Diabetes, Erectile Dysfunction, and Sleep-Related Erections

Transcription

1 Sleep 13(1):53--{)8, Raven Press, Ltd., New York 1990 Association of Professional Sleep Societies Diabetes, Erectile Dysfunction, and Sleep-Related Erections Max Hirshkowitz, Ismet Karacan, Kevin C. Rando, Robert L. Williams, and James W. Howell Sleep Disorders and Research Center, Department of Psychiatry, Baylor College of Medicine, and Sleep Research Laboratory, Research Service, Veterans Affairs Medical Center, Houston, Texas, U.S.A. Summary: Sleep-related erections were assessed in conjunction with polysomnography in 100 diabetic and 400 nondiabetic men with complaints of erectile problems. We also measured bulbocavernosus reflex latency, heart rate response to deep breathing, postural-related blood pressure changes, penile arterial sufficiency, and brachial blood pressures. To investigate the relationship between diabetes and erectile capacity, the results obtained from men with and without diabetes were compared. Men with diabetes had fewer sleep-related erections, less tumescence time, diminished penile circumference increase, and lower penile rigidity than nondiabetic men. These diabetes-related differences were found regardless of the maximum penile rigidity observed. The diabetic group had less heart rate response to deep breathing and lower penile blood pressures than the nondiabetic group, but only among men with maximum penile rigidity <500 g. These data indicate that both neurological and vascular mechanisms are involved to a greater degree in organic diabetic impotence than in the organic erectile dysfunction that occurs in nondiabetic men. Finally, the pattern of lower values for measures of nocturnal tumescence among diabetic men, compared to nondiabetic men, occurred in all age groups, except the oldest. Among impotent men, age 65 years or older, no difference was found between men with and without diabetes. This suggests that diabetes may foreshadow some of the age-related pathophysiological processes associated with erectile dysfunction. Key Words: Impotence-Sexual dysfunction Sleep-related erections-nocturnal penile tumescence. The manner by which diabetes adversely affects sexual function is a question many investigators have tried to answer. In the past, erectile problems concomitant with diabetes were often treated as a psychological problem. Erectile dysfunction, however, is a complex disorder and can certainly have more than one causal component. In much of the work in this area, clinical evaluation of erectile dysfunction relies on nonobjec- Accepted for publication August Address correspondence and reprint requests to Dr. Max Hirshkowitz at VA Medical Center-Research 151,2002 Holcombe Blvd., Houston, TX 77030, U.S.A. 53

2 54 M. HIRSHKOWITZ ET AL. tive means for diagnosis. Psychological components of erectile dysfunction are troublesome to delineate in any form of impotence. During sleep, penile erections naturally occur approximately every min, usually in conjunction with rapid eye movement (REM) sleep (1,2). By monitoring nocturnal penile tumescence (NPT), a patient's physiological erectile capacity can be assessed in a simple and nonstressful way (3). As a result, findings from NPT monitoring are helpful in differentiating psychogenic from organic impotence (4). Recent studies indicate that reduced NPT may accompany the loss of libido associated with depression in men (5,6). In interpreting this finding, the authors emphasize that, "Whether or not these observations lead to the conclusion that depression is a cause of 'false positive' NPT studies depends on whether one construes an abnormal NPT profile as indicative of an irreversible lesion and/or whether depression is viewed as, at least in part, involving organic processes" (6). Differences in sleep-related erections between patients with diabetes and erectile dysfunction and normal control subjects have been described in a number of studies. Karacan and colleagues (4) compared 35 impotent men with diabetes to 35 age-matched healthy controls. They found fewer NPT episodes, less total tumescence time, and lower maximal circumference increases among men with diabetes. Similar results were reported by Murray and coworkers, who studied 28 diabetic impotent men and 15 normal controls (7), and Hosking and colleagues (8), who studied 30 impotent diabetics and 11 healthy volunteers. Diminished NPT associated with diabetes has also been described in studies comparing impotent men with and without diabetes. Schiavi and Fisher (9) evaluated NPT in 44 subjects, 15 of whom had diabetes. Of the 15,6 had erectile problems. The diabetic impotent group had significantly fewer NPT episodes, less total tumescence time, and fewer minutes of maximal erection than any of the other groups. Ficher and associates (10) recorded NPT in three groups of men: (a) 22 healthy control subjects, (b) 31 nondiabetic impotent men, and (c) 47 impotent men with diabetes. These data were also presented as part of another report (11). Clear differences between the two impotent groups were found for the number, duration, and magnitude of erections on two of the three nights. Upon examination of the data, it appears that on the first recording night, some NPT measures were diminished, especially in the nondiabetic impotent and control group. The values for the men with diabetes remained relatively low from night to night. The authors allude to the "first-night effect," which is known to reduce REM sleep, mention the importance of evaluating sleep stages, and indicate that sleep changes were not found. Quantitative information about sleep, however, is not provided. Nevertheless, the authors conclude that "diabetic dysfunctionals showed significantly weaker NPT response." Maatman and colleagues (12) studied a large number of men with a complaint of erectile dysfunction. Of the 497 men tested, 161 had abnormal glucose metabolism. Moreover, 91% of this subgroup had "organic pattern impotence. " NPT, however, was evaluated with snap gauges; consequently, quantitative data concerning the frequency, magnitude, and duration of sleep-related erections are not available. In current clinical practice, penile rigidity during representative sleep-related erections is perhaps the single most important parameter of NPT. It has been empirically estimated that the average minimal rigidity requisite for vaginal penetration is 500 g (13). Penile rigidity, therefore, provides an index for erectile functionality. Notwith-

