O39 VTE risk in the general population: applying context to the impact of JAK inhibition
Abstract Background Trial data indicate an increased risk of VTE with JAK inhibition. The EMA and FDA have given regulatory advice for their use as a result. To understand the context of any association, we need to understand VTE risk in the background population. Our objective was to describe the changing epidemiology of VTE risk including a breakdown of DVT and PE in the general population in England, over a 20-year period. Methods We utilised hospital episodes statistics data to study all patients presenting for hospital care in England between 1 January 1998 and 31 December 2017. We identified VTE events using ICD10 codes I26.0, I26.9, I80.1, I80.2 and I80.3. To estimate annualised event rates, we used the number of finished consultant episodes for admitted care where the primary diagnosis was VTE for the numerator; the denominator was the whole population estimate, from the office for National Statistics, for the relevant year for England. Rates were calculated per 100,000 population, and the Cochrane Armitage test was used to evaluate statistical significance of trends over time. Linear regression allowed modelling of estimates over time, with model assumptions tested using residual versus fitted plots. To explore non-linearity, we used a cubic spline model. Results The average VTE rate across the 20 years was 127/100,000. VTE rates have increased over time, climbing from 108/100,000 in 1998, to 151/100,000 in 2017. The relative frequency of DVTs and PEs has changed over this time: in 1998 DVT was more frequent than PE (rates 68/100,000 and 40/100,000 respectively). By 2018 this ratio had reversed (DVT 52/100,000: PE 98/100,000). This change was statistically significant (p < 0.0001). DVT rates declined in a linear manner over time. PE rates increased with a non-linear pattern, with a sharp rise apparent between 2008 and 2010. Conclusion The inversion in the frequency of DVTs and PEs is consistent with previous findings using European data. One possible explanation is changing diagnostics (CTPA replacing V/Q), capturing previously subclinical PE events. Limitations include our ecologic design as well as lack of linkage to mortality records. The presumed VTE risk associated with JAK inhibition is small. It is plausible that future shifts in population wide patterns of VTE could obscure or amplify any effect attributable to JAK inhibitors. As longer-term observational data become available, it is crucial they are interpreted in the context of background population trends. Disclosures S. Baroncini: None. M. Yates: None. K. Bechman: None. S. Patel: None. A. Rutherford: None. A. Kleymann: None. J. Galloway: None.