- Email: [email protected]
Identifying predictors for good lithium response – A retrospective analysis of 100 patients with bipolar disorder using a life-charting method
Available online at
www.sciencedirect.com European Psychiatry 24 (2009) 171e177
Original article
Identifying predictors for good lithium response e A retrospective analysis of 100 patients with bipolar disorder using a life-charting method ˚ gren c Lena Backlund a,*, Anna Ehnvall b, Jerker Hetta a, Go¨ran Isacsson a, Hans A a
Karolinska Institute, Department of Clinical Neuroscience, Section of Psychiatry, Karolinska University Hospital, Huddinge M57, SE-141 86 Stockholm, Sweden b Institute for Clinical Neuroscience, Gothenburg and Psychiatric Out-patient Clinic, Varberg, Swedan c Sahlgrenska Academy at the University of Gothenburg, Institute of Neuroscience and Physiology, Department of Psychiatry and Neurochemistry, Gothenburg, Sweden Received 18 August 2008; received in revised form 17 December 2008; accepted 30 December 2008 Available online 14 March 2009
Abstract Purpose. e Our aim was to investigate bipolar patients in order to test the validity of various outcome measures and to identify prognostic predictors for pharmacological treatment. Material and method. e One hundred patients were interviewed using a computerized life-charting program in a descriptive, retrospective analysis. The concept ‘‘Burden of illness’’ was defined as a combination of severity and duration of episodes. Response to treatment was defined as the difference in burden before and after treatment, a low burden during treatment, and freedom of episodes for at least 3 years after insertion of treatment. Results. e The absence of mixed episodes and a high initial burden predicted a good response measured as the difference in burden. If remission for 3 years or a low burden during lithium treatment was used, the absence of rapid cycling and of mixed episodes were the most important predictors. The severity of illness before treatment had no impact. Discussion and conclusion. e We suggest the use of absolute measures of severity during treatment as the most appropriate measure of the outcome. Furthermore, our data provide corroboration that treatment with lithium ameliorates the prognosis of the illness, but that mixed episodes and rapid cycling predict a poorer response to lithium. Ó 2009 Elsevier Masson SAS. All rights reserved. Keywords: Bipolar disorder; Treatment outcome; Lithium
1. Introduction Bipolar affective disorder is a common psychiatric disease. The lifetime prevalence of classic bipolar I disorder is estimated to be 0.9e1.6% [20,27]. The median age at onset of illness is 21 years and more than 80% of the patients with a first episode will recur [36]. In most cases, the course of the disease is progressive [2,24]. Suicide rates are more than 20fold higher than in the general population [26,33,34]. Several investigations have shown that there is on average a delay of 5 years between a patient’s first illness episode and * Corresponding author. Tel./fax: þ46 8 585 857 08. E-mail address: [email protected] (L. Backlund). 0924-9338/$ - see front matter Ó 2009 Elsevier Masson SAS. All rights reserved. doi:10.1016/j.eurpsy.2008.12.009
the first consultation of a physician, and a further 8 years or so (range 7.8e9.3) until the correct diagnosis is established [3,5,32]. There is also a delay of several years until adequate mood-stabilizing medication is prescribed [8]. This delay may be ominous, since the early diagnosis and instigation of pharmacological treatment may be crucial for a good prognosis [16,32]. The aetiology of bipolar illness is multifactorial. Genetic factors [11,25,28] and pharmacologic maintenance treatment [5,6,23] seem to be crucial factors for the prognosis. Abilities to cope with stress, sustain close friendship and good relationships with family as well as a high adherence to the treatment [7,35] predict a good prognosis, while mixed episodes [1], a poorer social function, comorbidity with
172
L. Backlund et al. / European Psychiatry 24 (2009) 171e177
borderline personality disorders [12,13] or substance-use disorder [31] may have a more negative effect on the outcome. The need for continuous mood-stabilizing medication is wellknown, but the importance of its early instigation is still a matter of discussion. Franchini and Tondo found a relationship between the latency before treatment and the reduction of severity after prophylactic treatment [16,32]. Baethge showed a positive correlation between the grade of severity before treatment and the reduction of severity [3]. Later, Baldessarini investigated this systematically, and concluded that ‘‘prior episode counts and treatment delay have little association with morbidity during prophylaxis with mood-stabilizing agents. Comparisons of morbidity during vs. before treatment in episodic disorders are misleading’’ [4,5]. He also pointed out the need for investigations where functional disability is a part of the outcome. In weekly assessments during 13.4 years, Judd et al. found a high frequency of sub-syndromal affective symptoms between the major affective episodes and discussed the functional disability between episodes as an important outcome [18]. 1.1. Aims of the study This study focuses on the temporal patterns in bipolar disorders. The aims were to investigate retrospectively a clinical sample as a foundation for future prospective analyses, and to identify possible prognostic predictors with an emphasis on pharmacological treatment. We also wished to test innovative methods for measuring the burden of illness as a combination of the duration and the severity of episodes [15], and to test the validity of different outcome measures. 2. Subjects and methods
(41 males and 59 females) out of 150 invited had agreed and given informed consent, were included in the study. No patient was excluded by the investigators. The participants in the study group did not differ from those who refused to participate with respect to gender, diagnosis, or age. The mean age at the time of the first interview was 46.8 years (Md ¼ 48.5) and they had on average been ill for 23.8 years (Md ¼ 23.7). Eighty-eight had been diagnosed with bipolar I and the remaining 12 with bipolar II disorder. Fifty-six were married, 2 were widowed, 27 were divorced, and 16 had never had any long relationship. All except one had passed elementary school. Forty-five had completed high school, and 33 had graduated from university. Ninety-two had been hospitalized at least once, and 58 of these had been subjected to compulsory admission. 2.2. Procedures All the patients were interviewed between January 2001 and February 2004 by a psychiatrist (LB), a psychiatric nurse, or a psychologist, using a semi-structured interview according to a life-charting protocol [9,10,22] previously developed for ˚ gren [14]. For diagnostic the purpose by Ehnvall and A confirmation of bipolar comorbid disorders, the Mini International Neuropsychiatric Interview (M.I.N.I) was used [30]. The participants were carefully interviewed about their family history, life events, social history, previous depressions, manias, mixed episodes occurring since early childhood, and earlier and current treatments. The accurate determination of dates of each event was a primary concern. The DSM-IV criteria were used to fix the diagnoses of manic and depressive episodes. All available medical records were searched to validate the verbal reports. When the 2 sources of data were inconsistent, the patient was asked more detailed questions to provide clarification.
2.1. Patients 2.3. Potential predictors selected The Affective Disorders Unit at the Karolinska University Hospital, Huddinge, is an outpatient tertiary remittal service unit for patients diagnosed with bipolar disorders. The catchment area is Southwest Greater Stockholm. All patients with bipolar I disorder may, however, be attached to the Unit ‘‘life-long’’ (currently about 650 patients), while other affective patients are returned to secondary care after consultation, and this means that the bipolar I patients are not selected for treatment resistance or severity because they were more treatment-resistant or exhibited a more severe course of the illness. Almost all patients at the unit (>95%) are treated with mood-stabilizers, most frequently lithium. The patients are monitored by blood tests to ensure lithium serum levels of 0.5e0.9 mmol/l at least 3 times a year. They have regular appointments with a psychiatrist at least once a year. About 1/12 (n z 50) of the patients are monitored by blood tests every month and the first 13 of the patients planned to be monitored each month were invited by mail to participate in the study up to a total of 150. One hundred
2.3.1. Primary retrospective variables Heredityddefined as parents, siblings, or children, diagnosed with bipolar disorder. ‘‘Age at onset’’, defined as the first estimated DSM-IVdefined affective episode. Comorbidity of substance-use disorder, anxiety disorder, or attention deficit/hyperactivity disorder, according to DSM-IV. Any episode with psychotic features. Any mixed episode, defined by at least 3 symptoms of depression in a manic episode/full blown mania, or at least 3 symptoms of mania in a Major Depressive Episode. Any period of defined rapid cycling. 2.3.2. Calculated retrospective variables Latency between the first DSM-IV affective episode and the first consultation.
