Prednisone Use and the Risk of Cataract Formation in Persons with Rheumatoid Arthritis
Liron
1Washington
1, Caplan
Kaleb
2, Michaud
Frederick
2 Wolfe
Results (continued)
To addess the possibility that previously-identified cataract risk factors were not uniformly distributed, multivariate Cox proportional hazard modeling was performed. The relative strength of association between prior prednisone use and cataract formation (HR 1.66 [CI 1.25-2.20] was greater than that of patients’ age (HR 1.08 [CI 1.07-1.09]), educational level (HR 1.07 [CI 1.02,1.14]), and total income, but not as strong as the association of gender (HR 2.02 [CI 1.48-2.75]). Interestingly, history of smoking, duration of rheumatoid arthritis, and diabetes were not significant risk factors (Figure 3).
University, St. Louis, MO, USA; 2National Data Bank for Rheumatic Diseases, Wichita, KS, USA
Methods (continued) Results
The basic demographic information for the entire study cohort is represented in Table 1. The initial analyses included all patients, regardless of whether prednisone exposure occurred before or after enrollment in the longitudinal study. For this combined population, the incidence density of cataracts was slightly higher among women (30.1/100 person-yr [CI 29.4-30.9]) than men (27.7 person-yr [CI 26.429.0]) and there were age-specific differences between patients on prednisone and those never exposed to prednisone (Table 2 and Figure 1). This finding was also reflected in the odds of forming cataracts among patients receiving non-continuous prednisone (OR 1.3 [CI 1.2-1.5]) and continuous prednisone (OR 1.5 [CI 1.4-1.7]), after controlling for age, sex, and educational level. The population attributable fraction of cataracts associated with prednisone use amounted to 14% [CI 11-17%]. These cataracts resulted in a 3.7% decline in Quality Of Life (QOL), as measured by a 100 point Visual Analog Scale (VAS). Covariates accounted for in this regression analysis include: age, gender, Heath Assessment Questionnaire (HAQ) disability index, duration of RA, educational level, co morbidity score, and presence of diabetes. Among the entire cohort of RA patients, 25.1% acknowledge having difficulty in seeing/recognizing people and/or reading a newspaper. The odds of having these visual difficulties was more than doubled in RA patients with cataracts (OR 2.2 [CI 1.8-2.8]. Similarly, the odds of complaining of blurry vision was doubled in RA patients with cataracts (OR 2.0 [CI 1.7-2.3]), when compared to those without cataracts. In a subset of 1,013 RA patients with cataracts, 3.7% of patients described a limitation in their activities due to cataracts. To reduce the potential for recall bias, we restricted further analysis to those patients who were steroid-naïve at the time of enrollment in the prospective longitudinal study. Among this cohort of 3,794 patients, the risk of cataracts was calculated using multivariate regression. Patients with any prednisone exposure during the observation period were again more than twice as likely to develop cataracts as rheumatoid arthritis patients without any prednisone exposure (OR 2.3 [CI 1.73.1]). When analyzed according to dosage categories, similar odds ratios were present at all doses, even below 5mg/day (Figure 2). Duration of prednisone exposure also represented a significant risk factor, with each year of prednisone usage elevating the chance of developing cataracts by 26% (CI 11.4-44%). These regressions projected a population attributable fraction of 8.0% for prednisoneassociated cataracts. The associated deterioration in QOL was 2.1% on a VAS, controlling for age, gender, HAQ, and RA duration.
Abstract
Purpose/Objectives: Although prednisone is a commonly used therapy in rheumatoid arthritis (RA), adverse effects associated with prednisone have not been accurately determined. Prednisone use is a known risk factor for cataract formation in the general population. In this study, we utilize a large cohort of RA patients to determine the prevalence of cataracts and the contribution of prednisone to cataract development. We also estimate the effect of cataracts on healthcare costs and quality of life (QOL) in RA patients. Methods: 21,212 patients with RA completed semiannual questionnaires beginning in 1998. Cataracts were determined by self-report. To evaluate the effect of prednisone more accurately, patients with exposure to prednisone at the time of study enrollment were evaluated separately from those patients whose first exposure to prednisone occurred during the study. Logistic regression, adjusted for duration of study participation, was used to determine risk factors. Results: Over their lifetime, 69.5% of patients used prednisone. The fraction of cataracts attributed to prednisone (attributable fraction) was 14%, and the decrease in QOL was 3.7%. In the subset of 3794 patients who had never previously been on prednisone, the age and sex adjusted odds ratio for the effect of subsequent prednisone use on cataract was 2.3 (CI 1.7-3.1), and the attributable risk was 8.0%. The risk of developing cataracts increased by 26% (CI 11.4-44%) for each year on prednisone. A dose effect was not seen. Adjusted for age, sex and functional disability, QOL was 2.1% lower in those patients with cataracts. Prednisone use resulted in a hazard ratio of 1.7 (CI 1.2-2.2) for cataract formation by Cox proportional hazard modeling. Our results paralleled those of prior studies in identifying gender, age, educational level, and income as cataract risk factors. Conclusion: Prednisone use doubles the risk of cataract formation, and the population attributable fraction for prednisone is 8.0-14%. Cataracts occur at all doses of prednisone. These cataracts cause a small, but statistically significant reduction in QOL.
