Understanding and treating ejaculatory dysfunction in men with diabetes mellitus

Diabetes mellitus is a rapidly rising metabolic disorder with important systemic complications. Global figures have demonstrated the prevalence of diabetes mellitus has almost quadrupled from 108 million in 1980 to 422 million in 2014, with a current prevalence of over 525 million. Of the male sexual dysfunction resulting from diabetes mellitus, significant focus is afforded to erectile dysfunction. Nevertheless, ejaculatory dysfunction constitutes important sexual sequelae in diabetic men, with up to 35%–50% of men with diabetes mellitus suffering from ejaculatory dysfunction. Despite this, aspects of its pathophysiology and treatment are less well understood than erectile dysfunction. The main disorders of ejaculation include premature ejaculation, delayed ejaculation, anejaculation and retrograde ejaculation.

Aside from the obvious cardiovascular morbidity and mortality attributable to DM and other chronic ailments caused by it, male sexual dysfunction is a common sequalae. Overall, up to 52% of men with DM will suffer erectile dysfunction (ED). 4 However, less attention is afforded to EjD in the clinical assessment of these men despite there being a close interplay between ED and EjD. 5 Furthermore, EjD has also been reported as having a significant impact on quality of life 6 and can negatively impact on natural conception, accounting for up to 2% of all infertility cases. 7 More important to consider in this context is the finding that sexual dysfunction can be an early marker of more significant cardiovascular disease and subsequent cardiovascular events in diabetic patients. 8 Whilst the strong aetiological link between DM and ED is well established, less focus is placed on its role in EjD. Nevertheless, DM is a known aetiological factor contributing to EjD. 1,9 Whilst disorders of ejaculation are commonly considered concurrently alongside other aspects of sexual dysfunction (largely ED), in this review, we will focus on the ejaculatory aspect of sexual dysfunction, in particular PE, RE, DE and AE as these are the most common ejaculatory disorders associated with DM. There are multiple pathological mechanisms proposed that can contribute to the aetiology of EjD in DM and a knowledge of these can aid the clinician in understanding the various therapies available for diabetic men with ejaculatory disorders. This review will seek to explain the normal neurophysiology of ejaculation and the different ways this can be affected in DM, as well as the treatments available to this group of men.
causes can be divided into pharmacological, neurogenic or anatomical. 7 The true prevalence of RE has been difficult to ascertain, and despite the strong association with DM, its prevalence in DM is also highly variable amongst different populations/studies ranging from 6% to 34%. 20 Patients suspected to have RE may complain of a low volume or absent ('dry') ejaculate or may complain of seeing cloudy urine on post-orgasmic micturition. Diagnosis is confirmed by the presence of spermatozoa on a post-orgasmic urine sample and the two wellknown criteria for a confirmatory diagnosis are 10-15 spermatozoa per high power field 21 or >1 million spermatozoa in the post-orgasmic urine sample. 12 The problem with such criteria arises when men are azoospermic and have concomitant RE. In such scenarios, detection of fructose in the post-orgasmic urine sample would allow an inference of RE. 22 Other than the psychosexual distress caused by experiencing an unwanted 'dry' ejaculate, RE can also be distressing because of the impairment in fertility.
Delayed ejaculation is thought to be on the same spectrum as AE (complete failure of ejaculation). It is much less common than PE, 23,24 with an estimated prevalence ranging from 1% (lifelong) to 4% (acquired) 20 and is less clearly defined. The definition outlined in DSM-V states there must be a 'marked delay' or 'marked infrequency of absence of ejaculation' , where the delay in ejaculation is not desired and persistent for 6 months or more, resulting in significant distress to the individual. 13 However, unlike PE, it does not give a specific temporal definition for the IELT. Attempts have been made to define DE using an IELT-the third International Consultation on Sexual Medicine defined DE as an IELT threshold beyond 20-25 min. 9,19 This time cut-off was chosen as it was well above two standard deviations of the mean IELT based on population studies. 9 However, this definition has been criticised by some groups and other IELT cut-offs have been proposed, for example, IELT ≥10 min, based on 10 min being the upper limit of the range of IELT for men from population-based studies. 9,25,26 Regardless of this variation in the definition of DE, from a practical standpoint, the common criterion for diagnosing it is an unwanted delay in ejaculation, which results in significant distress in the individual.

