- Email: [email protected]
Discrimination between psychogenic and organic erectile dysfunction
Joumal of Psychosomrrc Printed in Great Bnfain.
Research, Vol. 37, No. 2, pp. 135-145.
1993. 0
DISCRIMINATION
0022-3999/93 $6.00+.00 1993 Pergamon Press Ltd
BETWEEN PSYCHOGENIC AND ORGANIC ERECTILE DYSFUNCTION
ANNE E. M. SPECKENS,* MICHIEL W. HENGEVELD,* Guus
A. B. LYCKLAMA A NrJEHor_-r,t
ALBERT M. VAN
HEMERT* and
KEITH E. HAWTON$ (Received 25 February 1992; accepted in revised form 15 July
1992)
Abstract-The aim of this study was to develop a screening test based on the Leiden Impotence Questionnaire (LIQ) in order to assist in the difficult process of differentiating between psychogenic and organic erectile dysfunction (ED). The main sample consisted of 176 patients with ED, which was classified according to the results of the urological and psychiatric assessment as either organic, in 109 (62%) patients, or psychogenic, in 67 (38%) patients. A logistic regression model including six general items from the LIQ correctly identified psychogenic ED in 62 % of the cases, and organic ED in 86%) with an overall correct classification rate of 76%. Adding information regarding sexual intercourse and the relationship in patients who had a partner and were having sexual intercourse the correct classification rates were: psychogenic 77 %, organic 94%) and overall 87 % Discrimination between psychogenic and organic ED is improved when more information concerning sexual activity can be assessed.
INTRODUCTION ERECTILE dysfunction (ED) is a common problem among men in the general population and in medical settings [ 1,2] . Until recently the vast majority of patients who presented with ED were thought to have a psychogenic disorder. Although prevalence estimates have probably been greatly influenced by the clinical setting in which they occur [ 31 , it appears that organic factors are relevant in many more patients than was believed previously [4] . Recent advances in both psychological and medical treatments for ED have made the accurate differential diagnosis of psychogenic and organic contributions to aetiology increasingly important. The most widely accepted of the diagnostic techniques used to distinguish between organic and psychogenic ED is nocturnal penile tumescence monitoring (NPT). Despite its widespread use, some investigators have raised questions about its diagnostic utility and accuracy. In a review in which the several technical and practical problems which could affect the use and interpretation of NPT results were discussed, Meisler and Carey [5] concluded that NPT may misdiagnose as many as 20% of patients and that more sophisticated validation studies are needed. In addition, NPT is a time-consuming, inconvenient and expensive procedure that requires facilities not available in most centres. In clinical practice its use might best be reserved for cases in which there are clear indications of a likely physical aetiology.
*Department of Psychiatry, University Hospital Leiden, The Netherlands. TDepartment of Urology, University Hospital Leiden, The Netherlands. SUniversity Department of Psychiatry, Warneford Hospital, Oxford, U.K. Address for correspondence: Anne E. M. Speckens, Department of Psychiatry, Leiden, Bl-P, P.O. Box 9600, 2300 RC Leiden, The Netherlands. 135
University
Hospital
136
A. E. M.
SPECKENS
et al
These factors have led to the search for more cost-efficient methods to assist in the discrimination between psychogenic and organic erectile problems. Initially, several studies focused on personality variables. Beutler and colleagues [6] described the use of the Minnesota Multiphasic Personality Inventory (MMPI) and Male Impotence Test (MIT) with 22 impotent men whose diagnosis was confirmed by NPT. Although the ability of the MIT to identify biogenic and psychogenic ED was quite low, the MMPI appropriately classified 90% of the cases. However, subsequent studies concluded that the MMPI and the decision rules Beutler and colleagues derived from MMPI scores were ineffective in differentiating between functional and organic ED [7-121. Camic [ 131 examined the Millon Behavioral Health Inventory (MBHI) scores of 40 impotent subjects and 40 control subjects and developed two discriminate functions which differentiated the organic, psychogenic subjects and normal controls with an overall correct classification rate of 81%. In a subsequent study, following discriminant analysis of the MBHI scores in 70 male patients, the proportion of correctly classified patients dropped to 47% in the psychogenic, 65% in the mixed, and 71% in the organic group [ 141. Another approach to the problem has been to assess sexual characteristics that might differentiate organic from psychogenic ED. Derogatis and colleagues [ 151 investigated the efficiency of the Derogatis Sexual Functioning Inventory (DSFI) in 28 impotent males whose diagnosis was established by clinical judgment. Relying on the Gender Role Definition subtest of the DSFI, they reported the overall correct classification rate as 89%. However, later studies failed to replicate these findings [ 16,171 , and in a recent study in 33 diabetic patients the Florida Sexual History Questionnaire did not succeed in discriminating between organic and psychogenic ED [18]. More encouraging results have been reported with questionnaires that closely examine the sexual symptomatology of ED. In a study of 42 male patients, 6 out of 80 sexual history questions effectively differentiated men with organic ED from men with psychogenic ED [ 191 and in 60 males with diabetes mellitus seven questions significantly discriminated between men with differing aetiologies to their erectile problems [20] . From a study of 32 patients, Segraves and colleagues [21] concluded that three aspects of symptomatology effectively discriminated between those with organic and those with psychogenic ED, namely the presence of adequate early morning erections, masturbatory erections and non-coital erections. Hatch and colleagues [22] examined the discriminative validity of information collected during a clinical interview with 56 patients and found that this correctly classified all of them into organic and psychogenic groups. Increasingly clinicians are recognizing that many cases of ED are not either entirely psychogenic or entirely organic, but that both psychogenic and organic factors may be relevant in the individual case [23] . Therefore the performance of possible discriminating measures is bound to be limited. Nevertheless, it is important when possible to determine the main aetiological influence early in the assessment to guide the use of specific diagnostic tests. In addition, with the availability of effective psychological and physical methods of treatment this may be of assistence in the formulation of a treatment plan. We have conducted a study of a large sample of men with ED using a questionnaire which assesses sexual functioning, particularly specific sexual symptoms. The aim was to compare the discrimination between
Erectile dysfunction
137
primarily psychogenic and primarily organic ED of this paper and pencil test with a classification based on an extensive urological and psychiatric evaluation. METHOD Study population
Subjects eligible for the study consisted of consecutive patients with ED who were either referred to the Urology Clinic (84.4%) or to the Sexual Dysfunction Clinic in the Department of Psychiatry (16.6%) of Leiden University Hospital between April 1985 and November 1988. For the purposes of the study, all patients underwent assessment in both clinics. Some patients were referred in order to evaluate their appropriateness for penile prosthesis surgery (13.3 %) or to receive psychosexual treatment (13.3%). However, most patients (73.3%) were evaluated for diagnostic reasons. Physical
assessment
including those physical diagnostic tests thought All patients underwent a urological assessment, necessary by the urologist to assess in what extent organic factors contributed to the aetiology of the ED. Although in this way a full investigation did not take place in the whole population, all patients had a basic assessment consisting of a medical history, a physical examination of the external genitals and prostate and measurement of blood pressure, peripheral pulses and bulbocavernosus-reflex. Laboratory studies included full blood count, glucose, blood urea and nitrogen, electrolytes, liver and cardiac enzymes, serum testosterone, FSH, LH and serum prolactin. Based on the results of penile blood pressure measurements the penile brachial index (PBI) was calculated [24]. In the first 60 patients, nocturnal penile tumescence was measured with an erectometer [25] In subsequent patients, the tumescence and rigidity of penile erections were recorded utilizing Rigiscan assessment during visual sexual stimulation in combination with an intracavernous injection with papaverine [26]. If required, nocturnal penile erections were recorded at the patient’s home with the Rigiscan. In addition, nerve conduct studies, cavemosography, cavemosometry and selective angiography were conducted if indicated. Psychiatric
and psychological
assessments
