Content uploaded by Rune Midgard
Author content
All content in this area was uploaded by Rune Midgard on Apr 29, 2016
Content may be subject to copyright.
Aeta
Neurol Scand
1996:
93:
307-314
Printed in
UK
-
all
rights reserved
Copyright
0
Munksgaard
1996
ACTA
NEUROLOGICA
SCANDINAVICA
ISSN
0001-6311
Disability and mortality in multiple sclerosis
in Western Norway
Midgard
R,
Riise T, Kvile
G,
Nyland
H.
Disability and mortality
in
multiple sclerosis in Western Norway
Acta Neurol Scand 1996: 93: 307-314.
0
Munksgaard 1996.
Introduction
-
Continued studies of frequency trends in carefully selected
sites around
the
world can provide clues
to
the
cause
of
multiple sclerosis
(MS).
Material and methods
-
Based
on information
from
three different,
semi-independant sources
of
infomation,
we
have examined the temporal
trends
in
the average annual age-adjusted rates
of
disability pension
incidence,
mortality,
and incidence
of
MS
from
1966
to
1991 in Merre and
Romsdal County, Norway.
Results
-
The average annual age-adjusted
disability pension incidence rates (1966-68
=
3.62/100,000; 1990-91
=
7.33/100,000), the mortality rates (1966-68 =0.91/100,000; 1990-91
=
1.88/100,000), and
the
incidence rates (1966-68 =4.22/100,000; 1990-91
=
5.02/100,000) all
showed
a
statistically significant increase. The difference
in the development of MS-specific disability pension prevalence rates in
the county compared
to
the
nation
is
notable.
Conclusions
-
We
consider
that
the
increase
in
disability
pension incidence, mortality, and incidence
of
MS
is
of biological significance.Thus three different sources of
information corroborate corresponding trends indicating that better case
ascertainment and improved diagnostic facilities only partially can explain
the reported MS increase
in
western Norway.
The mortality time trend of MS (1-3) and the
prevalence of MS patients drawing a disablement
benefit in Norway have shown a stable geographic
distribution with high frequencies in eastern parts
of the country and lower frequencies in western and
northern parts
(2).
Prevalence studies in Norwegian
counties show a corresponding geographic differ-
ence (4,
5).
Recent studies of the prevalence and
incidence rates of MS in Hordaland and Msre and
Romsdal counties in western Norway, however,
show a two- to threefold increase in the disease
frequency (4-6).
Although indirect, the annual mortality rates are
helpful in studying temporal trends and geograph-
ical differences in
MS
frequency (7-9). A steady
decline has been reported in western industrialized
nations (10, 1
I)
in contrast to the stable mortality
rate in Norway (1-3).
The purpose of the present study was to examine
if the observed rise in MS frequency during the past
40 years in Msre and Romsdal County is reflected
in the frequency of the disability pensions and the
mortality rates.
R.
Midgard
',
T.
Riise
',
G.
Kv%le
',
H.
Nyland
'
'Departments of'Neurology Molde County Hospital,
'Occupational Medicine, 3Epidemiology. 'Neurology.
University
of
Bergen. Norway.
Key
words
multiple sclerosis. disability. mortality,
incidence, time trends
R
Midgard, Department
of
Neurology, Molde
County Hospital, N-6400 Molde. Norway
Accepted for publication November
23.
1995
Material and methods
Disability
pension
A disability pension
is
given to
a
person between
16 and 66 years of age whose working capacity is
permanently reduced by at least
50%
-
due to
illness, injury or defect
-
after the termination of
medical treatment, vocational training, education
or other appropriate forms of rehabilitation. The
disability pension was legally fully implemented in
Norway in 1967. Initially persons aged 18-69 years
could obtain the pensions, but from 1973 the age
span changed to 16-66 years. Persons with chronic
diseases
of
the central nervous system are as a rule
examined by a neurologist before a disability pen-
sion is granted.
