LimphoTEC: Ретроспективное исследование реального мира по восстановлению лимфоцитов после лимфопенсии у пациентов с множественным склерозом, обработанных диметилФумаратом во Франции.

Key Summary Points

Why carry out this study?
Dimethyl fumarate (DMF) has demonstrated a favorable benefit–risk profile in the treatment of patients with relapsing–remitting multiple sclerosis (RRMS); however, DMF is associated with decreases in absolute lymphocyte counts (ALCs) and, in clinical trials, ~ 2% of patients experienced prolonged, severe lymphopenia. LymphoTEC was a retrospective, multicenter study of the Observatoire Français de la Sclérose en Plaques (OFSEP) registry, to assess time to ALC reconstitution and time to lymphopenia onset, and to evaluate factors associated with ALC reconstitution and lymphopenia in patients with RRMS treated with DMF.
What was learned from the study?
Overall, ALC decreased by 33% within the first year and reached a plateau, and median time to ALC reconstitution was 3.8 months after discontinuation of DMF; no overall increased risk of serious or opportunistic infections was observed in patients with lymphopenia. Rates of lymphopenia on DMF treatment were consistent with those reported previously; likelihood of lymphopenia increased with age at DMF initiation. ALC reconstitution was more rapid with shorter DMF duration after lymphopenia, therefore supporting the prompt discontinuation of DMF after the onset of lymphopenia.

Introduction

Dimethyl fumarate (DMF; Tecfidera; Biogen, Cambridge, MA, USA), a disease-modifying therapy (DMT), is indicated in patients with relapsing–remitting multiple sclerosis (RRMS) [ 1 ]. Significant, clinically meaningful, sustained efficacy and a favorable benefit–risk profile for DMF have been demonstrated in the pivotal phase 3 DEFINE and CONFIRM studies and in real-world studies [ 2 – 9 ]. However, DMF is associated with decreases in absolute lymphocyte counts (ALCs) [ 10 ]. In clinical trials and an observational study, ALC decreased by 30–40% during the first year of DMF treatment, then plateaued, remaining at or above the lower limit of normal (LLN; 0.91 × 10 9 /L) [ 3 , 11 , 12 ]. An integrated analysis of DMF studies showed a 2.2% rate of severe lymphopenia (ALC < 0.5 × 10 9 /L) that persisted for 6 months among patients treated for ≥ 6 months [ 3 ]. DMF labels recommend monitoring ALC after DMF initiation and considering discontinuation or interruption for patients with prolonged, severe lymphopenia [ 1 ].

Little is known regarding risk factors for lymphopenia in patients receiving DMF treatment. A few studies have looked at predictors of DMF-induced lymphopenia but, due to their small sizes, no conclusions can be drawn [ 13 – 15 ]. A study of 1258 DMF-treated patients from DEFINE and CONFIRM found no genetic predictors of moderate or severe lymphopenia but identified potential relationships with age and lower baseline ALC (there was insufficient power to detect a true difference between patients with and without lymphopenia) [ 16 ]. Thus, the study further supported ALC monitoring to identify at-risk patients.

Data are available concerning lymphocyte reconstitution after DMF discontinuation: most patients who discontinued DMF due to lymphopenia experienced ALC reconstitution within 2–4 months following DMF discontinuation [ 17 ]. Prolonged lymphopenia on DMF is associated with slower lymphocyte reconstitution after DMF discontinuation. Conclusions from studies that have examined potential predictors of delayed ALC recovery are limited due to their small sizes [ 15 , 18 ]. Data indicate that ALCs generally increase following DMF discontinuation; however, lymphocyte reconstitution time varies between patients [ 17 ].

LymphoTEC was a real-world study of records of DMF-treated patients enrolled in the Observatoire Français de la Sclérose en Plaques (OFSEP) cohort, the largest French national multiple sclerosis (MS) cohort (EDMUS database) [ 19 , 20 ]. The study investigated evolution of ALC and frequency, dynamics, and duration of lymphopenia in patients on DMF or after DMF discontinuation. The primary objective was to describe ALC reconstitution after lymphopenia in RRMS patients who discontinued DMF. Secondary objectives were to describe frequency, dynamics, and duration of lymphopenia in patients on DMF or after DMF discontinuation and to identify associations of baseline clinical and demographic factors with ALC reconstitution after DMF discontinuation and with occurrence of lymphopenia.

Methods

Study Design

This was a retrospective, observational, multicenter study of data from the OFSEP ( www.ofsep.org ) database from January 1, 2016, to June 15, 2021. Research was conducted within the framework of the OFSEP database [ 19 , 20 ], in which data were collected retrospectively at the first visit and, thereafter, prospectively at least once a year during routine clinical care using a standardized clinical form. Visits occurred based on routine site and physician practices. Individual case reports included demographic and MS disease characteristics, clinical and magnetic resonance imaging data, and treatment information. Database extraction was performed on June 15, 2021.

Eligibility Criteria

Patients with RRMS were eligible if DMF (Tecfidera; Biogen, Cambridge, MA, USA) was initiated between January 1, 2016, and December 15, 2020, DMF treatment lasted ≥ 3 continuous months, and they had ≥ 1 ALC measurement documented in the 6 months prior to or close to DMF initiation and ≥ 1 ALC measurement documented during DMF treatment.

Data Collection

The following data were collected based on availability in OFSEP: sex, age at MS onset and at DMF initiation, DMT usage before DMF initiation, smoking habits and presence of diabetes at DMF initiation, number of relapses in the prior year, Expanded Disability Status Scale (EDSS) score at DMF initiation, date of DMF discontinuation and reasons for discontinuation, ALCs, and serious and opportunistic infections coded with Medical Dictionary for Regulatory Activities version 24.

