- Email: [email protected]
Lithium prophylaxis of recurrent bipolar affective disorder: Long-term outcome and its psychosocial correlates
Journal of Affective Disorders 54 (1999) 87–96
Research report
Lithium prophylaxis of recurrent bipolar affective disorder: Long-term outcome and its psychosocial correlates a, a a a Parmanand Kulhara *, Debasish Basu , Surendra K. Mattoo , Pratap Sharan , b Rajni Chopra a
Department of Psychiatry, Postgraduate Institute of Medical Education and Research, Chandigarh 160 012, India b Institute of Health and Family Welfare, New Delhi, India Received 26 May 1998; received in revised form 27 June 1998; accepted 13 July 1998
Abstract Background: Discrepancy between efficacy of prophylactic lithium and its effectiveness in ordinary clinical practice necessitates long-term follow-up data from specialised lithium clinics. Also, role of psychosocial factors in influencing the outcome is unclear. Methods: One hundred and eighteen patients of bipolar affective disorder attending a lithium clinic were followed-up for | 11 years (range 2–27 years). Demographic and clinical data, measures of social support and psychosocial stress were obtained at the intake in 1989–1990. Study design combined retrospective chart-review (till the time of intake) with prospective follow-up till July 1995. Results: On lithium, the patients had a mean of 0.43 relapses per year (manic, 0.26; depressive, 0.17) which was significantly less ( p , 0.01) than the pre-lithium episode frequency. The figure for entirely relapse-free patients was 24%, and 62% had relapses up to one episode per year (median 5 0.3 per year). Fifty-eight (49%) patients were good responders to lithium (relapses # 0.30 per year). In comparison to good responders, partial / poor responders had a significantly greater number of pre-lithium depressive episodes, poor lithium compliance, more psychosocial stress and lower social support at intake. These variables correlated well with relapses and explained 32% of the variance of the data. Conclusions: Lithium had a definite prophylactic effect on long-term outcome. Social support and stressful life events are significant correlates of response to lithium. Clinical Implications: Lithium prophylaxis of bipolar affective disorders seems justified though psychosocial factors appear to modulate its effectiveness. Limitations: Other psychotropic medications were used during relapse and the assessment of psychosocial factors was cross-sectional. 1999 Elsevier Science B.V. All rights reserved. Keywords: Lithium; Bipolar affective disorders; Prophylaxis; Psychosocial factors
1. Introduction *Corresponding author. Tel.: 1 91-0172-541469-73, ext. 247; fax: 1 91-0172-540-401. E-mail address: [email protected] (P. Kulhara)
Following serendipitous discovery of the efficacy of lithium salts in controlling ‘psychotic excitement’
0165-0327 / 99 / $ – see front matter 1999 Elsevier Science B.V. All rights reserved. PII: S0165-0327( 98 )00145-1
88
P. Kulhara et al. / Journal of Affective Disorders 54 (1999) 87 – 96
(Cade, 1949) and more specifically manic psychosis (Schou et al., 1954), many more studies firmly established the prophylactic efficacy of lithium in recurrent affective disorders (Angst et al., 1970; Coppen et al., 1971; Prien et al., 1973). Studies from India also are in general agreement with this conclusion (Ghosh and Wig, 1977; Venkoba Rao and Hariharasubramanian, 1978; Tandon et al., 1981; Venkoba Rao et al., 1982; Prakash et al., 1978; Narayanan et al., 1979; Sethi et al., 1984; Khandelwal et al., 1984; Venkoba Rao, 1984; Mathew et al., 1995). Most of these Indian studies, however, were based on small samples, for example, 23 in Narayanan et al. (1979), 47 in Venkoba Rao et al. (1982), 71 in Venkoba Rao (1984), and 44 in the latest published study (Mathew et al., 1995), the largest number being 122 (Sethi et al., 1984). A small sample size obviously restricts the limits to which the findings can be generalised and extrapolated. The other major problem with these studies is relatively short duration of follow-up of the patients on lithium. Clearly, in any study on prophylactic efficacy of an agent, the longer the follow-up the better. Most of the studies cited above had follow-up periods ranging from 2 to 5 years. In the study by Sethi et al. (1984), out of the 122 patients studied, only 12 had follow-up for 5–7 years. Further, the mean or median duration of follow-up was not mentioned in most of the above studies. Our report aims to improve these lacunae. Another unresolved but important issue in this area is the role of psychosocial variables (stressor events, social support etc.) in modulating the response to lithium prophylaxis. Although the role of such social factors has been extensively studied in relation to recurrent depressive disorder and less in relation to bipolar disorder in general (O’Connell and Mayo, 1988; Silverstone and Romans-Clarkson, 1989), little research has been done so far in studying the role of these factors in modifying the effectiveness of lithium prophylaxis in bipolar affective disorders. Only a few very recent studies have looked at this issue systematically (Solomon et al., 1996; Stefos et al., 1996). More investigation as to how psychosocial variables influence outcome of lithium prophylaxis is called for (Schou, 1997), and this is the second major aim of our study. Incidentally, this study is from an institute credited
to have started one of the earliest lithium clinics in India (Khandelwal, 1991). The Department of Psychiatry of the Institute started using lithium as early as 1968 and the lithium clinic was formally started in 1975. Patients are followed-up regularly and serum lithium estimated by flame photometry method of Amdisen (1967). In view of the recently voiced doubt regarding efficacy of lithium in the prophylaxis of affective disorders when used in ordinary clinical practice (Dickson and Kendell, 1986; Markar and Mander, 1989; O’Connell et al., 1991; Guscott and Taylor, 1994) and even in classic double-blind studies of lithium (Moncrieff, 1995), it was felt appropriate at this point to make a critical and long-term assessment of the prophylactic benefits of lithium for patients of recurrent affective disorder. The protesting repercussion of an editorial (Goodwin, 1995) further sets the pace of this study. This paper describes the demographic and clinical characteristics, and outcome data of 122 patients from our lithium clinic over follow-up periods ranging from 2 to 27 years. The correlates of outcome patterns are analysed and discussed.
2. Material and methods Patients of recurrent affective disorder (manicdepressive psychosis according to International Classification of Diseases, 9th revision of the World Health Organisation, 1978, WHO) attending the Lithium Clinic, Department of Psychiatry, PGIMER, Chandigarh, between July 1989 and February 1990 comprised the sample. Most of these patients were registered in the clinic years before, including patients from 1968 onwards. All relevant data regarding the patients (demographic, clinical and therapeutic), charted in the case notes ever since registration through repeated (usually monthly) follow-up visits, were extracted. The accuracy of recording of demographic, diagnostic and clinical aspects were verified by two of us (PK and SKM) to the extent possible. It is worth pointing out that the routine case records of the patients at our lithium clinic are specifically designed for extracting data for research purposes. Other than these, the patients at intake (i.e. between July, 1989 and February, 1990) were also administered: (a) the Social Support Questionnaire
P. Kulhara et al. / Journal of Affective Disorders 54 (1999) 87 – 96
(Pollack and Harris, 1983, modified by Nehra and Kulhara, 1987), and (b) the Presumptive Stressful Life Events Scale (Singh et al., 1984), for measuring their perceived social support and life event-related stresses, respectively, prior to the study period. Both these instruments have sound psychometric properties and are widely used in Indian psychiatric studies. The scales to measure social support and life stresses were administered by one of us (RC), after obtaining informed consent. Following this, the patients were prospectively followed up at the lithium clinic and still continue to be in active follow-up. The follow-up data regarding the course of affective disorder, treatment compliance and serum lithium level monitoring were truncated in July 1995 for the purpose of this study, although patient care obviously continues. Thus, with regard to design, this study is a combination of retrospective chart-review and prospective follow-up of the lithium cohort. All of the patients were seen by at least one of the three consultant psychiatrists (PK, SKM and PS) during the period of enquiry on follow-up visits usually at 1–2 month intervals, their clinical status was monitored and details of relapses during the follow-up period and in between visits were recorded. The clinical examination followed an eclectic approach and was not structured in any standardised way. A relapse was operationally defined as emergence of full clinical picture of either depression or mania at or in between follow-up visits. If the patient presented with features of depression or mania at the time of follow-up visit, then the determination of relapse was done on the basis of clinical interview and mental status examination at that point in time. At the same time, the onset of the relapse was ascertained. However, if the patient and / or the key informant mentioned features of depression or mania in between follow-up visits, then the onset, nature and duration of relapse were assessed on the basis of detailed clinical interview. Similarly, the total duration of relapse was quantified on the basis of clinical interviews on subsequent visits. Throughout the study period serum lithium levels were regularly monitored and lithium dosage adjusted accordingly so as to achieve serum lithium levels between 0.4 and 1.2 mEq / l. Any reading below or above the stipulated levels were recorded in
89
the case notes. Other than lithium, antidepressants, neuroleptics and sedatives, hypnotic or other symptomatic treatments were also given whenever clinically warranted. Since some of these medicaments would affect the duration of a relapse episode, the latter was not considered as an outcome variable. Non-psychopharmacologic modes used were patient and family counselling including psycho-education and psychological support. No formal cognitive, behavioural or other psychotherapies were administered. Apart from descriptive statistics used for characterising the cohort and delineating outcome groups, inferential statistics used were the x 2 test, Students t-test, the Mann-Whitney U Test and Spearman’s Rank Order Correlation. Stepwise Multiple Regression Analyses were performed to ascertain relative contributions of various study variables in explaining response to lithium prophylaxis. A p value of , 0.05 was considered statistically significant.