The patient and his partner (if available) were interviewed separately and together by a psychiatrist. The psychiatric assessment included an extensive psychosexual history. The I.&den Imporence Questionnaire (LZQ) was administered. This was developed by the second author as an instrument to assess sexual function in patients with ED and their partners. Items of interest were selected on the basis of previous research and clinical experience. The LIQ consists of 72 items which are answered on 3-5-point Likert scales. The areas covered include: current sexual functioning (17 items), sexual functioning before the onset of the ED (19 items), aetiology of the ED (5 items), consequences of the ED (16 items) and expectations with regard to treatment options (15 items). Copies of the full version of the LIQ are available from the authors on request. The Sexual Experience Scale [ 271, Body Image Scale of the Derogatis Sexual Functioning Inventory [ 151, Zilbergeld’s Myths Questionnaire [28] and Symptom Check List-90 [29] were also administered. However, results regarding the psychological tests other than the LIQ will not be discussed further in this paper. Class#cation
of patients
On the basis of the results of all the findings of the multidisciplinary evaluation, patients were classified by one of the urologists and a psychiatrist (MWH) into four aetiological groups: (i) probable organic ED-a medical cause could be established for the ED and relevant psychogenic factors were absent; (ii) probable psychogenic ED-no organic cause could be found by the physician and psychogenic factors of sufficient severity to account for the onset of ED were present; (iii) mixed organic and psychogenic ED-organic and psychogenic factors were both significant; and (iv) ED of uncertain aetiology-neither organic nor psychogenic factors relevant to the ED could be found. The last group also included those subjects who did not comply sufficiently with the evaluation to allow a definite diagnosis. Patients in the fourth category were excluded from further analyses. Statistical
analyses
Prior to the statistical tests the items in the LIQ were dichotomized. Individual items were examined according to their relative frequency in the organic and psychogenic ED groups using the chi-square test. As it is known for dichotomous independent variables the discriminant function estimators will overestimate the magnitude of the association, categorical modeling with logistic regression analysis was used to assess the independent contribution of each of the individual items to the discrimination between
138
A. E. M. SPECKENS et (11
the diagnostic groups [ 301 In order to restrict the number of missing values inherent in a questionnaire concerning different types of sexual behaviour, the items were divided into six clusters: general, orgasm, masturbation, non-coital sex, sexual intercourse and the patient’s relationship. Data in each cluster were introduced into the logistic regression model only for patients in whom the particular cluster applied. For example, whereas logistic regression analysis of the general items was conducted in the whole sample, items concerning sexual intercourse were only included for those subjects who reported having sexual intercourse. To establish the discriminating power of the final logistic models a predicted probability of caseness was calculated for all subjects. This score was calculated as follows: predicted probability of caseness = l/(1 + em’) where A represents the sum of the constant and the regression coefficients of the selected items which are present in the individual. Subsequently, a cut-off value was chosen that maximized the correct classification rate. At this cut-off level were recorded: (a) the percentage of cases correctly classified as psychogenic; (b) the percentage of cases correctly classified as organic; and (c) the overall correct classification. The performance of the discriminatory items of the LIQ in the group of patients with apparent mixed aetiology was examined in a subsidary analysis.
RESULTS
Subjects Overall, 217 patients with ED were entered into the study. Eight patients failed to complete the LIQ and were therefore excluded. According to the multidisciplinary evaluation of the remaining 209 patients, in 109 (52%) the ED was considered to have an organic aetiology, 67 (32%) were assigned to the psychogenic group, 22 (11%) were classified as mixed organic and psychogenic and in 11 (5 %) patients the aetiology of the ED was uncertain. The patients in the last category were excluded from further analyses. The sample included in the main analysis thus consisted of 176 patients, 109 (62%) patients in the organic group and 67 (38%) patients considered to have a psychogenic aetiology. The patients with organic ED were significantly older than the patients with psychogenic ED, but the two groups were similar with regard to the proportions who had partners and duration of the ED (Table I). TABLE
I.-CHARACIERISTICS
OF WE
Organic ED (N = 109) Age: Mean (SD) Having a partner Duration ED: Mean (so)
52.5
(13.0) yr 81% 6.9 (7.8) yr
SAMPLE
Psychogenic ED (N = 67) 44.6
(11.7) yr 80% 5.7 (7.4) yr
p < o.ot* N.S. N.S.*
*r-test
Leiden Impotence
Questionnaire
According to the chi-square analysis, 17 of the 72 items of the LIQ discriminated between the psychogenic and organic groups. They are listed in Table II, together with two non-significant items (mode of onset and thoughts about separation) which proved to be relevant in the multivariate analysis. Logistic regression analysis was used to determine the independent contributions of the items to the distinction between the two groups. The item concerning ED with a different partner was omitted because few patients were able to answer this question. Applying logistic regression analysis to the whole sample, six items appeared
Erectile TABLE
II.-LEIDEN
IMPOTENCE
SUBJECTS
Items (shown in the LIQ)
according
WIT”
dysfunction
QUESTIONNAIRE ORGANIC
to groupings
AND
(LIQ): THOSE
of questions
General questions 1. Financial problems Current sexual functioning 2. Sudden onset of ED 3 Presence morning erections 4. Rigidity morning erections > =50% of previous normal erections 5. Rigidity masturbatory erections > =50% of previous normal erections 6. Rigidity coital erections > =50% of previous normal erections 7. Adequate erections with different partner 8. Perceived contractions during orgasm 9. Premature ejaculation Sexual functioning before onset of ED 10. Frequency coitus > l/month 11. Duration coitus > 10 minutes 12. Perceived contractions during orgasm Consequences of ED 13. Reduced size of penis 14. Decline in sexual interest 15. Partner regards patient as not being a real man 16. Tension in relationship 17. Thoughts about separation Patient’s opinion about the cause of ED 18. Psychogenic aetiology 19. Associated with psychological problems *Chi-square
139 ITEMS
WITH
DIFFERENTIATING
PSYCHOGENIC
BETWEEN
ED
Organic ED N= 109 (%)
Psychogenic ED N=67 (%)
P*
42 (40)
37 (56)
< 0.05
31 (31) 76 (72)
28 (42) 59 (91)
N.S. < 0.01
40 (53)
44 (75)
< 0.01
32 (39)
36 (71)
< 0.001
41 3 41 35
40 7 13 38
< < < <
(44) ( 9) (46) (45)
(70) (33) (24) (69)
0.01 0.05 0.01 0.01
88 (90) 22 (23) 61 (67)
45 (74) 25 (40) 23 (41)
< 0.01 < 0.05 < 0.01
69 40 38 34 23
30 37 33 30 9
(46) (58) (58) (52) (17)
< 0.05
35 (54) 45 (68)
< 0.01 < 0.001
(65) (39) (38) (33) (22)
30 (29) 38 (36)
< 0.05 < 0.05 < 0.05 N.S.
tests.
to contribute significantly to the discrimination between patients with psychogenic ED and those with organic ED. The most important discriminators were the presence (item 3) and rigidity of morning erections (item 4). In addition, a decline in sexual interest (item 14) and a belief that the ED was associuted withpsychologicul problems (item 19) were related to psychogenic ED. Perceived contractions during orgasm before the onset of ED (item 12) and reduction in penile size (item 13) were associated with organic ED. In contrast to organic ED, psychogenic ED tended to have a sudden onset (item 2), but this difference did not reach significance. Consequently, logistic regression analysis was applied to sub-samples of patients for which a particular cluster of items was relevant. In the group of 106 patients reporting masturbation the rigidity of the erection during masturbation (item 5) independently discriminated between psychogenic and organic ED. Items concerning non-coital sex did not differentiate between the two groups. In the 131 patients having sexual intercourse, premature ejaculation (item 9) and a lower coitul fiequency before the onset of the ED (item 10) were associated with psychogenic ED. In the 139 patients who had a partner, tension in the relationship (item 16) discriminated between psychogenic and organic ED. We did not include perceived contractions during orgasm before onset of the ED in the final models, because the basis on which this item differentiated between the two groups was not apparent. The nature of onset of ED was included because this
140
A. E. M.
TABLE
III.-MULTIVARIATE
LOGISTIC
SPECKENS
REGRESSION
VARIOUS
Models
COEFFICIENTS
2
- 1.545 0.537 2.039** 1.300 1.088** 1.256** -0.882*
ACCORDING
TO
THE
MODELS
1
Constant Onset Presence morning erections Rigidity morning erections Decline sexual interest Psychogenic aetiology Reduced penile size Rigidity masturbatory erections Tension relationship Thoughts separation Premature ejaculation Coital frequency before ED
et ul.
