Article Text

Download PDFPDF

Extended report
Remission in systemic lupus erythematosus: durable remission is rare
  1. Theresa R Wilhelm1,2,
  2. Laurence S Magder3,
  3. Michelle Petri1
  1. 1Department of Rheumatology, Johns Hopkins University School of Medicine, Baltimore, Maryland, USA
  2. 2Videncenter for Reumatologi og Rygsygdomme, Rigshospitalet Glostrup, Glostrup, Denmark
  3. 3Department of Epidemiology and Public Health, University of Maryland, Baltimore, Maryland, USA
  1. Correspondence to
    Theresa R Wilhelm, Videncenter for Reumatologi og Rygsygdomme, Rigshospitalet Glostrup, Nordre Ringvej 57, Glostrup 2600, Denmark; theresa.rita.wilhelm{at}regionh.dk

Abstract

Introduction Remission is the ultimate goal in systemic lupus erythematosus (SLE). In this study, we applied four definitions of remission agreed on by an international collaboration (Definitions of Remission in SLE, DORIS) to a large clinical cohort to estimate rates and predictors of remission.

Methods We applied the DORIS definitions of Clinical Remission, Complete Remission (requiring negative serologies), Clinical Remission on treatment (ROT) and Complete ROT. 2307 patients entered the cohort from 1987 to 2014 and were seen at least quarterly. Patients not in remission at cohort entry were followed prospectively. We used the Kaplan-Meier approach to estimate the time to remission and the time from remission to relapse. Cox regression was used to identify baseline factors associated with time to remission, adjusting for baseline disease activity and baseline treatment.

Results The median time to remission was 8.7, 11.0, 1.8 and 3.1 years for Clinical Remission, Complete Remission, Clinical ROT and Complete ROT, respectively. High baseline treatment was the major predictor of a longer time to remission, followed by high baseline activity. The median duration of remission for all definitions was 3 months. African-American ethnicity, baseline low C3 and baseline haematological activity were associated with longer time to remission for all definitions. Baseline anti-dsDNA and baseline low C4 were associated with longer time to Complete Remission and Complete ROT. Baseline low C4 was also negatively associated with Clinical Remission.

Conclusions Our results provide further insights into the frequency and duration of remission in SLE and call attention to the major role of baseline activity and baseline treatment in predicting remission.

  • Outcomes research
  • Systemic Lupus Erythematosus
  • Disease Activity
View Full Text

Statistics from Altmetric.com

Introduction

Systemic lupus erythematosus (SLE) occurs in three different patterns: relapsing-remitting, chronically active and remission.1 A successful outcome of treatment for the relapsing-remitting pattern might be reduction/prevention of flares, whereas, for the chronic active pattern, reduction or elimination of disease activity is the goal.

In 2010, Nossent et al analysed a multicenter European cohort and found that early remission predicts a better disease outcome.2 Recently, Zen et al and Steiman et al showed that patients with prolonged remission had a lower burden of damage.3 ,4 Thus, as in rheumatoid arthritis,5 remission is the ultimate goal in SLE.

To apply the principle of treating-to-target to SLE, it is necessary to precisely define what remission in SLE means. Accordingly, developing definition(s) of remission in SLE became the primary recommendation on the research agenda of the treat-to-target international task force.6 Investigation of remission in SLE has been ongoing for >50 years,7–11 yet no agreement on the definition of remission has been widely accepted. There are few reports on remission and poor agreement on criteria, thus yielding varying results. Online supplementary table S1 summarises selected studies from the last 20 years investigating remission. These studies largely agree that no signs of clinical disease activity should be present in remission. However, disagreement starts when it comes to the definition of ‘no disease activity’. Early studies of remission were imprecise in the definition of ‘asymptomatic’,9 ‘complete remission of clinical and laboratory features of disease’10 or ‘asymptomatic and without obvious active organ involvement’.11 Later studies applied the Systemic Lupus Disease Activity Index (SLEDAI) as an instrument to measure activity. Barr et al and Nossent et al used the physician global assessment (PGA) in their respective definitions of remission, as the SLEDAI is less able to detect mild degrees of disease activity.1 So far, SLEDAI and PGA have not been used in combination to study remission.

