Risk of MS is not associated with exposure to crude oil, but increases with low level of education

http://msj.sagepub.com/Multiple Sclerosis Journal

http://msj.sagepub.com/content/17/7/780The online version of this article can be found at:

 DOI: 10.1177/1352458510397686

2011 17: 780 originally published online 22 February 2011Mult SclerJørgen Mørland, Wenche Telstad, Per Tore Førland and Kjell-Morten Myhr

Trond Riise, Jorunn Kirkeleit, Jan Harald Aarseth, Elisabeth Farbu, Rune Midgard, Åse Mygland, Randi Eikeland, ToreRisk of MS is not associated with exposure to crude oil, but increases with low level of education

Published by:

http://www.sagepublications.com

On behalf of: 

European Committee for Treatment and Research in Multiple Sclerosis

Americas Committee for Treatment and Research in Multiple Sclerosis

Pan-Asian Committee for Treatment and Research in Multiple Sclerosis

Latin American Committee on Treatment and Research of Multiple Sclerosis

can be found at:Multiple Sclerosis JournalAdditional services and information for    

  http://msj.sagepub.com/cgi/alertsEmail Alerts:

http://msj.sagepub.com/subscriptionsSubscriptions:  

http://www.sagepub.com/journalsReprints.navReprints:  

http://www.sagepub.com/journalsPermissions.navPermissions:  

at Universitetsbiblioteket i Bergen on August 1, 2014msj.sagepub.comDownloaded from at Universitetsbiblioteket i Bergen on August 1, 2014msj.sagepub.comDownloaded from

What is This? 

- Feb 22, 2011 OnlineFirst Version of Record 

- Jun 20, 2011Version of Record >>

at Universitetsbiblioteket i Bergen on August 1, 2014msj.sagepub.comDownloaded from at Universitetsbiblioteket i Bergen on August 1, 2014msj.sagepub.comDownloaded from

Research Paper

Risk of MS is not associated withexposure to crude oil, but increaseswith low level of education

Trond Riise1,2, Jorunn Kirkeleit1, Jan Harald Aarseth2,Elisabeth Farbu3, Rune Midgard4, Ase Mygland5,6,Randi Eikeland5, Tore Jørgen Mørland7, Wenche Telstad8,Per Tore Førland3 and Kjell-Morten Myhr2,6

Abstract

Background: Offshore workers in the Norwegian upstream petroleum industry are exposed to a number of chemicals

such as organic solvents, mineral oils and other hydrocarbons, possibly contributing to an increased risk of multiple

sclerosis (MS).

Objective: To estimate the risk of MS in this population compared with the general working population in Norway,

adjusting for education.

Methods: Using the Norwegian Registry of Employers and Employees we included all 27,900 offshore workers regis-

tered from 1981 to 2003 and 366,805 referents from the general working population matched by gender, age and

community of residence. The cohort was linked to the Norwegian MS Registry and the Norwegian Education Registry.

Results: There was no increased risk of MS among the offshore workers. We found a marked and linear inverse

relationship between level of education and the risk of MS in the total study population, with a rate ratio of 0.48

(95% CI, 0.53 to 0.88) for workers with a graduate degree compared to workers with elementary school only.

Conclusions: These findings do not support a major aetiological role of petroleum-based products, but rather point to

smoking and other lifestyle factors related to the level of education as being important for the risk of MS.

Keywords

aetiology, chemical factors, cohort, education, multiple sclerosis, risk factors

Date received: 13th September 2010; revised: 16th November 2010; 8th December 2010; accepted: 9th December 2010

Introduction

Environmental factors are assumed to play an impor-tant role in the development of MS.1 In addition to anexpected role of an infectious agent,2 smoking3 andvitamin D,4 occupational factors such as exposure toorganic solvents,5 mineral oils,6 heavy metals7 andambient air pollution8 have been suggested. Offshoreworkers on the Norwegian shelf are exposed to anumber of chemicals including crude oil, organicsolvents and products based on mineral oils.9

Recently the media, the authorities, and the petroleumindustry in Norway have been concerned aboutoffshore workers who have been exposed to hydrocar-bons and later diagnosed with MS.