Endpoints

The primary endpoint was time to ALC reconstitution (≥ 0.91 × 10 9 /L) after DMF discontinuation. Secondary endpoints included time from DMF initiation to lymphopenia, time from DMF initiation to DMF discontinuation, evolution of ALC after DMF initiation, and occurrence of serious or opportunistic infections.

ALC was characterized using Common Terminology Criteria for Adverse Events version 5.0 (Grade 1: 0.8 × 10 9 /L ≤ ALC < 0.91 × 10 9 /L; Grade 2: 0.5 × 10 9 /L ≤ ALC < 0.8 ×  10 9 /L; Grade 3: 0.2 × 10 9 /L ≤ ALC < 0.5 × 10 9 /L; Grade 4: ALC < 0.2 × 10 9 /L).

Statistical Analysis

Study Populations and General Considerations

The full analysis set (FAS) included all patients who met the eligibility criteria. Population A included FAS patients who developed lymphopenia, population B included FAS patients who developed lymphopenia and discontinued DMF, and population C included FAS patients who discontinued DMF for any reason. Statistical analyses were performed with SAS 9.4 software (SAS Institute, Cary, NC, USA) and R 4.0.5 software (with survival package; R Foundation for Statistical Computing, Vienna, Austria). Data were summarized using descriptive analysis including medians and interquartile ranges (IQRs).

Survival Analyses and Evolution of ALC

Time to ALC reconstitution, time to lymphopenia, and time to DMF discontinuation (survival analyses) were estimated using the Kaplan–Meier method. Cumulative incidence curves with two-sided 95% confidence intervals (CIs; Greenwood formula for estimation of standard error) are displayed. For ALC reconstitution after DMF discontinuation, data were censored at the time of new DMT initiation after DMF discontinuation or the last available ALC, whichever event occurred first. For the first occurrence of lymphopenia and monthly ALC decreases, time was included in a linear mixed model with an individual-level random effect and ALC measures as the dependent variable; data were censored at DMF discontinuation or at the last determination of ALC, whichever event occurred first. For DMF discontinuation, data were censored at the last clinical evaluation.

Mean ALC evolution was approximated by a smoothed function resulting from a nonparametric regression (locally estimated scatterplot smoothing regression) with two-sided 95% CI. Data were censored 24 months after DMF initiation.

The cumulative probabilities of the first occurrence of lymphopenia stratified by grade and cumulative probabilities of DMF discontinuation stratified by reasons were estimated with the Aalen–Johansen estimator.

Infections

Total numbers of serious and opportunistic infections and unadjusted rates per 100 patient-years are displayed with 95% CIs, assuming a Poisson distribution. Rates were calculated by dividing the total number of serious adverse events (SAEs) by the exposure period. The DMF exposure period was defined as the time from DMF initiation until either the last clinical evaluation or DMF discontinuation, whichever occurred first; the exposure period after discontinuation was defined as the time from DMF discontinuation until either the last clinical evaluation or the next DMT, whichever occurred first.

Risk Factors

For analyses of factors associated with ALC reconstitution after DMF discontinuation or lymphopenia after DMF initiation, all covariates with p  < 0.20 in univariate analyses were included in multivariate analyses. A backward stepwise selection was performed, and the final model included all factors. Hazard ratios (HRs) and two-sided 95% CIs were estimated using a Cox model. Factors studied are provided in Supplementary Materials.

Ethics

In accordance with French legislation, the OFSEP cohort was approved by the national data protection agency (Commission Nationale de l’Informatique et des Libertés; authorization request 914066v3) and the French ethical committee (Comité de Protection des Personnes: reference 2019-A03066-51). All patients in the OFSEP cohort provided informed consent to have their medical data shared anonymously for research purposes. Appropriate research information on the OFSEP registry research cohort is available on the OFSEP website [ 19 , 20 ]. After 1 month without objections from the patient, patient data were extracted for analysis. As per French legislation (“Loi Informatique et Libertés” No. 78–17 on January 6, 1978, modified by Law No. 2016–41 on January 26, 2016), LymphoTEC was registered on ClinicalTrials.gov ( NCT04756687 ) and was compliant with the “French Methodology of reference MR-004”. The study was conducted in accordance with the Declaration of Helsinki.

Results

Population Description

Overall, 1429 patients with RRMS were included in the FAS (Fig.  1 ); 356 developed lymphopenia (population A) and 599 discontinued DMF (population C) (Fig.  2 ). A total of 183 patients with lymphopenia discontinued DMF (population B).

Flowchart of the study population. ALC absolute lymphocyte count, DMF dimethyl fumarate, MS multiple sclerosis, OFSEP Observatoire Français de la Sclérose en Plaques

Study populations. DMF dimethyl fumarate, FAS full analysis set. a The FAS includes 620 non-lymphopenic patients treated with DMF, 30 patients with lymphopenia before DMF initiation, and 7 patients with lymphopenia before DMF treatment but with lymphocyte reconstitution before or close to DMF initiation

Patient characteristics of the FAS and population B are displayed in Table  1 . Of note, the first ALC close to DMF initiation could occur before or after DMF initiation. Overall, 754 patients had their first ALC in the 6 months preceding DMF initiation {median [IQR] (max) time between first ALC and DMF initiation: 0.6 [0.1–1.4] (5.8) months}, and 675 patients had their first ALC after DMF initiation {median [IQR] (max) time between DMF initiation and first ALC: 2.0 [0.7–5.7] (42.4) months}.

