In the last decade, the number of patients having a lower joint replacement (LJR), i.e., hip and knee, either for the first time (primary replacement) or a corrective procedure after the first surgery (revision surgery), increased substantially. Previous studies showed that non-steroidal anti-inflammatory drugs (NSAIDs) might increase the risk of broken bone, non-fusion broken bone, and joint replacement. In contrast, other studies concluded that NSAIDs did not affect the contact between bone and artificial materials implanted into joints, and the movement of these materials after hip replacement (HR). Only a few studies investigated the association between NSAIDs and the risk of revision surgery for specific joints namely HR and knee replacement (KR). Thus, we aim to investigate the risk of revision surgery for LJR, i.e., HR and KR for NSAID users and to evaluate whether the duration of NSAID use, different sub-groups of NSAID (either conventional NSAIDs or selective COX-2 inhibitors), and joint replaced (either knee or hip) lead to different risk of revision surgery. Our findings could be a consideration for physicians to prescribe NSAIDs for patients with primary joint replacement.
We aim to evaluate the association between NSAIDs and the risk of revision of LJR. We will conduct a retrospective cohort study in patients who underwent primary LJR, i.e., HR and KR surgery using the CPRD data. We will limit our data to most likely an elective primary surgery. The exposure for this study will be either current, recent, or past use of NSAIDs. Nonusers will be defined as patients without any prescription of NSAIDs for the entire follow-up time. For NSAID users, they might also be defined as non-users if a total duration of either single or consecutive prescription of any NSAIDs were <3 weeks, or if they have no NSAID prescription following past use, or during a 3-week period following primary LJR. We will stratify our analyses based on the sub-groups of NSAIDs, the duration of current NSAID use, and joint replaced. The outcome will be the revision of LJR. We will consider age, sex, body mass index (BMI), co-medications, co-morbidities, and lifestyle factors as potential confounders. The risk of revision of LJR for NSAID use will be estimated by using Cox proportional hazard with time-dependent covariates. All potential confounders will be taken into account to estimate the adjusted hazard ratios.
The outcome will be the revision of LJR, i.e., the revision of primary HR or KR which will be based on the CPRD Read Codes (Annex 1).
Olaf Klungel - Chief Investigator - Utrecht University
Patrick Souverein - Corresponding Applicant - Utrecht University
Anthonius de Boer - Collaborator - Utrecht University
Arief Lalmohamed - Collaborator - Utrecht University
Mohammad Bakhriansyah - Collaborator - Utrecht University
Arief Lalmohamed - Collaborator - Utrecht University