A case–control study of risk factors for multiple sclerosis in Iran

Introduction

Multiple sclerosis (MS) is a chronic inflammatory disease of the central nervous system (CNS) of unknown etiology. Previous research has found that exposure to infectious agents, such as Epstein–Barr virus, and noninfectious factors, such as sunlight exposure, vitamin D, and smoking, might modify the risk of developing MS. ^1,2 Most epidemiologic research in MS, however, has been conducted in North America and Europe. ³ Studies conducted in other regions, with different patterns of exposure, might offer new clues about the genetic and environmental risk factors potentially involved in the etiology of MS.

No previous epidemiologic studies have explored risk factors for MS in Iran. Iran has been traditionally considered an area with low prevalence of MS, with a recent study reporting a prevalence of 5–13 MS cases per 100,000 in south-east Iran. ⁴ However, other studies, carried out in central parts of Iran, have found a higher prevalence (of 35.5–51.9 per 100,000), ^{5 – 7} with a possible increase of the female to male ratio in recent years. ⁸

In recent decades, Iran has experienced an important degree of development, with increased urbanization, leading to major lifestyle changes. Whether these changes affect the risk of MS remains to be studied. Thus, we have conducted a multicenter case–control study in Iran, with the aim of exploring the association of lifestyles, environmental exposures, and previous medical history with the occurrence of MS.

Methods

Results

The final sample included 394 MS cases and 394 controls. Selected characteristics of cases and controls are presented in Table 1. Age, education, and ethnicity were similar in both groups, with a higher prevalence of females among cases.

OPEN IN VIEWER

Table 1.

Sociodemographic characteristics of cases and controls

Characteristics	Cases (n = 394)	Controls (n = 394)
Age, years	30.9 (8.9)	31.0 (10.1)
Age at first symptoms, years	24.9 (8.9)	–
Gender, female	78.7	61.4
Education, some college	25.5	26.1
Ethnicity:
Persian	79.2	74.4
Azeri	5.3	4.8
Kurd	9.1	8.9
Other	6.4	11.9
Smoking, ever	9.9	8.1
MS subtype:
Relapsing–remitting	83.7	–
Secondary progressive   (with relapses)	13.7	–
Secondary progressive   (without relapses)	0.3	–
Primary progressive	2.1	–
Progressive relapsing	0.3	–
Time from onset to diagnosis, years	6.2 (5.4)	–

Higher levels of sunlight exposure were associated with a lower risk of MS (Table 2). The odds ratio (OR) and 95% confidence interval (CI) of MS associated with one additional daily hour of sunlight exposure was 0.64 and 0.55–0.74 (p < 0.0001), after adjustment for age and sex. Results did not appreciably change after adjustment for smoking, family history of MS, and previous history of infections and surgery (OR: 0.62, 95% CI: 0.53–0.73, p < 0.0001). Compared with those with sunlight exposure of 45 min per day or less, those reporting sunlight exposure greater than 2.5 h per day had 66% reduced odds of having MS (OR: 0.34, 95% CI: 0.19–0.60, p = 0.0002; see Figure 1). The inverse association was similar in men and women (p = 0.24 for interaction).

OPEN IN VIEWER

Figure 1.

Odds ratios (95% confidence intervals) of MS by approximate quartiles of daily sunlight exposure. Conditional logistic regression model adjusted for age and gender.

OPEN IN VIEWER

Table 2.

Association of selected characteristics with the incidence of multiple sclerosis

	Cases	Controls	OR (95% CI) 1	OR (95% CI) 2
Sunlight exposure, 1 h	NA	NA	0.64 (0.55–0.74)	0.62 (0.53–0.73)
Smoker, ever	39	32	1.69 (0.93–3.07)	1.72 (0.90–3.30)
Family history of MS	90	66	1.51 (1.02–2.23)	1.52 (0.99–2.31)
History of:
Tonsillectomy	23	25	0.78 (0.41–1.47)	0.65 (0.32–1.34)
Appendectomy	16	15	1.70 (0.73–4.00)	2.32 (0.89–6.05)
Hernia repair	8	9	1.28 (0.41–3.99)	1.22 (0.37–3.97)
Measles	59	44	1.28 (0.80–2.04)	1.12 (0.66–1.90)
Mumps	72	48	1.71 (1.12–2.62)	1.52 (0.93–2.49)
Chickenpox	121	111	1.13 (0.81–1.59)	1.15 (0.78–1.69)
Rubella	29	17	1.63 (0.85–3.13)	1.63 (0.81–3.28)
Hepatitis 3	3	3	1.05 (0.19–5.92)	0.65 (0.10–4.18)

1

Conditional logistic regression adjusting for age and gender.

