Retrospective observational study based on the review of the clinical files of RA patients who had been treated with tocilizumab in the University Regional Hospital of Malaga, Spain. The study was approved by the Clinical Research Ethics Committee of the hospital (CEIC).

All the patients treated with tocilizumab between January 2009 and December 2015 were recruited in the study. The patients were assessed on initiating treatment with tocilizumab and after 24 and 52 weeks of follow-up. The eligibility criteria were: age ≥18 years, RA according to the criteria of the American College of Rheumatology/European League against Rheumatism 201016 and having received an iv infusion with tocilizumab. Patients with any type of inflammatory or rheumatic disease which was not RA were excluded (except for secondary Sjögren syndrome).

All patients were normally prospectively followed up in a specific biologic therapy unit (BTU) in compliance with the pre-established protocol of systematic data collection. This protocol includes, among other variables, data on the disease activity, physical function and adverse effects. The BTU reviewed the patients treated with sc bDMARDs every 3 months in general and specific (of sc biologics) consultations, alternating between them. The patients with iv bDMARDs were reviewed every time the drug was administered. Five days after the 1st infusion or in the 2nd dose of the sc bDMARD the patients received a call from the nursing staff to confirm that everything had gone well. The patients initially received 8mg/kg tocilizumab every 4 weeks. During the treatment period with tocilizumab, there were no restrictions in the concomitant use of the DMARDs (i.e. methotrexate or leflunomide). The tocilizumab dose was optimised in several patients from 8 to 4mg/kg per infusion. The decision regarding the dose or intervals was initially based on the product data sheet but was changed in accordance with the practitioner's criteria. These changes also included the change from iv to sc administration, when the latter was available.

The main outcome was the effectiveness of tocilizumab at week 24. The secondary outcome variables included: effectiveness at week 52, treatment retention time, physical function and safety. Effectiveness of dose optimization was also measured, and the change of formulation from iv to sc after 3 and 6 months.

The effectiveness of tocilizumab was measured by using the proportion of patients with a low disease activity, measured by the disease activity score in 28 joints with erythrocyte sedimentation rate (DAS28-ESR<3.2) and remission (DAS28-ESR<2.6)17 at weeks 24 and 52. The following were also used as measurements of effectiveness: the retention rate of tocilizumab (i.e. the proportion of patients whose biologic treatment was not discontinued) at week 52 and physical function, measured with the health assessment questionnaire (HAQ) (range 0–3)18 at week 24 and 52. Safety was assessed by the total and serious adverse effects incidence rate between 2009 and 2015. This was conducted by dividing the total number of adverse effects by the sum follow-up time of all patients in years (number of events/patients–year). The serious adverse effects were defined as life-threatening or resulting in death, hospitalisation or long term disability. Serious infections were defined as requiring iv antibiotics or leading to hospitalisation or prolonged hospital stay or death. The effectiveness of optimization and change of iv route to sc route in treatment with tocilizumab was evaluated using the differences in DAS28-ESR, HAQ, C-reactive protein (CRP) and ESR at 3 and 6 months compared with the baseline value.

Other recorded variables include age (years), gender, weight (kg), height (cm), body mass index (BMI=weight/height2) and smoker habits (smoker, non smoker). RA variables included duration of symptoms, number of swollen joints (28 joints), number of painful joints (28 joints) and general assessment by the patient of their activity (measured on a visual analogue scale of 0–100mm). High, moderate and low activity, and also remission were identified as scores DAS28-ESR15 of >5.1, 3.2–5.1, 2.6–3.2 and <2.6, respectively. Concomitant use of DMARDs was recorded.

