Alzheimer’s & Dementia: Diagnosis, Assessment & Disease Monitoring 11 (2019) 493-499 Diagnostic Assessment & Prognosis Short height and poor education increase the risk of dementia in Nigerian type 2 diabetic women Efosa Kenneth Oghagbona,b, Lydia Gimenez-Llortc,d,* a Faculty of Basic & Allied Medical Sciences, Department of Chemical Pathology, College of Health Sciences, Benue State University, Makurdi, Nigeria b Department of Chemical Pathology, Benue State University Teaching Hospital, Makurdi, Nigeria c Department of Psychiatry and Forensic Medicine, School of Medicine, Universitat Autonoma de Barcelona, Barcelona, Spain d Institut de Neurociencies, Universitat Autonoma de Barcelona, Barcelona, Spain Introduction: There is urgent need to investigate type 2 diabetes and dementia crosstalk in sub- Saharan African countries with special attention to women who have higher vulnerability. Nigeria which has the highest number of diabetics on the African continent is a good location for the inves- tigation. Methods: Biophysical parameters, occupation, education, burden of diabetes mellitus, cardiovascu- lar health, and cognition were evaluated in 102 type 2 diabetics and 99 controls. Results: Short physical stature and lower level of education were hallmarks of diabetes in females. Two dementia scales (Mini–Mental State Examination and six-item Cognitive Impairment Test) showed cognitive impairment status, with the six-item Cognitive Impairment Test scale being more specific and sensitive. Both scales showed correlations with age, education, weight, height, and disease onset, whereas fasten blood glucose was negatively correlated with height and their blood pressure was normal. Discussion: Height, an easy-to-measure parameter in Nigeria, may reveal increased risk of dementia in poorly educated female Nigerian diabetics, thus helping to improve preventive and therapeutic interventions. Ó 2019 The Authors. Published by Elsevier Inc. on behalf of the Alzheimer’s Association. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). 1. Introduction the prevalence of DM2 will increase astronomically in the next few years; with 75% of the cases living in low-to- Nigeria has the highest number of diabetics on the Afri- middle–income countries (LMICs), including Nigeria [1]. can continent [1], and this is due to rising mean age of the The World Alzheimer Report 2015 states that 58% of the population and prevalence of obesity, hypertension, and global 46.8 million dementia patients live in LMICs [9], widespread adoption of sedentary lifestyle. Diabetes melli- and this is expected to reach 63% in 12 years with a concur- tus type 2 (DM2) in midlife is a risk factor for dementia, rent huge socioeconomic cost [9,10]. This cost, which is and this association is stronger in persons of African reported to be enormous in developed economies, will be ancestry [2–4]. Occurrence of DM2 in midlife and other devastating and unsustainable for the health care system of factors listed previously are known to accelerate Nigeria; a country that is witnessing rapid rise in diabetes development of dementia in Africa Americans [2,5] prevalence [10,11]. Between the 1960s and recent time, compared with Caucasians [2–4,6–8]. It is worrisome that the prevalence of diabetes increased from less than 1% to about 10% in Nigeria [11]. Some investigations have shown that dementia is 1.5 The authors declare no competing financial interests. times more common in females and it is even worse in sub- *Corresponding author. Tel: 134-5812378; Fax: 134-935811435. Saharan Africa women with a two- to eight-fold increased E-mail address: lidia.gimenez@uab.cat risk [12]. A combination of aforementioned factors could https://doi.org/10.1016/j.dadm.2019.05.006 2352-8729/ Ó 2019 The Authors. Published by Elsevier Inc. on behalf of the Alzheimer’s Association. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). 