Table 1 Patient characteristics of the FAS and population B Overall FAS ( n  = 1429) Lymphopenic population with discontinuation (population B; n  = 183) Female 1030 (72.1) 128 (70.0) Age at MS onset, years 29.7 (23.8–37.6) 32.9 (26.7–41.6) Age at DMF initiation, years 36.1 (28.9–44.7) 42.7 (33.6–51.9) MS duration at DMF initiation, years 2.9 (0.6–9.1) 5.7 (1.0–13.9) Duration of DMF at discontinuation, years 1.7 (0.9–2.9) 1.3 (0.8–2.2) Diabetes at DMF initiation  Yes 38 (2.7) 9 (4.9)  No 1054 (73.8) 129 (70.5)  Unknown 337 (23.6) 45 (24.6) Smoking habits at DMF initiation  Never 629 (44.0) 87 (47.5)  Current 244 (17.1) 25 (13.7)  Former 93 (6.5) 13 (7.1)  Unknown 463 (32.4) 58 (31.7)  Treatment naive before DMF 806 (56.4) 94 (51.4) Number of relapses in prior year  1 515 (36.0) 50 (27.3)  ≥ 2 914 (64.0) 133 (72.7) EDSS at DMF initiation  0.0–3.0 883 (61.8) 103 (56.3)  3.5–9.5 126 (8.8) 24 (13.1)  Unknown 420 (29.4) 56 (30.6) First ALC close to DMF initiation (× 10 9 /L) a 1.9 (1.5–2.4) 1.6 (1.0–2.1)  No lymphopenia 1349 (94.4) 154 (84.2)  Grade 1 27 (1.9) 6 (3.3)  Grade 2 45 (3.2) 22 (12.0)  Grade 3 6 (0.4) 1 (0.6)  Grade 4 2 (0.1) 0 (0.0) Last ALC under DMF (× 10 9 /L) 1.5 (1.0, 2.0) 0.7 (0.6, 0.8)  No lymphopenia 1172 (82.0) 25 (13.7)  Grade 1 77 (5.4) 33 (18.0)  Grade 2 153 (10.7) 100 (54.6)  Grade 3 27 (1.9) 25 (13.7)  Grade 4 0 (0.0) 0 (0.0) Data are presented as n (%) or median (IQR) ALC absolute lymphocyte count, DMF dimethyl fumarate, EDSS Expanded Disability Status Scale, FAS full analysis set, MS  multiple sclerosis a Within 6 months before or after DMF initiation, in patients whose first ALC was after DMF initiation, median (range) time to first ALC was 2.0 (0.03–42.4) months in the FAS and 2.7 (0.03–23.7) months in population B

Patient characteristics of the FAS and population B

Data are presented as n (%) or median (IQR)

ALC absolute lymphocyte count, DMF dimethyl fumarate, EDSS Expanded Disability Status Scale, FAS full analysis set, MS  multiple sclerosis

a Within 6 months before or after DMF initiation, in patients whose first ALC was after DMF initiation, median (range) time to first ALC was 2.0 (0.03–42.4) months in the FAS and 2.7 (0.03–23.7) months in population B

Occurrence of First Lymphopenia After DMF Initiation

Patients who had lymphopenia before DMF initiation ( n  = 30; 2.2%) were not considered in this analysis, which, therefore, included 1399 patients. In the FAS, ALC decreased by 33% within the first year on DMF (mean [95% CI]: 2.11 [2.08–2.15] × 10 9 cells/L to 1.41 [1.38–1.44] × 10 9 cells/L) and thereafter reached a plateau. The cumulative probability of developing lymphopenia was 7.8% (95% CI 6.3–9.2) at 6 months (Fig.  3 ; Table  2 ); corresponding probabilities classified by grade of lymphopenia were 3.7% (Grade 1), 3.7% (Grade 2), and 0.4% (Grade 3). At 12 months, the cumulative probability of developing lymphopenia was 18.2% (95% CI 16.1–20.3); classified by grade, probabilities were 8.0% (Grade 1), 9.6% (Grade 2), and 0.6% (Grade 3). First lymphopenia after DMF initiation was classified as Grade 1 in 173 patients, Grade 2 in 173 patients, and Grade 3 in 10 patients (Supplementary Material Table S1 ). No first lymphopenia event was Grade 4. Maximal lymphopenia grade was Grade 2 in 57.9% ( n  = 206), Grade 1 in 28.1% ( n  = 100), Grade 3 in 13.8% ( n  = 49), and Grade 4 in < 1% ( n  = 1) of the 356 lymphopenic patients (population A).

Cumulative probabilities of occurrence of the first lymphopenia after DMF initiation (FAS). Thirty patients developed lymphopenia before DMF initiation and were not considered from the analysis. DMF dimethyl fumarate, FAS full analysis set

Table 2 Estimates of cumulative probabilities of the first lymphopenia after DMF initiation by grade (FAS) Time, months P Grade1 , % 95% CI P Grade2 , % 95% CI P Grade3 , % 95% CI 6 3.7 2.8–4.9 3.7 2.8–4.8 0.4 0.2–0.9 12 8.0 6.6–9.6 9.6 8.1–11.4 0.6 0.3–1.2 Estimates of probabilities were evaluated in the FAS ( n  = 1429) CI confidence interval, DMF dimethyl fumarate, FAS full analysis set, P Grade1 probability of the first lymphopenia being Grade 1, P Grade2 probability of the first lymphopenia being Grade 2, P Grade3 probability of the first lymphopenia being Grade 3

Estimates of cumulative probabilities of the first lymphopenia after DMF initiation by grade (FAS)

3.7

2.8–4.9

3.7

2.8–4.8

0.4

0.2–0.9

8.0

6.6–9.6

9.6

8.1–11.4

0.6

0.3–1.2

Estimates of probabilities were evaluated in the FAS ( n  = 1429)

CI confidence interval, DMF dimethyl fumarate, FAS full analysis set, P Grade1 probability of the first lymphopenia being Grade 1, P Grade2 probability of the first lymphopenia being Grade 2, P Grade3 probability of the first lymphopenia being Grade 3

Based on a linear mixed model, ALC decreased by 2.9% (95% CI 1.9–3.9) per month from first ALC after DMF initiation in patients with Grade 1 lymphopenia who had stopped treatment, 2.3% (95% CI 1.4–3.2) per month in patients with Grade 1 lymphopenia and who were ongoing on DMF at the last visit, 4.8% (95% CI 3.5–6.2) per month in patients with Grade > 1 lymphopenia who had stopped treatment, and 2.2% (95% CI 1.4–3.0) per month in patients with Grade 1 lymphopenia and who were ongoing on DMF. Mean evolution of ALC by grade and DMF status from DMF initiation to last observed ALC on DMF is shown in Supplementary Material Fig. S1 . Individual ALCs over time for patients with lymphopenia who remained on DMF are shown in Supplementary Material Fig. S2 .

