MS in a snapshot: Global Burden of Disease study 2016


Trying to understand disease, especially rare disorders, at a global level is a Herculean task; not only from a resources perspective, but also from the cost and logistics involved. Before, the Global Burden of Disease study presented here, there was the Multiple Sclerosis International Federation that provided up to date figures on the global epidemiology of MS back in 2008. So, almost a decade on where do we stand?

Well, MS prevalence is on the up. According to the GBD study the prevalence of multiple sclerosis in 2016 was 2.22 million, representing a 10.4% increase in the age-standardised prevalence since 1990. The greatest rise in figures were seen in the east Asia region (44.8% up) and Canada (81.9% up). While, the greatest overall prevalence was in North-America (164.6 per 100,000), followed by western Europe (127 per 100,000) and then Australasia (91.1 per 100,000); see figure below.

Age-standardised MS prevalence per 100 000 population in 2016 for both sexes, by location

One country stands out above the rest, and that is Greenland. As noted by Stenager in his comment to the journal, there have been no reports of MS among the Inuit inhabitants that make up the majority in Greenland. By his calculation there were approximately 10-12 people with MS living in Greenland, giving rise to a prevalence of 22 cases per 100,000 population, placing Greenland amongst the lowest prevalence countries, rather than indicated by the GBD study.

Like all studies of this ilk, there will be some limitations. Case accrual/reporting is always going to be an issue, particularly from the low income countries, such as Africa and Asia. Conversely, some of the rise in prevalence within high income countries may just represent better access to healthcare services.

What about specifics in the study? An interesting finding for me was that the years of life lost (YLL) due to premature death and disability was greatest in the sixth decade of life (see figure below). While, the years of life lived with disability (YLD) rises to a peak at age 55, then plateaus before climbing again in the 80s. The latter is not to say that MS increases in those older than 80!

YLD, YLL by age

Secondly, the female and male prevalence. The two in fact seem to diverge in adolescence, while among the preteen children, the occurrence of MS is similar to that between boys and girls (see figure below). Puberty comes to mind as a possible explanation.

Age-standardised prevalence of MS by sex

Multiple sclerosis is the most common inflammatory neurological disease in young adults. The Global Burden of Diseases, Injuries, and Risk Factors Study (GBD) provides a systematic method of quantifying various effects of a given condition by demographic variables and geography. In this systematic analysis, we quantified the global burden of multiple sclerosis and its relationship with country development level.

We assessed the epidemiology of multiple sclerosis from 1990 to 2016. Epidemiological outcomes for multiple sclerosis were modelled with DisMod-MR version 2.1, a Bayesian meta-regression framework widely used in GBD epidemiological modelling. Assessment of multiple sclerosis as the cause of death was based on 13 110 site-years of vital registration data analysed in the GBD’s cause of death ensemble modelling module, which is designed to choose the optimum combination of mathematical models and predictive covariates based on out-of-sample predictive validity testing. Data on prevalence and deaths are summarised in the indicator, disability-adjusted life-years (DALYs), which was calculated as the sum of years of life lost (YLLs) and years of life lived with a disability. We used the Socio-demographic Index, a composite indicator of income per person, years of education, and fertility, to assess relations with development level.

In 2016, there were 2 221 188 prevalent cases of multiple sclerosis (95% uncertainty interval [UI] 2 033 866–2 436 858) globally, which corresponded to a 10·4% (9·1 to 11·8) increase in the age-standardised prevalence since 1990. The highest age-standardised multiple sclerosis prevalence estimates per 100 000 population were in high-income North America (164·6, 95% UI, 153·2 to 177·1), western Europe (127·0, 115·4 to 139·6), and Australasia (91·1, 81·5 to 101·7), and the lowest were in eastern sub-Saharan Africa (3·3, 2·9–3·8), central sub-Saharan African (2·8, 2·4 to 3·1), and Oceania (2·0, 1·71 to 2·29). There were 18 932 deaths due to multiple sclerosis (95% UI 16 577 to 21 033) and 1 151 478 DALYs (968 605 to 1 345 776) due to multiple sclerosis in 2016. Globally, age-standardised death rates decreased significantly (change −11·5%, 95% UI −35·4 to −4·7), whereas the change in age-standardised DALYs was not significant (−4·2%, −16·4 to 0·8). YLLs due to premature death were greatest in the sixth decade of life (22·05, 95% UI 19·08 to 25·34). Changes in age-standardised DALYs assessed with the Socio-demographic Index between 1990 and 2016 were variable.

Multiple sclerosis is not common but is a potentially severe cause of neurological disability throughout adult life. Prevalence has increased substantially in many regions since 1990. These findings will be useful for resource allocation and planning in health services. Many regions worldwide have few or no epidemiological data on multiple sclerosis, and more studies are needed to make more accurate estimates.