L. Backlund et al. / European Psychiatry 24 (2009) 171e177
Latency between the first affective episode and the first instigation of a mood-stabilizer (‘‘Latency of treatment’’). Separate calculations were also carried out on the group treated only with lithium (‘‘Latency of lithium’’). Latency between the first consultation by a physician and the start of mood-stabilizer treatment (‘‘Doctors latency’’). Separate calculations were also carried out on the group treated only with lithium. Early onset of illness, i.e. before 20 years of age. The percentage of days spent in episodes, depressions and manias from onset of illness (‘‘Part of time in.’’). The number of depressive and manic episodes per year since the onset of illness. Burden of illness (burden of all episodes, of manias, and of depressions, respectively)di.e. the annual burden calculated as the accumulated sum of the length of each episode multiplied by an estimate of its functional disability as mild, moderate, or severe (¼1, 2, or 3) [15]. 2.3.3. Potential outcome variables In order to be able to evaluate the treatment received, we tested several methods of estimating outcome measures. Basically, we compared different ‘‘illness variables’’ before and after the introduction of mood-stabilizers: Difference in the number of episodes/year before and after the start of lithium/mood-stabilizer. Difference in the annual burden of illness before and after the start of lithium/mood-stabilizer. Total remission for more than 3 years after the start of lithium maintenance treatment. Burden of illness after the introduction of pharmacological treatment.
2.4. Statistical analyses We used a computer program ‘Chronos’ [14] further developed in-house to create a graphic and numerical ‘life chart’ for each patient. SPSS software was used for all the statistical analyses. The distribution of all variables was examined by frequency histograms. Analyses were performed to ensure that there was no violation of the assumptions of normality, linearity and homoscedasticity. In the case of distinctly non-normal continuous distributions, the variables were normalised using square roots, log10, or natural logarithmic transformations. Student’s t-tests were used to compare means of continuous variables, the Chi sq-tests to compare dichotomous variables and pair-wise t-test comparisons to calculate the effects of lithium usage. Relationships between continuous variables were investigated using Pearson’s correlation coefficient. Multiple and logistic regression analyses were used to compare the relative influences of different predictors on the effect of the mood-stabilizers. The study was approved by the Ethics Committee of Karolinska Institute.
173
3. Results 3.1. Description of the sample The average time from the onset of illness to interview was 23.8 13.7 years (1e65, Md ¼ 24). The patients had spent 20% of their life in episodes. Ninety-two patients had been hospitalized for an affective episode at least once. Eleven had the sequence of DepressioneManiaeFree interval, while 41 had the sequence ManiaeDepressioneFree interval, and 45 had an Irregular Course [21]. One of the latter patients lacked free intervals, and showed a chronic continuous course. Twelve patients had a prominent seasonal variationdthere were no differences between the 4 different types of sequenced patients. Three patients could not be categorized in this way due to the small number of episodes. Thirty-seven participants had experienced mixed episodes, and 16 had showed rapid cycling at least once since the onset of illness. Eighty-two participants had at least one relative of first-degree with an affective disorder (38 unipolar and 34 bipolar) (Table 1). 3.2. Early onset of illness The mean age of onset was 23 years (Md ¼ 20) and 44 had an early onset (i.e. under 20 years). The patients with early onset had waited longer until maintenance with a moodstabilizer were introduced than those aged 18 years or more at onset (delay 18.1 vs. 10.7 years, t[98] ¼ 3.6, p < 0.001). A similar comparison with regard to the time from first consultation to the introduction of a mood-stabilizer (‘‘doctor’s delay’’) also showed a significant difference (12.4 vs. 8.1 years, t[98] ¼ 2.4, p < 0.02). In the total sample of patients, we found a highly significant negative correlation between age at onset and age at introduction of a mood-stabilizer
Table 1 Morbidity and treatment events.
Age at first episode Age from onset Age at first mania Age at first hospitalization (n ¼ 92) Latency before consultation Latency before mood-stabilizer (n ¼ 99) Latency between first mania and mood-stabilizer (n ¼ 99) Fraction of time in episodes (%) Fraction of time in mania (%) Fraction of time in depression (%) Number of episodes/year Number of manias/year Number of depressions/year Burden of illness Burden of depression Burden of mania
Mean
Min
Max
Percentiles 25
50
75
23.0 23.8 28.5 29.4
4.0 1.0 7.0 9.0
52.0 65.0 63.0 54.0
17.0 14.0 20.0 20.3
20.0 23.7 26.0 27.5
26.8 33.5 30.0 34.0
3.8 14.1
0.0 0.1
38.8 44.8
0.0 4.3
0.2 12.2
4.7 21.8
9.3
0.0
40.8
1.1
6.1
14.7
20.6 6.4 14.1 1.1 0.5 0.6 13.1 19.7 11.9
1.9 0.1 0.0 0.1 0.02 0.0 1.4 0.0 0.1
94.5 34.4 60.1 13.3 6.7 6.7 63.8 84.9 99.9
9.4 2.3 5.5 0.4 0.1 0.2 5.7 5.9 2.6
17.5 4.3 11.6 0.7 0.4 0.4 10.5 17.4 7.1
28.8 8.8 20.0 1.3 0.5 0.8 17.4 27.9 14.1
L. Backlund et al. / European Psychiatry 24 (2009) 171e177
174
(r ¼ 0.4, n ¼ 100, p < 0.001). Furthermore, there was a negative correlation between the age of onset and the reduction of episodes per year (r ¼ 0.3, n ¼ 100, p < 0.01) but no significant correlation between age of onset and different aspects of severity during treatment with moodstabilizer. 3.3. Latency before treatment Ninety-nine of the participants were on maintenance treatment with mood-stabilizer(s), while one had refused pharmacological treatment. Sixty-seven of the 100 patients had their first clinical visit for affective symptoms within one year from onset. In spite of that short latency they had spent on average 14.1 years (Md ¼ 12.2) from the first symptoms to the introduction of mood-stabilizer, during which time they experienced 12.3 episodes (6.6 depressive and 5.6 manic). Only 5 patients had never experienced any manic episode at the time of the start of their treatment with lithium, and 21 started within one year after their first mania. 3.4. Effectiveness of mood-stabilizers The average duration of the maintenance treatment was 9.7 years (Md ¼ 5.8). Ninety-one of the patients were prescribed lithium. Of these 91, 72 had lithium in monotherapy, while 12 had valproic acid, 6 lamotrigine, and 1 both drugs in addition to lithium. In 7 cases, lithium had been discontinued and replaced by monotherapy with valproic acid (4 cases), valproic acid combined with lamotrigine (1 case), and topiramate in monotherapy (2 cases). Of those 9 patients who currently were not on lithium, 5 had previously had lithium maintenance treatment for at least one year (4 for more than 3 years). In 4 cases, the patients had refused lithium treatment, in 3 cases, lithium was discontinued due to side effects, and in 2 due to lack of effect. Only one patient had no prophylactic treatment at all. The patients who had lithium in combination with other mood-stabilizers (n ¼ 28) had more often comorbidity for anxiety disorder (48% vs. 23%, c2 ¼ 5.8, p < 0.05). Comparing the course of illness before and during prophylactic treatment with mood-stabilizers, we found distinct improvements in certain aspectsdthe number of depressive episodes per year, and the combination of duration and severity (i.e. burden) of both manic and depressive
episodes (Table 2). In those who had lithium in monotherapy there were significant decreases ( p < 0.001) in all measured variables (i.e. difference of episodes, manias and depressions, part of time in episodes, manias and depressions and burden of episodes, manias and depressions). Clinical improvements were seen in all those patients who had started lithium treatment within 4.1 years from the onset of illness (n ¼ 25), and in 39 of the 42 who had started lithium treatment within 10 years, whereas those who started later showed a poorer response (Fig. 1). There was a negative correlation between latency before the introduction of lithium and reduction of the number of episodes per year, as well as the part of time in episodes and, foremost, the burden of episodes (r ¼ 0.4, n ¼ 86, p < 0.001). The relative importance of different predictors for lithium response was determined by logistic regression. The predictors were first identified in univariate analyses. We selected the patients who had been on lithium treatment for at least one year (n ¼ 86), and chose the difference in the burden of illness, dichotomized at the median (4.2), as the outcome measure. A good outcome was predicted by a high burden of mania (OR ¼ 4.2, CI 1.5e12.2), a high burden of depression (OR ¼ 3.4, CI 1.1e10.4) and the absence of mixed episodes (OR ¼ 3.8, CI 1.3e11.3). A short latency before the introduction of treatment was not a significant predictor in this analysis. Outcome measures comparing different aspects of severity of illness before and after the introduction of lithium treatment depend per definition, however, not only on the outcome but also on the severity of the illness before lithium. We, therefore, also used an absolute outcome measure, independent of severity before lithium: the absence of affective episodes
Table 2 Aspects of illness before and during lithium maintenance treatment (n ¼ 86).