3 DIABETES AND ERECTILE DYSFUNCTION 55 standing discussion of the importance of rigidity measurement in several of the abovementioned reports (6,9), rigidity data have not been reported. The purpose of the present study was to replicate previous work in a large sample of men with erectile dysfunction and to further describe sleep-related erection differences between men with and without diabetes who are experiencing erectile difficulties. We provide detailed quantitative information about the NPT phases, its coordination with sleep states, and changes associated with age. Furthermore, we have explored some aspects of the etiology of erectile dysfunction in diabetes and have presented new data concerning penile rigidity. Subjects METHODS All subjects in this sample were male patients with erectile difficulties who were evaluated at our sleep disorders and research center during a 40-month period. Patients were given detailed explanations of the procedures, and written informed consent was obtained. Polysomnography was performed on 500 men to assess their erectile capacity. This group included 100 diabetic and 400 nondiabetic men. The polysomnographic data presented in this paper are calculated from the 999 continuous all-night tracings made during the first and second night that each man slept in the laboratory (one patient slept only a single night). Age of the men with diabetes ranged from 25 to 76 years (mean 54). Nondiabetic men ranged in age from 20 to 76 (mean 53). Age distributions were equivalent for the diabetic and nondiabetic groups, and means did not differ significantly (t = 0.60; p = 0.55). Both diabetic and nondiabetic men had a variety of concurrent medical conditions. Table 1 shows the number of patients with different concurrent conditions and their age (penile rigidity is also shown, but will be discussed in the Results section). Chi-square analysis was used to compare the prevalence of each condition between men with and without diabetes. We found that hypertension, heart disease, stroke, and sleep apnea were more common among impotent men with diabetes than among those without diabetes. In contrast, genital trauma, back injury, and alcoholism occurred more frequently in nondiabetic patients with erectile dysfunction. No significant differences (p > 0.10) in age were found between diabetic and nondiabetic men with concurrent medical conditions. However, diabetic men who had also had a stroke and diabetic men with concomitant sleep apnea were marginally (0.05 < P < 0.10) younger than their nondiabetic counterparts. Men with diabetes and psychiatric disorders were marginally older than nondiabetic men with psychiatric disorders. As our patients had an assortment of concurrent medical problems, a wide variety of medications were in use at the time of study. Table 2 lists the number and percentage of men in the diabetic group and nondiabetic group taking various types of medicines. Except, of course, for insulin, the group percentages are fairly balanced. It is notable, however, that concomitant treatment for hypertension more often included diuretics among diabetic than nondiabetic men, although the combined percentage for treatment with diuretic and other antihypertensives was similar for men with and without diabetes.

4 56 M. HIRSHKOWITZ ET AL. TABLE 1. Concurrent medical conditions for diabetic and nondiabetic men with erectile dysfunction Diabetics Nondiabetics Concurrent condition n Age Q Rigidity" n Age Ridigity Genital trauma 3 d 46 ± ± 132 c ± ± 290 Hernia ± ± ± ± 268 Prostate problems ± ± 297 C ± ± 262 Vasectomy ± ± ± ± 261 Hypertension 54 d 56 ± ± 294 d ± ± 282 Heart disease 31 d 57 ± ± 275 c ± ± 264 Bypass surgery ± ± ± ± 280 Pulmonary disease ± ± 256 d ± ± 265 Stroke 7 c 55 ± 5 b 362 ± ± ± 311 Arthritis 8 57 ± ± ± ± 278 Back injury IO c 59 ± ± ± ± 275 Alcoholism ll C 50 ± ± 250 b ± ± 283 Psychiatric disorder 9 55 ± 5 b 364 ± ± ± 280 Sleep apnea 50 d 55 ± lib 333 ± 249 d ± ± 275 Periodic leg movements ± ± 291 d ± ± 271 Total for entire sample ± ± 281 d ± ± 281 Q Mean ± SD. b P < 0.10 for diabetic versus nondiabetic contrast. c p < 0.05 for diabetic versus nondiabetic contrast. d p < 0.01 for diabetic versus nondiabetic contrast. Neurovascular genital examination procedures In addition to standard clinical interviews, medical history taking, and a thorough physical examination, a special neurovascular genital examination was performed. This examination included assessment of bulbocavernosus reflex latency, heart rate response to deep breathing, supine-to-standing blood pressure change, and penile arterial sufficiency. Bulbocavernosus reflex latency was determined in a standardized manner (14) using a TECA electromyograph. The penis was stimulated electrically at the glans, and the response was recorded with a concentric needle electrode inserted percutaneously directly into the bulbocavernosus muscle. The reflex latency was measured from the evoked response. Autonomic function was assessed by evaluating heart rate response to deep breathing and supine-to-standing systolic blood pressure change. We recorded the duration of each successive R-R interval during 1 min of normal breathing. Patients were then instructed to breath deeply in cadence to a metronome that paced each breath at approximately 10 s. For each 10 successive heartbeats, the longest R-R interval was divided by the shortest R-R interval. The average of this ratio was calculated for periods of normal and deep breathing. To evaluate postural-related blood pressure changes, blood pressures were first auscultated in standard fashion while the patient lay supine on a tilt table. The table was then brought to an upright position and blood pressures were measured again. Penile blood pressures in left dorsal, right dorsal, left cavernosus, and right cavernosus arteries were measured using a doppler system. Ultrasonic measurements of brachial pressure were also made. To calculate penile/brachial index, penile arterial pressures were divided by the brachial pressure and multiplied by 100. Sleep. Vol. 13, No. 1,1990

