Patient controlled analgesia (PCA) is commonly used after TKR. Prolonged use of PCA may however have a negative impact on patients delaying their rehabilitation and therefore discharge. We aimed to evaluate the effect of the duration of PCA on the hospital length of stay (LOS) in patients who undergo TKR. We reviewed the casenotes of all patients who underwent a primary TKR in two South Australian teaching hospitals between 2006 and 2007. After excluding patients whose LOS was determined by placement issues and patients who developed intra-hospital post-operative complications, a number of 345 patients were included in this study. Data collected included: age, gender, ASA grade, regional blocks used, duration of postoperative PCA (< 24 hours, 24-48 hours, > 48 hours) and hospital LOS. Using SAS Version 9.2 statistical analysis software the data was analysed using univariate and multivariate Poisson regression models. Risk ratios, confidence intervals and P values were calculated.

Univariate regression models showed that there was a significant difference in length of stay between the three PCA groups (p < 0.0001). Post hoc tests revealed that the length of stay was longer in the 24 to 48 hours and > 48 hours groups compared to the < 24 hours group (p < 0.0001). There was also a significant difference in hospital LOS between males and females (p = 0.0049) with females expected to stay on average 9.7% longer (risk ratio = 1.097, 95% CI 1.028, 1.169). Patients in the ASA categories (1 and 2) recorded shorter lengths of stay than patients in the ASA categories (3 and 4) (p < 0.0001). Also patients treated at one hospital had longer LOS than the patients treated at the other hospital (risk ratio = 1.122, p = 0.0001). There was no evidence for a relationship between the patients' ages, and use or type of regional block used and the hospital LOS. Results from the multivariate regression models showed that each of the four variables found to influence LOS significantly, did so independent of the other variables. Therefore, duration of PCA, gender, ASA and hospital were all independent predictors of hospital LOS after primary TKR.

Longer administration of PCA, higher ASA grades and female gender are associated with longer hospital LOS after TKR. Reducing postoperative PCA, as well as improving ASA grading, could reduce LOS in these patients. A multimodal pain management strategy that shortens PCA use could reduce hospital LOS and costs after TKR.