ALC Reconstitution After DMF Discontinuation

Median time to ALC reconstitution after DMF discontinuation in population B was 3.8 (95% CI 3.0–6.3) months (Fig.  4 ); 25 patients were not considered because ALC reconstitution occurred before DMF discontinuation. At 12 months, 83.0% (95% CI 70.8–90.1) of the 158 patients included in the analysis had achieved ALC reconstitution.

Cumulative probabilities of the first ALC reconstitution in the lymphopenic population with DMF discontinuation (population B). Twenty-five patients were not considered from the analysis because ALC reconstitution occurred before DMF discontinuation. The dotted lines represent 95% Cis. ALC absolute lymphocyte count, CI confidence interval, DMF dimethyl fumarate

Median (95% CI) time to reconstitution after DMF discontinuation was 4.0 (3.2–11.2) months for the 33 patients with Grade 1 lymphopenia, 3.0 (2.2–5.5) months for the 100 patients with Grade 2 lymphopenia, and 9.7 (4.2–not estimable) months for the 25 patients with Grade 3 lymphopenia at time of discontinuation (log-rank test; p  = 0.127; Supplementary Material Fig. S3 ).

DMF Discontinuation

Among patients who discontinued DMF (population C), median time to discontinuation was 1.2 (IQR: 0.6–2.0) years. Cumulative probabilities of discontinuation were 18.5% (95% CI 16.4–20.6) at 12 months, 36.3% (95% CI 33.5–38.9) at 24 months, and 57.8% (95% CI 53.9–61.4) at 48 months.

After 48 months, the main reasons for discontinuation were lack of efficacy (22.5%) and intolerance (22.0%) (Fig.  5 ). Lymphopenia was categorized under biological intolerance; however, the database did not separate lymphopenia into a discrete reason for discontinuation and, therefore, lymphopenia alone could not be confirmed as a cause of discontinuation. After DMF discontinuation, 462 (77.1%; population C) patients initiated a new treatment.

Cumulative probability of DMF discontinuation by reason (FAS). DMF dimethyl fumarate, FAS full analysis set, SAE serious adverse event

Infections

Overall, 23 patients (1.7%) presented a total of 24 serious or opportunistic infections on DMF including 8 SARS-COV2 infections, 2 shingles, 1 urinary tract infection, 1 herpes virus infection, 1 hepatitis infection, and 11 other infections. The incidence rate of serious or opportunistic infections on DMF were 0.8 per 100 person-years (95% CI 0.6–1.3); corresponding rates were 3.4 per 100 person-years (95% CI 1.8–6.5) after DMF discontinuation (population C) and 0.68 per 100 person-years (95% CI 0.32–1.42) after first occurrence of lymphopenia after DMF initiation in patients with lymphopenia (population A).

No cases of progressive multifocal leukoencephalopathy (PML) were recorded in the study database.

Risk Factors

Factors Associated With ALC Reconstitution After DMF Discontinuation (Population B)

The univariate analysis showed that diabetes, treatment before DMF, DMF duration, DMF duration until lymphopenia, DMF duration after lymphopenia, and grade of lymphopenia could be introduced into the multivariate model ( p  < 0.2; Table  3 ).

Table 3 Univariate analysis of associations between clinical and biological characteristics at DMF discontinuation and the first ALC reconstitution after DMF discontinuation (population B) Variable HR 95% CI p value Age at DMF discontinuation, years 0.289  < 35 – –  35 to < 45 1.06 0.60–1.90  45 to < 55 0.59 0.30–1.18  ≥ 55 0.95 0.50–1.80 Sex 0.288  Male – –  Female 0.77 0.48–1.25 Diabetes 0.024  No – –  Yes 5.34 1.89–15.12 0.001  Unknown 0.84 0.51–1.38 0.487 Smoking habits 0.613  Never – –  Current 1.06 0.52–2.16  Former 1.73 0.80–3.78  Unknown 1.01 0.61–1.67 Last ALC during DMF 2.56 0.58–11.37 0.215 EDSS at DMF discontinuation 0.515  0.0–3.0 – –  3.5–9.5 1.26 0.64–2.49  Unknown 0.80 0.46–1.40 Number of relapses in the year before DMF initiation 0.263  1 – –  ≥ 2 1.35 0.80–2.29 Disease duration at DMF discontinuation, years 0.449  0 to < 3 – –  3 to < 5 0.76 0.28–2.01  5 to < 10  ≥ 10 1.40 0.82 0.73–2.69 0.50–1.36 Treatment before DMF 0.197  No – –  Yes 0.74 0.48–1.16 DMF duration, years 0.027  0 to < 1 – –  1 to < 2 0.48 0.28–0.84 0.010  2 to < 3 1.01 0.57–1.80 0.957  ≥ 3 1.09 0.48–2.49 0.834 DMF duration up to lymphopenia, months 1.02 0.99–1.04 0.103 DMF duration after lymphopenia, months 0.087  0 to < 0.1 – –  0.1 to < 0.5 1.12 0.62–2.03  0.5 to < 1.5 1.23 0.65–2.31  ≥ 1.5 0.47 0.23–0.96 Grade of lymphopenia 0.111  Grade 1 – –  Grade 2 1.32 0.75–2.32  Grade 3 0.68 0.32–1.47 ALC absolute lymphocyte count, CI confidence interval, DMF dimethyl fumarate, EDSS Expanded Disability Status Scale, HR hazard ratio