1Department of Public Health and Primary Health Care, University of

Bergen, Norway.2The Norwegian Multiple Sclerosis Competence Centre, Haukeland

University Hospital, Norway.3Department of Neurology, Stavanger University Hospital, Norway.4Department of Neurology, Molde Hospital, Molde, Norway.5Department of Neurology, Sørlandet Hospital, Norway.6Department of Clinical Medicine, University of Bergen, Norway.7Department of Neurology, Telemark Hospital, Norway.8Department of Neurology, Central Hospital of Sogn and Fjordane,

Norway.

Corresponding author:

Professor Trond Riise, Department of Public Health and Primary Health

Care, University of Bergen, Kalfarveien 31, N-5018 Bergen, Norway

Email: [email protected]

Multiple Sclerosis Journal

17(7) 780–787

! The Author(s) 2011

Reprints and permissions:

sagepub.co.uk/journalsPermissions.nav

DOI: 10.1177/1352458510397686

msj.sagepub.com

at Universitetsbiblioteket i Bergen on August 1, 2014msj.sagepub.comDownloaded from

In a historical cohort study of petroleum workers,we previously found that petroleum workers employedon Norway’s continental shelf from 1981 to 2003 had ahigher risk of developing acute myelogenous leukae-mia, multiple myeloma and oesophageal adenocarci-noma than the general working population.10,11 Theincreased risk was found in the job category assumedto have the most extensive contact with crude oil and itsderivatives. We wanted to investigate whether the riskof MS was increased among workers exposed to thesuspected chemical risk factors compared to that ofthe general working population. In the study we alsoadjusted for educational level, a factor previouslyshown to be associated with the risk of MS.

Methods

Study population and study design

The cohort was established using information from theNorwegian Registry of Employers and Employees,containing employments from 1981 up until today. AllNorwegian employers are required to register theiremployees with a personal identification number, indus-trial classification code (International StandardIndustrial Classification [ISIC] or Classification ofEconomic Activities in the European Union [NACE]),county of work and the first and last date of all theirwork engagements. The Registry requires that allengagements with a minimum of 4 h of work per weekand duration of at least 6 weeks must be registered. Theestablishment of the cohort has been reportedpreviously.10,11 In short, we included all individualworkers registered with any of the offshore-relatedindustrial classification codes or having Norway’scontinental shelf (North Sea) as their work location.

The category ‘upstream operator offshore’ containssolely workers registered with the NACE and ISICcode ‘extraction of crude oil and natural gas’, andincludes job titles such as process technicians,laboratory engineers, control operators, and other jobgroups involved in the production process, such asmechanics, electricians, turbine operators, hydraulicstechnicians and other support personnel. The category‘drilling and well maintenance offshore’ includes theISIC code 50230 (oil drilling) and NACE code 11200(service activities incidental to oil and gas extractionexcluding surveying). The latter comprises activitiessuch as drilling of wells and installation, disassembling,and maintenance of drilling towers at site on contract.The category ‘catering offshore’ includes the job groupscatering crew and housekeeping personnel. The cate-gory ‘others offshore’ includes miscellaneous industrialcodes and comprises activities contracted out to oil field

service companies, such as construction and mainte-nance personnel, and logistics and technical consul-tancy activities. Since there are no onshore oil fieldsin Norway, ‘petroleum workers onshore’ containsmainly workers involved in administering, planning,and coordinating the activities offshore.

