Clinic Speak: horses for courses; which DMT will you switch to?

Horses for courses: if you are failing a 1st-line DMT which DMT will you switch to? #ClinicSpeak #MSBlog #MSResearch

"Data-mining is becoming increasingly ubiquitous and is being used by health economists to assess how healthcare interventions are performing in real-life and to make comparisons between drugs. The study below compares MSers who have failed and interferon and are then switched to either fingolimod or GA. Surprise, surprise those switched to fingolimod were ~ 60% less likely to have a relapse compared to those switched to GA. This tells us that fingolimod is a better DMT for treating interferon failures than GA and would support the relative efficacy results of the phase three studies of these two compounds."

"The data in this study is not high quality in the sense that it is not from a randomised controlled trial. One could argue that some systematic factor may be at play that determines who is switched to GA and this factor could explain factor the differences in efficacy between the two groups of MSers. This is what we call a bias and is the main reason we do randomised trials. Randomisation ensures, at least most of the time, that baseline characteristics or biases are neutralised by being equally distributed between the arms of the study. The other factor that needs to be kept in mind is that this study was funded and performed by Novartis, the company who produces and markets fingolimod. Therefore it is not surprising that this study was positive. Do you think Novartis would publish or release negative data on fingolimod? These results could therefore represent the phenomenon called publication bias; only positive studies get published. Publication bias is a common problem and can only be counteracted by initiatives that ensure all studies are registered and the investigators are forced to publish their results. Saying this this will only work for clinical trials; I can’t see this working for data-mining studies as the one below. If Novartis purchases data from a healthcare company, or the NHS for that matter, who is going to know? These deals are usually confidential. It is however reassuring that data from the large MSBase project has been presented confirming these results; if you switch from interferon to GA or from GA to interferon you are less likely to be relapse free than if you switch to fingolimod.”

“Despite these concerns payers and neurologists will take this data and assimilate it with other data and come to the conclusion that fingolimod is more effective than GA.”

“I would like to remind you that it is horses for courses. Some MSers who switch from an interferon to GA may respond very well and have NEDA (no evident disease activity). I agree that this will be a minority of cases and will be a lower proportion than those on fingolimod. It is a great pity we can't predict up front who will be a responder or non-responder to GA so that we can make a better decision. At present we need to monitor MSers on GA for 9 to 21 months, and beyond, to find out whether or not they are responders or non-responders. I say 9 months as this is the time point I used for rebaseling MSers on GA. Why 9 months? This is how long it takes for GA to reach its maximal levels of efficacy on MRI. So if someone has a Gd-enhancing lesion or lesion on their 9 month rebaselining scan I would recommend switching them to a more effective treatment. If there are no Gd-enhancing lesions I would rescan them in 12 months; the 21 month scan will then be compared to the 9 month scan and if there were any new T2 lesions or a Gd-enhancing lesion I would switch them. Obviously an objective clinical relapse in this period would trump the MRI activity.”

“This is how we are now using MRI to monitor and assess response to treatment. My big concern about this approach is that if you have smouldering, or subclinical, disease activity 21 months is a long time to have a shredder active in your brain. This is why we need to push for better metrics to assess response and non-response to treatments at an earlier stage. We need to be able to make the call within 6 months of starting a treatment. To do this we need to validate other biomarkers in our treatment algorithms. We were hoping to this as part of a large UK study, but the MRC decided not to fund the programme. I think we have missed a big opportunity to improve the way we treat MS in this country. At the moment there are no evidence-based guidelines to inform us on how to sequence DMT treatments to get the best outcomes for MSers."

"If you are failing interferon therapy which DMT are you going to switch to? This is not a trivial question and depends on a large number of factors. I am hoping to develop a decision aid to help you with this decision. Would you be interested in using it?"


Bergvall et al Relapse Rates in Patients with Multiple Sclerosis Interferon Switching from Glatiramer Acetate or to Fingolimod: A U.S. Claims Database Study. PLoS One 6 February 2014, 9 (2): e88472.

BACKGROUND: Approximately one-third of patients with multiple sclerosis (MS) are unresponsive to, or intolerant of, interferon (IFN) therapy, prompting a switch to other disease-modifying therapies. Clinical outcomes of switching therapy are unknown. This retrospective study assessed differences in relapse rates among patients with MS switching from IFN to fingolimod or glatiramer acetate (GA) in a real-world setting.

METHODS: US administrative claims data from the PharMetrics Plus™ database were used to identify patients with MS who switched from IFN to fingolimod or GA between October 1, 2010 and March 31, 2012. Patients were matched 1∶1 using propensity scores within strata (number of pre-index relapses) on demographic (e.g. age and gender) and disease (e.g. timing of pre-index relapse, comorbidities and symptoms) characteristics. A claims-based algorithm was used to identify relapses while patients were persistent with therapy over 360 days post-switch. Differences in both the probability of experiencing a relapse and the annualized relapse rate (ARR) while persistent with therapy were assessed.

RESULTS: The matched sample population contained 264 patients (n = 132 in each cohort). Before switching, 33.3% of patients in both cohorts had experienced at least one relapse. During the post-index persistence period, the proportion of patients with at least one relapse was lower in the fingolimod cohort (12.9%) than in the GA cohort (25.0%), and ARRs were lower with fingolimod (0.19) than with GA (0.51). Patients treated with fingolimod had a 59% lower probability of relapse (odds ratio, 0.41; 95% confidence interval [CI], 0.21-0.80; p = 0.0091) and 62% fewer relapses per year (rate ratio, 0.38; 95% CI, 0.21-0.68; p = 0.0013) compared with those treated with GA.

CONCLUSIONS: In a real-world setting, patients with MS who switched from IFNs to fingolimod were significantly less likely to experience relapses than those who switched to GA.

CoI: multiple

Labels: , , , , ,