-
3
I. 736 0.739 2.584** 1.910 0.975 1.224*
-0.932 0.869
4
5
_ -2.033
-2.056 0.875 2.533** 1.593 0.717 1.218* -0.983*
2.223* 1.881 0.931 2.345** - I .608**
-3.284 1.689* 3.120* 2.618* 0.471 2.301** - 1.725*
1.581* 2.082*
0.964 -0.348 I .787* 2.202*
I. 192*
0.958 -1.151
*p < 0.05; **p < 0.01. finding has support from some previous research [ 221. The constant and the logistic regression coefficients of the items as analysed in the final logistic models are provided in Table III. On the basis of these models a predicted probability of caseness was calculated for all subjects. Consequently, cut-off levels were chosen which maximized the correct classification rate. Table IV shows the cut-off levels and the corresponding percentages of cases correctly classified as psychogenic or organic, and overall, when different clusters were entered into the analysis. A mode1 using the basic cluster of symptoms, namely presence and rigidity of morning erections, decline in sexual interest, perceived connection of the ED with psychological problems, reduction in size of the penis and gradual or sudden onset of the ED (cluster l), correctly classified 62% of the psychogenic and 86% of the organic cases, with an overall correct classification rate of 76%. If only the two items, presence and rigidity of morning erections, were used, the correct classification rates were: psychogenic 68 % , organic 62 %, and overall 64%. TABLE
IV.-DISCRIMINATION
ED
BETWEEN ACCORDING
TO
GROUPS
OF
ANALYSES
PATIENTS
WITH
WITH
DIFFERENT
PSYCHOGENIC
% (IV) of cases correctly Clusters 1. Gradual Presence Rigidity Decline ED due Reduced
Organic or sudden onset of ED of morning erections of morning erections in sexual interest to psychological problems size of penis
2. Rigidity of masturbatory (+ cluster 1)
erections
3. Tension in relationship Thoughts about separation (+ cluster I) 4. Premature ejaculation Coital frequency before (+ cluster 1) 5. Cluster
1 + 3 + 4
Psychogenic
AND
ORGANIC
CLUSTERS
classified
Overall
Cut-off
86%
(92)
62%
(60)
76%
(152)
0.51
79%
(61)
80%
(40)
79%
(101)
0.42
82%
(76)
74%
(47)
79%
(123)
0.43
85%
(59)
83%
(47)
84%
(106)
0.51
94%
(54)
77%
(39)
87%
( 93)
0.56
onset of ED
level
Erectile dysfunction
141
Adding the rigidity of erections during masturbation in those who reported masturbation (cluster 2), the percentage of cases correctly classified as psychogenic improved over that in the initial model, although the overall correct classification rate barely changed (Table IV). The same was found when tension in the relationship and thoughts about separation in those with a regular partner (cluster 3) were introduced into the basic model (cluster 1). However, including premature ejaculation and coital frequency before the onset of ED in patients having sexual intercourse (cluster 4), both the percentage of cases correctly classified as psychogenic and the overall correct classification rate substantially improved. The best overall classification rate was produced when items concerning sexual intercourse (cluster 4) as well as the relationship items (cluster 3) in those for whom both clusters were relevant were added to the basic model (cluster l), with a particularly high rate of correct classification of the organic cases. Adding items concerning masturbation did not significantly improve the overall correct classification. Applying the basic model (cluster 1) to the 22 patients with ED based on a mixed aetiology, 11 patients would have been classified as having organic and 7 as having psychogenic ED. DISCUSSION
The aim of this study was to compose a screening test from a questionnaire which assesses sexual functioning in patients with ED in order to assist in the discrimination between psychogenic and organic ED. The full list of questions from the Leiden Impotence Questionnaire for which statistical differences were found between the psychogenic and organic ED groups on the basis of the logistic regression analysis is presented in the Appendix. A logistic regression model which included seven general items from the LIQ correctly identified aetiology of ED in 76 % of the cases, with the percentages correctly classified as psychogenic and organic being 62% and 86% respectively. Adding items regarding sexual intercourse and the relationship in patients who had a partner and were having sexual intercourse resulted in 77% of cases being correctly classified as psychogenic, 94% as organic, and 87 % overall. Thus discrimination between psychogenic and organic ED is improved when more items concerning sexual experience can be assessed. Although this may seem an obvious observation, it is striking that no previous studies have taken account of it. Since the two diagnostic categories formed the main dependent variable, the multidisciplinary evaluation and criteria used to assign patients to the aetiological groups are very important. In keeping with many previous studies the evaluative procedure was not standardized and included only those diagnostic tests considered necessary by the urologist to allow an appropriate diagnosis, NPT recordings only being conducted if indicated [ 8-10,14,21]. Both financial and ethical constraints prohibited the application of all diagnostic tests in the whole population. Since an infallible instrument for the differential diagnosis of ED is not (and probably will never be) available, we believe that our decision to rely on clinical judgement based on an extensive urological and psychiatric evaluation is justifiable. The selected items will have to be applied as a separate screening questionnaire to derive the weights of the scores and cut-off values which can be used clinically. In addition, the replication of our findings in other study populations is required to
142
A. E. M. SPECKENS etul
confirm the discriminatory power of the items selected [ 3 1 ] . It must be recognized that because a fairly large number of analyses were carried out in this study the overall discriminatory power found is probably too optimistic. As already mentioned, several previous screening questionnaires appeared to discriminate well in the study in which they were originated, but disappoint in later studies [6-171 However, most of the differentiating items we found have been identified in previous research. The importance of presence and quality of morning erections as well as masturbatory erections in differentiating psychogenic and organic ED has been reported by several authors [ 19-21,231. According to Hatch and colleagues [ 221 the degree of sexual interest also contributed to the discrimination in their study, being lower in the psychogenic group. The attribution of the ED to psychological causes in men with psychogenic ED was established in two earlier studies [ 22,33 ] . Hatch and colleagues [22] found that sudden or gradual onset of ED differentiated psychogenic from organic ED, although that finding has not been confirmed by other investigators [ 14,321 Abel and colleagues [ 201 reported that premature ejaculation distinguished psychogenic from organic ED. Although lower coital frequency prior to psychogenic ED has not previously been reported, in clinical practice this is not an unusual observation and concords with the reduced sexual interest found prior to ED in this group. The association of tension in the relationship with psychogenic ED is in keeping with that of a previous retrospective study in which relationship difficulties were more common in patients with psychogenic ED [ 341 As already mentioned in the introduction, clinicians increasingly emphasize that in many cases classification of ED as either purely psychogenic or purely organic is not possible. Some authors even argue that considering all the conflicting and inconsistent reports in the literature on how best to differentiate organic from psychogenic ED the concept underlying this issue should be critically re-examined [ 23,351 . Erectile functioning may be best viewed as the final common pathway of various individual systems. Assessment methodology should reflect this through the use of specific tests designed to detect specific pathology e.g. vascular, neurological or psychological factors [ 51 . Identifying features in which patients with primarily psychogenic ED differ from patients with primarily organic ED is a valuable first step towards developing a screening questionnaire specifically for identifying patients in which psychological factors play an important aetiological role. The selection of items indicating the absence of organic pathology, such as rigidity of morning erections, in addition to items related to possible psychological aetiological factors, for example relationship problems, is inherent to the design of this study. One should also not lose sight of the fact that diagnosis is not an end in itself, but rather a process of gathering information useful in the formulation of a treatment plan. For this purpose, one should take into consideration the degree and nature of the contributory physiological and psychological factors together with prognostic indicators [ 36,371. More research is needed to determine whether, and to what extent, the success of various methods of treatment is related to the different types of aetiological factors. In anticipation of further research, we believe our findings represent a useful contribution to the multifactorial assessment of ED. A~know/ec(KrmPnt.s-The
version
of the manuscript,
authors wish to thank Dr John Bancroft for his helpful comments and the patients for their cooperation.
on an earlier
Erectile
dysfunction
143
REFERENCES 1. SPECTOR KR, BOYLE M. The prevalence and perceived aetiology of male sexual problems in a nonclinical sample. Br J Med Psycho1 1986; 59: 351-358. 2. SLAG MF, MORLEY JE, ELSON MK, TRENCE DL, NEL.SON CJ, NELSON AE, KINLAW WB, BEYER HS, NUTTALL FQ, SHAFER RB. Impotence in medical clinic outpatients. JAMA 1983; 249:
1736-1740. 3. SEGRAVES RT, 4. 5. 6. 7.
8.
9. 10. 11. 12.
13. 14.
15.
16. 17. 18. 19.
20.
21.
22. 23.
