To evaluate an intervention to reduce the nocebo effect (NE) when switching from the originator infliximab (OI) to the infliximab biosimilar SB2 in chronic inflammatory rheumatic disease (CIRD).
An intervention was built with healthcare professionals (HPs) and a patient representative, based on a systematic review of interventions reducing the NE in musculoskeletal diseases and semi-directed questioning of five patients. Our strategy consisted of training HPs, switch information given by the nurses, a consistent vocabulary. All CIRD patients switched from OI to SB2 were included for the intervention. The primary outcome was the SB2 retention rate (RR) at 34 weeks. Secondary outcomes were the SB2 RR at 12 months, discontinuation rates due to a possible NE and comparison with a historical cohort of CIRD patients receiving the OI and 6 published European cohorts.
45 patients were included from March 2018 (rheumatoid arthritis, n=17, spondylarthritis, n=28). After 34 weeks, the SB2 RR was 91.2%, similar to the historical cohort RR (p=0.41) but higher than the 3 European cohort RRs (p<0.05). At 12 months, the SB2 RR was 84.5% vs 88.4% for the historical cohort (p=0.52). SB2 discontinuation due to a possible NE was 6.6% after 12 months.
A tailored communication with a prominent role of nurses reduced the NE in non-medical switches from the OI to SB2 as compared to published results. The RR was similar to the historical cohort RR. The methodology used to construct this intervention may help improve the outcomes of switches with upcoming biosimilars.
Owing to cost-effectiveness considerations, health authorities promote switches from originator biologics to biosimilars. Lower retention rates of biosimilas in real life studies compared to double blinded controlled trials have been observed and may be explained by a nocebo effect.
This study describes an intervention that reduced the nocebo effect when switching from the originator infliximab to an infliximab biosimilar.
The methodology used in the present study ie, building an intervention after investigation of patient's perspective, and ensuring consitent communication from all healthcare professionnal - may be used for future switches to other biologic agent biosimilars.
Biologic drugs (BDs) have transformed disease outcome for patients with chronic inflammatory rheumatic diseases (CIRDs), but at high cost for healthcare systems.
Double-blind-randomised control trials (RCTs) assessing a switch from the originator infliximab (OI) to an infliximab BS have reported comparable efficacy, safety and immunogenicity, with no difference in discontinuation rate,
The NE, an emerging concept in rheumatology,
Some recommendations,
Because the healthcare professional–patient interaction affects healthcare outcomes,
Step 1: a systematic literature review (inception to 15 November 2018) was conducted to identify risk factors (RFs) and intervention strategies to reduce the NE in musculoskeletal diseases, and more specifically in switches from originators to BSs. The method, flow chart and full results are available in
Six studies collected RFs of withdrawal after a switch. However, conclusions could not be drawn because the RFs identified were markers of higher disease activity and treatment failure (