Normal physiology of ejaculation
Normal ejaculation comprises a highly coordinated sequence of neuronal signalling, largely mediated via autonomic and somatic innervation of the penis and pelvic musculature to release semen from the external urethral meatus. 27 In addition, there is important central modulation of the whole sexual response of erection followed by ejaculation. 28 In normal physiology, an erection is mediated by parasympathetic stimulation via the cavernous nerves, travelling alongside the pelvic nerve, which arises from spinal segments S2-4. 27 The cavernous nerves innervate the corpora cavernosa and elicit non-adrenergic, non-cholinergic smooth muscle relaxation. Furthermore, arterial vasodilatation and endothelium-dependent relaxation of the corpus cavernosum lead to penile erection. 29 This is usually in response to sexual stimulation; visual or penile tactile stimulus.
The latter involves sensory signals conveyed to the central nervous system (CNS) via sensory afferent nerve fibres mainly in the dorsal nerve of the penis via the pudendal nerve to S2-S4. 27 Rat studies have identified a spinal ejaculatory generator (SEG), 30 which has also been corroborated in humans from post-mortem studies. 31 This SEG comprises a specialised group of spinothalamic neurons relatively recently found to be located at L3-L5 spinal cord segments in humans. 31 This is important in triggering ejaculation by synchronised and coordinated activation of the autonomic and somatic spinal efferents to the genital tract mentioned above. 31,32 Various centres in the brain summate all signals from cognitive, visual and tactile sexual stimuli during an erection, which then feed into the SEG. 30 When the intensity of these signals increases with sexual activity/stimulation, it nears a threshold with maximal activation of SEG, thereby triggering ejaculation. 30,33,34 This normally occurs concurrently with an orgasm, which is a euphoric event experienced centrally. It is important to distinguish between orgasm and ejaculation. Orgasm is a CNS phenomenon experienced by the individual as the result of complex cerebral processing of signals received from spinal centres and peripheral sensory feedback, which is associated with the pleasurable sensation of seminal tract, urethral bulb and bulbospongiosus contraction. 27,29,35 Whereas ejaculation is simply the physical process of expelling semen from genital tract.
Ejaculation can be divided into two phases-emission followed by ejection/expulsion, mediated by the sympathetic and somatic nervous system, respectively. 36  Once the semen is deposited at the posterior urethra (emission), the ejection phase is mediated by impulses travelling along somatic motor nerve fibres (efferents) of the pudendal nerve in a pulsatile fashion, which bring about rhythmic contractions of the pelvic floor muscles as well as bulbospongiosus and bulbocavernosus muscles surrounding the urethra. Cell bodies of motor fibres in the pudendal nerve arise in the ventral horn of spinal segments S2-S4, also known as Onuf's nucleus. 32 The ejection part of ejaculation involves acetylcholine neurotransmission at the pudendal nerve acting at target striated muscles such as bulbospongiosus and bulbocavernosus to elicit contraction and antegrade propulsion of semen. Hence, ejection is triggered by the presence of semen in the posterior urethra, sensed by sensory afferents, bringing about reflex contraction the of pubo-perineal muscles via somatic motor efferents travelling in the pudendal nerve. These neuronal pathways governing the emission and expulsion phases of ejaculation are depicted in Figure 1.

Central modulation of ejaculation
Ejaculation is under strong influence of central neuronal pathways, which in turn receive input from other centres within the CNS as well as peripheral afferents. Ultimately, central and peripheral signals feed into to the SEG which coordinates neurotransmission to the genital tract to bring about ejaculation. Important centres in the brain implicated in the process of ejaculation include the medial pre-optic area (MPOA) and paraventricular nucleus (PVN), both found within the hypothalamus. 28 Both of these are involved in promoting ejaculation, whereas the nucleus paragigantocellularis located at the medulla is thought to have an suppressive role, sending inhibitory signals to the SEG. 28,32 The thalamus, posteromedial bed nucleus of stria terminalis and posterodorsal medial amygdaloid nucleus are centres involved in receiving sensory signals related to sexual stimulation and processing the information before modulating the excitatory or inhibitory signals, which are relayed to the spinal centres either via the hypothalamus or brainstem centres. 32 The SEG also sends regulatory signals to the thalamus. This is depicted in Figure 2.
A multitude of neurotransmitters have been found to be important in providing central regulation of ejaculatory function; however, rodent studies have identified dopamine, serotonin (5-HT) and oxytocin as the most important. 28,[40][41][42][43][44] Dopamine is thought to promote ejaculation via D2 41 and D3 42 receptors at centres in the hypothalamus, whereas serotonin largely exerts an inhibitory role via its action at 5-HT 2C receptors. 43 However, it has also been found that agonist action at 5-HT 1A can have pro-ejaculatory effects. 43 Animal studies have shown oxytocinergic neurons projecting to the SEG to largely have a pro-ejaculatory effect. 44

