Relationship of UV exposure to prevalence
of multiple sclerosis in England
A.E. Handel, BMBCh
G. Giovannoni, MD
G.C. Ebers, MD,
G. Chaplin, MSc
Objective: To assess the potential relationship of ultraviolet B radiation (UVB) and Epstein-Barr
virus (EBV) exposure in explaining the period prevalence of multiple sclerosis (MS) in England.
Methods: English national Hospital Episode Statistics covering all admissions to National Health
Service hospitals in England in the 7 years from 1998 to 2005 were used to obtain the period
prevalences of MS and infectious mononucleosis (IM) in England. The United States National
Aeronautics and Space Administration’s data on UVB intensity for England from the Nimbus 7
satellite was collected. The relationships among the 3 variables (MS prevalence, IM prevalence,
and UVB intensity) were investigated.
Results: The regression of MS against UVB intensity for all seasons had an r2of 0.61; when
including the interaction of IM with seasonal UVB, the r2rose to 0.72.
Conclusions: UVB exposure and IM together can explain a substantial proportion of the variance
of MS. The effect of UVB on generating vitamin D seems the most likely candidate for explaining
its relationship with MS. There is a pressing need to investigate the role of vitamin D and EBV and
how they might interact to influence MS risk to identify potential prevention strategies.
EBV ? Epstein-Barr virus; HES ? Hospital Episode Statistics; ICD ? International Classification of Diseases; IM ? infectious
mononucleosis; MS ? multiple sclerosis; NHS ? National Health Service; NASA ? National Aeronautics and Space Adminis-
tration; SHA ? English Strategic Health Authorities; UVB ? ultraviolet B radiation.
Multiple sclerosis (MS) is the most common disease of the CNS in young adults of Northern
European descent.1MS seems unlikely to result from a single causative event; instead, the
disease seems to develop in genetically susceptible populations as a result of environmental
The most striking illustration of the importance of the environment in MS is its geographi-
cal distribution.3Within regions of temperate climate, MS incidence and prevalence are
thought to increase with latitude.3,4The latitudinal trend starts around 42 degrees of latitude
North.5A cutoff at this latitude implicates solar wavelengths in the shorter UV range (ultravi-
olet B [UVB]) as UVB is strongly affected by the solar zenith angle.6There is also a month of
birth effect for MS,7which is also indicative of solar-correlated processes, and this effect is itself
We have recently reported period prevalence values for MS across England.8Epstein-Barr
virus (EBV) is a B-lymphotropic human DNA herpesvirus that infects most individuals asymp-
tomatically but in some people causes infectious mononucleosis (IM) upon infection.9Because
the risk of MS is increased in individuals with a clinical history of IM,10-12we also compared the
geography of MS with the distribution of IM, and observed a strong correlation, but this was
From the Wellcome Trust Centre for Human Genetics (S.V.R., A.E.H., G.C.E.) and Department of Clinical Neurology (S.V.R., A.E.H., G.C.E.),
University of Oxford, Oxford; Blizard Institute of Cell and Molecular Science (S.V.R., G.G.), Queen Mary University of London, Barts and The
London School of Medicine and Dentistry, London, UK; and Department of Anthropology (S.R.S., G.C.), The Pennsylvania State University,
Study funding: Supported by the Multiple Sclerosis Society of Great Britain and Northern Ireland, the Medical Research Council, and the Wellcome
Disclosure: Author disclosures are provided at the end of the article.
Address correspondence and reprint
requests to Dr. George Chaplin,
Department of Anthropology, 519
Carpenter Bldg., The Pennsylvania
State University, University Park,
Dr. George C. Ebers, University
Department of Clinical Neurology,
Level 3, West Wing, John Radcliffe
Hospital, Oxford, OX3 9DU, UK
Copyright © 2011 by AAN Enterprises, Inc.
not enough to fully account for the variance
of MS. Recent studies have documented the
inverse relationship between MS prevalence
and UV exposure.13In the present study, we
assess how the prevalence of MS and IM re-
lates to UVB exposure.