3. Results The patients were predominantly male (88), (male:female ratio 2.5:1), their ages ranging from 17 to 80 years (mean 5 41.3, S.D. 5 13.1) although the majority of them (70%) were 30–60 years old. Almost 20% of them were young (below 30 years). Eighty percent of the patients were married. Most of the females (26 out of 34, 76.5%) were housewives, and most others were working at the time of intake (unemployed, 2.5%; retired 7.4%). Nearly 10% of them were engaged in professional or semi-professional work. Ten percent of the patients were illiterate, the rest had received some formal education, 18% had Master’s or professional-level degrees. Half of the patients were from nuclear families, and the majority were Hindu (71.3%) and urban (72.1%); 44% were local, i.e. they were residents of the Union Territory of Chandigarh, though 13% came even from a distance of more than 160 km. In the majority of the patients (60.6%), the affective disorder started with a manic episode. While 20 patients (16.4%) continued to have recurrent mania or hypomania, all others developed ‘circular’ (as per ICD-9, WHO, 1978) illness. Thus, in current terminology, all the patients had recurrent
90
P. Kulhara et al. / Journal of Affective Disorders 54 (1999) 87 – 96
Table 1 Relapse per year while on lithium (N 5 118) Variable
N
Percentage
Number of relapses ( per year) Nil Up to 0.2 0.4 0.6 0.8 1.0 1.2 1.4 . 1.4 (max. 2.2) At least 1 mania / hypomania At least 1 depression
28 17 27 12 12 6 7 3 6 75 a 66 b
24 14 23 10 10 5 6 2 6 64 56
a b
Of these, 34 (45%) had only hypomanic symptoms. Of these, 25 (38%) had only mild depressive symptoms.
bipolar affective disorder. Seasonal pattern, rapid cycling pattern, associated alcoholism, and atypical features (obsessional, schizoaffective) were noted in 4.9, 3.3, 3.3 and 6.6% of the patients, respectively. Appreciable psychosocial stressor was identified in 9% of the patients. The cohort had a high family loading of mental illnesses (44.3%). The duration of illness in the patients ranged from 1 to 25 years (pre-lithium) with a mean of 8.4 (S.D. 5 6.7) years. Most of the patients at intake were judged to be regular on follow-up (88%) and had an overall good drug compliance (90%) as per chart review (including lithium levels) and direct corroboration on interview. After institution of lithium prophylaxis, the follow-up period (inclusive of both the retrospective period up to intake and the prospective period up to July 1995) ranged from 2 to 27 (mean 5 11.1, S.D. 5 5.0) years. At the last follow-up in July 1995, data
could be gathered on 118 (97%) of the initial 122 patients. For the purposes of analysis of relapse episodes while on lithium, the particular episode during which lithium was started was counted as a pre-lithium episode and not as a relapse while on lithium. The main outcome data regarding relapses of affective episodes per year while on lithium are shown in Table 1. Of 118 patients, 28 (24%) did not have any relapse while on lithium prophylaxis. All others (90 cases, 76%) suffered varying frequencies of relapses ranging from 1 to an extreme of 34 relapses suffered by a patient. However, 62% of the patients had up to 1 relapse episode per year and the remaining 14% had more than 1 relapse per year. The median number of relapses was 0.3 per year. The mean of relapses was slightly higher (0.43) because of the effect of outliers. After adjusting for the latter, the mean of relapses came down to 0.36 relapses per year. Although 64% and 56% of the patients had manic and depressive relapses, respectively, many of these relapses while on lithium were clinically noted to be mild in nature (hypomania, 45%; mild depression, 38%). Lithium was combined with carbamazepine later in the course of illness in 13 patients, who were clinically considered to be non-responders to lithium alone and 12 of these maintained improvement thereafter. All these patients were on a lithium-carbamazepine combination till the last follow-up visit. Table 2 shows the data regarding affective episodes per year and hospitalisation pre- and onlithium prophylaxis. Compared to the pre-lithium course of affective disorder, the on-lithium course showed a significantly fewer mean number of episodes per year, both for manic and for depressive
Table 2 Affective episodes (per year) before and after starting lithium prophylaxis (N 5 118) Before lithium
Any affective episode Manic episodes Depressive episodes Hospitalisations a Duration of hospital stay in days
On lithium
Mann-Whitney U Test (Z values)
Mean
S.D.
Mean
S.D.
1.43 0.79 0.64 0.62 21.45
2.00 1.23 1.55 1.12
0.43 0.26 0.17 0.43 20.91
0.46 0.31 0.24 1.27
8.02* 6.36* 5.12* 0.76
* p , 0.01, df 5 116. a Twenty-six patients needed hospitalisation while on lithium compared to 49 patients before lithium ( x 2 5 5.17, df 5 1, p , 0.05).
P. Kulhara et al. / Journal of Affective Disorders 54 (1999) 87 – 96
episodes. Although no statistical significance emerged for the number of hospital admissions, significantly fewer patients needed hospitalisation after starting lithium prophylaxis than before (26 vs. 49, x 2 5 5.17, df 5 1, p , 0.05). Following these main outcome measures, our next task was to delineate the correlates and characteristics of good or poor response. This was done in two ways. Firstly, the patients were categorised as good responders or partial / poor responders to lithium, based on the median frequency of relapses (per year) suffered by patients while on lithium. Thus, 58 patients with up to 0.3 episodes of relapse per year of treatment (range 0–0.3), constituting 49% of the sample, were categorised as good responders and the rest 60 (51%) as partial / poor responders to lithium (number of relapses . 0.3 per year, max. 2.2). These two groups were then compared with respect to demographic, clinical, social support, life event and stress variables, by means of Student’s t-test, MannWhitney U test and x 2 wherever appropriate. The two groups of patients did not differ significantly with regard to socio-demographic variables. Similarly, the presence or absence of family history of mental illness was non-contributory. Data on selected clinical and psychosocial variables are shown in Table 3. The partial responders had a significantly higher number of depressive episodes prior to lithium therapy. Also, serum lithium levels dropped below 0.4 mEq / l significantly more often in partial /
91
poor responders than in good responders. With regard to psychosocial variables, both the number of life events and the total stress scores were significantly more in partial responders who also scored significantly less on social support than good responders. The two groups did not differ significantly on other measures. Secondly, various continuous measures at intake (demographic, clinical, social support, life event and stress scores) were studied in relation to the number of affective relapses per year suffered while on lithium by the entire sample without further categorisation. The results are presented in Table 4. The variables that had significant correlation with the number of relapses per year were the same as those Table 4 Correlates of frequency of relapses while on lithium, using Spearman’s Rank-Order Correlation (r) Variables
rs
Duration of illness (pre-lithium) No. of episodes (pre-lithium) No. of manic episodes (pre-lithium) No. of depressive episodes (pre-lithium) Total duration of illness (including lithium period) No. of times Li level , 0.4 mEq / l No. of life events Total stress score Social support score
0.048 0.057 0.011 0.200* 2 0.007 0.296** 0.222* 0.254** 2 0.174*
* p , 0.05; df 5 116. ** p , 0.01; df 5 116.
Table 3 Comparison of good responders (N 5 58) and partial / poor responders (N 5 60) to lithium on selected clinical and psychosocial variables Variable
Pre-lithium illness duration (years) No. of episodes (pre-lithium) No. of manic episodes (pre-lithium) No. of depressive episodes (pre-lithium) No. of hospital admissions (pre-lithium) Hospital days (pre-lithium) Total duration of lithium (years) No. of time Li , 0.4 mEq / l No. of times Li . 1.2 mEq / l No. of life events Total stress score Social support score * p , 0.05, df 5 116. b p , 0.01, df 5 116.
Good responders
Poor responders
Mean
S.D.
Mean
S.D.
7.99 1.59 0.90 0.69 0.74 25.19 10.89 1.91 0.52 5.09 212.76 56.81
5.98 2.59 1.64 0.59 1.04 43.49 4.54 2.34 1.37 3.02 131.07 11.26
9.24 1.27 0.68 2.11 0.57 20.95 11.37 3.20 0.68 6.53 280.30 51.95
6.52 1.17 0.61 0.68 0.79 30.66 5.40 2.52 1.25 3.12 134.26 11.87
Student t-test / Mann-Whitney U Test
1.10 0.26 0.29 2.48* 1.62 0.44 0.51 4.23** 0.69 2.56* 2.76* 2.28*
92
P. Kulhara et al. / Journal of Affective Disorders 54 (1999) 87 – 96
Table 5 Correlates of response to lithium prophylaxis, relative importance of variables a Variable
Multiple R
R2
Percentage variance explained
(1) No. of times lithium level , 0.4 mEq / l (2) Total stress score (3) No. of life events (4) Social support (5) No. of depressive episodes Total variance explained(%) 5 31.9
0.317 0.289 0.279 0.237 0.026
0.101 0.083 0.078 0.056 0.001
10.1 8.3 7.8 5.6 0.1
a
Stepwise multiple regression using affective episodes per year after commencing lithium as the dependent variable.
differing significantly between good and poor responders: number of depressive episodes prior to lithium prophylaxis, number of times lithium level dropped below 0.4 mEq / l, number of life events and total stress score (positive correlation), and social support score (negative correlation). Having identified significant correlates of lithium response in this bivariate analysis, these variables were subjected to stepwise multiple regression analysis in order to explore the strength and direction of the relationships of these variables. For these regression analyses, number of relapses per year while on lithium prophylaxis was considered to be the dependent variable and the other variables were treated as independent variables. The results of these analyses are depicted in Table 5. The total cumulative variance explained by these variable was estimated to be nearly 32% of which about one third of the contribution was by the variable ‘Number of times lithium level , 0.4 mEq / l’. Psychosocial variables ranked second in importance. Each of the three psychosocial variables—total stress score, number of life events and social support—had modest contribution, whereas ‘Number of depressive episodes’ before lithium therapy had an almost negligible contribution.