PATHOPHYSIOLOGY OF EjD IN DM
Peripheral neuropathy is a well-established sequelae of DM. 45 The basic tenet of diabetic peripheral neuropathy is based on the detrimental effects of chronic hyperglycaemia and insulin resistance on the structure and function of neurons and accompanying microvasculature. The neuropathy is believed to arise from a combination of oxidative stress, build-up of cytokines inducing neuronal inflammation, as well as mitochondrial and cellular dysfunction, eventually lead-ing to neuronal cell death. 45,46 When this implicates the autonomic nervous system, various aspects of sexual dysfunction, 1 alongside autoregulatory dysfunction, can manifest. It is widely believed that the EjD encountered by diabetic men is related to this; however, multiple animal and human studies have provided evidence for other, less well-recognised, pathophysiological mechanisms. 1

Peripheral mechanisms
Historical studies on rats with drug-induced diabetes have showed reduced sympathetic innervation of the VD and SVs by peripheral noradrenergic neurons, resulting in impaired contraction of the seminal tract during emission. 47 Interestingly, this was only shown in chronically diabetic rats (more than 6 months) and no difference in noradrenergic neurotransmission was seen at the neuromuscular junction (NMJ) of rats with a shorter duration of diabetes. 47 This may manifest as a reduced force/volume of ejaculate. An apparent paradox to this impaired noradrenergic stimulation of the seminal tract is the finding that there can be a heightened sensitivity to exogenous noradrenaline in chronically diabetic rats. 47,48 This increased sensitivity has been inferred because of a higher frequency of action potentials measured at the NMJ; however, this does not mean increased force of contraction. 48 This may partly explain PE, albeit with reduced volume of the seminal fluid on emission in diabetics. The increased frequency of these impulses, and hence increased sensitivity to noradrenergic neurotransmission may be explained by upregulation of calcium ion (Ca 2+ ) channels in these neurons in chronic diabetic states. 48 This hypothesis has been supported by other early rodent studies. 49,50 The formation of free radicals and reactive oxygen species (ROS) is a well-known feature of diabetic neuropathy. 45,46,51 In the context of EjD, this has also been demonstrated as a factor associated with the reduced sympathetic drive seen in diabetic rats. 52,53 The specific mechanism involves deposition of advanced glycation end-products (AGEs) because of serum proteins and amino acids being chronically exposed to a hyperglycaemic state. Adavanced glcation end-products interact with the receptor for AGEs on the membrane of neurons, leading to upregulation of pro-inflammatory cytokines and ROS which promote cellular dysfunction and apoptosis. 54 Diabetic neuropathy has also been attributed to the effects of the closely associated Schwann cells, responsible for myelination of axons and optimising neurotransmission. 45,46 It has been shown that the hyperglycaemia of DM can reduce Schwann cell proliferation and migration, thereby resulting in axonal dysfunction. 55 Animal studies have shown this may be because of upregulation of phosphodiesterase enzyme subtype 5 (PDE5) in response to the hyperglycaemia and results in reduced myelin sheath thickness as well as nerve conduction velocity. 56 Nevertheless, alterations in myelination of peripheral nerves is thought to occur at advanced stages of poorly controlled diabetes. 39 Impaired sympathetic activation during emission phase of ejaculation also leads to failure of the internal urethral sphincter/bladder neck to relax. 12,57 Human studies have supported this as a mechanism