The National Insurance Administration provided
information concerning each year's new recipients
of
disability pension due to
MS
among the inhabit-
ants living in Msre and Romsdal County from 1966
to 199 1. Name, 1 1 -digit person identification
number, and month and year of granted disability
pension due to
MS
as the primary and secondary
diagnoses classified according
to
ICD-9
were given
307
Midgard
et
al.
for each patient. Two-hundred and eighty-nine MS
patients,
11 1
men and 178 women, had received
disability pension due to
MS
during the study
period. The average annual age-adjusted disability
pension incidence rates for Msre and Romsdal
County were calculated for each 3-year period from
1966-68 to 1987-89 and for the period 1990-91
separately. We also calculated the crude annual rate
of disability pension incidence from 1968 to 1990
as 3-year centered moving averages.
The total number of disability pension recipients
comprising all medical diagnoses by the end of each
year in Msre and Romsdal County and in Norway
during the same period was also provided. The
prevalence rates
of
disability pensions by the end
of
each year in More and Romsdal County and in
Norway due to MS; the MS-specific disability pen-
sion prevalence rate, and to all causes; the total
disability pension prevalence rate, were calculated
for each year from 1967 to 1991.
Mortality
Information concerning deaths with MS as under-
lying or contributing cause was supplied by the
Registry of Death Causes, Statistics Norway. Name,
1
1-digit person identification number, date
of
death,
and diagnostic classification on the death certificates
were supplied for each patient. The International
Classification of Diseases in different revisions (ICD
7-8-9) have been used during the study period with
no significant change in the mortality coding sys-
tems being used. Less than 0.25%
of
all deaths in
Norway are classified as of unknown cause due to
lack of information. One-hundred and thirteen
patients, 53 men and 60 women, died from 1966 to
1991 with MS as underlying or contributing cause
of death. The average annual age-adjusted mortality
rates in More and Romsdal County due to MS
were calculated for each 3-year period from 1966-68
to 1987-89 and for the period 1990-91 separately.
Crude annual death rates from 1968 to
1990
were
also calculated as 3-year centered moving averages.
The Registry of Death Causes, Statistics Norway
provided the ICD-codes from the death certificates
of all deceased patients as assessed in the clinical
material in the county. A comparison between
deaths with MS in the clinical material and deaths
with MS coded on their death certificates according
to the official mortality statistics was performed to
examine the completeness of recording on the death
certificates. The sensitivity of the clinical assess-
ments vs. the death certificates was also evaluated,
and the positive predictive value
(PPV)
of the death
certificate diagnoses was estimated.
Incidence
The previous studies of incidence of
MS
in More
and Romsdal County (5,12) were updated. All new
cases with clinical onset
of
MS before 31 December
1991 coming to our attention before 31 August
1992 were included. Onset
of
disease was defined as
the year of onset of symptoms. Year of diagnosis
was defined as the year when the diagnosis of MS
based on clinical criteria was made by a neurologist.
The sources
of
information are outlined in the most
recent survey of the county
(5).
The official files of
the National Insurance Administration and the
Registry of Death Causes, Statistics Norway served
as supplementary sources in the final case ascertain-
ment. Three-hundred and twenty-six MS patients,
130 men and 196 women, with onset of disease in
Marre and Romsdal County from 1966 to 1991 were
included. According to McAlpine’s clinical criteria
(
13) 189 were definite, 68 probable, and 69 possible
MS. The age-adjusted average annual incidence
rates in Msre and Romsdal County due to MS
were calculated for each 3-year period from 1966-68
to 1987-89 and for the period 1990-91 separately
including all diagnostic categories. We also calcu-
lated the annual incidence rates in the period
1968-90 as 3-year centered moving averages.
Statistics
The Chi-square test for linear trend (Extended
Mantel-Haenszel)( 14) was used to test the statist-
ical significance of the increase in disability pension
rates, mortality, and incidence. The tests for linear
trend were performed on the crude annual rates
from 1967 to 1991. The Chi-square test for 2-by-2
contingency tables was used to test the statistical
difference between the disability pension prevalence
rates in the county and the nation
in
1967 and in
1991. The statistical analyses were performed by
means
of
the BMDP software packages (1
5),
and
the Epi Info v.5 (16).
