Browse Subject Areas
?

Click through the PLOS taxonomy to find articles in your field.

# Correlation of Alzheimer’s Disease Death Rates with Historical Per Capita Personal Income in the USA

• Dariusz Stępkowski,
• Grażyna Woźniak,
• Marcin Studnicki
x

## Methods

The PCPI data for each state for years 1929–2011 were taken from the FRED St. Louis data base (http://research.stlouisfed.org/series) and compiled into an Excel worksheet (available in S1 File). For each year, linear regression was performed with AADRs for years: 2000, 2005 and 2008 using Origin 7.5 software (AADRs were taken from National Vital Statistics Reports, Deaths final data for 1998–2010). For all three sets of data on AADR, the highest correlation was obtained with PCPI for the year 1970. Regressions were done using PCPI for each state normalized to PCPI in 2011 using the following formula: PCPI in a given year * Fn where Fn = PCPI (2011) (average US in 2011) /PCPI in a given year (average US in that year). This normalization allows for comparison of the slope for each year, without affecting the correlation coefficient R—an indicator of the strength of correlation. This normalization is different from the recalculation of the income in constant dollars, which gives real values of the income but does not permit to compare slopes of correlations from different years. The absolute value of the slope of the correlation between PCPI and AADR can be interpreted as the number of dollars by which a decrease in PCPI causes one additional AD-linked death per 100 000 population. The higher the absolute value of the slope the steeper the relation of death rates with PCPI. It means that an increase in the death rate by one victim of AD per 100 000 population is related to a higher decrease in PCPI. The normalization used by us allows the slopes of the correlations from different years to be compared with each other despite the differences of the real value of dollars corresponding to those slopes. Using constant dollars recalculated to the value of 2011 dollars would not allow such a comparison due to differences in the level of income. The average US PCPI in constant dollars for each year in the period 1929–2011 was calculated using the US inflation calculator (http://www.usinflationcalculator.com/) based on the Consumer Price Index for urban consumers prepared by the U.S. Department of Labor, Bureau of Labor Statistic. Pearson’s coefficient R was chosen as the measure of the strength of correlation over R2 since it differentiates between negative and positive correlation and offers higher sensitivity (due to the wider range of variability (-1,1) than that of R2 (0,1)). Additionally, unlike R2, R changes linearly with the change in data values therefore, R is more suitable for use to relate the variability of R with the susceptibility of population to AD.

### Statistical methods

A crucial question about the statistical significance of the oscillatory nature of R variability and the trend lines (for the correlations between PCPI and age-adjusted AD death rates (AADRs) in 2000, 2005,2008) was addressed by dividing the studied period into seven segments (seasons). These segments correspond to periods between the peaks of consecutive highest and lowest correlations (extremes). The first segment spans the period between 1929 and 1936, the second from 1937 to 1950, the third from 1951 to 1969, the fourth from 1970 to 1980, the fifth from 1981 to 1991, the sixth from 1992 to 2002 and the seventh the period from 2003 to 2011. For each period the Mann-Kendall test and value of tau correlation were calculated. These tests were used to statistically assess the monotonicity of upward or downward trends of R variability in time. The results of these tests are presented in Supporting Information 2. Furthermore, the 95% confidence interval from polynomial regressions (degree 16) was used to determine the statistical significance of the differences between the extreme (minima and maxima) values of the regression curve. The polynomial regressions for AADR in 2000, 2005 and 2008 were performed using the R 3.1.2 software with lm function. The 95% confidence intervals are presented as the upper and lower confidence borders for the designated polynomial functions and were based on Student's t distribution. The pattern of division of the total studied period and the validity of extremes of the regression curve were confirmed by differential analysis. Differential analysis was performed by subtraction of the R value for a given year from the R value of the preceding year. These differentials were plotted against time and regressed by polynomial regression procedure similarly as for correlations between PCPI and AADR. The first derivative function obtained in this way zeroed in most cases of the extreme values of the original curve, confirming both the validity of segmental division and of the extreme points of the original correlations (minima and maxima).

## Discussion

Crucial to the evaluation of the quality of the presented results is their potential susceptibility to many possible sources of bias such as, for example, migration of patients between states. Such migration has been estimated to be rather high [31]. However, our results, showing a strong correlation of PCPI with AADRs, suggest only a limited significance of this level of migration. Other possible limitations of using local death rates for analysis of geographical distribution of a disease impact have been discussed by Glymour et al. [31]. Those authors considered them as insignificant. Factors unrelated to personal income such as differences in climate, rural to urban population ratio, stratification of personal income in the population, or average educational attainment or overestimation of death rates due to higher awareness of AD impact on society can also, due to the same reasons as in the case of migration, be considered of limited significance.