METHODS Standard protocol approvals, registra-
tions, and patient consents. The English NHS Central Of-
fice for Research Ethics approved this study.
MS and IM prevalence data. As published, we analyzed data
on hospital admissions for MS from linked hospital admission
statistics, assembled from the English national Hospital Episode
Statistics (HES) system.8This covered all admissions to National
Health Service (NHS) hospitals in England in the 7 years from
April 1, 1998, to March 31, 2005 (population: 51 million). We
identified cases of MS as code G35 in the 10th revision of the
International Classification of Diseases (ICD). The English na-
tional linked HES database includes information about all peo-
ple who are admitted to hospital (including day care as well as
overnight stays). Using data linkage, we identified each person
only once for MS, regardless of how many admissions each per-
son had, and recorded their residence out of 28 English Strategic
Health Authorities (SHAs) at first known admission for MS.
Admission rates were calculated using numbers of admissions for
residents of each area as the numerators and the total resident
populations of the area as the denominators. To adjust for differ-
ences in the age structure of different areas, age standardization
was undertaken using the indirect method and the age-specific
rates in 5-year age groups in the English population as the stan-
dard. We undertook the same calculations for hospital admis-
sionsfor peoplewith IM,
cytomegalovirus infections using the ICD codes B27, B01, and
B25 to identify them.
varicella infection, and
UV. We used the US National Aeronautics and Space Adminis-
tration’s (NASA’s) data for the single wavelength of 305 nm at
noon (J.R. Herman, 2003, NASA Goddard Centre, Greenbelt,
MD). This single wavelength data slightly overestimates the
amount of UVB at its prime physiologically active wavelengths
which are 290–295 nm. The data also tend to overestimate
ground availability of UVB by approximately 15% in areas
where there is dense optical pollution from particulates and aero-
sols formed from oxides of sulfur and nitrogen. The data model
has allowances for natural clouds but the sensor will read the
reflection from the top of polluted cloud formations, thereby
overestimating ground values.14The magnitude of this error is
not expected to be large for the United Kingdom because it is on
the oceanic side of the prevailing wind and so does not carry
much pollution. The error will be greater in urban areas. The
NASA data for the single wavelength of 305 nm at noon col-
lected on the Nimbus 7 satellite from 1978–1992 was sampled
for 5 days centered on the 22nd of each month and averaged
over the whole period. This data reduction was necessary due to
the enormous size of the dataset comprising more than 60,000
readings a day for 14 years. This solar month sampling also has
the advantage of catching the full range of the annual variation
because it captures the solstices and overcomes the problem of
differing numbers of days per calendar month. The data were
further averaged into seasons and integrated for the area of each
of the 28 SHAs using ArcGIS (ESRI, 2009, Redlands, CA, 9.3.1
ArcInfo). See figure 1 for a map of spring season UVB 305 nm
for the SHAs of England.
Analysis. The data were subjected to exploratory data analysis,
consisting of inspecting the correlation matrix and evaluation
using principal component analysis. The final analysis was a
least-squares regression analysis. To test for geographic depen-
dencies the data were tested by looking at quantile–quantile
plots of the residuals, and performing geographically weighted
regression. The data were considered suitable for ordinary least
squares regression, ArcGIS, and Geospatial Analyst (ESRI 2009)
and S-Plus (Insightful Corp., S-PLUS 6.2 , Insightful
Corp., Seattle, WA).
RESULTS Correlations between seasonal UVB and
MS and IM. The correlations between seasonal UVB
intensity with latitude and both MS and IM preva-
lence are shown in table 1. As shown previously, MS
was highly correlated with IM (0.69). As a control,
we also examined the correlations of MS with cyto-
megalovirus prevalence (0.05), and MS with varicella
prevalence (0.21), which were both nonsignificant.
