You are here
Prof. Bernard Combe discusses safe use of glucocorticoid therapy
Interview with Prof. Bernard Combe
There is a wide debate in the medical community regarding the balance between efficacy and safety of glucocorticoid use. Prof. Bernard Combe defines what he considers as the spectrum of reasonably safe use, citing data from the ESPOIR cohort.
The use of glucocorticoid in inflammatory arthritis is very important for daily practice. Many patients are on glucocorticoid and we have concerns – there is a wide debate in the medical community regarding the balance between efficacy and safety of glucocorticoid. This was also included in the new EULAR recommendations on arthritis but also on the management of rheumatoid arthritis. We now have evidence – more evidence than we had before – supporting the efficacy, both on clinical and … outcome and, secondly, by contrast, we also have evidence that the safety profile in the long-term of glucocorticoid, given at quite low dose, is not very good.
We wanted to try to answer this question on safety in our National Early Arthritis Cohort, which is the ESPOIR cohort, which was set up in 2002. We recruited patients in 2002 and 2005, more than 800 patients. We wanted to look at the safety profile of glucocorticoid in this early arthritis cohort, from the beginning and for seven years. The advantage of this cohort is that patients at baseline were glucocorticoid-naïve – this was one of the inclusion criteria.
We then wanted only to keep on board patients with early rheumatoid arthritis – patients who may not have had rheumatoid arthritis at the start, but who will have developed rheumatoid arthritis during the follow-up. Secondly, we excluded patients who already had important safety concerns with cardiovascular disease, severe infection, or bone fracture. These patients were excluded from the analysis. At the end, we also excluded patients who dropped out within the first year because it wasn’t possible to have the correct follow-up.
In the end, we kept more than 600 patients; more than 600 patients were evaluated. We separated these into two groups: patients who had received systemic glucocorticoid, 380 patients; and patients who had never received any systemic glucocorticoid, 220 patients. In addition to the comparison between these two groups, using the Cox model, we developed the propensity score, which was defined by the willingness to use glucocorticoid. We included this propensity score in the Cox model, so we performed two analyses – a crude analysis and one with the propensity score.
What was the main result? First, the outcome variable that we used was a composite score, a composite measurement, which included deaths, major cardiovascular disease, severe infection and bone fracture. We had 65 events, so almost 10 per cent of the cohort. We considered also that the propensity score was important because the two groups, as expected, were different. The patients with glucocorticoid had more active disease at the start and they were also more at par, and also they were most focussed. It was really necessary to adjust and to use this propensity score.
The main result was that, from the univariate analysis before adjustment, there was no significant difference between the two groups. This means that there was no significant increase with the glucocorticoid groups compared to the others. When including the propensity score, it was the same, with no significant difference between the two groups.
What I forgot to tell you, however, is that in this cohort, prednisone was used at very low dosage. The mean dosage for the well cohort during these seven years was 3.2mg/day, the mean, and the median was around 3mg also. That is very important. Even the duration of corticoid was quite long, at around three years for the mean and around two years for the median. That was quite long-term use but at very low dose. The conclusion for us is that, during seven years, there was no evidence of increased serious risk of safety when using a very low dose of glucocorticoids.
Q. So basically then, the ESPOIR cohort data goes into the spectrum of reasonably safe use, and anything above - ?
A. Safe use, if you use very low dose. Very low dose. That is consistent with what has been - - but this is the first time that this has been reported in a prospective cohort.
Q. Yes, that is very helpful. I think this will be at least an important focal point for practising rheumatologists to work with their patients, to get to the lowest dose.
A. Yes. This supports the current EULAR recommendation and it is interesting that, when we proposed this recommendation, we didn’t know the results of this study.
Q. Thank you very much. That was a very important addition.
A. Thank you very much.