Low-dose oral prednisone improves clinical and ultrasonographic remission rates in early rheumatoid arthritis: results of a 12-month open-label randomised study

Glucocorticoids (GC) in association with conventional disease-modifying anti-rheumatic drugs (DMARDs) are currently recommended for the treatment of rheumatoid arthritis (RA), after a careful evaluation of the patient's risk-to-benefit ratio [1].

In RA, GC show a rapid symptomatic effect on pain and general well-being and help in controlling objective signs of inflammation, such as tender and swollen joints [2‐4]. Several studies also support disease-modifying properties for GC, in terms of prevention of future joint damage in early RA, both for initial short-term high dosages followed by step-down schedules and for low-dose oral therapy regimens [3‐8].

On the whole, in clinical trials involving early-onset RA patients treated with low-dose oral GC in association with DMARDs, clinical and functional benefits seem to last only few months after the start of treatment, fading away thereafter, whilst prevention of structural damage may persist over a longer time [5, 6]. These studies suggest a sort of dissociation between a short-lasting clinical improvement and a long-lasting effect on radiographic progression [9]. Whether the effect on structural damage is just related to a faster clinical response or may be associated with a persistently deeper control of joint inflammation, not fully detectable by clinical tools, is still to be ascertained.

Sensitive imaging techniques may help in testing these hypotheses. Recent imaging studies on RA have demonstrated that the structural progression observed in patients in clinical remission might be explained by the persistence of subclinical signs of synovitis detected by magnetic resonance imaging or ultrasonography (US) [10]. In particular, power Doppler (PD) US has been demonstrated to be a valid and reliable tool in detecting subclinical inflammation in RA, and is believed to be a measure of activity of joint inflammation [11, 12]. At present there are no data on the effect of low-dose oral GC on US-detected synovial inflammation in early RA.

This study aimed to verify the effect of low dosages of oral prednisone (PDN) on clinical and US outcomes over 12 months of follow-up in patients with recent-onset RA. For this purpose we randomly allocated patients to receive low-dose PDN or not within a structured treat-to-target protocol aiming for low disease activity (LDA) [13]. Disease activity was prospectively assessed both clinically and by US.

Our data indicate that low-dose oral PDN leads to higher probability of clinical remission and PD negativity at 1 year of follow-up in a LDA-driven therapeutic protocol in early-onset RA patients. We found that the percentage of subjects achieving LDA at 1 year was not statistically different between the two treatment arms. This is not surprising with a treat-to-target therapeutic approach along with a higher frequency of therapeutic adjustments in the MTX monotherapy group. Nevertheless, GC co-medication gave an earlier and faster clinical response in terms of DAS28 values over time, with significant intergroup difference already detectable after 2 months of treatment, persisting up to 6 months.

The prompt clinical effect of GC co-medication in early RA has already been reported by several authors [2‐5]. In this regard, Kirwan presented no sustained clinical advantage on GC co-medication in early RA patients since the addition of prednisolone to DMARDs accelerated, but did not enhance, the clinical response [2]. The clinical benefits on singular clinical parameters under a very low dose (5 mg/day) of GC co-medication also did not continue beyond the first year of follow up in Wassenberg and colleagues' early RA cohort [5]. Only Svensson and colleagues found a sustained clinical effect on low-dose oral prednisolone (7.5 mg/day) as an additional therapy to DMARDs in early RA patients [7].

Analysing in our study the trend of each singular clinical and laboratory parameter over time, we observed only an initial substantial advantage in adding GC compared with MTX monotherapy in clinical variables, with no persistent differences. Most of the benefits associated with GC co-medication on DAS28 response seems to be explained by significantly greater, faster and persistent control over time on acute-phase reactant serum levels (erythrocyte sedimentation rate and C-reactive protein). This might also have indirect relevant implications, considering their prognostic value, especially for C-reactive protein levels over time, in predicting subsequent structural damage both in early and established RA [24, 25].

The analysis of the US data strengthens this result. There is evidence that a subgroup of RA patients with drug-induced sustained remission still have detectable active subclinical synovitis [10, 23]. We analysed the percentage of PD negativity in early-onset disease after achieving DAS28 remission. The analysis revealed that PD negativity was almost 1.5-fold higher in the PDN co-medication group than in the MTX monotherapy group (Additional file 1 shows this in more detail).

Previous studies on the effect of steroid medication on US measures in RA are scanty and mostly refer to the favourable short-term effect of a single intravenous, intra-articular or high-dose oral administration in established disease [26‐29].

In our study, median US score values significantly decreased during follow-up in the whole series of patients, but GC-treated patients presented a trend toward a greater and more rapid response, so that GC-treated patients more frequently reached a completely negative PD status at the end of follow-up. Nevertheless, in both treatment arms, PD score trends over 6 months of follow-up were not predictive of the subsequent achievement of clinical remission. This finding is particularly important due to the predictive role of PD active synovitis on relevant disease outcomes [10, 22]. As the persistency of a positive PD signal is the main predictor of early relapse and radiographic progression in patients in clinical remission, the higher effect of GC on PD may explain the long-term beneficial effect on structural damage.

The particular setting of an inceptional cohort enabled us to analyse the specific effect of a structured therapeutic protocol including or not including low-dose PDN in untreated patients from the onset of the disease. Furthermore, this study is the first controlled clinical trial exploring the effect of low-dose oral GC on US outcomes in early RA. The blinded assessment of US outcome strengthens the validity of these results.

The present study has some limitations. The sample size was able to detect only large differences between the two therapeutic arms, so the study was underpowered for subgroup analyses - such as the specific effect of GC on the patients achieving clinical remission. Furthermore, caution should be taken when interpreting the positive results in terms of the combined clinical and US variables, as this combined outcome was not prespecified. A reduced number of joints (n = 12) was assessed by US in our study. Although a similar US evaluation has been extensively applied and validated in RA, we are aware that a small percentage of joints with signs of active synovitis might be lost in this way [30‐32]. Moreover, the therapeutic strategy applied in our study is less aggressive than currently proposed MTX escalation schedules [33, 34]. This characteristic is related to the time when this study was designed and might have increased the relative effect of combination therapy over MTX monotherapy. Finally, the sample size and length of follow-up did not allow us to measure the effect of GC on rare events, such as the appearance of new radiographic erosions or some side effects such as osteoporotic fractures. In relation to the safety profile, data from our analyses described few relevant side effects related to low-dose PDN medication, in accordance with a recent meta-analysis presenting a very good safety profile [35, 36]. In general, patients on GC were more likely to withdraw the treatment protocol due to unwillingness to proceed with GC medication, as also reported in a previous cross-sectional survey [37].

Our results suggest that in patients with early-onset RA, the addition of low-dose oral PDN to a MTX background in the contest of a treat-to-target strategy may ensure a faster and better control of disease activity, with higher rates of both clinical remission and PD signal negativity at 1 year.

To our knowledge, this study shows for the first time that a greater clinical and subclinical effect of low-dose GC co-medication over DMARD monotherapy was demonstrated in early RA patients in a controlled manner. Further data on subsequent structural damage might confirm and clarify the role of such deeper control of inflammation on GC therapy. The acceptable safety profile along with the positive clinical and imaging results should allow us to suggest such a treatment strategy in the management of early disease.