4. Discussion In this, probably the longest follow-up of study of lithium prophylaxis from India, 118 patients of recurrent bipolar affective disorder maintained on lithium prophylaxis were followed-up for a mean period of 11 years (range 2–27 years). Of the initial cohort of 122 patients, data of only 4 patients were not available. Compliance was generally good and attendance to lithium clinic was regular in most
patients. Serum lithium levels were monitored by flame photometry (Amdisen, 1967) on a regular periodic basis. Experienced consultant psychiatrists evaluated patients. All these factors combined together provide a unique opportunity to comment on the value of lithium prophylaxis in patients of recurrent bipolar affective disorder from India. Only 24% of the patients did not relapse while on lithium and thus responded best to lithium. Considering the median number of relapses per year, however, it was estimated that 49% of the patients were good responders to lithium. Venkoba Rao and Hariharasubramanian (1978) noted 41.7% of the cases to have good response, and in their later paper on 47 patients, 13 (27.6%) did not have a single relapse in 3–6 years (Venkoba Rao et al., 1982). In the series reported by Sethi et al. (1984), 32.7% did not report any relapse during the follow-up period of 6 months to 7 years. The somewhat lower proportion (24%) of non-relapsers in our cohort is partly explained by the much longer follow-up period in our study which might have allowed for more relapse episodes to surface. A recent study from Pondicherry, (Mathew et al., 1995) reported 45% of the 44 cases to be relapse-free on lithium. This study failed to mention the range and mean follow-up period although it was stated to be minimally 18–24 months. Aagaard and Vestergaard (1990) found that overall response to lithium prophylaxis was poor, with 40% of their patients experiencing two or more re-admissions during a 2-year follow-up. Other more recent prospective follow-up studies elsewhere have shown similar results (Peselow et al., 1994; Winokur et al., 1994). Indeed, recently there is a renewed interest and some controversy over the efficacy and clinical effectiveness of lithium prophylaxis (Guscott and
P. Kulhara et al. / Journal of Affective Disorders 54 (1999) 87 – 96
Taylor, 1994; Moncrieff, 1995; Goodwin, 1995). Results of the classic placebo-controlled prospective randomised clinical trials of lithium prophylaxis in recurrent affective disorder (Coppen et al., 1971; Prien et al., 1973) have been questioned on the basis of re-examination of the study designs (Moncrieff, 1995), followed by a rebuttal by Goodwin (1995). Earlier, Schou (1989) too addressed the issue of emerging scepticism about the magnitude of lithium prophylaxis, and maintained that about two-thirds of the patients of recurrent bipolar affective disorder would respond favourably to lithium. While this controversy continues, it is becoming progressively clear that the actual prophylactic effectiveness of lithium in normal clinical practice is considerably less than its efficacy, i.e., the potential for prophylaxis in strictly controlled and methodologically rigorous trials (Guscott and Taylor, 1994). Thus, there are a number of ‘naturalistic’ studies, both epidemiological and management trials, which raise serious concern about the benefits of lithium in ordinary clinical practice (Dickson and Kendell, 1986; Page et al., 1987; Maj et al., 1989; Markar and Mander, 1989; Aagaard and Vestergaard, 1990; O’Connell et al., 1991). Based on these and other considerations, Guscott and Taylor (1994) have proposed that ‘specialised lithium or mood disorders clinics have the potential to narrow the gap between efficacy and effectiveness–efficiency’. Our report from one such specialised lithium clinics becomes all the more relevant in this context. Two specific inferences may be drawn regarding lithium’s prophylactic effectiveness from the present study. First, lithium was definitely effective in reducing the mean number of episodes of affective disorder. This effect was observed for both manic and depressive episodes, and statistically significant results were in evidence for both types of episodes. Further, significantly fewer patients needed hospitalisation for treatment of affective disorder while on lithium compared to the pre-lithium period. Also, manic and depressive relapses, when they did occur, were of mild nature as judged clinically. All these data provide evidence for clinical effectiveness of lithium prophylaxis of bipolar affective disorder, and are in line with the consensus view put forward by Schou (1997). The second inference is that the magnitude of such
93
effect is, at best, modest. Contrary to earlier claims of an excellent effect of prophylactic lithium (Angst et al., 1970; Coppen et al., 1971; Prien et al., 1973; Narayanan et al., 1979), only 24% of our patients remained entirely relapse-free while 62% suffered up to one relapse per year while on lithium. In fact, 14% of the patients experienced more than one relapses per year during the follow-up period. There was no significant difference in number of hospital admissions or duration of hospitalisation, before or after lithium. Carbamazepine had to be added in 13 patients (and sodium valproate in one). Thus, in a group of generally compliant patients, lithium exerted a definite but modest effect in prophylaxis of recurrent affective disorder. These findings are largely in harmony with the recent ‘naturalistic’ studies cited above (e.g., Dickson and Kendell, 1986; Markar and Mander, 1989; Page et al., 1987; O’Connell et al., 1991; Peselow et al., 1994; Winokur et al., 1994). Long-term follow-up studies are especially useful for documentation of this, since it has been shown that patients initially categorised as ‘complete responders’ to lithium may relapse subsequently while on adequate lithium prophylaxis (Maj et al., 1989). This brings us to the discussion of the second aim of this study: correlates of lithium response. There is not much consensus about various factors claimed to be associated with a good or poor response (Khandelwal, 1991; Mathew et al., 1995; Aagaard and Vestergaard, 1990). In our study, both approaches adopted to address this issue identified the following common correlates of poor response to lithium: a higher number of depressive episodes prior to lithium prophylaxis, a higher number of life events and total psychosocial stress, a low degree of social support and relatively poor compliance to lithium as judged by the higher frequency of sub-prophylactic serum lithium levels. Stepwise multiple regression analyses revealed ‘Number of times lithium levels dropped below 0.4 mEq / l’ as the most important variable followed by other variables such as total stress score, number of life events and social support. This constellation of variables in the order as seen in our study in explaining lithium response is not surprising. In fact, it has been suggested that poor compliance with lithium therapy may be a major factor in accounting for the discrepancy between
94
P. Kulhara et al. / Journal of Affective Disorders 54 (1999) 87 – 96
lithium’s known prophylactic efficacy and its low actual effectiveness in clinical settings (Schou, 1993; Guscott and Taylor, 1994; Gershon and Soares, 1997). Our study in a way helps this notion by quantifying the magnitude of this ‘compliance’ effect, something that has not been done earlier (Gershon and Soares, 1997). What is impressive from the results, however, is the documentation of the role of psychosocial factors as correlates of response to lithium prophylaxis. Psychosocial stress and life events have long been identified as precipitating factors for bipolar affective disorder (Dunner et al., 1979). In this connection, it has even been proposed that lithium may exercise its suppressive prophylactic action by raising the patient’s threshold to stressful life situations (Prien, 1979), though this proposition is not well substantiated (Aronson and Shukla, 1987; Silverstone and Romans-Clarkson, 1989). Alternatively, since social support and psychosocial stress are often negatively correlated (Kulhara and Chopra, 1996), a combination of these factors may further enhance the probability of a relapse by temporarily decreasing the patient’s lithium compliance. An earlier study in this regard did find social support to be an important factor modifying lithium compliance and outcome (O’Connell et al., 1985). Such ‘intermittent noncompliance’ may actually spark off rebound episodes, increase episode frequency or cause lithium resistance (Guscott and Taylor, 1994), and lithium discontinuation may clearly precipitate rebound manic episodes (Suppes et al., 1991). Thus, a vicious cycle may ensue where psychosocial stress, other than directly precipitating affective episodes, may also lower social support, reduce lithium compliance, even if temporarily, and indirectly precipitate a rebound episode. Poor responders in our cohort often had this combination of events. A significant association between low serum lithium levels, poor psychosocial functioning and relapse of affective episodes in bipolar I disorders has also been noted in a recent study (Solomon et al., 1996). Hence, a study of lithium response should also focus on these psychosocial variables. Similar opinion is voiced in recent studies from the USA (O’Connell et al., 1991) and Belgium (Stefos et al., 1996). The other correlate identified in this study was the
number of depressive episodes prior to starting lithium. Given that the patients were all bipolar patients, this would mean that such patients with more depressive episodes are poor responders to lithium. This is a new finding and needs replication. In any case, it emerged as a weak variable as evinced by the multivariate analyses that were performed by us. There is, however, some suggestion that bipolar II patients (patients with predominantly recurrent depressive episodes with hypomania) may be less responsive to lithium prophylaxis (Joyce and Paykel, 1989). Other controversial predictors like family history, sex of patient and illness duration were not supported in this study. Some methodological caveats of the present work should be kept in mind while interpreting the results. Psychotropic medicines other than lithium, e.g. antidepressants, anxiolytics, neuroleptics and sedativehypnotics, were prescribed to the patients particularly at the time of relapse. Our study design did not have the latitude to extract these data pertaining to use of other psychotropic medication during the long years of follow-up; therefore, we have not specifically analysed the data keeping this other psychotropic medication variable in mind. Assessment of social support and life events was cross-sectional and done once only. Serial assessments were not performed. Thus, the association of these variables with the outcome variable should be viewed and interpreted as significant correlates of outcome. Whether these variables are also aetiologically meaningful predictor variables for relapse can not be commented on in this study design. It should also be appreciated that the study variables only explained a certain percentage of variance, leaving a large proportion of variance unexplained. This suggests that there are variables not identified and studied by us that may be influencing response to lithium and are linked to relapses while on lithium prophylaxis. More investigations are needed to explore these uncharted areas. In conclusion, this long-term lithium prophylaxis study shows that lithium exerts a definite, though modest, prophylactic effect on the course of recurrent bipolar affective disorder. Other than drug compliance itself, psychosocial variables, such as life stresses and social support, are important correlates of this effect.