Number of episodes/year Number of depressions/year Number of manias/year Part of time in episodes (%) Part of time in depression (%) Part of time in mania (%) Burden of illness Burden of depression Burden of mania
Before treatment
During treatment
t
p
1.4 1.4 0.7 27.1 16.8 11.0 19.8 11.2 8.4
0.8 0.3 0.4 15.6 9.8 5.3 9.4 5.6 3.7
7.664 5.515 0.167 1.541 0.979 1.501 7.931 2.448 4.727
<0.001 <0.001 n.s. n.s. n.s. n.s. <0.001 <0.02 <0.001
Fig. 1. The reduction in the number of episodes per year in relation to the latency before lithium maintenance treatment. Among the 25 patients who had lithium within 10 years after the onset, there were only 3 ‘‘non-responders’’. These are indicated with arrows. They started on lithium after 4.1, 5.6, and 8.7 years of latency, respectively.
L. Backlund et al. / European Psychiatry 24 (2009) 171e177 Table 3 A comparison of the most interesting predictors for the alternative outcomes of lithium response, a. no episodes within 3 years of lithium treatment and a b. low burden during treatment (median-split 7.5) selected by cross tabulation and the Chi sq-tests. The continuous variables were dichotomized according to the median. The comparisons were made by logistic regressions. a. Factor
Risk Ratio (CI 95%)
Onset of illness after 20 No comorbidity No mixed episodes before lithium No periods of rapid cycling before lithium High burden of mania before lithium (median-split) High burden of depression before lithium (median-split)
3.4 3.3 3.5 7.3 1.0 0.8
b. Factor
Risk Ratio (CI 95%)
Onset of illness after 20 No comorbidity before lithium No mixed episodes before lithium No periods of rapid cycling before lithium High burden of mania before lithium (median-split) High burden of depression before lithium (median-split)
1.1 3.8 2.8 9.6 0.9 1.2
(1.156e10.197) (1.056e10.230) (1.161e10.477) (1.280e41.180) (0.303e3.385) (0.237e2.482)
(0.404e2.869) (1.378e4.998) (1.008e7.843) (1.800e51.551) (0.324e3.048) (0.420e3.312)
during the first 3 years after the initiation of lithium treatment. Thirty-five of the 77 (45%) participants who had had lithium for at least 3 years were responders in this sense. In a logistic regression analysis with this absolute outcome measure, the absence of mixed episodes, rapid cycling, comorbidity or early onset predicted a good response to lithium (Table 3). The absence of rapid cycling was the most important predictor (OR ¼ 7.3). The latency or burdens of mania or depression before lithium did not affect the outcome. We also used another absolute measure of the outcome of lithium treatmentdthe burden after introduction of lithium, dichotomized according to the median. With the exception of early onset, a good response was predicted by the same variables (Table 3).
4. Discussion One hundred bipolar patients were investigated in this retrospective study. After describing the sample in detail, we attempted to test the usefulness of different predictors and outcome variables for the evaluation of the response to treatment, of potential use in planned prospective studies. A limitation was the retrospective design where the important measures of outcome, the functional recovery and subsyndromal symptoms between episodes were impossible to establish. In the on-going prospective investigation we can study this in detail. In our sample, the average percentage of time in major depression was 14% and in mania 6.4%, which is close to data from other investigations [2,4,19]. The relatively low average age at the onset of illness (23.8 years, Md ¼ 20) was similar to pooled data from 15 studies published after 1990 [17]. The skewed gender ratio with a slight majority of women (59%) concurs with ordinary clinical experience, and we consider
175
that our sample is representative of bipolar patients treated in general psychiatry. We note that there was a remarkable latency between the first affective episode and the introduction of mood-stabilizing medicationdnearly one and a half decade, on average. During this time delay, the patients had on average experienced 7 depressive and 6 manic episodes. This latency is almost twice that found in several other investigations [3,5,29]. One reason for this discrepancy may be the high mean age of our sample (46.8 years), which means that several patients experienced the onset of illness before lithium was available as a medication. We found a strong negative correlation (r ¼ 0.7, p < 0.001) between the calendar year of onset and the latency before treatment, which supported this interpretation. The patients with onset of illness in the last 2 decades had shorter latencies (Md ¼ 4.2 years), in agreement with data reported in other modern investigations. Patients with an early onset of illness were subjected to latencies of almost 20 years before they received a moodstabilizer. This predicted a worse response to lithium according to one of our absolute outcome measures, and the latency certainly constitutes a prolonged time of suffering, functional disability, and risk of suicide. Eighty-two participants had at least one relative of the firstdegree who suffered from an affective disorder. In 38 cases, a first-degree relative had a unipolar and in 34 cases a bipolar illness. This well-known hereditary overlap between unipolar and bipolar disorders was not the focus of this particular study, but it will be considered in more detail in another on-going investigation of the relationship between clinical phenotypes and genetic factors. To analyse the effects of lithium treatment, monotherapy would be preferable. Selecting the patients with monotherapy would, however, introduce bias since in a naturalistic study combination therapy is associated with an insufficient effect of lithium alone. All but 2 patients had started with lithium monotherapy. We, therefore, preferred to include in the analyses all the patients who had been treated with lithium for at least one year. We found that the introduction of lithium was associated with a decrease in all measured outcome variables, particularly in the burden of manic and depressives episodes, with the number of manias per year as the only exception. We believe this exception might be because our relatively small sample is sensitive to random events (type II error). We suggest that the combination of duration and severity of episodes, as first described by Ehnvall et al. in 2003 [15], is the most valid outcome measure of functional disability due to affective episodes. In agreement with Baethge and Baldessarini, we found that different aspects of severity pre-treatment, but not latency, were important predictors for the reduction of severity. Since the difference in severity before and after the introduction of a mood-stabilizer depends on the severity of the illness before the introduction it is not a good measure of the effect of mood-stabilizers [29]. Therefore, we also tested 2 absolute measures of treatment response, independent of pre-
176
L. Backlund et al. / European Psychiatry 24 (2009) 171e177
treatment severity: (1) whether or not the patients experienced a total remission during the 3 years after the introduction of lithium, and (2) the burden of illness after the introduction of lithium. When a good outcome was defined as total remission for at least 3 years after the start of lithium treatment, the absence of previous mixed episodes as well as the absence of rapid cycling predicted a good response. The latency prior to lithium treatment did not, however, influence the outcome. When a good outcome was defined as a lower than median burden of illness during treatment, the result was the same, except that it was not influenced by early onset. 4.1. Conclusions Since severity pre-treatment and reduction of severity are interrelated, a comparison between the severity before and after treatment is not an adequate measure of the outcome for the estimation of treatment effects. The use absolute measures of treatment effects are preferable. Our results are in accordance with the clinical observation that prophylactic treatment with lithium is related to a good prognosis. The burden of illness, as a combination of severity and duration of episodes, was the most valid outcome measure. With this measure, the absence of rapid cycling, mixed episodes and possibly an onset of illness at 20 years of age or later predicted a good outcome of lithium treatment.
