- Email: [email protected]
Effects of clonidine in schizophrenic patients with primary polydipsia
Progress in Neuro-Psychopharmacology & Biological Psychiatry 26 (2002) 387 – 392
Effects of clonidine in schizophrenic patients with primary polydipsia Three single case studies Nicholas J. Delvaa,b,*, Anna Changb, Emily R. Hawkenb, J. Stuart Lawsona,b,c, James A. Owena,b,d a Department of Psychiatry, Queen’s University, Kingston, Ontario, Canada Kingston Psychiatric Hospital, Queen’s University, Kingston, Ontario, Canada c Department of Psychology, Queen’s University, Kingston, Ontario, Canada d Department of Pharmacology and Toxicology, Queen’s University, Kingston, Ontario, Canada b
Abstract A pilot study was conducted in schizophrenic patients with primary polydipsia to determine the tolerability of adding clonidine to an existing antipsychotic drug regimen and to seek evidence of an antidipsic effect. Three patients with chronic schizophrenia and primary polydipsia underwent open controlled prospective trials of treatment with clonidine in doses of up to 800 mg/day. The trials lasted from 2 to 5 months each, and analysis of variance was used to test for changes in dependent variables on a case-by-case basis. Blood pressure and pulse declined significantly in a dose-dependent manner, but fluid intake, as assessed by measurements of weight and 24-h urine volume, was not affected. Hypotension and bradycardia limited the extent to which the dose of clonidine could be increased. The lack of evident effect of clonidine on polydipsia in this small sample and the inconsistent results of two other recent studies of clonidine in patients with schizophrenia and primary polydipsia provide little overall support for the effectiveness of clonidine treatment in primary polydipsia associated with schizophrenia. D 2001 Elsevier Science Inc. All rights reserved. Keywords: Antidipsic drugs; Clonidine; Primary polydipsia; Schizophrenia; Thirst
1. Introduction Severe primary polydipsia accompanied by hyponatremia is a common finding in mental hospital patients. In their review of the epidemiological literature in this area, De Leon et al. (1994) concluded that more than 20% of chronic inpatients suffer from polydipsia; water intoxication occurs in about one of four experiencing polydipsia, and schizophrenic patients account for 80% of those with water intoxication. The most recently published epidemiological study (De Leon et al., 1996) identified 26% of chronic inpatients at a US state mental hospital as polydipsic, with water intoxication observed in about one of five polydipsic patients. Patients with primary polydipsia may suffer from a number of sequelae apart from water intoxication, which
Abbreviations: ANOVA, One-way analysis of variance * Corresponding author. Department of Psychiatry, Queen’s University, 72 Barrie Street, Kingston, Ontario, Canada K7L 3J7. Tel.: +1-613-5482372; fax: +1-613-548-6095.
itself may be fatal. These include bladder dilation, enuresis, incontinence, hydronephrosis, renal failure, gastrointestinal dilation and hypotonicity, hernias, hypocalcemia, osteoporosis, hypothermia, and congestive heart failure (Blum and Friedland, 1983; Vieweg et al., 1984; Koczapski et al., 1987; Zubenko, 1987; Delva et al., 1989). Approaches to treatment have included weight monitoring with fluid restriction (Delva and Crammer, 1988), behavioral methods (Vieweg, 1994), electroconvulsive therapy (Greer and Stewart, 1993), and medication. Drug therapy has been directed toward reduction of drinking on the one hand and increasing water clearance on the other. Antipsychotic medication can reduce both psychotic symptoms and polydipsia (Vieweg, 1994), and some polydipsic patients who fail to benefit from typical neuroleptics may respond to atypical neuroleptics such as clozapine (Spears et al., 1996) or risperidone (Landry, 1995) with a decrease in drinking as well as better control of the psychosis, although Millson et al. (1992) failed to find an antidipsic effect from risperidone in a study on eight patients. Other treatments assessed for potential antidipsic efficacy have
0278-5846/02/$ – see front matter D 2001 Elsevier Science Inc. All rights reserved. PII: S 0 2 7 8 - 5 8 4 6 ( 0 1 ) 0 0 2 4 6 - 9
388
N.J. Delva et al. / Progress in Neuro-Psychopharmacology & Biological Psychiatry 26 (2002) 387–392
included fluoxetine (Goldman and Janecek, 1991), propranolol (Shevitz et al., 1980; Kathol et al., 1986), angiotensinconverting enzyme inhibitors (Kathol et al., 1986; Lawson et al., 1988; Sebastian et al., 1995), and naloxone (Nishikawa et al., 1994). While some case reports have appeared promising, the better controlled trials employing larger numbers of patients have, in general, been disappointing. Treatments directed at increasing water clearance have included the combination of lithium and phenytoin, and demeclocycline (Nixon et al., 1982; Vieweg, 1994). While the lithium/phenytoin treatment may be helpful, a controlled trial has shown no benefit from demeclocycline (Alexander et al., 1991). Clonidine, an a2 adrenergic agonist, was thought to be of potential value in treating primary polydipsia because it reduces drinking in animal models of thirst. Fregly et al. have shown that in rats clonidine attenuates water intake induced by angiotensin II, isoproterenol, pilocarpine, hypertonic saline, dehydration, or 5-hydroxytryptophan but fails to affect water intake significantly when given alone (Fregly and Kelleher, 1980; Fregly et al., 1981, 1984a,b,c; Wilson et al., 1984). On the basis of these findings and the current limitations of treatment for this serious condition, two recent studies have evaluated the effectiveness of clonidine in polydipsia patients, with contradictory results: Hayashi et al. (1997) reported that patients became more severely polydispic while receiving clonidine but Greendyke et al. (1998) found that clonidine had an antidipsic effect. The authors studied the effectiveness of clonidine in three patients with schizophrenia and significant primary polydipsia using a single case study design for each patient.
2. Methods 2.1. Subjects Patients with chronic polydipsia were identified by a survey. Three male patients consented to undergo a trial of clonidine, which had been approved by the institutional Research Ethics Board; written informed consent for the trial was obtained in each case. By DSM-IV criteria (American Psychiatric Association, 1994), all three suffered from chronic schizophrenia; one suffered a significant head injury after the onset of the schizophrenia. The mean age was 40.7 years (range 39 – 44) and mean duration of polydipsia, defined as the number of years since hyposthenuria (urine specific gravity 1.003), hyponatremia, obvious polydipsia, or any other complication was first noted, was 9.7 years (range 5 – 12). Complications related to polydipsia were present in all three subjects. All had osteoporosis and incontinence. Two had also suffered hyponatremic seizures. They all received calcium (0.22 g elemental calcium by mouth daily) and ergocalciferol (50,000 IU by mouth once weekly) for osteoporosis and neuroleptics (listed in Table 1).