The age adjustment was done with the method
of direct standardization with the European
Standard as the reference population (17). The
population in the county in the middle of each
triennial was provided by Statistics Norway. The
population in the study increased from 221,173 in
1966-68 to 238,810 in 1990-91.
Results
Disability
The total number
of
MS cases receiving disablement
benefits, the surveyed population, and the average
annual age-adjusted disability pension incidence
rates due to MS in 3-years periods are summarized
MS
disability and mortality
Table 1. Multiple sclerosis in Msre and Romsdal. Norway 1966-91
Average annual age-adjusted' rates' in disability pension incidence, mortality,
and incidence.
Disability
County pension
population
incidence
Mortality Incidence
Mean in the period
No Rate No Rate
No Rate
~~~
1966-68
1969-71
1972-74
1975-77
1978-80
1981 -83
1984-86
1987-89
1990-91
Chi square for
linear trend
1966-91
221,173
223,360
227.737
231,944
2
3 4,9 9 3
236,062
237,396
237.873
238,810
______
24
362
8 091 28 422
31 463
12 131 30 448
32 468
11 135 19 278
28 402 11 121 43 618
35 496
15 185 47
667
38 537 8 086 47
664
30 421 13 152
45 632
36 504 26 273 43 603
35 733 9 188 24
502
p=OO2
p=005 p=002
'Age-adjusted by the direct standardization method with the European standard as
the reference population.
'Rates per 100.000 per year.
in Table
1.
The observed increase in average annual
age-adjusted disability pension incidence rates was
statistically significant when tested for linear trend
(Chi square for linear trend
=
5.39; p =0.02).
The average annual age-adjusted disability pen-
sion incidence rates by sex from 1966 to 1991
showed fluctuations with rate variation from 2.4 to
5.1 per 100,000 per year for men, and from 4.2 to
10.6 per 100,000 per year for women.
The total disability pension prevalence rate includ-
ing all medical diagnoses by the end
of
1967
was statistically significantly higher in Mare
and Romsdal County compared to Norway (Chi-
square=41.45; p<O.OOOl) (Fig. 1). From 1967 to
1972 the national and the county prevalence rates
showed a similar increase. The abrupt and parallel
decrease from 1972 to 1973 was due to the change
in age groups made in 1973. From 1974 the national
rates increased to a larger extent than the county
rates when comparing the annual end-of-year total
disability pension prevalence rates (Fig. 1). By the
end
of
1991 the difference between the total disability
pension prevalence rates was still highly statistically
significant (Chi-square= 171.3; p<O.OOOl), but by
this time the national rates had become the highest.
The MS-specific disability pension prevalence
rates by the end of each year, however, showed
exactly the opposite trend when comparing the
county to the country (Fig. 2). The difference
between the MS-specific disability pension preval-
ence rates in Mare and Romsdal and in Norway
by the end of 1967 did not reach statistical signific-
ance (Chi-square
=
3.39;
p
=
0.07). From 1967 to
198
1
the MS-specific disability pension prevalence
rates ran almost parallel in the county and in the
nation with the national rates slightly higher. From
1982 a steady increase in the MS-specific disability
pension prevalence rates in the county was observed,
and from 1983 the county rates exceed the national
rates in sharp contrast to the trend in the total
I
-Msre
and Romsdal
1
OC
----+--
+
--
.-+
i-
-,-,
---
+
+-p,
196'1
1968
1969
1970
1971
I972
1973
1974
197s
1976
19?7 1978
1979 19wI
1981
I982
1983
1984
1983
1986
19a
1)tll
1989
1990
1991
Years
Fig.
1.
Total
disability pension prevalence rates in Mme and Romsdal County and in Norway by the end
of
each year (rates/
l,OOO/
year) in 1967-91.
309
Midgard
et
al.