[10]
[11]
[12]
[13]
[14]
[15]
[16]
[17]
[18]
[19]
Acknowledgements [20]
We thank Birgitta Lindberg and Bo¨rje Suvinen for aid in collecting patient data. References [1] Akiskal HS. The prevalent clinical spectrum of bipolar disorders: beyond DSM-IV. J Clin Psychopharmacol 1996;16:4Se14S. [2] Angst J, Sellaro R. Historical perspectives and natural history of bipolar disorder. Biol Psychiatry 2000;48:445e57. [3] Baethge C, Smolka MN, Gruschka P, Berghofer A, Schlattmann P, Bauer M, et al. Does prophylaxis-delay in bipolar disorder influence outcome? Results from a long-term study of 147 patients. Acta Psychiatr Scand 2003;107:260e7. [4] Baldessarini RJ, Tondo L, Baethge CJ, Lepri B, Bratti IM. Effects of treatment latency on response to maintenance treatment in manicdepressive disorders. Bipolar Disord 2007;9:386e93. [5] Baldessarini RJ, Tondo L, Hennen J. Treatment-latency and previous episodes: relationships to pretreatment morbidity and response to maintenance treatment in bipolar I and II disorders. Bipolar Disord 2003; 5:169e79. [6] Baldessarini RJ, Tondo L, Viguera AC. Discontinuing lithium maintenance treatment in bipolar disorders: risks and implications. Bipolar Disord 1999;1:17e24. [7] Berk M, Hallam K, Lucas N, Hasty M, McNeil CA, Conus P, et al. Early intervention in bipolar disorders: opportunities and pitfalls. Med J Aust 2007;187:S11e4. [8] Bryant-Comstock L, Stender M, Devercelli G. Health care utilization and costs among privately insured patients with bipolar I disorder. Bipolar Disord 2002;4:398e405. [9] Denicoff KD, Ali SO, Sollinger AB, Smith-Jackson EE, Leverich GS. Post RM utility of the daily prospective National Institute of Mental
[21]
[22]
[23]
[24]
[25] [26]
[27]
[28] [29]
[30]
Health Life-Chart Method (NIMH-LCM-p) ratings in clinical trials of bipolar disorder. Depress Anxiety 2002;15:1e9. Denicoff KD, Leverich GS, Nolen WA, Rush AJ, McElroy SL, Keck PE, et al. Validation of the prospective NIMH-Life-Chart Method (NIMHLCM-p) for longitudinal assessment of bipolar illness. Psychol Med 2000;30:1391e7. Duffy A, Grof P, Robertson C, Alda M. The implications of genetics studies of major mood disorders for clinical practice. J Clin Psychiatry 2000;61:630e7. Dunayevich E, Sax KW, Keck Jr PE, McElroy SL, Sorter MT, McConville BJ, et al. Twelve-month outcome in bipolar patients with and without personality disorders. J Clin Psychiatry 2000;61:134e9. Dunayevich E, Strakowski SM, Sax KW, Sorter MT, Keck Jr PE, McElroy SL, et al. Personality disorders in first- and multiple-episode mania. Psychiatry Res 1996;64:69e75. ˚ gren H. Patterns of sensitisation in the course of affective Ehnvall A, A illness. A life-charting study of treatment-refractory depressed patients. J Affect Disord 2002;70:67e75. ˚ gren H. Lifetime burden of Ehnvall A, Sjo¨gren M, Zachrisson OC, A mood swings and activation of brain norepinephrine turnover in patients with treatment-refractory depressive illness. J Affect Disord 2003;74: 185e9. Franchini L, Zanardi R, Smeraldi E, Gasperini M. Early onset of lithium prophylaxis as a predictor of good long-term outcome. Eur Arch Psychiatry Clin Neurosci 1999;249:227e30. Goodwin FK, Jamison KR. Manic-depressive illness: bipolar disorders and recurrent depression. New York: Oxford University Press; 2007. p. 120e5. Judd LL, Akiskal HS, Schettler PJ, Coryell W, Endicott J, Maser JD, et al. A prospective investigation of the natural history of the long-term weekly symptomatic status of bipolar II disorder. Arch Gen Psychiatry 2003;60:261e9. Judd LL, Akiskal HS, Schettler PJ, Endicott J, Maser J, Solomon DA, et al. The long-term natural history of the weekly symptomatic status of bipolar I disorder. Arch Gen Psychiatry 2002;59:530e7. Kessler RC, McGonagle KA, Zhao S, Nelson CB, Hughes M, Eshleman S, et al. Lifetime and 12-month prevalence of DSM-III-R psychiatric disorders in the United States. Results from the National Comorbidity Survey. Arch Gen Psychiatry 1994;51:8e19. Koukopoulos A, Sani G, Koukopoulos AE, Giraldi P. Cyclicity and manic-depressive illness. Dordrecht: Kluwer Academics; 2000. Chapter in a book. Leverich GS, Nolen WA, Rush AJ, McElroy SL, Keck PE, Denicoff KD, et al. The Stanley foundation bipolar treatment outcome network. I. Longitudinal methodology. J Affect Disord 2001;67:33e44. Maj M. The impact of lithium prophylaxis on the course of bipolar disorder: a review of the research evidence. Bipolar Disord 2000;2: 93e101. Marneros A, Deister A, Rohde A. Affective, schizoaffective and schizophrenic psychoses. A comparative long-term study. Monogr Gesamtgeb Psychiatr Psychiatry Ser 1991;65:1e454. Miklowitz DJ, Johnson SL. The psychopathology and treatment of bipolar disorder. Annu Rev Clin Psychol 2006;2:199e235. ¨ sby U, Brandt L, Correia N, Ekbom A, Spare´n P. Excess mortality in O bipolar and unipolar disorder in Sweden. Arch Gen Psychiatry 2001;58: 844e50. Regier DA, Boyd JH, Burke Jr JD, Rae DS, Myers JK, Kramer M, et al. One-month prevalence of mental disorders in the United States. Based on five epidemiologic catchment area sites. Arch Gen Psychiatry 1988;45: 977e86. Rohayem J, Bayle JF, Richa S. Predictors of prophylactic response to lithium. Encephale 2008;34:394e9. Sharma P, Basu D, Thapar S. Latency and outcome of prophylaxis in bipolar disorder: role of severity as a confounding variable? Acta Psychiatr Scand 2004;109:157 [author reply 158]. Sheehan DV, Lecrubier Y, Harnett-Sheehan K. Reliability and validity of the M.I.N.I International Neuropsychiatric Interview (M.I.N.I.): according to the SCID-P. Eur Psychiatry 1997;12:232e41.
L. Backlund et al. / European Psychiatry 24 (2009) 171e177 [31] Skeppar P, Adolfsson R. Bipolar II and the bipolar spectrum. Nord J Psychiatry 2006;60:7e26. [32] Tondo L, Baldessarini RJ, Hennen J, Floris G. Lithium maintenance treatment of depression and mania in bipolar I and bipolar II disorders. Am J Psychiatry 1998;155:638e45. [33] Tondo L, Baldessarini RJ, Hennen J, Minnai GP, Salis P, Scamonatti L, et al. Suicide attempts in major affective disorder patients with comorbid substance use disorders. J Clin Psychiatry 1999;60(Suppl. 2):63e9 [discussion 75e6, 113e6].