Central mechanisms
A well-documented finding in diabetics is the dysfunctional dopaminergic CNS neurotransmission. [67][68][69][70] Whilst this has important applications in the therapy for other important metabolic and biochemical sequelae of diabetes, it may also contribute to EjD in diabetic men, given dopamine acting at D2 and D3 receptors is important in signalling to the SEG to stimulate ejaculation. 41,42 It is conceivable then that this may manifest as DE or AE in DM.
Serotonin (5-HT) neuro-signalling has also been found to have important behavioural and physiological roles, with 5-HT's plethora of effects exerted by its many receptor subtypes. 71 This can be pathologically implicated in DM. 72,73 In fact, murine studies have demonstrated impaired glycaemic control and insulin resistance, leading to DM as well as eating disorders leading to obesity in mice with deletions in the 5HT 2C receptor gene, 74,75 indicating a pathogenic mechanism for this particular receptor in T2DM. Supportive of this is the finding that supplementing diabetic mice with a 5HT 2C receptor agonist improves insulin resistance and glycaemic control. 76 Pertinent to this review, it has also been established that 5HT 2C receptor signalling is important in inhibiting ejaculation. 43 Hence, the dysfunction in serotonergic signalling, more specifically at the 5HT 2C receptor, encountered in DM may relate to the acquired PE seen in this cohort of patients. This has been reported as a plausible mechanism for PE in the literature. 77 Oxytocin is a neuropeptide released by the posterior pituitary and has an important mode of action both centrally and peripherally in the mediation of male sexual behaviour, namely promoting libido, erection and ejaculation. 44,78,79 Oxytocinergic neurons arising at the PVN in the hypothalamus project to other sites in brain, and the spinal cord and are involved in the signalling of the initial stages of ejaculation. The SEG in the lumbar spinal cord of rats have gastric-releasing peptide (GRP) neurons, thought to be important in triggering ejaculation, 80 and these GRP neurons at the SEG express oxytocin receptors. 81 Receptor activation here by oxytocin has also been shown to be instrumental in coordinating ejaculation. 81 Supportive of the pro-ejaculatory effect of centrally released oxytocin is the finding that increased local concentration of oxytocin at the SEG coincides with a reduced number of intromissions in rodents before ejaculation, effectively reducing the ejaculatory latency. 44,81 This effect has also been shown, by the same group of authors, on systemic administration of oxytocin. 81 Conversely, administration of an oxytocin antagonist centrally prolongs the ejaculatory latency. 44 This clearly has important potential applications for the treatment of PE, DE and AE. With respect to diabetes, there has been a strong negative correlation shown in circulating levels of oxytocin and hyperglycaemia. 81 Thus, this can play a potential role in DM-related EjD.

POSSIBLE PREDICTORS OF EjD IN DM
Although many studies have investigated the association between various metabolic, socioeconomic and psychological factors, and ejaculatory disorders in diabetic men, it must be remembered these studies are largely observational/cross-sectional and retrospective. Hence, these associations, even if deemed statistically significant, cannot be assumed to be causative. Nevertheless, they are important to consider as they may be influential in the work-up and possible treatment of diabetics with EjD.

Age
Age has been found in epidemiological and cross-sectional studies to correlate with the prevalence of EjD, in particular PE, amongst diabetics. 72,82 Diabetic men older than 50 years have been found to be twice as likely to suffer from PE than their younger counterparts. 82 However, another cross-sectional study found no correlation although this study involved a much smaller sample size. 83

Duration of diabetes
The above association with age may be supported by reports of an increased prevalence of PE in those with a longer duration of DM.
In particular, El-Sakka 82 found in a cohort of more than 600 diabetic patients, those with a longer than 10-year history of DM were nearly three times more likely to have PE than those having been diagnosed less than 5 years. Diabetes being a chronic condition, lends this finding to being interpreted as the result of long-term changes imposed on the aforementioned structures involved in ejaculation by hyperglycaemia and/or insulin insufficiency. This draws parallels to findings in animal studies where the supplementation of insulin in rats during the later stages of T2DM fails to restore emission, whereas administration at earlier stages has the potential to restore seminal fluid during emission. 63 Similarly, a randomised control study of T1DM and T2DM men found those with RE had a longer duration of diabetes than those without. 12 Nevertheless, other observational studies have failed to show this link between duration of DM and ejaculatory disorders such as PE. 72

Glycaemic control
The correlation between poor glycaemic control and microvascular complications of DM are well documented. 72,84,85 However, the link between glycaemic control or glycaemic variability and EjD in diabetics is less well established. The previously mentioned observational study by El-Sakka 82 reported a significantly higher propensity of PE in T2DM with poor glycaemic control (HbA1c >7%) than those with lower HbA1c-it was reported those with poor glycaemic control would be up to 9.6 times more likely to have PE than those with good glycaemic control. This association was also supported by another study in T1DM patients which reported a statistically significant negative correlation between HbA1c and IELT time. 86 Similarly, a rat study in which allantoin was supplemented to diabetic rats showed an improvement in PE (diabetic-induced reduction in ejaculatory latency time).
This study also observed a reduction in HbA1c alongside this change in sexual dysfunction. 87 However, another observational study on T2DM patients with PE failed to show this correlation. 72 Erratic blood sugars, that is, high glucose variability (GV), can also be considered poor glycaemic control. Observational studies on T1DM men have failed to show any significant difference in PE between the high GV and low GV groups, suggesting other non-glycaemic aetiological factors may be responsible for PE. 88 Interestingly, it has also been shown that low blood glucose indices (increased risk of hypoglycaemia) are associated with greater likelihood of PE in T1DM. 86 It is postulated that this may be due increased sympathetic tone in states of hypoglycaemia, which promotes earlier emission. This re-iterates the complex multifactorial aetiology of PE. It is also important to remem-

