Reducing lupus flares: should we be more careful about stopping glucocorticoids?

KEYWORDS:

• Systemic lupus erythematosus
• relapse
• flare
• prednisone
• remission
• low disease activity

1. Introduction

Systemic lupus erythematosus (SLE) is a chronic relapsing-remitting disease characterized by an unpredictable course. Given the increased risk of organ damage due to persistent or fluctuating disease activity over time, durable remission is a very desirable outcome. However, an active disease-free state is generally maintained only when patients are on medication which often leads to treatment-related complications. Therefore, once remission has been achieved, prolonged maintenance treatment inevitably requires a regimen of drug de-escalation. Several factors have to be taken into account during this decision-making process, with both long-term toxicity and the ability of the drug to prevent relapse being crucial factors. The procedure for drug withdrawal is even more complicated by the observation that sustained drug-free remission, defined as a period of at least 5 years without active disease in the absence of any medication, including glucocorticoids (GCs), hydroxychloroquine (HCQ), and immunosuppressant (ISS), is infrequent in SLE, ranging only from 3% to 7% [1–6].

The latter drugs have been used for nearly 50 years now to maintain remission in SLE, in spite of the fact that their relative efficacy and toxicity are still poorly known for this indication. Accordingly, the duration and dose of GC, HCQ, and ISS therapy, after achievement of disease remission, remain undefined in SLE and it is only with medication tapering that the physician is able to determine whether the disease was truly in remission or maintained in remission as a result of medication.

2. Clinicians’ approaches to drug withdrawal in SLE patients in remission: unfulfilled wishes

An internet-based survey of 130 clinicians from 30 countries has shown that prednisolone was by far the first medication suggested by physicians to be reduced or withdrawn during remission from any therapeutic regimen. The wish to withdraw GC first was irrespective of persistent serological abnormalities, remission duration, minor or major organ involvement, and use of prednisolone with HCQ alone, or as part of a treatment regimen involving HCQ and an ISS as well [7]. Concurrently, even with mild disease, normal serology, and a 5-year clinical remission, almost 90% of the clinicians continue to prescribe HCQ [7]. Accordingly, the recent EULAR recommendations for the treatment of SLE during chronic maintenance treatment advocate that GCs should be, when possible, withdrawn [8]. However, in routine practice, a significant proportion of treating physicians prefers to continue a low-dose GC regimen, despite clinical remission, which is most likely due to the fear that withdrawal of low-dose GCs may lead to a severe flare, even after very long intervals of remission. This apprehension is especially heightened if there is a previous history of major organ involvement, as observed in lupus nephritis or neuropsychiatric SLE [7]. Consequently, it has been reported that between 57% and 86% of SLE patients undergo long-term low-dose GC treatment, in spite of the uncertainties surrounding the ability of low-dose GCs to maintain remission in SLE [9,10].

3. Ability of low-dose corticosteroids to maintain remission in SLE

Recent data suggest that treatment with low-dose GCs prevents relapse in about a fifth to a third of patients with SLE. In particular, Tani et al. reported the results of a longitudinal study of a cohort of 148 Italian patients with SLE among which 91 attempted interruption of GCs intake during the 6 years of follow-up with 74% of patients treated with HCQ. A total of 77 patients (85%) successfully stopped GC intake while 14 patients (15%) were not able to discontinue treatment due to flare symptoms. For those patients who were successfully withdrawn from GC, 18 flares (23%) were recorded, the time lapse of quiescent disease being the sole predictor of flare onset. Even if 72% of flares were mild, the reintroduction of GCs was necessary in almost all patients [11]. Similarly, Goswami et al. reported that 21% of Indian patients in remission, and treated with HCQ, experienced an exacerbation of the disease after GC withdrawal [12]. Unlike the work of Tani et al., many flares were deemed major. In this study, duration of disease, duration of GCs before interruption, and an ISS were independent predictors of flare-free survival [12].