P. Kulhara et al. / Journal of Affective Disorders 54 (1999) 87 – 96
References Aagaard, J., Vestergaard, P., 1990. Predictors of outcome in prophylactic lithium treatment: A 2-year prospective study. J. Affect. Disord. 18, 259–266. Amdisen, A., 1967. Serum lithium determination for clinical use. Scand. J. Clin. Lab. Invest. 20, 104–106. Angst, J., Weis, P., Grof, P., 1970. Lithium prophylaxis in recurrent affective disorders. Br. J. Psychiatry 116, 604–614. Aronson, T.A., Shukla, S., 1987. Life events and relapse in bipolar disorder: The impact of a catastrophic event. Acta Psychiat. Scand. 75, 571–576. Cade, J.F., 1949. Lithium salts in the treatment of psychotic excitement. Med. J. Aust. 36, 349–352. Coppen, A., Noguera, R., Bailey, J., 1971. Prophylactic lithium in affective disorders. Lancet ii, 275–279. Dickson, W.E., Kendell, R.E., 1986. Does maintenance lithium therapy prevent recurrences of mania under ordinary clinical conditions?. Psychol. Med. 16, 521–530. Dunner, D.L., Patrick, V., Fieve, R.R., 1979. Life events at the onset of bipolar affective illness. Am. J. Psychiatry 136, 508– 511. Gershon, S., Soares, J.C., 1997. Current therapeutic profile of lithium. Arch. Gen. Psychiatry 54, 16–20. Ghosh, A., Wig, N.N., 1977. Prophylactic value of lithium in manic-depressive psychosis-genetic aspects. Ind. J. Psychiatry 19, 73–76. Goodwin, G.M., 1995. Lithium revisited: A reply. Br. J. Psychiatry 167, 573–574. Guscott, R., Taylor, L., 1994. Lithium prophylaxis in recurrent affective illness: Efficacy, effectiveness and efficiency. Br. J. Psychiatry 164, 741–746. Joyce, P.R., Paykel, E.S., 1989. Predictors of drug response in depression. Arch. Gen. Psychiatry 46, 89–99. Khandelwal, S.K., 1991. Lithium in psychiatry: Therapeutic efficacy, side effects and practical issues. Ind. J. Psychiatry 33, 161–175. Khandelwal, S.K., Varma, V.K., Srinivasa Murthy, R., Khare, C.B., 1984. Lithium carbonate in the treatment of manic depressive psychosis in children. Ind. J. Psychiatry 26, 194–199. Kulhara, P., Chopra, R., 1996. Social support, social dysfunction and stressful life events in neurotic patients. Ind. J. Psychiatry 38, 23–29. Maj, M., Pirozzi, R., Kemali, D., 1989. Long-term outcome of lithium prophylaxis in patients initially classified as complete responders. Psychopharmacol. 98, 535–538. Markar, H.R., Mander, A.J., 1989. Efficacy of lithium prophylaxis in clinical practice. Br. J. Psychiatry 55, 496–500. Mathew, M.R.K., Chandrasekaran, R., Shreeram, S.S., Anand, I., 1995. Effect of lithium prophylaxis in bipolar affective disorder. Ind. J. Psychiatry 37, 5–8. Moncrieff, J., 1995. Lithium revisited: A re-examination of the placebo-controlled trials of lithium prophylaxis in manic-depressive disorder. Br. J. Psychiatry 67, 569–573. Narayanan, H.S., Malikarjunaiah, S., Rama Rao, B.S.S., 1979. A
95
clinical and follow-up study of 23 manic-depressive cases treated with lithium. Ind. J. Psychiatry 21, 77–79. Nehra, R., Kulhara, P., 1987. Development of a scale for the assessment of social support: Initial try-out in Indian setting. Ind. J. Soc. Psychiatry 4, 353–359. O’Connell, R.A., Mayo, J.A., 1988. The role of social factors in affective disorders: A review. Hosp. Comm. Psychiatry 39, 842–851. O’Connell, R.A., Mayo, J.A., Eng, L.K., Jones, J.S., Gabel, R.M., 1985. Social support and long-term lithium outcome. Br. J. Psychiatry 147, 272–275. O’Connell, R.A., Mayo, J.A., Flatow, L., Cuthbertson, B., O’Brien, B.E., 1991. Outcome of bipolar disorder on long-term treatment with lithium. Br. J. Psychiatry 159, 123–129. Page, C., Benaim, S., Lappin, F., 1987. A long-term retrospective follow-up of patients treated with prophylactic lithium carbonate. Br. J. Psychiatry 150, 175–179. Peselow, E.D., Fieve, R.R., Difglia, C., Sanfilipo, M.P., 1994. Lithium prophylaxis of bipolar illness. Br. J. Psychiatry 164, 208–214. Pollack, L., Harris, R., 1983. Measurement of social support. Psychol. Rep. 53, 446–449. Prakash, R., Sethi, N., Sethi, B.B., 1978. Lithium in affective disorders. Ind. J. Psychiatry 20, 380–383. Prien, R.F., 1979. Lithium in the prophylactic treatment of affective disorders. Arch. Gen. Psychiatry 36, 847–848. Prien, R.F., Caffey, E.M., Kleit, C.J., 1973. Prophylactic efficacy of lithium in manic-depressive illness. Arch. Gen. Psychiatry 28, 337–341. Schou, M., 1989. Lithium prophylaxis: Myths and realities. Am. J. Psychiatry 146, 573–576. Schou, M., 1993. Lithium prophylaxis: About ‘naturalistic’ or ‘clinical practice’ studies. Lithium 4, 77–81. Schou, M., 1997. Forty years of lithium treatment. Arch. Gen. Psychiatry 54, 9–13. Schou, M., Jael-Nielsen, N., Stromgren, E., 1954. The treatment of manic psychosis by the administration of lithium salts. J. Neurol. Neurosurg. Psychiatry 17, 250–260. Sethi, B.B., Dalal, P.K., Trivedi, J.K., 1984. Lithium in affective disorders: A seven-year observation of a lithium clinic. Ind. J. Psychiatry 26, 377–385. Silverstone, T., Romans-Clarkson, S., 1989. Bipolar affective disorder: Causes and prevention of relapse. Br. J. Psychiatry 154, 321–335. Singh, G., Kaur, D., Kaur, H., 1984. Presumptive Stressful Life Events Scale—A new stressful life events scale for use in India. Ind. J. Psychiatry 26, 107–114. Solomon, D.A., Ristow, W.R., Keller, M.B., Kane, J.M., Gelenberg, A.J., Rosenbaum, J.F., Warshaw, M.G., 1996. Serum lithium levels and psychosocial function in patients with bipolar I disorder. Am. J. Psychiatry 153, 1301–1307. Stefos, G., Bauwens, F., Staner, L., Pardeon, D., Mendlewicz, J., 1996. Psychosocial predictors of major affective recurrences in bipolar disorder: A 4-year longitudinal study of patients on prophylactic treatment. Acta Psychiat. Scand. 93, 420–426. Suppes, T., Baldessarini, R.J., Faedda, G.L., Tohen, M., 1991.
96
P. Kulhara et al. / Journal of Affective Disorders 54 (1999) 87 – 96
Risk of recurrence following discontinuation of lithium treatment in bipolar disorder. Arch. Gen. Psychiatry 48, 1082– 1088. Tandon, A.K., Asare, R., Saxena, V.C., 1981. Lithium treatment in affective disorders. Ind. J. Psychiatry 23, 58–62. Venkoba Rao, A., 1984. Depressive illness in India. Ind. J. Psychiatry 26, 301–311. Venkoba Rao, A., Hariharasubramanian, N., 1978. Lithium treatment in affective disorders: Certain observations. Ind. J. Psychiatry 20, 304–309. Venkoba Rao, A., Hariharasubramanian, N., Parvathi Devi, S.,
Sugumar, A., Srinivasan, V., 1982. Lithium prophylaxis in affective disorder. Ind. J. Psychiatry 24, 22–30. Winokur, G., Coryell, W., Akiskal, H.S., Endicott, J., Keller, M., Muller, T., 1994. Manic-depressive (bipolar) disorder: The course in light of a prospective ten-year follow-up of 131 patients. Acta Psychiat. Scand. 89, 102–110. World Health Organisation, 1978. Mental Disorders: Glossary and Guide to their Classification in Accordance with the 9th Revision of the International Classification of Diseases, WHO, Geneva.