Univariate analysis of associations between clinical and biological characteristics at DMF discontinuation and the first ALC reconstitution after DMF discontinuation (population B)

5 to < 10

≥ 10

1.40

0.82

0.73–2.69

0.50–1.36

ALC absolute lymphocyte count, CI confidence interval, DMF dimethyl fumarate, EDSS Expanded Disability Status Scale, HR hazard ratio

After adjustment, multivariate analysis showed the first ALC reconstitution after DMF discontinuation in patients with lymphopenia was associated with presence of diabetes, DMF duration (1 to < 2 years), DMF duration up to lymphopenia, and DMF duration after lymphopenia (0.1 to < 0.5 months; and 0.5 to < 1.5 months; Table  4 ). Thus, in patients with lymphopenia who discontinued DMF, the longer the time to lymphopenia onset, the shorter the time to ALC reconstitution (i.e., for every additional month of DMF treatment before lymphopenia occurrence, the likelihood of ALC reconstitution increased by 6%).

Table 4 Multivariate analysis of associations between clinical and biological characteristics at DMF discontinuation and the first ALC reconstitution after DMF discontinuation (population B) Variable HR 95% CI p value Diabetes 0.003  No – –  Yes 9.23 2.88–29.57  < 0.001  Unknown 0.71 0.42–1.20 0.201 DMF duration, years 0.002  0 to < 1 –  1 to < 2 0.27 0.12–0.62 0.002  2 to < 3 0.41 0.09–1.85 0.248  ≥ 3 0.15 0.01–1.53 0.109 DMF duration up to lymphopenia, months 1.06 1.00–1.13 0.034 DMF duration after lymphopenia, days 0.043  0 to < 3 1.74 0.75–4.02 0.197  3 to < 15 2.45 1.06–5.64 0.035  15 to < 45 2.86 1.31–6.24 0.008  ≥ 45 – The multivariate model selected was the one including all covariates with a p value < 0.2 in univariate analysis, which minimizes the AIC AIC Akaike information criterion, ALC absolute lymphocyte count, CI confidence interval, DMF dimethyl fumarate, HR hazard ratio

Multivariate analysis of associations between clinical and biological characteristics at DMF discontinuation and the first ALC reconstitution after DMF discontinuation (population B)

The multivariate model selected was the one including all covariates with a p value < 0.2 in univariate analysis, which minimizes the AIC

AIC Akaike information criterion, ALC absolute lymphocyte count, CI confidence interval, DMF dimethyl fumarate, HR hazard ratio

Risk Factors Associated With Lymphopenia After DMF Initiation (FAS)

The univariate analysis showed that age at DMF initiation, first ALC close to DMF initiation, number of relapses in the year before DMF initiation, disease duration at DMF initiation, and treatment before DMF were associated with lymphopenia development (Table  5 ) and could be introduced into the multivariate model selection.

Table 5 Univariate analysis of associations between clinical and biological characteristics at DMF initiation and the first occurrence of lymphopenia after DMF initiation (FAS) Variable HR 95% CI p value Age at DMF initiation, years < 0.001  < 25 0.35 0.21–0.56 < 0.001  25 to < 35 0.71 0.53–0.93 0.015  35 to < 45 – –  45 to < 55 1.26 0.96–1.67 0.101  ≥ 55 2.61 1.86–3.65 < 0.001 Sex 0.884  Male – –  Female 1.02 0.81–1.28 Diabetes 0.203  No –  Yes 1.68 0.98–2.88 0.059  Unknown 1.07 0.84–1.36 0.600 Smoking habits 0.473  Never – –  Current 0.79 0.58–1.09 0.150  Former 0.96 0.62–1.47 0.849  Unknown 1.00 0.80–1.27 0.955 First ALC close to DMF initiation 0.31 0.26–0.38 < 0.001 EDSS 0.584  0.0–3.0 – –  3.5–9.5 1.21 0.85–1.71 0.289  Unknown 1.03 0.81–1.30 0.787 Number of relapses in the prior year 0.004  1 – –  ≥ 2 1.39 1.10; 1.74 0.005 Disease duration at DMF initiation, years < 0.001  0 to < 3 – –  3 to < 5 0.90 0.60–1.34 0.594  5 to < 10 0.94 0.69–1.30 0.719  ≥ 10 1.77 1.39–2.24 < 0.001 Treatment before DMF 0.009  No – –  Yes 1.32 1.07–1.62 ALC absolute lymphocyte count, CI confidence interval, DMF dimethyl fumarate, EDSS Expanded Disability Status Scale, HR hazard ratio

Univariate analysis of associations between clinical and biological characteristics at DMF initiation and the first occurrence of lymphopenia after DMF initiation (FAS)

ALC absolute lymphocyte count, CI confidence interval, DMF dimethyl fumarate, EDSS Expanded Disability Status Scale, HR hazard ratio

After adjustment from the univariate analysis, multivariate analysis showed the first occurrence of lymphopenia after DMF initiation was associated with age (45 to < 55 years; and ≥ 55 years), and ALC at DMF initiation (Table  6 ).

