Elsevier

Journal of Affective Disorders

Volume 57, Issues 1–3, January–March 2000, Pages 139-145
Journal of Affective Disorders

Research report
Predictors of recurrence in affective disorder — analyses accounting for individual heterogeneity

https://doi.org/10.1016/S0165-0327(99)00080-4Get rights and content

Abstract

Background: Previous studies suggest that gender, age at onset, and marital status act as risk factors for further recurrence initially during the course of affective disorder but not at a later stage. These studies did, however, not take the individual liability to recurrence into account. Method: The effect of predictors of recurrence was estimated with the use of generalised linear mixed models in a case register study including a random sample of all patients admitted with primary affective disorder in Denmark during 1971–1993. Results: In total, 7047 first admission patients with a diagnosis of affective disorder, depressive or manic/circular type were included in the analyses. The study confirmed that the effect of the type of disorder, age at first admission, and never being married decreased during the course of illness even when the individual liability to recurrence was taken into account. No differences in the effect of gender and in the effect of a recent divorce were found between early and later episodes and the effect of a recent death of a spouse seemed to increase during the course of illness. The risk of recurrence increased with every new episode for all sub-groups of patients. Conclusion: The effect of some, but not all, predictors of recurrence decline during the course of affective illness. The number of previous episodes predicts recurrence in most subgroups of patients. Limitation: The data relate to re-admissions rather than recurrence. Clinical relevance: The study underscores the importance of the illness process itself.

Introduction

Episodes in unipolar (recurrent depressive disorder) and bipolar affective disorders may develop in the interplay between biological, psychological and social factors. Psychosocial stress and major life events may precipitate affective episodes in susceptible individuals. Some evidence suggests that the individual vulnerability increases with the number of affective episodes as more psychosocial stressors have been found to be involved in the first episode than in subsequent episodes of major affective disorder (review by Goodwin and Jamison, 1990, pp. 142–146; Post, 1992). However, this relation between the prevalence of precipitating events and the course of episodes has not been found in all studies of unipolar disorder (Perris, 1984, Bidzinska, 1984) or bipolar disorder (Bidzinska, 1984, Ellicott et al., 1990, Hammen and Gitlin, 1997). Furthermore, it has been objected that, in the primary retrospective studies of precipitating events, patients are more likely to search for stressors prior to the onset of the illness, whereas once the illness is accepted less attention is paid to stressors (Goodwin and Jamison, 1990, p. 143; Johnson and Roberts, 1995).

Studies with prospective rigorous assessment of stressors are few and include a short follow-up period only (up to 3 years, Ellicott et al., 1990, Pardoen et al., 1996, Paykel et al., 1996, Monroe et al., 1996, Hammen and Gitlin, 1997). Thus, no study encompasses appropriately long follow-up time to compare the presence of prospectively assessed life events before early episodes and before later episodes, as pointed out by Hammen and Gitlin (1997).

In this regard, register data offer some remarkable advantages. They often include a large number of patients followed for a long period of time and recall bias is completely avoided since data are ‘hard facts’ collected routinely and independently of researchers.

We have previously used data from the the nationwide Danish Psychiatric Central Register to identify predictors of recurrence at successive episodes in affective disorder (Kessing et al., 1998, Kessing, 1998). In these studies, the risk of recurrence was expressed as the risk of readmission after successive discharges in affective disorder. Initially in the course of unipolar and bipolar illness, sociodemographic variables such as sex, age (partly) and marital status were found to predict recurrence whereas these variables had no effect later in the course of the illnesses. Patients who had never been married had greater risk of recurrence after the first and the second episode but not after subsequent episodes, and divorced/separated patients had greater risk of recurrence after the first episode but not after subsequent episodes.

It is, however, a major statistical problem in every study of the effect of psychosocial stressors during the course of affective illness that it is difficult to assess whether the difference in effect between early and late episodes is statistically significant. Analyses at successive episodes are not independent of each other as patients with many episodes will be a subgroup of patients with few episodes, and hence patients with many episodes and great liability to recurrence will tend to dominate the analyses.

However, it is possible to account for this individual liability to recurrence with the use of either frailty models (Andersen et al., 1993) or generalised linear mixed models (Breslow and Clayton, 1993; see also Kessing et al., 1999). The present study used models of the latter type to reanalyse the effect of age, gender and marital status on the risk of recurrence during the course of illness with the use of data from the Danish Psychiatric Case Register.

