Objectives Medication persistence, defined as the duration of time from its initiation to its discontinuation, is a surrogate for treatment effectiveness. The aim of the study was to evaluate persistence and causes of biological therapy (BT) suspension in patients with chronic inflammatory arthropathies: rheumatoid arthritis, ankylosing spondylitis and psoriatic arthritis.
Methods Single institution, descriptive, retrospective cohort study. Adult patients with chronic inflammatory arthropathies on BT between January 2009 and December 2016 were included. Persistence to BT was compared considering the type of pathology and treatment. The Kaplan–Meier test was used to analyse medication persistanence and factors associated with it. An analysis of reasons for therapy discontinuation was performed.
Results Three hundred and sixty-two patients were included in the study, which comprised 478 BT lines. For all patients, the 12-month persistence rate was 71.3% (341 out of 478). At the end of the study, 45.2% of the patients continued on their initial BT. Median treatment persistence was 1489 days (CI 95% 1195 to 1783). Longer BT persistence was associated with naïve BT patients: 1945 days (95% CI 1523 to 2367; P<0.001) and ankylosing spondylitis diagnosis: 2402 days (95% CI 1604 to 3200; P=0.014). The most frequent causes of treatment discontinuation were therapeutic failure (47.6%) and adverse drug events (28.2%).
Conclusions We found good long-term persistence in patients with chronic inflammatory arthropathies treated with BT. Patients with rheumatoid arthritis had significantly shorter persistence compared with those with ankylosing spondylitis and psoriatic arthritis. Naïve BT was associated with longer persistence. Therapeutic failure was the main cause of BT withdrawal.
- clinical pharmacy
- health economics
- public health
- side effects of drugs
Statistics from Altmetric.com
Rheumatoid arthritis (RA), ankylosing spondylitis (AS) and psoriatic arthritis (PsA) are chronic inflammatory arthropathies. The commercialisation of biological therapies (BT) for more than 15 years has led to a significant improvement in chronic inflammatory arthropathies' prognosis and in patients' quality of life. However, these are not curative therapies but a lifelong treatment. As a result, life expectancy has also been extended. This means that the number of patients on BT has grown exponentially in recent times, thus generating a very high consumption of resources, which raises great concern among health system managers.1
Several studies have demonstrated that inadequate adherence and non-persistence with respect to prescribed medication regimens result in increased morbidity and mortality from a wide variety of illnesses, as well as higher healthcare costs.2–4
Medication persistence, defined as the duration of time from its initiation to its discontinuation, is a measure of drug effectiveness, safety and real-world clinical utility.5 Over the past decade, some studies have been published showing that approximately one-third of patients discontinue BT in the first year of treatment and this has resulted in higher healthcare costs and resources use.6 Biological therapy persistence has been reported in several registries,7–10 and some reports in the real-world setting,11–15 but there are limited data on the influence of the type of chronic inflammatory arthropathies, number of BT lines or the type of BT in the persistence. There is also little information about the causes of drug discontinuation.
The objective of our study was to evaluate the persistence of BT in patients with chronic inflammatory arthropathies as well as the causes for its discontinuation.
Single institution, descriptive, retrospective cohort study. Eligible study population included patients treated with BT between January 2009 and December 2016.
Persistence was retrospectively calculated as the time period from BT initiation to its definitive discontinuation. Suspensions were considered definitive when no consecutive re-introduction of treatment was reported according to pharmacy electronic medication dispensing programme and medical record. The reasons for BT withdrawal were classified in six categories: therapeutic failure, adverse events, non-adherence, patient voluntary suspension, clinical remission and others.
The study included adult patients diagnoses with RA, AS or PsA who fulfilled the 1987 American College of Rheumatology classification criteria,16 the modified New York/Assessment of Spondylarthritis International Society classification criteria17 18 and the CASPAR classification.19 Patients should be being treated by the Rheumatology Clinic and started a BT with abatacept, adalimumab, certolizumab, etanercept, golimumab, infliximab, tocilizumab or ustekinumab between 1 January 2009 and 6 months before the study end date (31 December 2016). Patients participating in clinical trials and those not fulfilling the selection criteria were excluded from the study.
Patient recruiment was stopped in June 2016, in order to assure a minimun of 180 days of monitoring.
The following analytical parameters were collected at the beginning of the BT: C-reactive protein (CRP), erythrocyte sedimentation rate (ESR) and haemoglobin.
Patients were classified into naïve BT patients if they had not been on BT before (first BT line) and non-naïve patients if they had already been before (second or successive BT lines).
The following BT-related variables were collected: concomitant treatments at the start of the BT (methotrexate, leflunomide and glucocorticoids) and BT type (abatacept, adalimumab, certolizumab, etanercept, golimumab, infliximab, tocilizumab and ustekinumab).
Adverse effect was considered as any undesirable event affecting patients during treatment suspect of being caused by the BT. Data were collected from the clinical records of the patients. When infections were identified, their location was specified.
For the statistical analysis, the Statistical Package for the Social Sciences (SPSS) programme was used. A descriptive analysis per disease and biological therapy was made which included all variables considered. Quantitative variables were expressed as mean value ±SD in the case of normal distribution and as median and interquartile range if distribution was not normal. Categorical variables were shown as absolute values and percentages. Survival curves for the persistence analysis were determined applying the Kaplan–Meier method and the log-rank test was used for comparison between groups.
All data were logged onto an Excel spreadsheet of which only study researchers were allowed access. Every BT line was assigned a code number to make patient identification impossible. The study complied with the Organic Law 15/1999 of 13 December on the protection of personal data and received the approval of the Research Ethics Committee from Galicia.