2

Conditional logistic regression adjusting for age, gender and all the variables in the table.

3

Type not specified.

CI, confidence interval; NA, Not applicable (continuous variable); OR, odds ratio.

Family history of MS and smoking were each associated with a higher risk of MS. The corresponding OR of MS in those reporting a family history of MS was 1.51 (95% CI: 1.02–2.23, p = 0.04). Smokers had a 69% increased risk of MS compared with non-smokers (OR: 1.69, 95% CI: 0.93–3.07, p = 0.08). Gender modified the association of smoking with MS (Figure 2). In men, OR (95% CI) of MS comparing smokers vs. nonsmokers was 0.72 (0.31–1.68), p = 0.45, while the corresponding figure in women was 6.48 (1.46–28.78), p = 0.01, and p = 0.002 for interaction.

OPEN IN VIEWER

Figure 2.

Odds ratio of multiple sclerosis by smoking and gender. Results from logistic regression model adjusted for age and education level.

A medical history of common surgical procedures or several infectious disorders was not associated with MS risk in a multivariable analysis (Table 2), though a history of mumps was associated with risk of MS in a model adjusted for age and gender only. A history of infectious mononucleosis was reported by only three controls and no cases (Fisher exact test, p = 0.25). Finally, exposure to different pets or farm animals was not associated with MS risk (Table 3).

OPEN IN VIEWER

Table 3.

Exposure to pets or farm animal ownership and MS incidence

	Cases	Controls	OR (95% CI) 1
Pet/farm animal ownership
Dogs	24	39	0.57 (0.30–1.06)
Cats	22	25	0.67 (0.35–1.31)
Birds	78	77	1.11 (0.73–1.68)
Sheep	27	31	0.70 (0.32–1.54)
Goats	18	20	0.97 (0.36–2.59)
Cows	23	28	0.64 (0.30–1.40)
No. of pets / farm animals
0	284	280	1 (ref.)
1	73	61	1.27 (0.83–1.94)
2	16	27	0.55 (0.27–1.14)
3+	21	26	0.56 (0.23–1.36)

Discussion

In this case–control study performed in Iran, we have observed that sunlight exposure was associated with a lower risk of MS, while both family history of MS and smoking (only among women) were associated with a higher risk of MS. History of infectious diseases or exposure to pets and farm animals was not associated with the risk of MS.

Few studies of risk factors for MS have been conducted in Southwest and Southern Asia. In a small case–control study conducted in Israel, higher socioeconomic status in childhood was associated with increased risk of MS later in life. ¹⁰ In India, a case–control study including 56 cases and 147 controls found that ownership of dogs, but not of other animals, and a history of measles were associated with higher MS risk. ¹¹ No previous studies have evaluated risk factors for MS in the Iranian population. To our knowledge, our epidemiologic study on the risk factors for MS in this geographic area is the largest to date.

The observed inverse association between sunlight exposure and MS risk in this Iranian population is consistent with previous reports. ² Exposure of at least 45 min per day was associated with a 70% lower risk of MS. The most likely mechanism linking sunlight exposure with MS incidence is through vitamin D production. Synthesis of vitamin D increases with sunlight exposure, and vitamin D has immunomodulatory actions that might prevent the development of MS. ¹² In addition, strong epidemiologic evidence, including prospective studies, suggests that higher levels of vitamin D are related to a lower risk of MS. ¹² The increasing sex ratio of MS in Iran has also been linked to vitamin D deficiency among young women, with sun avoidance being a major cause, among other factors. ⁸ Urbanization has been shown to be a risk factor for vitamin D deficiency ¹³ and, given increased urbanization during recent years in Iran, this might in turn have increased the risk of vitamin D deficiency, resulting in increased risk of MS among susceptible individuals. ¹⁴

