Reduced frequency of ALS in an ethnically mixed population

ALS is a progressive neurodegenerative disease of middle and late life. The onset of ALS is age-related, with the highest rate of onset between ages 55 and 75 years.¹ A recent systematic review of ALS has found a relatively uniform incidence across the white populations of Europe and North America, and a lower incidence among African, Asian, and Hispanic ethnicities. The degree to which this difference is due to reduced ascertainment and lower life expectancy remains unclear.²

Mortality is a valuable indicator of ALS incidence.³ Lower mortality rates have been consistently observed in Hispanic and Asian groups compared with similar studies in white populations.² Some of the differences may be accounted for by poor ascertainment in developing countries and by differential access to health services. Nonetheless, such variation in incidence would also support the notion that ancestral origin may play a role in disease pathogenesis.^4,5 We undertook a mortality study of ALS in Cuba over a 6-year period. Cuba was selected because the population is genetically admixed and the life expectancy is similar to that of the developed world.^6,7 Medical care in Cuba is of a high standard, is free, and is available equally to all Cuban citizens.^7–9 The Cuban census defines the population as those self-reporting to be of Spanish (65%), mixed (24%), and African origin (10%).⁶ Cuba is thus ideally suited to determine whether ancestral origin affects disease susceptibility.

METHODS

After ethical approval from the Cuban Ministry of Health and the Institute of Neurology and Neurosurgery, the Cuban Centre for Health Statistics was searched for the years 2001 through 2006 for all death records in Cuba containing international disease classification codes that correspond to ALS, primary lateral sclerosis, progressive bulbar palsy, spinal muscular atrophy, and motor neuron disease (MND). All mortality rates were calculated using records with a coding for ALS/MND listed anywhere on the death certificate. Variables obtained for each record included age at death, sex, skin color, and province and municipality of residence and place of death.

Fifteen age groups were divided into 5-year age groups ranging from 15 to 85+ years. Mortality rates were age-adjusted using the 2002 Cuban census, standardized and adjusted to the US population census of 2000. The urban population was derived from the population of cities greater than 1 million inhabitants (Havana city, Holguin, and Santiago de Cuba), and the rural population was derived from provinces with low population densities, including Cienfuegos and Sancti Spiritus.

Standardized adjusted mortality rates and confidence intervals (CIs) were stratified by sex, ethnicity, age, and geographic region at time of death. CIs were calculated assuming a Poisson distribution. Differences in mortality rates across ethnic groups were evaluated using the two-tailed Z test. Rate ratios (RRs) with white or black as the reference group were also calculated with 95% CI. Statistical analyses were conducted using STATA 9 and Epiinfo version 3.5.1 (Centers for Disease Control and Prevention, Atlanta, GA).

RESULTS

From 2001 to 2006, a total of 432 deaths were reported in Cuba based on death certificates containing codes corresponding to ALS/MND.

The overall crude mortality rate per 100,000 person-years for the Cuban population older than 15 years was 0.67 (CI 0.6–0.72), and the adjusted mortality rate was 0.83 person-years (CI 0.72–0.8). Overall, there was a slight preponderance of men (1.1:1). The peak age of death was slightly older that in white populations. A decline in incidence in the very old, observed in white populations, was less evident (figure).

• Download figure
• Open in new tab
• Download powerpoint

Figure Age-specific mortality from ALS in the Cuban population, stratified by sex and ancestral origin

The age distribution of mortality from ALS is similar to that of other populations.¹ The increasing mortality in the elderly, especially among black and mixed populations, provides evidence against differential reporting in death certificates.

The adjusted mortality rate per 100,000 was lower in the mixed population (0.55; CI 0.4–0.72) than in whites (0.93; CI 0.83–1.03) and blacks (0.87; CI 0.62–1.17) (table 1). These differences were highly statistically significant when white and mixed populations were compared (z score: 5.8; p < 0.0000001; RR [white as reference]: 0.44; 95% CI 0.33–0.59) and when black and mixed populations were compared (z score: 6.2; p < 0.0000001; RR [black as reference]: 0.31; 95% CI 0.21–0.45). The difference in rates between black and white populations was not statistically significant (z score: 0.82; p = 0.41; RR [white as reference]: 0.88; 95% CI 0.64–1.2).

When mortality rates of the ethic subpopulations were compared within the 45- to 74-year age group, which is at highest risk of developing ALS, the difference between the white and mixed groups remained statistically significant, but differences between the black and mixed groups did not reach statistical significance, possibly because of small numbers (table 1).

In whites, ALS was slightly more common in men than in women (1.25:1). This ratio was reversed in the black population, where the male to female ratio was 1:1.4, whereas in the mixed population the ratio was 1:1 (figure).

With the exception of Havana province, there was no difference in the mortality rates within each of the 14 Cuban provinces (table 2), nor was there a difference in crude and adjusted rates between urban and rural dwellers, with the exception of rural-dwelling females (table 3). There was no correlation between the number of neurologists in each region and the mortality rate from ALS (r = 0.268, p = 0.335).

DISCUSSION

This analysis of mortality from ALS in Cuba shows that rates are considerably lower than those reported in populations of European extraction. The unique population structure of Cuba, coupled with the excellent health care system and census collection system, has permitted more detailed population-based subanalysis than has been the case for mortality studies of other non-European populations, or of mixed populations in US studies.