One patient received diazepam on an as-needed basis (10 mg for agitation); this was given only three times in the course of his clonidine trial and did not have any effect on water intake. Each patient smoked approximately one-half package of cigarettes per day. 2.2. Procedures The study design consisted of the nonblind administration of clonidine in gradually increasing (to assess potential benefit and side effects) and then gradually decreasing doses (to avoid rebound hypertension) while other medication, other treatment, and ward environment were held constant. The length of study varied from subject to subject depending on the results of treatment. Each patient was allowed to drink freely unless his weight exceeded 7% over that while normally hydrated. Above this threshold, drinking was discouraged according to our usual practice (Delva and Crammer, 1988). Regular measurements were made of pulse rate, blood pressure, body weight, urine volume, and urine creatinine. These variables were measured before and during administration of clonidine (range: 50– 800 mg/day) over a time period ranging from 9 to 23 weeks. Clonidine was administered in divided doses, from two to four times each day, in order to minimize fluctuations in plasma levels. Pulse, blood pressure, and weight were measured daily, and urine was collected every third to fourth day. In this way, adequate data were collected for each dose of clonidine in each patient. Pulse and blood pressure were measured at 08:00 and 18:00 h; this was done after the patient had been lying for 5 min, sitting for 2 min, and standing for 2 min. Patients were weighed hourly while awake, with an empty bladder and in the same clothes. In order to ensure that the urine collections were complete, nursing staff observed the patient constantly during each urine collection; urine was collected over a 24-h period starting at 06:00 h, once every hour while awake and every 2 h during sleep. A sample of each 24-h urine collection was assayed for creatinine; the ratio of urine volume/creatinine was used to correct for any possible losses of urine during the collections. 2.3. Data analysis Mean arterial pressure was calculated as the diastolic pressure (represented by the point at which the Korotkoff sound completely disappeared) plus one-third of the difference between the systolic and diastolic pressures. Diurnal weight gain was calculated as the difference between the mean of the weights measured at 18:00, 19:00, and 20:00 h (representing the highest weights of the day) and the mean of those at 06:00 and 07:00 h (representing the lowest). In the statistical analysis, each case was examined separately. The phase of increasing dosage of clonidine
N.J. Delva et al. / Progress in Neuro-Psychopharmacology & Biological Psychiatry 26 (2002) 387–392
389
Table 1 Duration of study, number of 24-h urine collections, neuroleptic treatment, maximum dose of clonidine given, and effects of clonidine treatment on pulse and arterial pressure
Duration of study (days) Number of urine collections Neuroleptic/dose (mg/day) Maximum dose of clonidine (mg/day) Pulse, lying (bpm ± S.D.) Baseline/maximum dose Pulse, standing (bpm ± S.D.) Baseline/maximum dose MAP, lying (mmHg ± S.D.) Baseline/maximum dose MAP, standing (mmHg ± S.D.) Baseline/maximum dose
Patient 1
Patient 2
Patient 3
107 35 Haloperidol 25 800 77 ± 5/60 ± 5z,a 88 ± 8/72 ± 7z,d 92 ± 3/82 ± 6**,g 96 ± 5/82 ± 7y,j
159 40 Haloperidol 25 800 79 ± 6/69 ± 4**,b 88 ± 11/75 ± 10*,e 89 ± 5/83 ± 0y,h 91 ± 15/72 ± 1y,k
65 20 Thioridazine 300 250 71 ± 6/66 ± 13NS,c 83 ± 8/79 ± 4NS,f 92 ± 7/82 ± 6y,i 92 ± 7/75 ± 3z,l
MAP, mean arterial pressure; bpm, beats per minute. * P < .05. ** P < .01. y P < .001. z P < .0001. NS Not significant. a ANOVA for phase of increasing dose of clonidine: F7,56 = 9.29. b ANOVA for phase of increasing dose of clonidine: F8,107 = 3.50. c ANOVA for phase of increasing dose of clonidine: F5,50 = 0.968. d ANOVA for phase of increasing dose of clonidine: F7,54 = 6.13. e ANOVA for phase of increasing dose of clonidine: F8,104 = 2.04. f ANOVA for phase of increasing dose of clonidine: F5,50 = 1.19. g ANOVA for phase of increasing dose of clonidine: F7,55 = 3.36. h ANOVA for phase of increasing dose of clonidine: F8,107 = 4.37. i ANOVA for phase of increasing dose of clonidine: F5,51 = 3.80. j ANOVA for phase of increasing dose of clonidine: F7,56 = 4.32. k ANOVA for phase of increasing dose of clonidine: F8,107 = 3.68. l ANOVA for phase of increasing dose of clonidine: F5,49 = 7.72.
was treated separately from that of decreasing dosage. Oneway analysis of variance (ANOVA) was employed for each dependent variable (percentage diurnal weight gain; 24-h urine volume; urine volume/creatinine ratio; pulse, standing; pulse, lying; blood pressure, standing; blood pressure, lying) during each phase, with the dosage of clonidine administered being the independent variable. Thus, separate ANOVAs were performed for the phase in which the dose of clonidine increased and that during which it decreased. If the ANOVA yielded statistically significant results, a poly-
nomial regression analysis of the dependent variable on dosage was performed.
3. Results Although the patients knew that their drinking behaviour was under study, and the investigators and ward staff had hopes that clonidine would prove to be an effective antidipsic agent, there was no evidence that the patients were
Table 2 Body weight, diurnal weight gain, urine volume, and urine volume/creatinine ratio during entire period of study
Weight (normal hydration) Diurnal weight gain (kg ± S.D.) Percentage diurnal weight gain (% ± S.D.) Urine volume (l/24 h ± S.D.) Urine: ratio of volume to creatinine (ml/mg ± S.D.) NS a b c d e f g h i
Not significant. ANOVA for phase of increasing dose of clonidine: F7,57 = 0.46. ANOVA for phase of increasing dose of clonidine: F8,106 = 1.43. ANOVA for phase of increasing dose of clonidine: F5,54 = 1.179. ANOVA for phase of increasing dose of clonidine: F7,17 = 2.13. ANOVA for phase of increasing dose of clonidine: F8,25 = 0.58. ANOVA for phase of increasing dose of clonidine: F3,13 = 0.171. ANOVA for phase of increasing dose of clonidine: F7,16 = 1.26. ANOVA for phase of increasing dose of clonidine: F8,25 = 0.74. ANOVA for phase of increasing dose of clonidine: F3,13 = 1.40.
Patient 1
Patient 2
Patient 3
61 kg 2.5 ± 0.7 4 ± 1NS,a 7.4 ± 2.4NS,d 740 ± 313NS,g
53 kg 1.9 ± 0.6 4 ± 1NS,b 7.6 ± 2.2NS,e 627 ± 197NS,h
75 kg 2.5 ± 1.0 3 ± 1NS,c 4.9 ± 2.4NS,f 524 ± 260NS,i
390
N.J. Delva et al. / Progress in Neuro-Psychopharmacology & Biological Psychiatry 26 (2002) 387–392
motivated to alter their drinking behaviour either to please or oppose the investigators and staff. Table 1 shows the duration of the trial in each patient, the number of urine collections, concurrent neuroleptic treatment, the maximum dose of clonidine administered, the effects of clonidine on cardiovascular function, and the ANOVAs for the period of increasing dose; there were no substantial differences for the period of decreasing dose. All three patients experienced clinically and statistically significant reductions in arterial pressure as the dose of clonidine increased. Patients 1 and 2 manifested significant reductions in pulse rate. At maximal doses of clonidine, the overall mean reductions in standing pulse and mean arterial pressure were 11 beats per minute and 17 mmHg, respectively. For each case in which there was a statistically significant drug effect, there was a significant linear relationship with dose, but there was no consistent contribution by nonlinear terms up to the quintic. The administration of clonidine did not affect polydipsia as assessed by diurnal weight gain, urine volume, or urine volume/creatinine ratio. Mean values of these variables and the ANOVAs for the period of increasing dose are shown in Table 2; there were no substantial differences for the period of decreasing dose. A moderate to severe degree of polydipsia is evident.