Supplementary table

A Selection Of Past Studies Investigating Remission In SLE

Varying importance has been attributed to serological abnormalities. A few study definitions required serological quiescence in addition to clinical quiescence,3 ,12 but most accept serological activity.1 ,3 ,4 ,12–15 In 1979, Gladman et al first described a small subset of patients showing serological activity, evidenced by elevated anti-dsDNA levels and/or hypocomplementaemia, despite clinical quiescence.9 In 2012, Steiman et al showed that patients who achieved a serologically active, clinically quiescent period lasting for at least 2 years accrued less damage over a decade compared with matched controls.16

The minimum duration of quiescence required to fulfil the definition of remission has varied. Some studies have chosen a time criterion of 5 years in line with oncological remission criteria. Other common minimal durations are 1 and 2 years. Different studies have shown quite consistent results for the 1 year remission rates, with ranges between 18.9% and 28.4% if serological activity was allowed.1 ,2 ,12 ,13 ,15 In one study where serological activity was not allowed, the 1 year remission rate was just 6.5%.12 The 2, 3 and 5 years remission rates were generally much lower, with the 5 years rates between 1.7% and 2.4%.4 ,12 An exception was Zen et al’s recent study with far higher 5 years remission rates.3

Another important aspect of the definition of remission is whether remission requires withdrawal of all treatments. Whereas older studies did not allow any treatment, more recent ones permit treatment with antimalarials.

A very important, but yet poorly studied group of patients in studies of remission are those who are asymptomatic but requiring corticosteroids and/or immunosuppressives. In 2014, Steiman et al suggested that the group of patients achieving remission under coverage of treatment with corticosteroids and/or immunosuppressives may consist of one subset in whom medication can be tapered and withdrawn (thus being in true remission) and another subset in whom disease was merely suppressed.4

Besides studying the frequency of remission and its duration, an issue of great interest is what characterises patients who achieve remission (see online supplementary table S1). So far, few studies have investigated predictors of remission and reported significant differences between the patients who experienced remission and the ones who did not.2–4 ,12 ,15 Zen et als recent study identified glomerulonephritis, vasculitis and haematological manifestations to be negative predictors of remission in multivariable models.3

Differences in the cohorts may play an essential role in explaining the varying results. Ethnicity has been proven to have an effect on disease manifestations, severity and prognosis of SLE.17 Nevertheless, not all papers give the information on the ethnic groups in their cohorts.

In this study, we used the definitions of remission agreed on by a large multiparty international collaboration from the DORIS (Definitions Of Remission In SLE) working group that aims to find consensus definitions of remission in SLE (van Vollenhoven RF, Aranow C, Bertsias G, et al. Remission in SLE, consensus findings from a large international task force on definitions of remission in SLE (DORIS). submitted).18 Applying these definitions to a large, closely monitored clinical cohort, this study aimed to determine the time to remission, the duration of remission and the predictors of remission in SLE.

Methods

Four different definitions have been considered: Clinical Remission, Complete Remission, Clinical ROT and Complete ROT, where ROT stands for remission on treatment (table 1). To fulfil any of these definitions, the clinical SLEDAI (cSLEDAI), meaning without serology, has to be 0 and the PGA<0.5 on a 0–3 visual analogue scale. Prednisone is not allowed for a patient to be considered in Remission. The daily dose of prednisone has to be ≤5 mg to be considered in ROT. Immunosuppressive drugs are not allowed for remission, but they are allowed for ROT. What distinguishes Clinical and Complete Remission and ROT is the serology. Clinical Remission and ROT are regardless of serology, whereas Complete Remission and Complete ROT also require negative anti-dsDNA and normal complement. Hydroxychloroquine is allowed for both Remission and ROT.

Table 1

DORIS definitions of remission

We applied four DORIS definitions of remission to the Hopkins Lupus Cohort, which is approved on an annual basis by the Johns Hopkins Institutional Review Board. Patients gave informed consent before taking part in the cohort. A total of 2307 patients were included in our analyses. Patients entered the cohort from 1987 to 2014. Patients were seen quarterly, or more often if warranted. At each clinic visit a large amount of clinical information was collected, including information on SLE disease activity and relevant serologies (anti-dsDNA via Crithidia, complement).

In this analysis, for each of the four definitions of remission, we considered the experience of patients who did not satisfy that definition at baseline, defined as cohort entry. These patients were followed prospectively until either they satisfied that definition, they had a gap of nine or more months in their follow-up or they dropped out of the study. For 89% of consecutive visits, the time between visits was <4 months. The time to remission was ascribed to the first clinic visit in which the specific definition of remission was satisfied.