Three hundred and sixty-two patients were included in the study, which comprised 478 BT lines (250 lines in RA patients, 119 lines in AS patients and 109 lines in PsA patients). The characteristics of the sample are shown in table 1 divided per disease.
1A value of P<0.05 was considered to be statistically relevant.
2All glucocorticoids doses were converted to equivalent prednisone dose.
3Test could not be made because there are some values under 5.
For all patients, the 12-month persistence rate was 71.3% (341 out of 478). Median treatment persistence was 1489 days (CI 95% 1195 to 1783).
By chronic inflammatory arthropathies type, after 1 year, the RA persistence was 168 out of 250 biological therapy lines (67.2%), 90 out of 119 BT lines (75.6%) for AS and 83 out of 109 biological therapy lines (76.1%) for PsA (P=0.014).
First biological therapy line, AE and PsA patients had longer persistence (P<0.0001). (see figure 1).
At the time of the study closure, nearly half of the BT lines had been discontinued (45.2%–46.3%, 210 BT lines). Major BT discontinuation causes per chronic inflammatory arthropathies were therapeutic failures in RA (44.1%) and PsA (60.8%) patients. In AS patients, adverse drug events (44.2%) and therapeutic failure (44.2%) caused similar suspension of treatment (table 2).
Table 3 shows that the most frequent adverse effects leading to BT withdrawal were infections, gastrointestinal disorders, skin and subcutaneous tissue disorders, and site-of-administration reactions.
The most frequent treatment discontinuation cause was therapeutic failure for all biological therapy analysed except for certolizumab and abatacept, in which most frequent reason for suspension was adverse drug effects (table 4).
Data obtained in our study on BT persistence in different CIA are similar to those from other studies which conclude that treatment persistence in RA patients is lower than in patients with others CIA.7 12 A potential explanation of that may have been the fact that RA-diagnosed patients were older, had more comorbidities, higher inflammation activity and required more frequently concomitant methotrexate, leflunomide and glucocorticoids. These factors can contribute to the presentation of adverse effects that lead to an earlier treatment suspension.
Regarding the persistence of BT after 1 year, most of the information in the literature refers to RA. In our study, the percentage of patients continuing on the same BT after 1 year is similar to that reported in the US CORRONA RA Registry.8 In both cases, about one-third of RA patients discontinued BT therapies after 1 year, most commonly due to loss of effectiveness.
Treatment persistence was longer in the first BT line than in successive ones. However, Thomas el al12 did not find this difference when analysing the factors contributing to a higher persistence to golimumab.
Our study analysed persistence per BT type in line with other published works. In 2017, Favalli et al13 published a 10 years' retrospective study analysing the persistence of a tumour necrosis factor (TNF)-inhibitor in patients with chronic inflammatory arthropathies. Etanercept showed the longest medication persistence in an adult population in comparison to adalimumab and infliximab (P<0.001) and infliximab the lowest persistence in a juvenile population (P=0.005 vs adalimumab and P<0.001 vs etanercept). Jones et al14 described persistence rates of firstline BT of 79% (tocilizumab and abatacept), 64% (subcutaneous TNF-inhibitor) and 13% (infliximab), after 1 year of treatment. Similar results were also found in the study by Takabayashi et al15, which analysed 4970 cases in 13 national university hospitals in Japan. But these results are not easily comparable with ours due to differences in population and study period, and to non-statistically significant results given the width of CIs and the difference in line numbers among different BT. Therefore, tocilizumab did not have longer persistence in our study which may be attributed to the fact that this drug is not usually used as a first BT line in our hospital.
The most frequent cause for BT discontinuation was therapeutic failure except for AS patients. In this case, BT withdrawal was caused by both treatment failure and adverse effects in equal degree. In the same way, in the Spanish Registry for Adverse Events of Biological Therapy in Rheumatic Diseases, treatment failure also appeared to be the most frequent cause for suspension.9
A different study showed that the most frequent cause for BT discontinuation in over 65-year-old patients were adverse drug effects, while in younger patients it was treatment failure.10 However, according to Strand et al8, the most frequent causes for BT suspension were failure (35.8%), physician's preference (27.8%), safety (20.1%), patient's preference (17.9%) and problems related to access to treatment (9.0%).
RA patients show significantly shorter persistence when compared with AS and PsA patients.
Second and/or successive therapy lines have shorter persistence than first line.
Therapeutic failure was the main cause of biological withdrawal.
Infusion/injection-related response and infections were the adverse effects that caused more BT discontinuations.
What is already known on this subject?
The features of treatment (eg, persistence, switch rate and drug consumption) seem to directly influence its health costs in chronic inflammatory arthropaties.
Better knowledge of pre-existing prescription patterns for this chronic illness and greater awareness of the economic burden of the disease could stimulate planning of healthcare interventions aimed at improving public health services for the management of this disease.
What this study adds?
This real-world data study provides knowledge to stratify and prioritise follow-up strategies in chronic inflammatory arthropathies.
RA patients with second and/or successive therapy line due to therapeutic failure or infusion/injection-related response and infections show shorter persistence.
The authors wish to acknowledge the work of Cristina Martínez-Reglero as responsible for the statistical analysis. An unrestricted grant from Pfizer has been used for the development of this study.
EAHP Statement 5: Patient Safety and Quality Assurance.
Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.
Competing interests None declared.
Patient consent for publication Not required.
Provenance and peer review Not commissioned; externally peer reviewed.
Data availability statement No data are available.
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.