4. Discussion In the single case approach used in the present study, the dependent measures were sampled prospectively over long periods of time, with each patient representing a ‘‘population’’ from the statistical point of view. The measures included both direct (timed urine collections) and indirect (hourly body weights) assessments. This is the only study to date on the use of clonidine for the treatment of primary polydipsia in which actual rather than estimated urine values were employed. We can say with confidence that in these three polydipsic patients, the administration of clonidine at doses sufficient to produce clinically significant hypotension and bradycardia failed to affect their daily pattern of excessive drinking during the waking hours and normalization of hydration during sleep. 4.1. Effects of clonidine on heart rate and blood pressure and in animal models of thirst Because the hypotensive effect of clonidine appears to be mediated mainly by stimulation of central a2 adrenergic receptors (Kobinger, 1978; Langer et al., 1980; Lowenthal et al., 1988), the profound cardiovascular effects induced by clonidine indicate significant central nervous system penetration in our three patients. Clonidine reduces heart rate and blood pressure by its action within the nucleus tractus solitarius of the medulla; clonidine appears to produce its antidipsic effect in animal models of thirst by action on the central nervous system, possibly also in the nucleus tractus
solitarius (Fregly et al., 1980; Wilson et al., 1984). It is possible that higher doses of clonidine may have resulted in an effect on the water intake of our patients but this was not possible to assess without jeopardising patient safety. The concurrent treatment with neuroleptic medication possessing a1 adrenergic antagonist effects may have increased the degree of hypotension observed. In the studies of Fregly’s group (Fregly and Kelleher, 1980; Fregly et al., 1981, 1984a,b,c; Wilson et al., 1984), antidipsic effects were observed in rats with doses of clonidine as low as 6 mg/kg given intraperitoneally, although fourfold higher doses were required for antagonism of dehydration-induced drinking (Fregly and Kelleher, 1980; Fregly et al., 1981). In our study, the maximum doses ranged from 3 to 15 mg/kg per os. Given that the oral bioavailability of clonidine in humans is 75 – 95% (Lowenthal et al., 1988), the doses of clonidine in our study are in approximately the same range of dosage as those used in the rat model studies of Fregly’s group. Apart from the issue of dose, it is also possible that the difference in effect of clonidine in our study in comparison with those of Fregly’s group might be the result of the difference in species evaluated. 4.2. Studies on the treatment of polydipsia with clonidine If examined in isolation, the small size of our patient sample might limit the generalizability of our findings to the population of chronically psychotic patients with primary polydipsia, but when looked at in the context of the other two studies of the effects clonidine in these patients (Greendyke et al., 1998; Hayashi et al., 1997), it adds support to the view that clonidine is not likely to be an effective antidipsic drug in polydipsic patients. Hayashi et al. (1997), in an open-label controlled study, found that all four of their polydispic patients experienced increased drinking with increasing doses of clonidine over a 6-week treatment course (a fifth patient was withdrawn from the study), but Greendyke et al. (1998), in a double-blind, placebo-controlled study, found that 7 of their 14 patients showed improvement after receiving clonidine for 4 weeks. Hayashi et al. (1997) examined the effect of clonidine in polydipsic patients, expecting that it would reduce drinking, but found instead that it exacerbated the condition, as measured by diurnal weight gain. An unanticipated and apparently strong contrary effect would generally be expected to be a genuine finding, especially in an openlabel study. Furthermore, treatment with mianserin, an a2-adrenergic antagonist, resulted in reduced diurnal weight gain. These findings would, however, be more convincing if the withdrawal phases had produced results opposite to the treatment phases: clonidine is known to produce marked cardiovascular withdrawal symptoms, but the withdrawal phase in the study of Hayashi et al. (1997) was associated with a continued prodipsic effect. Similarly, the withdrawal phase of mianerin treatment was associated with an antidipsic effect in three of the four patients. One
N.J. Delva et al. / Progress in Neuro-Psychopharmacology & Biological Psychiatry 26 (2002) 387–392
patient showed increased drinking with both clonidine and mianserin. Vital signs were measured but not reported in this study. The study by Greendyke et al. (1998) has the advantage of being double blind. The method of estimating urine output employed in this study was based on that of Vieweg et al. (1986, 1989, 1992) but involved only a single sample of urine, taken at 3 p.m., whereas the method of Vieweg et al. (1986, 1989, 1992) used the mean of morning and afternoon urine samples taken at 7 a.m. and 4 p.m. Thus, the method used in this study to estimate 24-h urine volumes cannot be considered to have been adequately validated. Further, doubt regarding the validity of the measures used arises from a reanalysis of the case by case data, which were provided. The authors were able to examine the relationships between the variables used to estimate the extent of drinking during the baseline period: calculated urine output showed no significant correlation with either percent diurnal weight gain (r = .2075, P = .476) or serum sodium (r = .2329, P = .423). We were also able to examine the relationships between the indices of improvement provided in this paper: percentage improvement in calculated urine output, percentage improvement in diurnal weight gain, percentage improvement in urine osmolality, and absolute improvement in serum sodium. There were no significant correlations between these measures, except for that between the improvements in urine osmolality and calculated urine output (r = .5301, P = .051), but these two measures were both based on the afternoon urine samples, one relying on urine creatinine concentration and the other being urine osmolality, measures highly correlated. These concerns about the validity of the measures used in the study by Greendyke et al. (1998) serve to weaken the conclusions of this paper that clonidine is an effective antidipsic agent. 4.3. A comparison of studies conducted to date The authors tested the hypothesis that the response rates reported in the three studies were equal, using Fisher’s Exact Test for a three outcomes (‘‘improved,’’ ‘‘no response,’’ and ‘‘worse’’) by three studies contingency table. Four patients worsened in the study by Hayashi et al. (1997), three patients failed to respond in our study, and seven patients improved in the study Greendyke et al. (1998). If one categorizes the seven patients who failed to respond positively to clonidine in the Greendyke et al.’s (1998) study as simply showing no response, the hypothesis is strongly rejected ( P < .0001) by the Fisher’s Exact Test. The criterion for improvement in the study by Greendyke et al. (1998) was a mean change in the calculated urine output and the urine osmolality of greater than 30%; if we employ a similar criterion for worsening (i.e., a mean change in these two variables of greater than 30%), one of their patients can be said to have worsened while receiving the clonidine treatment. If we repeat the Fisher’s Exact Test using the following numbers for the study by
391
Greendyke et al. (1998), seven improved, six no response, and one worse, the P value was only slightly increased ( P < .001). This analysis provides clear evidence of significantly different response rates in the three studies. Of the 21 patients with schizophrenia and primary polydipsia treated to date with clonidine, seven showed improvement, five showed worsening, and nine showed no change. The failure of clonidine to show robust antidipsic effects in our patients and in the other 18 patients who have been examined in the studies of Hayashi et al. (1997) and Greendyke et al. (1998) suggests that the brain mechanisms controlling thirst may not be disturbed in this disorder, but that instead primary polydipsia of this type may be a result of dysfunction in brain areas other than those controlling thirst. Goldman and Janecek (1991), in a retrospective open study on five patients, found that fluoxetine did not affect water intake in schizophrenic patients with histories of polydipsia and water intoxication. The investigators argued that because fluoxetine and other serotonin reuptake inhibitors are effective in the treatment of obsessive-compulsive disorder, the failure of this drug to reduce drinking in their patients was evidence against this type of primary polydipsia being an obsessive-compulsive phenomenon. In a similar fashion, it could be argued that the failure of clonidine to affect drinking in a reliable fashion is evidence that mechanisms of thirst are not fundamentally affected in these patients. Thus, our findings suggest that the neuropatholophysiological basis for this disorder may lie in other areas. In fact, thirst is not the reason for drinking usually advanced by patients (Crammer, 1991; Millson et al., 1992). To date, the most effective pharmacological treatments for the primary polydipsia seen in chronic mental patients have been the conventional (Vieweg, 1994) and, more recently, the atypical (Spears et al., 1996) neuroleptics rather than drugs found to manifest antidipsic effects in laboratory animals.
5. Conclusions Our finding of a lack of effect of clonidine on drinking in patients with primary polydipsia and the inconsistent results of the studies by Hayashi et al. (1997) and Greendyke et al. (1998) provide little overall support for the effectiveness of clonidine treatment in primary polydipsia associated with schizophrenia.
Acknowledgments This work was supported by a bequest and the Queen’s University Work Study Program, which funded Ms. A. Chang. The authors wish to thank the patients who participated in the study and the diligent staff of Kingston Psychiatric Hospital.