Demographic characteristics recorded include: age, gender, ethnicity, educational level attained, marital status, and total income. The available health information contain a thorough past medical history, extensive pharmaceutical data (including dosages and frequency of use), and smoking history. Cataracts were determined by participant self-report. Functional disability was assessed by the Heath Assessment Questionnaire (HAQ), and a comorbidity score documents the presence of 13 conditions, thereby allowing for estimates for the effect of comorbid illnesses. For the purpose of this report, QOL was measured by a 100mm VAS scale, as most QOL questionnaires do not specifically assess vision. Statistical analysis: The entire RA cohort was initially evaluated in order to clarify the relationship between prednisone use and the incidence of cataracts. A more rigorous standard was applied to subsequent analyses, which eliminated patients with any prednisone exposure prior to enrollment in the NDB. Univariate analysis was succeeded by multivariate regression analysis, which progressed in an additive, step-wise fashion. Exact confidence intervals in the age-specific analysis were computed based upon a Poisson-distribution. All confidence intervals were established at 95% and the standard for significance was α=0.05. Computations were performed using the Stata/SE statistical package, version 8.2 and all tests were two-tailed. Table 2: Incidence rates for cataracts among patients with rheumatoid arthritis
Category Cataract Exposure Incidence Frequency (years) Rate† 5549 18781.0 95% C.I.
Discussion
Prednisone use doubles the risk of cataract formation in patients with RA. In contrast to previous studies, this increased risk is present even at average daily doses below 5mg. While duration of prednisone treatment was associated with an increased risk of cataracts, there was only a subtle trend towards a dose-response effect, which did not reach statistical significance. The sample size may have been insufficiently powered to detect such a dose effect, as only 10% of prednisone-naïve patients began to use prednisone during follow-up. Our regression models corroborated some putative cataract risk factors (age, educational level, gender, income), while rejecting others (history of diabetes and history of smoking). The population attributable fraction of cataracts ascribed to prednisone use varies from 8.0-14%, depending on the method of statistical evaluation. These cataracts lead to small, but statistically significant, declines in QOL and patients’ perceived autonomy. Assuming only 50% of these cataracts require surgery (avg. cost $2500), direct costs of correcting cataracts due to prednisone are projected at $2.1-3.9 million/yr (or $2852 million for the entire current RA population in the U.S.).
All Patients Gender Female Male Age group (years) <20 20-29 30-39 40-49 50-59 60-69 70-79 80-89 >= 90
29.6 (28.9,30.2)
4344 1205
4351.0 14430.0
30.1 (29.4,30.9) 27.7 (26.4,29.0)
Methods
Source of the Data: The data collected for this analysis was assembled under the auspices of a longitudinal study examining outcomes in patients with rheumatoid arthritis. Known as the National Data Bank for Rheumatic Diseases (NDB), this observational study has collected semi-annual, detailed, 28-page questionnaires from patients treated by rheumatologists throughout the United States. In the current investigation, 21,212 patients with RA submitted extensive health information over a period of up to 6 years, ending in early 2004. All patients consented to participate in accordance with the standards of the appropriate ethics committees. Table 1: Demographic and clinical features of 21,212 RA patients
Variable Age (years) Sex (% male) Ethnic origin White, not of Hispanic origin (%) Black, not of Hispanic origin (%) Asian or Pacific Islander (%) American Indian or Alaskan Native (%) Hispanic (%) Other (%) Married (%) Education category (yrs completed) 0-8 (%) 8-11 (%) 12 (%) 13 15 (%) 16 or > (%) Disease duration (years) HAQ (0-3) Lifetime comorbidity score (0-11) Diabetic (%) Prior prednisone use (%) Mean 61.6 23.4 89.3 4.9 1.2 1.1 2.9 0.8 61.7 3.4 9.2 37.5 25.6 24.3 15.0 1.1 4.17 9.6 69.5 (S.D.) ±13.6
Author Contact Information
Liron Caplan, MD lcaplan@im.wustl.edu 314-454-7279
0 9 32 180 681 1693 2281 636 19
7.0 262.0 945.0 2624.0 4582.0 5085.0 4207.0 977.0 28.0
0.0 3.4 3.4 6.9 14.5 33.3 54.2 65.1 67.9
(0,0) (1.2,5.7) (2.2.4.5) (5.9,7.8) (13.8,15.9) (32.0,34.6) (52.7,55.7) (62.1,68.1) (49.4,86.3)
Author Disclosures:
Centocor Health Outcomes in Rheumatic Diseases Fellowship; 2 The NDB has received grant support from Centocor, Bristol-Myers-Squibb, Aventis, Abbott, Amgen, Pfizer and Merck pharmaceutical companies Figure 3: Hazard ratios for potential predictors of cataract formation
3
1
Figure 1: Age-specific rate of cataracts in RA patients according to prednisone exposure
Figure 2: Odds of developing cataracts according to average daily dose of prednisone.
±10.9 ±0.8 ±3.2
at
>=
20
30
40
50
60
70
80
0
0.1-5
5.1-10
10.1-20
>20
Age (years)
Average daily prednisone dosage (mg)
Variable in Cox Proportional Hazard Regression
Ed uc
<2
-2
-3
-4
-5
-6
-7
-8
Pr e
90
0
9
9
9
9
9
9
9
dn
0
io
iso
100 90 80 70 60 50 40 30 20 10 0
Incidence Rate (per 100 pt yrs)
Patients with prednisone use Patients with no prednisone use
Hazard Ratio (95% CI)
8
Odds Ratio (95% CI)
2.5 2 1.5 1 0.5 0
A ge er r tic be D ia ev e en ll Sm ok ne U se de l na
7 6 5 4 3 2 1
G