The variability of the correlations of PCPI between 1929 and 2005 and the death rates is rather smooth despite the fairly narrow time window of the income data used (one-year state average PCPIs). This indicates that the observed oscillatory character of R changes over time reflects gradual processes which occur in a time span longer than one year. Figs 36 indicate that the most plausible interpretation of this phenomenon is a significant influence of lifestyle-related events which occurred in the past rather than temporal economic fluctuations. Among the possible biological factors influencing this correlation, infectious diseases affecting large segments of the population in the past and historical changes in the lifestyle, among them diet, are most likely. Diet, unlike infections, is one of the factors related significantly to the income level and affects statistically the majority of a population. Our studies point to the fact that historical diet changes had a profound influence on the susceptibility of the population to AD. This is not a new observation since the influence of diet on the risk of developing AD is well documented [12,13,14,15]. Our analysis, however, offers some advantages over those earlier studies because, we use whole population data and thus we avoid the problem of selection and recall errors which may bias some of the epidemiological studies.

The mechanism by which life style influences the susceptibility to LOAD most probably relies on epigenetic changes caused by certain diet components and other life style factors [32,33,34]). Since epigenetic changes are reversible, the potential interventional space is likely to be achievable through reversing the undesired epigenetic changes, as has been suggested by some authors [32,34,35,36,37].

In the light of our results, the differences in AD mortality rates between races mentioned in the first paragraph of Rationale can be largely explained by differences in per capita income, since populations of African-Americans and Hispanic-origin Americans have lower per capita incomes than whites (US Census Bureau 2010 ACS 1 Year Estimates). Race-specific diets and other lifestyle factors, e.g., educational attainment (US Census Bureau 2010 ACS 1 Year Estimates) or others, related to the income may contribute to racial differences in AD mortality rates.

We have shown that the novel whole-population approach used here gives valuable results allowing identification of the long-gone causes of the enormous increase in the prevalence of AD in the USA in recent years. This was possible owing to the use of a general parameter of the healthiness of the lifestyle of a population; in our case the historical data of PCPI, and their correlation with death rates. Changing the habits of a population toward more healthy nutrition and other beneficial lifestyle behaviors, such as physical, mental and social activity, appear as the most effective measures to be undertaken to combat this disease. A similar strategy of reducing AD prevalence in the population by reducing the influence of certain risk factors has been proposed by Barnes and Yaffe [41]. A recently published paper by Grant [42] uses a similar logic to ours of analyzing the effects of historical changes of diet on the present AD mortality in Japan and eight developing countries. The author concludes, in line with our results, that the dietary history of these countries has affected the present AD prevalence. Our approach of relating semi-quantitatively the history of a population’s lifestyle with the present mortality data could be easily applied to diseases other than AD, such as cancer, diabetes, cardiovascular disease and others in the hope of finding more efficient strategies of prevention.

## Supporting Information

### S1 File. Dataset with Excel spreadsheet containing PCPI and normalized PCPI data

https://doi.org/10.1371/journal.pone.0126139.s001

(XLS)

### S2 File. Fig A in S2 File Variability in time of correlation coefficients R for correlations between US state PCPIs in the period 1929–2005 and age-adjusted AD death rates (AADRs) for the respective states in 2000.

Fig B in S2 File Variability in time of correlation coefficients R for correlations of US states’ PCPIs in period 1929–2005 against age-adjusted AD death rates (AADRs) for these states in 2008. Table A in S2 File Parameters of correlation between PCPIs for each state of the USA and AADRs in 2005. Table B. The Kendall’s Tau correlation coefficients and results of the Mann-Kendall trend test for the correlations between PCPI and age-adjusted AD death rates (AADRs) in 2005 (p value)

https://doi.org/10.1371/journal.pone.0126139.s002

(DOC)

## Acknowledgments

Jan Fronk, Wojciech Otto and Tomasz M. Stępkowski are gratefully acknowledged for critical reading of the manuscript.

## Author Contributions

Conceived and designed the experiments: DS. Performed the experiments: DS GW MS. Analyzed the data: DS GW MS. Wrote the paper: DS MS.