Table 6 Multivariate analysis of associations between clinical and biological characteristics at DMF initiation and the occurrence of the first lymphopenia after DMF initiation (FAS) Variable HR 95% CI p value Age at DMF initiation, years < 0.001  < 25 0.37 0.23–0.59 < 0.001  25 to < 35 0.79 0.59–1.04 0.092  35 to < 45 – –  45 to < 55 1.34 1.01–1.77 0.040  ≥ 55 2.32 1.65–3.25 < 0.001 First ALC close to DMF initiation 0.33 0.27–0.40 < 0.001 The multivariate model selected was the one including all covariates with a p value < 0.2 in univariate analysis which minimizes the AIC AIC Akaike information criterion, ALC absolute lymphocyte count, CI confidence interval, DMF dimethyl fumarate, HR hazard ratio

Multivariate analysis of associations between clinical and biological characteristics at DMF initiation and the occurrence of the first lymphopenia after DMF initiation (FAS)

The multivariate model selected was the one including all covariates with a p value < 0.2 in univariate analysis which minimizes the AIC

AIC Akaike information criterion, ALC absolute lymphocyte count, CI confidence interval, DMF dimethyl fumarate, HR hazard ratio

Discussion

Data for this study were obtained from the OFSEP database, comprising more than 50% of identified MS cases in France [ 20 ]. To our knowledge, LymphoTEC is the largest real-world study to describe ALC changes in patients with MS during DMF treatment and after DMF discontinuation.

Median time to ALC reconstitution after DMF discontinuation was 3.8 months in patients with lymphopenia (population B), consistent with the time to ALC reconstitution (≥ 0.8 × 10 9 /L) of 2–4 months after DMF discontinuation described in a previous real-world study [ 17 ]. Shorter times to reconstitution of 2 months have been observed in clinical trials [ 4 , 17 ]. This may be due to differences in data collection timepoints between clinical trials and real-world studies: clinical trials have protocol-determined, scheduled intervals for data collection, whereas real-world studies have data collection intervals that vary and are based on the physician’s decision.

Rates of lymphopenia were similar to those reported in clinical trials and real-world studies [ 3 , 11 , 12 ]. Patients in LymphoTEC had a mean decrease in ALC of 33% from DMF initiation to Year 1, before reaching a plateau. The probability of developing lymphopenia was 18.2% at Year 1. Similar to an integrated analysis of phase 2b and 3 DMF studies which reported a 21% incidence of Grade 2 lymphopenia [ 3 ], patients in LymphoTEC who developed lymphopenia were more likely to have Grade 2 lymphopenia than either Grades 1 or 3.

First documented ALC after DMF initiation occurred after a median of 2 months and occurred within 6 months in 75% of patients. As stated in the Summary of Product Characteristics, a complete blood count must be performed prior to treatment initiation, and low ALCs require caution if treatment is initiated [ 1 ].

Multivariate analysis in LymphoTEC indicated that the greater the patient’s age at initiation, the higher the likelihood of lymphopenia; risk was doubled for patients aged > 55 years versus those aged 35 to < 45 years. Lower ALC at baseline was associated with development of lymphopenia. Both of these findings corroborated those of a smaller retrospective study [ 21 ].

Overall, 599 (41.9%) patients discontinued DMF (population C), with median time to discontinuation of 1.2 years. The main reasons for discontinuation were intolerance (general, local, or biological) and lack of efficacy. As lymphopenia was categorized under biological intolerance, it was not possible to determine whether discontinuations in the lymphopenic population were due to lymphopenia. The probability of discontinuing DMF after lymphopenia in the FAS was 57.8% at 4 years.

Consistent with earlier data [ 17 ], the probability of ALC reconstitution 1 year after DMF discontinuation was numerically higher for Grade 1 than for Grade 3 lymphopenia (85.4% vs. 61.3%), although not significantly so, and the median time to reconstitution in patients with prolonged lymphopenia was longer than in patients with lymphopenia for < 6 months. This indicates that to achieve good reconstitution, DMF must be stopped quickly after lymphopenia. One report suggested that, rather than ALC at discontinuation, duration of lymphopenia while on DMF was associated with a lower chance of ALC reconstitution [ 17 ]. Furthermore, patients with longer durations and greater severity of lymphopenia during DMF treatment experienced extended times to ALC reconstitution [ 17 ]. In LymphoTEC, patients with diabetes were found to have faster ALC reconstitution. However, diabetes status was missing for almost 25% of patients.

Earlier reports indicated that lower ALC at DMF discontinuation, longer disease duration, and prior exposure to MS treatments were predictive of delayed ALC recovery in patients who discontinued DMF and achieved ALC reconstitution after prolonged lymphopenia [ 15 ]. Multivariate analysis in LymphoTEC showed associations between more rapid ALC reconstitution after DMF discontinuation and presence of diabetes, longer duration of DMF treatment before lymphopenia, and shorter DMF treatment duration after lymphopenia. The probability of reconstitution was almost tripled when DMF was stopped between 3 days (0.1 months) and 1.5 months after lymphopenia onset compared with those who stopped ≥ 1.5 months after. This further supports the prompt discontinuation of DMF after onset of lymphopenia in order to achieve reconstitution. These associations with ALC reconstitution have not been previously reported; further studies are needed to corroborate these factors as predictors of ALC reconstitution after DMF discontinuation.

DMF-induced lymphopenia could increase the risk of opportunistic infections, including the rare and serious viral infection, PML. As of December 31, 2021, there were a total of 12 cases of PML in > 560,000 DMF-treated patients with MS, representing > 1,190,000 person-years of exposure worldwide; most cases (9/12) were in the setting of prolonged (> 6 months), moderate (≥ 0.5 × 10 9 /L to < 0.8 × 10 9 /L) to severe (< 0.5 × 10 9 /L) lymphopenia, while 3/12 cases occurred in the setting of mild, non-prolonged (< 6 months) lymphopenia [ 22 ]. In LymphoTEC, there were no cases of PML reported and no increased risks of serious or opportunistic infections during DMF treatment. Furthermore, no increases in the rates of serious or opportunistic infections in lymphopenic patients were observed. However, most participants who developed lymphopenia reached only mild or moderate grades (Grades 1 or 2) and discontinuation of DMF was not delayed for longer than a few months after the occurrence of lymphopenia. This context is important since our observations regarding infection rates in lymphopenic patients may not be applicable to cases of severe lymphopenia.