5 DIABETES AND ERECTILE DYSFUNCTION 57 TABLE 2. Concomitant medications taken by diabetic and nondiabetic patients: general categories of medications, the number of patients taking the medication, and the group percentage Diabetics N ondiabetics Type of drug n and percent n Percent Diabetes treatments Insulin injections Oral Cardiovascular Angina treatments Antiarrhythmics Antihypertensives Diuretics Psychotropics Anxiolytics Antidepressants Antipsychotics Anticonvulsants Hypnotics Analgesics Antibiotics Bronchodilators Gout medications Thyroid supplements Testosterone injections Polysomnographic procedures After reporting to the sleep laboratory, patients had monitoring devices attached and completed a brief presleep questionnaire. Surface electrodes were placed on the scalp and mastoid (C3-A2) to record electroencephalographic activity; at the outer canthi of the left and right eyes to record eye movements; on the chin to record submental electromyographic activity; and on the chest (VI-V5) to record heart rate. Penile circumference change was monitored with mercury-filled strain gauges placed just below the glans at the coronal sulcus and at the base of the penis. Bulbocavernosusischiocavernosus muscle activity and segmental penile blood flow were also recorded in some patients, but will not be reported here [see Karacan et al. (3) for details]. Patients slept in climate-controlled, sound-attenuated, private bedrooms. Continuous all-night recordings were made using Grass Model 78 polygraph systems. Penile circumference was amplified using American Medical Systems or Texas Medical Electronics NPT monitors interfaced to the polygraph. Penile rigidity was measured by awakening the patient during a representative sleep-related erection and applying a gradually increasing force to the tip of the penis. Penile rigidity is the force at which the shaft of the penis buckles. If a force of 1,000 g is reached with no bending of the penile shaft, the value 1,000 is assigned. Polygraph tracings were marked at the time of awakening and when the device was applied. The erect penis was also photographed. The polygraphic record and photograph were used to help validate the rigidity estimate (3). At the end of the sleep period, patients had the monitoring devices removed and completed a short post sleep questionnaire.

6 58 M. HIRSHKOWITZ ET AL. RESULTS The results of the neurovascular genital examination are shown in Table 3. The most robust finding ofthis examination was that, as a group, the men with diabetes had lower penilefbrachial blood pressure index values. We also found that the deep breathing was less successful at inducing heart rate variability in the diabetic than in the nondiabetic group and that supine blood pressures were higher among the men with diabetes. We subdivided the diabetic group and nondiabetic group on the basis of maximum penile rigidity. Of the 100 men with diabetes, 28 had a maximum rigidity of 500 g or more, whereas the remaining 72 had <500 g. Of the 400 nondiabetic patients, 202 had at least one erection with a rigidity of 500 g or more. Included in Table 3 are the results of the neurovascular genital examination for patients with a maximum penile rigidity ;?;500 g and <500 g. When diabetic and nondiabetic men with rigidity <500 g were compared, diabetics had less heart rate response to deep breathing (p = ), marginally greater orthostatic blood pressure drop (p = 0.09), higher systolic blood pressure (p = 0.05), and lower penile/brachial indices (p = ). In contrast, neurovascular genital examination findings for men with rigidity ;?; 500 g differed between diabetic and nondiabetic groups only for systolic blood pressure (p = 0.003). Penile rigidity was calculated for patients with various concurrent medical conditions TABLE 3. Results of neurovascular genital examination performed on men with and without diabetes who had complaints of erectile problems Values for all patients a Neurological Indices Bulbocavernosus reflex latency (msec) Heart rate response to deep breathing Supine-to-standing systolic drop (mm Hg) Blood pressures Brachial systolic blood pressure (mm Hg) Brachial diastolic blood pressure (mm Hg) Average penile/brachial index Patients with rigidity ~500 Neurological indices Bulbocavernosus reflex latency (msec) Heart rate response to deep breathing Supine-to-standing systolic drop (mm Hg) Blood pressures Brachial systolic blood pressure (mm Hg) Brachial diastolic blood pressure (mm Hg) Average penile/bnlchial index Patients with rigidity <500 Neurological indices Bulbocavernosus reflex latency (msec) Heart rate response to deep breathing Supine-to-standing systolic drop (mm Hg) Blood pressures Brachial systolic blood pressure (mm Hg) Brachial diastolic blood pressure (mm Hg) Average penile/brachial index a Mean ± SO. b P < 0.10 for diabetic versus nondiabetic contrast. c p < 0.05 for diabetic versus nondiabetic contrast. d p < 0.01 for diabetic versus nondiabetic contrast. e p < for diabetic versus nondiabetic contrast. Diabetics 40.9 ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± 14.9 Nondiabetics 39.3 ± ± 0.10e 3.3 ± ± 19.8 e 84.2 ± 11.5 c 82.5 ± 12.7e 38.6 ± ± ± ± 18.7 d 83.9 ± ± ± ± O.W 3.2 ± l1.3 b ± 20.7 c 84.4 ± ± 12.6 e