177
[34] Tondo L, Isacsson G, Baldessarini R. Suicidal behaviour in bipolar disorder: risk and prevention. CNS Drugs 2003;17:491e511. [35] Vieta E, Pacchiarotti I, Scott J, Sanchez-Moreno J, Di Marzo S, Colom F. Evidence-based research on the efficacy of psychological interventions in bipolar disorders: a critical review. Curr Psychiatry Rep 2005;7:449e55. [36] Winokur G, Coryell W, Keller M, Endicott J, Akiskal H. A prospective follow-up of patients with bipolar and primary unipolar affective disorder. Arch Gen Psychiatry 1993;50:457e65.
www.sciencedirect.com European Psychiatry 24 (2009) 171e177
Original article
Identifying predictors for good lithium response e A retrospective analysis of 100 patients with bipolar disorder using a life-charting method ˚ gren c Lena Backlund a,*, Anna Ehnvall b, Jerker Hetta a, Go¨ran Isacsson a, Hans A a
Karolinska Institute, Department of Clinical Neuroscience, Section of Psychiatry, Karolinska University Hospital, Huddinge M57, SE-141 86 Stockholm, Sweden b Institute for Clinical Neuroscience, Gothenburg and Psychiatric Out-patient Clinic, Varberg, Swedan c Sahlgrenska Academy at the University of Gothenburg, Institute of Neuroscience and Physiology, Department of Psychiatry and Neurochemistry, Gothenburg, Sweden Received 18 August 2008; received in revised form 17 December 2008; accepted 30 December 2008 Available online 14 March 2009
Abstract Purpose. e Our aim was to investigate bipolar patients in order to test the validity of various outcome measures and to identify prognostic predictors for pharmacological treatment. Material and method. e One hundred patients were interviewed using a computerized life-charting program in a descriptive, retrospective analysis. The concept ‘‘Burden of illness’’ was defined as a combination of severity and duration of episodes. Response to treatment was defined as the difference in burden before and after treatment, a low burden during treatment, and freedom of episodes for at least 3 years after insertion of treatment. Results. e The absence of mixed episodes and a high initial burden predicted a good response measured as the difference in burden. If remission for 3 years or a low burden during lithium treatment was used, the absence of rapid cycling and of mixed episodes were the most important predictors. The severity of illness before treatment had no impact. Discussion and conclusion. e We suggest the use of absolute measures of severity during treatment as the most appropriate measure of the outcome. Furthermore, our data provide corroboration that treatment with lithium ameliorates the prognosis of the illness, but that mixed episodes and rapid cycling predict a poorer response to lithium. Ó 2009 Elsevier Masson SAS. All rights reserved. Keywords: Bipolar disorder; Treatment outcome; Lithium
1. Introduction Bipolar affective disorder is a common psychiatric disease. The lifetime prevalence of classic bipolar I disorder is estimated to be 0.9e1.6% [20,27]. The median age at onset of illness is 21 years and more than 80% of the patients with a first episode will recur [36]. In most cases, the course of the disease is progressive [2,24]. Suicide rates are more than 20fold higher than in the general population [26,33,34]. Several investigations have shown that there is on average a delay of 5 years between a patient’s first illness episode and * Corresponding author. Tel./fax: þ46 8 585 857 08. E-mail address: [email protected] (L. Backlund). 0924-9338/$ - see front matter Ó 2009 Elsevier Masson SAS. All rights reserved. doi:10.1016/j.eurpsy.2008.12.009
the first consultation of a physician, and a further 8 years or so (range 7.8e9.3) until the correct diagnosis is established [3,5,32]. There is also a delay of several years until adequate mood-stabilizing medication is prescribed [8]. This delay may be ominous, since the early diagnosis and instigation of pharmacological treatment may be crucial for a good prognosis [16,32]. The aetiology of bipolar illness is multifactorial. Genetic factors [11,25,28] and pharmacologic maintenance treatment [5,6,23] seem to be crucial factors for the prognosis. Abilities to cope with stress, sustain close friendship and good relationships with family as well as a high adherence to the treatment [7,35] predict a good prognosis, while mixed episodes [1], a poorer social function, comorbidity with
172
L. Backlund et al. / European Psychiatry 24 (2009) 171e177
borderline personality disorders [12,13] or substance-use disorder [31] may have a more negative effect on the outcome. The need for continuous mood-stabilizing medication is wellknown, but the importance of its early instigation is still a matter of discussion. Franchini and Tondo found a relationship between the latency before treatment and the reduction of severity after prophylactic treatment [16,32]. Baethge showed a positive correlation between the grade of severity before treatment and the reduction of severity [3]. Later, Baldessarini investigated this systematically, and concluded that ‘‘prior episode counts and treatment delay have little association with morbidity during prophylaxis with mood-stabilizing agents. Comparisons of morbidity during vs. before treatment in episodic disorders are misleading’’ [4,5]. He also pointed out the need for investigations where functional disability is a part of the outcome. In weekly assessments during 13.4 years, Judd et al. found a high frequency of sub-syndromal affective symptoms between the major affective episodes and discussed the functional disability between episodes as an important outcome [18]. 1.1. Aims of the study This study focuses on the temporal patterns in bipolar disorders. The aims were to investigate retrospectively a clinical sample as a foundation for future prospective analyses, and to identify possible prognostic predictors with an emphasis on pharmacological treatment. We also wished to test innovative methods for measuring the burden of illness as a combination of the duration and the severity of episodes [15], and to test the validity of different outcome measures. 2. Subjects and methods
(41 males and 59 females) out of 150 invited had agreed and given informed consent, were included in the study. No patient was excluded by the investigators. The participants in the study group did not differ from those who refused to participate with respect to gender, diagnosis, or age. The mean age at the time of the first interview was 46.8 years (Md ¼ 48.5) and they had on average been ill for 23.8 years (Md ¼ 23.7). Eighty-eight had been diagnosed with bipolar I and the remaining 12 with bipolar II disorder. Fifty-six were married, 2 were widowed, 27 were divorced, and 16 had never had any long relationship. All except one had passed elementary school. Forty-five had completed high school, and 33 had graduated from university. Ninety-two had been hospitalized at least once, and 58 of these had been subjected to compulsory admission. 2.2. Procedures All the patients were interviewed between January 2001 and February 2004 by a psychiatrist (LB), a psychiatric nurse, or a psychologist, using a semi-structured interview according to a life-charting protocol [9,10,22] previously developed for ˚ gren [14]. For diagnostic the purpose by Ehnvall and A confirmation of bipolar comorbid disorders, the Mini International Neuropsychiatric Interview (M.I.N.I) was used [30]. The participants were carefully interviewed about their family history, life events, social history, previous depressions, manias, mixed episodes occurring since early childhood, and earlier and current treatments. The accurate determination of dates of each event was a primary concern. The DSM-IV criteria were used to fix the diagnoses of manic and depressive episodes. All available medical records were searched to validate the verbal reports. When the 2 sources of data were inconsistent, the patient was asked more detailed questions to provide clarification.
2.1. Patients 2.3. Potential predictors selected The Affective Disorders Unit at the Karolinska University Hospital, Huddinge, is an outpatient tertiary remittal service unit for patients diagnosed with bipolar disorders. The catchment area is Southwest Greater Stockholm. All patients with bipolar I disorder may, however, be attached to the Unit ‘‘life-long’’ (currently about 650 patients), while other affective patients are returned to secondary care after consultation, and this means that the bipolar I patients are not selected for treatment resistance or severity because they were more treatment-resistant or exhibited a more severe course of the illness. Almost all patients at the unit (>95%) are treated with mood-stabilizers, most frequently lithium. The patients are monitored by blood tests to ensure lithium serum levels of 0.5e0.9 mmol/l at least 3 times a year. They have regular appointments with a psychiatrist at least once a year. About 1/12 (n z 50) of the patients are monitored by blood tests every month and the first 13 of the patients planned to be monitored each month were invited by mail to participate in the study up to a total of 150. One hundred
2.3.1. Primary retrospective variables Heredityddefined as parents, siblings, or children, diagnosed with bipolar disorder. ‘‘Age at onset’’, defined as the first estimated DSM-IVdefined affective episode. Comorbidity of substance-use disorder, anxiety disorder, or attention deficit/hyperactivity disorder, according to DSM-IV. Any episode with psychotic features. Any mixed episode, defined by at least 3 symptoms of depression in a manic episode/full blown mania, or at least 3 symptoms of mania in a Major Depressive Episode. Any period of defined rapid cycling. 2.3.2. Calculated retrospective variables Latency between the first DSM-IV affective episode and the first consultation.