Descriptive analysis was made expressing the quantitative variables as measurements of central and dispersion tendencies and qualitative variables as absolute and relative frequencies. The Kolgomorov–Smirnov test was used to confirm the fit to normality of continuous variables. The Student's t-test was used for independent samples or the Mann–Whitney test in cases of non normality between the DAS28-ESR, HAQ, CRP, ESR mean at 24 and 52 weeks. It was also used for the study of dose optimization and the change of formula at 12 and 24 weeks compared with baseline. The survival time of tocilizumab was analysed using the Kaplan–Meier curves. Survival probabilities and their 95% confidence interval (95% CI) were calculated. For all analysis a value of P<.05 was considered significant. Data were analysed with the IBM SPSS Statistics version 19 software.

The main patient characteristics are described in Table 1. The majority were women around 53 years of age and with a seropositive disease, which was erosive and long term. The great majority had high inflammatory activity of the disease at the beginning of treatment. Mean exposure (SD) to tocilizumab was 44.8 (19.4) months. The majority received tocilizumab in combination with a DMARD.

Out of the 53 patients treated with tocilizumab, 51 (96.2%) continued receiving tocilizumab in week 24 and 52. None of the patients discontinued tocilizumab due to insufficient response during the observation period at 52 weeks. As shown in Table 2, the patients treated with tocilizumab had substantially improved by week 24. The proportion of patients who achieved remission or low activity of the disease in weeks 24 and 52 was 75.5% and 87.3%, respectively. Furthermore, patients also demonstrated better disease control at week 52, measured by DAS28-ESR, ESR, CRP and HAQ. Mean retention time (95% CI) was 81.7 months (76.6–86.7). Patients with moderate to high disease activity in week 52 were associated with a better BMI compared with patients with low activity/remission (25.6 vs 28.9; P=.02).

Twenty one of the 53 patients (39.6%) optimised the tocilizumab dose, 19 with iv and 2 with sc. All the patients in iv treatment reduced the dose of each infusion and maintained the interval between doses, whilst all patients with sc tocilizumab extended the interval between doses and maintained the dose. Two patients who had reduced doses with iv tocilizumab changed directly to sc with the reduced dose. The main baseline characteristics of the patients who reduced the doses are listed in Table 3. After 6 months, no differences were observed in efficacy results (Table 3), patient in remission percentage and low baseline activity of the disease and at 6 months it was situated at 94.7% and 89.4% respectively. Only one patient regressed to a complete dose of tocilizumab after 37 months of reduced dose due to clinical worsening of symptoms. No adverse effects were observed during follow-up.

Thirty five patients changed to sc formulation from iv and depending on the previous dose of iv tocilizumab (i.e. 8mg/kg, 6mg/kg or 4mg/kg) which was being received they were prescribed with the following sc tocilizumab dose: 162mg/week (n=30), 162mg/10 days (n=4) or 162mg/15 days (n=1), respectively. At 6 months, no differences were observed in the efficacy results when efficacy of sc formulation was compared with iv formation. The percentage of patients in remission and low baseline disease activity at 6 months was 88.6% and 77.3%, respectively. Four (11.4%) patients was readministered iv tocilizumab due to subjective worsening of symptoms (increase in the number of painful joints and assessment by the patients, without any increase in acute phase reactant tests, n=2) or for non-serious side effects (n=2).

As shown in Table 4, half of the patients presented with an adverse effect (13.6 events/100 patients per year), but only 2 of them were serious (1.0 case/100 patients per year), which required treatment withdrawal. The majority of side effects were infections. During treatment with sc tocilizumab, 12 (34.3%, .55 events/100 patients per year) patients had non serious adverse effects: 8 injection site reactions (22.8), 2 respiratory infections (5.71), one patient presented with hypertransaminasemia (2.85) and one patient presented with a vasovagal reaction (2.85). There were no serious adverse effects. There were no differences regarding the number of adverse effects between iv and sc tocilizumab (15 vs 12, P=.158).

The authors declare that for this research no experiments have been carried out on humans or animals.

The authors declare that they have adhered to the protocols of their centre of work on the publication of patient data.

The authors declare that no patient data appears in this article.