Research report
Lithium prophylaxis of recurrent bipolar affective disorder: Long-term outcome and its psychosocial correlates a, a a a Parmanand Kulhara *, Debasish Basu , Surendra K. Mattoo , Pratap Sharan , b Rajni Chopra a
Department of Psychiatry, Postgraduate Institute of Medical Education and Research, Chandigarh 160 012, India b Institute of Health and Family Welfare, New Delhi, India Received 26 May 1998; received in revised form 27 June 1998; accepted 13 July 1998
Abstract Background: Discrepancy between efficacy of prophylactic lithium and its effectiveness in ordinary clinical practice necessitates long-term follow-up data from specialised lithium clinics. Also, role of psychosocial factors in influencing the outcome is unclear. Methods: One hundred and eighteen patients of bipolar affective disorder attending a lithium clinic were followed-up for | 11 years (range 2–27 years). Demographic and clinical data, measures of social support and psychosocial stress were obtained at the intake in 1989–1990. Study design combined retrospective chart-review (till the time of intake) with prospective follow-up till July 1995. Results: On lithium, the patients had a mean of 0.43 relapses per year (manic, 0.26; depressive, 0.17) which was significantly less ( p , 0.01) than the pre-lithium episode frequency. The figure for entirely relapse-free patients was 24%, and 62% had relapses up to one episode per year (median 5 0.3 per year). Fifty-eight (49%) patients were good responders to lithium (relapses # 0.30 per year). In comparison to good responders, partial / poor responders had a significantly greater number of pre-lithium depressive episodes, poor lithium compliance, more psychosocial stress and lower social support at intake. These variables correlated well with relapses and explained 32% of the variance of the data. Conclusions: Lithium had a definite prophylactic effect on long-term outcome. Social support and stressful life events are significant correlates of response to lithium. Clinical Implications: Lithium prophylaxis of bipolar affective disorders seems justified though psychosocial factors appear to modulate its effectiveness. Limitations: Other psychotropic medications were used during relapse and the assessment of psychosocial factors was cross-sectional. 1999 Elsevier Science B.V. All rights reserved. Keywords: Lithium; Bipolar affective disorders; Prophylaxis; Psychosocial factors
1. Introduction *Corresponding author. Tel.: 1 91-0172-541469-73, ext. 247; fax: 1 91-0172-540-401. E-mail address: [email protected] (P. Kulhara)
Following serendipitous discovery of the efficacy of lithium salts in controlling ‘psychotic excitement’
0165-0327 / 99 / $ – see front matter 1999 Elsevier Science B.V. All rights reserved. PII: S0165-0327( 98 )00145-1
88
P. Kulhara et al. / Journal of Affective Disorders 54 (1999) 87 – 96
(Cade, 1949) and more specifically manic psychosis (Schou et al., 1954), many more studies firmly established the prophylactic efficacy of lithium in recurrent affective disorders (Angst et al., 1970; Coppen et al., 1971; Prien et al., 1973). Studies from India also are in general agreement with this conclusion (Ghosh and Wig, 1977; Venkoba Rao and Hariharasubramanian, 1978; Tandon et al., 1981; Venkoba Rao et al., 1982; Prakash et al., 1978; Narayanan et al., 1979; Sethi et al., 1984; Khandelwal et al., 1984; Venkoba Rao, 1984; Mathew et al., 1995). Most of these Indian studies, however, were based on small samples, for example, 23 in Narayanan et al. (1979), 47 in Venkoba Rao et al. (1982), 71 in Venkoba Rao (1984), and 44 in the latest published study (Mathew et al., 1995), the largest number being 122 (Sethi et al., 1984). A small sample size obviously restricts the limits to which the findings can be generalised and extrapolated. The other major problem with these studies is relatively short duration of follow-up of the patients on lithium. Clearly, in any study on prophylactic efficacy of an agent, the longer the follow-up the better. Most of the studies cited above had follow-up periods ranging from 2 to 5 years. In the study by Sethi et al. (1984), out of the 122 patients studied, only 12 had follow-up for 5–7 years. Further, the mean or median duration of follow-up was not mentioned in most of the above studies. Our report aims to improve these lacunae. Another unresolved but important issue in this area is the role of psychosocial variables (stressor events, social support etc.) in modulating the response to lithium prophylaxis. Although the role of such social factors has been extensively studied in relation to recurrent depressive disorder and less in relation to bipolar disorder in general (O’Connell and Mayo, 1988; Silverstone and Romans-Clarkson, 1989), little research has been done so far in studying the role of these factors in modifying the effectiveness of lithium prophylaxis in bipolar affective disorders. Only a few very recent studies have looked at this issue systematically (Solomon et al., 1996; Stefos et al., 1996). More investigation as to how psychosocial variables influence outcome of lithium prophylaxis is called for (Schou, 1997), and this is the second major aim of our study. Incidentally, this study is from an institute credited
to have started one of the earliest lithium clinics in India (Khandelwal, 1991). The Department of Psychiatry of the Institute started using lithium as early as 1968 and the lithium clinic was formally started in 1975. Patients are followed-up regularly and serum lithium estimated by flame photometry method of Amdisen (1967). In view of the recently voiced doubt regarding efficacy of lithium in the prophylaxis of affective disorders when used in ordinary clinical practice (Dickson and Kendell, 1986; Markar and Mander, 1989; O’Connell et al., 1991; Guscott and Taylor, 1994) and even in classic double-blind studies of lithium (Moncrieff, 1995), it was felt appropriate at this point to make a critical and long-term assessment of the prophylactic benefits of lithium for patients of recurrent affective disorder. The protesting repercussion of an editorial (Goodwin, 1995) further sets the pace of this study. This paper describes the demographic and clinical characteristics, and outcome data of 122 patients from our lithium clinic over follow-up periods ranging from 2 to 27 years. The correlates of outcome patterns are analysed and discussed.
2. Material and methods Patients of recurrent affective disorder (manicdepressive psychosis according to International Classification of Diseases, 9th revision of the World Health Organisation, 1978, WHO) attending the Lithium Clinic, Department of Psychiatry, PGIMER, Chandigarh, between July 1989 and February 1990 comprised the sample. Most of these patients were registered in the clinic years before, including patients from 1968 onwards. All relevant data regarding the patients (demographic, clinical and therapeutic), charted in the case notes ever since registration through repeated (usually monthly) follow-up visits, were extracted. The accuracy of recording of demographic, diagnostic and clinical aspects were verified by two of us (PK and SKM) to the extent possible. It is worth pointing out that the routine case records of the patients at our lithium clinic are specifically designed for extracting data for research purposes. Other than these, the patients at intake (i.e. between July, 1989 and February, 1990) were also administered: (a) the Social Support Questionnaire
P. Kulhara et al. / Journal of Affective Disorders 54 (1999) 87 – 96
(Pollack and Harris, 1983, modified by Nehra and Kulhara, 1987), and (b) the Presumptive Stressful Life Events Scale (Singh et al., 1984), for measuring their perceived social support and life event-related stresses, respectively, prior to the study period. Both these instruments have sound psychometric properties and are widely used in Indian psychiatric studies. The scales to measure social support and life stresses were administered by one of us (RC), after obtaining informed consent. Following this, the patients were prospectively followed up at the lithium clinic and still continue to be in active follow-up. The follow-up data regarding the course of affective disorder, treatment compliance and serum lithium level monitoring were truncated in July 1995 for the purpose of this study, although patient care obviously continues. Thus, with regard to design, this study is a combination of retrospective chart-review and prospective follow-up of the lithium cohort. All of the patients were seen by at least one of the three consultant psychiatrists (PK, SKM and PS) during the period of enquiry on follow-up visits usually at 1–2 month intervals, their clinical status was monitored and details of relapses during the follow-up period and in between visits were recorded. The clinical examination followed an eclectic approach and was not structured in any standardised way. A relapse was operationally defined as emergence of full clinical picture of either depression or mania at or in between follow-up visits. If the patient presented with features of depression or mania at the time of follow-up visit, then the determination of relapse was done on the basis of clinical interview and mental status examination at that point in time. At the same time, the onset of the relapse was ascertained. However, if the patient and / or the key informant mentioned features of depression or mania in between follow-up visits, then the onset, nature and duration of relapse were assessed on the basis of detailed clinical interview. Similarly, the total duration of relapse was quantified on the basis of clinical interviews on subsequent visits. Throughout the study period serum lithium levels were regularly monitored and lithium dosage adjusted accordingly so as to achieve serum lithium levels between 0.4 and 1.2 mEq / l. Any reading below or above the stipulated levels were recorded in
89
the case notes. Other than lithium, antidepressants, neuroleptics and sedatives, hypnotic or other symptomatic treatments were also given whenever clinically warranted. Since some of these medicaments would affect the duration of a relapse episode, the latter was not considered as an outcome variable. Non-psychopharmacologic modes used were patient and family counselling including psycho-education and psychological support. No formal cognitive, behavioural or other psychotherapies were administered. Apart from descriptive statistics used for characterising the cohort and delineating outcome groups, inferential statistics used were the x 2 test, Students t-test, the Mann-Whitney U Test and Spearman’s Rank Order Correlation. Stepwise Multiple Regression Analyses were performed to ascertain relative contributions of various study variables in explaining response to lithium prophylaxis. A p value of , 0.05 was considered statistically significant.