Up to six referents per petroleum worker wererandomly drawn from the general working population,using the same Norwegian Registry of Employers andEmployees and the same year of first engagement asthat of the corresponding petroleum worker. StatisticsNorway did this by randomly selecting individualsavailable in the Registry having the same gender, ageand community of residence as the petroleum worker inquestion. The crude historical cohort included 71,018workers from the petroleum industry (‘at risk’) and424,584 referents. We excluded subjects from thecohort if they had had an onset of MS before theyear they were first registered in the cohort. We alsoexcluded referents if they had an earlier engagement inthe petroleum industry before they were drawn asreferents, even if they were not considered to havebeen exposed during the engagement. We allowedsubjects to serve as referents for more than one ‘subjectat risk’. The final cohort included 365,805 referents,42,657 onshore petroleum workers and 27,900 offshoreworkers, in the four offshore job categories, comprisinga total of 6,836,188 person-years of follow-up (Table 1).

We created an ‘area of residence variable’ consistingof the 19 counties in Norway. Since the number ofworkers was small in some of these counties, we col-lapsed all counties except Oslo and Telemark in EasternNorway into one category (n¼ 49,066), the counties inNorthern Norway including Trøndelag into onecategory (n¼ 30,487) and kept the rest of the countiesas separate categories; Oslo (n¼ 22,766), Telemark(n¼ 20,759), Aust-Agder (n¼ 11,198), Vest-Agder(n¼ 17,851), Rogaland (n¼ 172,057), Hordaland(n¼ 79,983), Sogn og Fjordane (n¼ 8673) and Møreog Romsdal (n¼ 15,506).

Statistics Norway linked the total cohort tothe nationwide Norwegian MS Registry inSeptember 2008, including all cases of MS reportedup to 31 December 2007 with information on date ofonset and initial clinical course of the disease. This reg-istry was established in 2001, but includes cases withonset from the early 1980s. Nevertheless, since theNorwegian MS Registry still does not include allcases in Norway, a diagnosis of MS among these work-ers was also ascertained by linkage to discharge filesfrom neurological departments in the counties wherethe majority of the workers resided, resulting in atotal of 648 patients with an onset of disease after thestart of their working engagement. Statistics Norwayalso linked the cohort to the Norwegian Cause of

Riise et al. 781

at Universitetsbiblioteket i Bergen on August 1, 2014msj.sagepub.comDownloaded from

Death Registry, the Cancer Registry of Norway andthe Norwegian Education Registry, including the vari-able ‘highest completed education’, ranging from 1(no education or pre-school education) to 8 (PhDdegree).

Statistical analysis

We estimated rate ratios (RR) (hazard ratios) compar-ing the various job categories with the general workingpopulation using the Cox proportional hazard regres-sion model, and censored subjects at the end of follow-up or at the date of death, whichever occurred first. Theproportional hazards assumption was checked by com-paring the estimated ln(–ln) survival curves for the dif-ferent groups. No marked deviation from theproportional hazards assumption was found.

Since we used the total reference population foranalysing risk in each specific job category, there wasno longer an identical distribution of gender, age, areaof residence and year of first registered engagementbetween the reference group and the job categories.We therefore performed multivariate analysis includingthese independent covariates, in addition to educationallevel on an aggregated three-point scale. Age wasdefined at the time of entering into the cohort (timeof first registered engagement) and used as a categoricalvariable in the model in 5-year intervals. We also ran amodel including education on a five-point scale. Sincethe mean age at entering the cohort was 33.8 years andthe mean age of onset of MS in Norway is estimated tobe around 32 years,12 we also repeated the analysesstratifying by age at entering the cohort: above and

below 30 years of age. We performed all analysesusing SPSS 15.0 (SPSS Inc., Chicago, IL, USA).

Ethical considerations

We conducted the study with the approval of theRegional Committee for Medical Research Ethics ofWestern Norway, the Norwegian Data Inspectorateand the Directorate for Health.

Results

There was no increased risk of MS among the offshoreworkers compared to the general working population(Table 2). The workers in the two job categoriesassumed to have the most extensive exposure to hydro-carbons (production and drilling) had a non-significantdecreased risk (RR¼ 0.52, 95% CI 0.23–1.16 andRR¼ 0.90, 95% CI 0.45–1.81). We repeated the analy-sis stratifying by age at inclusion into the cohort (aboveand below 30 years). The results were similar for thetwo sub-populations (data not shown). We alsorepeated the analysis including exposure (workoffshore) only up to 3 years prior to onset of MS.Again, only minor changes in the estimated risks werefound (data not shown).