We used the Kaplan-Meier approach to estimate the distribution of time to remission after cohort entry censoring patients who had a gap in their follow-up or dropped out before satisfying the definition of remission. We stratified by baseline disease activity (low activity: PGA<1 and SLEDAI<3) and baseline treatment (low treatment: prednisone <5 mg/day, no use of immunosuppressive drugs). Once remission was achieved, patients were followed prospectively until relapse occurred. The Kaplan-Meier approach was used to estimate the distribution of time from remission to relapse. Finally, Cox regression was used to identify patient characteristics at baseline that were associated with time to remission, adjusting for baseline disease activity and baseline treatment. Using this approach, we examined demographic and immunological characteristics as well as specific types of disease activity, one variable at a time. Those variables found significantly associated with remission were then entered into multivariable models and those that remained significant were retained in the final models.

Results

Patient characteristics

Table 2 shows the demographic and clinical characteristics of the patients included in the analysis for each definition. In general, 92% were female, and most were either Caucasian American or African-American. There was a wide age span from <30 to >60. Many (39%) had been diagnosed with SLE within the last year, but 35% had SLE for five or more years.

Table 2

Clinical and demographic characteristics of the patients in the Hopkins lupus cohort who were not in remission at baseline, by each of the four definitions

Time to remission

Table 3 shows the median time to remission for all the definitions, overall and stratified by baseline disease activity and baseline treatment. We found that baseline disease activity and baseline treatment were strongly associated with time to remission. Patients with low baseline activity (PGA<1 and SLEDAI<3) and low treatment (prednisone <5 mg/day, no use of immunosuppressive drugs) achieved remission faster than those with high disease activity and high treatment. High treatment was the major predictor of a longer time to remission, followed by high activity. These finding are further illustrated in figure 1, which shows the Kaplan-Meier curves for time to remission overall and in subgroups defined by baseline treatment and disease activity levels. By 1 year after cohort entry, for example, the probability of achieving Clinical Remission was only 3% for patients with high disease activity and high treatment, but it was 42% for patients with low disease activity and low treatment.

Table 3

Median time (years) until remission for each definition and by baseline disease activity and baseline treatment strata

Figure 1

Kaplan-Meier curves with time to remission for unstratified and stratified data. The y-axis shows the probability of continuously not being in remission. The x-axis shows the time in days. Low activity was defined as physician global assessment <1 and Systemic Lupus Disease Activity Index <3. If the activity was not low, we called it high. Low treatment was defined as prednisone ≤5 mg/day and no use of immunosuppressive drugs. If the treatment was not low, we called it high. All patients are shown in blue. Patients with high systemic lupus erythematosus (SLE) activity and high treatment are shown in red. Patients with high SLE activity and low treatment are shown in green. Patients with low SLE activity and high treatment are shown in brown. Patients with low SLE activity and low treatment are shown in violet. ROT, remission on treatment.

Duration of remission

We found the median duration of remission to be about 3 months. This was quite similar for all definitions with a median duration of 105, 101, 109 and 101 days for Clinical Remission, Complete Remission, Clinical ROT and Complete ROT, respectively. This reflected the fact that most patients were seen quarterly. Based on Kaplan-Meier estimates (curves not shown), we determined the durability of remission by specified times as shown in table 4. At 1 year, for example, only 13.2% who entered Clinical Remission were still in it. At 5 years time, only 1.2% of the patients were still in Clinical Remission and at 10 years only 0.4%.

Table 4

Per cent of patients still in remission, by time since start of remission

Predictors of remission

Table 5 shows characteristics of patients at cohort entry that were independently associated with time until remission based on multivariable Cox regression models. We found that African-American ethnicity and low C3 levels at baseline were associated with lower rates of remission for all four definitions. Baseline anti-dsDNA and baseline low C4 levels were associated with lower rates of Complete Remission and Complete ROT. Baseline low C4 levels were also negatively associated with Clinical Remission. The only clinical manifestation that was related to longer time to remission was haematological disease activity. Other clinical manifestations (musculoskeletal, cutaneous and renal disease activity) were not significantly associated with longer time to remission.