494 E.K. Oghagbon and L. Gimenez-Llort / Alzheimer’s & Dementia: Diagnosis, Assessment & Disease Monitoring 11 (2019) 493-499 lead to a “dementia boom” in Nigeria and the subregion, if ters) were recorded in both patients with DM2 and con- they are not proactively managed. Hence, there is urgent trols. Their body mass index (BMI) was calculated by need to investigate the factors associated with the crosstalk weight (kg)/height (meters square), and waist-hip ratio between DM2 and dementia in persons of African ancestry (WHR) was determined from WC/HC, in both groups. with special attention to women who are more susceptible The burden of DM2 was measured as fasting blood to dementia. Such efforts will help in the development of glucose (FBG), duration and onset of disease. Systolic evidence-based management approaches in relevant clinics (SBP) and diastolic pressures (DBP) were measured for and communities of Nigeria and the subregion. the evaluation of the cardiovascular health of the sub- jects. Six-item Cognitive Impairment Test (6-CIT) and Mini– 2. Methods Mental State Examination (MMSE) scales were used for A total of 208 adult female and male subjects—109 (62 the comparative assessment of cognitive status of the sub- female, 47 male) with DM2 and 99 (53 female, 46 male) jects. Anonymity of the participants was guaranteed, and healthy controls—were recruited, at the Diabetic Clinic of the statistical analyses were carried out by a researcher blind the Benue State University Teaching Hospital, Makurdi, to the experiment. The study was approved by the Ethics Nigeria, and among Benue State University Teaching Hospi- Committee of the Benue State University Teaching Hospital, tal and Benue State University staff, respectively, from Makurdi, Nigeria. All the participants provided signed and October 2017 to November 2018. written informed consent. Subjects who were illiterate had Standard anthropometric measures of body weight (ki- the study plans interpreted in the local languages (Tiv, logram), height (centimeters), waist circumference (WC Idoma, Igede, Hausa) before their signed consent was taken in centimeters) and hip circumference (HC in centime- and included in the study. Fig. 1. Tag clouds of the occupational status of the four groups in the Nigerian sample population. E.K. Oghagbon and L. Gimenez-Llort / Alzheimer’s & Dementia: Diagnosis, Assessment & Disease Monitoring 11 (2019) 493-499 495 Statistical analyses were performed using SPP15.0. A ment and activities. The enhanced cognitive reserve helps two-factor (DM2 and sex) factorial analysis was used to the patient to compensate for greater degrees of neuropa- analyze the effect of the disease and sex interactions in the thology that may occur in later life; hence, education is iden- population. Two-way analysis of variance followed by Dun- tified as a proxy for such reserve in various investigations can’s post hoc or Student t-test was used to analyze differ- [13]. Education in early life appears to disconnect observed ences between the groups. Pearson’s correlation was used degree of neuropathology and cognitive impairment in later to identify relationship between demographic, anthropo- life. This disconnection is reiterated by findings in some in- metric, and disease variables with statistical significance dividuals able to tolerate greater disease burden without cor- considered at P , .05. responding evidence of clinical dementia [14]. Furthermore, in addition to age and complications of diabetes, education is 3. Results and discussion an integral component of a key predictive tool for a 10-year The recruitment of patients attending a diabetic clinic and likelihood of dementia [15]. controls from two adjacent establishments provided a sam- Age was negatively correlated with education in this ple population with different occupations and cadres study. The mean age of the subjects in the DM2 group was (Fig. 1). Around 45.2% of the study subjects were either civil higher than that of the controls, possibly due to recruitment servants or its retirees. The other subjects were from varied bias. The authors relate this finding to the higher number of vocations, including farming, clergy, cleaning services, retired subjects in the diabetic group. There was no signifi- technicians, and housewives. Among the DM2 group, the cant age difference between the sexes, although the age most common occupation was farming (females: 30.7%), was lower in the females. This can be interpreted to suggest followed by retirees (males: 36.2%). In contrast, civil service that the Nigerian females become diabetic at earlier ages. was the most common occupation of the control subjects (fe- Sex is a key factor for education achievement in this pop- males: 47.2%; males: 21.7%). Similar to previous reports in ulation as indicated previously. In our opinion, early child- hospital-based studies in Nigeria, patients with DM2 in this hood education or even adulthood education activities, study were 5-6 years older than their healthy controls especially for females, may be a