## References

1. 1. Brookmeyer R, Johnson E, Ziegler-Graham K, Arrighi HM. Forecasting the global burden of Alzheimer’s disease. Alzheimers Dement 2007; 3: 186–191. pmid:19595937
2. 2. Brookmeyer R, Evans DA, Hebert L, Langa KM, Heeringa SG, Plassman BL, et al. National estimates of the prevalence of Alzheimer’s disease in the United States. Alzheimers Dement 2011; 7: 61–73. pmid:21255744
3. 3. Sosa-Ortiz AL, Acosta-Castillo I, Prince MJ. Epidemiology of Dementias and Alzheimer’s Disease. Arch Med Res 2012; 43: 600–608. pmid:23159715
4. 4. Hebert LE, Scherr PA, Beckett LA, Albert MS, Pilgrim DM, Chown MJ, et al. Age-specific incidence of Alzheimer’s disease in a community population. J Am Med Assoc 1995; 273: 1354–1359.
5. 5. Alzheimer’s Association. 2013 Alzheimer’s disease facts and figures. Alzheimers Dement 2013; 9: 208–245. pmid:23507120
6. 6. De Felice FG. Connecting Type 2 diabetes to Alzheimer’s disease. Expert Rev Neurother 2013; 13: 1297–1299. pmid:24236899
7. 7. Crane PK, Walker R, Hubbard RA, Li G, Nathan DM, Zheng H, et al. Glucose Levels and Risk of Dementia. N Engl J Med 2013; 369: 540–548. pmid:23924004
8. 8. Kivipelto M, Helkala E-L, Laakso MP, Hänninen T, Hallikainen M, Alhainen K, et al. Midlife vascular risk factors and Alzheimer’s disease in later life: longitudinal, population based study. BMJ 2001; 322: 1447–1451. pmid:11408299
9. 9. Exalto LG, Quesenberry CP, Barnes D, Kivipelto M, Biessels GJ, Whitmer RA. Midlife risk score for the prediction of dementia four decades later. Alzheimers Dement 2014; 10: 562–570. pmid:24035147
10. 10. Ballard C, O’Sullivan MJ. Alzheimer disease and stroke: Cognitive and neuroimaging predictors of AD and stroke. Nat Rev Neurol 2013; 9: 605–606. pmid:24145371
11. 11. Tolppanen A-M, Ngandu T, Kaareholt I, Laatikainen T, Rusanen M, Soininen H, et al. (2013) Midlife and Late-Life Body Mass Index and Late-Life Dementia: Results from a Prospective Population-Based Cohort. J Alzheimers Dis 2014; 38: 201–209. pmid:23948937
12. 12. Panza F, Capurso C, D’Introno A, Colacicco AM, Parigi AD, Gagliardi G, et al. Mediterranean diet, mild cognitive impairment, and Alzheimer’s disease. Exp Gerontol 2007; 42: 6–7. pmid:17092680
13. 13. Féart C, Samieri C, Rondeau V, Amieva H, Portet F, Dartigues JF, et al. Adherence to a Mediterranean diet, cognitive decline, and risk of dementia. J Am Med Assoc 2009; 302: 638–648.
14. 14. Scarmeas N, Luchsinger JA, Schupf N, Brickman AM, Cosentino S, Tang MX, et al. Physical activity, diet, and risk of Alzheimer disease. J Am Med Assoc 2009; 302: 627–637.
15. 15. Wengreen H, Munger RG, Cutler A, Quach A, Bowles A, Corcoran C, et al. Prospective study of Dietary Approaches to Stop Hypertension- and Mediterranean-style dietary patterns and age-related cognitive change: the Cache County Study on Memory, Health and Aging. Am J Clin Nutr 2013; 98: 1263–1271. pmid:24047922
16. 16. Zeki Al Hazzouri A, Vittinghoff E, Byers A, Covinsky K, Blazer D, Diem S, et al. Long-term Cumulative Depressive Symptom Burden and Risk of Cognitive Decline and Dementia Among Very Old Women. J Gerontol A Biol Sci Med Sci. 2014; May; 69(5):595–601. pmid:24097423
17. 17. Breunig JJ, Guillot-Sestier M-V, Town T. Brain injury, neuroinflammation and Alzheimer’s disease. Front Aging Neurosci 2013; 5:26. pmid:23874297
18. 18. Richardson JR, Roy A, Shalat SL, von Stein RT, Hossain MM, Buckley B, et al. Elevated Serum Pesticide Levels and Risk for Alzheimer Disease. JAMA Neurol. 2014 Mar; 71(3):284–90. pmid:24473795
19. 19. Fratiglioni L, Wang H-X. Smoking and Parkinson’s and Alzheimer’s disease: review of the epidemiological studies. Behav Brain Res 2000; 113: 117–120. pmid:10942038
20. 20. Moceri VM, Kukull WA, Emanuel I, van Belle G, Larson EB. Early-life risk factors and the development of Alzheimer’s disease. Neurology 2000; 54: 415–420. pmid:10668705
21. 21. Fischer C, Yeung E, Hansen T, Gibbons S, Fornazzari L, Ringer L, et al. Impact of socioeconomic status on the prevalence of dementia in an inner city memory disorders clinic. Int Psychogeriatr 2009; 21: 1096. pmid:19712540