We found a marked and linear inverse relationshipbetween the level of education and the risk of MS (Table2). A repeated analysis including level of education on afive-point scale showed a rate ratio of 0.43 (95% CI0.27–0.66) for workers with a graduate degree comparedto workers with elementary school only (Figure 1).

Table 1. Characteristics of the study population examined to determine the risk of developing multiple sclerosis among workers

employed in Norwegian upstream offshore petroleum industry

Variable Referents

Onshore

petroleum

workers

Offshore

workers (all)

Offshore workers

Upstream

operators

Drilling and well

maintenance

Catering

personnel

Others

offshore

Number of subjects 365,805 42,657 27,900 6,733 7,038 2,414 11,715

Male (%) 282,773 (77.3) 30,595 (71.7) 24,750 (88.7) 5,852 (86.9) 6,713 (95.4) 1,229 (50.9) 10,956 (93.5)

Female (%) 83,032 (22.7) 12,061 (28.3) 3,150 (11.3) 881 (13.1) 325 (4.6) 1,185 (49.1) 759 (6.5)

Age, mean (SD)

At inclusion in the cohort 33.8 (9.5) 33.9 (9.7) 33.6 33.8 (8.1) 32.6 (8.6) 32.6 (10.0) 34.3 (9.6)

At the end of follow-up 49.5 (11.6) 49.3 (11.6) 49.7 52.6 (11.3) 46.4 (11.2) 50.8 (11.5) 49.7 (10.7)

Education level

Tertiary education (%) 28.4 46.8 15.3 18.4 13.9 8.7 15.7

Intermediate education (%) 59.0 41.7 74.3 72.5 74.4 73.6 70.8

Compulsory education (%) 11.4 4.3 9.7 7.2 9.6 16.4 9.8

Average follow-up (years) 15.7 15.4 16.1 18.8 13.8 18.2 15.4

Person-years of follow-up 5,730,257 657,432 448,499 126,600 97,192 44,010 180,697

782 Multiple Sclerosis Journal 17(7)

at Universitetsbiblioteket i Bergen on August 1, 2014msj.sagepub.comDownloaded from

To investigate whether the lower level of educationamong those who developed MS could have beencaused by MS patients being less likely to go back toschool after onset of the disease, we repeated the anal-ysis including only those who had finished their educa-tion at least 3 years before entering into the study(n¼ 301,130; 70.2%). This gave a slightly stronger esti-mate, with RR¼ 0.37 (95% CI 0.21–0.67) for workerswith a graduate degree compared to workers with ele-mentary school only.

To enable a better interpretation of the associationbetween education and the risk of MS we analysed theassociation between education and three other diseaseswith known lifestyle risk factors, using data on cancerand mortality previously published.10,11 We found astrong inverse association between the level of educa-tion and the risk of bronchus and lung cancer, andwith the overall mortality due to vascular diseases(Figure 2). There was further a marked reduced riskof colorectal cancer for those with the highest level ofeducation.

Discussion

Offshore workers did not have any excess risk of MScompared to the general working population.Furthermore, we found that the risk of MS decreasedin the total study population with an increasing level ofeducation.

Occupational exposure

A number of previous studies have reported an associ-ation between various chemical factors and anincreased risk of MS.13 One case-control study of17 workers exhibiting neurological symptoms resem-bling MS and 58 controls at an automotive carburetorplant found an association with exposure to hydraulicand machining fluids.6 Several studies have reported alink between MS and exposure to organic solvents.An early review of predominantly case-control studiesfound a pooled risk estimate of 1.7.14 Further, occupa-tional exposure to mineral oil has been reported to be