preventive policy to adapt (P 5 .000). There was no age difference between sexes in for mitigation of later life cognitive impairment in this pop- both diabetic and control populations (Fig. 2A). ulation. The factors of age and education correlated with The distribution of level of education in the sample pop- onset of disease, systolic pressure, and cognitive impairment ulation (Fig. 2C) did not show a standard normal probability when measured by both dementia scales used in the study. distribution. When the study groups were combined, ana- Also, the factors are related to anthropometric measures of lyses showed that more than 50% of the subjects obtained WC and WHR, parameters which clinical guidelines advise a tertiary level of education (females-DM2: 53.2%, males- should be measured in diabetic clinics. Similarly, education DM2: 66.0%, females-controls: 62.3%, and males- was related to body weight and height in the whole study controls: 58.7%). However, the distribution of participants population. with low levels of education differed between sexes. The Anthropometric measures showed a strong relationship sex of the subjects were related to their level of education with diabetes (Fig. 2D–G). The BMI (P 5 .004)—but not attainment, with diabetic women being least educated body weight—WC (P 5 .000), and WHR (P 5 .000) were compared with the other groups (Fig. 2B,C). Illiteracy, significantly associated with FBG in DM2, but not in the which means the absence of any formal education, was healthy controls (see Supplementary Table 1). Body height more common in the female patients with DM2. The males, was the most specific and distinguishing variable of the irrespective of disease status, had a higher level of education different groups, especially by DM2 (P 5 .000) and sex than the females (P 5.030). In contrast, a statistically signif- (P 5 .000) status. The diabetic female group had the shortest icant higher percentage of female subjects were illiterate height or stature. Reduced height in females was associated compared with males. Illiteracy among diabetic females (fe- with DM2 in this Nigerian population. BMI (P 5 .001) and males-DM2: 22.6%) was increased two-fold compared with HC (P 5.002) also showed differences between sexes. Height their control counterpart (females-controls: 11.3%) but was was strongly related to education, body weight, and BMI, but 11-fold higher than diabetic males (males-DM2: 2.2%) and not to WC, HC, or WHR. Short height or stature in the Niger- controls (males-controls: 2.1%). Thus, the diabetic females ian females may be due to exposure to worse early-life stress had the lowest level of education, whether at a primary or and poor nutrition, compared with the males. This is because, secondary level. Therefore, sex may be a key factor for ed- besides genetics, the above discussed risk factors are major ucation achievement in Nigeria. The association of educa- contributors to adult height and other biophysical indices tional attainment and dementia is long established, even in [16]. Short stature and patient sex may be dependent on the places where there is lower access to education [13]. It is environment, as diabetic adult males were reported to be known that exposure to education increases the cognitive shorter in Israel [16], unlike what is observed in the females reserve of individuals. This is achieved through neuroplas- of this Nigerian population. The relevance of early-life nutri- ticity and development of functional and complex neural tion interventions is supported by evidence that it increases networks that come with a higher level of education attain- the brain reserves and protect against later-life 496 E.K. Oghagbon and L. Gimenez-Llort / Alzheimer’s & Dementia: Diagnosis, Assessment & Disease Monitoring 11 (2019) 493-499 Fig. 2. Age, education, anthropometric measurements, and burden of diseases in the Nigerian sample population. (A) Age of the sample population. Education: (B) mean education level and (C) percentage of subjects at each level of education, from nil or no education to tertiary level education. Anthropometric mea- surements: (D) body weight and BMI, (E) waist circumference and hip circumferences, (F) WHR, and (G) height. Burden of DM2: (H) fasting blood glucose, (I) E.K. Oghagbon and L. Gimenez-Llort / Alzheimer’s & Dementia: Diagnosis, Assessment & Disease Monitoring 11 (2019) 493-499 497 neurodegeneration, through decrease in cardiometabolic risk MMSE (age, r 5 2.388, P , .01; education, r 5 .743, factors [17]. This is consistent with the Barker hypothesis, P ,.01) scales. Age was also related to anthropometric mea- which says that optimal prenatal and early-life environment sures of WC (r 5 .206, P , .01) and WHR (r 5 .271, increases the risk of adult disease [18]. The sub-Saharan Af- P , .01), and these are parameters expected to be measured rica region has a high prevalence of low–birth weight babies. in diabetic clinics. Education was related to body weight Low–birth weight babies have been shown to have poorer (r 5 .197, P , .01) and height (r 5 .258, P , .01) in the cognitive development, which may lead to even poor cogni- whole study population. The FBG only correlated with sys- tive function in adulthood [19]. The present study is unable tolic (r 5 2.162, P , .05) and diastolic (r 5 2.170, P , .05) to confirm if the prevalence of low birth weight is more com- measures for cardiovascular health. Onset of disease corre- mon in the female population of this study, or in the general lated with age (r 5 .820, P , .01), education (r 5 2.293, population. However, a recent Nigerian study suggests that P , .01), systolic pressure (r 5 .219, P , .05), and cognitive an equal number of male and female children are born with impairment as measured by 6-CIT (r 5 .267, P , .01) and low birth weight [20]. Thus, the observed sex differences in MMSE (r 5 2.297, P , .01). height could be environmentally determined. The association of height with education attainment and The burden of DM2 (Fig. 2H,I) measured as FBG levels the dementia scales in this population may be related to (FBG, P 5 .000 vs. controls) and prevalence of disease (fe- early-life circumstances, which were discussed previously. males-DM2: 21.3%; males-DM2: 23.4%) were similar in It is known that optimal nutrition in early infant life can both sexes. However, duration of disease (P 5 .000) was improve cognitive and brain development throughout child- longer in females than males, thus supporting earlier position hood and adolescence; periods that have significant implica- that disease occurs earlier in women. On the average, DM2 tion for later-life biophysical and educational attainment occurred 4 years earlier in the females (P 5 .05, one-tailed). [21]. The authors are of the opinion that if Nigerian public The two scales administered to assess dementia (Fig. 2L– policy targets improvements of early-life nutrition and edu- O) showed a strong effect of sex (6-CIT, P 5 .005; MMSE, cation programs, with emphasis on the girl child, the burden P 5 .004) with diabetic females showing poorer cognitive of dementia in adult diabetic females may be mitigated. performances than their male counterparts. A lower percent- In conclusion, short body height or stature and lower ed- age of diabetic females were categorized as normal cogni- ucation levels are hallmarks of diabetes in females in this Ni- tion compared with the others (Fig. 2M,O). Conversely, gerian population. Most importantly, these factors are more diabetic females were found in the moderate to signif- associated with poorer cognitive performances measured icant cognitive impairment category using the 6-CIT and by 6-CIT and MMSE scales. The distinct sex/gender factor MMSE scales. The 6-CIT scale was more discriminative reported in this Nigerian DM2 population brings attention than MMSE as it showed a disease and sex impact on mea- to the impact of short stature and poor education on the surement of cognition in this population. The MMSE only mental health of diabetic patients in Nigeria, especially in showed sex impact with the mildly impaired diabetic fe- the females. We here propose that height, an easy-to- males in contrast to the diabetic males and controls found measure parameter in most clinical settings in Nigeria, to be cognitively normal. Correlation analysis is shown in may reveal increased risk of dementia among poorly Supplementary Table 1. The two dementia scales showed educated Nigerian diabetic females. When such simple mea- correlations (n 5 208) with age (6-CIT, r 5 .347, P , .01; surement is carried out routinely in Nigeria, it may help in- MMSE, r 5 2.388, P , .01), education (6-CIT, crease detection of cognitive impairment in susceptible r 5 2.750, P , .01; MMSE, r 5 .743, P , .01), weight patients and doing so improves preventive and therapeutic (6-CIT, r 5 2.230, P , .01; MMSE, r 5 2.278, P , .01), interventions in the country. Public policy targeting im- height (6-CIT, r 