3. Results The patients were predominantly male (88), (male:female ratio 2.5:1), their ages ranging from 17 to 80 years (mean 5 41.3, S.D. 5 13.1) although the majority of them (70%) were 30–60 years old. Almost 20% of them were young (below 30 years). Eighty percent of the patients were married. Most of the females (26 out of 34, 76.5%) were housewives, and most others were working at the time of intake (unemployed, 2.5%; retired 7.4%). Nearly 10% of them were engaged in professional or semi-professional work. Ten percent of the patients were illiterate, the rest had received some formal education, 18% had Master’s or professional-level degrees. Half of the patients were from nuclear families, and the majority were Hindu (71.3%) and urban (72.1%); 44% were local, i.e. they were residents of the Union Territory of Chandigarh, though 13% came even from a distance of more than 160 km. In the majority of the patients (60.6%), the affective disorder started with a manic episode. While 20 patients (16.4%) continued to have recurrent mania or hypomania, all others developed ‘circular’ (as per ICD-9, WHO, 1978) illness. Thus, in current terminology, all the patients had recurrent
90
P. Kulhara et al. / Journal of Affective Disorders 54 (1999) 87 – 96
Table 1 Relapse per year while on lithium (N 5 118) Variable
N
Percentage
Number of relapses ( per year) Nil Up to 0.2 0.4 0.6 0.8 1.0 1.2 1.4 . 1.4 (max. 2.2) At least 1 mania / hypomania At least 1 depression
28 17 27 12 12 6 7 3 6 75 a 66 b
24 14 23 10 10 5 6 2 6 64 56
a b
Of these, 34 (45%) had only hypomanic symptoms. Of these, 25 (38%) had only mild depressive symptoms.
bipolar affective disorder. Seasonal pattern, rapid cycling pattern, associated alcoholism, and atypical features (obsessional, schizoaffective) were noted in 4.9, 3.3, 3.3 and 6.6% of the patients, respectively. Appreciable psychosocial stressor was identified in 9% of the patients. The cohort had a high family loading of mental illnesses (44.3%). The duration of illness in the patients ranged from 1 to 25 years (pre-lithium) with a mean of 8.4 (S.D. 5 6.7) years. Most of the patients at intake were judged to be regular on follow-up (88%) and had an overall good drug compliance (90%) as per chart review (including lithium levels) and direct corroboration on interview. After institution of lithium prophylaxis, the follow-up period (inclusive of both the retrospective period up to intake and the prospective period up to July 1995) ranged from 2 to 27 (mean 5 11.1, S.D. 5 5.0) years. At the last follow-up in July 1995, data
could be gathered on 118 (97%) of the initial 122 patients. For the purposes of analysis of relapse episodes while on lithium, the particular episode during which lithium was started was counted as a pre-lithium episode and not as a relapse while on lithium. The main outcome data regarding relapses of affective episodes per year while on lithium are shown in Table 1. Of 118 patients, 28 (24%) did not have any relapse while on lithium prophylaxis. All others (90 cases, 76%) suffered varying frequencies of relapses ranging from 1 to an extreme of 34 relapses suffered by a patient. However, 62% of the patients had up to 1 relapse episode per year and the remaining 14% had more than 1 relapse per year. The median number of relapses was 0.3 per year. The mean of relapses was slightly higher (0.43) because of the effect of outliers. After adjusting for the latter, the mean of relapses came down to 0.36 relapses per year. Although 64% and 56% of the patients had manic and depressive relapses, respectively, many of these relapses while on lithium were clinically noted to be mild in nature (hypomania, 45%; mild depression, 38%). Lithium was combined with carbamazepine later in the course of illness in 13 patients, who were clinically considered to be non-responders to lithium alone and 12 of these maintained improvement thereafter. All these patients were on a lithium-carbamazepine combination till the last follow-up visit. Table 2 shows the data regarding affective episodes per year and hospitalisation pre- and onlithium prophylaxis. Compared to the pre-lithium course of affective disorder, the on-lithium course showed a significantly fewer mean number of episodes per year, both for manic and for depressive
Table 2 Affective episodes (per year) before and after starting lithium prophylaxis (N 5 118) Before lithium
Any affective episode Manic episodes Depressive episodes Hospitalisations a Duration of hospital stay in days
On lithium
Mann-Whitney U Test (Z values)
Mean
S.D.
Mean
S.D.
1.43 0.79 0.64 0.62 21.45
2.00 1.23 1.55 1.12
0.43 0.26 0.17 0.43 20.91
0.46 0.31 0.24 1.27
8.02* 6.36* 5.12* 0.76
* p , 0.01, df 5 116. a Twenty-six patients needed hospitalisation while on lithium compared to 49 patients before lithium ( x 2 5 5.17, df 5 1, p , 0.05).
P. Kulhara et al. / Journal of Affective Disorders 54 (1999) 87 – 96
episodes. Although no statistical significance emerged for the number of hospital admissions, significantly fewer patients needed hospitalisation after starting lithium prophylaxis than before (26 vs. 49, x 2 5 5.17, df 5 1, p , 0.05). Following these main outcome measures, our next task was to delineate the correlates and characteristics of good or poor response. This was done in two ways. Firstly, the patients were categorised as good responders or partial / poor responders to lithium, based on the median frequency of relapses (per year) suffered by patients while on lithium. Thus, 58 patients with up to 0.3 episodes of relapse per year of treatment (range 0–0.3), constituting 49% of the sample, were categorised as good responders and the rest 60 (51%) as partial / poor responders to lithium (number of relapses . 0.3 per year, max. 2.2). These two groups were then compared with respect to demographic, clinical, social support, life event and stress variables, by means of Student’s t-test, MannWhitney U test and x 2 wherever appropriate. The two groups of patients did not differ significantly with regard to socio-demographic variables. Similarly, the presence or absence of family history of mental illness was non-contributory. Data on selected clinical and psychosocial variables are shown in Table 3. The partial responders had a significantly higher number of depressive episodes prior to lithium therapy. Also, serum lithium levels dropped below 0.4 mEq / l significantly more often in partial /
91
poor responders than in good responders. With regard to psychosocial variables, both the number of life events and the total stress scores were significantly more in partial responders who also scored significantly less on social support than good responders. The two groups did not differ significantly on other measures. Secondly, various continuous measures at intake (demographic, clinical, social support, life event and stress scores) were studied in relation to the number of affective relapses per year suffered while on lithium by the entire sample without further categorisation. The results are presented in Table 4. The variables that had significant correlation with the number of relapses per year were the same as those Table 4 Correlates of frequency of relapses while on lithium, using Spearman’s Rank-Order Correlation (r) Variables
rs
Duration of illness (pre-lithium) No. of episodes (pre-lithium) No. of manic episodes (pre-lithium) No. of depressive episodes (pre-lithium) Total duration of illness (including lithium period) No. of times Li level , 0.4 mEq / l No. of life events Total stress score Social support score
0.048 0.057 0.011 0.200* 2 0.007 0.296** 0.222* 0.254** 2 0.174*
* p , 0.05; df 5 116. ** p , 0.01; df 5 116.
Table 3 Comparison of good responders (N 5 58) and partial / poor responders (N 5 60) to lithium on selected clinical and psychosocial variables Variable
Pre-lithium illness duration (years) No. of episodes (pre-lithium) No. of manic episodes (pre-lithium) No. of depressive episodes (pre-lithium) No. of hospital admissions (pre-lithium) Hospital days (pre-lithium) Total duration of lithium (years) No. of time Li , 0.4 mEq / l No. of times Li . 1.2 mEq / l No. of life events Total stress score Social support score * p , 0.05, df 5 116. b p , 0.01, df 5 116.
Good responders
Poor responders
Mean
S.D.
Mean
S.D.
7.99 1.59 0.90 0.69 0.74 25.19 10.89 1.91 0.52 5.09 212.76 56.81
5.98 2.59 1.64 0.59 1.04 43.49 4.54 2.34 1.37 3.02 131.07 11.26
9.24 1.27 0.68 2.11 0.57 20.95 11.37 3.20 0.68 6.53 280.30 51.95
6.52 1.17 0.61 0.68 0.79 30.66 5.40 2.52 1.25 3.12 134.26 11.87
Student t-test / Mann-Whitney U Test
1.10 0.26 0.29 2.48* 1.62 0.44 0.51 4.23** 0.69 2.56* 2.76* 2.28*
92
P. Kulhara et al. / Journal of Affective Disorders 54 (1999) 87 – 96
Table 5 Correlates of response to lithium prophylaxis, relative importance of variables a Variable
Multiple R
R2
Percentage variance explained
(1) No. of times lithium level , 0.4 mEq / l (2) Total stress score (3) No. of life events (4) Social support (5) No. of depressive episodes Total variance explained(%) 5 31.9
0.317 0.289 0.279 0.237 0.026
0.101 0.083 0.078 0.056 0.001
10.1 8.3 7.8 5.6 0.1
a
Stepwise multiple regression using affective episodes per year after commencing lithium as the dependent variable.
differing significantly between good and poor responders: number of depressive episodes prior to lithium prophylaxis, number of times lithium level dropped below 0.4 mEq / l, number of life events and total stress score (positive correlation), and social support score (negative correlation). Having identified significant correlates of lithium response in this bivariate analysis, these variables were subjected to stepwise multiple regression analysis in order to explore the strength and direction of the relationships of these variables. For these regression analyses, number of relapses per year while on lithium prophylaxis was considered to be the dependent variable and the other variables were treated as independent variables. The results of these analyses are depicted in Table 5. The total cumulative variance explained by these variable was estimated to be nearly 32% of which about one third of the contribution was by the variable ‘Number of times lithium level , 0.4 mEq / l’. Psychosocial variables ranked second in importance. Each of the three psychosocial variables—total stress score, number of life events and social support—had modest contribution, whereas ‘Number of depressive episodes’ before lithium therapy had an almost negligible contribution.