Table 2. Rate ratios for multiple sclerosis with 95% confidence intervals (CI) for workers in the Norwegian upstream petroleum

industry versus referents from the general working population, adjusted for gender, age at inclusion into the cohort, year of first

exposure, area of residence and education

Total number* Number of cases Rate ratio§ 95% CI p-value

Occupation 0.77

Referents 361,599 554 1.00 –

Onshore workers 39,583 58 0.99 0.75–1.30

Upstream operators 6,603 6 0.52 0.23–1.16

Drilling and well maintenance 6,886 8 0.90 0.45–1.81

Catering personnel 2,381 5 0.87 0.36–2.10

Others offshore 11,772 17 1.03 0.64–1.68

Sex <0.001

Men 331,250 412 1.00 –

Women 97,096 236 1.94 1.64–2.30

Age at inclusion (years) <0.001

15–20 19,646 35 1.00 –

21–25 66,608 117 0.94 0.65–1.38

26–30 99,203 181 1.18 0.82–1.70

31–35 82,630 149 1.21 0.83–1.76

36–40 63,246 90 0.93 0.63–1.39

41–45 43,006 46 0.70 0.45–1.10

46þ 54,007 30 0.39 0.24–0.64

Education 0.009

Compulsory education 46,220 80 1.00 –

Intermediate education 253,901 401 0.88 0.69–1.13

Tertiary education 128,225 167 0.72 0.55–0.94

*Information on education was missing for 8004 individuals. These cases were excluded from the regression analysis.

§Cox proportional hazards model.

Riise et al. 783

at Universitetsbiblioteket i Bergen on August 1, 2014msj.sagepub.comDownloaded from

Graduate levelUndergraduate level

Final year upper

seconday

Basic upper secondary

Compulsory education

1.2

1.0

0.8

0.6

0.4

n=46220 n=104424 n=149477 n=91482 n=36743

Figure 1. Rate ratio, with 95% confidence intervals, between the level of education and the risk of multiple sclerosis in a cohort

study of 428,346 Norwegian workers using the lowest level of education as reference category. Cox proportional hazard regression

model adjusted for gender, age at inclusion into the cohort, year of first exposure, area of residence and education.

Graduate levelUndergraduate level

Final year upper

secondary

Basic upper secondary

Compulsoryeducation

1.2

1.0

0.8

0.6

0.4

0.2

Colorectal cancer

Cancer of bronchus/lung

Mortality from vascular diseases

Figure 2. Rate ratio, with 95% confidence intervals, between the level of education and the risk of colorectal cancer, bronchus/lung

cancer and mortality due to vascular diseases in a cohort study of 428,346 Norwegian workers using the lowest level of education as

reference category. Cox proportional hazard regression model adjusted for gender, age at inclusion into the cohort, year of first

exposure, area of residence and education.

784 Multiple Sclerosis Journal 17(7)

at Universitetsbiblioteket i Bergen on August 1, 2014msj.sagepub.comDownloaded from

directly linked to rheumatoid arthritis, another inflam-matory, autoimmune disease.15

We found no increased risk of MS among theoffshore workers on the Norwegian shelf. A retrospec-tive exposure assessment of carcinogens performed inNorway’s petroleum industry offshore concluded thatoffshore workers are potentially exposed to all of thechemical agents that have been suggested as risk factorsof MS, either through their contact with the petroleumstream (hydrocarbons and heavy metals such asmercury) or agents used in the production process(organic solvents, synthetic and mineral oil-basedfluids and lead).9 However, the same study concludedthat information about exposure levels is scarce.

We found a non-significant lower risk of MS for theupstream operators, the group of workers assumed tobe most exposed to petroleum-derived hydrocarbons.These workers are exposed to the different phases ofthe petroleum stream, a large number of syntheticand mineral oil-based fluids and organic solventsused as cleaning and metal-degreasing agents.9 Wedid not have individual exposure data, but these work-ers have in this same cohort shown an increased risk ofhaematological malignancies and oesophagealcancer.10,11 The increased risks of leukaemia andmultiple myeloma in particular were interpreted to beassociated with exposure to the aromatic hydrocarbonbenzene that is a natural component of crude oil andnatural gas.9,16

We lack individual data on possible confoundingfactors such as smoking. Nevertheless, in a previousstudy of this population we found that the incidencerates of smoking-related cancers did not differ mark-edly between the petroleum workers and the refer-ents.10,11 This indicates that smoking has not had amajor influence on the differences in rates of MSbetween the worker categories. The mean age at inclu-sion into our prospective study population was 33.8years. Since the mean age of onset of MS in general isaround 30 years, we may have missed an effect onlyassociated with an early onset. However, the analysisstratified by age above and below 30 years of age didnot suggest such an effect.

