When MS Risk Factors Multiply With Genotype Instead of Adding
Most reviews of multiple sclerosis (MS) risk factors list them. This one is built around a different idea: that certain environmental exposures multiply with HLA genotype rather than add to it, and that the size of the combined effect is itself evidence about mechanism. Olsson and colleagues assemble 13 lifestyle and environmental factors into a single table carrying, for each one, the odds ratio, whether an interaction with HLA genes exists, the combined odds ratio when both are present, whether the effect falls during adolescence, and a graded level of evidence. They also state their inclusion rule up front, concentrating on factors from adequately powered studies and particularly those independently confirmed, and they list the biases that make observational work in this area hard: reverse causation, where the disease produces the association; residual confounding; recall bias in how exposures are reported; and selection bias.
Three Rows Carry the Whole Argument
Smoking on its own gives an odds ratio of about 1.6, which is a modest number. In a person carrying HLA-DRB1*15:01 and lacking the protective HLA-A*02, the combined odds ratio is 14. Adolescent obesity is about 2 alone and roughly 15 combined. Epstein-Barr virus seropositivity is about 3.6 alone and roughly 15 combined. Infectious mononucleosis gives 2 alone and 7 combined, and even passive smoking moves from 1.3 to 6. For scale, HLA-DRB1*15:01 by itself carries an odds ratio near 3, and combining it with the absence of HLA-A*02 reaches about 5. The exposures are individually unremarkable and the genotype alone is moderate; the large numbers appear only when the two are present together. That pattern is the review's evidence that the exposure and the risk allele act on the same pathogenic pathway rather than on two separate ones.
What Does Not Interact, and Why That Is Informative
The table is as useful for its blank cells. Vitamin D below 50 nM carries an odds ratio around 1.4 with no HLA interaction recorded. Low sun exposure sits at about 2, night work at 1.7, cytomegalovirus seropositivity at 0.7, oral tobacco at 0.5, alcohol at 0.6, and coffee at 0.7, none of them with an interaction. That division matters under the review's own logic, since interaction with HLA is read as a signal that a factor works through adaptive immunity, which is what HLA molecules serve. Vitamin D was in fact proposed as the first example of a gene-environment interaction involving HLA-DRB1*15:01, on the basis of in vitro work, and the review says plainly that the finding was not reproduced in case-control studies. Leaving that on the record rather than quietly dropping it is the right handling of a claim that did not survive.
The Timing Argument, and an Internal Control on Smoking
Several of these factors only work inside a window. Obesity at a body mass index above 27 around age 20 raises risk, while high body mass index at diagnosis has no influence and high body mass index at age 10 is not associated with future disease. Epstein-Barr virus matters when infection arrives in adolescence or later, not in childhood. Shift work before age 20 raises risk about 1.7-fold, with weaker associations after 20. Migration data point the same way, since people moving from a low-risk to a high-risk country before adolescence carry a risk resembling those born there. The smoking story comes with its own internal control. Oral tobacco, common in Sweden and therefore separable from smoke in that population, is associated with reduced risk in a dose-dependent way at an odds ratio of 0.5. Nicotine may be protective through the α7 nicotinic receptor subunit on immune cells. The contrast between inhaled smoke and oral nicotine is what supports lung irritation, rather than nicotine, as the mechanism driving risk upward.
What Follows for Prevention, With the Numbers Attached
This is where the review stops describing and starts recommending, and the attributable fractions are why. On Swedish data, 20.4% of all MS cases are estimated attributable to tobacco smoke exposure. Among people carrying HLA-DRB1*15:01 and lacking HLA-A*02, 41% of cases are attributable to smoking. The recommendations follow from those two figures. People with a family history of MS should be informed about the effect of smoking and about protecting their children from passive smoke. Vitamin D supplementation at relatively high doses is described as recommendable for all adolescents, on the grounds that it is non-toxic up to high levels with no upper limit yet defined. Avoiding night shift work is named as a further modifiable factor for those at high risk. The authors are equally clear about what they will not recommend, writing that the data give no argument for people at risk to abstain from alcohol, and that the coffee evidence is too weak to support any recommendation.
Where the Review Pushes Back on Its Own Field
Two passages are worth reading for their restraint. On dietary salt, which had a prominent run as a proposed MS trigger, the review notes that the salt intake used in the mouse experiments corresponded to more than 500 g of salt daily in humans, a point raised in a published comment on the original paper, and that human studies of salt intake are notoriously difficult to design. On Epstein-Barr virus it lays out the competing reading in full: raised antibody titres against EBNA1 could be a consequence of poor viral elimination owing to insufficient cell-mediated immunity in people whose immune-genetic setup predisposes them to MS, rather than evidence that the virus triggers the disease. It adds that EBV-related RNA found in central nervous system infiltrates in one study was only partially reproduced by two independent groups. On gut microbiota, human data are described as largely lacking, and any organism found in patients as arguable epiphenomenon or consequence of disease.
How Much of the Risk This Accounts For, and What It Rests On
One number in the review deserves more attention than it usually gets. Sibling relative risk in MS was previously estimated at almost 17-fold; recent population-based studies put it at 5 to 7-fold, which the authors read as showing a much lower importance of genetic predisposition and a larger role for environment. They add that the rise in MS among women across recent decades has been too fast for changes in genetic composition to explain, pointing instead to smoking, obesity, and changes in reproductive behaviour. Two things a reader should weigh against all of the above. The evidence grades in the table are uneven: four factors carry the top grade, four the middle, and five sit at the lowest grade, which the authors define explicitly as observations not yet independently confirmed, included to enable further study. And a large share of the interaction findings, the smoking and obesity numbers among them, come from Scandinavian case-control studies run by these authors' own groups. That does not weaken the findings, but it does make independent confirmation in other populations the obvious next requirement rather than an optional extra.
Disclaimer: This blog post is based on the cited review article and is intended for informational purposes only. It is not intended to provide medical advice. Please consult with a healthcare professional for any health concerns.
Reference:
Olsson, T., Barcellos, L. F., & Alfredsson, L. (2017). Interactions between genetic, lifestyle and environmental risk factors for multiple sclerosis. Nature Reviews Neurology, 13(1), 25–36. https://doi.org/10.1038/nrneurol.2016.187
Primary studies named above are cited as reported and discussed within this review; see its reference list for full details.