SCHOENBERG HW, ZARINS CK, CAMIC P, KNOPF J. Characteristics of erectile dysfunction as a function of medical care system entry point. Psychosom Med 198 1; 43: 227-234. MELMAN A, TIEFER L, PEDERSEN R. Evaluation of first 406 patients in urology department based Center for Male Sexual Dysfunction. Urology 1988; 32: 6-10. MEISLER AW, CAREY MP. A critical re-evaluation of nocturnal penile tumescence monitoring in the diagnosis of erectile dysfunction. J Nerv Men? Dis 1990; 178: 78-89. BEUTLER LE, KARACAN I, ANCH AM, SALIS PJ, SCOTT FB, WILLIAMS RL. MMPI an MIT discriminators of biogenic and psychogenic impotence. J Consult Clin Psycho1 1975; 43: 899-903. JEFFERSON TW, GLAROS A, SPEVACK M, BOAZ TL, MURRAY FT. An evaluation of the Minnesota Multiphasic Personality Inventory as a discriminator of primary organic and primary psychogenic impotence in diabetic males. Arch Sex Behav 1989; 18: 117-126. LEVENSON H, OLKIN R, HERZOFF N, DELANCY M. MMPI evaluation of erectile dysfunction: failure of organic versus psychogenic decision rules. J Clin Psycho1 1986; 42: 752-754. MARSHALL P, SURRIDGE D, DELVA N. Differentiation of organic and psychogenic impotence on the basis of MMPI decision rules. J Consulr Clin Psycho1 1980; 48: 407-408. MARTIN LM, RODGERS DA, MONTAGUE DK. Psychometric differentiation of biogenic and psychogenic impotence. Arch Sex Behav 1983; 12: 475-485. ROBINER WN, GODEC CJ, CASS AS, MEYER JJ. The role of Minnesota Multiphasic Personality Inventory in evaluation of erectile dysfunction. J Ural 1982; 128: 487-488. STAPLES RB, FICHER IV, SHAPIRO M, MARTIN K, GONICK P. A re-evaluation of MMPI discriminators of biogenic and psychogenic impotence. J Consult Clin Psycho1 1980; 48: 543-545. CAMIC PM. Differentiating organic and psychogenic erectile dysfunction with the Millon Behavioral Health Inventory. Sex Disabil 1983; 6: 145-149. LANTINGA LJ, KRAUSS DJ, CLARK MM, SCHELL RE. Differentiating the etiology of male erectile dysfunction using the Millon Behavioral Health Inventory and self-report measures. J Sex Mariral Ther 1988; 14: 263-268. DEROGATIS LR, MEYER JK, DUPKIN CN. Discrimination of organic versus psychogenic impotence with the DSFI. J Sex Marital Ther 1976; 2: 229-240. SEGRAVES RT, SCHOENBERG HW, ZARINS CK, KNOPF J, CAMIC P. Discrimination of organic versus psychological impotence with the DSFI: a failure to replicate. JSexMariral Ther 1981; 7: 230-238. MELMAN A, REDFIELD J. Evaluation of the DSFI as a test of organic impotence. Sex Disabil 1981; 4: 108-l 14. GEISSER ME, JEFFERSON TW, SPEVAK M, BOAZ T. THOMAS RG, MURRAY FT. Reliability and validity of the Florida sexual history questionnaire. J Chin Psychology 1991; 47: 519-528. KOCKOTT G, FEIL W, REVENSTORF D, ALDENHOFF J, BESINGER U. Symptomatology and psychological aspects of male sexual inadequacy: results of an experimental study. Arch Sex Behav 1980; 9: 457-475. ABEL GG, BECKER JV, CUNNINGHAM-RATHNER J, MITTELMAN M, PRIMACK M. Differential diagnosis of impotence in diabetics: the validity of sexual symptomatology. Neurourol Urodynam 1982; 1: 57-69. SEGRAVES KA, SFGRAVES RT, SCHOENBERG HW. Use of sexual history to differentiate organic from psychogenic impotence. Arch Sex Behav 1987; 16: 125-137. HATCH JP, DE LE PENA AM, FISHER JG. Psychometric differentiation of psychogenic and organic erectile disorders. J Ural 1987; 138: 781-783. CONTE HR. Multivariate assessment of sexual dysfunction. J ConsuO Chin Psycho! 1986; 54:
149-157. 24. PADMA-NATHAN H, GOLDSTEIN I, KRANE RJ. Evaluation 25. 26.
of the impotent patient. Semin Ural 1986; 4: 225-232. SLOB AK, BLOM JH, VAN DER WERFF TEN BOSCH JJ. Erection problems in medical practice: differential diagnosis with relatively simple method J Ural 1990; 143: 46-50. GIESBERS AA, KRAMER AE, JONAS U. Differential diagnosis in erectile impotence. Intracorporeal papaverine injection and RigiScan tumescence and rigidity monitoring. Actu Ural Belg 1988; 56:
192-201. 27. FRENKEN J, VENNIX P. Sexual Experience 1978.
Scales: Manual
(in Dutch).
Lisse:
Swets & Zeitlinger,
144
A. E. M. SPECKENS et al
28. BAKER CD, DE SILVA P. The relationship between male sexual dysfunction and belief in Zilbergeld’s myths: an empirical investigation. Sex Marif 7&r 1988; 3: 229-238. 29. DEROGATE LR, MEYER JK, KOURLESIS S. Psychiatric diagnosis and psychological symptoms in impotence. Hillside J Clin Psychiar 1985; 7: 120-133. 30. HOSMER DW, LEMESHOW S. Applied Logistic Regression. New York: Wiley, 1989. 31. WASSON JH, Sox HC, NEFF RK, GOLDMAN L. Clinical prediction rules. Applications and methodological standards. N Engl J Med 1985; 313: 793-799. 32. CONDRA M, MORALES A, SURRIDCE D, MARSHALL P, FENEMORE J. Evaluation of the urological assessment in impotence: findings with a new diagnostic rating scale. J CJrol 1984; 131: 486-490. 33. PLANTE TG, KERNS RD. YELLIG W, HAYTHORNTHWAITE J. Using the biopsychosocial model to predict nocturnal penile rigidity in men with erectile dysfunction. J Sex Marital 7her 1989; 15: 247-254. 34. REID K, SURRIDGE D, HARRIS C, CONDRA M, FENEMORE J, OWEN J, MORALES A. The psychological correlates of psychogenic impotence: a retrospective comparison of psychologenically and organically impotent men. Sex Disabil 1987; 8: 3-16. 35. WILLIAMS W. Psychogenic erectile impotence-a useful or a misleading concept‘? Aust N 2 J Psychiat 1985; 19: 77-82.23. 36. HENGEVELD MW. Erectile dysfunction: diagnosis and choice of therapy. World J Ural 1986; 3: 249-252. 37. MOHR DC, BEU~LER LE. Erectile dysfunction: a review of diagnostic and treatment procedures. Clin Psycho1 Rev 1990; 10: 123-150.
APPENDIX Items from the L.&den Impotence psychogenic and organic ED
Questionnaire
which in the multivariate
analyses
distinguished
between
1. Did your erection problem start suddenly or did it develop over a period of time’? (a) developed over weeks (b) developed over months 2. Has your penis become smaller since the development of your erection problem? (a) yes (b) no 3. How often have you noticed morning-erections during the last 6 months? (a) never (continue with question 5) (b) sometimes 4. How stiff has your erection been at its best (at night or in the morning) during the last 6 months? (a) more than 50% of how stiff it used to be (b) 50% or less of how stiff it used to be 5. How stiff did your erection become at its best when you last masturbated (if you have tried in the past 6 months)? (a) I did not try to masturbate (b) more than 50% of how stiff it used to be (c) 50% or less of how stiff it used to be 6. Before the development of your erection problem, how often were you having sexual intercourse (if you had sexual intercourse)? (a) I did not have sexual intercourse (b) once a month or less often (c) more often than once a month 7. Did you have problems with ejaculating too quickly when you last had sexual intercourse (if you have tried during the past 6 months)? (a) I did not try to have sexual intercourse (b) yes (c) no 8. If you have a regular partner, does your erection problem ever result in tension between you and your partner? (a) 1 don’t have a regular partner (continue with question IO) (b) yes (c) no
Erectile
dysfunction
145
9. Have you ever thought about separation or divorce after your erection problem began? (a) yes (b) no 10. During the last 6 months, has your interest in sex been less than before your erection problem began? (a) yes (b) no 11. Do you think your erection problem has something to do with any psychological problems or worries you have?