There are several other studies that do not find anassociation between MS and occupational exposure topetroleum-derived hydrocarbons. A mortality study ofworkers in the US petroleum industry showed lowerrates of MS for men17 and women.18 A study of USveterans of the Persian Gulf War potentially exposed tosmoke from oil well fires found a non-significantincreased risk of brain cancer, but not of MS.19

Further, the prevalence of MS was not increased in aresidential area bordering on an oil refinery withdocumented spills and leaks of crude oil and otherchemical products.20

The two studies reporting lower rates of MS amongthe workers in the petroleum industry also reported alower overall mortality,17,18 a finding also reported inother studies of petroleum workers.21,22 The results inthese studies may reflect a ‘healthy worker effect’ sincethe mortality rate was calculated by comparison withthe general population. We aimed to reduce this effectin our study by using a prospective design and selectingour referents from the general working populationwithin the same cohort.

If individuals with preclinical symptoms of MS areless likely to choose a physically demanding joboffshore, this might result in an underestimation ofthe rate of MS among the offshore workers in thepresent study. No change in the rates was seen whenanalysing the observation period until 3 years priorto the onset of MS only, and this argues against suchan effect.

Education

A striking finding in this study was the marked inverserelationship between the level of education and the riskof MS. The sub-analysis of individuals who hadfinished their education at least 3 years before enteringthe study showed that this finding does not reflect alower probability of MS patients going back to schoolafter onset of the disease.

Education is generally associated with markeddifferences in lifestyle factors.23 Smoking is a well-established risk factor for MS with risk estimatesreported between 1.5 and 2.0.3,24 Annual surveys ofsmoking habits in Norway show a decline from 42%daily smokers in 1973 to 21% in 2009.25 The level ofeducation correlated highly with smoking habits in2009, with only 7% of those with a university andcollege education smoking daily, compared to 26% ofthose with only an elementary school education.Furthermore, in our own study population we foundthat the level of education was strongly and inverselyrelated to the risk of bronchus and lung cancer. Thus,smoking may explain much of the association betweenMS and education.

Further, a diet low in vitamin D4 and a high bodymass index,26 suggested as possible risk factors in MS,may also be related to education. The risk of vasculardiseases was markedly associated with education in ourcohort. This may reflect an association between educa-tion and known risk factors related to diet and obesity,although smoking is also an important risk factor forthese diseases. A recent review of colorectal cancerpoints to a low vitamin D status as an important riskfactor.27 We found a markedly reduced risk of thiscancer type among individuals with the highest levelof education, who also showed the lowest risk of MS.

Riise et al. 785

at Universitetsbiblioteket i Bergen on August 1, 2014msj.sagepub.comDownloaded from

Except for a case-control study from Canada show-ing a similar inverse association between the level ofeducation and the risk of MS,28 and a Swedish studywhere no association was found,29 the findings inthis study are in contrast to some early studies.An increased risk of MS among individuals with ahigher level of education and a higher social class wasnoted in a study of 271 cases in 1960,30 and laterreported in a smaller hospital-based series of 103patients.31 Such a positive correlation was also foundin two prevalence studies of 2307 cases from 1980 inAustralia32 and 5305 US veterans from the SecondWorld War and the Korean War with prevalencedates in 1956 and 1969.33 These findings have beeninterpreted as compatible with the so-called hygienehypothesis or ‘polio hypothesis’ where assuminglypoorer standards of hygiene among those of lowersocio-economic status could result in relative immunityto a virus infection that might trigger disease inlater life.