392
N.J. Delva et al. / Progress in Neuro-Psychopharmacology & Biological Psychiatry 26 (2002) 387–392
References Alexander, R.C., Ilkowsky Karp, B., Thompson, S., Khot, V., Kirch, D.G., 1991. A double blind, placebo-controlled trial of demeclocycline treatment of polydipsia-hyponatremia in chronically psychotic patients. Biol. Psychiatry 30, 417 – 420. American Psychiatric Association, 1994. Diagnostic and Statistical Manual of Mental Disorders, fourth ed. American Psychiatric Press, Washington, DC. Blum, A., Friedland, G.W., 1983. Urinary tract abnormalities due to chronic psychogenic polydipsia. Am. J. Psychiatry 140, 915 – 916. Crammer, J.L., 1991. Drinking, thirst and water intoxication. Br. J. Psychiatry 159, 83 – 89. De Leon, J., Verghese, C., Tracy, J.I., Josiassen, R.C., Simpson, G.M., 1994. Polydipsia and water intoxication in psychiatric patients: a review of the epidemiological literature. Biol. Psychiatry 35, 408 – 419. De Leon, J., Dadvand, M., Canuso, C., Odom-White, A., Stanilla, J., Simpson, G.M., 1996. Polydipsia and water intoxication in a longterm psychiatric hospital. Biol. Psychiatry 40, 28 – 34. Delva, N.J., Crammer, J.L., 1988. Polydipsia in chronic psychiatric patients — body weight and plasma sodium. Br. J. Psychiatry 152, 242 – 245. Delva, N.J., Crammer, J.L., Jarzylo, S.V., Lawson, J.S., Owen, J.A., Sribney, M., Weir, B.J., Yendt, E.R., 1989. Osteopenia, pathological fractures, and increased urinary calcium excretion in schizophrenic patients with polydipsia. Biol. Psychiatry 26, 781 – 793. Fregly, M.J., Kelleher, D.L., 1980. Antidipsogenic effect of clonidine on isoproterenol-induced water intake. Appetite 1, 279 – 289. Fregly, M.J., Kelleher, D.L., Greenleaf, J.E., 1981. Antidipsogenic effect of clonidine on angiotensin II-, hypertonic saline-, pilocarpine- and dehydration-induced water intakes. Brain Res. Bull. 7, 661 – 664. Fregly, M.J., Rowland, N.E., Greenleaf, J.E., 1984a. Clonidine antagonism of angiotensin-related drinking: a central site of action. Brain Res. 298, 321 – 327. Fregly, M.J., Rowland, N.E., Greenleaf, J.E., 1984b. A role for presynaptic a2-adrenoceptors in angiotensin II-induced drinking in rats. Brain Res. Bull. 12, 393 – 398. Fregly, M.J., Rowland, N.E., Williams, C.M., Greenleaf, J.E., 1984c. Effect of intracerebroventricularly administered octopamines and synephrines on angiotensin II-induced water intake in rats. Brain Res. Bull. 13, 293 – 297. Goldman, M.B., Janecek, H.M., 1991. Is compulsive drinking a compulsive behaviour? A pilot study. Biol. Psychiatry 29, 503 – 505. Greendyke, R.M., Bernhardt, A.J., Tasbas, H.E., Lewandowski, K.S., 1998. Polydipsia in chronic psychiatric patients: therapeutic trials of clonidine and enalapril. Neuropsychopharmacology 18, 272 – 281. Greer, R.A., Stewart, R.B., 1993. Hyponatremia and ECT. Am. J. Psychiatry 150, 1272. Hayashi, T., Nishikawa, T., Koga, I., Uchida, Y., Horiguchi, J., Yamawaki, S., 1997. Involvement of the alpha2-adrenergic system in polydipsia in schizophrenic patients: a pilot study. Psychopharmacology (Berlin) 130, 382 – 386. Kathol, R.G., Wilcox, J.A., Turner, R.D., Kronfol, Z., Olson, S.C., 1986. Pharmacologic approaches to psychogenic polydipsia: case reports. Prog. Neuro-Psychopharmacol. Biol. Psychiatry 10, 95 – 100.
Kobinger, W., 1978. Central a-adrenergic systems as targets for hypotensive drugs. Rev. Physiol. Biochem. Pharmacol. 81, 39 – 100. Koczapski, A.B., Ashby, Y.T., Ibraheem, S., Paredes, J., Jones, B.D., Ancill, R., 1987. ‘‘Afternoon radiator-sitting syndrome’’: hypothermia and early diagnosis of self-induced water intoxication. Br. J. Psychiatry 151, 133 – 134. Landry, P., 1995. Effect of risperidone on polydipsia and hyponatremia in schizophrenia. Can. J. Psychiatry 40, 566 – 567. Langer, S.Z., Cavero, I., Massingham, R., 1980. Recent developments in noradrenergic neurotransmission and its relevance to the mechanism of action of certain antihypertensive agents. Hypertension 2, 372 – 382. Lawson, W.B., Williams, B., Pasion, R., 1988. Effects of captopril on psychosis and disturbed water regulation. Psychopharmacol. Bull. 24, 176 – 178. Lowenthal, D.T., Matzek, K.M., MacGregor, T.R., 1988. Clinical pharmacokinetics of clonidine. Clin. Pharmacokinet. 14, 287 – 310. Millson, R.C., Koczapski, A.B., Cook, M.I., Daszkiewicz, M., 1992. A survey of patient attitudes toward self-induced water intoxication. Can. J. Psychiatry 37, 46 – 47. Nishikawa, T., Tsuda, A., Tanaka, M., Nishikawa, M., Koga, I., Uchida, Y., 1994. Naloxone attenuates drinking behaviour in psychiatric patients displaying self-induced water intoxication. Prog. Neuro-Psychopharmacol. Biol. Psychiatry 18, 149 – 153. Nixon, R.A., Rothman, J.S., Chin, W., 1982. Demeclocycline in the prophylaxis of self-induced water intoxication. Am. J. Psychiatry 139, 828 – 830. Sebastian, C.S., Sinha, D., Gulati, N.K., 1995. Serum sodium stabilizing effect of enalapril after standard water load in a patient with a history of self-induced water intoxication. Prog. Neuro-Psychopharmacol. Biol. Psychiatry 19, 655 – 663. Shevitz, S.A., Jameison, R.C., Petrie, W.M., Crook, J.E., 1980. Compulsive water drinking treated with high dose propranolol. J. Nerv. Ment. Dis. 168, 246 – 248. Spears, N.M., Leadbetter, R.A., Shutty, M.S., 1996. Clozapine treatment in polydipsia and intermittent hyponatremia. J. Clin. Psychiatry 57, 123 – 128. Vieweg, W.V.R., 1994. Treatment strategies in the polydipsia-hyponatremia syndrome. J. Clin. Psychiatry 55, 154 – 160. Vieweg, W.V.R., Rowe, W.T., David, J.J., Suteer, L.H., Spradlin, W.W., 1984. Evaluation of patients with self-induced water intoxication and schizophrenic disorders (SIWIS). J. Nerv. Ment. Dis. 172, 552 – 555. Vieweg, W.V.R., David, J.J., Rowe, W.T., Yank, G.R., Spradlin, W.W., 1986. Diurnal variation of urinary excretion for patients with psychosis, intermittent hyponatremia, and polydipsia (PIP syndrome). Biol. Psychiatry 21, 1031 – 1042. Vieweg, W.V.R., Godleski, L.S., Hundley, P.L., Yank, G.R., 1989. Survey of diurnal weight gain and urine volume in chronic schizophrenia. Can. J. Psychiatry 34, 779 – 784. Vieweg, W.V.R., Pandurangi, A.K., Pelonero, A.L., 1992. Estimating daily urine volume in chronic psychosis. Schizophr. Res. 8, 89 – 91. Wilson, K.M., Rowland, N., Fregly, M.J., 1984. Drinking: a final common pathway? Appetite 5, 31 – 38. Zubenko, G.S., 1987. Water homeostasis in psychiatric patients. Biol. Psychiatry 22, 121 – 125.
