly regulated network that functions to maintain eugonadal systemic testosterone exposure in order to support libido, mood, muscle mass, adiposity, erectile function and fertility.

• The signaling required to regulate this network requires pulsatile (ultradian) secretion of gonadotropin releasing hormone (GnRH), luteinizing hormone (LH) and testosterone, which underpins the diurnal concentrations of these hormones.

• Restricting sleep decreases testosterone in young and older men, and preliminary studies show that restricted sleep disrupts the network regulation of testosterone by LH in a time-of-day dependent manner.

• Misaligning sleep with night shift work does not seem to alter mean testosterone levels, but available studies have not properly controlled putative environmental and behavioral confounders.

• Disrupted sleep from obstructive sleep apnea is associated with reduced testosterone, but this relationship may be due to concomitant obesity.