Baseline characteristics of the SB2 cohort and the historical originator infliximab (OI) cohort
OI→SB2 (n=45) | Historical OI cohort (n=52) | P value | |
---|---|---|---|
Age, mean (SD) | 53.2 (2,1) | 50.25 (1.2) | 0.194 |
Female (n, %) | 25 (55%) | 31 (59%) | 0.690 |
| 28 (62%) | 33 (62%) | 0.901 |
Type of SpA | |||
Axial SpA (n, %) | 25 (89%) | 32 (97%) | 0.234 |
Peripherical SpA including PsA (n, %) | 16 (35.5) | 22 (42.3%) | 0.821 |
MRI lesions (n, %) | 17 (61%) | 20 (60%) | 0.811 |
HLA B27+ (n, %) | 19 (68%) | 20 (61%) | 0.335 |
Inflammatory bowel disease (n, %) | 7 (25%) | 8 (24%) | 0.947 |
Uveitis (n, %) | 6 (21%) | 4 (12%) | 0.363 |
Psoriasis (n, %) | 7 (25%) | 10 (30%) | 0.599 |
Methotrexate | 6 (86%) | 7 (100) | 0.613 |
BASDAI, mean (SD) | 27.7 (4.6) | 31.3 (4.9) | 0.554 |
CRP, mg/l, mean, (SD) | 5 (1.9) | 5.5 (2.2) | 0.907 |
PGS/100, mean (SD) | 32.1 (5.4) | 38.7 (5.5) | 0.907 |
Psoriatic arthritis DAS28, n, mean (SD) | 5, 1.9 (0.2) | 7, 2.3 (0.5) | 1 |
| 17 (38%) | 20 (38%) | 0.946 |
ACPA+ (n, %) | 16 (94%) | 19 (95%) | 0.909 |
Erosive disease (n, %) | 13 (76%) | 14 (70%) | 0.669 |
Corticosteroids (n, %) | 6 (35.2%) | 8 (40%) | 0.776 |
Dose corticosteroids | 4.9 (0.8) | 5.5 (0.5) | 0.429 |
Methotrexate | 10 (100%) | 14 (100%) | 0.437 |
DAS28 CRP, mean (SD) | 2.5 (0.3) | 2.8 (0.26) | 0.511 |
CRP, mg/l, mean (SD) | 3.8 (1.7) | 2.1 (0.7) | 0.560 |
PGS/100, mean (SD) | 36 (7.4) | 49.5 (6) | 0.276 |
Swollen joint count, mean (SD) | 0.2 (0.1) | 1.2 (0.5) | 0.916 |
Fibromyalgia (n, %) | 3 (6.6%) | 3 (5.8%) | 0.876 |
Anxiety-depression (n, %) | 6 (13.3%) | 8 (15.4%) | 0.748 |
| |||
No. of BD lines before IFX, mean (SD) | 0.4 (0.1) | 0.4 (0.1) | 0.791 |
Mean time under OI (month), mean (SD) | 113.5 (9.3) | 94.8 (9.4) | 0.066 |
Number of previous injections | 64.8 (5.4) | 54.1 (4.7) | 0.104 |
Dose (mg/kg), mean (SD) | 4.5 (0.1) | 4.6 (0.1) | 0.512 |
Infusion rhythm, mean (SD) | 7.5 (0.2) | 7.3 (0.2) | 0.559 |
ACPA, Anti-citrullinated protein antibody; BASDAI, Bath Ankylosing Spondylitis Disease Activity Index; BD, biologic drug; CRP, C-reactive protein; DAS28, Disease Activity Score in 28 joints; IFX, Infliximab; OI, Originator infliximab Remicade; PGS, Patient global score; SpA, Spondylarthropathy.
Five studies aimed to reduce the NE during switches by an intervention targeted to patients: two studies used a shared decision-making method and a structured communication strategy,
Retention rate (RR) at 34 weeks. Comparison with published European cohorts with endpoints at 34 weeks
Study | Follow-up | n | RR (%) (p value*) | Historical cohort RR | Withdrawal due to NSAE/NSS (%) | NSAE/NSS definition |
---|---|---|---|---|---|---|
Tweehuysen, 2018 | 6 months | 192 | 75.5% | None | 2% | NSAE not mentioned by the author but calculated by from the article tables (NSAE considered if no objective adverse event, no worsening of disease activity score or increased ESR or CRP level) |
Scherlinger, 2018 | 33 weeks | 89 | 72% | 72/82 (88%) | 12.5% | Requested to switch back to OI although they did not present clinical activity defined by a worsening of disease activity scores. |
Avouac, 2018 | 34 weeks | 182 | 73.5% | None | Unknown | |
This study | 34 weeks | 45 | 91.1% | Unknown | 2% | Cf. Methods |
*p<0.05 compared with this study, Fisher’s exact test.