7 DIABETES AND ERECTILE DYSFUNCTION 59 (see Table 1). Men with and without diabetes with each condition were compared. Significantly lower rigidity was found (p < 0.05) for diabetics with genital trauma, prostate problems, hypertension, heart disease, pulmonary disease, sleep apnea, and periodic leg movements in sleep. Diabetic men with alcoholism had marginally lower rigidity than nondiabetic men with alcoholism. Polysomnograms were scored for sleep stage and NPT phase according to standard technique (3,15). Each minute of recording was classified as awake (stage 0), or sleep stage 1,2, 3, 4, or REM. Likewise, each minute was categorized as T-Up (ascending tumescence), T-Max (maximal tumescence), T-Down (descending tumescence), or T Zero (no tumescence). This scoring was done within the context of tumescence episodes to designate the phase of the NPT cycle. Tracings from both gauges were scored in this manner, and each episode's maximum increase was tabulated. Tables 4 and 5 contain the results for the parameters we routinely examine and consider to be of primary importance when evaluating sleep-related erectile activity. The sleep summary variables provide information necessary to validate the procedure and to insure that NPT results are not a secondary effect of sleep differences. Dramatic and consistent differences were found for all tumescence measures between the groups of diabetic and nondiabetic men. Tumescence was markedly diminished among the men with diabetes in terms of frequency (number of episodes), duration (total tumescence time), and magnitude (maximum circumference increase). This was the case for gauges placed at both the coronal sulcus (CS) and the penile base (PB), and for both recording nights. In contrast, the sleep context in which the erections occurred differed slightly between groups only on the second night and only with respect to the number of REM periods (fewer for diabetic group) and REM efficiency (greater for diabetic group). Patients with diabetes had significantly less REM-related NPT per night (17 min at TABLE 4. Polysomnographic values for men with and without diabetes who have a complaint of erectile dysfunction Diabetic Sleep summary variables a Total sleep time (TST) ± 74.2 Number of REM periods 3.3 ± 1.1 REM percentage of TST 18.3 ± 7.5 Minutes per REM period 22.4 ± 10.5 REM efficiency index 87.3 ± 17.6 REM fragmentation index 0.6 ± 0.6 Tumescence summary variables a Number of episodes Coronal sulcus 1.6 ± 1.6 Penile base 1.7 ± 1.6 Maximum increase (mm) Coronal sulcus 7.2 ± 8.9 Penile base 9.6 ± 10.1 Total NPT time (min) Coronal sulcus 33.5 ± 38.6 Penile base 38.5 ± 4l.l Night I Nondiabetic ± ± ± ± ± ± ± 1.8 d 2.6 ± 1.8 d 10.8 ± 9.lc 15.2 ± 10.5 d 58.8 ± 50.ld 68.7 ± 56.9 d Diabetic Night 2 Nondiabetic ± ± ± ± 1.1 b 21.0 ± ± ± ± ± ± l2.4 b 0.7 ± ± ± ± 1.8 d 1.7 ± ± 1.8 d 6.7 ± ± 9.9 d 9.6 ± ± 10.9 d 37.6 ± ± 57.ld 41.0 ± ± 61.6 d a Mean ± SD. b P < 0.05 for diabetic versus nondiabetic contrast. C p < for diabetic versus nondiabetic contrast. d p < for diabetic versus nondiabetic contrast. Sleep. Vol. 13. No