80
70
60
I.
h
L
W
P
0 0
J
50
3
40
-
L
t
3
2
30
20
10
0
1967 1969 1971 1973
1975
1977 1979 1981 1983 1985 1987 1989
1991
Years
I
+Nomay
mere
and Romsdal
~
Fig.
2.
MS-specific disability pension prevalence rates in M0re and Romsdal County and in Norway by the end
of
each year (rates/
lOO,OOO/ year) in
1967-91.
disability pension prevalence rates.
By
the end of
1991 the MS-specific disability pension prevalence
rate in Msre and Romsdal County was statistically
significantly higher compared to the corresponding
rate in Norway (Chi-square
=
7.45; p =0.006).
Mortality
The total number of deceased MS patients, and the
average annual age-adjusted mortality rates in
3-years periods are shown in Table 1. When tested
for linear trend, the increase in crude annual
MS-specific mortality rate from 1967 to 1991
reached statistical significance (Chi-square for linear
trend
=
3.82; p
=0.05).
The linear trend test did not
show statistical significance when applied to the
national rates from 1967 to 1991 (p=0.46) con-
trasting the significant increase in linear trend in
the MS-specific mortality in Msre and Romsdal
County. Due to larger numbers the national crude
annual MS-specific mortality rates indicate a more
stable trend throughout the period, while large
Auctuations are seen in the crude annual mortality
rates of More and Romsdal County (Fig. 3).
When comparing the mortality in the clinically
assessed material with the Registry of Death Causes,
Statistics Norway, 120 deceased
MS
cases were
found in the clinical material, but only 91 of these
cases (75.8%) had MS coded as underlying or
contributing cause of disease
on
the death certific-
ate. In
29
cases
(24.2%)
MS
was not mentioned as
a cause
of
death. The search into the Registry
of
Death Causes revealed 103 cases with the diagnosis
of
MS
on the death certificate from 1966 to 1991.
Ten of these were not known from the clinical
material, but could be traced in the hospital record
files. The 10 cases that were lacking in the clinical
material represent sporadically distributed cases
from 1969 to 1987. Two of the 103 cases that could
be tracked in the hospital record files were found
to be misclassified on their death certificates
(1
muscle dystrophy, 1 cerebral palsy). After this sup-
plement 101 (77.7%) of the 130 clinically defined
cases were coded with
MS
on the death certificate.
Thus, the sensitivity of the death certificates was
0.777 (101/130). Reversely, taking the death certi-
ficate diagnosis as “the gold standard”, the sensitiv-
ity of the clinical assessments was 0.901 (91/101).
The positive predictive value of
a
death certificate
diagnosis was 0.981 (101/103).
Incidence
The number
of
new MS cases, and the average
annual age-adjusted incidence rates in 3-years
310
MS
disability
and
mortality
1
1
i
0
ITI-r
1
-
7-
T--v
-7-
1968
1971
1974 1977
1980
1983 1986
1989
Calender
year
Fig.
3.
MS-specific crude annual mortality rates in Merre and Romsdal County and in Norway
calculated as 3-year centered moving averages.
periods are shown in Table 1. When tested for linear
trend, the increase in crude annual incidence rates
from 1967 to 1991 reached statistical significance
(Chi square for linear trend=5.68; p=0.02). The
incidence rates both for men (p
=
0.03) and women
(p
=
0.04)
showed a statistically significant rise in
linear trend, but the difference between the sexes
did not reach statistical significance (p=0.51).
Temporal
trends
The crude annual rates of incidence, disability pen-
sion incidence, and mortality in the period from
1968 to 1990 expressed as 3-year centered moving
averages are presented in Fig. 4. The highest con-
tinuous increase in incidence was observed from
1974 (2.9/ 100,00O/yr) to 1977 (6.7/ 100,00O/yr). A
similar, but lower increase in disability pension
incidence was observed from 1987 (3.6/ 100,00O/yr)
to 1989 (5.7/ 100,00O/yr). The most prominent
increase in annual death rates was seen from 1984
(1.4/ 100,00O/yr) to 1988 (3.7/ 100,00O/yr).