Annual average UVB was not correlated with IM
(0.02) nor with cytomegalovirus (0.19), and was
only moderately correlated with varicella (?0.41),
whereas the annual average UVB is more weakly cor-
related with MS (?0.29). The correlation between
MS and latitude was stronger than the correlation
between IM and latitude (0.46 and 0.16, respec-
Figure 1Map of ultraviolet B radiation 305 nm spring from total ozone
mapping spectrometer averaged for each of 28 English Strategic
Neurology 76 April 19, 2011
tively). UVB in any season was more highly corre-
lated with MS than with IM. Therefore, the seasonal
pattern of UVB is more important in MS. The most
highly correlated UVB seasons with MS distinguish
between onset and prevalence were spring and au-
tumn (?0.46 and ?0.48, respectively).
Principal components analysis. The first 2 compo-
nents in the principal components analysis together
accounted for 99.7% of the variance for MS (table
2). The interaction term for MS and IM together
with spring UVB were the most important factors in
determining MS variance. Next were the interaction
terms for MS and IM together with summer and
autumn UVB. Of more minor importance were the
interaction terms for MS on its own with spring,
summer, or autumn UVB as well as IM with spring
UVB. The way MS and IM interact with UVB is
more important to MS prevalence than just MS and
Multivariate regression. TheregressionofMSagainsta
single season UVB had r2values ranging between 0.03
3). The regression of MS against UVB for all seasons
had an r2of 0.61 (p ? 0.0001). When including the
(p ? 0.001). The regression line for the relationship of
UVB to MS for all seasons is a very good fit to the data
(figure 2A) and the residuals show little nonrandom
been explained by this model.
DISCUSSION We show that the distribution of MS
across England is explained both by UVB exposure
and the prevalence of IM. In England, there is a def-
inite latitudinal cline in MS occurrence but a weaker
cline in IM, and as expected, IM prevalence could
not entirely explain the variance of MS. Incorporat-
ing a linear model for MS with an interaction term
for IM and UVB could explain 72% of the variance
of MS prevalence across England. Interestingly, it is
spring UVB that is most strongly associated. Lower
levels of UVB in the spring season would coincide
with late gestation for offspring born in late spring/
early summer. This corresponds with the time for
peak risk of MS by month of birth.7
There are several possible mechanisms through
which UVB radiation could be mediating the effect
Table 1Correlations between seasonal UVB with latitude, MS, and
?0.170.101.00 0.880.95 0.90
?0.130.88 1.000.95 0.96
?0.07 0.950.951.000 0.97
Abbreviations: IM ? infectious mononucleosis; MS ? multiple sclerosis; UVB ? ultraviolet B
Table 2Principal component analysis for MS
MS: IM: winter
MS: IM: spring
MS: IM: summer
MS: IM: autumn
Abbreviations: IM ? infectious mononucleosis; MS ? multi-
Table 3 Regression resultsa
Abbreviation: IM ? infectious mononucleosis.
aRegression model multiple sclerosis ? IM ? (winter ?
spring ? summer ? autumn). Multiple R2: 0.72 p value
Neurology 76April 19, 2011
on MS risk. The most attractive explanation is UVB-
induced synthesis of cutaneous vitamin D, which is
the principal source of this important sterol hor-
mone.15,16It has been suggested that low vitamin D
levels may lead to an increase in EBV infection. The
pleiotropic roles of vitamin D on the immune sys-
tem17may lead to an abnormal or variant response to
EBV infection as manifested by IM when an individ-
ual is vitamin D deficient.18
There are limitations to our study. Limitations of
record linkage studies using routinely collected ad-
ministrative data are well-known, and include the
facts that the data are limited to hospitalized patients
and that information about some variables of poten-
tial interest, such as social circumstances and ethnic-
ity, are generally unavailable. The pattern of MS or
of UVB and MS is confounded in the large metro-
politan areas of Merseyside and Manchester through
to Leeds, together with Birmingham and London.
Geographically weighted regression showed these re-
gions to be 2 standard deviations below the main
trend. These conurbations have larger populations of
people with lower genetic susceptibility to MS.