Step 2: Patient’s perspective: a trained rheumatologist conducted semi-directive interviews with five patients who received intravenous abatacept or tocilizumab for rheumatoid arthritis. Those were not immediately concerned by a switch but might be in the future. The aim was to clarify patients’ knowledge and perceptions of BS, the information needed, and how patients expected the information to be delivered. The interview schedule and patients’ demographics and responses are available in
Step 3: Construction of the intervention in a multidisciplinary fashion, based on the results of steps 1 and 2 and patients’ and rheumatologists ‘perception of BSs in published surveys.
Switch modalities and messaging and vocabulary tools (
The switch modality was as follows: on infusion day, before the medical consultation and BS prescription, the day-care nurses provided oral information (a simple sentence mentioning a change of name) and written information (letter,
Step 4: Implementation of the intervention: a non-medical switch (ie, not initiated by the patient or the physician) from OI to the infliximab BS SB2 was requested by hospital directors for all patients receiving OI. All patients with CIRD (rheumatoid arthritis [RA], axial or peripheric spondylarthritis [SpA], including psoriatic arthritis [PSA]) currently receiving OI received the support strategy described above and were included. Switches started in March 2018 until every day-care patient was seen at least once; follow-up lasted 1 year. Intervention implementation was not monitored.
Patients were followed up in routine care; treatment change was decided by the rheumatologists in charge. Drug maintenance, side effects and disease activity were routinely assessed at each infusion. Data were analysed from medical records.
The primary endpoint was the RR of SB2 after 34 weeks. Secondary endpoints were the SB2 RR at 12 months; SB2 discontinuations due to NE/NSAE/NSS at 34 weeks and 12 months; comparison of RR and discontinuations due to NSAE/NSS with data from a historical cohort and with RRs from published European studies; and assessment of RFs of withdrawals.
Because the NE had been revealed in real-life studies vs RCTs, comparison with other real-life studies and cohorts was considered adequate for the study purpose. Therefore, we compared data with that from a historical cohort and the literature. The historical cohort consisted of all patients with CIRD receiving IO at 1 year before the switch in the same rheumatology department. Data were collected from patients’ files.
For comparison with the literature, we used European cohorts with similar follow-up: Tweehuysen
For each biosimilar withdrawal, the patient’s file was examined by an independent evaluator to assess NE/NSAE/NSS. In case of doubt, consensus was achieved with a second independent evaluator.
An NE leading to BS withdrawal was defined as lack of efficacy with no objective criteria for increased inflammation (ie, CIRD worsening or end-of-dose wear-off without increased level [CRP] or erythrocyte sedimentation rate [ESR], number of swollen joints, or flare of a concurrent disease [inflammatory bowel disease, uveitis or psoriasis]) or non-objective and non-specific side effects (ie, fatigue, headache, anxiety). NE was retained if the symptoms occurred after the switch and disappeared after a back-switch or change of BD. Criteria for NSAE/NSS in the historical cohort were the same lack of efficacy or subjective adverse events and disappearance after change of BD.
The number of patients to include to detect a 50% increase in RR as compared with the only French study of non-medical switches
For risk factor analysis, we used descriptive statistics and tested significance with the χ² or Fisher exact test as appropriate. For quantitative variables, we used the Student’s t-test or a non-parametric test according to the data distribution. P<0.05 was considered statistically significant.
Patients were involved in the construction of the intervention, not in the recruitment. Results will be shared via conference presentation.
This study is an observational study registered at the national agency for drug safety biological research and collections registration service (l’Agence nationale de sécurité du médicament [ANSM]) as part of a search in routine care assessment of the registered and approved patient education programme in the rheumatology department (RCB ID: 2018-A00695-50) and the IT and freedom commission (Commission Nationale Informatique et Liberté; CNIL no. 1734260).