8 60 M. HIRSHKOWITZ ET AL. CS; 19 min at PB) than their nondiabetic counterparts (31 min at CS; 35 min at PB). A similar pattern of non-rem (NREM)-associated tumescence was found when the two groups were compared. Group differences were statistically reliable (p < 0.05) for both REM and NREM comparisons. A differential pattern in the proportion ofnpt in REM and NREM was found between men with rigidity ~500 g and rigidity <500 g. Figure 1 illustrates these results. More base gauge NPT occurred during both REM and NREM in the diabetic group than in the nondiabetic group, regardless of the maximum rigidity. However, NPT recorded from the coronal sulcus and occurring during stage REM sleep and during NREM stages was less for diabetics than nondiabetics only for those men with maximum rigidity <500 g. Durations of CS gauge REM-related NPT and NREMrelated NPT did not differ for men with rigidities of 500 g or more. To examine the internal architecture of tumescence episodes, we calculated the average duration of each phase oftumescence. All phases (T-Up, T-Max, and T-Down) were significantly shorter (p < 0.05) for the diabetic group when the entire sample was considered. However, when the sample was subdivided using the 500-g rigidity criteria, diabetes-related decrements were found only among those men with <500 g rigidity. No diabetes-related decline in average NPT episode phase durations was found for men with rigidity ~500 g. Also, diabetes was associated mostly with reductions in the maintenance phases of tumescence (T-Max and T-Down) and less with the acquisition phase (T-Up). Figure 2 illustates these results. Two analyses were performed for rigidity. In the first, the value 0 was assigned for penile rigidity in patients with no NPT. U sing this correction factor, the mean rigidity was 344 g for men with diabetes and 533 g for men without diabetes. This difference was significant (p < ). Notwithstanding the suggestion that absent NPT implies a rigidity of 0, we performed another analysis that avoided the controversy of this as- TABLE 5. Polysomnographic values for men with and without diabetes who have a complaint of erectile dysfunction. Mean values across nights 1 and 2 are given for patients with a maximum rigidity of 500 g or more and <500 g Sleep summary variables a Total sleep time (TST) Number of REM periods REM percentage of TST Minutes per REM period REM efficiency index REM fragmentation index Tumescence summary variables a Number of episodes Coronal sulcus Penile base Maximum increase (mm) Coronal sulcus Penile base Total NPT time (min) Coronal sulcus Penile base a Mean:!: SD. Diabetic :!: :!: :!: :!: :!: :!: ± :!: :!: :!: :!: :!: 39.7 Rigidity ;;.500 g b P < 0.05 for diabetic versus nondiabetic contrast. c p < 0.01 for diabetic versus nondiabetic contrast. d p < for diabetic versus nondiabetic contrast. Nondiabetic 342.7:!: :!: ± A:!: ± :!: :!: lab 3.2 :!: l.4 b 15.2 ± 8.9 b 20.3 :!: 9.2c 82.1 :!: 46.9 b 91.5 :!: 49.8 c Diabetic :!: ± :!: ± :!: :!: :!: :!: :!: :!: :!: :!: 33.3 Rigidity <500 g Nondiabetic ± :!: ± ± ± ± ± lac 2.0 ± 1.6 c 7.2±6.1b 10.7 ± 7.7 c 41.8 ± 37.7 d 51.9 ± 48.9 d Sleep. Vol. 13. No

9 DIABETES AND ERECTILE DYSFUNCTION 61 Minutes of NPT 80 RIGIDITY? 500 Gms CS Gauge PB Gauge o o RIGIDITY < 500 Gms CS Gauge PB Gauge DrAB NOND DIAB NOND FIG. 1. This figure illustrates the minutes of NPT occurring during REM (_) and NREM (0) sleep in a group of men with diabetes (DIAB) and a nondiabetic group (NOND). Significant reductions between groups are shown (* p < 0.05; ** p < 0.01; *** p < 0.001) for tumescence time recorded at the coronal sulcus (CS) and penile base (PB). sumption. For the men with diabetes who had at least one NPT episode, rigidity ranged from 40 to 1,000 g, with a mean of 465 g. Nondiabetic men with at least one NPT episode had a mean rigidity of 560 g, with a range of 60-1,000g. This difference was also statistically significant (p < 0.004). Figure 3 illustrates penile rigidity differences both across age groupings (Fig. 3A) and for subgrouping according to maximum rigidity (Fig. 3B). It is evident that the difference in rigidity between diabetic and nondiabetic groups principally reflects a difference among patients with impaired rigidity «500 g). The analysis of rigidity by age groupings (Fig. 3A) reveals significant differences between diabetics and nondiabetics in all age groups except the oldest. After age 65, rigidities were equivalent for diabetic and nondiabetic groups. Age group differences in frequency, magnitude, and duration of NPT were evaluated. The following age groups were formed: Group 1: <45 years (16 diabetics and 81 nondiabetics) Group 2: years (29 diabetics and 106 nondiabetics) Group 3: 55-M years (41 diabetics and 155 nondiabetics) Group 4: 65 years or older (14 diabetics, 58 nondiabetics)

10 62 M. HIRSHKOWlTZ ET AL. Minutes of NPT 30 CS Gauge GIDITY > 500 Gms PB Gauge o o RIGIDITY < 500 Gms CS Gauge PB Gauge DlAB NOND DlAB NOND FIG. 2. This figure shows the average durations per episode of the ascending (T-Up ), maximal (T-Max, D), and descending (T-Down, D) phases of sleep-related erections in a group of men with diabetes (DIAB) and a nondiabetic (NOND) group. Significant reductions between groups are shown (* p < 0.05; ** p < 0.01; *** p < 0.001) for tumescence time recorded at the coronal sulcus (CS) and penile base (PB). Figure 4 shows NPT characteristics for men with and without diabetes in the different age groups. The diabetic groups had fewer NPT episodes than the nondiabetic groups at ages 45-54, and Episode frequency did not differ between men with and without diabetes in the oldest and youngest groups. Maximum circumference increase and total tumescence time were greater among nondiabetics than diabetics in all age groups except the oldest. Penile rigidity in men with diabetes was correlated with measures taken during the neurovascular genital examination. Significant correlations were found between rigidity and left dorsal (r = 0.22, p = 0.04) and right dorsal (r = 0.25, p = 0.02) penilelbrachial indices. The correlation between blood pressure response to standing among men with diabetes and corrected rigidity (with zero replacement for no NPT) was marginal (r = , p = 0.06); however, uncorrected rigidities and blood pressure drop correlated significantly (r = , p = 0.04). In men with diabetes and erectile dysfunction, no relationships were found between rigidity and bulbocavernosus reflex latency, rigidity and heart rate response to deep breathing, or rigidity and cavernosus penilelbrachial blood pressure indices.