Section snippets

The register

In Denmark, all psychiatric admissions have been registered in a nationwide register (Dupont, 1983; Munk-Jørgensen and Mortensen, 1997) with the same diagnostic system, the International Classification of Diseases, 8th Revision (National Board of Health, Denmark, 1971) from 1 April 1969 to 31 December 1993. For various reasons, and to achieve better diagnostic reliability over time, it was decided not to change to ICD, 9th revision in 1978. Since there are no private psychiatric hospitals or

Results

From 1971 to 1993, 17 447 patients got a diagnosis of depression (code 296.09 and 296.29) and 2903 patients got a diagnosis of mania or circular episode (code 296.19 and 296.39) at first admission. The entire sample was, however, not included in analyses due to limitations in computer memory. Thus, a random sample of 7047 patients (4558 women and 2489 men) among the 20 350 patients was selected using the random number generator in sas, version 6.12.

Table 1 presents the effect of predictors for

Discussion

To our knowledge this study and the studies previously reported (Kessing et al., 1998, Kessing, 1998) are the first studies to present within-patient analyses in which the effect of sociodemographic variables and life events for the individual is studied at early episodes and at later episodes. Long-term prospective studies of the impact of psychosocial stressors on the course of affective illness may be difficult to perform in a valid way since they may be hampered by selective drop-out.

Acknowledgements

This work was supported by The Foundation for Psychiatric Research (Denmark).

References (23)

  • L.V. Kessing et al.

    Predictors of recurrence in affective disorder — a case register study

    J. Affect. Disord.

    (1998)
  • H. Perris

    Life events and depression. Part 2. Results in diagnostic subgroups, and in relation to the recurrence of depression

    J. Affect. Disord.

    (1984)
  • P.K. Andersen et al.

    Statistical Models Based on Counting Processes

    (1993)
  • E.J. Bidzinska

    Stress factors in affective diseases

    Br. J. Psychiatry

    (1984)
  • N.E. Breslow et al.

    Approximate inference in generalized linear models

    J. Am. Stat. Assoc.

    (1993)
  • G.W. Brown et al.

    Life events and endogenous depression. A puzzle reexamined

    Arch. Gen. Psychiatry

    (1994)
  • A. Dupont

    A national psychiatric case register as a tool for mental health planning, research, and administration. The Danish model

  • A. Ellicott et al.

    Life events and the course of bipolar disorder

    Am. J. Psychiatry

    (1990)
  • F.K. Goodwin et al.

    Manic-Depressive Illness

    (1990)
  • C. Hammen et al.

    Stress reactivity in bipolar patients and its relation to prior history of disorder

    Am. J. Psychiatry

    (1997)
  • S.L. Johnson et al.

    Life events and bipolar disorder: Implications from biological theories

    Psychol. Bull.

    (1995)
  • Cited by (30)

    • Neuropathological and neuromorphometric abnormalities in bipolar disorder: View from the medial prefrontal cortical network

      2014, Neuroscience and Biobehavioral Reviews
      Citation Excerpt :

      Nevertheless, if progressive damage to the mPFC (and other brain regions) is indeed occurring over time then this putative pathological process may be responsible for the progressively deteriorating clinical course, worsening of cognitive function, and comorbidity observed in some BD patients. For instance, treatment resistance and shortening of the inter-episode interval have been reported over time in parallel with an increase in the number of discrete episodes of depression and/or mania (Berk et al., 2011; Kessing et al., 2000) leading to attempts to define several clinical stages of BD based on neuroimaging, cognitive, and molecular biomarkers (Berk et al., 2007; Kapczinski et al., 2009; Lin et al., 2013; McNamara et al., 2010). These data raise the possibility that early diagnosis and treatment of both BD and co-occurring medical illnesses may help to reduce the morbidity associated with BD.

    • Clinical differences between first and recurrent episodes in depressive patients

      2011, Comprehensive Psychiatry
      Citation Excerpt :

      Despite of women being at increased risk of developing depression and relapse [22], female sex is not a significant risk factor for recurrence and the profile of symptoms is similar in both sexes according to different studies [23,24]. The only exception is the inpatient study of Kessing et al [25], with women at high risk of recurrence. Our results related to sex are similar to the results of the later study: there was a significant difference in the number of women in the recurrent group experiencing depression than the first episode group.

    • Risk for recurrence in depression

      2007, Clinical Psychology Review
    • Clinical predictors of response to lamotrigine and gabapentin monotherapy in refractory affective disorders

      2002, Biological Psychiatry
      Citation Excerpt :

      Our study would suggest the potential importance of studying gabapentin in younger individuals as well as in less treatment-refractory populations. Others have shown that more lifetime episodes are associated with increased rate of recurrence and progressive illness despite treatment (Kessing et al 2000). Should the finding that age is inversely correlated with responsiveness to gabapentin be confirmed in subsequent studies, such differential response could be linked to the well-described developmental changes in GABAergic biochemistry and function in the central nervous system (Follesa et al 1999; Gravielle et al 1998; Pouzat and Marty 1999; Stefanova et al 1998).

    View all citing articles on Scopus

    This study was conducted in cooperation with the Department of Psychiatric Demography, University of Aarhus, Psychiatric Hospital, Risskov, Denmark.

    View full text