Table 2. Use of glucocorticoids and risk of skin cancer and non-Hodgkin's lymphoma.
|
All prescriptions before diagnosis date
|
Prescriptions>1 year before diagnosis date | Prescriptions>5 years before diagnosis date | |||||
|---|---|---|---|---|---|---|---|
| Cases N (%) | Controls N (%) | IRR age and gender adjusted (95% CI) | IRR adjusteda (95% CI) | P-value for trend test (adjusted) | IRR adjusteda (95% CI) | IRR adjusteda (95% CI) | |
| Basal cell carcinoma | |||||||
| No prescriptions | 4300 (79%) | 18278 (81%) | 1.00 | 1.00 | — | 1.00 | 1.00 |
| Any prescription | 1122 (21%) | 4179 (19%) | 1.17 (1.08–1.27) | 1.15 (1.07–1.25) | 1.17 (1.08–1.28) | 1.22 (1.09–1.36) | |
| Risk increase per 10 000 mgb | 1.10 (0.98–1.22) | 1.07 (0.96–1.20) | 0.2 | ||||
| Squamous cell carcinoma | |||||||
| No prescriptions | 732 (78%) | 3379 (81%) | 1.00 | 1.00 | — | 1.00 | 1.00 |
| Any prescription | 203 (22%) | 787 (19%) | 1.21 (1.01–1.46) | 1.14 (0.94–1.39) | 1.09 (0.88–1.34) | 1.11 (0.85–1.45) | |
| Risk increase per 10 000 mgb | 1.23 (0.98–1.54) | 1.12 (0.88–1.43)c | 0.4 | ||||
| Malignant melanoma | |||||||
| No prescriptions | 820 (83%) | 3696 (84%) | 1.00 | 1.00 | — | 1.00 | 1.00 |
| Any prescription | 163 (17%) | 710 (16%) | 1.09 (0.90–1.33) | 1.15 (0.94–1.41) | 1.12 (0.90–1.39) | 1.00 (0.75–1.36) | |
| Risk increase per 10 000 mgb | 1.11 (0.91–1.37) | 0.94 (0.65–1.37) | 0.8 | ||||
| Non-Hodgkin's lymphoma | |||||||
| No prescriptions | 383 (80%) | 1648 (82%) | 1.00 | 1.00 | — | 1.00 | 1.00 |
| Any prescription | 98 (20%) | 356 (18%) | 1.20 (0.92–1.56) | 1.11 (0.85–1.46) | 1.08 (0.81–1.44) | 1.04 (0.71–1.54) | |
| Risk increase per 10 000 mgb | 2.51 (1.41–4.49) | 2.26 (1.20–4.26) | 0.01 | ||||
CI=confidence interval; IRR=incidence rate ratio; SCC=squamous cell carcinoma.
a
Conditional logistic regression was used to estimate IRRs and 95% CI, adjustments were made for a prior hospitalisation for selected chronic diseases and use of methotrexate and azathioprine.
b
In this model we assumed a linear effect of the exposure.
c
The restricted cubic spline showed the best fit to the data (Harre et al, 1988), with a statistically significant increased risk of SCC with increasing amounts of glucocorticoids prescribed (cumulative IRR per 10 000 mg=7.40 (95% CI: 1.80–30.4), P=0.001).