JNNP: heterogeneity of long-term disability in patients with secondary progressive MS

Secondary progressive multiple sclerosis (SPMS) is defined as a clinically progressive course
.
Studies have classified SPMS and primary progressive MS (PPMS) as progressive MS, and the concept of disease activity has been introduced to better determine the exact course of
the disease.
In general, the transition from RRMS to SPMS remains difficult to determine, and is usually retrospectively determined
after a period of uncertain diagnosis.
To overcome the lack of a consistent definition of SPMS, Lorscheider et al.
defined an objective definition
of SPMS in 2016 based on the Extended Disability Status Scale (EDSS) and information related to past relapses.
This definition requires a short confirmation period and can be traced back to longitudinal data
.
The criteria included in this definition are Disability Progression 1 EDSS score (EDSS≤5.
5) or 0.
5 EDSS points (EDSS≥6.
0).

Some natural history cohort studies have shown similar
rates of deterioration between SPMS and PPMS.
However, in a study in the Mayo Clinic group, PPMS took longer from the onset of progressive MS to EDSS 7 times, and the onset of progressive disease was shortened to EDSS
in women and after age 50 years compared to SPMS, although it recurred before and after the onset of progression.
Another study from the British Columbia cohort showed that male sexuality and motor symptoms were associated
with a shorter onset and younger age of SPMS.
In a previous study analyzing the trajectory of PPMS progression, Signori et al.
10 confirmed significant heterogeneity in the progression of long-term disability
.
In addition, three groups of patients
with very different EDSS trajectories were identified by unsupervised clustering.
The primary objective of this study was to investigate heterogeneity of long-term disability in SPMS patients using the same methodology
.
This article was published in the Journal of Neurology, Neurosurgery & Psychiatry
.

The project was conducted using data from four clinical MS registries: the MS Registry in Italy, OFSEP in France, the National MS Registry in Sweden and the International MSBase Registry
.
Minimum dataset requirements include demographics (age and gender), clinical and disease characteristics (date of first symptom, date of MS diagnosis, EDSS).

Data on all disease-adjusted therapy (DMT) exposures (including start and end dates) and clinical recurrence were
also extracted.
LCGA was used to model longitudinal EDSS scores
.
LCGA is a method to analyze heterogeneity between patients by dividing them into unobserved groups (potential categories); Time index models with more and more potential classes and different polynomial functions are fitted to the data, and the best model
is selected according to the Akaike information criterion and Bayesian information criterion (BIC) fit exponent, as well as the simplicity and clinical interpretability of the data.
The posterior classification probability and entropy of the correct latent class are calculated to measure the classification error, where higher values correspond to lower classification error and better
separation between latent classes.
Identify members
of a particular class (trajectories) by calculating the posterior probability of belonging to a category and assigning the patient to the category with the highest probability.
The median posterior probability for each patient category is reported as a measure
of goodness.
The median time and age
of an irreversible EDSS of 6 was calculated using the Kaplan-Meier method.

Longitudinal EDSS trajectories of patients diagnosed by doctors with SPMS, graded by different severity

In the primary analysis, in order to compare the demographic and clinical characteristics of different categories, traditional categorical analysis methods were used, χ2 tests were performed on categorical variables, ANOVA (age) or Kruskal-Wallis test (disease duration, baseline year, annual number of visits)
were performed for continuous variables.
To directly observe the effects of these covariates on members, the most important variables such as age, course of disease, sex, country, and recurrence status were included in the top 24 months as time-stable variables in the LCGA model
.
Results for each grade are reported as OR relative to the first group (mild), indicating a higher or lower
probability of being classified as moderate or severe based on baseline variables.
Wald's test is used to assess the effect of each variable on category members, reporting OR
only for variables with significant (p<0.
05) Wald test.
Class member frequencies
in unconditional (variableless) and conditional (variable-based) models are also compared.

Cohort 1 has a total of 3613 patients with
SPMS.
LCGA detected three different subgroups of patients with mild (n=1297; 35.
9%), moderate (n=1936; 53.
6%), and severe (n=380; 10.
5%) disability
.
The median time to EDSS 6 in the three groups was 12.
1 years, 5.
0 years and 1.
7 years, respectively, and the probability of reaching EDSS 6 in 8 years was 14.
4%, 78.
4% and 98.
3%,
respectively.
Similar results
were found in 7613 patients who met Lorscheider's criteria.
The EDSS time course of patients in the mild group is indicated by the trajectory of reaching EDSS 10 after 6 years of irreversible EDSS from 3-4 years, and only about 30% of patients at enrolled follow-up meet this threshold; And the most serious trajectory reached EDSS 6
within a few years.
Patients tend to be classified as having a more severe disability trajectory
when the Lorscheider criteria are applied compared to SPMS based on a physician's diagnosis.
This may indicate that patients in the cohort who met Lorscheider's criteria were, on average, milder disease, or that the time to diagnosis was earlier than those
diagnosed by a doctor.
In both cohorts, a shorter course of irreversible EDSS in the range of 3-4 at first visit was associated
with the most severe and rare longitudinal EDSS trajectories.
This was also confirmed by secondary analyses, where direct inclusion of the course of the disease into the latency model showed a reduced likelihood of classifying patients with a
longer course of disease into moderate (first cohort) or severe (two cohorts).
This may indicate that patients have begun a rapid secondary progression phase early in the course of the disease and continue
as they progress to the baseline EDSS range we set.

The rate of progression in SPMS patients varies
widely.
The identification of different disability progressions could guide future Phase 3 SPMS clinical trials
.
In addition, different trajectories of progression can reflect the progression of
different pathological mechanisms.