Limitations include that no ALC was documented before DMF initiation in 675 patients, and therefore lymphopenia prior to DMF in those patients cannot be ruled out. Indeed, 30 patients already had lymphopenia documented prior to DMF initiation. However, as the first documented ALC could occur well after DMF initiation (≤ 3.5 years), time to lymphopenia was likely overestimated. In addition, after DMF discontinuation, data were censored at the time of initiation of a new DMT ( n  = 43) or the last available ALC ( n  = 35), whichever occurred first, which may have shortened the evaluation period for ALC reconstitution. A further limitation was the small sample size of patients used in the analysis of infection rates. However, our results align with other reports showing no association between lymphopenia and an overall increased risk of serious infections, including opportunistic infections [ 9 , 17 ], and that no relationship was found between infections and the lymphopenia grade [ 3 ]. Our analyses also did not assess the potential influence of comorbidities (aside from diabetes) or types of post-DMF treatments on ALC reconstitution.

Conclusions

These results provide real-world data on lymphopenia in patients with MS who received DMF and support previous reports of DMF-induced lymphopenia, indicating a 33% decrease in ALC in the first year before plateauing. The probability of developing lymphopenia was 18.2% at Year 1, and median time to ALC reconstitution was 3.8 months after DMF discontinuation, although patients with shorter and milder lymphopenia had faster reconstitution times. Lymphopenia was not associated with an overall increase in the risk of serious or opportunistic infectious diseases, and no cases of PML were reported.

Supplementary Information

Below is the link to the electronic supplementary material.

Supplementary file1 (PDF 325 KB)

Acknowledgements

The authors thank the patients, investigator sites, and the Observatoire Français de la Sclérose en Plaques (OFSEP) for their participation in this study, without whom this effort would not succeed.

Medical Writing/Editorial Assistance

Biogen provided funding for medical writing and editorial support in the development of this manuscript. Medical writing support was provided by Meryl Mandle, BS, and David Pertab, PhD, from Excel Scientific Solutions, and copyediting was provided by Cara Farrell from Excel Scientific Solutions.

Author Contributions

Jérôme de Sèze, Pierre Labauge, Roland Liblau, Mikel Martinez, Thibault Moreau, Laurent Suchet, Patrick Vermersch, Sandra Vukusic, Guillaume Mathey, Laure Michel, Jonathan Ciron, Aurelie Ruet, Elisabeth Maillart, Helene Zephir, Caroline Papeix, Gilles Defer, Mikael Cohen, David Axel Laplaud, Eric Berger, Pierre Clavelou, Eric Thouvenot, Olivier Heinzlef, Jean Pelletier, Claire Giannesini, Olivier Casez, Bertrand Bourre, Abir Wahab, Laurent Magy, Jean-Philippe Camdessanché, Inès Doghri, Céline Labeyrie, Karolina Hankiewicz, Jean-Philippe Neau, Corinne Pottier, Pamela Dobay, and Seth Levin: Investigation; Writing – review and editing; Marilyn Gros, Marta Ruiz: Conceptualization; formal analysis; investigation; methodology; writing – review and editing; Hanyue Li: Data curation; formal analysis; writing – review and editing; Fabien Rollot: Data curation; formal analysis; investigation; writing – review and editing.

Funding

Data collection has been supported by a grant provided by the French State and handled by the Agence Nationale de la Recherche, within the framework of the France 2030 program, under the reference ANR-10-COHO-002, Observatoire Français de la Sclérose en Plaques (OFSEP). The Eugène Devic EDMUS Foundation Against Multiple Sclerosis, Lyon, France provided support with site coordination, data management, and statistical analysis. Biogen France SAS (Paris, France) sponsored the LymphoTEC study, provided funding for medical writing and editorial support in the development of this paper, and provided funding for the Rapid Service Fee.

Data Availability

OFSEP is the sole proprietor of the study database, and access to data is governed by the rules and guidelines of the steering committee. Patient-level data for the current study were neither directly accessible nor transferred to the sponsor.