The largest ancestral population in Cuba is drawn from Spanish immigrants from the Canary Islands, Catalonia, Andalucia, Galicia, and Asturias and is classified as white in the Cuban census. The black population is descended from Africans of Yoruba, Igbo, Kongo, and other tribes. The indigenous aboriginal Cuban population no longer exists as a defined group.

Although self-reported skin color, which is the method of classification used by the Cuban census,⁶ has been considered a reliable proxy of ancestral origin,¹⁰ it is also the case that a very high proportion of the Cuban population is genetically admixed. Indeed, a portion of the “white” population in Cuba would likely be classified as Hispanic were they to reside in the United States. A recent admixture study of 206 normal subjects from Havana city using 15 ancestral informative markers demonstrated 15% non-European ancestry in the self-reported white group, 55% European and 45% African ancestry in the self-reported mixed group, and 74% West African contribution in the self-reported black group (M. Nazabal, personal communication, 2009).

The low overall adjusted mortality rate of 0.8 per 100,000 for the Cuban population is consistent with previous reports of lower rates of ALS in Mexico and Chile and among the Hispanic population in the United States^2,11–13 (table e-1 on the Neurology® Web site at www.neurology.org). Disparities in previously reported incidence and prevalence of ALS in non-European/North American regions have been attributed in part to low ascertainment as a result of socioeconomic status and poor access to health care. A further confounding factor has been the age profiles of some of the study populations, rendering cross-study comparisons difficult in the absence of age-specific data from each epidemiologic study.² Neither of these confounders pertain in the study; indeed, the age distribution, coupled with a trend toward an increase in mortality from ALS in the very old, supports the likelihood of compete ascertainment with the population. This finding contrasts with the possible underascertainment of the very old in European populations.¹⁴

A striking feature of this study was the low adjusted mortality rates in the mixed group compared with the white and black groups. These lower rates are unlikely to reflect an ascertainment bias, because access to health care in Cuba is not stratified according to ethnicity or ancestral origin^7–9 and regional mortality rates did not correlate with the number of neurologists in each region. It is also unlikely to reflect differences in environmental exposure or behavior, because ethnic groups are no longer stratified by socioeconomic factors in Cuba.⁹ With the exception of the province of Havana (in which adjusted mortality rates were higher with a female predominance), the lower rates were observed across all provinces of Cuba, and there was no observed difference between urban and rural dwellers. Furthermore, the reduced mortality rate in the mixed population was present in all ages, indicating that the effect is not a function of underlying differences in the population structure at different age intervals. The most likely explanation is that the observation is a valid reflection of the occurrence of ALS in Cuba, implying that disease susceptibility is reduced in the ethnically mixed population.

Existing evidence supports the concept of ALS as a complex genetic disease caused by multiple susceptibility genes interacting with a variety of environmental risks.¹⁵ European populations share common ancestral origins and, depending on the degree of relatedness, are likely to share a variety of rare “at-risk” genes, combinations of which may increase susceptibility to disease. Moreover, the population frequency of some rare at-risk genes may be higher in more homogenous populations, as has been demonstrated in Irish and Scottish ALS studies.^4,5

Conversely, admixed populations, containing a much wider variety and different combinations of at-risk alleles, are therefore likely to experience a lower overall risk of developing the disease. Although scant attention has been paid to variations in the frequency and natural history of neurodegenerative diseases outside ancestral European populations, there is evolving evidence that the prevalence of Parkinson disease may be reduced in admixed populations,^16,17 and that those of non-European or admixed origin with Alzheimer disease may differ from ancestral Europeans with respect to age at onset, phenotype, and survival.^18–20

Because this study was based on mortality data, it is not clear whether the phenotype of ALS in Cuba differs from that observed in less admixed populations. In our study, the male-to-female ratio for ALS is inverted in the black population, which differs from most other populations. The reasons for this are unclear and require further investigation, as does the apparently lower mortality rates in rural-dwelling females and the higher mortality rates in the province of Havana. The trend for mortality rates to continue to increase in the very old differs from European studies¹; however, the numbers are small, and this observation will require verification in a prospective population-based incidence study.

In diseases of short duration such as ALS, although mortality can be viewed as an indirect measure of incidence, mortality studies are invariably limited by their retrospective nature. The degree of underestimation of the true frequency of ALS in Cuba in this study is difficult to determine. Medical training and data collection in Cuba are of a high standard; autopsies are frequently performed,²¹ and all patients require death certification before interment.⁷ The number of deaths attributed to ill-defined causes is very low (0.7%), and, in contrast to other Caribbean and Latin American countries, extensive mortality and morbidity data by cause and province are collected and published.⁷ Furthermore, because of the homogeneity of the Cuban health care system and the uniform access to health care for all ethnic groups,^7–9 errors in reporting are expected to be the same for all the population, and it is unlikely that this factor should account for the differences between the ancestral groups observed within the Cuban population.

Given the overlap between ALS and other forms of neurodegeneration,²² the relative protection conferred by genetic admixture identified in this study may also be of wider relevance.

AUTHOR CONTRIBUTIONS

Statistical analysis was performed by M. Carbonara.