The authors of these studies pointed to methodolog-ical issues such as possible differences in access tomedical services between social groups and an effectof differential mortality due to the use of prevalencedata. Further, it is quite probable that the associationbetween smoking and other lifestyle factors and thelevel of education assumed to explain the findings ofthe current study could have been less marked, or eventhe other way around, in the years of putative riskfor the studies mentioned above, covering mainly the40 s, 50 s and 60 s. None of the studies investigating theassociation between level of education and the risk ofMS have made any attempt to adjust for smoking orother lifestyle factors.

In conclusion, the findings of the present study donot support a major aetiological role of crude oil,organic solvents or other petroleum-based products,but rather point to smoking and other lifestyle factorsrelated to the level of education as important for therisk of MS.

Acknowledgement

We acknowledge Katrine Hveem of Statistics Norway forlinking the cohort to the Norwegian MS Registry.

Funding

This work was supported by the Research Council of Norway

under the PETROMAKS programme, the Department ofHealth (UniHealth) of UniResearch, and the NorwegianEXTRAfoundation for Health and Rehabilitation throughEXTRA funds and the Norwegian Cancer Society.

Conflict of interest statement

None declared.

References

1. Ascherio A and Munger KL. Environmental risk factors

for multiple sclerosis. Part II: Noninfectious factors.Ann Neurol 2007; 61: 504–513.

2. Ascherio A and Munger KL. Environmental risk factorsfor multiple sclerosis. Part I: The role of infection. Ann

Neurol 2007; 61: 288–299.3. Hernan MA, Olek MJ and Ascherio A. Cigarette smok-

ing and incidence of multiple sclerosis. Am J Epidemiol

2001; 154: 69–74.4. Ascherio A, Munger KL and Simon KC. Vitamin D and

multiple sclerosis. Lancet Neurol 2010; 9: 599–612.

5. Amaducci L, Arfaioli C, Inzitari D and Marchi M.Multiple sclerosis among shoe and leather workers: anepidemiological survey in Florence. Acta Neurol Scand1982; 65: 94–103.

6. Krebs JM, Park RM and Boal WL. A neurological dis-ease cluster at a manufacturing plant. Arch Env Health1995; 50: 190–195.

7. Henry JP, Williamson DM, Schiffer R, Wagner L, Shire Jand Garabedian M. Investigation of a cluster of multiplesclerosis in two elementary school cohorts. J Environ

Health 2007; 69: 34–38.8. Gregory II AC, Shendell DG, Okosun IS and Gieseker

KE. Multiple sclerosis disease distribution and potential

impact of environmental air pollutants in Georgia. SciTotal Environ 2008; 396: 42–51.

9. Steinsvag K, Bratveit M and Moen BE. Exposure to car-cinogens for defined job categories in Norway’s offshore

petroleum industry, 1970 to 2005. Occup Environ Med2007; 64: 250–258.

10. Kirkeleit J, Riise T, Bratveit M and Moen BE. Increased

risk of acute myelogenous leukemia and multiple

myeloma in a historical cohort of upstream petro-

leum workers offshore. Cancer Causes Control 2008; 19:

13–23.

11. Kirkeleit J, Riise T, Bjørge T, Moen BE, Bratveit M and

Christiani DC. Increased risk of oesophageal adenocarci-noma among upstream petroleum workers. OccupEnviron Med 2010; 67: 335–340.

12. Grytten N, Glad SB, Aarseth JH, Nyland H, Midgard Rand Myhr KM. A 50-year follow-up of the incidence ofmultiple sclerosis in Hordaland County, Norway.Neurology 2006; 66: 182–186.

13. Marrie AM. Environmental risk factors in multiple scle-rosis aetiology. Lancet Neurol 2004; 3: 709–718.

14. Landtblom AM, Flodin U, Soderfeldt B, Wolfson C and

Axelson O. Organic solvents andmultiple sclerosis: a synthe-

sis of the current evidence. Epidemiology 1996; 7: 429–433.