Concomitant ED
One of the most common associations for EjD, largely PE, is that with concurrent ED in diabetics. 72,82,92,93 Even the two large observational studies that reported differing associations between PE and glycaemic control had concordant findings of a positive correlation between DMrelated PE and ED. 72,82 Supporting this, are reports from a recently published cross-sectional study finding a greater than fourfold prevalence of PE in those who had DM-related ED than those without ED. 83 Given that the majority of ED in DM is attributed to DM-related neuropathy and endothelial dysfunction, it supports the aetiological role of neuropathy in DM-related EjD. Hence, in diabetic patients who report ED (the more frequently reported sexual dysfunction in DM), EjD should also be actively enquired about.
The close relationship between ED and acquired PE in DM also suggests that these two common disorders of sexual function have overlapping pathophysiology. The chances of those suffering from PE having concomitant ED are significantly high (odds ratio 3.68), without factoring the presence/absence of DM. 5 Both have variable degrees of psychogenic and organic pathophysiology and may be part of a viscous cycle where one propagates the other. For example, one of the theories behind this close association is that these men may subconsciously try to hasten sexual activity and intercourse to mitigate against the failure to maintain an erection because of their DM-related ED. 83,94 The PE, in turn, adds to future performance anxiety and can worsen the ED/PE by adding to the psychogenic component of these sexual disorders. Conversely, others have suggested men with PE may try to offset/prolong the early IELT by trying to reduce sexual arousal and hence this may inadvertently lead to detumescence and failure to maintain an erection, further compounding the ED. 95

Serum testosterone
Hypogonadism is accepted as an important feature that may coincide with the metabolic syndrome in men, 96 Indeed, it has been postulated that differing levels of T may affect ejaculatory function. 104 Conflicting data have been presented in reference to T levels and PE, 105 with some studies reporting lower T levels associated with PE, 87,106,107 whereas others reporting PE to be associated with high T levels. [108][109][110] The latter has been the more common finding and has been supported by many in the literature. 29,111,112 It is hypothesised that high levels of T may be involved in PE through the stimulatory effects on emission mediated via AR expressed centrally at the MPOA. 111 This may lead one to speculate the commonly observed hypogonadal state in DM and metabolic syndrome may result in DE. However, a randomised control study in hypogonadal men comparing testosterone replacement to placebo has failed to show an improvement in ejaculatory dysfunction. 113 In support of this, a large multicentre observational study surveying almost 1000 men failed to show any association between IELT, that is, possible DE or PE, and serum T level. 114 Additionally, a rodent study specifically investigating the effects of T in diabetic rats failed to show any direct effect of T on EjD. 115 In contrast to this, another rodent study found supplementation of allantoin (a naturally occurring substance with proposed health benefits) prolonged a curtailed ejaculatory latency time in diabetic rats amongst other domains of sexual function. 87 The authors concluded this was because of a rise in serum T driven by allantoin. However, it has been shown that allantoin also upregulates NOS and improves endothelial dysfunction, alongside improving other aspects of sexual dysfunction such ED and sexual desire in these subjects; hence, it is difficult to accurately attribute the improvement in PE to a rise in T alone. In view of the contrasting findings of other studies, a more plausible inference is likely to be that serum T has more profound effects on libido/sexual desire 105,116 and even erectile function rather than direct effects on ejaculatory function in diabetics, especially when compared to the aforementioned pathophysiological mechanisms responsible for diabetic neuropathy. 105,117

Genetic pre-disposition
It has been shown that DM patients with PE have a strong association with a polymorphism constituting a long trinucleotide (CAG triplet) repeat sequence in the gene coding for the AR, compared to diabetics without PE. 93 The DM-related PE in this study was diagnosed/measured using the IELT and score on the premature ejaculation diagnostic tool, both of which had significant correlation with the length of the CAG trinucleotide repeats in the AR gene. 118 The authors postulated this polymorphism may reflect hormonal modulation in the patient that mediates the PE; more specifically a raised T level and low oxytocin levels. The former would seem to support studies proposing raised T levels increase the risk of PE. 108,110 Related to the possibility of a genetic pre-disposition to EjD, is the suggestion that men of Asian ethnicity are more likely to have PE.
This observation has been made by a large observational study that found both Asian diabetic and non-diabetic men were significantly much more likely to have PE than their European counterparts. 103 Interestingly, the same study did not find any difference in prevalence of ED between Asian and European diabetic men. Contrary to the above study, a recent systematic review of studies on sexual dysfunction between Asian and European populations actually found PE to be higher amongst Europeans. 119 However, this was not specific to diabetics.
When considering PE not specific to the diabetic population, a recent systematic review by Mostafa et al. 120