About the author

Neuro Doc Gnanapavan


  • Just maybe these figures will make the WHO wake up to the global pandemic of MS and allow our proposed DMTs onto their EML? What do you think?

    • Most definitely, I’m also of the opinion that Pharma should actively push for MS treatments to make it onto the WHO Essential Medicines List.

  • Do Norse / “Viking” genes (part of my genetic heritage and presumably part of my PPMS story) still stand out as risky for MS these days? Or are environmental / lifestyle factors overtaking genetic factors for MS risk?

    • The latitude effect still exists, so yes genes matter. But I don’t think the changes we’re seeing in Asia for instance are genetic…

  • You’ve advised us all that it’s best to use the most effective treatments as early as possible in the disease course but realistically, how many people with MS are on any sort of treatment within the first year after diagnosis? Would five years still be classed as early enough? Due to the complex and variable symptoms involved, it would seem that MS can be so easily misdiagnosed by GPs/physios and the waiting lists to see a neurologist can be long (currently five months in this area and then several weeks after that to get an MRI and follow on appointment).

    • According to most natural history cohorts the first 5y seem to be important. So as long as treatment is commenced then it would have a sizeable impact (the caveat is that there’s not already irreversible disability at presentation). We’re now finding that DMTs even work in the progressive phase, this raises the question of when to try these treatments up to. What worries me are countries that have no treatment, and allowed to get progressively worse.

      • if a DMT works in the progressive phase then why aren’t you working harder to make them available people with SPMS and PPMS. Once you are progressive in the UK then there no hope of any medical assistance to stop the disease advancing. if MS is one disease then moves ought to be made to make treatment available to all. My wheelchair awaits me, currently held up by the snow

  • The map is horrible. An epidemiological study with sense would avoid the borders by countries and introduce data with a little more sense if you want to draw conclusions. I suppose that these data are only used to estimate possible health costs by country. I am geoinformatic and the truth that I would love to be able to make a map with some sense to draw conclusions, but the access to the data is very complicate

    • Yes, it’s difficult to coordinate something like this at a large scale. It’s also why so many years go by before someone else does it. This one was funded by the Gates foundation (Bill & Melinda Gates). I suppose it would be unwise to interpret beyond the most basic of findings!

      • It would be a beautiful project. Collect as much information as possible to draw conclusions through bigdata.
        To establish individuals geographically, where they were from childhood, where they were when they were diagnosed, to be able to cross it with other possible diseases, hours of sun, degree of incidence of UVA rays, general atmospheric currents with main pollutants and a thousand other factors.
        I would like to start a project like that, but I would not know where to start.

  • Whilst it’s good to have this updated study, just how reliable and accurate is it?
    Without countries with the capacity having registers how is it possible to be sure of the numbers involved?
    Even here in the UK whilst there’s a Scottish MS Register there’s no such provision in the rest of the country.
    How can MS be understood and addressed as a global phenomenon,that’s on the increase, without a maximum amount of figures and data?
    Should we be asking charities to campaign for registers?

    • Registers are helpful, but they need to be manadatory. Charities, have advocated this in the UK but may not have the money to sustain it. I’m afraid this needs to be more a national project such as NHS England or equivalent, depending on the country. Again, the elephant in the room is going to be the poor nations that no one wants to challenge.

  • One point about MS prevalence in India:

    The overall low prevalence is misleading, because of how Indian society is structured. Indian society consists of groups that have been endogamous for long periods.

    So, within India there are certain to be communities with a very high prevalence (I’m from such a group, I’m sure)

    Dr David Reich from Harvard has been using genetics to map ancient migrations, prehistory, etc
    He repeatedly makes this point about diseases in India /south Asia: the people are not a single large population like the Chinese. Rather, they’re a more extreme version of the Askenazi Jews of Europe.

    Paper ” The promise of discovering population-specific disease-associated genes in South Asia”

    “The more than 1.5 billion people who live in South Asia are correctly viewed not as a single large population but as many small endogamous groups. We assembled genome-wide data from over 2,800 individuals from over 260 distinct South Asian groups. We identified 81 unique groups, 14 of which had estimated census sizes of more than 1 million, that descend from founder events more extreme than those in Ashkenazi Jews and Finns, both of which have high rates of recessive disease due to founder events. We identified multiple examples of recessive diseases in South Asia that are the result of such founder events. This study highlights an underappreciated opportunity for decreasing disease burden among South Asians through discovery of and testing for recessive disease-associated genes.”

    The above by the way is a limited study. It does not mean there are only 81 such groups. What it means is that other groups have not been studied



Recent Posts

Recent Comments