Research, Vol. 37, No. 2, pp. 135-145.
1993. 0
DISCRIMINATION
0022-3999/93 $6.00+.00 1993 Pergamon Press Ltd
BETWEEN PSYCHOGENIC AND ORGANIC ERECTILE DYSFUNCTION
ANNE E. M. SPECKENS,* MICHIEL W. HENGEVELD,* Guus
A. B. LYCKLAMA A NrJEHor_-r,t
ALBERT M. VAN
HEMERT* and
KEITH E. HAWTON$ (Received 25 February 1992; accepted in revised form 15 July
1992)
Abstract-The aim of this study was to develop a screening test based on the Leiden Impotence Questionnaire (LIQ) in order to assist in the difficult process of differentiating between psychogenic and organic erectile dysfunction (ED). The main sample consisted of 176 patients with ED, which was classified according to the results of the urological and psychiatric assessment as either organic, in 109 (62%) patients, or psychogenic, in 67 (38%) patients. A logistic regression model including six general items from the LIQ correctly identified psychogenic ED in 62 % of the cases, and organic ED in 86%) with an overall correct classification rate of 76%. Adding information regarding sexual intercourse and the relationship in patients who had a partner and were having sexual intercourse the correct classification rates were: psychogenic 77 %, organic 94%) and overall 87 % Discrimination between psychogenic and organic ED is improved when more information concerning sexual activity can be assessed.
INTRODUCTION ERECTILE dysfunction (ED) is a common problem among men in the general population and in medical settings [ 1,2] . Until recently the vast majority of patients who presented with ED were thought to have a psychogenic disorder. Although prevalence estimates have probably been greatly influenced by the clinical setting in which they occur [ 31 , it appears that organic factors are relevant in many more patients than was believed previously [4] . Recent advances in both psychological and medical treatments for ED have made the accurate differential diagnosis of psychogenic and organic contributions to aetiology increasingly important. The most widely accepted of the diagnostic techniques used to distinguish between organic and psychogenic ED is nocturnal penile tumescence monitoring (NPT). Despite its widespread use, some investigators have raised questions about its diagnostic utility and accuracy. In a review in which the several technical and practical problems which could affect the use and interpretation of NPT results were discussed, Meisler and Carey [5] concluded that NPT may misdiagnose as many as 20% of patients and that more sophisticated validation studies are needed. In addition, NPT is a time-consuming, inconvenient and expensive procedure that requires facilities not available in most centres. In clinical practice its use might best be reserved for cases in which there are clear indications of a likely physical aetiology.
*Department of Psychiatry, University Hospital Leiden, The Netherlands. TDepartment of Urology, University Hospital Leiden, The Netherlands. SUniversity Department of Psychiatry, Warneford Hospital, Oxford, U.K. Address for correspondence: Anne E. M. Speckens, Department of Psychiatry, Leiden, Bl-P, P.O. Box 9600, 2300 RC Leiden, The Netherlands. 135
University
Hospital
136
A. E. M.
SPECKENS
et al
These factors have led to the search for more cost-efficient methods to assist in the discrimination between psychogenic and organic erectile problems. Initially, several studies focused on personality variables. Beutler and colleagues [6] described the use of the Minnesota Multiphasic Personality Inventory (MMPI) and Male Impotence Test (MIT) with 22 impotent men whose diagnosis was confirmed by NPT. Although the ability of the MIT to identify biogenic and psychogenic ED was quite low, the MMPI appropriately classified 90% of the cases. However, subsequent studies concluded that the MMPI and the decision rules Beutler and colleagues derived from MMPI scores were ineffective in differentiating between functional and organic ED [7-121. Camic [ 131 examined the Millon Behavioral Health Inventory (MBHI) scores of 40 impotent subjects and 40 control subjects and developed two discriminate functions which differentiated the organic, psychogenic subjects and normal controls with an overall correct classification rate of 81%. In a subsequent study, following discriminant analysis of the MBHI scores in 70 male patients, the proportion of correctly classified patients dropped to 47% in the psychogenic, 65% in the mixed, and 71% in the organic group [ 141. Another approach to the problem has been to assess sexual characteristics that might differentiate organic from psychogenic ED. Derogatis and colleagues [ 151 investigated the efficiency of the Derogatis Sexual Functioning Inventory (DSFI) in 28 impotent males whose diagnosis was established by clinical judgment. Relying on the Gender Role Definition subtest of the DSFI, they reported the overall correct classification rate as 89%. However, later studies failed to replicate these findings [ 16,171 , and in a recent study in 33 diabetic patients the Florida Sexual History Questionnaire did not succeed in discriminating between organic and psychogenic ED [18]. More encouraging results have been reported with questionnaires that closely examine the sexual symptomatology of ED. In a study of 42 male patients, 6 out of 80 sexual history questions effectively differentiated men with organic ED from men with psychogenic ED [ 191 and in 60 males with diabetes mellitus seven questions significantly discriminated between men with differing aetiologies to their erectile problems [20] . From a study of 32 patients, Segraves and colleagues [21] concluded that three aspects of symptomatology effectively discriminated between those with organic and those with psychogenic ED, namely the presence of adequate early morning erections, masturbatory erections and non-coital erections. Hatch and colleagues [22] examined the discriminative validity of information collected during a clinical interview with 56 patients and found that this correctly classified all of them into organic and psychogenic groups. Increasingly clinicians are recognizing that many cases of ED are not either entirely psychogenic or entirely organic, but that both psychogenic and organic factors may be relevant in the individual case [23] . Therefore the performance of possible discriminating measures is bound to be limited. Nevertheless, it is important when possible to determine the main aetiological influence early in the assessment to guide the use of specific diagnostic tests. In addition, with the availability of effective psychological and physical methods of treatment this may be of assistence in the formulation of a treatment plan. We have conducted a study of a large sample of men with ED using a questionnaire which assesses sexual functioning, particularly specific sexual symptoms. The aim was to compare the discrimination between
Erectile dysfunction
137
primarily psychogenic and primarily organic ED of this paper and pencil test with a classification based on an extensive urological and psychiatric evaluation. METHOD Study population
Subjects eligible for the study consisted of consecutive patients with ED who were either referred to the Urology Clinic (84.4%) or to the Sexual Dysfunction Clinic in the Department of Psychiatry (16.6%) of Leiden University Hospital between April 1985 and November 1988. For the purposes of the study, all patients underwent assessment in both clinics. Some patients were referred in order to evaluate their appropriateness for penile prosthesis surgery (13.3 %) or to receive psychosexual treatment (13.3%). However, most patients (73.3%) were evaluated for diagnostic reasons. Physical
assessment
including those physical diagnostic tests thought All patients underwent a urological assessment, necessary by the urologist to assess in what extent organic factors contributed to the aetiology of the ED. Although in this way a full investigation did not take place in the whole population, all patients had a basic assessment consisting of a medical history, a physical examination of the external genitals and prostate and measurement of blood pressure, peripheral pulses and bulbocavernosus-reflex. Laboratory studies included full blood count, glucose, blood urea and nitrogen, electrolytes, liver and cardiac enzymes, serum testosterone, FSH, LH and serum prolactin. Based on the results of penile blood pressure measurements the penile brachial index (PBI) was calculated [24]. In the first 60 patients, nocturnal penile tumescence was measured with an erectometer [25] In subsequent patients, the tumescence and rigidity of penile erections were recorded utilizing Rigiscan assessment during visual sexual stimulation in combination with an intracavernous injection with papaverine [26]. If required, nocturnal penile erections were recorded at the patient’s home with the Rigiscan. In addition, nerve conduct studies, cavemosography, cavemosometry and selective angiography were conducted if indicated. Psychiatric
and psychological
assessments