Effects of clonidine in schizophrenic patients with primary polydipsia Three single case studies Nicholas J. Delvaa,b,*, Anna Changb, Emily R. Hawkenb, J. Stuart Lawsona,b,c, James A. Owena,b,d a Department of Psychiatry, Queen’s University, Kingston, Ontario, Canada Kingston Psychiatric Hospital, Queen’s University, Kingston, Ontario, Canada c Department of Psychology, Queen’s University, Kingston, Ontario, Canada d Department of Pharmacology and Toxicology, Queen’s University, Kingston, Ontario, Canada b
Abstract A pilot study was conducted in schizophrenic patients with primary polydipsia to determine the tolerability of adding clonidine to an existing antipsychotic drug regimen and to seek evidence of an antidipsic effect. Three patients with chronic schizophrenia and primary polydipsia underwent open controlled prospective trials of treatment with clonidine in doses of up to 800 mg/day. The trials lasted from 2 to 5 months each, and analysis of variance was used to test for changes in dependent variables on a case-by-case basis. Blood pressure and pulse declined significantly in a dose-dependent manner, but fluid intake, as assessed by measurements of weight and 24-h urine volume, was not affected. Hypotension and bradycardia limited the extent to which the dose of clonidine could be increased. The lack of evident effect of clonidine on polydipsia in this small sample and the inconsistent results of two other recent studies of clonidine in patients with schizophrenia and primary polydipsia provide little overall support for the effectiveness of clonidine treatment in primary polydipsia associated with schizophrenia. D 2001 Elsevier Science Inc. All rights reserved. Keywords: Antidipsic drugs; Clonidine; Primary polydipsia; Schizophrenia; Thirst
1. Introduction Severe primary polydipsia accompanied by hyponatremia is a common finding in mental hospital patients. In their review of the epidemiological literature in this area, De Leon et al. (1994) concluded that more than 20% of chronic inpatients suffer from polydipsia; water intoxication occurs in about one of four experiencing polydipsia, and schizophrenic patients account for 80% of those with water intoxication. The most recently published epidemiological study (De Leon et al., 1996) identified 26% of chronic inpatients at a US state mental hospital as polydipsic, with water intoxication observed in about one of five polydipsic patients. Patients with primary polydipsia may suffer from a number of sequelae apart from water intoxication, which
Abbreviations: ANOVA, One-way analysis of variance * Corresponding author. Department of Psychiatry, Queen’s University, 72 Barrie Street, Kingston, Ontario, Canada K7L 3J7. Tel.: +1-613-5482372; fax: +1-613-548-6095.
itself may be fatal. These include bladder dilation, enuresis, incontinence, hydronephrosis, renal failure, gastrointestinal dilation and hypotonicity, hernias, hypocalcemia, osteoporosis, hypothermia, and congestive heart failure (Blum and Friedland, 1983; Vieweg et al., 1984; Koczapski et al., 1987; Zubenko, 1987; Delva et al., 1989). Approaches to treatment have included weight monitoring with fluid restriction (Delva and Crammer, 1988), behavioral methods (Vieweg, 1994), electroconvulsive therapy (Greer and Stewart, 1993), and medication. Drug therapy has been directed toward reduction of drinking on the one hand and increasing water clearance on the other. Antipsychotic medication can reduce both psychotic symptoms and polydipsia (Vieweg, 1994), and some polydipsic patients who fail to benefit from typical neuroleptics may respond to atypical neuroleptics such as clozapine (Spears et al., 1996) or risperidone (Landry, 1995) with a decrease in drinking as well as better control of the psychosis, although Millson et al. (1992) failed to find an antidipsic effect from risperidone in a study on eight patients. Other treatments assessed for potential antidipsic efficacy have
0278-5846/02/$ – see front matter D 2001 Elsevier Science Inc. All rights reserved. PII: S 0 2 7 8 - 5 8 4 6 ( 0 1 ) 0 0 2 4 6 - 9
388
N.J. Delva et al. / Progress in Neuro-Psychopharmacology & Biological Psychiatry 26 (2002) 387–392
included fluoxetine (Goldman and Janecek, 1991), propranolol (Shevitz et al., 1980; Kathol et al., 1986), angiotensinconverting enzyme inhibitors (Kathol et al., 1986; Lawson et al., 1988; Sebastian et al., 1995), and naloxone (Nishikawa et al., 1994). While some case reports have appeared promising, the better controlled trials employing larger numbers of patients have, in general, been disappointing. Treatments directed at increasing water clearance have included the combination of lithium and phenytoin, and demeclocycline (Nixon et al., 1982; Vieweg, 1994). While the lithium/phenytoin treatment may be helpful, a controlled trial has shown no benefit from demeclocycline (Alexander et al., 1991). Clonidine, an a2 adrenergic agonist, was thought to be of potential value in treating primary polydipsia because it reduces drinking in animal models of thirst. Fregly et al. have shown that in rats clonidine attenuates water intake induced by angiotensin II, isoproterenol, pilocarpine, hypertonic saline, dehydration, or 5-hydroxytryptophan but fails to affect water intake significantly when given alone (Fregly and Kelleher, 1980; Fregly et al., 1981, 1984a,b,c; Wilson et al., 1984). On the basis of these findings and the current limitations of treatment for this serious condition, two recent studies have evaluated the effectiveness of clonidine in polydipsia patients, with contradictory results: Hayashi et al. (1997) reported that patients became more severely polydispic while receiving clonidine but Greendyke et al. (1998) found that clonidine had an antidipsic effect. The authors studied the effectiveness of clonidine in three patients with schizophrenia and significant primary polydipsia using a single case study design for each patient.
2. Methods 2.1. Subjects Patients with chronic polydipsia were identified by a survey. Three male patients consented to undergo a trial of clonidine, which had been approved by the institutional Research Ethics Board; written informed consent for the trial was obtained in each case. By DSM-IV criteria (American Psychiatric Association, 1994), all three suffered from chronic schizophrenia; one suffered a significant head injury after the onset of the schizophrenia. The mean age was 40.7 years (range 39 – 44) and mean duration of polydipsia, defined as the number of years since hyposthenuria (urine specific gravity 1.003), hyponatremia, obvious polydipsia, or any other complication was first noted, was 9.7 years (range 5 – 12). Complications related to polydipsia were present in all three subjects. All had osteoporosis and incontinence. Two had also suffered hyponatremic seizures. They all received calcium (0.22 g elemental calcium by mouth daily) and ergocalciferol (50,000 IU by mouth once weekly) for osteoporosis and neuroleptics (listed in Table 1).