We included 45 patients with a switch from March 2018, who were followed up until August 2019. Initially, no patient refused SB2 treatment. Every patient had a consultation with a rheumatologist (resident or attending) before BS prescription, as part of routine care. No patient required a specific consultation with a rheumatologist to discuss switch beforehand. Only one patient, who described loss of efficacy after the switch was addressed to the rheumatologist trained in motivational communication, but she refused to resume SB2. Three patients were excluded: one for pregnancy, one because follow-up was planned in another hospital and one with SAPHO (synovitis, acne, pustulosis, hyperostosis, osteitis). In the historical cohort, 52 patients were included between December 2016 and February 2018. Baseline characteristics of the cohorts were similar (
Mean follow-up at first evaluation (34 weeks) was 235 (SD 5.4) and 238 (SD 0) days in the SB2 and historical cohorts. RRs were 41/45 (91.2%) and 50/52 (96.2%) (p=0.41). SB2 RR was significantly higher than in other European cohorts with similar follow-up (p<0.05) (
NSAE: non-specific side events. NSS: non-specific symptoms. OI: originator infliximab. RR: retention rate. CRP: ESR: erythrocyte sedimentation rate.
Mean follow-up at 12 months was 368 days (SD 14.6, 95% CI [CI] 339.3-396.8) and 410 (SD 10.8, 388.8-431.2) for the SB2 and historical cohorts. RRs were 38/45 (84.4%) and 46/52 (88.5%). Withdrawal-free survival did not differ between the two cohorts (p=0.520) (
Retention rate 1 years after the switch, comparison with published European cohorts
Study | Follow-up | n | RR (%) | Historical cohort RR | Withdrawal due to NSAE/NSS (%) | NSAE/NSS definition |
---|---|---|---|---|---|---|
Nikiphorou, 2015 | 7.5–13 months | 39 | 71.8% | None | 15.3% | Discontinuation for subjective reasons without objective deterioration of disease. |
Boone, 2017 | 9 months | 24 | 87.5% | None | 12.5% | NE response: unexplained, unfavourable therapeutic effect subsequent to a non-medical switch from OI to BS with regaining of the beneficial effects after reinitiating the OI |
Glintborg | 413 days | 802 | 83.5% | 86.8% | Data not available | |
This study | 368 days | 45 | 84.5% | 88.4% | 6.6% | Cf. methods |
*p<0.05 compared with this study, **χ² test and *** Fisher’s exact test.
BS, Biosimilars; NSAE, Non-specific side events; NSS, Non-specific symptoms; NE, Nocebo effect; OI, Originator infliximab; RR, Retention rate.
Treatment withdrawal-free survival (SB2 in switched cohort and originator infliximab [OI] in historical cohort).
Kaplan-Meier survival curves. Comparison with log-rank test for OI to SB2 and historical OI, p=0.520.
Reasons for OI discontinuation are detailed in
Reasons for SB2 in switched cohort and historical OI cohort
Groups | Follow-up | n | Number of withdrawals | Reasons for discontinuation | Treatment after discontinuation |
---|---|---|---|---|---|
OI to SB2 | 34 weeks | 45 | 4 | NE (pain, fatigue) | Backswitch to OI |
1 year | 41 | 3 | Uveitis | Backswitch to OI | |
Historical OI | 34 weeks | 52 | 2 | Loss of efficacy | Istekinumab |
1 year | 50 | 4 | Loss of efficacy | Abatacept |
Patient global score, disease activity score in 28 joints and Bath Ankylosing Spondylitis Disease Activity Index were unchanged between baseline and 1-year follow-up (p=0.316, p=0.369 and p=0.390, respectively); CRP level was significantly higher at baseline than at 1 year (mean 4.5 [SD 78] vs 2.5 [SD 8.7] mg/l, p<0.001). No adverse event causing hospitalisation occurred after the switch.
NE: nocebo effect. NSAE: non-specific side events. NSS: non-specific symptoms. OI: originator infliximab.