Discussion
This study shows a significant rise in the average
annual age-adjusted incidence
of
MS patients receiv-
ing disability pension while resident in More and
Romsdal County (Table
(rates/100,000/yr) in
1968-90
1).
After the full legal
implementation
df
the disability pensions in Norway
in 1967, the incidence of disability pensions in the
years immediately following might reflect an accu-
mulation of cases from the previous years. Thus,
the observed disability pension incidence in the
years 1967-71 might be artificially high indicating
that the true incidence increase might be even larger
than the one reported here.
The difference in the development of MS-specific
disability pension prevalence rates in the county
and in the nation is notable (Fig.
2).
Although the
national trend also increased, the rise in More and
Romsdal County was higher, especially from 1983
and onwards. The disparity between the trends in
the disability pension rates due
to
all causes (Fig.
1
)
and the MS-specific disability pension rates (Fig. 2)
when comparing the county to the nation is remark-
able. The findings represent a contrast to previous
studies by Westlund
(1,2)
and Kurtzke (18). Based
on data from the same source (The National
Insurance Administration), these studies show a
long-term stability in the geographical pattern in
Norway. The age and sex-adjusted ratio Observed/
Expected in disability pension due to MS in More
and Romsdal compared to the national ratio was
0.97
on
1
January 1966 and 0.90 on
1
January 1978
according to Westlund
(
1,2).
31
1
Midgard
et
al.
8
7-7
Ti,,
,
-7-r-----
0
-
-j
~
_~--~
, ,
, ,
,
1968 1971 1974 1977 1980 1983 1986 1989
Calender
year
Fig.
4.
Crude annual rates
of
incidence, disability pension incidence, and mortality in More and Romsdal, Norway
1968-90
calculated as 3-year centered moving averages (rates/
lOO,OOO/
yr).
Several factors influence the disability pension
rates such as the general legislation regulating the
pensions, the national economy, the unemployment
rates, and the attitudes towards receiving a pension
among professionals and the public. The overall
increasing trend in total disability pension preval-
ence rates offers only a partial explanation to the
increase in MS-specific disability pension prevalence
in the county. The diagnosis of MS may be more
accurate in the disability cases than on the death
certificates. The increased MS-specific disability
pension prevalence rate reflects partly the raised
incidence, partly a longer survival, and consequently
an accumulation of cases from year to year (19).
We report a statistically significant increase in
average annual age-adjusted mortality rates due to
MS in Marre and Romsdal County from 1966 to
199
1
(Table
1
).
The increasing trend in the mortality
rate in Marre and Romsdal County contrasts the
more stable trend in the country which is also
previously reported in Norway and other western
industrialized nations
(8,
10). The mortality rate in
MS in another county in Western Norway (3) also
shows stability over time. The statistically significant
increase in the mortality rate in Marre and Romsdal
County underscore
MS
as
a
fatal disease with a
considerable
loss
of lifetime for patients, families,
and society. No major changes in the mortality
coding systems that might affect the mortality rates
have taken place.
According to Westlund
&
Kurtzke
(1,2,18)
the
average annual mortality rate from MS in Merre
and Romsdal County was 1.16 per 100,000
(37
MS-deaths) in the period from 1951 to 1965, while
the national rate in the same period was 1.85 per
100,000
(997 MS-deaths) per yr. Thus, the mortality
rate in the county during this period was below the
national rate. When the distribution of MS in Marre
and Romsdal County in 1951-65 is expressed as an
approximate percentage of the national rate, the
cumulative death rate of
MS
as underlying or
contributing cause of death is 63% (18).