One patient received diazepam on an as-needed basis (10 mg for agitation); this was given only three times in the course of his clonidine trial and did not have any effect on water intake. Each patient smoked approximately one-half package of cigarettes per day. 2.2. Procedures The study design consisted of the nonblind administration of clonidine in gradually increasing (to assess potential benefit and side effects) and then gradually decreasing doses (to avoid rebound hypertension) while other medication, other treatment, and ward environment were held constant. The length of study varied from subject to subject depending on the results of treatment. Each patient was allowed to drink freely unless his weight exceeded 7% over that while normally hydrated. Above this threshold, drinking was discouraged according to our usual practice (Delva and Crammer, 1988). Regular measurements were made of pulse rate, blood pressure, body weight, urine volume, and urine creatinine. These variables were measured before and during administration of clonidine (range: 50– 800 mg/day) over a time period ranging from 9 to 23 weeks. Clonidine was administered in divided doses, from two to four times each day, in order to minimize fluctuations in plasma levels. Pulse, blood pressure, and weight were measured daily, and urine was collected every third to fourth day. In this way, adequate data were collected for each dose of clonidine in each patient. Pulse and blood pressure were measured at 08:00 and 18:00 h; this was done after the patient had been lying for 5 min, sitting for 2 min, and standing for 2 min. Patients were weighed hourly while awake, with an empty bladder and in the same clothes. In order to ensure that the urine collections were complete, nursing staff observed the patient constantly during each urine collection; urine was collected over a 24-h period starting at 06:00 h, once every hour while awake and every 2 h during sleep. A sample of each 24-h urine collection was assayed for creatinine; the ratio of urine volume/creatinine was used to correct for any possible losses of urine during the collections. 2.3. Data analysis Mean arterial pressure was calculated as the diastolic pressure (represented by the point at which the Korotkoff sound completely disappeared) plus one-third of the difference between the systolic and diastolic pressures. Diurnal weight gain was calculated as the difference between the mean of the weights measured at 18:00, 19:00, and 20:00 h (representing the highest weights of the day) and the mean of those at 06:00 and 07:00 h (representing the lowest). In the statistical analysis, each case was examined separately. The phase of increasing dosage of clonidine
N.J. Delva et al. / Progress in Neuro-Psychopharmacology & Biological Psychiatry 26 (2002) 387–392
389
Table 1 Duration of study, number of 24-h urine collections, neuroleptic treatment, maximum dose of clonidine given, and effects of clonidine treatment on pulse and arterial pressure
Duration of study (days) Number of urine collections Neuroleptic/dose (mg/day) Maximum dose of clonidine (mg/day) Pulse, lying (bpm ± S.D.) Baseline/maximum dose Pulse, standing (bpm ± S.D.) Baseline/maximum dose MAP, lying (mmHg ± S.D.) Baseline/maximum dose MAP, standing (mmHg ± S.D.) Baseline/maximum dose
Patient 1
Patient 2
Patient 3
107 35 Haloperidol 25 800 77 ± 5/60 ± 5z,a 88 ± 8/72 ± 7z,d 92 ± 3/82 ± 6**,g 96 ± 5/82 ± 7y,j
159 40 Haloperidol 25 800 79 ± 6/69 ± 4**,b 88 ± 11/75 ± 10*,e 89 ± 5/83 ± 0y,h 91 ± 15/72 ± 1y,k
65 20 Thioridazine 300 250 71 ± 6/66 ± 13NS,c 83 ± 8/79 ± 4NS,f 92 ± 7/82 ± 6y,i 92 ± 7/75 ± 3z,l
MAP, mean arterial pressure; bpm, beats per minute. * P < .05. ** P < .01. y P < .001. z P < .0001. NS Not significant. a ANOVA for phase of increasing dose of clonidine: F7,56 = 9.29. b ANOVA for phase of increasing dose of clonidine: F8,107 = 3.50. c ANOVA for phase of increasing dose of clonidine: F5,50 = 0.968. d ANOVA for phase of increasing dose of clonidine: F7,54 = 6.13. e ANOVA for phase of increasing dose of clonidine: F8,104 = 2.04. f ANOVA for phase of increasing dose of clonidine: F5,50 = 1.19. g ANOVA for phase of increasing dose of clonidine: F7,55 = 3.36. h ANOVA for phase of increasing dose of clonidine: F8,107 = 4.37. i ANOVA for phase of increasing dose of clonidine: F5,51 = 3.80. j ANOVA for phase of increasing dose of clonidine: F7,56 = 4.32. k ANOVA for phase of increasing dose of clonidine: F8,107 = 3.68. l ANOVA for phase of increasing dose of clonidine: F5,49 = 7.72.
was treated separately from that of decreasing dosage. Oneway analysis of variance (ANOVA) was employed for each dependent variable (percentage diurnal weight gain; 24-h urine volume; urine volume/creatinine ratio; pulse, standing; pulse, lying; blood pressure, standing; blood pressure, lying) during each phase, with the dosage of clonidine administered being the independent variable. Thus, separate ANOVAs were performed for the phase in which the dose of clonidine increased and that during which it decreased. If the ANOVA yielded statistically significant results, a poly-
nomial regression analysis of the dependent variable on dosage was performed.
3. Results Although the patients knew that their drinking behaviour was under study, and the investigators and ward staff had hopes that clonidine would prove to be an effective antidipsic agent, there was no evidence that the patients were
Table 2 Body weight, diurnal weight gain, urine volume, and urine volume/creatinine ratio during entire period of study
Weight (normal hydration) Diurnal weight gain (kg ± S.D.) Percentage diurnal weight gain (% ± S.D.) Urine volume (l/24 h ± S.D.) Urine: ratio of volume to creatinine (ml/mg ± S.D.) NS a b c d e f g h i
Not significant. ANOVA for phase of increasing dose of clonidine: F7,57 = 0.46. ANOVA for phase of increasing dose of clonidine: F8,106 = 1.43. ANOVA for phase of increasing dose of clonidine: F5,54 = 1.179. ANOVA for phase of increasing dose of clonidine: F7,17 = 2.13. ANOVA for phase of increasing dose of clonidine: F8,25 = 0.58. ANOVA for phase of increasing dose of clonidine: F3,13 = 0.171. ANOVA for phase of increasing dose of clonidine: F7,16 = 1.26. ANOVA for phase of increasing dose of clonidine: F8,25 = 0.74. ANOVA for phase of increasing dose of clonidine: F3,13 = 1.40.
Patient 1
Patient 2
Patient 3
61 kg 2.5 ± 0.7 4 ± 1NS,a 7.4 ± 2.4NS,d 740 ± 313NS,g
53 kg 1.9 ± 0.6 4 ± 1NS,b 7.6 ± 2.2NS,e 627 ± 197NS,h
75 kg 2.5 ± 1.0 3 ± 1NS,c 4.9 ± 2.4NS,f 524 ± 260NS,i
390
N.J. Delva et al. / Progress in Neuro-Psychopharmacology & Biological Psychiatry 26 (2002) 387–392
motivated to alter their drinking behaviour either to please or oppose the investigators and staff. Table 1 shows the duration of the trial in each patient, the number of urine collections, concurrent neuroleptic treatment, the maximum dose of clonidine administered, the effects of clonidine on cardiovascular function, and the ANOVAs for the period of increasing dose; there were no substantial differences for the period of decreasing dose. All three patients experienced clinically and statistically significant reductions in arterial pressure as the dose of clonidine increased. Patients 1 and 2 manifested significant reductions in pulse rate. At maximal doses of clonidine, the overall mean reductions in standing pulse and mean arterial pressure were 11 beats per minute and 17 mmHg, respectively. For each case in which there was a statistically significant drug effect, there was a significant linear relationship with dose, but there was no consistent contribution by nonlinear terms up to the quintic. The administration of clonidine did not affect polydipsia as assessed by diurnal weight gain, urine volume, or urine volume/creatinine ratio. Mean values of these variables and the ANOVAs for the period of increasing dose are shown in Table 2; there were no substantial differences for the period of decreasing dose. A moderate to severe degree of polydipsia is evident.
