Findings from a real-world “comparative effectiveness study,” showed relative risk for 2 years of relapse freedom in patients treated with natalizumab compared with those treated with rituximab was 0.903. This indicates that natalizumab is associated with a higher likelihood of relapse within 2 years, investigators note.
“What differentiates our research from all the prior observational studies comparing disease-modifying therapy efficacy is that we adjusted for confounding using the full electronic health record [EHR] data at the patient level,” co-senior author Zongqi Xia, MD, PhD, associate professor of neurology at the University of Pittsburgh, Pittsburgh, Pennsylvania, told Medscape Medical News.
This technique helped correct for some of the bias associated with observational studies, Xia added.
The findings were published online November 16 in JAMA Network Open.
The introduction of new disease-modifying therapies (DMTs) for MS in recent years has broadened the range of treatment options — but also made therapeutic decision making more difficult.
Data about the comparative efficacy of various DMTs could help neurologists choose the appropriate therapy for a given patient. Although head-to-head DMT comparisons are uncommon, researchers have previously used observational data to assess these questions.
The current investigators sought to compare two standard-efficacy DMTs (dimethyl fumarate and fingolimod) and two high-efficacy DMTs (rituximab and natalizumab). They examined registry and EHR data for patients in the CLIMB study, which is being conducted at Brigham and Women’s Hospital in Boston.
Eligible participants for the current analysis were patients aged 18 years or older with a diagnosis of MS who began DMT treatment between January 1, 2006 and December 31, 2016.
Patients who received chemotherapy before DMT initiation were excluded. In the comparison of standard-efficacy DMTs, patients who previously had received high-efficacy DMTs were also excluded.
The researchers gleaned relapse date and type (clinical or radiologic) from the CLIMB registry. Study outcomes included relapse rate at 1 year after DMT initiation, relapse rate at 2 years after DMT initiation, and time to relapse after DMT initiation. These outcomes were similar to the typical endpoints of clinical trials examining drug efficacy.
To account for potential confounders, the investigators analyzed demographic features, clinical features, and expert-defined EHR features. Among the demographics examined were age, sex, and race. Clinical features included follow-up, duration of previous DMT use, and number of previous relapses. Expert-defined EHR features included diagnostic codes and healthcare use.
In total, 204 participants received natalizumab, 115 received rituximab, 260 received dimethyl fumarate, and 267 received fingolimod. In each two-drug pair, patients did not differ significantly by sex or ethnicity but did differ by age.
Mean age at diagnosis was 37.2 years for the group receiving natalizumab vs 44.1 for the rituximab group (P < .001). Similarly, mean age was 41.7 for the dimethyl fumarate group vs 37.9 for the group receiving fingolimod (P < .001).
After covariate adjustments were made, the natalizumab group showed a higher 1-year relapse rate (doubly robust [DR] estimate, 0.080), higher 2-year relapse rate (DR estimate, 0.132), and shorter time to relapse (DR estimate, 0.903) than the rituximab group.
These consistent findings supported an association between rituximab and a lower relapse rate compared with natalizumab, the investigators note.
They found no significant difference between dimethyl fumarate and fingolimod in 1-year relapse rate (0.028) or time to relapse (0.957).
Although patients receiving fingolimod had a higher 2-year relapse rate than those receiving dimethyl fumarate (DR estimate, 0.071), the adjusted P value of .08 did not meet the threshold for multiple testing.
Ambiguities about the comparative efficacy of commonly prescribed DMTs were a motivation for the current study, said Xia.
“None of the prior studies have balanced patient confounders using EHR data,” he added.
The researchers used a tool called the E value to assess the strength of unmeasured confounding that would change the direction or negate the significance of an observed association. “It’s a way for us to demonstrate the value of using EHR data as a high-dimensional feature set for balancing patient characteristics,” Xia said.
Previous research has suggested that B-cell-depletion drugs, such as rituximab, are associated with a higher risk for infection.
Xia noted clinicians need to balance risks against benefits. “If B-cell-depletion drugs such as rituximab have higher risk, how do we justify using that class of drugs?” he asked.
“Having higher efficacy would be one reason, particularly for people who may be at a higher risk for more aggressive inflammatory disease activity,” Xia said.
Commenting on the study for Medscape Medical News, John R. Corboy, MD, director of the Rocky Mountain MS Center at the University of Colorado in Westminster, said the results largely confirm findings from previous comparisons of DMTs.
In addition, the investigators used the well-characterized CLIMB database and faced difficulties common to real-world studies, said Corboy, who was not involved in the research.
“When you do real-world data, you have to come up with a way to adjust for all the differences in the groups. They used a relatively novel way of adjusting for all the differences,” he said.
Study limitations, however, include the lack of data on disability, DMT discontinuation, and safety, Corboy noted. The patient populations also were relatively small.
In addition, the researchers’ failure to compare oral drugs to IV drugs is “a missed opportunity,” said Corboy.
“There are essentially no data from phase 3, controlled clinical trials in this realm, and limited data in the real-world studies.” He also noted the researchers did not compare the oral and injectable DMTs.
“Our job is to come up with the best plan to keep the patient healthy, and any data that helps us understand what the differences are between the drugs is helpful,” Corboy concluded.
The study was funded by the National Institute of Neurological Disorders and Stroke (NINDS). Xia reported having received grants from NINDS during the conduct of the study and serving on a scientific steering committee for Roche/Genentech outside the submitted work. Corboy receives research funding from Novartis, which manufactures fingolimod.
JAMA Netw Open. Published online November 16, 2021. Full text
Follow Erik Greb on Twitter: @MedscapeErik .