Tobacco Use in Alzheimer??s Disease

Smoking and Alzheimer's Disease: A Review of the Epidemiological Evidence

“…A total of 19 case-control studies have been identified [18,[26][27][28][29][30][31][32][33][34][35][36][37][38][39][40][41][42][43][44], Table 2 lists, for each study, authors, reference, location, period, sexes involved and the numbers of cases and controls. 10 of the studies were conducted in the USA, 6 in Europe, with 1 each in Austra lia, Japan and Colombia.…”

“…Another study [36] com- [27] hospital spouse Brenner et al [28] population population/random (no dementia) Broe et al [29] hospital and general practitioner population/general practitioner Chandra et al [30] hospital hospital (no dementia) Dewey et al [31] general practitioner general practitioner (no dementia) Ferini-Strarabi et al [32] hospital population/neighbourhood French etal. [33] veterans' hospital (a) hospital (no dementia) (b) population/neighbourhood Graves et al [34] hospital friend or non-blood relative Grossberg et al [35] hospital hospital (no dementia) Hebert et al [36] population -performed well on memory test in 1982, probably AD on retest in 1985 possible AD or unaffected on retest Heyman et al [37] hospital population/random digit dialling Hofman and Van Duijn [18,38] population population/random Jones et al [39] hospital local luncheon club and care home Joya et al [40] not stated not stated Katzman et al [41] those developing AD in 5-year prospective study of volunteer subjects those not developing AD Rondo et al [42] hospital population/neighbourhood Shalat et al [43] veterans' hospital population/neighbourhood Soininen and Heinoncn [44] population nursing home population [27] all Brenner et al [28] 60+ Broe et al [29] all Chandra et al [30] >70 Dewey etal. [31] 65+ Ferini-Strambi et al [32] <65 French et al [33] all Graves et al [34] all Grossberg et al [35] all Hebert et al [36] 65+ Heyman et al [37] <67 Hofman and van Duijn [18,38] <65 Jones et al [39] >65 Joya et al [40] not stated Katzman et al ...…”

“…The same problem arises with the studies which included patients from nursing and chronic care homes as controls. One study [27] used spouses as the control group, and 1 [34] used a friend or non-blood relative. Both these control groups also seemed somewhat unsatisfactory because of the known concordance of smoking habits be tween husband and wife and between a per son and his or her friends and relations.…”

“…The majority of the studies specifically matched controls for age and sex. Exceptions were the study of Barclay and Kheyfets [27], where controls were spouses; the studies of Dewey et al [31] and Hebert et al [36], where age and sex were adjusted for in the analysis; the study of Jones et al [39], where only age was noted to be a matching factor, and the study of Katzman et al [41], which followed an initially healthy group of subjects aged 75-85 and compared those who subsequently got AD with a group of those who did not. Some studies matched on other criteria, such as race, area of residence and social status.…”

“…In the study of Kondo et al [42], data for the controls were collected both from the subjects and the next of kin in order to test answer comparability. Unfortunately, it was not made clear which answers were used in the comparisons with [27] spouse controls inappropriate Brenner et al [28] none Broe et al [29] none Chandrae! al.…”