The size of the correlations can be misleading al-
though they are strong for UVB. In any case where
there are putative gene–environment interactions,
and where both elements are required, there will in-
evitably be the presence of one or other in the ab-
sence of the outcome being studied.
It is unclear as to how generalizable our data are to
genetically dissimilar regions or to different insola-
tion climates across the world. There are well-known
exceptions to the latitude gradient, for example, Sar-
dinia.3The relationships shown represent associa-
tions and causative inferences must be tentative. In
the case of IM in particular, the age dependency of
the IM phenotype intersects with early life hygiene,
social class, and other variables that may account for
secondary associations, leaving the primary associa-
tion speculative. However, when our data are taken
in combination with others,13,19,20it gives confidence
that there is a pressing need to investigate the role of
vitamin D and IM and their interaction in the patho-
genesis of MS. At present there are no effective vac-
cines or treatments targeting EBV, but their
development could complement vitamin D supple-
mentation as potential interventions that might re-
sult in a reduction in the prevalence of this often
Statistical analysis was conducted by Dr. Sreeram Ramagopalan.
Dr. Ramagopalan receives research support from the Multiple Sclerosis Soci-
ety of Canada Scientific Research Foundation and the Multiple Sclerosis So-
ciety of the United Kingdom. Dr. Handel reports no disclosures. Dr.
Idec and Vertex Pharmaceuticals; served on the editorial board of Multiple
Sclerosis; has received speaker honoraria from Bayer Schering Pharma, Merck
Serono, Biogen Idec, Pfizer Inc, Teva Pharmaceutical Industries Ltd.–sanofi-
aventis, Vertex Pharmaceuticals, Genzyme Corporation, Ironwood, and No-
vartis; has served as a consultant for Bayer Schering Pharma, Biogen Idec,
GlaxoSmithKline, Merck Serono, Protein Discovery Laboratories, Teva
GW Pharma, Novartis, and FivePrime; serves on the speakers bureau for
Merck Serono; and has received research support from Bayer Schering
Pharma, Biogen Idec, Merck Serono, Novartis, UCB, Merz Pharmaceuticals,
LLC, Teva Pharmaceutical Industries Ltd.–sanofi-aventis, GW Pharma, and
Ironwood. Dr. Siegel reports no disclosures. Dr. Ebers serves on the editorial
boards of the International Multiple Sclerosis Journal and Multiple Sclerosis and
speaker honoraria from Bayer Schering Pharma, sanofi-aventis, Roche, and
UCB; has served as a consultant to Biopartners, Bayer Schering Pharma,
Howrey LLP, Heron Health, and Eli Lilly and Company; and receives re-
Figure 2 Response of multiple sclerosis to the effect of infectious
mononucleosis and ultraviolet B radiation (A) and residuals to the
effect of infectious mononucleosis and ultraviolet B radiation
against a QQ normal plot (B)
Neurology 76 April 19, 2011
the United Kingdom, and the Multiple Sclerosis Society of Canada Scientific
Research Foundation. G. Chaplin reports no disclosures.
Received November 9, 2010. Accepted in final form January 6, 2011.
1. Noseworthy JH, Lucchinetti C, Rodriguez M, Weinshen-
ker BG. Multiple sclerosis. N Engl J Med 2000;343:938–
2. Ebers GC. Environmental factors and multiple sclerosis.
Lancet Neurol 2008;7:268–277.
3.Pugliatti M, Sotgiu S, Rosati G. The worldwide prevalence
of multiple sclerosis. Clin Neurol Neurosurg 2002;104:
4.Taylor BV, Pearson JF, Clarke G, et al. MS prevalence in
New Zealand, an ethnically and latitudinally diverse coun-
try. Mult Scler 2010;16:1422–1431.
5.Ascherio A, Munger KL. Environmental risk factors for
multiple sclerosis: part II: noninfectious factors. Ann Neu-
6. Johnson FS, Mo T, Green AES. Average latitudinal varia-
tion in ultraviolet radiation at the Earth’s surface. Pho-
tochem Photobiol 1976;23:179–188.