22. 22. Verghese J, Lipton RB, Katz MJ, Hall CB, Derby CA, Kuslansky G, et al. Leisure activities and the risk of dementia in the elderly. N Engl J Med 2003; 348: 2508–2516. pmid:12815136
23. 23. Middleton LE, Barnes DE, Lui L-Y, Yaffe K. Physical Activity Over the Life Course and Its Association with Cognitive Performance and Impairment in Old Age: LIFE COURSE PHYSICAL ACTIVITY AND COGNITION. J Am Geriatr Soc 2010; 58: 1322–1326. pmid:20609030
24. 24. Barnes DE. The Mental Activity and eXercise (MAX) Trial: A Randomized Controlled Trial to Enhance Cognitive Function in Older Adults. JAMA Intern Med 2013; 173: 797. pmid:23545598
25. 25. Glymour MM, Tzourio C, Dufouil C. Is Cognitive Aging Predicted by One’s Own or One’s Parents’ Educational Level? Results From the Three-City Study. Am J Epidemiol 2012; 175: 750–759. pmid:22472116
26. 26. Rogers MAM, Plassman BL, Kabeto M, Fisher GG, McArdle JJ, Llewellyn DJ, et al. Parental Education and Late-life Dementia in the United States. J Geriatr Psychiatry Neurol 2008; 22: 71–80. pmid:19073840
27. 27. Mangialasche F, Kivipelto M, Solomon A, Fratiglioni L. Dementia prevention: current epidemiological evidence and future perspective. Alzheimers Res Ther 2012; 4: 1–8. pmid:22182734
28. 28. Daviglus ML, Bell CC, Berrettini W, Bowen PE, Connolly ES Jr, Cox NJ, et al. National Institutes of Health State-of-the-Science Conference statement: preventing alzheimer disease and cognitive decline. Ann Intern Med 2010; 153: 176–181. pmid:20547888
29. 29. Moceri VM, Kukull WA, Emanual I, van Belle G, Starr JR, Schellenberg GD, et al. Using census data and birth certificates to reconstruct the early-life socioeconomic environment and the relation to the development of Alzheimer’s disease. Epidemiology 2001; 12: 383–389. pmid:11416775
30. 30. Chin AL, Negash S, Hamilton R. Diversity and Disparity in Dementia: The Impact of Ethnoracial Differences in Alzheimer Disease. Alzheimer Dis Assoc Disord 2011; 25: 187–195. pmid:21399486
31. 31. Glymour MM, Kosheleva A, Wadley VG, Weiss C, Manly JJ. Geographic Distribution of Dementia Mortality: Elevated Mortality Rates for Black and White Americans by Place of Birth. Alzheimer Dis Assoc Disord 2011; 25: 196–202. pmid:21297428
32. 32. Kwok JB. Role of epigenetics in Alzheimer’s and Parkinson’s disease. Epigenomics 2010; 2: 671–682. pmid:22122050
33. 33. McKay JA, Mathers JC. Diet induced epigenetic changes and their implications for health: Nutrition, epigenetics and health. Acta Physiol 2011; 202: 103–118. pmid:21401888
34. 34. Alegría-Torres JA, Baccarelli A, Bollati V. Epigenetics and lifestyle. Epigenomics 2011; 3: 267–277. pmid:22122337
35. 35. Balazs R, Vernon J, Hardy J. Epigenetic mechanisms in Alzheimer’s disease: progress but much to do. Neurobiol Aging 2011; 32: 1181–1187. pmid:21669333
36. 36. Gray SG. Epigenetic treatment of neurological disease. Epigenomics 2011; 3: 431–450. pmid:22126204
37. 37. Kim M, Bae M, Na H, Yang M. Environmental Toxicants—Induced Epigenetic Alterations and Their Reversers. J Environ Sci Health Part C 2012; 30: 323–367. pmid:23167630
38. 38. Roseboom TJ, Van Der Meulen JH, Ravelli AC, Osmond C, Barker DJ, Bleker OP Effects of prenatal exposure to the Dutch famine on adult disease in later life: an overview. Mol Cell Endocrinol 2001; 185: 93–98. pmid:11738798
39. 39. Altmann S, Murani E, Schwerin M, Metges CC, Wimmers K, Ponsuksili S. Maternal dietary protein restriction and excess affects offspring gene expression and methylation of non-SMC subunits of condensin I in liver and skeletal muscle. Epigenetics 2012; 7: 239–252. pmid:22430800
40. 40. Rizzuto D, Orsini N, Qiu C, Wang H-X, Fratiglioni L. Lifestyle, social factors, and survival after age 75: population based study. BMJ 2012; 345: e5568–e5568. pmid:22936786
41. 41. Barnes DE, Yaffe K. The projected effect of risk factor reduction on Alzheimer’s disease prevalence. Lancet Neurol 2011; 10: 819–828. pmid:21775213
42. 42. Grant WB. Trends in diet and Alzheimer’s disease during the nutrition transition in Japan and developing countries. J Alzheimers Dis 2014; 38: 611–620. pmid:24037034