4. Discussion In the single case approach used in the present study, the dependent measures were sampled prospectively over long periods of time, with each patient representing a ‘‘population’’ from the statistical point of view. The measures included both direct (timed urine collections) and indirect (hourly body weights) assessments. This is the only study to date on the use of clonidine for the treatment of primary polydipsia in which actual rather than estimated urine values were employed. We can say with confidence that in these three polydipsic patients, the administration of clonidine at doses sufficient to produce clinically significant hypotension and bradycardia failed to affect their daily pattern of excessive drinking during the waking hours and normalization of hydration during sleep. 4.1. Effects of clonidine on heart rate and blood pressure and in animal models of thirst Because the hypotensive effect of clonidine appears to be mediated mainly by stimulation of central a2 adrenergic receptors (Kobinger, 1978; Langer et al., 1980; Lowenthal et al., 1988), the profound cardiovascular effects induced by clonidine indicate significant central nervous system penetration in our three patients. Clonidine reduces heart rate and blood pressure by its action within the nucleus tractus solitarius of the medulla; clonidine appears to produce its antidipsic effect in animal models of thirst by action on the central nervous system, possibly also in the nucleus tractus
solitarius (Fregly et al., 1980; Wilson et al., 1984). It is possible that higher doses of clonidine may have resulted in an effect on the water intake of our patients but this was not possible to assess without jeopardising patient safety. The concurrent treatment with neuroleptic medication possessing a1 adrenergic antagonist effects may have increased the degree of hypotension observed. In the studies of Fregly’s group (Fregly and Kelleher, 1980; Fregly et al., 1981, 1984a,b,c; Wilson et al., 1984), antidipsic effects were observed in rats with doses of clonidine as low as 6 mg/kg given intraperitoneally, although fourfold higher doses were required for antagonism of dehydration-induced drinking (Fregly and Kelleher, 1980; Fregly et al., 1981). In our study, the maximum doses ranged from 3 to 15 mg/kg per os. Given that the oral bioavailability of clonidine in humans is 75 – 95% (Lowenthal et al., 1988), the doses of clonidine in our study are in approximately the same range of dosage as those used in the rat model studies of Fregly’s group. Apart from the issue of dose, it is also possible that the difference in effect of clonidine in our study in comparison with those of Fregly’s group might be the result of the difference in species evaluated. 4.2. Studies on the treatment of polydipsia with clonidine If examined in isolation, the small size of our patient sample might limit the generalizability of our findings to the population of chronically psychotic patients with primary polydipsia, but when looked at in the context of the other two studies of the effects clonidine in these patients (Greendyke et al., 1998; Hayashi et al., 1997), it adds support to the view that clonidine is not likely to be an effective antidipsic drug in polydipsic patients. Hayashi et al. (1997), in an open-label controlled study, found that all four of their polydispic patients experienced increased drinking with increasing doses of clonidine over a 6-week treatment course (a fifth patient was withdrawn from the study), but Greendyke et al. (1998), in a double-blind, placebo-controlled study, found that 7 of their 14 patients showed improvement after receiving clonidine for 4 weeks. Hayashi et al. (1997) examined the effect of clonidine in polydipsic patients, expecting that it would reduce drinking, but found instead that it exacerbated the condition, as measured by diurnal weight gain. An unanticipated and apparently strong contrary effect would generally be expected to be a genuine finding, especially in an openlabel study. Furthermore, treatment with mianserin, an a2-adrenergic antagonist, resulted in reduced diurnal weight gain. These findings would, however, be more convincing if the withdrawal phases had produced results opposite to the treatment phases: clonidine is known to produce marked cardiovascular withdrawal symptoms, but the withdrawal phase in the study of Hayashi et al. (1997) was associated with a continued prodipsic effect. Similarly, the withdrawal phase of mianerin treatment was associated with an antidipsic effect in three of the four patients. One
N.J. Delva et al. / Progress in Neuro-Psychopharmacology & Biological Psychiatry 26 (2002) 387–392
patient showed increased drinking with both clonidine and mianserin. Vital signs were measured but not reported in this study. The study by Greendyke et al. (1998) has the advantage of being double blind. The method of estimating urine output employed in this study was based on that of Vieweg et al. (1986, 1989, 1992) but involved only a single sample of urine, taken at 3 p.m., whereas the method of Vieweg et al. (1986, 1989, 1992) used the mean of morning and afternoon urine samples taken at 7 a.m. and 4 p.m. Thus, the method used in this study to estimate 24-h urine volumes cannot be considered to have been adequately validated. Further, doubt regarding the validity of the measures used arises from a reanalysis of the case by case data, which were provided. The authors were able to examine the relationships between the variables used to estimate the extent of drinking during the baseline period: calculated urine output showed no significant correlation with either percent diurnal weight gain (r = .2075, P = .476) or serum sodium (r = .2329, P = .423). We were also able to examine the relationships between the indices of improvement provided in this paper: percentage improvement in calculated urine output, percentage improvement in diurnal weight gain, percentage improvement in urine osmolality, and absolute improvement in serum sodium. There were no significant correlations between these measures, except for that between the improvements in urine osmolality and calculated urine output (r = .5301, P = .051), but these two measures were both based on the afternoon urine samples, one relying on urine creatinine concentration and the other being urine osmolality, measures highly correlated. These concerns about the validity of the measures used in the study by Greendyke et al. (1998) serve to weaken the conclusions of this paper that clonidine is an effective antidipsic agent. 4.3. A comparison of studies conducted to date The authors tested the hypothesis that the response rates reported in the three studies were equal, using Fisher’s Exact Test for a three outcomes (‘‘improved,’’ ‘‘no response,’’ and ‘‘worse’’) by three studies contingency table. Four patients worsened in the study by Hayashi et al. (1997), three patients failed to respond in our study, and seven patients improved in the study Greendyke et al. (1998). If one categorizes the seven patients who failed to respond positively to clonidine in the Greendyke et al.’s (1998) study as simply showing no response, the hypothesis is strongly rejected ( P < .0001) by the Fisher’s Exact Test. The criterion for improvement in the study by Greendyke et al. (1998) was a mean change in the calculated urine output and the urine osmolality of greater than 30%; if we employ a similar criterion for worsening (i.e., a mean change in these two variables of greater than 30%), one of their patients can be said to have worsened while receiving the clonidine treatment. If we repeat the Fisher’s Exact Test using the following numbers for the study by
391
Greendyke et al. (1998), seven improved, six no response, and one worse, the P value was only slightly increased ( P < .001). This analysis provides clear evidence of significantly different response rates in the three studies. Of the 21 patients with schizophrenia and primary polydipsia treated to date with clonidine, seven showed improvement, five showed worsening, and nine showed no change. The failure of clonidine to show robust antidipsic effects in our patients and in the other 18 patients who have been examined in the studies of Hayashi et al. (1997) and Greendyke et al. (1998) suggests that the brain mechanisms controlling thirst may not be disturbed in this disorder, but that instead primary polydipsia of this type may be a result of dysfunction in brain areas other than those controlling thirst. Goldman and Janecek (1991), in a retrospective open study on five patients, found that fluoxetine did not affect water intake in schizophrenic patients with histories of polydipsia and water intoxication. The investigators argued that because fluoxetine and other serotonin reuptake inhibitors are effective in the treatment of obsessive-compulsive disorder, the failure of this drug to reduce drinking in their patients was evidence against this type of primary polydipsia being an obsessive-compulsive phenomenon. In a similar fashion, it could be argued that the failure of clonidine to affect drinking in a reliable fashion is evidence that mechanisms of thirst are not fundamentally affected in these patients. Thus, our findings suggest that the neuropatholophysiological basis for this disorder may lie in other areas. In fact, thirst is not the reason for drinking usually advanced by patients (Crammer, 1991; Millson et al., 1992). To date, the most effective pharmacological treatments for the primary polydipsia seen in chronic mental patients have been the conventional (Vieweg, 1994) and, more recently, the atypical (Spears et al., 1996) neuroleptics rather than drugs found to manifest antidipsic effects in laboratory animals.
5. Conclusions Our finding of a lack of effect of clonidine on drinking in patients with primary polydipsia and the inconsistent results of the studies by Hayashi et al. (1997) and Greendyke et al. (1998) provide little overall support for the effectiveness of clonidine treatment in primary polydipsia associated with schizophrenia.
Acknowledgments This work was supported by a bequest and the Queen’s University Work Study Program, which funded Ms. A. Chang. The authors wish to thank the patients who participated in the study and the diligent staff of Kingston Psychiatric Hospital.
