Search

The association of extrapyramidal symptoms (EPS) with the use of antipsychotics was first discovered in the 1950s. To our knowledge, little research has been conducted with any retrospective observational database to evaluate the comparative risk of EPS between first-generation antipsychotic (FGA) and second-generation antipsychotic (SGA) users in the U.S. population. The purpose of the study was to compare EPS risk between FGA and SGA users using propensity score-matching (PSM) and instrumental variable (IV) analyses. A retrospective cohort design with an intention-to-treat (ITT) analysis (where the patients included in the cohort were assumed to take the medications without switching or dropping out) was chosen to examine the relationship between antipsychotic treatment and EPS risk. First-time antipsychotic users (as identified in the MEPS database), during the time frame 2002 to 2009, were included in the cohort. All subjects included in the cohort were followed over time to assess EPS risk. Propensity score-based logistic regression (using the Greedy 5[right arrow]1 digit match technique) was used to compare EPS risk between the SGA and FGA users after adjusting for demographic variables and risk factors (associated with EPS). In order to identify the presence of unobserved confounding, an instrumental variable analysis was attempted. Based on previous research findings, "delay in obtaining prescribed medicines" was selected as the instrument to evaluate the relationship between EPS risk and antipsychotic type. The feasibility of instrumental variable analyses was examined by evaluating the strength of the chosen instrument. The propensity score-based logistic regression analysis showed no difference in EPS risk [OR = 1.77, 95% CI = (0.49, 6.40)] between FGA and SGA users. The strength (partial r² = 0.0002) of the instrument was low. A weak instrument used in a regression model may produce biased estimates while evaluating treatment/outcome relationship; therefore, instrumental variable analysis was not conducted. EPS risk was found not to differ between the FGA and SGA users. However, clinicans may choose to evaluate other side-effects (in addition to EPS) of antipsychotics while making treatment decisions. Evaluation of economic, clinical and humanistic outcomes associated with treatment of antipsychotic-related side-effects (besides EPS) can provide clinicans the rationale for selecting one class of antipsychotics over the other. Interpretation of the study findings should be considered in light of the limitations of the MEPS database. Future research is necessary to identify a strong instrument to assess the presencec of unobserved confounding between antipsychotic exposure and EPS risk. Furthermore, additional research is warranted to assess differences in time to development of EPS between the FGA and SGA users.

The association of extrapyramidal symptoms (EPS) with the use of antipsychotics was first discovered in the 1950s. To our knowledge, little research has been conducted with any retrospective observational database to evaluate the comparative risk of EPS between first-generation antipsychotic (FGA) and second-generation antipsychotic (SGA) users in the U.S. population. The purpose of the study was to compare EPS risk between FGA and SGA users using propensity score-matching (PSM) and instrumental variable (IV) analyses. A retrospective cohort design with an intention-to-treat (ITT) analysis (where the patients included in the cohort were assumed to take the medications without switching or dropping out) was chosen to examine the relationship between antipsychotic treatment and EPS risk. First-time antipsychotic users (as identified in the MEPS database), during the time frame 2002 to 2009, were included in the cohort. All subjects included in the cohort were followed over time to assess EPS risk. Propensity score-based logistic regression (using the Greedy 5[right arrow]1 digit match technique) was used to compare EPS risk between the SGA and FGA users after adjusting for demographic variables and risk factors (associated with EPS). In order to identify the presence of unobserved confounding, an instrumental variable analysis was attempted. Based on previous research findings, "delay in obtaining prescribed medicines" was selected as the instrument to evaluate the relationship between EPS risk and antipsychotic type. The feasibility of instrumental variable analyses was examined by evaluating the strength of the chosen instrument. The propensity score-based logistic regression analysis showed no difference in EPS risk [OR = 1.77, 95% CI = (0.49, 6.40)] between FGA and SGA users. The strength (partial r² = 0.0002) of the instrument was low. A weak instrument used in a regression model may produce biased estimates while evaluating treatment/outcome relationship; therefore, instrumental variable analysis was not conducted. EPS risk was found not to differ between the FGA and SGA users. However, clinicans may choose to evaluate other side-effects (in addition to EPS) of antipsychotics while making treatment decisions. Evaluation of economic, clinical and humanistic outcomes associated with treatment of antipsychotic-related side-effects (besides EPS) can provide clinicans the rationale for selecting one class of antipsychotics over the other. Interpretation of the study findings should be considered in light of the limitations of the MEPS database. Future research is necessary to identify a strong instrument to assess the presencec of unobserved confounding between antipsychotic exposure and EPS risk. Furthermore, additional research is warranted to assess differences in time to development of EPS between the FGA and SGA users.