In Norway both primary and secondary causes
of death are assigned to each death certificate. The
official mortality statistics therefore constitute valu-
able information, but when dealing with time trends
of
MS
caution must be taken in the interpretation
of data. Only 101
(77.7%)
of the 130 deceased
patients in the clinically assessed material were
coded with MS in the Registry
of
Death Causes,
Statistics Norway, thus the MS frequency at death
was underrated in the official appraisal. The magni-
tude of the underestimation is in accordance with
previously published reports (3,20,21). 103
MS
deaths were recorded in the Registry of Death
Causes, Statistics Norway. Two were wrongly
31
2
MS
disability
and
mortality
diagnosed, and ten deceased
MS
patients randomly
scattered throughout the period 1969-87 were not
ascertained in the epidemiological study. The ten
cases revealed by the Registry of Death Causes not
previously recognized in repeated assessments of
the incidence rate indicate one of the pitfalls and
the difficulties met when evaluating incidence
figures. The significance
of
the thoroughly scrutiny
of all possible sources of information is obvious.
In this study we report a statistically significant
increase in average annual age-adjusted incidence
of
MS
in More and Romsdal County from 1966 to
1991. In addition to an update of the
MS
incidence
in the county from 1984 to 1991, we have reorgan-
ized and recalculated the incidence figures from
previous studies
(12,5)
showing a highly significant
change in
MS
frequency compared to the figures of
Swank
(22).
Better case ascertainment may be of
importance and partly explain the secular changes
in incidence rates (23).
We have studied the trends in the disease fre-
quency of multiple sclerosis using three different,
but not entirely independent sources. The increase
in the average annual age-adjusted incidence rates
is seen from 1975-77. Approximately six years later,
in
1981-83, a rise in the average annual age-adjusted
disability pension incidence rate in Mrare and
Romsdal County appeared. In the triennial
1987-89-about twelve years after the observed
increase in the average annual age-adjusted incid-
ence, an increase in the average annual age-adjusted
mortality can be observed (Table 1).
When the three different annual rates were calcu-
lated as 3-year centered moving averages the largest
differences in the disease frequency were observed
in 1974-77 (incidence), 1987-90 (disability), and
1984-88 (mortality) (Fig. 4). Thus, there was a
thirteen years interval from the observed peak in
incidence rate to the corresponding peak in disabil-
ity pension incidence rate, and an eleven years
interval to the observed peak in the mortality rate.
Acknowledging the limitations of the prevalence
rates in the disability pensions by the end of each
year used as a frequency measure, the significant
differences in Mrare and Romsdal County compared
to Norway indicate that the observed increase in
the frequency of
MS
represent more than a better
case ascertainment. The observed alterations in the
previously reported stable rates of disability pension
and mortality might indicate a change in environ-
mental factors that determine the frequency of
multiple sclerosis as population genetics shape the
distribution of disease more slowly.
We consider that this increase in disability pen-
sion incidence and mortality confirms the previously
reported increase in prevalence and incidence. Thus,
three different sources
of
information corroborate
corresponding trends indicating that case ascertain-
ment only partially can explain the reported increase
in this area
of
Norway.
Acknowledgements
We gratefully acknowledge valuable suggestions and comments
from Klaus Lauer. The study was supported by grants from the
Legacy of Fritz and Ingrid Nilsen, the Odd Fellow Order, the
Norwegian MS Society, and the Norwegian Council for Science
and the Humanities.
References
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17
WESTLUND K. Distribution and mortality time trend of
multiple sclerosis and some other diseases in Norway. Acta
Neurol Scand 1970: 46: 455-83.
WESTLUND
K.
Recent Statistical Data on Multiple Sclerosis
and some other diseases in Norway. Nordic Council Arct
Med Res Rep 1982: 32: 19-29.
LARSEN JP, KVALE G,
AARLI
JA. Multiple sclerosis and
mortality statistics. Acta Neurol Scand 1985: 71: 237-41.
LARSEN JP,
AARLI
JA, NYLAND H, RIISE
T.
Western Norway,
a high-risk area for multiple sclerosis:
A
prevalence/ incid-
ence study in the county
of
Hordaland. Neurology 1984:
34: 1202-7.