Table 5

Baseline variables independently associated with delayed remission based on multivariable Cox models controlling for baseline levels of disease activity and treatments

Discussion

In this study, we applied four different definitions of remission agreed on by an international collaboration from the DORIS working group to a large cohort of patients with SLE. We determined time to remission, durability of remission and predictors of remission. As one might expect, it was easier to reach ROT than Remission. Baseline treatment and baseline disease activity were strongly associated with the time to remission for all definitions. We found that the durability of remission was very short, regardless of definition. African-American ethnicity, baseline low C3 and baseline haematological activity were associated with a longer time to remission. Baseline anti-dsDNA and baseline low C4 were negative predictors for Complete Remission and for Complete ROT. Baseline low C4 was also a negative predictor for Clinical Remission. Our results provide further insights into the frequency and durability of remission in SLE and call attention to the major role of baseline activity and baseline treatment in predicting remission.

Our results concerning durability of remission show the relapsing-remitting nature of SLE. The median duration of remission was only about 3 months for all definitions. This was the time to the next quarterly cohort visit. Even though achieving remission was frequent, durable remission was rare. Notably, even though the time to remission was very variable among the four different definitions, the median duration of remission was very similar, thus underlining the need for further treatment options to reach sustained remission.

Looking at the previous studies on remission (see online supplementary table S1), one clearly sees how the duration of remission varies according to the minimal time criterion set in the definition. Omitting an arbitrary time criterion in the definitions of remission used in this study, though, allowed us to better approximate the true nature of remission in SLE, which is of a rather short duration. In other chronic inflammatory diseases, such as Rheumatoid Arthritis and Crohn’s disease, the definitions of remission do not include a time criterion.19 ,20 Nevertheless, it can be argued that remission needs to be a durable state to be considered a desirable treatment outcome. In childhood SLE, the preliminary definition of remission agreed on by an international collaboration required a duration of at least 6 months.21

Very few studies have obtained data on patients who achieve remission while taking medications other than hydroxychloroquine. Their results, though, are in good agreement with our findings that only few patients maintain durable remission. At 1 year, 13.4% and 9.3% of patients in our cohort were still in clinical and Complete ROT, respectively. Our data further shows that 0.6% and 0.7% of patients had a clinical and Complete ROT, respectively, that lasted 5 years. In Steiman et al’s study from 2014, 2.1% of the patients achieved remission when defined with a minimal durability of 5 years and all medications allowed.4 Urowitz et al observed 18.9% of patients to go into remission for at least 1 year and 1.8% of patients to achieve remission of at least a 5-year duration when all medications were allowed.12 The slightly higher remission frequencies in these studies may be explained by the allowance of corticosteroid at unlimited dose. Notably, in Zen et al’s purely Caucasian cohort, considerably higher remission frequencies were found.3 About 15.6% of patients had a 5-year remission with corticosteroids up to 5 mg/day and all other treatments allowed.

The question whether the definition of remission should include patients on treatment is a controversy, as prednisone, immunosuppressives and biologicals can cause damage. Corticosteroids are widely used in the management of SLE, but the price patients pay for reliance on prednisone is high. Above 6 mg/day, the risk of later organ damage increases by 50%,22 but even lower doses may have serious side effects including osteoporosis, hyperglycaemia, Cushing syndrome and increased risk of infections.23 Doses above 20 mg/day lead to a fivefold increase in the rate of cardiovascular events after adjustment for age.24 It is not possible to decide which patients on medication would continue in remission when reducing/omitting medications, and in which patients disease activity is just suppressed. Having the treat-to-target approach in mind, it is still of great interest to know which patient characteristics may predict if a patient has a high or low likelihood of going into remission.

Not surprisingly, we found that the level of baseline treatment and baseline disease activity were strongly associated with the time to remission for all definitions. In accordance with this, Steiman et al found that when comparing patients who achieved prolonged remission (≥5 years) to matched and unmatched controls, the former had lower disease activity measured in SLEDAI-2K at various time points of their disease course.4 Nossent et al described that patients who experienced remission had lower corticosteroid doses in the first year after cohort entry.2 Steiman et al also found patients who achieved prolonged remission to have lower prednisone and immunosuppressive drug use and a lower cumulative prednisone dose than controls (measured at the start of their remission period).4 We found that when pooling all patients together, the time to remission depended on the case mix. Thus, these findings are of great importance when the results on remission frequency and on time to remission in different cohorts are compared, or in clinical trials.