The patient and his partner (if available) were interviewed separately and together by a psychiatrist. The psychiatric assessment included an extensive psychosexual history. The I.&den Imporence Questionnaire (LZQ) was administered. This was developed by the second author as an instrument to assess sexual function in patients with ED and their partners. Items of interest were selected on the basis of previous research and clinical experience. The LIQ consists of 72 items which are answered on 3-5-point Likert scales. The areas covered include: current sexual functioning (17 items), sexual functioning before the onset of the ED (19 items), aetiology of the ED (5 items), consequences of the ED (16 items) and expectations with regard to treatment options (15 items). Copies of the full version of the LIQ are available from the authors on request. The Sexual Experience Scale [ 271, Body Image Scale of the Derogatis Sexual Functioning Inventory [ 151, Zilbergeld’s Myths Questionnaire [28] and Symptom Check List-90 [29] were also administered. However, results regarding the psychological tests other than the LIQ will not be discussed further in this paper. Class#cation
of patients
On the basis of the results of all the findings of the multidisciplinary evaluation, patients were classified by one of the urologists and a psychiatrist (MWH) into four aetiological groups: (i) probable organic ED-a medical cause could be established for the ED and relevant psychogenic factors were absent; (ii) probable psychogenic ED-no organic cause could be found by the physician and psychogenic factors of sufficient severity to account for the onset of ED were present; (iii) mixed organic and psychogenic ED-organic and psychogenic factors were both significant; and (iv) ED of uncertain aetiology-neither organic nor psychogenic factors relevant to the ED could be found. The last group also included those subjects who did not comply sufficiently with the evaluation to allow a definite diagnosis. Patients in the fourth category were excluded from further analyses. Statistical
analyses
Prior to the statistical tests the items in the LIQ were dichotomized. Individual items were examined according to their relative frequency in the organic and psychogenic ED groups using the chi-square test. As it is known for dichotomous independent variables the discriminant function estimators will overestimate the magnitude of the association, categorical modeling with logistic regression analysis was used to assess the independent contribution of each of the individual items to the discrimination between
138
A. E. M. SPECKENS et (11
the diagnostic groups [ 301 In order to restrict the number of missing values inherent in a questionnaire concerning different types of sexual behaviour, the items were divided into six clusters: general, orgasm, masturbation, non-coital sex, sexual intercourse and the patient’s relationship. Data in each cluster were introduced into the logistic regression model only for patients in whom the particular cluster applied. For example, whereas logistic regression analysis of the general items was conducted in the whole sample, items concerning sexual intercourse were only included for those subjects who reported having sexual intercourse. To establish the discriminating power of the final logistic models a predicted probability of caseness was calculated for all subjects. This score was calculated as follows: predicted probability of caseness = l/(1 + em’) where A represents the sum of the constant and the regression coefficients of the selected items which are present in the individual. Subsequently, a cut-off value was chosen that maximized the correct classification rate. At this cut-off level were recorded: (a) the percentage of cases correctly classified as psychogenic; (b) the percentage of cases correctly classified as organic; and (c) the overall correct classification. The performance of the discriminatory items of the LIQ in the group of patients with apparent mixed aetiology was examined in a subsidary analysis.
RESULTS
Subjects Overall, 217 patients with ED were entered into the study. Eight patients failed to complete the LIQ and were therefore excluded. According to the multidisciplinary evaluation of the remaining 209 patients, in 109 (52%) the ED was considered to have an organic aetiology, 67 (32%) were assigned to the psychogenic group, 22 (11%) were classified as mixed organic and psychogenic and in 11 (5 %) patients the aetiology of the ED was uncertain. The patients in the last category were excluded from further analyses. The sample included in the main analysis thus consisted of 176 patients, 109 (62%) patients in the organic group and 67 (38%) patients considered to have a psychogenic aetiology. The patients with organic ED were significantly older than the patients with psychogenic ED, but the two groups were similar with regard to the proportions who had partners and duration of the ED (Table I). TABLE
I.-CHARACIERISTICS
OF WE
Organic ED (N = 109) Age: Mean (SD) Having a partner Duration ED: Mean (so)
52.5
(13.0) yr 81% 6.9 (7.8) yr
SAMPLE
Psychogenic ED (N = 67) 44.6
(11.7) yr 80% 5.7 (7.4) yr
p < o.ot* N.S. N.S.*
*r-test
Leiden Impotence
Questionnaire
According to the chi-square analysis, 17 of the 72 items of the LIQ discriminated between the psychogenic and organic groups. They are listed in Table II, together with two non-significant items (mode of onset and thoughts about separation) which proved to be relevant in the multivariate analysis. Logistic regression analysis was used to determine the independent contributions of the items to the distinction between the two groups. The item concerning ED with a different partner was omitted because few patients were able to answer this question. Applying logistic regression analysis to the whole sample, six items appeared
Erectile TABLE
II.-LEIDEN
IMPOTENCE
SUBJECTS
Items (shown in the LIQ)
according
WIT”
dysfunction
QUESTIONNAIRE ORGANIC
to groupings
AND
(LIQ): THOSE
of questions
General questions 1. Financial problems Current sexual functioning 2. Sudden onset of ED 3 Presence morning erections 4. Rigidity morning erections > =50% of previous normal erections 5. Rigidity masturbatory erections > =50% of previous normal erections 6. Rigidity coital erections > =50% of previous normal erections 7. Adequate erections with different partner 8. Perceived contractions during orgasm 9. Premature ejaculation Sexual functioning before onset of ED 10. Frequency coitus > l/month 11. Duration coitus > 10 minutes 12. Perceived contractions during orgasm Consequences of ED 13. Reduced size of penis 14. Decline in sexual interest 15. Partner regards patient as not being a real man 16. Tension in relationship 17. Thoughts about separation Patient’s opinion about the cause of ED 18. Psychogenic aetiology 19. Associated with psychological problems *Chi-square
139 ITEMS
WITH
DIFFERENTIATING
PSYCHOGENIC
BETWEEN
ED
Organic ED N= 109 (%)
Psychogenic ED N=67 (%)
P*
42 (40)
37 (56)
< 0.05
31 (31) 76 (72)
28 (42) 59 (91)
N.S. < 0.01
40 (53)
44 (75)
< 0.01
32 (39)
36 (71)
< 0.001
41 3 41 35
40 7 13 38
< < < <
(44) ( 9) (46) (45)
(70) (33) (24) (69)
0.01 0.05 0.01 0.01
88 (90) 22 (23) 61 (67)
45 (74) 25 (40) 23 (41)
< 0.01 < 0.05 < 0.01
69 40 38 34 23
30 37 33 30 9
(46) (58) (58) (52) (17)
< 0.05
35 (54) 45 (68)
< 0.01 < 0.001
(65) (39) (38) (33) (22)
30 (29) 38 (36)
< 0.05 < 0.05 < 0.05 N.S.
tests.
to contribute significantly to the discrimination between patients with psychogenic ED and those with organic ED. The most important discriminators were the presence (item 3) and rigidity of morning erections (item 4). In addition, a decline in sexual interest (item 14) and a belief that the ED was associuted withpsychologicul problems (item 19) were related to psychogenic ED. Perceived contractions during orgasm before the onset of ED (item 12) and reduction in penile size (item 13) were associated with organic ED. In contrast to organic ED, psychogenic ED tended to have a sudden onset (item 2), but this difference did not reach significance. Consequently, logistic regression analysis was applied to sub-samples of patients for which a particular cluster of items was relevant. In the group of 106 patients reporting masturbation the rigidity of the erection during masturbation (item 5) independently discriminated between psychogenic and organic ED. Items concerning non-coital sex did not differentiate between the two groups. In the 131 patients having sexual intercourse, premature ejaculation (item 9) and a lower coitul fiequency before the onset of the ED (item 10) were associated with psychogenic ED. In the 139 patients who had a partner, tension in the relationship (item 16) discriminated between psychogenic and organic ED. We did not include perceived contractions during orgasm before onset of the ED in the final models, because the basis on which this item differentiated between the two groups was not apparent. The nature of onset of ED was included because this
140
A. E. M.
TABLE
III.-MULTIVARIATE
LOGISTIC
SPECKENS
REGRESSION
VARIOUS
Models
COEFFICIENTS
2
- 1.545 0.537 2.039** 1.300 1.088** 1.256** -0.882*
ACCORDING
TO
THE
MODELS
1
Constant Onset Presence morning erections Rigidity morning erections Decline sexual interest Psychogenic aetiology Reduced penile size Rigidity masturbatory erections Tension relationship Thoughts separation Premature ejaculation Coital frequency before ED
et ul.