7. Willer CJ, Dyment DA, Sadovnick AD, Rothwell PM,
Murray TJ, Ebers GC. Timing of birth and risk of multi-
ple sclerosis: population based study. BMJ 2005;330:120.
8.Ramagopalan SV, Hoang U, Seagroatt V, et al. Geography
of hospital admissions for multiple sclerosis in England
and comparison with the geography of hospital admissions
for infectious mononucleosis: a descriptive study. J Neurol
Neurosurg Psychiatry Epub 2011 Jan 6.
9. Crawford DH. Biology and disease associations of Epstein-
Barr virus. Philos Trans R Soc Lond B Biol Sci 2001;356:
10. Handel AE, Williamson AJ, Disanto G, Handunnetthi L,
Giovannoni G, Ramagopalan SV. An updated meta-
analysis of risk of multiple sclerosis following infectious
mononucleosis. PLoS One 2010;5.
Goldacre MJ, Wotton CJ, Seagroatt V, Yeates D. Multiple
sclerosis after infectious mononucleosis: record linkage
study. J Epidemiol Community Health 2004;58:1032–
Lindberg C, Andersen O, Vahlne A, Dalton M, Run-
marker B. Epidemiological investigation of the association
between infectious mononucleosis and multiple sclerosis.
Sloka JS, Pryse-Phillips WE, Stefanelli M. The relation of
ultraviolet radiation and multiple sclerosis in Newfound-
land. Can J Neurol Sci 2008;35:69–74.
Xu M, Liang XZ, Gao W, Krotkov N. Comparison of
TOMS retrievals and UVMRP measurements of surface
spectral UV radiation in the United States. Atmos Chem
Phys Discuss 2010;10:10969–11008.
Handel AE, Giovannoni G, Ebers GC, Ramagopalan SV.
Environmental factors and their timing in adult-onset
multiple sclerosis. Nat Rev Neurol 2010;6:156–166.
Holick MF. Vitamin D deficiency. N Engl J Med 2007;
Ramagopalan SV, Heger A, Berlanga AJ, et al. A ChIP-seq
defined genome-wide map of vitamin D receptor binding:
associations with disease and evolution. Genome Res
Grant WB. Latitude and multiple sclerosis prevalence: vi-
tamin D reduces risk of Epstein-Barr virus infection. Mult
Scler 2010;16:373; author reply 374–375.
van der Mei IA, Ponsonby AL, Blizzard L, Dwyer T. Re-
gional variation in multiple sclerosis prevalence in Austra-
lia and its association with ambient ultraviolet radiation.
Beretich BD, Beretich TM. Explaining multiple sclerosis
prevalence by ultraviolet exposure: a geospatial analysis.
Mult Scler 2009;15:891–898.
Ascherio A, Munger KL, Simon KC. Vitamin D and mul-
tiple sclerosis. Lancet Neurol 2010;9:599–612.
Historical Abstract: April 1, 1985
NEURONAL ANTINUCLEAR ANTIBODY IN SENSORY NEURONOPATHY FROM LUNG CANCER
Francesc Graus, Carlos Cordon-Cardo, and Jerome B. Posner
We found an antinuclear antibody highly restricted to nuclei of neurons in two patients with subacute sensory neuronopathy
complicating oat cell carcinoma of the lung. Serum was tested by indirect immunofluorescence and immunoperoxidase staining. At
low concentrations of antibody, only the nuclei of the neurons were stained. At high concentrations, there was also staining of the
nuclei of glial cells and fetal nonneural tissues. The cytoplasm of most neurons was stained with the immunoperoxidase method.
Free Access to this article at www.neurology.org/content/35/4/538
Comment from Richard M. Ransohoff, MD, Associate Editor: A groundbreaking demonstration that paraneoplastic
syndromes were associated with antibodies to neuronal components.
Neurology 76April 19, 2011