L. Backlund et al. / European Psychiatry 24 (2009) 171e177
Latency between the first affective episode and the first instigation of a mood-stabilizer (‘‘Latency of treatment’’). Separate calculations were also carried out on the group treated only with lithium (‘‘Latency of lithium’’). Latency between the first consultation by a physician and the start of mood-stabilizer treatment (‘‘Doctors latency’’). Separate calculations were also carried out on the group treated only with lithium. Early onset of illness, i.e. before 20 years of age. The percentage of days spent in episodes, depressions and manias from onset of illness (‘‘Part of time in.’’). The number of depressive and manic episodes per year since the onset of illness. Burden of illness (burden of all episodes, of manias, and of depressions, respectively)di.e. the annual burden calculated as the accumulated sum of the length of each episode multiplied by an estimate of its functional disability as mild, moderate, or severe (¼1, 2, or 3) [15]. 2.3.3. Potential outcome variables In order to be able to evaluate the treatment received, we tested several methods of estimating outcome measures. Basically, we compared different ‘‘illness variables’’ before and after the introduction of mood-stabilizers: Difference in the number of episodes/year before and after the start of lithium/mood-stabilizer. Difference in the annual burden of illness before and after the start of lithium/mood-stabilizer. Total remission for more than 3 years after the start of lithium maintenance treatment. Burden of illness after the introduction of pharmacological treatment.
2.4. Statistical analyses We used a computer program ‘Chronos’ [14] further developed in-house to create a graphic and numerical ‘life chart’ for each patient. SPSS software was used for all the statistical analyses. The distribution of all variables was examined by frequency histograms. Analyses were performed to ensure that there was no violation of the assumptions of normality, linearity and homoscedasticity. In the case of distinctly non-normal continuous distributions, the variables were normalised using square roots, log10, or natural logarithmic transformations. Student’s t-tests were used to compare means of continuous variables, the Chi sq-tests to compare dichotomous variables and pair-wise t-test comparisons to calculate the effects of lithium usage. Relationships between continuous variables were investigated using Pearson’s correlation coefficient. Multiple and logistic regression analyses were used to compare the relative influences of different predictors on the effect of the mood-stabilizers. The study was approved by the Ethics Committee of Karolinska Institute.
173
3. Results 3.1. Description of the sample The average time from the onset of illness to interview was 23.8 13.7 years (1e65, Md ¼ 24). The patients had spent 20% of their life in episodes. Ninety-two patients had been hospitalized for an affective episode at least once. Eleven had the sequence of DepressioneManiaeFree interval, while 41 had the sequence ManiaeDepressioneFree interval, and 45 had an Irregular Course [21]. One of the latter patients lacked free intervals, and showed a chronic continuous course. Twelve patients had a prominent seasonal variationdthere were no differences between the 4 different types of sequenced patients. Three patients could not be categorized in this way due to the small number of episodes. Thirty-seven participants had experienced mixed episodes, and 16 had showed rapid cycling at least once since the onset of illness. Eighty-two participants had at least one relative of first-degree with an affective disorder (38 unipolar and 34 bipolar) (Table 1). 3.2. Early onset of illness The mean age of onset was 23 years (Md ¼ 20) and 44 had an early onset (i.e. under 20 years). The patients with early onset had waited longer until maintenance with a moodstabilizer were introduced than those aged 18 years or more at onset (delay 18.1 vs. 10.7 years, t[98] ¼ 3.6, p < 0.001). A similar comparison with regard to the time from first consultation to the introduction of a mood-stabilizer (‘‘doctor’s delay’’) also showed a significant difference (12.4 vs. 8.1 years, t[98] ¼ 2.4, p < 0.02). In the total sample of patients, we found a highly significant negative correlation between age at onset and age at introduction of a mood-stabilizer
Table 1 Morbidity and treatment events.
Age at first episode Age from onset Age at first mania Age at first hospitalization (n ¼ 92) Latency before consultation Latency before mood-stabilizer (n ¼ 99) Latency between first mania and mood-stabilizer (n ¼ 99) Fraction of time in episodes (%) Fraction of time in mania (%) Fraction of time in depression (%) Number of episodes/year Number of manias/year Number of depressions/year Burden of illness Burden of depression Burden of mania
Mean
Min
Max
Percentiles 25
50
75
23.0 23.8 28.5 29.4
4.0 1.0 7.0 9.0
52.0 65.0 63.0 54.0
17.0 14.0 20.0 20.3
20.0 23.7 26.0 27.5
26.8 33.5 30.0 34.0
3.8 14.1
0.0 0.1
38.8 44.8
0.0 4.3
0.2 12.2
4.7 21.8
9.3
0.0
40.8
1.1
6.1
14.7
20.6 6.4 14.1 1.1 0.5 0.6 13.1 19.7 11.9
1.9 0.1 0.0 0.1 0.02 0.0 1.4 0.0 0.1
94.5 34.4 60.1 13.3 6.7 6.7 63.8 84.9 99.9
9.4 2.3 5.5 0.4 0.1 0.2 5.7 5.9 2.6
17.5 4.3 11.6 0.7 0.4 0.4 10.5 17.4 7.1
28.8 8.8 20.0 1.3 0.5 0.8 17.4 27.9 14.1
L. Backlund et al. / European Psychiatry 24 (2009) 171e177
174
(r ¼ 0.4, n ¼ 100, p < 0.001). Furthermore, there was a negative correlation between the age of onset and the reduction of episodes per year (r ¼ 0.3, n ¼ 100, p < 0.01) but no significant correlation between age of onset and different aspects of severity during treatment with moodstabilizer. 3.3. Latency before treatment Ninety-nine of the participants were on maintenance treatment with mood-stabilizer(s), while one had refused pharmacological treatment. Sixty-seven of the 100 patients had their first clinical visit for affective symptoms within one year from onset. In spite of that short latency they had spent on average 14.1 years (Md ¼ 12.2) from the first symptoms to the introduction of mood-stabilizer, during which time they experienced 12.3 episodes (6.6 depressive and 5.6 manic). Only 5 patients had never experienced any manic episode at the time of the start of their treatment with lithium, and 21 started within one year after their first mania. 3.4. Effectiveness of mood-stabilizers The average duration of the maintenance treatment was 9.7 years (Md ¼ 5.8). Ninety-one of the patients were prescribed lithium. Of these 91, 72 had lithium in monotherapy, while 12 had valproic acid, 6 lamotrigine, and 1 both drugs in addition to lithium. In 7 cases, lithium had been discontinued and replaced by monotherapy with valproic acid (4 cases), valproic acid combined with lamotrigine (1 case), and topiramate in monotherapy (2 cases). Of those 9 patients who currently were not on lithium, 5 had previously had lithium maintenance treatment for at least one year (4 for more than 3 years). In 4 cases, the patients had refused lithium treatment, in 3 cases, lithium was discontinued due to side effects, and in 2 due to lack of effect. Only one patient had no prophylactic treatment at all. The patients who had lithium in combination with other mood-stabilizers (n ¼ 28) had more often comorbidity for anxiety disorder (48% vs. 23%, c2 ¼ 5.8, p < 0.05). Comparing the course of illness before and during prophylactic treatment with mood-stabilizers, we found distinct improvements in certain aspectsdthe number of depressive episodes per year, and the combination of duration and severity (i.e. burden) of both manic and depressive
episodes (Table 2). In those who had lithium in monotherapy there were significant decreases ( p < 0.001) in all measured variables (i.e. difference of episodes, manias and depressions, part of time in episodes, manias and depressions and burden of episodes, manias and depressions). Clinical improvements were seen in all those patients who had started lithium treatment within 4.1 years from the onset of illness (n ¼ 25), and in 39 of the 42 who had started lithium treatment within 10 years, whereas those who started later showed a poorer response (Fig. 1). There was a negative correlation between latency before the introduction of lithium and reduction of the number of episodes per year, as well as the part of time in episodes and, foremost, the burden of episodes (r ¼ 0.4, n ¼ 86, p < 0.001). The relative importance of different predictors for lithium response was determined by logistic regression. The predictors were first identified in univariate analyses. We selected the patients who had been on lithium treatment for at least one year (n ¼ 86), and chose the difference in the burden of illness, dichotomized at the median (4.2), as the outcome measure. A good outcome was predicted by a high burden of mania (OR ¼ 4.2, CI 1.5e12.2), a high burden of depression (OR ¼ 3.4, CI 1.1e10.4) and the absence of mixed episodes (OR ¼ 3.8, CI 1.3e11.3). A short latency before the introduction of treatment was not a significant predictor in this analysis. Outcome measures comparing different aspects of severity of illness before and after the introduction of lithium treatment depend per definition, however, not only on the outcome but also on the severity of the illness before lithium. We, therefore, also used an absolute outcome measure, independent of severity before lithium: the absence of affective episodes
Table 2 Aspects of illness before and during lithium maintenance treatment (n ¼ 86).