Co-existing depression/anxiety
Whilst sexual dysfunction has been shown to be strongly linked with psychological disorders, 122,123 the presence of depression/anxiety in patients specifically with DM has been found to increase the proba-bility of concurrent EjD, namely PE. 118,124,125 Furthermore, a recent observational study investigating aetiological factors of acquired PE found just over 30% of patients with DM-related PE have either depression or anxiety. 125 Supporting this strong relationship between sexual dysfunction and depression, a recent meta-analysis revealed a bidirectional association between the two, such that those with depression are more likely to exhibit sexual dysfunction (adjusted odds ratio of 1.7) and those with sexual dysfunction are more likely to have depression (adjusted odds ratio of 3.2). 126 These data reinforce the common assertion that EjD even in diabetics, especially PE, has a complex multifactorial aetiology and is influenced by non-organic pathology.
Studies reporting on the association of all the above factors with EjD in DM are shown in Table A1.

Type 1 DM versus type 2 DM
Differences in the aetiology of sexual dysfunction in T1DM and T2DM have been postulated from the well-documented differences in ED in observational studies 127 and a recent meta-analysis 4 between these two cohorts of diabetic men. It is known that the pathogenesis in T1DM is that of insulin deficiency, which may have an autoimmune aetiology, hence, tends to present at a younger age, whereas T2DM is largely attributable to the metabolic syndrome and insulin resistance.
With the latter, there is also a strong association with chronic inflammation impacting the microvasculature and neuronal signalling. The latter, bearing a strong correlation with the metabolic syndrome, can also potentially be compounded by adverse cardiometabolic factors (hypertension, obesity, hypogonadism and hyperlipidaemia) which may contribute to the EjD.
A recent large meta-analysis has shown the prevalence of ED is higher in those with insulin resistance (T2DM) than those with T1DM. 4 Less robust data have been reported for differences in EjD between these two diabetic groups. Interestingly, a study comparing PE in T1DM to non-diabetic men found no significant difference in prevalence of PE between both groups; however, it did find a significant association between PE and glycaemic control in T1DM. 86 Conversely, in the previously discussed prospective study comparing PE in T2DM to healthy non-diabetics, PE was significantly higher amongst the diabetic group. 72 Although there is no robust randomised controlled data comparing EjD between T1DM and T2DM, these two studies 72

TREATMENT STRATEGIES FOR EjD IN DIABETIC MEN
The different disorders of ejaculation in diabetic men are distinguished as acquired, rather than primary or lifelong. It is key to remember that EjD can be multifactorial, and especially in T2DM, the underlying causes of the EjD can have contributions from other co-existing cardiovascular, endocrine and metabolic disorders. Hence, careful investigation and optimisation of these is a paramount part of managing the comorbid patient with EjD. Moreover, there can be complex psychological/central interaction. Once an acquired form of EjD has been diagnosed and attributed to DM, a treatment strategy can be formulated.