-
3
I. 736 0.739 2.584** 1.910 0.975 1.224*
-0.932 0.869
4
5
_ -2.033
-2.056 0.875 2.533** 1.593 0.717 1.218* -0.983*
2.223* 1.881 0.931 2.345** - I .608**
-3.284 1.689* 3.120* 2.618* 0.471 2.301** - 1.725*
1.581* 2.082*
0.964 -0.348 I .787* 2.202*
I. 192*
0.958 -1.151
*p < 0.05; **p < 0.01. finding has support from some previous research [ 221. The constant and the logistic regression coefficients of the items as analysed in the final logistic models are provided in Table III. On the basis of these models a predicted probability of caseness was calculated for all subjects. Consequently, cut-off levels were chosen which maximized the correct classification rate. Table IV shows the cut-off levels and the corresponding percentages of cases correctly classified as psychogenic or organic, and overall, when different clusters were entered into the analysis. A mode1 using the basic cluster of symptoms, namely presence and rigidity of morning erections, decline in sexual interest, perceived connection of the ED with psychological problems, reduction in size of the penis and gradual or sudden onset of the ED (cluster l), correctly classified 62% of the psychogenic and 86% of the organic cases, with an overall correct classification rate of 76%. If only the two items, presence and rigidity of morning erections, were used, the correct classification rates were: psychogenic 68 % , organic 62 %, and overall 64%. TABLE
IV.-DISCRIMINATION
ED
BETWEEN ACCORDING
TO
GROUPS
OF
ANALYSES
PATIENTS
WITH
WITH
DIFFERENT
PSYCHOGENIC
% (IV) of cases correctly Clusters 1. Gradual Presence Rigidity Decline ED due Reduced
Organic or sudden onset of ED of morning erections of morning erections in sexual interest to psychological problems size of penis
2. Rigidity of masturbatory (+ cluster 1)
erections
3. Tension in relationship Thoughts about separation (+ cluster I) 4. Premature ejaculation Coital frequency before (+ cluster 1) 5. Cluster
1 + 3 + 4
Psychogenic
AND
ORGANIC
CLUSTERS
classified
Overall
Cut-off
86%
(92)
62%
(60)
76%
(152)
0.51
79%
(61)
80%
(40)
79%
(101)
0.42
82%
(76)
74%
(47)
79%
(123)
0.43
85%
(59)
83%
(47)
84%
(106)
0.51
94%
(54)
77%
(39)
87%
( 93)
0.56
onset of ED
level
Erectile dysfunction
141
Adding the rigidity of erections during masturbation in those who reported masturbation (cluster 2), the percentage of cases correctly classified as psychogenic improved over that in the initial model, although the overall correct classification rate barely changed (Table IV). The same was found when tension in the relationship and thoughts about separation in those with a regular partner (cluster 3) were introduced into the basic model (cluster 1). However, including premature ejaculation and coital frequency before the onset of ED in patients having sexual intercourse (cluster 4), both the percentage of cases correctly classified as psychogenic and the overall correct classification rate substantially improved. The best overall classification rate was produced when items concerning sexual intercourse (cluster 4) as well as the relationship items (cluster 3) in those for whom both clusters were relevant were added to the basic model (cluster l), with a particularly high rate of correct classification of the organic cases. Adding items concerning masturbation did not significantly improve the overall correct classification. Applying the basic model (cluster 1) to the 22 patients with ED based on a mixed aetiology, 11 patients would have been classified as having organic and 7 as having psychogenic ED. DISCUSSION
The aim of this study was to compose a screening test from a questionnaire which assesses sexual functioning in patients with ED in order to assist in the discrimination between psychogenic and organic ED. The full list of questions from the Leiden Impotence Questionnaire for which statistical differences were found between the psychogenic and organic ED groups on the basis of the logistic regression analysis is presented in the Appendix. A logistic regression model which included seven general items from the LIQ correctly identified aetiology of ED in 76 % of the cases, with the percentages correctly classified as psychogenic and organic being 62% and 86% respectively. Adding items regarding sexual intercourse and the relationship in patients who had a partner and were having sexual intercourse resulted in 77% of cases being correctly classified as psychogenic, 94% as organic, and 87 % overall. Thus discrimination between psychogenic and organic ED is improved when more items concerning sexual experience can be assessed. Although this may seem an obvious observation, it is striking that no previous studies have taken account of it. Since the two diagnostic categories formed the main dependent variable, the multidisciplinary evaluation and criteria used to assign patients to the aetiological groups are very important. In keeping with many previous studies the evaluative procedure was not standardized and included only those diagnostic tests considered necessary by the urologist to allow an appropriate diagnosis, NPT recordings only being conducted if indicated [ 8-10,14,21]. Both financial and ethical constraints prohibited the application of all diagnostic tests in the whole population. Since an infallible instrument for the differential diagnosis of ED is not (and probably will never be) available, we believe that our decision to rely on clinical judgement based on an extensive urological and psychiatric evaluation is justifiable. The selected items will have to be applied as a separate screening questionnaire to derive the weights of the scores and cut-off values which can be used clinically. In addition, the replication of our findings in other study populations is required to
142
A. E. M. SPECKENS etul
confirm the discriminatory power of the items selected [ 3 1 ] . It must be recognized that because a fairly large number of analyses were carried out in this study the overall discriminatory power found is probably too optimistic. As already mentioned, several previous screening questionnaires appeared to discriminate well in the study in which they were originated, but disappoint in later studies [6-171 However, most of the differentiating items we found have been identified in previous research. The importance of presence and quality of morning erections as well as masturbatory erections in differentiating psychogenic and organic ED has been reported by several authors [ 19-21,231. According to Hatch and colleagues [ 221 the degree of sexual interest also contributed to the discrimination in their study, being lower in the psychogenic group. The attribution of the ED to psychological causes in men with psychogenic ED was established in two earlier studies [ 22,33 ] . Hatch and colleagues [22] found that sudden or gradual onset of ED differentiated psychogenic from organic ED, although that finding has not been confirmed by other investigators [ 14,321 Abel and colleagues [ 201 reported that premature ejaculation distinguished psychogenic from organic ED. Although lower coital frequency prior to psychogenic ED has not previously been reported, in clinical practice this is not an unusual observation and concords with the reduced sexual interest found prior to ED in this group. The association of tension in the relationship with psychogenic ED is in keeping with that of a previous retrospective study in which relationship difficulties were more common in patients with psychogenic ED [ 341 As already mentioned in the introduction, clinicians increasingly emphasize that in many cases classification of ED as either purely psychogenic or purely organic is not possible. Some authors even argue that considering all the conflicting and inconsistent reports in the literature on how best to differentiate organic from psychogenic ED the concept underlying this issue should be critically re-examined [ 23,351 . Erectile functioning may be best viewed as the final common pathway of various individual systems. Assessment methodology should reflect this through the use of specific tests designed to detect specific pathology e.g. vascular, neurological or psychological factors [ 51 . Identifying features in which patients with primarily psychogenic ED differ from patients with primarily organic ED is a valuable first step towards developing a screening questionnaire specifically for identifying patients in which psychological factors play an important aetiological role. The selection of items indicating the absence of organic pathology, such as rigidity of morning erections, in addition to items related to possible psychological aetiological factors, for example relationship problems, is inherent to the design of this study. One should also not lose sight of the fact that diagnosis is not an end in itself, but rather a process of gathering information useful in the formulation of a treatment plan. For this purpose, one should take into consideration the degree and nature of the contributory physiological and psychological factors together with prognostic indicators [ 36,371. More research is needed to determine whether, and to what extent, the success of various methods of treatment is related to the different types of aetiological factors. In anticipation of further research, we believe our findings represent a useful contribution to the multifactorial assessment of ED. A~know/ec(KrmPnt.s-The
version
of the manuscript,
authors wish to thank Dr John Bancroft for his helpful comments and the patients for their cooperation.
on an earlier
Erectile
dysfunction
143
REFERENCES 1. SPECTOR KR, BOYLE M. The prevalence and perceived aetiology of male sexual problems in a nonclinical sample. Br J Med Psycho1 1986; 59: 351-358. 2. SLAG MF, MORLEY JE, ELSON MK, TRENCE DL, NEL.SON CJ, NELSON AE, KINLAW WB, BEYER HS, NUTTALL FQ, SHAFER RB. Impotence in medical clinic outpatients. JAMA 1983; 249:
1736-1740. 3. SEGRAVES RT, 4. 5. 6. 7.
8.
9. 10. 11. 12.
13. 14.
15.
16. 17. 18. 19.
20.
21.
22. 23.