Previous studies have found smoking to be an independent risk factor for MS, ^2,15,16 but none reported an interaction between smoking and gender. The association of smoking with MS by gender in our study might be due to diversity in smoking habits between men and women in Iran. In our sample, prevalence of smoking was 3.4% among women versus 22.0% among men, similar to that reported by another study conducted in Iran (3.6% vs. 26% in women and men, respectively). ¹⁷ However, interactions between smoking and gender have been reported for other disorders, including amyotrophic lateral sclerosis, ^18,19 chronic obstructive pulmonary disease, ²⁰ and thyroid disorders. ²¹ These differences might be related to the differential metabolism of smoke toxicants between men and women. ²⁰ An alternative explanation, though, is that the association of smoking with MS in women is confounded by other lifestyle variables. In addition, since the number of smokers in our study was small, this could have influenced the analysis. Future studies in this population should determine whether smoking affects progression of MS, as observed in Western populations. ^16,22

In contrast with other studies that have found higher risk of MS associated with exposure to dogs, ¹¹ cats, ²³ or birds, ²⁴ we did not observe any association between previous exposure to farm animals or pets and the risk of MS. Previous studies have tried to explain these associations by referencing transmission of infectious agents from animals to humans. However, the only infectious agent that has been clearly associated with MS is Epstein–Barr virus, and no unequivocal evidence has been provided linking animal infections with MS etiology. ^1,25,26 Our results offer support for this lack of association.

Vaccinations have been linked with the risk of MS in some studies. ^27,28 In our population, self-reported information on vaccinations was available. However, due to the lack of access to vaccination records and the potential for differential recall in cases and controls, we decided against including associations of vaccinations with MS risk. In future work, the validity of self-reported vaccinations should be determined. This limitation also applies to the assessment of history of infections, which was probably subject to recall bias. This might be the reason why a history of infectious mononucleosis was underrepresented in our cohort and was not associated with the risk of MS as shown by others. ²⁹

In the last decades, Iran has experienced a number of important changes. The growth of the urban population has been approximately twice as fast as the general population growth. ³⁰ Simultaneously, fertility rates have declined from seven children per woman in 1986 to two children per woman in 2000, in parallel with use of contraception ³¹ (though limited evidence exists of oral contraception affecting MS risk). ³² Therefore, a large proportion of the population is young and in the age range at risk for MS. In addition, it has been shown that a large proportion of Iranian women suffer from vitamin D deficiency, which is mainly due to avoidance of sun in combination with cultural and lifestyle factors. ^33,34 Moreover, Iranians are mainly Caucasian, and, according to phylogenetic studies, they are the closest ethnic group to the European ancestry, which is the most susceptible to MS. ³⁵ A combination of these and related variables might be responsible for recent increases in MS incidence in Iran, and in the specific pattern of risk factors for MS observed in this study. ^4,5,8 However, other factors, such as enhanced diagnosis of MS due to advanced diagnostic criteria, more widespread use of MRI, increased number of neurologists, and increased awareness by physicians, could in part be responsible for this increase. ³⁶

Strengths of the current study include the relatively large sample size, the availability of information on potential confounders, including family history of MS, and the quality of the diagnosis of MS, with all cases meeting McDonald criteria. However, our study has some limitations. Exposure information was obtained retrospectively, with some of the exposures referring to a period many years before the study. Exposure misclassification, therefore, may have occurred. It is possible that the quality of exposure information is different between cases and controls, which could lead to differential measurement error. As in most case–control studies, selection bias is a threat if controls are not representative of the population from which the cases arose. In this particular study, controls were specifically selected from among friends and relatives of the cases, guaranteeing that both cases and controls originated from the same source population. The matched analysis helped to control for this design characteristic. Pets such as dogs are not common in Iranian culture, and exposure to farm animals was not highly prevalent in our sample due to the urban population residence. Further studies looking at the association between exposure to pets and farm animals and risk of MS in the Iranian population are warranted. Confounding could also explain the observed associations. Nonetheless, we adjusted for factors potentially related to smoking and sunlight exposure, such as age, gender, and education. Finally, the participating clinics were not a random sample of all the MS clinics in Iran, but patients attending these clinics were similar to the general population of MS patients (e.g. gender and age distribution, other sociodemographic variables), suggesting that our results are generalizable to all MS patients in Iran.

In conclusion, our results provide additional evidence of a potential protective effect of sunlight exposure, possibly mediated through vitamin D levels, and of smoking as a risk factor for MS. Future studies should determine whether a true interaction exists between smoking, gender, and the risk of MS.