4. Discussion In this, probably the longest follow-up of study of lithium prophylaxis from India, 118 patients of recurrent bipolar affective disorder maintained on lithium prophylaxis were followed-up for a mean period of 11 years (range 2–27 years). Of the initial cohort of 122 patients, data of only 4 patients were not available. Compliance was generally good and attendance to lithium clinic was regular in most
patients. Serum lithium levels were monitored by flame photometry (Amdisen, 1967) on a regular periodic basis. Experienced consultant psychiatrists evaluated patients. All these factors combined together provide a unique opportunity to comment on the value of lithium prophylaxis in patients of recurrent bipolar affective disorder from India. Only 24% of the patients did not relapse while on lithium and thus responded best to lithium. Considering the median number of relapses per year, however, it was estimated that 49% of the patients were good responders to lithium. Venkoba Rao and Hariharasubramanian (1978) noted 41.7% of the cases to have good response, and in their later paper on 47 patients, 13 (27.6%) did not have a single relapse in 3–6 years (Venkoba Rao et al., 1982). In the series reported by Sethi et al. (1984), 32.7% did not report any relapse during the follow-up period of 6 months to 7 years. The somewhat lower proportion (24%) of non-relapsers in our cohort is partly explained by the much longer follow-up period in our study which might have allowed for more relapse episodes to surface. A recent study from Pondicherry, (Mathew et al., 1995) reported 45% of the 44 cases to be relapse-free on lithium. This study failed to mention the range and mean follow-up period although it was stated to be minimally 18–24 months. Aagaard and Vestergaard (1990) found that overall response to lithium prophylaxis was poor, with 40% of their patients experiencing two or more re-admissions during a 2-year follow-up. Other more recent prospective follow-up studies elsewhere have shown similar results (Peselow et al., 1994; Winokur et al., 1994). Indeed, recently there is a renewed interest and some controversy over the efficacy and clinical effectiveness of lithium prophylaxis (Guscott and
P. Kulhara et al. / Journal of Affective Disorders 54 (1999) 87 – 96
Taylor, 1994; Moncrieff, 1995; Goodwin, 1995). Results of the classic placebo-controlled prospective randomised clinical trials of lithium prophylaxis in recurrent affective disorder (Coppen et al., 1971; Prien et al., 1973) have been questioned on the basis of re-examination of the study designs (Moncrieff, 1995), followed by a rebuttal by Goodwin (1995). Earlier, Schou (1989) too addressed the issue of emerging scepticism about the magnitude of lithium prophylaxis, and maintained that about two-thirds of the patients of recurrent bipolar affective disorder would respond favourably to lithium. While this controversy continues, it is becoming progressively clear that the actual prophylactic effectiveness of lithium in normal clinical practice is considerably less than its efficacy, i.e., the potential for prophylaxis in strictly controlled and methodologically rigorous trials (Guscott and Taylor, 1994). Thus, there are a number of ‘naturalistic’ studies, both epidemiological and management trials, which raise serious concern about the benefits of lithium in ordinary clinical practice (Dickson and Kendell, 1986; Page et al., 1987; Maj et al., 1989; Markar and Mander, 1989; Aagaard and Vestergaard, 1990; O’Connell et al., 1991). Based on these and other considerations, Guscott and Taylor (1994) have proposed that ‘specialised lithium or mood disorders clinics have the potential to narrow the gap between efficacy and effectiveness–efficiency’. Our report from one such specialised lithium clinics becomes all the more relevant in this context. Two specific inferences may be drawn regarding lithium’s prophylactic effectiveness from the present study. First, lithium was definitely effective in reducing the mean number of episodes of affective disorder. This effect was observed for both manic and depressive episodes, and statistically significant results were in evidence for both types of episodes. Further, significantly fewer patients needed hospitalisation for treatment of affective disorder while on lithium compared to the pre-lithium period. Also, manic and depressive relapses, when they did occur, were of mild nature as judged clinically. All these data provide evidence for clinical effectiveness of lithium prophylaxis of bipolar affective disorder, and are in line with the consensus view put forward by Schou (1997). The second inference is that the magnitude of such
93
effect is, at best, modest. Contrary to earlier claims of an excellent effect of prophylactic lithium (Angst et al., 1970; Coppen et al., 1971; Prien et al., 1973; Narayanan et al., 1979), only 24% of our patients remained entirely relapse-free while 62% suffered up to one relapse per year while on lithium. In fact, 14% of the patients experienced more than one relapses per year during the follow-up period. There was no significant difference in number of hospital admissions or duration of hospitalisation, before or after lithium. Carbamazepine had to be added in 13 patients (and sodium valproate in one). Thus, in a group of generally compliant patients, lithium exerted a definite but modest effect in prophylaxis of recurrent affective disorder. These findings are largely in harmony with the recent ‘naturalistic’ studies cited above (e.g., Dickson and Kendell, 1986; Markar and Mander, 1989; Page et al., 1987; O’Connell et al., 1991; Peselow et al., 1994; Winokur et al., 1994). Long-term follow-up studies are especially useful for documentation of this, since it has been shown that patients initially categorised as ‘complete responders’ to lithium may relapse subsequently while on adequate lithium prophylaxis (Maj et al., 1989). This brings us to the discussion of the second aim of this study: correlates of lithium response. There is not much consensus about various factors claimed to be associated with a good or poor response (Khandelwal, 1991; Mathew et al., 1995; Aagaard and Vestergaard, 1990). In our study, both approaches adopted to address this issue identified the following common correlates of poor response to lithium: a higher number of depressive episodes prior to lithium prophylaxis, a higher number of life events and total psychosocial stress, a low degree of social support and relatively poor compliance to lithium as judged by the higher frequency of sub-prophylactic serum lithium levels. Stepwise multiple regression analyses revealed ‘Number of times lithium levels dropped below 0.4 mEq / l’ as the most important variable followed by other variables such as total stress score, number of life events and social support. This constellation of variables in the order as seen in our study in explaining lithium response is not surprising. In fact, it has been suggested that poor compliance with lithium therapy may be a major factor in accounting for the discrepancy between
94
P. Kulhara et al. / Journal of Affective Disorders 54 (1999) 87 – 96
lithium’s known prophylactic efficacy and its low actual effectiveness in clinical settings (Schou, 1993; Guscott and Taylor, 1994; Gershon and Soares, 1997). Our study in a way helps this notion by quantifying the magnitude of this ‘compliance’ effect, something that has not been done earlier (Gershon and Soares, 1997). What is impressive from the results, however, is the documentation of the role of psychosocial factors as correlates of response to lithium prophylaxis. Psychosocial stress and life events have long been identified as precipitating factors for bipolar affective disorder (Dunner et al., 1979). In this connection, it has even been proposed that lithium may exercise its suppressive prophylactic action by raising the patient’s threshold to stressful life situations (Prien, 1979), though this proposition is not well substantiated (Aronson and Shukla, 1987; Silverstone and Romans-Clarkson, 1989). Alternatively, since social support and psychosocial stress are often negatively correlated (Kulhara and Chopra, 1996), a combination of these factors may further enhance the probability of a relapse by temporarily decreasing the patient’s lithium compliance. An earlier study in this regard did find social support to be an important factor modifying lithium compliance and outcome (O’Connell et al., 1985). Such ‘intermittent noncompliance’ may actually spark off rebound episodes, increase episode frequency or cause lithium resistance (Guscott and Taylor, 1994), and lithium discontinuation may clearly precipitate rebound manic episodes (Suppes et al., 1991). Thus, a vicious cycle may ensue where psychosocial stress, other than directly precipitating affective episodes, may also lower social support, reduce lithium compliance, even if temporarily, and indirectly precipitate a rebound episode. Poor responders in our cohort often had this combination of events. A significant association between low serum lithium levels, poor psychosocial functioning and relapse of affective episodes in bipolar I disorders has also been noted in a recent study (Solomon et al., 1996). Hence, a study of lithium response should also focus on these psychosocial variables. Similar opinion is voiced in recent studies from the USA (O’Connell et al., 1991) and Belgium (Stefos et al., 1996). The other correlate identified in this study was the
number of depressive episodes prior to starting lithium. Given that the patients were all bipolar patients, this would mean that such patients with more depressive episodes are poor responders to lithium. This is a new finding and needs replication. In any case, it emerged as a weak variable as evinced by the multivariate analyses that were performed by us. There is, however, some suggestion that bipolar II patients (patients with predominantly recurrent depressive episodes with hypomania) may be less responsive to lithium prophylaxis (Joyce and Paykel, 1989). Other controversial predictors like family history, sex of patient and illness duration were not supported in this study. Some methodological caveats of the present work should be kept in mind while interpreting the results. Psychotropic medicines other than lithium, e.g. antidepressants, anxiolytics, neuroleptics and sedativehypnotics, were prescribed to the patients particularly at the time of relapse. Our study design did not have the latitude to extract these data pertaining to use of other psychotropic medication during the long years of follow-up; therefore, we have not specifically analysed the data keeping this other psychotropic medication variable in mind. Assessment of social support and life events was cross-sectional and done once only. Serial assessments were not performed. Thus, the association of these variables with the outcome variable should be viewed and interpreted as significant correlates of outcome. Whether these variables are also aetiologically meaningful predictor variables for relapse can not be commented on in this study design. It should also be appreciated that the study variables only explained a certain percentage of variance, leaving a large proportion of variance unexplained. This suggests that there are variables not identified and studied by us that may be influencing response to lithium and are linked to relapses while on lithium prophylaxis. More investigations are needed to explore these uncharted areas. In conclusion, this long-term lithium prophylaxis study shows that lithium exerts a definite, though modest, prophylactic effect on the course of recurrent bipolar affective disorder. Other than drug compliance itself, psychosocial variables, such as life stresses and social support, are important correlates of this effect.