15. Sverdrup B, Kallberg H, Bengtsson C, Lundberg I,

Padtukov L, Alfredsson L, et al. The Epidemiological

Investigation of Rheumatoid Arthritis study group.

Association between occupational exposure to mineral

oil and rheumatoid arthritis: results from the Swedish

EIRA case-control study. Arthritis Res Ther 2005; 7:

R1296–R1303.

16. Kirkeleit J, Riise T, Bratveit M and Moen BE. Exposure

to benzene on a crude oil production ship. Ann Occup

Hyg 2006; 50: 123–129.

786 Multiple Sclerosis Journal 17(7)

at Universitetsbiblioteket i Bergen on August 1, 2014msj.sagepub.comDownloaded from

17. Huebner WW, Wojcik NC, Jorgensen G, Marcella SPand Nicolich MJ. Mortality patterns and trends among127,266 U.S.-based men in a petroleum company: update

1979-2000. J Occup Environ Med 2009; 51: 1333–1348.18. Huebner WW, Wojcik NC, Jorgensen G, Marcella SP

and Nicolich MJ. Mortality patterns and trends among49,705 U.S.-based women in a petroleum company:

update 1979-2000. J Occup Environ Med 2010; 52:99–108.

19. Barth SK, Kang HK, Bullman TA and Wallin MT.

Neurological mortality among US Veterans of thePersian Gulf War: 13-year follow up. Am J Ind Med2009; 52: 663–670.

20. Neuberger JS, Lynch SG, Sutton ML, Hall SB, Feng Cand Schmidt WR. Prevalence of multiple sclerosis in aresidential area bordering an oil refinery. Neurology

2004; 63: 1796–1802.21. Wong O and Raabe GK. A critical review of cancer epi-

demiology in the petroleum industry, with a meta-analy-sis of a combined database of more than 350,000 workers.

Regul Toxicol Pharmacol 2000; 32: 78–98.22. Gun RT, Pratt N, Ryan P and Roder D. Update of mor-

tality and cancer incidence in the Australian petroleum

industry cohort. Occup Environ Med 2006; 63: 476–481.23. Jha P. Social inequalities in male mortality, and in male

mortality from smoking: indirect estimation from

national death rates in England and Wales, Poland, andNorth America. Lancet 2006; 368: 367–370.

24. Riise T, Nortvedt MW and Ascherio A. Smoking is a riskfactor for multiple sclerosis. Neurology 2003; 61:1122–1124.

25. Statistics Norway. http://www.ssb.no/english/subjects/03/01/royk_en/ (accessed 31 January 2011).

26. Munger KL, Chitnis T and Ascherio A. Body size andrisk of MS in two cohorts of US women. Neurology 2009;

73: 1543–1550.27. Giovannucci E. Epidemiology of vitamin D and colorec-

tal cancer: causal or causal link? J Steroid Biochem Mol

Biol 2010; 121: 349–354.28. Ghadirian P, Dadgostar B, Azani R and Maisonneuve P.

A case-control study of the association between sociode-

mographic, lifestyle and medical history factors and mul-tiple sclerosis. Can J Public Health 2001; 92: 281–285.

29. Montgomery SM, Lambe M, Olsson T and Ekbom A.

Parental age, family size, and risk of multiple sclerosis.Epidemiology 2004; 15: 717–723.

30. Miller H, Ridley A and Shapira K. Multiple sclerosis.A note on social incidence. BMJ 1960; 2: 343–345.

31. Russell WR. Multiple sclerosis: occupational and socialgroup at onset. Lancet 1971; 2: 832–834.

32. Hammond SR, McLeod JG, Macaskill AC and English

DR. Multiple sclerosis in Australia: socio-economic fac-tors. J Neurol Neurosurg Psychiatry 1996; 61: 311–313.

33. Kurtzke JF. Epidemiology of multiple sclerosis in US

veterans: VII. Risk factors for MS. Neurology 1997; 48:204–213.

Riise et al. 787