The comprehensive record-keeping and public health research of the Scandinavian countries have produced many valuable epidemiologic studies, perhaps most famously the identical-twins studies that have led to estimates of genetic and environmental contributions to diseases ranging from schizophrenia to Alzheimer's. A study of Swedish twins, including many over the age of 80, now adds weight to the argument that the genetic contribution to late-onset Alzheimer's disease is not as great as previously estimated.

Writing in the early online edition of the Annals of Neurology, Nancy Pedersen of the Karolinska Institute in Stockholm, Sweden, and her colleagues in Sweden and the United States report on the first quantitative genetic study of identical twins that focuses on incident cases (arising during the course of the study period) rather than prevalent cases (a “snapshot” of cases at a given time). Earlier twin studies using prevalence designs had estimated that genes account for the majority of the risk for Alzheimer's, even among late-onset cases. In some studies, the estimate of genetic versus environmental risk was as high as 75 percent. However, subsequent studies have suggested that this figure is too high (see for example, the findings of Jeremy Silverman and colleagues from earlier this year.)

The current study, which had an average of five years of follow-up, was notable for the large number of twin pairs over the age of 80 (297 of the total 662 initially enrolled). Pedersen and colleagues found that among 221 pairs of monozygotic twins, 21 had one sibling afflicted with Alzheimer's, and five had both siblings with the disease (probandwise concordance rate, 32.2 percent). The corresponding figures for 329 dizygotic twins were 42 discordant pairs, but only two concordant pairs (probandwise concordance rate, 8.7 percent). Structural modeling of genetic versus environmental contributions indicated that the genetic contribution to AD in this sample was 48 percent—considerably less than seen in the earlier prevalence studies. Subjects over the age of 80 did not have a significantly greater environmental risk relative to those under 80.

Much recent news from the epidemiology of Alzheimer's disease leaves an impression that if we haven't all earned advanced degrees and been highly active early in life, we have lost our chance to counteract our genetic fate. By contrast, this study carries good news for preventive measures late in life. According to the authors, incidence design tends to evaluate recent environmental influences, rather than cumulative lifetime exposures. For example, the study design excludes pairs in which a sibling developed the disease earlier in life.

"If supported by further epidemiological findings, one might speculate that late-life interventions that help the individual to maintain cognitive status may help reduce the total population burden of Alzheimer morbidity before death," the authors write. Among the commonly cited risk factors that might be relevant, they suggest "vitamin B-12 deficits, drug-induced increments or decrements, recent head trauma, and other late-life exposures that are sufficient to reduce cognitive reserve past the threshold for clinical manifestation of the disease."

The researchers also opine that their results bear upon efforts to identify further genes affecting AD risk; for example, identifying sibling pairs whose age of onset is close, within five to six years of each other, should increase the likelihood of finding genetically "loaded" pairs. “Furthermore, comparisons of sibpairs initially screened when they are young-old and those screened when they are old-old are likely to help focus findings on genes of importance for timing of onset and slope of decline rather than genes associated more generally with longevity and survival," the authors write.—Hakon Heimer


  1. This interesting paper raises a number of technical questions in my mind.

    For one, I noticed that in Table 2, variance explained by genetics was significant for the younger series of twins, but apparently not for the older twin series. This would be consistent with family data from my group and others.

    For another, it is unclear to me how much of a role low statistical power may have had in the failure to provide significant evidence that heritability is lower in those older than 80 years of age.

    A third issue is the more or less comparable concordance rates observed in the monozygotic (MZ) twins in the younger (probandwise: 33.3 percent) and older (31.6 percent) series contrasted with the big increase from 0 percent to 10.8 percent in the younger and older dizygotic (DZ) twins series. Assuming that the role of genetic factors is reduced as age at onset increases, one might hypothesize that concordance rates would be reduced in an older series of MZ twins versus a younger one. That was not the case in this sample. However, because AD is so highly age-related, with incidence doubling every five years, many more cases can be expected to emerge at these late ages, and this would tend to make the likelihood of the concordance rate between any two individuals (related or not) increase with age. So what is striking to me, then, is not so much the persistence of the MZ concordance rate in the older versus the younger series (in fact, there was a very slight decline) as the absence of an increase. And, as noted, this is in contrast to the DZ concordance rate, which does increase substantially in the older versus the younger twin series.

    Another curious result (related, I think, to the above) is the observed incidence rates in the younger and older MZ and DZ twin series. At 4.5 percent, the incidence rate among the younger MZ twins was more than twice that observed in the younger DZ twins. Given strong genetic factors, one would expect higher rates overall in comparably aged MZ over DZ twins, since the likelihood of concordance is higher in the MZ twins (i.e., when one case arises, there is an increased likelihood over DZ twins that a second one will arise). In the older series, however, the incidence rate for the MZ twins was 8 percent, substantially less than the 12.5 percent rate in the DZ twins. Where there was a greater than sixfold increased incidence in older DZ twins compared with younger DZ twins, there was only a 1.8-fold increased incidence in older MZ twins compared with younger MZ twins.


Make a Comment

To make a comment you must login or register.


No Available References

Further Reading


  1. . Familial patterns of risk in very late-onset Alzheimer disease. Arch Gen Psychiatry. 2003 Feb;60(2):190-7. PubMed.

Primary Papers

  1. . How heritable is Alzheimer's disease late in life? Findings from Swedish twins. Ann Neurol. 2004 Feb;55(2):180-5. PubMed.