11 DIABETES AND ERECTILE DYSFUNCTION QlI... ~ 400 '" 'Sn 0:: QlI , >.... :s 400.~ 0:: A B " : <45 Total Group >64 Age Group (Years) p=.2527 Rig >499 Subgroup Rig <500 Subgroup FIG. 3. Penile rigidity for men with and without diabetes. (A) Mean rigidity for diabetic (0) and nondiabetic (@) men subgrouped by age. Significantly lower rigidity (p < 0.05) was found in men with diabetes in all age groups, except the oldest. (B) The overall group rigidity (mean ± SE) for men with (0) and without (13) diabetes and for subgroups with maximum rigidity ;;,500 and <500 g. DISCUSSION The results of the present study clearly indicate that erectile impotence associated with diabetes has a different profile than erectile dysfunction in nondiabetic men. These findings also support the conclusions we reached in earlier papers on this subject (4,16,17). It should be noted that there is no overlap between the sample used in these previous reports and the data presented here. We have shown that the impaired NPT found in men with diabetes is not the result of disrupted sleep but relates directly to the underlying, presumably organic erectile dysfunction. We also found that diabetes in conjunction with other medical conditions is associated with greater impairment in the quality of sleep-related erections. Diabetics with prostate problems, hypertension, heart disease, pulmonary disease, sleep apnea, or periodic leg movements in sleep have lower maximum rigidities than nondiabetic men with these conditions. Additionally, the prevalence of hypertension, heart disease, and sleep apnea is greater among men with diabetes and erectile dysfunction than among nondiabetic impotent men. The frequency, magnitude, and duration of sleep-related erections are less for men Sleep.Yol. 13. No

12 64 M. HIRSHKOWITZ ET AL. Number of Episodes o Maxim um Increase (mm) 20~ , _--4)... e ,... B _- ~----~~ o -... ~ o ~--~ ~ ~ ~~ Minutes of NPT _... ~ G EJ. -" :l., ---~ t=j... ' Q "15 o~--~------~ ~------~~~ < >64 Age Group (Years) FIG. 4. Frequency, magnitude, and duration of NPT for groups of impotent men with and without diabetes as a function of age. Mean values across nights 1 and 2 are shown for penile base (PB) and coronal sulcus (CS) recordings. Nondiabetic, PB gauge:o- -0; nondiabetic, CS gauge:o- -0; diabetic, PB gauge:o-o; diabetic, CS gauge:d-o. with diabetes than for those without diabetes. This relationship exists even among those men with erectile quality adequate for penetration (rigidity ~500 g). Similarly, among patients with rigidity <500 g, diabetic men have fewer NPT episodes, smaller maximum increases, and fewer minutes of tumescence than nondiabetic men. Although these overall summary measures of NPT show the same diabetes-associated reduction for subgroups with rigidity ~500 and <500 g, there are differences in the architecture of NPT episodes and sleep stage associations. Individual phases of the sleep-related erection (T-Up, T-Max, and T-Down) differ between men with and without diabetes only within the reduced rigidity subgroup. Additionally, diabetics with <500 g rigidity have fewer minutes of NPT during REM and during NREM than nondiabetics for measures made at both the coronal sulcus and penile base. In contrast, diabetics with rigidity ~500 g had less REM-associated and NREM-associated NPT than nondiabetics

