Session Title: Epidemiology and Health Services II & III
Session Type: Abstract Submissions (ACR)
Background/Purpose: Oral glucocorticoid (GC) therapy is used in more than half of patients with rheumatoid arthritis (RA). Cataracts are a recognised treatment-related side effect of GC use. However, the magnitude of the risk, its relationship with dose, or whether screening for cataracts should be undertaken in patients on GC therapy is not clear. The aim of this study was to determine the incidence of cataracts in the active and placebo arms of RCTs of systemic GC therapy in patients with RA. Because of early concerns about possible under-reporting in the trials, a second aim was to compare the observed incidence in the RCTs to the incidence of cataracts in the general population.
Methods: : Twenty two RCTs of systemic GC therapy in patients with RA were reassessed to identify incident cataracts in the placebo and treatment arms. The number of person years at risk (pyr) and the resultant incidence rates/ 1000pyr were calculated for the combined RCT population, then for the GC and control groups separately. The incidence rate ratio for GC use was calculated using Poisson regression. Two large observational studies (The Beaver Dam Eye Study and The Blue Mountains Eye Study) reporting the incidence of cataract in the general population were identified. Both studies reported the age and gender-specific cumulative incidence of cataracts over 5 years and 10 years. The mean of the incidence rates across the two studies was used to generate a general population incidence rate for the age-specific strata. Indirect standardisation was used to calculate the expected number of cases in the RCTs, were the RCT patients to have had the same incidence rate as the general population. A standardised incidence ratio was then calculated using the number of observed and expected cases.
Results: In the 22 RCTs, there were 2250 patients with a mean age of 56 years (weighted by duration of follow up), contributing 3764pyr. There were 13 cases of cataract reported in 3 of the 22 RCTs, giving an overall incidence rate of 3.5 per 1000pyr. The RCTs contributed 1884pyr to the GC group and 1880pyr to the control group. There were 7 cases of cataract reported in the GC group and 6 in the control group, generating incidence rates of 3.7 and 3.2 / 1000pyr in the GC and control group, respectively. The incidence rate ratio for GC use was 1.16 (95%CI 0.39 – 3.46). Using the general population incidence for patients aged 55-64 years of 46.7 per 1000pyr, the expected number of cataracts was 176. This generates a standardised incidence ratio of 0.07 (95%CI 0.03 – 0.21).
Conclusion: The incidence of cataracts was 16% higher in GC-treated compared to control patients with RA using all available RCTs. However the lack of precision (wide confidence intervals) due to small numbers means that this calculation is inconclusive. The observed number of cataracts was around 15 times lower than expected in both arms, suggesting huge under-reporting. Such under-reporting results in low confidence of any measured risk from these studies emphasising the need for well designed observational studies and better reporting of steroid-side effects in future RCTS.
R. J. Black,
Supported jointly by an educational research grant by Australian Rheumatology Association and Roche Products Pty Limited.,
W. G. Dixon,
« Back to 2013 ACR/ARHP Annual Meeting
ACR Meeting Abstracts - https://acrabstracts.org/abstract/under-reporting-of-cataracts-in-randomised-controlled-trials-investigating-the-use-of-systemic-glucocorticoids-in-patients-with-rheumatoid-arthritis/