Premature ejaculation
Data concerning treatment of PE specifically in DM is scant; however, treatment of DM-related PE should begin by correcting/optimising glycaemic control and other possible contributory metabolic/cardiovascular factors; an approach which has been supported by international guidelines. 128 The studies suggesting that DM-related PE may be strongly associated with poor glycaemic con- Any co-existing ED is also important to treat and this should be instigated with PDE5i. 128 Indeed, some groups have also suggested monotherapy with PDE5i may improve perception of ejaculatory control 131,132 and in some cases have also been shown to increase IELT with reduced post-ejaculatory refractory time (time taken to achieve another erection following ejaculation). 130 In cases where the initial optimisation of diabetes or co-existing ED fails to ameliorate the PE, pharmacotherapy with selective serotonin re-uptake inhibitors (SSRIs) is often the next step. 128 These theoretically increase serotonergic neurotransmission at the 5HT 2C receptors which are inhibitory to ejaculation. There is strong evidence from randomised control trials (RCTs) for the use of SSRIs in PE, [133][134][135] and although the majority of these have been conducted in patients with primary PE, in practice their findings have been applied to the treatment of acquired PE. 136 Amongst the different SSRIs (paroxetine, fluoxetine, clomipramine and sertraline) superiority in prolonging IELT has been shown for paroxetine. 136 The treatment of PE with these SSRIs is considered 'off-label' , with the exception of dapoxetine, which has been licensed for this use by the European Medical Agency (EMA). It is rapid-acting and has a short half-life, making it ideal for on-demand use, usually taken at 30-60 mg, 2 h prior to planned intercourse. 126 A recent RCT studying the efficacy and safety of combined dapoxetine with PDE5i (sildenafil) in those with concomitant ED and PE showed safe use and superior efficacy of combined use comparedmonotherapy with SSRI in prolonging IELT in these patients. 137 The superior effect of combination treatment in prolonging IELT for patients suffering from ED and PE has also been supported by a meta-analysis. 138 Off-label treatment with other SSRIs such as paroxetine would involve daily drug administration rather than on-demand use as this yields the best effect; however, on demand use can be instituted at the expense of slightly reduced efficacy. 139 The side effects of SSRIs are self-limiting and the most common ones are often mild, for example nausea, headache and diarrhoea. Of pertinence to young diabetic patients with PE, for example those with T1DM, SSRI use has been associated with slightly higher risk of suicidal ideation amongst young patients; hence, caution is to be exercised when contemplating prescribing SSRI to young diabetics under the age of 25 years. 140 The use of topical anaesthetic gels such as prilocaine or lidocaine to attenuate sensitivity of the glans is an old treatment strategy for PE and has evidence from RCTs 141,142 and meta-analysis 143 demonstrating its efficacy in prolonging the IELT. However, its use is limited by the attenuated sensation and enjoyment of intercourse in the patient, and in the partner where there is vaginal absorption of the local anaesthetic (if a condom is not used). 139 The lidocaine/prilocaine metered dose spray (Fortacin) containing a pure mixture of lidocaine 150 mg/ml and prilocaine 50 mg/ml is the only EMA-approved topical anaesthetic agent approved for use in PE. 142 However, these topical agents should be avoided in couples wishing to conceive as they may have direct toxic effects on spermatozoa. 144 A proposed management algorithm for PE is shown in Figure 3.

Delayed or anejaculation
Like with other forms of EjD, careful attention must be given to establish a diagnosis of DE or AE from the history and clinical examination.
Both are rare relative to PE. AE is the complete lack of semen (whether antegrade or retrograde) but usually accompanied with orgasm; hence, it is important to exclude RE before diagnosing AE. Due to the relative rarity of DE, there is scarce data regarding treatments, let alone specific outcomes in diabetics.
The proposed mechanisms in DM-related DE or AE include: • A peripheral neuropathy leading to an attenuated autonomic neurotransmission at the seminal tract, impairing emission.

F I G U R E 3
Basic algorithm for managing diabetes mellitus (DM)-related premature ejaculation (PE).ˆShould be avoided in couples desiring fertility/conception. # Paroxetine given as on-demand regimen (cf. daily administration) may reduce efficacy. *Phosphodiesterase inhibitor (PDE5i) can be trialled as therapy for isolated PE without erectile dysfunction (ED) (given alone or with selective serotonin re-uptake inhibitor [SSRI]). 121 Like DM-related PE, DE can have a significant psychogenic component and careful psychosexual history with appropriate referral to a psychosexual counsellor may be necessary. Where this has failed or not deemed appropriate, there are many pharmacotherapies that can be offered; however, all available options are regarded off-licence because of lack of robust RCT placebo-controlled data. 128,146 Some of the most commonly utilised agents include cabergoline, bupropion, oxytocin 147 and α-1-adrenergic agonists (e.g., pseudoephedrine).
Cabergoline is a D2 receptor agonist and activation at this receptor promotes ejaculation. 146 The reduction in prolactin induced by cabergoline has also been shown to reduce IELT time. 148 Bupropion is an atypical antidepressant and classed as a dopamine and noradrenaline re-uptake inhibitor, thereby increasing dopaminergic and noradrenergic tone. It has been effectively used as an alternative or adjunct to other antidepressants in SSRIinduced sexual dysfunction and DE. 149 It has also been shown to increase epididymal and vas contraction in male rats by augmenting sympathetic tone at these locations, thereby promoting emission. 150,151 The use of penile vibratory stimulation (PVS) has shown some success in inducing ejaculation in patients with DE or AE. 152 The premise of PVS involves activating the ejaculatory reflex arc via the pudendal nerve, largely recruiting the lumbosacral spinal centres S2-S4 and the SEG at L3-L5. Sensory afferents stimulated within the pudendal nerve at the penis signal to lumbosacral segments to propagate sym-pathetic outflow to the seminal tract and induce ejaculation. 27 Despite the success of this technique in spinal cord injury patients, 153 it can be unsuccessful in diabetics. 154 This may be because of peripheral neuropathy impairing the sensory and/or motor arms of the spinal reflex arc. Nevertheless, trial of this technique tends to be reserved for those with AE in whom an ejaculated specimen is needed for assisted conception.

