Is the use of antidepressants associated with patient-reported outcomes following total hip replacement surgery?

Abstract

Background and purpose - Patients with anxiety and/or depression tend to report less pain reduction and less satisfaction with surgical treatment. We hypothesized that the use of antidepressants would be correlated to patient-reported outcomes (PROs) 1 year after total hip replacement (THR), where increased dosage or discontinuation would be associated with worse outcomes. Patients and methods - THR cases with pre- and postoperative patient-reported outcome measures (PROMs) were selected from the Swedish Hip Arthroplasty Register (n = 9,092; women: n = 5,106). The PROMs were EQ-5D, visual analog scale (VAS) for pain, Charnley class, and VAS for satisfaction after surgery. These cases were merged with a national database of prescription purchases to determine the prevalence of antidepressant purchases. Regression analyses were performed where PROs were dependent variables and sex, age, Charnley class, preoperative pain, preoperative health-related quality of life (HRQoL), patient-reported anxiety/depression, and antidepressant use were independent variables. Results - Antidepressants were used by 10% of the cases (n = 943). Patients using antidepressants had poorer HRQoL and higher levels of pain before and after surgery and they experienced less satisfaction. Preoperative antidepressant use was independently associated with PROs 1 year after THR regardless of patient-reported anxiety/depression. Interpretation - Antidepressant usage before surgery was associated with reduced PROs after THR. Cases at risk of poorer outcomes may be identified through review of the patient's medical record. Clinicians are encouraged to screen for antidepressant use preoperatively, because their use may be associated with PROs after THR.

Figure 1.

Flow chart showing the selection of patients from the Swedish Hip Arthroplasty Register based on the inclusion criteria of the study. ^aThe Swedish Hip Arthroplasty Register PROM program began in 2002 at 11 hospitals. Participation gradually increased until 2008 when it was active nationwide.

Figure 2.

Linear regression results of the independent categorical variables including the dichotomous antidepressant variable, where the points represent the slope coefficient with the 95% confidence interval (CI) for the dependent EQ-5D index variable. EQ-5D indices can range from –0.594 to 1.0. Any variable without a CI was the reference variable and any CI that did not include 0 represents a significant influence on the EQ-5D index. Preoperative EQ-5D index, EQ VAS, pain VAS, and age were the influential continuous variables on postoperative EQ-5D indices as indicated in Table 3.

Figure 3.

Linear regression results of the independent categorical variables including the dichotomous antidepressant variable, where the points represent the slope coefficient with the 95% confidence interval (CI) for the dependent EQ VAS variable. EQ VAS values can range from 0 to 100. Any variable without a CI was the reference variable and any CI that did not include 0 represents a significant influence on the EQ VAS. Preoperative EQ-5D index and EQ VAS scores and age were the influential continuous variables on postoperative EQ VAS scores as indicated in Table 3.

Figure 4.

Linear regression results of the independent categorical variables including the dichotomous antidepressant variable where the points represent the slope coefficient with the 95% confidence interval (CI) for the dependent pain VAS variable. Pain VAS values can range from 0 to 100. Any variable without a CI was the reference variable and any CI that did not include 0 represents a significant influence on the pain VAS. Preoperative EQ VAS and pain VAS scores and age were the influential continuous variables on postoperative pain VAS scores as indicated in Table 3.

Figure 5.

Linear regression results of the independent categorical variables including the dichotomous antidepressant variable where the points represent the slope coefficient with the 95% confidence interval (CI) for the dependent satisfaction VAS variable. Satisfaction VAS values can range from 0 to 100. Any variable without a CI was the reference variable and any CI that did not include 0 represents a significant influence on the satisfaction VAS. Preoperative EQ VAS and age were the influential continuous variables on satisfaction VAS scores as indicated in Table 3.