13 DIABETES AND ERECTILE DYSFUNCTION 65 at the base, but not at the coronal sulcus. The implication of this finding is unclear at present and therefore requires further investigation. It is apparent from the data, however, that men with diabetes with impaired rigidity are more severely impaired than their nondiabetic counterparts. The NPT results in the present study are consistent with those of previous reports (4,7-11). NPT measures of frequency, duration, and magnitude were all lower in impotent patients with diabetes. In previous investigations led by Karacan (4), Hosking (8), and Murray (7), impotent men with diabetes were compared to normal control subjects. In this study, as in those by Schiavi and Fisher (9) and Ficher et al. (10), impotent patients with and without diabetes were compared directly. The consistency of the NPT effects, despite vast age differences, is most remarkable. In Schiavi and Fisher's study, the average age was approximately 28 years, in that of Hosking et al. it was 33.5, and in the investigations lead by Ficher and Zuckerman it was 50. The average age in the present study was 53. Given our large sample, we were able to compare men with and without diabetes in different age groups. In general, men with diabetes had lower values on NPT measures than nondiabetic men in all age groups, except the oldest. In the group age 65 or older, the age-related reductions in NPT among nondiabetic men converged with the consistently reduced NPT seen across age groups among diabetics. This suggests that the effect of diabetes on the systems underlying erectile function is sufficient to produce changes regardless of age until some minimal level is reached. It is known that NPT parameters naturally change as a function of age (18); however, diabetes apparently foreshadows these changes. The contribution of the autonomic nervous system to erectile dysfunction associated with diabetes can be assessed with the Val salva maneuvre, heart rate response to deep breathing, and systolic blood pressure response to standing (19,20). Hosking and associates (8) evaluated autonomic function in men with diabetes and erectile dysfunction using all three procedures. They report autonomic dysfunction in 5 of the 15 patients who complained of "total" impotence. Cavan and coworkers (21) had patients perform the Valsalva maneuvre and a deep breathing test. Of the 45 impotent patients with diabetes tested, 9 had abnormal responses on both tests and 14 had an abnormal response on one test or the other. By contrast, none of the 45 potent patients with diabetes had abnormal results on both tests. These researchers conclude that diabetic impotence is associated with "definite neuropathy" in some patients. In the present study, we found significantly less heart rate response to deep breathing among men with diabetes. Reduced variation in heart rate has been reported to reflect early changes in peripheral autonomic neuropathy in men with diabetes (19,20). If calculations are based on the longest/shortest R-R interval, the expected normal value is ~ 1.20 or greater. Loss of heart rate variability reflects parasympathetic compromise, and in men with diabetes, this abnormal response may precede other signs of systemic pathology, neurovascular pathology, or both. Interestingly, in our study, reduced heart rate variability among men with diabetes was found only in the group with impaired rigidity. No R-R ratio difference was found in men with reduced NPT but with nocturnal penile rigidity ;;::500 g. Therefore, this parasympathetic effect does not precede decrements in sleep-related erections. Other evidence for neurological involvement in diabetic impotence has been published. Vacek and Lachman (22) found the bulbocavernosus reflex had a longer latency in impotent men with diabetes than in control subjects. This finding has been supported

14 66 M. HIRSHKOWITZ ET AL. by the work of other investigators (23). In the present study, no bulbocavernosus reflex latency prolongation was found. This may be due to the fact that the comparisons in this study were made between two patient groups, both of whom were impotent. Another major component of the neurovascular mechanisms governing erections in man involves penile hemodynamics. Notwithstanding the possible neurological control systems, a functional vascular substrate is essential for penile erection and rigidification. Pelvic or penile vascular insufficiency has been reported previously, both by us and others, in investigations of erectile dysfunction in diabetic men (24-27). In this study, we found clear evidence for vascular insufficiency associated with diabetic impotence, but only for those patients with impaired rigidity. Penile/brachial blood pressure indices were lower among men with diabetes who had rigidity <500 g than among nondiabetics with comparable rigidity. All of the arteries examined (left dorsal, right dorsal, left cavernosus, and right cavernosus) were affected. Although the penilelbrachial index means for both groups fell within the borderline range (65-91), group differences were highly significant (p < 0.001). These data suggest that a compromised penile vascular system is characteristic of organic diabetic impotence. Rigidity was more impaired among diabetic men than among nondiabetic men with maximal rigidity <500 g. Additionally, dorsal penilelbrachial blood pressure indices correlated significantly with penile rigidity in men with diabetes, suggesting that vascular factors are involved in the process of rigidification during penile erection. Lehman and Jacobs (26) studied 31 men with diabetes and sexual dysfunction and observed vascular problems in 68% of patients with neurological abnormalities. In contrast, only 28% of patients with vascular insufficiencies had abnormal pudendal nerve latencies. They contend that vascular occlusion is the most prevalent problem in diabetic impotence and that it may even predate neurological involvement. Over the entire sample, all measures of sleep-related tumescence were affected by diabetes, both during REM and NREM sleep. This result demonstrates that a sleep stage-specific dysfunction is not present. Likewise, the durations of all phases of sleeprelated erections were diminished in men with diabetes. From a clinical standpoint, it is important to emphasize that penile buckling force rigidity is perhaps the single most important measure of erectile capacity. It is a direct measure of whether or not an erection is adequate to achieve penetration. Typically, if no erection occurs, rigidity is assigned the value O. In this study, we found significantly diminished penile rigidity during NPT in men with diabetes. We also analyzed rigidity using only patients with at least one episode of tumescence-thus, no assignment of 0 rigidity. In evaluating rigidity, removal from the sample of those patients with the most deficit (no NPT), who are overrepresented in the diabetic group, biases the data against finding a difference. Nonetheless, buckling force rigidity was lower for men with diabetes than for those without diabetes. Maximum penile rigidity during sleep-related erections remained sensitive to organic erectile problems despite the inherent sampling problem that stacks the statistical deck against obtaining a significant difference. The analyses presented in this article provide some new information concerning the relationship between diabetes and erectile dysfunction. The use of a large sample, collection of quantitative NPT measures (from both the penile base and coronal sulcus), replication of NPT results from one night to the next, and the complimentary evaluation of autonomic, vascular, and sacral-spinal reflex parameters during the awake state allow us greater confidence in our results. The interrelationship between diabetes and