Our study identified predictors of remission in multivariable models. In our cohort, African-American ethnicity was associated with lower rates of Remission and ROT. Socioeconomic factors such as family income and education, however, were not significant after adjusting for ethnicity. Conversely to our results, Steiman et al did find no association between the presence of remission and race in a logistic regression model including all potential risk factors.4 They found, however, when comparing patients who achieved prolonged remission (≥5 years) to matched controls, that there were significantly more Caucasian cases among the remission group. Low levels of serum C3 and C4 are part of the Systemic Lupus International Collaborating Clinics (SLICC) classification criteria,25 but on their own show low sensitivity and specificity to predict disease flare.26 Our results show that baseline hypocomplementaemia is associated with a decreased likelihood of achieving any definition of remission. In accordance with Zen et al’s results,3 we identified haematological activity to be associated with a decreased likelihood of achieving remission. Urowitz et al compared a non-remission group to both a 1-year and a 5-year remission group and found anti-dsDNA antibodies were significantly more frequent in the non-remission group.12 We showed that anti-dsDNA at baseline predict against achieving Complete Remission and Complete ROT.

This was one of the largest cohort studies of remission in SLE, and the only cohort study in which patients were followed quarterly by protocol. The cohort represented the epidemiology of SLE regarding its sex distribution and is composed of a broad variety of incomes and educational levels. As patients in this cohort were seen at intervals of 3 months or more regularly and patients with a gap of >9 months were excluded from the analysis, the likelihood of detecting remission was high. However, there were some limitations of our study setting. It was a single-centre cohort, thus the remission parameters (frequency, time to remission and duration) reflected the patients coming to one centre and the treatment strategies used there in the last 30 years. Even though the cohort presented considerable ethnic variation, it lacked large numbers of patients with Asian and Latin American background. Further multicenter studies with different ethnic compositions would be desirable to generalise the results.

Remission is an emerging concept in SLE. Testing definitions of remission is of great relevance for clinical practice, as much as for clinical trials. Our future goal is to find out which definitions are most successful in predicting the best possible outcome for our patients.

References

View Abstract

Footnotes

  • Handling editor Tore K Kvien

  • Contributors TRW contributed to the design of the work, to the analysis and the interpretation of the data and she drafted the work. LSM contributed to the design of the work, developed statistical tools, analysed and interpreted the data and critically revised the manuscript. MP designed the work, acquired the data, interpreted the data and critically revised the manuscript. All the authors approved the final version of the manuscript and gave their agreement to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.

  • Funding The Hopkins Lupus Cohort was funded by the US National Institutes of Health: R0-1 AR 43727.

  • Competing interests None declared.

  • Patient consent Obtained.

  • Ethics approval Patients gave informed consents before taking part in the Lupus Hopkins Cohort. The Hopkins Lupus Cohort is approved on an annual basis by the Johns Hopkins Institutional Review Board.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Linked Articles

  • Clinical and epidemiological research
    Ronald van Vollenhoven Alexandre Voskuyl George Bertsias Cynthia Aranow Martin Aringer Laurent Arnaud Anca Askanase Petra Balážová Eloisa Bonfa Hendrika Bootsma Dimitrios Boumpas Ian Bruce Ricard Cervera Ann Clarke Cindy Coney Nathalie Costedoat-Chalumeau László Czirják Ronald Derksen Andrea Doria Thomas Dörner Rebecca Fischer-Betz Ruth Fritsch-Stork Caroline Gordon Winfried Graninger Noémi Györi Frédéric Houssiau David Isenberg Soren Jacobsen David Jayne Annegret Kuhn Veronique Le Guern Kirsten Lerstrøm Roger Levy Francinne Machado-Ribeiro Xavier Mariette Jamil Missaykeh Eric Morand Marta Mosca Murat Inanc Sandra Navarra Irmgard Neumann Marzena Olesinska Michelle Petri Anisur Rahman Ole Petter Rekvig Jozef Rovensky Yehuda Shoenfeld Josef Smolen Angela Tincani Murray Urowitz Bernadette van Leeuw Carlos Vasconcelos Anne Voss Victoria P Werth Helena Zakharova Asad Zoma Matthias Schneider Michael Ward
  • Clinical and epidemiological research
    Margherita Zen Luca Iaccarino Mariele Gatto Silvano Bettio Francesca Saccon Anna Ghirardello Leonardo Punzi Andrea Doria