Declarations

Conflict of Interest

Jérôme de Sèze: consultant, advisory boards, and scientific meetings for Alexion, Biogen, CSL Behring, Genzyme/Sanofi, Horizon Therapeutics, LFB, Janssen, Novartis, Roche, and Teva. Pierre Labauge: congresses and research grants from Biogen, BMS, Merck, Novartis, Sanofi Genzyme, and Teva. Roland Liblau: research grants from BMS, GSK, and Roche; speaker fees from Biogen, Merck, Novartis, Roche, and Sanofi Genzyme. Mikel Martinez: no conflicts of interest to disclose. Thibault Moreau: research support and consultancy fees from Biogen, Merck, Novartis, Roche, and Sanofi. Laurent Suchet: consulting fees from Biogen, BMS-Celgene, Merck, Novartis, Roche, and Sanofi Genzyme; honoraria from Novartis and Sanofi. Patrick Vermersch: honoraria from AB Science, Ad Scientiam, Biogen, BMS-Celgene, Imcyse, Janssen, Merck, Novartis, Roche, Sanofi Genzyme, and Teva; research support from Novartis, Roche, and Sanofi Genzyme. Sandra Vukusic: lecturing fees, travel grants, and research support from Biogen, BMS-Celgene, Janssen, Merck, Novartis, Roche, Sandoz, and Sanofi Genzyme. Guillaume Mathey: no conflicts of interest to disclose. Laure Michel: no conflicts of interest to disclose. Jonathan Ciron: consulting, serving on a scientific advisory board, speaking, or other activities with Alexion, Biogen, Horizon Therapeutics, Novartis, Merck, Roche, and Sanofi Genzyme. Aurelie Ruet: received honoraria for meeting speaking from Alexion, Horizon Therapeutics, Merck, and Sanofi Genzyme; received support for traveling from Biogen, Merck, and Novartis. Her institution received research grants from Biogen, BMS, Roche, and Sanofi Genzyme. Elisabeth Maillart: consulting and lecturing fees from Alexion, Biogen, Horizon, Janssen, Merck Serono, Novartis, Roche, Sandoz, Sanofi Genzyme, and Teva Pharmaceuticals; and research support from Biogen. Helene Zephir: consulting or lectures, and invitations for national and international congresses, from Bayer, Biogen, Merck, Novartis, Sanofi Genzyme, and Teva; research support from Roche and Teva; and academic research grants from Académie de Médecine, ARSEP Foundation, FHU Imminent, and LFSEP. Caroline Papeix: no conflicts of interest to disclose. Gilles Defer: consulting and lecturing fees for Biogen, Genzyme, Merck Serono, Novartis, Roche, and Teva; and funding for travel from Biogen, Merck Serono, Novartis, Sanofi Genzyme, and Teva. Institution granted for research supporting from Biogen, Genzyme, Merck Serono, and Novartis. Mikael Cohen: has received personal compensation for consulting, serving on a scientific advisory board, speaking, or other activities with Ad Scientiam, Alexion, Biogen, Celgene-BMS, Horizon Therapeutics, Janssen, Merck, Roche, and Sanofi. David Axel Laplaud: served on scientific advisory boards for Alexion, Biogen, BMS, Janssen, Merck, Novartis, Roche, and Sanofi; received conference travel support and/or speaker honoraria from Alexion, Biogen, BMS, Merck, Novartis, Roche, and Sanofi; and received research support from Agence Nationale de la Recherche, Fondation ARSEP, and Fondation EDMUS.​ Eric Berger: honoraria and consulting fees from Biogen, Genzyme, Novartis, Roche, Sanofi Aventis, and Teva Pharma. Pierre Clavelou: consulting and lecturing fees, travel grants, and unconditional research support from Biogen, Janssen, Medday, Merck, Novartis, Roche, Sanofi Genzyme, and Teva Pharma. Eric Thouvenot: consulting and lecturing fees, travel grants, or unconditional research support from Actelion, Biogen, Janssen, Merck Serono, Novartis, Roche, and Sanofi. Olivier Heinzlef: consulting and lecturing fees from Almirall, Bayer Schering, Biogen Idec, Genzyme, Merck, Novartis, and Teva; travel grants from Biogen Idec, Genzyme, Merck Serono, Novartis, and Teva; and research support from Roche, Merck, and Novartis. Jean Pelletier: no conflicts of interest to disclose. Claire Giannesini: no conflicts of interest to disclose. Olivier Casez: Biogen (personal fees), Janssen (personal fees), Merck (personal fees, non-financial support [traveling/congress]), Novartis (personal fees, non-financial support [traveling/congress]), Roche (personal fees, non-financial support [traveling/congress]), and Sanofi (personal fees). Bertrand Bourre: served on scientific advisory board for Alexion, Biogen, BMS, Horizon, Janssen, Merck, Novartis, Roche, Sandoz, and Sanofi; and received funding for travel and honoraria from Alexion, Janssen, Merck, Novartis, Roche, and Sanofi. Abir Wahab: provided expert testimony for Roche and received travel grants from Biogen. Laurent Magy: no conflicts of interest to disclose. Jean-Philippe Camdessanché: consulting and lecturing fees from Akcea, Alexion, Alnylam, Argenx, Biogen, Bristol Myers Squibb, CSL Behring, Genzyme, Grifols, Laboratoire Français des Biotechnologies, Merck Serono, Natus, Novartis, Pfizer, Pharmalliance, SNF-Floerger, Teva, and UCB Pharma; travel grants from Akcea, Alexion, Alnylam, Argenx, Biogen, CSL Behring, Genzyme, Grifols, Laboratoire Français des Biotechnologies, Merck Serono, Natus, Novartis, Pfizer, SNF-Floerger, and Teva. Inès Doghri: no conflicts of interest to disclose. Céline Labeyrie: no conflicts of interest to disclose. Karolina Hankiewicz: no conflicts of interest to disclose. Jean-Philippe Neau: no conflicts of interest to disclose. Corinne Pottier: support from Alexion, Biogen, Merck, Novartis, Roche, and Teva. Pamela Dobay, Seth Levin, Marilyn Gros, Marta Ruiz: employees of Biogen and may own stock in the company. Hanyue Li: employee of Biogen and may have owned stock in the company at the time this work was conducted. Fabien Rollot: no conflicts of interest to disclose.

Ethical Approval

In accordance with French legislation, the OFSEP cohort was approved by the national data protection agency (Commission Nationale de l’Informatique et des Libertés; authorization request 914066v3) and the French ethical committee (Comité de Protection des Personnes: reference 2019-A03066-51). All patients in the OFSEP cohort provided informed consent to have their medical data shared anonymously for research purposes. Appropriate research information on the OFSEP registry research cohort is available on the OFSEP website [ 19 , 20 ]. After 1 month without objections from the patient, patient data were extracted for analysis. As per French legislation (“Loi Informatique et Libertés” No. 78–17 on January 6, 1978, modified by Law No. 2016–41 on January 26, 2016), LymphoTEC was registered on ClinicalTrials.gov ( NCT04756687 ) and was compliant with the “French Methodology of reference MR-004”. The study was conducted in accordance with the Declaration of Helsinki.

Footnotes

Prior Presentation : The data in this manuscript were previously presented at Journeés de Neurologie de Langue Français (de Sèze J, Labauge P, Liblau R, et al. Résultats finaux de l’étude LymphoTEC: Données en vie réelle sur la reconstitution lymphocytaire chez des patients lymphopéniques traités par diméthyl fumarate; 4–7 April 2023, Lyon, France).

References

Associated Data

Supplementary Materials

Supplementary file1 (PDF 325 KB)

Data Availability Statement

OFSEP is the sole proprietor of the study database, and access to data is governed by the rules and guidelines of the steering committee. Patient-level data for the current study were neither directly accessible nor transferred to the sponsor.