SCHOENBERG HW, ZARINS CK, CAMIC P, KNOPF J. Characteristics of erectile dysfunction as a function of medical care system entry point. Psychosom Med 198 1; 43: 227-234. MELMAN A, TIEFER L, PEDERSEN R. Evaluation of first 406 patients in urology department based Center for Male Sexual Dysfunction. Urology 1988; 32: 6-10. MEISLER AW, CAREY MP. A critical re-evaluation of nocturnal penile tumescence monitoring in the diagnosis of erectile dysfunction. J Nerv Men? Dis 1990; 178: 78-89. BEUTLER LE, KARACAN I, ANCH AM, SALIS PJ, SCOTT FB, WILLIAMS RL. MMPI an MIT discriminators of biogenic and psychogenic impotence. J Consult Clin Psycho1 1975; 43: 899-903. JEFFERSON TW, GLAROS A, SPEVACK M, BOAZ TL, MURRAY FT. An evaluation of the Minnesota Multiphasic Personality Inventory as a discriminator of primary organic and primary psychogenic impotence in diabetic males. Arch Sex Behav 1989; 18: 117-126. LEVENSON H, OLKIN R, HERZOFF N, DELANCY M. MMPI evaluation of erectile dysfunction: failure of organic versus psychogenic decision rules. J Clin Psycho1 1986; 42: 752-754. MARSHALL P, SURRIDGE D, DELVA N. Differentiation of organic and psychogenic impotence on the basis of MMPI decision rules. J Consulr Clin Psycho1 1980; 48: 407-408. MARTIN LM, RODGERS DA, MONTAGUE DK. Psychometric differentiation of biogenic and psychogenic impotence. Arch Sex Behav 1983; 12: 475-485. ROBINER WN, GODEC CJ, CASS AS, MEYER JJ. The role of Minnesota Multiphasic Personality Inventory in evaluation of erectile dysfunction. J Ural 1982; 128: 487-488. STAPLES RB, FICHER IV, SHAPIRO M, MARTIN K, GONICK P. A re-evaluation of MMPI discriminators of biogenic and psychogenic impotence. J Consult Clin Psycho1 1980; 48: 543-545. CAMIC PM. Differentiating organic and psychogenic erectile dysfunction with the Millon Behavioral Health Inventory. Sex Disabil 1983; 6: 145-149. LANTINGA LJ, KRAUSS DJ, CLARK MM, SCHELL RE. Differentiating the etiology of male erectile dysfunction using the Millon Behavioral Health Inventory and self-report measures. J Sex Mariral Ther 1988; 14: 263-268. DEROGATIS LR, MEYER JK, DUPKIN CN. Discrimination of organic versus psychogenic impotence with the DSFI. J Sex Marital Ther 1976; 2: 229-240. SEGRAVES RT, SCHOENBERG HW, ZARINS CK, KNOPF J, CAMIC P. Discrimination of organic versus psychological impotence with the DSFI: a failure to replicate. JSexMariral Ther 1981; 7: 230-238. MELMAN A, REDFIELD J. Evaluation of the DSFI as a test of organic impotence. Sex Disabil 1981; 4: 108-l 14. GEISSER ME, JEFFERSON TW, SPEVAK M, BOAZ T. THOMAS RG, MURRAY FT. Reliability and validity of the Florida sexual history questionnaire. J Chin Psychology 1991; 47: 519-528. KOCKOTT G, FEIL W, REVENSTORF D, ALDENHOFF J, BESINGER U. Symptomatology and psychological aspects of male sexual inadequacy: results of an experimental study. Arch Sex Behav 1980; 9: 457-475. ABEL GG, BECKER JV, CUNNINGHAM-RATHNER J, MITTELMAN M, PRIMACK M. Differential diagnosis of impotence in diabetics: the validity of sexual symptomatology. Neurourol Urodynam 1982; 1: 57-69. SEGRAVES KA, SFGRAVES RT, SCHOENBERG HW. Use of sexual history to differentiate organic from psychogenic impotence. Arch Sex Behav 1987; 16: 125-137. HATCH JP, DE LE PENA AM, FISHER JG. Psychometric differentiation of psychogenic and organic erectile disorders. J Ural 1987; 138: 781-783. CONTE HR. Multivariate assessment of sexual dysfunction. J ConsuO Chin Psycho! 1986; 54:
149-157. 24. PADMA-NATHAN H, GOLDSTEIN I, KRANE RJ. Evaluation 25. 26.
of the impotent patient. Semin Ural 1986; 4: 225-232. SLOB AK, BLOM JH, VAN DER WERFF TEN BOSCH JJ. Erection problems in medical practice: differential diagnosis with relatively simple method J Ural 1990; 143: 46-50. GIESBERS AA, KRAMER AE, JONAS U. Differential diagnosis in erectile impotence. Intracorporeal papaverine injection and RigiScan tumescence and rigidity monitoring. Actu Ural Belg 1988; 56:
192-201. 27. FRENKEN J, VENNIX P. Sexual Experience 1978.
Scales: Manual
(in Dutch).
Lisse:
Swets & Zeitlinger,
144
A. E. M. SPECKENS et al
28. BAKER CD, DE SILVA P. The relationship between male sexual dysfunction and belief in Zilbergeld’s myths: an empirical investigation. Sex Marif 7&r 1988; 3: 229-238. 29. DEROGATE LR, MEYER JK, KOURLESIS S. Psychiatric diagnosis and psychological symptoms in impotence. Hillside J Clin Psychiar 1985; 7: 120-133. 30. HOSMER DW, LEMESHOW S. Applied Logistic Regression. New York: Wiley, 1989. 31. WASSON JH, Sox HC, NEFF RK, GOLDMAN L. Clinical prediction rules. Applications and methodological standards. N Engl J Med 1985; 313: 793-799. 32. CONDRA M, MORALES A, SURRIDCE D, MARSHALL P, FENEMORE J. Evaluation of the urological assessment in impotence: findings with a new diagnostic rating scale. J CJrol 1984; 131: 486-490. 33. PLANTE TG, KERNS RD. YELLIG W, HAYTHORNTHWAITE J. Using the biopsychosocial model to predict nocturnal penile rigidity in men with erectile dysfunction. J Sex Marital 7her 1989; 15: 247-254. 34. REID K, SURRIDGE D, HARRIS C, CONDRA M, FENEMORE J, OWEN J, MORALES A. The psychological correlates of psychogenic impotence: a retrospective comparison of psychologenically and organically impotent men. Sex Disabil 1987; 8: 3-16. 35. WILLIAMS W. Psychogenic erectile impotence-a useful or a misleading concept‘? Aust N 2 J Psychiat 1985; 19: 77-82.23. 36. HENGEVELD MW. Erectile dysfunction: diagnosis and choice of therapy. World J Ural 1986; 3: 249-252. 37. MOHR DC, BEU~LER LE. Erectile dysfunction: a review of diagnostic and treatment procedures. Clin Psycho1 Rev 1990; 10: 123-150.
APPENDIX Items from the L.&den Impotence psychogenic and organic ED
Questionnaire
which in the multivariate
analyses
distinguished
between
1. Did your erection problem start suddenly or did it develop over a period of time’? (a) developed over weeks (b) developed over months 2. Has your penis become smaller since the development of your erection problem? (a) yes (b) no 3. How often have you noticed morning-erections during the last 6 months? (a) never (continue with question 5) (b) sometimes 4. How stiff has your erection been at its best (at night or in the morning) during the last 6 months? (a) more than 50% of how stiff it used to be (b) 50% or less of how stiff it used to be 5. How stiff did your erection become at its best when you last masturbated (if you have tried in the past 6 months)? (a) I did not try to masturbate (b) more than 50% of how stiff it used to be (c) 50% or less of how stiff it used to be 6. Before the development of your erection problem, how often were you having sexual intercourse (if you had sexual intercourse)? (a) I did not have sexual intercourse (b) once a month or less often (c) more often than once a month 7. Did you have problems with ejaculating too quickly when you last had sexual intercourse (if you have tried during the past 6 months)? (a) I did not try to have sexual intercourse (b) yes (c) no 8. If you have a regular partner, does your erection problem ever result in tension between you and your partner? (a) 1 don’t have a regular partner (continue with question IO) (b) yes (c) no
Erectile
dysfunction
145
9. Have you ever thought about separation or divorce after your erection problem began? (a) yes (b) no 10. During the last 6 months, has your interest in sex been less than before your erection problem began? (a) yes (b) no 11. Do you think your erection problem has something to do with any psychological problems or worries you have?