Comorbid bipolar disorder (BD) and substance use disorder (SUD) have been found to be highly prevalent in both epidemiological and clinical studies, with rates of SUD in subjects with BD ranging from 35-60% [1–6]. The high prevalence is found across different age groups and also in first episode BD samples [7, 8].
So far, most studies in BD have investigated only substance use fulfilling SUD criteria. Investigating a broader range of substance use in BD could be relevant because people with severe mental disorders are more likely to experience negative consequences from using relatively small amounts of psychoactive substances . Moderate alcohol consumption in BD is associated with more severe manic symptoms compared to abstinence, and to poorer social and familial adjustment and increased health-care use . To the best of our knowledge, only one study assessed substance use in BD more globally, reporting that 46% had SUDs and 8% had SUD-subthreshold substance use. In addition, the authors indicated that another substantial proportion used illicit substances occasionally .
Clarifying whether there is an increased use of substances in BD may increase our understanding of the psychopathology underlying the increased risk of abuse or dependence. Although most studies show a large prevalence of BD and SUD comorbidity, the rates vary widely. This variation could be mirroring differences in substance use in the general population where the BD sample is recruited. In a smaller sample from an earlier part of our ongoing study, we showed elevated rates of lifetime use of illicit substances among patients with psychotic disorders (including BD) compared to the general population , and differences in patterns of substance use between schizophrenia and BD . Due to the small number of patients with BD included in our earlier report, a separate comparison of BD patients with the general population sample was not implementable. Thus, there is a need for studies comparing BD subjects with reference populations on substance use and they should be done with samples from the same geographical area within the same time period.
In the current literature, BD with comorbid SUD is consistently referred to as associated with a poorer disease course and with reduced functioning compared to BD without SUD. The findings regarding the effects of SUD on BD are however divergent. To explore this more thoroughly we did a search in PubMed (terms bipolar disorder, substance abuse and outcome), and in addition tracked all cited references in key publications (Additional file 1). The main finding from this search was that the only consistently reported findings were delayed recovery and lower remission rates [14–22] as well as faster relapses [14, 23–25] in groups of BD patients with SUD (both lifetime/current substance - and/or alcohol use disorders) compared to BD without SUD. Furthermore, there appears to be extensive evidence for elevated suicidality rates in BD with SUD compared to BD without [18, 20, 26–37], although several studies also report no significant differences [19, 38–42]. Medication compliance rates are also relatively consistently reported to be lower in BD with SUD compared to BD without [18, 19, 29, 43–46] although a few studies report lack of differences [38, 42]. Another consistent finding is that the prevalence of psychotic symptoms does not appear to be elevated among BD patients with SUD compared to patients without [18, 19, 28, 38, 47, 48], and there is neither a tendency towards increased numbers of affective episodes [19, 27, 31, 48, 49].
The findings are more divergent regarding rapid cycling; as some studies did [38, 40, 50–52] and some did not [19, 29, 53] find this to be more prevalent in the SUD patients. The same inconsistency is found for the prevalence of mixed episodes, some studies found this phenomenon to be more common [14, 18, 39, 50, 54] while others did not [17, 47, 55] in the SUD patients. There are also inconsistencies regarding age of onset for BD; here some report earlier onset for patients with SUD [26, 29–31, 50, 51, 56, 57] while others do not find any differences compared to BD patients without SUD [18, 19, 38, 47, 55, 58]. Studies also diverge as to whether affective symptoms are of increased severity in BD patients with SUD compared to BD patients without [18, 21, 26, 39, 42, 47, 49, 50, 59, 60]. Furthermore, the number of hospitalizations or days in hospital is found to be elevated in BD patients with SUD in some studies [29, 31, 50, 55, 61–64] as opposed to in others [18, 26–28, 38, 48, 56, 65].
Findings concerning other functional variables such as decreased global functioning [19, 26, 38, 39, 47, 48, 56, 60, 66], social functioning [20, 21, 27, 29, 38, 58, 60, 67], educational level [19, 20, 26, 31, 38, 50, 56, 60], and quality of life [20, 21, 26, 58, 60, 61] in BD with SUD also diverge. Finally, some studies find lower employment status in BD with SUD compared to BD without [21, 24, 29, 67] while others do not [28, 43, 50, 56], and two studies even find better employment rates in BD with SUD [19, 61]. The current evidence therefore suggests that BD with comorbid SUD is clearly associated with worsening of some clinical and functional characteristics: Length of affective episodes and relapse rates, risks of suicidality and compliance to medication. However, substance abuse does not appear to be as consistently associated with a more severe course and outcome as frequently indicated in the literature.
In the present study, we aim at investigating differences in relevant outcome variables in a sample of BD patients with and without substance use. The present paper is based on a cross-sectional study of consecutively referred patients with BD from a catchment-area based psychiatric service, and a population survey of the use of illicit substances in the same area within the same time period. Our aims were to answer the following questions:
1) Is the rate of lifetime use of illicit substances higher in the patient sample than in the reference population?
2) Do patients with and without excessive substance use, defined as SUD and/or excessive use, differ on clinical and functional characteristics, in terms of disease course variables, current symptom levels and functioning?