Attributable risk of co-morbid substance use disorder in poor Observance to pharmacological treatment and the occurrence of relapse in schizophrenia

Background. - There are numerous risk factors involved in poor (incomplete) compliance to pharmacological treatment, and the associated relapse risk, for patients with schizophrenia. Comorbid substance use disorders are considered as among the most important ones, although how much their presence increase the risk of poorer observance (and higher risk of relapse) has not been yet assessed. This measure would be important, especially if the published literature on the topic provides sufficient material to perform a meta-analysis and to assess different potential biases such as those related to time (new studies are easier to publish when positive) or sample size (small samples might drive the global positive conclusion). Method. - A PubMed(R) search was made, screening the following terms between 1996 and august 2014 "Addiction AND (Observance OR Adherence) AND schizophrenia AND (French OR English [Language])" and "(Substance Abuse OR substance dependance) AND Outcome AND schizophrenia AND (French OR English [Language])". Studies were included if they describe two patients groups (schizophrenia with and without present substance use disorder) and assess the studied outcome. MetaWin(R) version 2 was used for the meta-analysis, while publication time bias relied on non-parametric correlation and the one linked to sample size was assessed through normal quantile plots. An attributable risk was also computered, on the basis of the odds-ratio derived from the meta-analysis and the prevalence of the analyzed trait (associated substance use disorder). Results. - Eight studies could be included in the meta-analysis, showing that the presence of a substance use disorder significantly increases the risk of poor observance to pharmacological treatment (OR = 2.18 [1.84-2.58]), no significant bias being detected, either linked to time (rho = 0.287, P = 0.490) or sample size (Kendall's Tau = -0.286, P = 0.322). The related attributable risk is 18.50%. Only three studies could be used for the meta-analysis of the risk of relapse associated with the presence of substance use disorders. The corresponding odds-ratio is 1.52 [1.19-1.94], and the attributable risk is 31.20%, but the search for biases could not be performed because of the small number of studies. Conclusions. - These results shed light on the importance of comorbid substance use disorder to explain the poor observance frequently observed in patients with schizophrenia. Indeed, having an associated substance use disorder double the risk of poor compliance to pharmacological treatment, this comorbidity explaining a fifth of all factors involved. Although the number of available studies does not allow definite conclusions, the meta-analysis of prospective studies focusing this time of the risk of relapse requiring hospitalization is also in favor of a significant role of associated substance use disorder. These results argue in favor of developing specific strategies to better treat patients with dual diagnoses, i.e. schizophrenia and substance use disorder. (C) L'Encephale, Paris, 2015.