Early Life Circumstances and Late Life Alzheimer’s Disease

Dementia is the most frequent and devastating neurologic disorder in the elderly. Approximately two thirds of all patients who have symptoms of dementia are diagnosed with Alzheimer’s disease (AD). Despite intensive research efforts and accumulating insights into some of the pathophysiologic mechanisms involved, 1–3 its causes are still largely unknown. Alzheimer’s disease can manifest itself before 60 years of age, yet such instances are rare and concern usually monogenetic, autosomal dominant disorders. Most cases that occur in later life are of multifactorial and heterogeneous etiology. What we have come to diagnose clinically as Alzheimer’s disease may actually be a complex mixture of disease pathology rather than one disease. 4 Although the causes of Alzheimer’s disease are largely obscure, they must exert their effects before the onset of disease. Unfortunately, onset is difficult to pinpoint. It is commonly assumed that the preclinical stages of Alzheimer’s disease extend over years, maybe even decades. 5 The central culprit in the pathogenesis appears to be accumulation of the amyloid beta-protein (A beta), either through increased production or decreased clearance. 6 Theoretically, this abnormality suggests that any factor that directly or indirectly interferes with normal amyloid processing may contribute to risk of disease. 7 It also suggests that late onset Alzheimer’s disease is the result of life-long accumulation of damage from a combination of endogenous and exogenous factors. Postmortem studies have shown that many older individuals without dementia have sufficient plaques and tangles in their brain to meet neuropathologic criteria for Alzheimer’s disease. This finding has fueled the cognitive reserve hypothesis, which holds that “the more brain you have, the more you can lose.” A beta accumulation may be the underlying pathogenetic mechanism, but whether this accumulation results in clinically overt dementia depends on the initial number and quality of the neurons and interneuronal connections that one started out with. Since the brain grows and matures throughout childhood, an ensuing hypothesis is that Alzheimer’s disease may originate early in life. Reported associations between educational level, 8,9 young-adult cognitive performance, 10,11 and head circumference 12,13 and Alzheimer’s disease appear to support this view and have lead to increased interest in early life antecedents of Alzheimer’s disease. The growing interest in early life risk factors for Alzheimer’s disease coincides with more widespread hypotheses on early life determinants of diseases in adulthood. On the one hand there are studies that emphasize the importance of social determinants of health inequalities. Although the relation between socioeconomic status and disease is seen over the entire life course, adult life socioeconomic circumstances often have their roots in early life. 14 On the other hand there are studies that focus on early life growth patterns in the context of the fetal and postnatal programming hypothesis, 15 widely known as the Barker hypothesis. Although at times heavily criticized, 16,17 this hypothesis has gained large appeal. 18 Also, minor abnormalities in neuronal development may determine increased susceptibility to neurodegeneration. 19 In this issue, Moceri and colleagues report on a case-control study of Alzheimer’s disease that examines early-life socioeconomic determinants through U.S. census records and birth certificates of participants. 20 They found that, compared with subjects whose father held a nonmanual occupation, subjects whose father held a manual occupation had a higher risk for AD. The risk was particularly increased when the subject himself or herself carried an apolipoprotein E (APOE) ε4 allele. 20 However plausible and appealing the idea of early life risk factors for late life Alzheimer’s disease may be, do these new data support it? And do they advance our understanding of what causes Alzheimer’s disease? There are several issues to consider. The first issue concerns the validity of studies on early life risk factors of late life disease. Prospective studies are not feasible. Retrospective studies are usually limited by incomplete data and attrition and in particular the bias that may result from differentially missing data. Moceri and colleagues tried to beat time by choosing the more efficient case-control design and tracing back census records and birth certificates of cases and controls in their study. Still, this approach leaves open the possibility for selection bias. Indeed, eligible controls who refused participation had lower educational levels, and the cases for whom information could be found had a higher education than those whose information could not be retrieved. This selection was such that the investigators considered an association between lower education and Alzheimer’s disease to be spurious. Yet educational level is closely related to socioeconomic status, often even used as a proxy marker thereof. This connection raises questions about the validity of the findings regarding the association between that other marker of socioeconomic status, father’s occupation, and risk of Alzheimer’s disease in this dataset. Unfortunately, the authors do not address this issue. Second, what is the relevant underlying biological parameter when one considers father’s occupation in relation to late-life disease risk? Moceri et al. suggest that father’s occupation is an approximate measure of the early-life environment, but what is that? Childhood exposure to infections, or specific behavioral patterns, or nutritional status, or stress, or intellectual stimulation or yet something completely different? Alternatively, one might consider that parental occupation reflects to some extent genetic make-up rather than environment. Or that parental occupation marks some later-life corollaries, including a person’s own adult occupation and socioeconomic circumstances or cardiovascular risk profile. Father’s occupation is not a single biologic parameter, but may be read as a Rorschach blot that acquires its interpretation only in the eye of the beholder. The proposition that father’s occupation may mark determinants of disease that only occur later in life merits further attention. Moceri et al. assessed the socioeconomic measures for the childhood period only, and did not take into account what happened later on. Early life socioeconomic status is correlated with later life socioeconomic status. The relative importance of influences of socioeconomic factors at different stages during a life span varies for disease outcomes. 21 Socioeconomic status in middle age is associated with morbidity in old age, including poor mental health. 22 Socioeconomic status is strongly associated with cardiovascular risk and vascular factors have been invoked in the etiology of late life dementia and Alzheimer’s disease. 23 There is evidence that the socially patterned accumulation of cardiovascular risk begins in childhood and continues, according to socioeconomic position, during adulthood. Cardiovascular risk factors in adulthood are more strongly related to adult than to childhood socioeconomic position. 24 If the specific influence of socioeconomic environment in early life is to be investigated, data on circumstances in both early and later life are required. 14 When only socioeconomic status from childhood is considered, it cannot be concluded that the relevant underlying exposure that led to the increased risk for late life Alzheimer’s disease occurred during that childhood period. Last, when we do not even know the main effect measures, how can we meaningfully interpret interaction with a genetic risk factor for the disease? Hence, what does this study contribute? It emphasizes the importance of taking into account possible exposures and cumulative damage over the entire life span in the search for causes of late life disorders. It may be used by those who seek to promote equal access to public health and diminish health inequalities. It does not refute the idea that the origins of Alzheimer’s disease may start in childhood. But above all, it indicates the need to define exposures as precisely as possible with regard to the underlying phenomenon under investigation to be able to yield etiologically meaningful information.