Number of episodes/year Number of depressions/year Number of manias/year Part of time in episodes (%) Part of time in depression (%) Part of time in mania (%) Burden of illness Burden of depression Burden of mania
Before treatment
During treatment
t
p
1.4 1.4 0.7 27.1 16.8 11.0 19.8 11.2 8.4
0.8 0.3 0.4 15.6 9.8 5.3 9.4 5.6 3.7
7.664 5.515 0.167 1.541 0.979 1.501 7.931 2.448 4.727
<0.001 <0.001 n.s. n.s. n.s. n.s. <0.001 <0.02 <0.001
Fig. 1. The reduction in the number of episodes per year in relation to the latency before lithium maintenance treatment. Among the 25 patients who had lithium within 10 years after the onset, there were only 3 ‘‘non-responders’’. These are indicated with arrows. They started on lithium after 4.1, 5.6, and 8.7 years of latency, respectively.
L. Backlund et al. / European Psychiatry 24 (2009) 171e177 Table 3 A comparison of the most interesting predictors for the alternative outcomes of lithium response, a. no episodes within 3 years of lithium treatment and a b. low burden during treatment (median-split 7.5) selected by cross tabulation and the Chi sq-tests. The continuous variables were dichotomized according to the median. The comparisons were made by logistic regressions. a. Factor
Risk Ratio (CI 95%)
Onset of illness after 20 No comorbidity No mixed episodes before lithium No periods of rapid cycling before lithium High burden of mania before lithium (median-split) High burden of depression before lithium (median-split)
3.4 3.3 3.5 7.3 1.0 0.8
b. Factor
Risk Ratio (CI 95%)
Onset of illness after 20 No comorbidity before lithium No mixed episodes before lithium No periods of rapid cycling before lithium High burden of mania before lithium (median-split) High burden of depression before lithium (median-split)
1.1 3.8 2.8 9.6 0.9 1.2
(1.156e10.197) (1.056e10.230) (1.161e10.477) (1.280e41.180) (0.303e3.385) (0.237e2.482)
(0.404e2.869) (1.378e4.998) (1.008e7.843) (1.800e51.551) (0.324e3.048) (0.420e3.312)
during the first 3 years after the initiation of lithium treatment. Thirty-five of the 77 (45%) participants who had had lithium for at least 3 years were responders in this sense. In a logistic regression analysis with this absolute outcome measure, the absence of mixed episodes, rapid cycling, comorbidity or early onset predicted a good response to lithium (Table 3). The absence of rapid cycling was the most important predictor (OR ¼ 7.3). The latency or burdens of mania or depression before lithium did not affect the outcome. We also used another absolute measure of the outcome of lithium treatmentdthe burden after introduction of lithium, dichotomized according to the median. With the exception of early onset, a good response was predicted by the same variables (Table 3).
4. Discussion One hundred bipolar patients were investigated in this retrospective study. After describing the sample in detail, we attempted to test the usefulness of different predictors and outcome variables for the evaluation of the response to treatment, of potential use in planned prospective studies. A limitation was the retrospective design where the important measures of outcome, the functional recovery and subsyndromal symptoms between episodes were impossible to establish. In the on-going prospective investigation we can study this in detail. In our sample, the average percentage of time in major depression was 14% and in mania 6.4%, which is close to data from other investigations [2,4,19]. The relatively low average age at the onset of illness (23.8 years, Md ¼ 20) was similar to pooled data from 15 studies published after 1990 [17]. The skewed gender ratio with a slight majority of women (59%) concurs with ordinary clinical experience, and we consider
175
that our sample is representative of bipolar patients treated in general psychiatry. We note that there was a remarkable latency between the first affective episode and the introduction of mood-stabilizing medicationdnearly one and a half decade, on average. During this time delay, the patients had on average experienced 7 depressive and 6 manic episodes. This latency is almost twice that found in several other investigations [3,5,29]. One reason for this discrepancy may be the high mean age of our sample (46.8 years), which means that several patients experienced the onset of illness before lithium was available as a medication. We found a strong negative correlation (r ¼ 0.7, p < 0.001) between the calendar year of onset and the latency before treatment, which supported this interpretation. The patients with onset of illness in the last 2 decades had shorter latencies (Md ¼ 4.2 years), in agreement with data reported in other modern investigations. Patients with an early onset of illness were subjected to latencies of almost 20 years before they received a moodstabilizer. This predicted a worse response to lithium according to one of our absolute outcome measures, and the latency certainly constitutes a prolonged time of suffering, functional disability, and risk of suicide. Eighty-two participants had at least one relative of the firstdegree who suffered from an affective disorder. In 38 cases, a first-degree relative had a unipolar and in 34 cases a bipolar illness. This well-known hereditary overlap between unipolar and bipolar disorders was not the focus of this particular study, but it will be considered in more detail in another on-going investigation of the relationship between clinical phenotypes and genetic factors. To analyse the effects of lithium treatment, monotherapy would be preferable. Selecting the patients with monotherapy would, however, introduce bias since in a naturalistic study combination therapy is associated with an insufficient effect of lithium alone. All but 2 patients had started with lithium monotherapy. We, therefore, preferred to include in the analyses all the patients who had been treated with lithium for at least one year. We found that the introduction of lithium was associated with a decrease in all measured outcome variables, particularly in the burden of manic and depressives episodes, with the number of manias per year as the only exception. We believe this exception might be because our relatively small sample is sensitive to random events (type II error). We suggest that the combination of duration and severity of episodes, as first described by Ehnvall et al. in 2003 [15], is the most valid outcome measure of functional disability due to affective episodes. In agreement with Baethge and Baldessarini, we found that different aspects of severity pre-treatment, but not latency, were important predictors for the reduction of severity. Since the difference in severity before and after the introduction of a mood-stabilizer depends on the severity of the illness before the introduction it is not a good measure of the effect of mood-stabilizers [29]. Therefore, we also tested 2 absolute measures of treatment response, independent of pre-
176
L. Backlund et al. / European Psychiatry 24 (2009) 171e177
treatment severity: (1) whether or not the patients experienced a total remission during the 3 years after the introduction of lithium, and (2) the burden of illness after the introduction of lithium. When a good outcome was defined as total remission for at least 3 years after the start of lithium treatment, the absence of previous mixed episodes as well as the absence of rapid cycling predicted a good response. The latency prior to lithium treatment did not, however, influence the outcome. When a good outcome was defined as a lower than median burden of illness during treatment, the result was the same, except that it was not influenced by early onset. 4.1. Conclusions Since severity pre-treatment and reduction of severity are interrelated, a comparison between the severity before and after treatment is not an adequate measure of the outcome for the estimation of treatment effects. The use absolute measures of treatment effects are preferable. Our results are in accordance with the clinical observation that prophylactic treatment with lithium is related to a good prognosis. The burden of illness, as a combination of severity and duration of episodes, was the most valid outcome measure. With this measure, the absence of rapid cycling, mixed episodes and possibly an onset of illness at 20 years of age or later predicted a good outcome of lithium treatment.