Retrograde ejaculation
This is passage of semen from the posterior urethra, where it is deposited during emission backward into the bladder, which can be either complete or partial. Diagnosis is confirmed on detecting spermatozoa 12,19 or fructose 22  mechanisms. 155 This is shown in Figure 4.
Most of the studies evaluating the role of pharmacotherapy for RE are case series and these mostly comprise small patient groups. 1 Moreover, as with PE and DE, the studies focusing on or specific to diabetics are scant. 12,156,157 A summary of these limited case series is shown in Table A2. Two of these case series including diabetic subjects have shown efficacy for pseudoephedrine, either as monotherapy or combined with another agent, in treating partial or complete RE is in the order of 45%-60%. 157,158 A systematic review of all the different pharmacotherapies employed to treat RE has found sympathomimetics achieved an antegrade ejaculate in 28%, whereas for anticholinergics such as imipramine and brompheniramine this was 22%. 7 It is important to note that although insulin can directly improve male reproductive hormone levels in animal models [160][161][162] and sperm parameters in men, 163,164 and there is some evidence for such direct effects of oral anti-hyperglycaemic agents, [165][166][167]   Premature ejaculation is the most common sexual dysfunction in men; hence, treatment for this has been studied more extensively than for the other forms of EjD, with evidence from large robust placebocontrolled RCTs 132-134 and meta-analyses. 135,137 Despite this, these studies are not specific to DM and largely include primary/lifelong PE.

Management of infertility in RE
Retrograde ejaculation and DE/AE are much less common but are more likely to have an organic cause and more likely to cause subfertility.
Unfortunately, owing to their rarity in comparison to PE, data for the treatment of these are limited to case series and non-randomised studies with small sample sizes. Nevertheless, promising results have been reported for the various pharmacotherapeutic options for RE, as well as non-surgical and surgical options for sperm retrieval for ART where RE is compounded by infertility.

CONCLUSION
Although current literature provides a reasonable basic understanding of the pathophysiological mechanisms involved in diabetes mellitusrelated ejaculatory dysfunction, there is a scarcity of robust randomised data investigating the potential causative metabolic factors in diabetics, which may be directly contributing to the ejaculatory dysfunction. These would allow more accurate prediction of diabetics more likely to develop ejaculatory dysfunction and provide more rationale for primary prevention. There is a need for more robust placebo-controlled randomised trials in the investigation of treatments for delayed ejaculation and retrograde ejaculation, with a specific focus on diabetic patients.

Proposed factor influencing ejaculation Study Findings of study
Age EL-Sakka, 82 Majzoub et al. 72 Those >50 years twice as likely to have PE Olamoyegun et al. 83 Failed to find association between age and rate of PE EL-Sakka 82 HbA1c >7 resulted in OR of 9.6 in having PE Glycaemic control Bellastella et al. 86 Significant negative correlation between HbA1c and IELT (r = -0.3; p = 0.01) Olamoyegun et al. 83 Higher fasting blood glucose in PE than those without Fedder et al. 12 Found no difference in HbA1c between diabetics with and without RE High blood glucose variability and LBGI Bellastella et al. 86 Significant positive correlation between LBGI and severity of PE  83 Increased incidence of PE in diabetics with ED than those without (32% vs. 13%) Serum T Khan et al. 118 Significantly higher T in men with DM-related PE than controls (p < 0.001) Owiredu et al. 109 Negative correlation between T and short IELT (PE) in T2DM Corona et al. 104,110 Higher T seen with PE and this decreased with DE a Wu et al. 112 Significantly higher levels of T in those with PE than controls (p < 0.05) a Paduch et al. 113 T replacement was not associated with significant improvement in EjD (DE) in hypogonadal men a Morgantaler et al. 114 Failed to show any association between T and IELT Concomitant depression/anxiety (largely associated with PE) Corona et al. 124 Significant association of severe depressive symptoms and EjD in diabetics Khan et al. 118 Those with DM-related PE have more than ×5 higher depression score on validated questionnaire Abbreviations: DE, delayed ejaculation; DM, diabetes mellitus; ED, erectile dysfunction; IELT, intravaginal ejaculatory latency time; LBGI, low blood glucose index; OR, odds ratio; PE, premature ejaculation; RE, retrograde ejaculation; T, testosterone; T2DM, type 2 diabetes mellitus. a Not specific to diabetic men.