In a recent randomized controlled trial, we have shown that in SLE patients with quiescent disease and a stable treatment regimen for at least one year, withdrawal of 5 mg of prednisone was associated with a four-fold increase (i.e. 27%), in the risk of flare onset, as defined by the SELENA-SLEDAI flare index (SFI) and the British Isles Lupus Assessment Group (BILAG) index during a one-year follow-up [13]. Other SLE treatments remained unmodified during this study. In particular, at study entry, 91% and 27% of the patients were also treated with HCQ and an ISS, respectively. Therefore, although the one-time withdrawal of 5 mg/day could be considered by some physicians as too abrupt with the consequence of favoring the appearance of flares, it is important to consider that in our study, the vast majority of patients, following the interruption of GC intake, remained on long-term treatment with HCQ and that, therefore, their treatment for SLE was never abruptly stopped. Our results thus indicate that low-dose GCs even after several years still have anti-inflammatory properties, thereby emphasizing the interest of continuing a low dose of prednisone at long course to avoid relapse [13]. None of the factors identified in the previous observational cohorts, including also low complement C3 serum levels and the presence of anti-double-stranded DNA antibodies, were associated with the risk of flare after GC withdrawal, possibly due to the small number of patients included in these studies.

4. Expert opinion: How do we move forward?

The search for new treatments in SLE is ongoing and suitable drugs will certainly emerge in the coming months to facilitate the care of patients with active disease. Nevertheless, clinicians must also challenge in academic clinical trials the modalities of therapeutic de-escalation in patients in remission in view of the low number of work in this field [14]. At present, it is quite common that patients continue to receive, for indefinite periods of time, the same treatment regimen that initially led to remission, even after prolonged periods of remission and mild disease [7]. The usefulness and the toxicity of these drugs in the long term should be evaluated in order to maintain the safest possible therapeutic pressure in SLE while avoiding flares which will doubtlessly increase the disease burden. It should be stressed that the toxicity of HCQ in this indication is probably underestimated. For example, Petri et al. have recently reported that more than one-tenth of patients treated with HCQ, for periods as long as 16 years, presented with retinopathy [15]. In this study, it was also shown that higher blood levels of HCQ predicted later retinopathy which suggests that a decrease in the daily dosage of this drug could reduce the appearance of this ocular complication [15]. Yet, the effective dose of HCQ intake and the corresponding HCQ blood levels in remission are not known, an issue that should be addressed in clinical trials.

With regard to other treatments for SLE, several authors advocate the use of ISS rather than low-dose GCs for maintaining long-term remission, even if the benefit/risk ratio of ISS with regard to the infectious and oncological risk remains to be specified. Results from a case-cohort study have shown that long-term immunosuppressive therapy increases the probability of tumorigenesis, notably the occurrence of hematological malignancies, with a time lag of five years between drug intake and cancer onset [16]. On the other hand, enthusiasm for long-term prednisone treatment, even if effective, is understandably tempered by potential side-effects such as infections, diabetes mellitus, cataract, osteoporosis, gastrointestinal bleeding, and cardiovascular disease, leading to the development of irreversible organ damage [10,17–24]. However, if the toxicity of GCs at doses exceeding 7.5 mg/day is well known, that of doses inferior or equal to 5 mg/day remains to be established. Currently, the European League Against Rheumatism (EULAR) agrees that the degree of harm brought about by long-term GG administration is dose-dependent with dosages of less than 5 mg/day prednisone producing an acceptably low degree of harm, with the exception of patients at high cardiovascular risk [25]. Yet, the long-term tolerance to very small doses of prednisone (i.e. ≤5 mg per day) should be assessed from population-based studies, whereas the appropriate course of action to stop GCs, either abruptly or more slowly, should be addressed in large double-blind randomized controlled trials as well.

Finally, most of the time, only medication withdrawal will allow the physician to determine whether the disease is truly in remission or maintained in remission as a result of drug administration. Future studies should be undertaken to determine whether clinical characteristics and biomarkers, such as low IFN-α serum levels, could help to identify SLE patients who are at a lower risk of relapse and therefore could benefit from discontinuation of their chronic maintenance treatment [26].

5. Conclusion

Therapeutic intervention in patients in remission should be aimed at the prevention of SLE relapse while avoiding undesired drug toxicity. However, at present, pertinent information permitting to determine which regimen to interrupt first in the treatment of remission in SLE is still missing. Therefore, until the availability of effective drugs with little or no toxicity, it is recommended to not abandon the option of using very low doses of GCs (i.e. ≤5 mg prednisone) given their potential benefits in patients at low cardiovascular risk.

Declaration of interest

The authors have no relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending, or royalties.

Reviewer disclosures

One peer reviewer declares speaker fees from Eli Lilly and Company, Merck Sharpe and Dohme, Amgen, UCB Pharma, Roche, and Sanofi Genzyme and a consultant fee from Sanofi Genzyme (not related to the subject matter discussed in the manuscript reviewed). No other peer reviewers on this manuscript have no other relevant financial or other relationships to disclose.

Additional information

Funding

This paper was not funded.