Background/Purpose: In several developed countries it has been observed that the incidence of orthopaedic surgery among rheumatoid arthritis (RA) patients has been in decline, despite a concurrent increase among non-RA patients. Whilst the introduction of biologic therapies may be an explanatory factor for the divergent trends, specific data testing this hypothesis is lacking. We set out to estimate the impact of approval and national guidance (National Institute for Health and Care Excellence (NICE) TA36) on anti-tumour necrosis factor alpha (TNF-α) therapy use for RA on rates of total hip (THR) and knee replacement (TKR) among incident RA patients.

Methods: Primary care data (Clinical Practice Research Datalink (CPRD)) for the study period (1995-2014) were used to identify incident adult RA patients. The 5-year cumulative age and sex standardised incidence of THR and TKR was calculated for RA cohorts diagnosed in each six-months of the period 1^st April 1995 to 30^th September 2009. Interrupted time series analysis was used to estimate changes in level and trend following the publication of NICE TA36 in March 2002. In main analyses a 1-year time lag was used to allow for the delay associated with patients failing conventional DMARD therapy before initiating biologics. In sensitivity analyses a 2-year lag was used.

Results: We identified 17,505 incident RA patients. The number of THR and TKR events occurring within 5-years of diagnosis were 465 and 650, respectively. Throughout the pre-NICE TA36 period the cumulative incidence of THR remained stable at 2.48%. Following publication of guidance there was an immediate level increase (1.71%, P=0.001) but subsequent slope decrease (-0.21% per six months, P=0.001), equating to no significant difference in the mean absolute change in THR compared to that expected given prior level and trend only (using the mid-point of the post-NICE TA36 period). Conversely, the cumulative incidence of TKR was 2.61% at the beginning of the study period and increased by 0.13% per six months (P=0.005) throughout the pre-NICE TA36 period (figure 1). Publication of guidance was associated with a significant downward slope change (-0.29% per six months, P=0.002), equating to a mean absolute change in TKR of -1.90% (95% C.I. -2.98 to -0.82) compared to values expected given prior level and trend only (figure 1). This represented a relative reduction of 34%. In sensitivity analyses using a 2-year lag, the cumulative incidence of THR was flat for the whole study period whilst results for TKR remained unchanged from the main analysis.

Conclusion: Among incident RA patients in England and Wales, approval of anti-TNF-α therapies and related NICE guidance was associated with reduced 5-year rates of TKR but not THR. Future work to further elucidate the impact of biologic therapies on the need for joint replacement surgery among RA patients is required.