[10]
[11]
[12]
[13]
[14]
[15]
[16]
[17]
[18]
[19]
Acknowledgements [20]
We thank Birgitta Lindberg and Bo¨rje Suvinen for aid in collecting patient data. References [1] Akiskal HS. The prevalent clinical spectrum of bipolar disorders: beyond DSM-IV. J Clin Psychopharmacol 1996;16:4Se14S. [2] Angst J, Sellaro R. Historical perspectives and natural history of bipolar disorder. Biol Psychiatry 2000;48:445e57. [3] Baethge C, Smolka MN, Gruschka P, Berghofer A, Schlattmann P, Bauer M, et al. Does prophylaxis-delay in bipolar disorder influence outcome? Results from a long-term study of 147 patients. Acta Psychiatr Scand 2003;107:260e7. [4] Baldessarini RJ, Tondo L, Baethge CJ, Lepri B, Bratti IM. Effects of treatment latency on response to maintenance treatment in manicdepressive disorders. Bipolar Disord 2007;9:386e93. [5] Baldessarini RJ, Tondo L, Hennen J. Treatment-latency and previous episodes: relationships to pretreatment morbidity and response to maintenance treatment in bipolar I and II disorders. Bipolar Disord 2003; 5:169e79. [6] Baldessarini RJ, Tondo L, Viguera AC. Discontinuing lithium maintenance treatment in bipolar disorders: risks and implications. Bipolar Disord 1999;1:17e24. [7] Berk M, Hallam K, Lucas N, Hasty M, McNeil CA, Conus P, et al. Early intervention in bipolar disorders: opportunities and pitfalls. Med J Aust 2007;187:S11e4. [8] Bryant-Comstock L, Stender M, Devercelli G. Health care utilization and costs among privately insured patients with bipolar I disorder. Bipolar Disord 2002;4:398e405. [9] Denicoff KD, Ali SO, Sollinger AB, Smith-Jackson EE, Leverich GS. Post RM utility of the daily prospective National Institute of Mental
[21]
[22]
[23]
[24]
[25] [26]
[27]
[28] [29]
[30]
Health Life-Chart Method (NIMH-LCM-p) ratings in clinical trials of bipolar disorder. Depress Anxiety 2002;15:1e9. Denicoff KD, Leverich GS, Nolen WA, Rush AJ, McElroy SL, Keck PE, et al. Validation of the prospective NIMH-Life-Chart Method (NIMHLCM-p) for longitudinal assessment of bipolar illness. Psychol Med 2000;30:1391e7. Duffy A, Grof P, Robertson C, Alda M. The implications of genetics studies of major mood disorders for clinical practice. J Clin Psychiatry 2000;61:630e7. Dunayevich E, Sax KW, Keck Jr PE, McElroy SL, Sorter MT, McConville BJ, et al. Twelve-month outcome in bipolar patients with and without personality disorders. J Clin Psychiatry 2000;61:134e9. Dunayevich E, Strakowski SM, Sax KW, Sorter MT, Keck Jr PE, McElroy SL, et al. Personality disorders in first- and multiple-episode mania. Psychiatry Res 1996;64:69e75. ˚ gren H. Patterns of sensitisation in the course of affective Ehnvall A, A illness. A life-charting study of treatment-refractory depressed patients. J Affect Disord 2002;70:67e75. ˚ gren H. Lifetime burden of Ehnvall A, Sjo¨gren M, Zachrisson OC, A mood swings and activation of brain norepinephrine turnover in patients with treatment-refractory depressive illness. J Affect Disord 2003;74: 185e9. Franchini L, Zanardi R, Smeraldi E, Gasperini M. Early onset of lithium prophylaxis as a predictor of good long-term outcome. Eur Arch Psychiatry Clin Neurosci 1999;249:227e30. Goodwin FK, Jamison KR. Manic-depressive illness: bipolar disorders and recurrent depression. New York: Oxford University Press; 2007. p. 120e5. Judd LL, Akiskal HS, Schettler PJ, Coryell W, Endicott J, Maser JD, et al. A prospective investigation of the natural history of the long-term weekly symptomatic status of bipolar II disorder. Arch Gen Psychiatry 2003;60:261e9. Judd LL, Akiskal HS, Schettler PJ, Endicott J, Maser J, Solomon DA, et al. The long-term natural history of the weekly symptomatic status of bipolar I disorder. Arch Gen Psychiatry 2002;59:530e7. Kessler RC, McGonagle KA, Zhao S, Nelson CB, Hughes M, Eshleman S, et al. Lifetime and 12-month prevalence of DSM-III-R psychiatric disorders in the United States. Results from the National Comorbidity Survey. Arch Gen Psychiatry 1994;51:8e19. Koukopoulos A, Sani G, Koukopoulos AE, Giraldi P. Cyclicity and manic-depressive illness. Dordrecht: Kluwer Academics; 2000. Chapter in a book. Leverich GS, Nolen WA, Rush AJ, McElroy SL, Keck PE, Denicoff KD, et al. The Stanley foundation bipolar treatment outcome network. I. Longitudinal methodology. J Affect Disord 2001;67:33e44. Maj M. The impact of lithium prophylaxis on the course of bipolar disorder: a review of the research evidence. Bipolar Disord 2000;2: 93e101. Marneros A, Deister A, Rohde A. Affective, schizoaffective and schizophrenic psychoses. A comparative long-term study. Monogr Gesamtgeb Psychiatr Psychiatry Ser 1991;65:1e454. Miklowitz DJ, Johnson SL. The psychopathology and treatment of bipolar disorder. Annu Rev Clin Psychol 2006;2:199e235. ¨ sby U, Brandt L, Correia N, Ekbom A, Spare´n P. Excess mortality in O bipolar and unipolar disorder in Sweden. Arch Gen Psychiatry 2001;58: 844e50. Regier DA, Boyd JH, Burke Jr JD, Rae DS, Myers JK, Kramer M, et al. One-month prevalence of mental disorders in the United States. Based on five epidemiologic catchment area sites. Arch Gen Psychiatry 1988;45: 977e86. Rohayem J, Bayle JF, Richa S. Predictors of prophylactic response to lithium. Encephale 2008;34:394e9. Sharma P, Basu D, Thapar S. Latency and outcome of prophylaxis in bipolar disorder: role of severity as a confounding variable? Acta Psychiatr Scand 2004;109:157 [author reply 158]. Sheehan DV, Lecrubier Y, Harnett-Sheehan K. Reliability and validity of the M.I.N.I International Neuropsychiatric Interview (M.I.N.I.): according to the SCID-P. Eur Psychiatry 1997;12:232e41.
L. Backlund et al. / European Psychiatry 24 (2009) 171e177 [31] Skeppar P, Adolfsson R. Bipolar II and the bipolar spectrum. Nord J Psychiatry 2006;60:7e26. [32] Tondo L, Baldessarini RJ, Hennen J, Floris G. Lithium maintenance treatment of depression and mania in bipolar I and bipolar II disorders. Am J Psychiatry 1998;155:638e45. [33] Tondo L, Baldessarini RJ, Hennen J, Minnai GP, Salis P, Scamonatti L, et al. Suicide attempts in major affective disorder patients with comorbid substance use disorders. J Clin Psychiatry 1999;60(Suppl. 2):63e9 [discussion 75e6, 113e6].
177
[34] Tondo L, Isacsson G, Baldessarini R. Suicidal behaviour in bipolar disorder: risk and prevention. CNS Drugs 2003;17:491e511. [35] Vieta E, Pacchiarotti I, Scott J, Sanchez-Moreno J, Di Marzo S, Colom F. Evidence-based research on the efficacy of psychological interventions in bipolar disorders: a critical review. Curr Psychiatry Rep 2005;7:449e55. [36] Winokur G, Coryell W, Keller M, Endicott J, Akiskal H. A prospective follow-up of patients with bipolar and primary unipolar affective disorder. Arch Gen Psychiatry 1993;50:457e65.