At 34 weeks, withdrawals due to NE/NSAE/NSS were 1/45 (2.2%) in the switch group (
The number of withdrawals was too small for statistical analysis of risk factors. Baseline characteristics of patients who withdrew SB2 are in
Baseline characteristics among withdrawers in the 2 cohorts
OI→SB2 | Historical OI (n=6) | |
---|---|---|
Age, mean (SD) | 53.1 (4.9) | 47.1 (4.9) |
Female | 2 | 5 |
Spondylarthropathy | 6 | 4 |
BASDAI, mean (SD) | 26.3 (8.8) | 63 (9.6) |
Rheumatoid arthritis | 1 | 3 |
DAS28 CRP, mean (SD) | 0.96 | 2.9 (0.4) |
Methotrexate | 1 | 3 |
Fibromyalgia | 0 | 1 |
Anxiety-depression | 1 | 2 |
Time under OI (month), mean (SD) | 104.8 (22.6) | 49.3 (11.2) |
IFX dose (mg/kg), mean (SD) | 4.6 (0.08) | 4.6 (0.3) |
Infusion rhythm (weeks), mean (SD) | 7.5 (0.1) | 7 (0.44) |
IFX, Infliximab; OI, Originator infliximab Remicade.
In this real-life observational study, we monitored the efficacy of an intervention to limit NE and improve maintenance after a non-medical switch from OI to SB2. The intervention was constructed after a literature search and semi-directives interviews, which emphasised that patients had low knowledge of BS, needed information on the differences between BS and generics, had fears about their efficacy and tolerance and needed to be supported, with a prominent role of nurses. The intervention was standardised and involved all healthcare professionals. We found a significantly better RR at 34 weeks than that in three published European real-life studies
As concluded in our systematic review, interventions to reduce NE were scarce and did not provide positive outcomes as compared with historical cohorts.
Our structured communication was based on a prior investigation of patients’ needs and perspectives and the inclusion of a patients’ association representative when building the intervention. Interventions in the literature resulted mostly from agreement between healthcare professionals.
According to patients interviews in step 2, the nurses’ experiences on switches were particularly valued. Although the role of nurses in patient education and management of CIRD is well known,
We believe that our results are in favour of including of nurses in the intervention, as the first point of contact in announcing the switch to patients, since none of the patients requested a specific consultation to discuss the switch, in addition to the consultation with the rheumatologist or the resident in charge of the day-care unit.
A non-systematic literature search of the NE when constructing the intervention allowed for a closer insight into its determinants. Negative expectations are a known NE mechanism,
Given this tailored intervention, we obtained a good outcome in terms of RR at 34 weeks, which was the primary endpoint, and compared it with the literature and our historical cohort. Assessment at 34 weeks was considered appropriate because the few published intervention studies also had a follow-up at 6 and 7 months. These favourable results were obtained despite the injunction for a non-medical switch imposed by the hospital authority, preventing a shared decision-making process as recommended
At 12-month follow-up, our biosimilar RR did not significantly differ from that from the 3 other real-life published studies.
The strength of the study is an intervention close to patients’ perspectives. The conclusion of the qualitative step are consistent with international surveys on the patients’ perspective
Our study has several limitations: we had no control group but rather a historical cohort, with therefore limits in comparability. Moreover, some patients in the switch cohort received the OI in the historical cohort, so statistical comparison between groups was challenging. However, a real-life design was needed to study NE, which had not been identified in RCTs. Another limitation is the low number of patients (45). Nevertheless, our results showed a statistically significant improvement in RR as compared with another French cohort, with a predetermined number of patients to include of 52.
In conclusion, a tailored communication strategy focusing on a trustful relationship between patients and nurses and the implication of a multidisciplinary team showed improvement in outcomes after a switch from infliximab to a biosimilar, by reducing the nocebo effect. The intervention we describe may be useful for future switches of other intravenous BDs. The methodology may be used for developing interventions for all BD switches, focusing on the patient’s perceptive.
Dr. Margaux Boisson, Service de rhumatologie du Professeur Cahan, Hôpital Cochin, APHP.