15 DIABETES AND ERECTILE DYSFUNCTION 67 erectile dysfunction is not simple; our findings support a multifactor explanation that involves the autonomic nervous system and the penile vascular system. Additional careful work concerning the contribution of autonomic nervous system dysfunction is one area of immediate concern and interest. As our knowledge of the neurovascular mechanisms of erectile hemodynamic process increases and advances in recording technology are made, new opportunities to understand the interrelationship between diabetes and impotence will emerge. Diabetes-related decrements in measures of sleeprelated erections occur in subgroups with rigidity adequate for penetration in whom diurnal neurovascular tests do not distinguish diabetic from nondiabetic men. Therefore, NPT provides a uniquely sensitive opportunity to investigate subtle and presumably early changes associated with the pathophysiology of erectile dysfunction in diabetes. REFERENCES l. Fisher C, Gross J, Zuch J. Cycle of penile erection synchronous with dreaming. Arch Gen Psychiatry 1965;12: Karacan I, Goodenough DR, Shapiro A, Starker S. Erection cycle during sleep in relation to dream anxiety. Arch Gen Psychiatry 1966;15: Karacan I. Evaluation of nocturnal penile tumescence and impotence. In: Guilleminault C, ed. Sleeping and waking disorders: indications and techniques. Menlo Park, CA: Addison-Wesley, 1982: Karacan I, Salis PJ, Ware JC, et ai. Nocturnal penile tumescence and diagnosis in diabetic impotence. Am J Psychiatry 1978;135: Thase ME, Reynolds CF, Glanz LM, Jennings JR, Sewitch DE, Kupfer OJ, Frank E. Nocturnal penile tumescence in depressed men. Am J Psychiatry 1987;144: Thase ME, Reynolds CF, Jennings JR, et al. Nocturnal penile tumescence is diminished in depressed men. Bioi Psychiatry 1988;24: Murray FT, Wyss HU, Thomas RG, Spevack M, Glaros AG. Gonadal dysfunction in diabetic men with organic impotence. J Clin Endocrinol Metab 1987;65: Hosking OJ, Bennett T, Hampton JR, Evans OF, Clark AJ, Robertson G. Diabetic impotence: studies of nocturnal erection during REM sleep. Br Med J 1979;2: Schiavi RC, Fisher C. Assessment of diabetic impotence: measurement of nocturnal erections. Clin Endocrinol Metab 1982;11: Ficher M, Zuckerman M, Fishkin RE, et al. Do endocrines play an etiological role in diabetic and nondiabetic sexual dysfunction? J AndroI1984;5: Zuckerman M, Neeb M, Ficher M, et al. Nocturnal penile tumescence and penile response in the waking state in diabetic and nondiabetic sexual dysfunctionals. Arch Sexual Behav 1985;14: Maatman TJ, Montague OK, Martin LM. Erectile dysfunction in men with diabetes mellitus. Urology 1987;29: Karacan I, Moore CA, Sahmay S. Measurement of pressure necessary for vaginal penetration. Sleep Res 1985;14: Ertekin C, Reel F. Bulbocavernosus reflex in normal men and in patients with neurogenic bladder and/or impotence. J Neurol Sci 1976;28: Rechtschaffen A, Kales A. A manual of standardized terminology, techniques and scoring system for sleep stages in human subjects. NIH publication no Washington, DC: US Government Printing Office, Karacan I, Scott FB, Salis PJ, Attia SL, Ware JC, Altinel A, Williams RL. Nocturnal erections, differential diagnosis of impotence and diabetes. Bioi Psychiatry 1977;12: Karacan I, Ware JC, Dervent B, et al. Impotence and blood pressure in the flaccid penis: relationship to nocturnal penile tumescence. Sleep 1978;1: Karacan I, Salis PJ, Thornby n, Williams RL. The ontogeny of nocturnal penile tumescence. Waking Sleep 1976;1 : Clarke BF, Ewing OJ, Campbell IW. Diabetic autonomic neuropathy. Diabetologia 1979;17: Clarke BF, Ewing DJ. Cardiovascular reflex tests in the natural history of diabetic autonomic neuropathy. NY State J Med 1982;82: Cavan DA, Barnett AH, Leatherdale BA. Diabetic impotence: risk factors in a clinic population. Diabetes Res 1987;5:

16 68 M. HIRSHKOWITZ ET AL. 22. Vacek J, Lachman M. Bulbokavernozni reflex u diabetiku s poruchou erecktivity: klinicka a elektromyograficka studic. Cas Lek Cesk 1977;33: Sarica Y, Karacan I. Bulbocavernosus reflex to somatic and visceral nerve stimulation in normal subjects and in diabetics with erectile impotence. J Urol 1987;138: Jevitch MJ, Edson M, Jarman WD, Herrera HH. Vascular factor in erectile failure among diabetics. Urology 1982;19: Buvat J, Lemaire A, Buvat-Herbaut M, Guieu JD, Bailleul JP, Fossati P. Comparative investigations in 26 impotent and 26 nonimpotent diabetic patients. J Urol 1985;133: Lehman TP, Jacobs JA. Etiology of diabetic impotence. J Urol 1983;129: Karacan I. Diagnosis of erectile impotence in diabetes mellitus. Ann Intern Med 1980;92(part 2):334-7.