P. Kulhara et al. / Journal of Affective Disorders 54 (1999) 87 – 96
References Aagaard, J., Vestergaard, P., 1990. Predictors of outcome in prophylactic lithium treatment: A 2-year prospective study. J. Affect. Disord. 18, 259–266. Amdisen, A., 1967. Serum lithium determination for clinical use. Scand. J. Clin. Lab. Invest. 20, 104–106. Angst, J., Weis, P., Grof, P., 1970. Lithium prophylaxis in recurrent affective disorders. Br. J. Psychiatry 116, 604–614. Aronson, T.A., Shukla, S., 1987. Life events and relapse in bipolar disorder: The impact of a catastrophic event. Acta Psychiat. Scand. 75, 571–576. Cade, J.F., 1949. Lithium salts in the treatment of psychotic excitement. Med. J. Aust. 36, 349–352. Coppen, A., Noguera, R., Bailey, J., 1971. Prophylactic lithium in affective disorders. Lancet ii, 275–279. Dickson, W.E., Kendell, R.E., 1986. Does maintenance lithium therapy prevent recurrences of mania under ordinary clinical conditions?. Psychol. Med. 16, 521–530. Dunner, D.L., Patrick, V., Fieve, R.R., 1979. Life events at the onset of bipolar affective illness. Am. J. Psychiatry 136, 508– 511. Gershon, S., Soares, J.C., 1997. Current therapeutic profile of lithium. Arch. Gen. Psychiatry 54, 16–20. Ghosh, A., Wig, N.N., 1977. Prophylactic value of lithium in manic-depressive psychosis-genetic aspects. Ind. J. Psychiatry 19, 73–76. Goodwin, G.M., 1995. Lithium revisited: A reply. Br. J. Psychiatry 167, 573–574. Guscott, R., Taylor, L., 1994. Lithium prophylaxis in recurrent affective illness: Efficacy, effectiveness and efficiency. Br. J. Psychiatry 164, 741–746. Joyce, P.R., Paykel, E.S., 1989. Predictors of drug response in depression. Arch. Gen. Psychiatry 46, 89–99. Khandelwal, S.K., 1991. Lithium in psychiatry: Therapeutic efficacy, side effects and practical issues. Ind. J. Psychiatry 33, 161–175. Khandelwal, S.K., Varma, V.K., Srinivasa Murthy, R., Khare, C.B., 1984. Lithium carbonate in the treatment of manic depressive psychosis in children. Ind. J. Psychiatry 26, 194–199. Kulhara, P., Chopra, R., 1996. Social support, social dysfunction and stressful life events in neurotic patients. Ind. J. Psychiatry 38, 23–29. Maj, M., Pirozzi, R., Kemali, D., 1989. Long-term outcome of lithium prophylaxis in patients initially classified as complete responders. Psychopharmacol. 98, 535–538. Markar, H.R., Mander, A.J., 1989. Efficacy of lithium prophylaxis in clinical practice. Br. J. Psychiatry 55, 496–500. Mathew, M.R.K., Chandrasekaran, R., Shreeram, S.S., Anand, I., 1995. Effect of lithium prophylaxis in bipolar affective disorder. Ind. J. Psychiatry 37, 5–8. Moncrieff, J., 1995. Lithium revisited: A re-examination of the placebo-controlled trials of lithium prophylaxis in manic-depressive disorder. Br. J. Psychiatry 67, 569–573. Narayanan, H.S., Malikarjunaiah, S., Rama Rao, B.S.S., 1979. A
95
clinical and follow-up study of 23 manic-depressive cases treated with lithium. Ind. J. Psychiatry 21, 77–79. Nehra, R., Kulhara, P., 1987. Development of a scale for the assessment of social support: Initial try-out in Indian setting. Ind. J. Soc. Psychiatry 4, 353–359. O’Connell, R.A., Mayo, J.A., 1988. The role of social factors in affective disorders: A review. Hosp. Comm. Psychiatry 39, 842–851. O’Connell, R.A., Mayo, J.A., Eng, L.K., Jones, J.S., Gabel, R.M., 1985. Social support and long-term lithium outcome. Br. J. Psychiatry 147, 272–275. O’Connell, R.A., Mayo, J.A., Flatow, L., Cuthbertson, B., O’Brien, B.E., 1991. Outcome of bipolar disorder on long-term treatment with lithium. Br. J. Psychiatry 159, 123–129. Page, C., Benaim, S., Lappin, F., 1987. A long-term retrospective follow-up of patients treated with prophylactic lithium carbonate. Br. J. Psychiatry 150, 175–179. Peselow, E.D., Fieve, R.R., Difglia, C., Sanfilipo, M.P., 1994. Lithium prophylaxis of bipolar illness. Br. J. Psychiatry 164, 208–214. Pollack, L., Harris, R., 1983. Measurement of social support. Psychol. Rep. 53, 446–449. Prakash, R., Sethi, N., Sethi, B.B., 1978. Lithium in affective disorders. Ind. J. Psychiatry 20, 380–383. Prien, R.F., 1979. Lithium in the prophylactic treatment of affective disorders. Arch. Gen. Psychiatry 36, 847–848. Prien, R.F., Caffey, E.M., Kleit, C.J., 1973. Prophylactic efficacy of lithium in manic-depressive illness. Arch. Gen. Psychiatry 28, 337–341. Schou, M., 1989. Lithium prophylaxis: Myths and realities. Am. J. Psychiatry 146, 573–576. Schou, M., 1993. Lithium prophylaxis: About ‘naturalistic’ or ‘clinical practice’ studies. Lithium 4, 77–81. Schou, M., 1997. Forty years of lithium treatment. Arch. Gen. Psychiatry 54, 9–13. Schou, M., Jael-Nielsen, N., Stromgren, E., 1954. The treatment of manic psychosis by the administration of lithium salts. J. Neurol. Neurosurg. Psychiatry 17, 250–260. Sethi, B.B., Dalal, P.K., Trivedi, J.K., 1984. Lithium in affective disorders: A seven-year observation of a lithium clinic. Ind. J. Psychiatry 26, 377–385. Silverstone, T., Romans-Clarkson, S., 1989. Bipolar affective disorder: Causes and prevention of relapse. Br. J. Psychiatry 154, 321–335. Singh, G., Kaur, D., Kaur, H., 1984. Presumptive Stressful Life Events Scale—A new stressful life events scale for use in India. Ind. J. Psychiatry 26, 107–114. Solomon, D.A., Ristow, W.R., Keller, M.B., Kane, J.M., Gelenberg, A.J., Rosenbaum, J.F., Warshaw, M.G., 1996. Serum lithium levels and psychosocial function in patients with bipolar I disorder. Am. J. Psychiatry 153, 1301–1307. Stefos, G., Bauwens, F., Staner, L., Pardeon, D., Mendlewicz, J., 1996. Psychosocial predictors of major affective recurrences in bipolar disorder: A 4-year longitudinal study of patients on prophylactic treatment. Acta Psychiat. Scand. 93, 420–426. Suppes, T., Baldessarini, R.J., Faedda, G.L., Tohen, M., 1991.
96
P. Kulhara et al. / Journal of Affective Disorders 54 (1999) 87 – 96
Risk of recurrence following discontinuation of lithium treatment in bipolar disorder. Arch. Gen. Psychiatry 48, 1082– 1088. Tandon, A.K., Asare, R., Saxena, V.C., 1981. Lithium treatment in affective disorders. Ind. J. Psychiatry 23, 58–62. Venkoba Rao, A., 1984. Depressive illness in India. Ind. J. Psychiatry 26, 301–311. Venkoba Rao, A., Hariharasubramanian, N., 1978. Lithium treatment in affective disorders: Certain observations. Ind. J. Psychiatry 20, 304–309. Venkoba Rao, A., Hariharasubramanian, N., Parvathi Devi, S.,
Sugumar, A., Srinivasan, V., 1982. Lithium prophylaxis in affective disorder. Ind. J. Psychiatry 24, 22–30. Winokur, G., Coryell, W., Akiskal, H.S., Endicott, J., Keller, M., Muller, T., 1994. Manic-depressive (bipolar) disorder: The course in light of a prospective ten-year follow-up of 131 patients. Acta Psychiat. Scand. 89, 102–110. World Health Organisation, 1978. Mental Disorders: Glossary and Guide to their Classification in Accordance with the 9th Revision of the International Classification of Diseases, WHO, Geneva.