392
N.J. Delva et al. / Progress in Neuro-Psychopharmacology & Biological Psychiatry 26 (2002) 387–392
References Alexander, R.C., Ilkowsky Karp, B., Thompson, S., Khot, V., Kirch, D.G., 1991. A double blind, placebo-controlled trial of demeclocycline treatment of polydipsia-hyponatremia in chronically psychotic patients. Biol. Psychiatry 30, 417 – 420. American Psychiatric Association, 1994. Diagnostic and Statistical Manual of Mental Disorders, fourth ed. American Psychiatric Press, Washington, DC. Blum, A., Friedland, G.W., 1983. Urinary tract abnormalities due to chronic psychogenic polydipsia. Am. J. Psychiatry 140, 915 – 916. Crammer, J.L., 1991. Drinking, thirst and water intoxication. Br. J. Psychiatry 159, 83 – 89. De Leon, J., Verghese, C., Tracy, J.I., Josiassen, R.C., Simpson, G.M., 1994. Polydipsia and water intoxication in psychiatric patients: a review of the epidemiological literature. Biol. Psychiatry 35, 408 – 419. De Leon, J., Dadvand, M., Canuso, C., Odom-White, A., Stanilla, J., Simpson, G.M., 1996. Polydipsia and water intoxication in a longterm psychiatric hospital. Biol. Psychiatry 40, 28 – 34. Delva, N.J., Crammer, J.L., 1988. Polydipsia in chronic psychiatric patients — body weight and plasma sodium. Br. J. Psychiatry 152, 242 – 245. Delva, N.J., Crammer, J.L., Jarzylo, S.V., Lawson, J.S., Owen, J.A., Sribney, M., Weir, B.J., Yendt, E.R., 1989. Osteopenia, pathological fractures, and increased urinary calcium excretion in schizophrenic patients with polydipsia. Biol. Psychiatry 26, 781 – 793. Fregly, M.J., Kelleher, D.L., 1980. Antidipsogenic effect of clonidine on isoproterenol-induced water intake. Appetite 1, 279 – 289. Fregly, M.J., Kelleher, D.L., Greenleaf, J.E., 1981. Antidipsogenic effect of clonidine on angiotensin II-, hypertonic saline-, pilocarpine- and dehydration-induced water intakes. Brain Res. Bull. 7, 661 – 664. Fregly, M.J., Rowland, N.E., Greenleaf, J.E., 1984a. Clonidine antagonism of angiotensin-related drinking: a central site of action. Brain Res. 298, 321 – 327. Fregly, M.J., Rowland, N.E., Greenleaf, J.E., 1984b. A role for presynaptic a2-adrenoceptors in angiotensin II-induced drinking in rats. Brain Res. Bull. 12, 393 – 398. Fregly, M.J., Rowland, N.E., Williams, C.M., Greenleaf, J.E., 1984c. Effect of intracerebroventricularly administered octopamines and synephrines on angiotensin II-induced water intake in rats. Brain Res. Bull. 13, 293 – 297. Goldman, M.B., Janecek, H.M., 1991. Is compulsive drinking a compulsive behaviour? A pilot study. Biol. Psychiatry 29, 503 – 505. Greendyke, R.M., Bernhardt, A.J., Tasbas, H.E., Lewandowski, K.S., 1998. Polydipsia in chronic psychiatric patients: therapeutic trials of clonidine and enalapril. Neuropsychopharmacology 18, 272 – 281. Greer, R.A., Stewart, R.B., 1993. Hyponatremia and ECT. Am. J. Psychiatry 150, 1272. Hayashi, T., Nishikawa, T., Koga, I., Uchida, Y., Horiguchi, J., Yamawaki, S., 1997. Involvement of the alpha2-adrenergic system in polydipsia in schizophrenic patients: a pilot study. Psychopharmacology (Berlin) 130, 382 – 386. Kathol, R.G., Wilcox, J.A., Turner, R.D., Kronfol, Z., Olson, S.C., 1986. Pharmacologic approaches to psychogenic polydipsia: case reports. Prog. Neuro-Psychopharmacol. Biol. Psychiatry 10, 95 – 100.
Kobinger, W., 1978. Central a-adrenergic systems as targets for hypotensive drugs. Rev. Physiol. Biochem. Pharmacol. 81, 39 – 100. Koczapski, A.B., Ashby, Y.T., Ibraheem, S., Paredes, J., Jones, B.D., Ancill, R., 1987. ‘‘Afternoon radiator-sitting syndrome’’: hypothermia and early diagnosis of self-induced water intoxication. Br. J. Psychiatry 151, 133 – 134. Landry, P., 1995. Effect of risperidone on polydipsia and hyponatremia in schizophrenia. Can. J. Psychiatry 40, 566 – 567. Langer, S.Z., Cavero, I., Massingham, R., 1980. Recent developments in noradrenergic neurotransmission and its relevance to the mechanism of action of certain antihypertensive agents. Hypertension 2, 372 – 382. Lawson, W.B., Williams, B., Pasion, R., 1988. Effects of captopril on psychosis and disturbed water regulation. Psychopharmacol. Bull. 24, 176 – 178. Lowenthal, D.T., Matzek, K.M., MacGregor, T.R., 1988. Clinical pharmacokinetics of clonidine. Clin. Pharmacokinet. 14, 287 – 310. Millson, R.C., Koczapski, A.B., Cook, M.I., Daszkiewicz, M., 1992. A survey of patient attitudes toward self-induced water intoxication. Can. J. Psychiatry 37, 46 – 47. Nishikawa, T., Tsuda, A., Tanaka, M., Nishikawa, M., Koga, I., Uchida, Y., 1994. Naloxone attenuates drinking behaviour in psychiatric patients displaying self-induced water intoxication. Prog. Neuro-Psychopharmacol. Biol. Psychiatry 18, 149 – 153. Nixon, R.A., Rothman, J.S., Chin, W., 1982. Demeclocycline in the prophylaxis of self-induced water intoxication. Am. J. Psychiatry 139, 828 – 830. Sebastian, C.S., Sinha, D., Gulati, N.K., 1995. Serum sodium stabilizing effect of enalapril after standard water load in a patient with a history of self-induced water intoxication. Prog. Neuro-Psychopharmacol. Biol. Psychiatry 19, 655 – 663. Shevitz, S.A., Jameison, R.C., Petrie, W.M., Crook, J.E., 1980. Compulsive water drinking treated with high dose propranolol. J. Nerv. Ment. Dis. 168, 246 – 248. Spears, N.M., Leadbetter, R.A., Shutty, M.S., 1996. Clozapine treatment in polydipsia and intermittent hyponatremia. J. Clin. Psychiatry 57, 123 – 128. Vieweg, W.V.R., 1994. Treatment strategies in the polydipsia-hyponatremia syndrome. J. Clin. Psychiatry 55, 154 – 160. Vieweg, W.V.R., Rowe, W.T., David, J.J., Suteer, L.H., Spradlin, W.W., 1984. Evaluation of patients with self-induced water intoxication and schizophrenic disorders (SIWIS). J. Nerv. Ment. Dis. 172, 552 – 555. Vieweg, W.V.R., David, J.J., Rowe, W.T., Yank, G.R., Spradlin, W.W., 1986. Diurnal variation of urinary excretion for patients with psychosis, intermittent hyponatremia, and polydipsia (PIP syndrome). Biol. Psychiatry 21, 1031 – 1042. Vieweg, W.V.R., Godleski, L.S., Hundley, P.L., Yank, G.R., 1989. Survey of diurnal weight gain and urine volume in chronic schizophrenia. Can. J. Psychiatry 34, 779 – 784. Vieweg, W.V.R., Pandurangi, A.K., Pelonero, A.L., 1992. Estimating daily urine volume in chronic psychosis. Schizophr. Res. 8, 89 – 91. Wilson, K.M., Rowland, N., Fregly, M.J., 1984. Drinking: a final common pathway? Appetite 5, 31 – 38. Zubenko, G.S., 1987. Water homeostasis in psychiatric patients. Biol. Psychiatry 22, 121 – 125.