MIDGARD R, RIISE
T,
NYLAND H. Epidemiologic trends in
multiple sclerosis in Mare and Romsdal, Norway: A
prevalence/ incidence study in a stable population.
Neurology 1991: 41: 887-92.
GR0NNING M, RIISE T, KV~LE
G,
NYLAND
H,
LARSEN JP,
AARLI
JA. Incidence
of
Multiple Sclerosis in Hordaland,
Western Norway: A Fluctuating Pattern.
Neuroepidemiology 1991: 10: 53-61.
GOLDBERG
ID, KURLAND
LT.
Mortality in 33 countries
from diseases of the nervous system. Wld Neurol 1962:
MASSEY EW, SCHOENBERG BS. International patterns of
mortality from multiple sclerosis. Neuroepidemiology 1982:
KURTZKE JF, Lux JR. WE.
In
defense of death data: An
example of multiple sclerosis. Neurology 1985:
35:
1787-90.
LAI SM, ZHANG
ZX,
ALTER
M, SOBEL E. World-Wide
Trends in Multiple Sclerosis Mortality. Neuroepidemiology
WILLIAMS ES,
JONES
DR, MCKERAN RO. Mortality rates
from multiple sclerosis: geo-graphical and temporal vari-
ations revisited. J Neurol Neurosurg Psychiatry 1991:
54:
-THUS
J. Multiple Sclerosis in Msre and Romsdal
County, Norway. Acta Neurol Scand 1966: 42 (Suppl
MCALPINE
D.
The benign
form
of multiple sclerosis. A
study based on 241 cases seen within three years of onset
and followed up until the tenth year of more
of
disease.
Brain 1961: 84: 186-203.
ALTMAN
DG.
Practical Statistics for Medical Research.
London-Glasgow-New
York-Tokyo-Melbourne-Madras:
Chapman
&
Hall, 1992.
DIXON WJ,
BROWN
MB, ENGELMAN L, HILL MA, JENNRICH
RI. BMDP Statistical Software. Berkeley: University of
California Press, 1990.
DEAN AG, DEAN JA, BURTON AH, DICKER RC. EpiInfo,
Version
5:
a word processing, database, and statistics
program for epidemiology on microcomputers. Stone
Mountain, Georgia.
:
USD, Incorporated, 1990.
WATERHOUSE
J,
MUIR
C, CORREA
P, POWELL
J.
Cancer
313
3:
444-65.
1: 189-96.
1989:
8:
56-67.
104-9.
19): 12-18.
Midgard
et
al.
incidence in five continents. Volume 3. Lyon, International
Agency for Research in Cancer.: IARC Scientific
Publications, No
15,
1976.
18.
KURTZKE JF. Further features of the fennoscandian focus
of
multiple sclerosis. Acta Neurol Scand 1974:
50:
478-502.
19. RUSE
T,
GR0NNING M,
AARLI
JA, NYLAND
H,
LARSEN
JP,
EDLAND A. Prognostic factors for life expectancy in multiple
sclerosis analysed by Cox-models.
J
Clin Epidemiol 1988:
41: 1031-36.
20. HAMMOND SR,
ENGLISH
DR,
DEWYTT
C et al. The contribu-
tion of mortality statistics to the study of multiple sclerosis
in Australia. J Neurol Neurosurg Psychiatry 1989: 52:
1-7.
21. MALMGREN RM,
VALDIVIEZO
NL, VISSCHW BR et al.
Underlying cause of death
as
recorded for multiple sclerosis
patients: associated factors.
J
Chron Dis 1983: 36: 699-705.
22. SWANK RL, LERSTAD
0,
STR0M A,
BACKER
J.
Multiple
Sclerosis in Rural Norway. N Engl J Med 1952: 246: 721-8.
23.
WYNN
DR,
RODRIGUEZ
M,
O'FALLON
WM,
KURLAND
LT.
A reappraisal of the epidemiology of multiple sclerosis in
Olmsted County, Minnesota. Neurology 1990: 40: 780-6.
314