5 2.262, P , .01; MMSE, r 5 .267, provements of early-life nutrition and education programs, P , .01), and disease onset (6-CIT, r 5 .267, P , .01; with emphasis on the girl child, may mitigate the burden MMSE, r 5 2.297, P , .01). FBG showed a negative corre- of dementia in adult diabetic females. lation with height (r 5 2.104). The factors age and educa- tion correlated with onset of disease (age, r 5 .820, Acknowledgments P , .01; education, r 5 2.293, P , .01), systolic pressure This work was supported by personal fundings and FAS, (age, r 5 .306, P , .01; education, r 5 2.180, P , .01), Fundacio Autonoma Solidaria, Fons de Solidaritat de la and cognitive impairment measured by both 6-CIT (age, UAB, UAB Solidarity Fund, FSXXXIV-03, University r 5 .347, P , .01; education, r 5 2.750, P , .01) and cooperation projects for development of countries of the =onset of disease, and (J) duration of disease. Cardiovascular health: (K) systolic and diastolic pressures. Mental health: (L) mean 6-CIT scores, (M) percentage of subjects at each severity level of 6-CIT scale, (N) mean MMSE scores, and (O) percentage of subjects at each severity level of MMSE scale. Data are mean 6 SEM or percentages. Factorial analysis: D, DM2; S, sex; DxS, DM2 x sex interaction effects; Post hoc: *P , .05, **P , .01, ***P , .001 versus all the other groups; dP , .05, ddP,.01, dddP,.001 versus control group of the same sex; sP , .05, ssP , .01, sssP , .001 versus the opposite sex of the same group. #P ,.05, ##P ,.01, ###P ,.001 versus male control group. (*) One-tailed Student t-test, P ,.05. Abbreviations: 6-CIT, six-item Cognitive Impair- ment Test; BMI, body mass index; CI, cognitive impairment; DM2, type 2 diabetes mellitus; MMSE, Mini–Mental State Examination; WHR, waist-hip ratio. 498 E.K. Oghagbon and L. Gimenez-Llort / Alzheimer’s & Dementia: Diagnosis, Assessment & Disease Monitoring 11 (2019) 493-499 South: Prevalence of Cognitive Impairment in Nigerian Sub- References jects with Diabetes Mellitus Type 2 and Associated Comor- bidities to L.G.-L, with E.K.O as counterpart in Nigeria. The [1] International Diabetes Federation. IDF Diabetes Atlas. 7th ed.; 2015. authors sincerely thank Dr. Faeren Dogoh for his consistent www.idf.org. and skillful training and guidance of research assistants in [2] Wessels AM, Lane KA, Gao S, Hall KS, Unverzagt FW, the collection of data from the subjects at the Diabetic Clinic Hendrie HC. Diabetes and cognitive decline in elderly African of Benue State University Teaching Hospital, Makurdi, Americans: a 15 year follow up study. Alzheimers Dement 2011; 7:418–24. Nigeria. Also, the authors thank Professor Moses Msugh [3] Biessels GJ, Staekenborg S, Brunner E, Brayne C, Scheltens P. Risk of Kembe, Vice-Chancellor, Benue State University, Makurdi, dementia in diabetes mellitus: a systematic review. Lancet Neurol Nigeria, for institutional support for the project while it 2006;5:64–74. lasted. The authors are also grateful to the management [4] Tuligenga RH, Dugravot A, Tabak AG, Elbaz A, Brunner EJ, and staff of the Benue State University Teaching Hospital, Kivimaki M, et al. Midlife type 2 diabetes and poor glycaemic control as risk factors for cognitive decline in early old age: a post-hoc analysis Makurdi, Nigeria, for their support in subjects’ recruitment of the Whitehall II cohort study. Lancet Diabetes Endocrinol 2014; and blood analyses during the study. 2:228–35. Author’s contributions: E.K.O. and L.G.-L were responsible [5] Fan Y-C, Hsu J-L, Tung H-T, Chou C-C, Bai C-H. Increased dementia for the study design and data interpretation and wrote the risk predominantly in diabetes mellitus rather than in hypertension or manuscript. E.K.O. was responsible of data acquisition and hyperlipidemia: a population-based cohort study. Alzheimers Res Ther 2017;9:7. cognitive measures. L.G.-L. was responsible of data analysis [6] Potter GG, Plassman BL, Burke JR, Kabeto MU, Langa KM, and illustrations. Both authors approved the final version. Llewellyn DJ, et al. Cognitive performance and informant re- ports in the diagnosis of cognitive impairment and dementia in African Americans and whites. Alzheimers Dement 2009; Supplementary Data 5:445–53. Supplementary data related to this article can be found at [7] Rajan KB, Arvanitakis Z, Lynch EB, McAninch EA, Wilson RS, Weuve J, et al. Cognitive decline following incident and preexist- https://doi.org/10.1016/j.dadm.2019.05.006. ing diabetes mellitus in a population sample. Neurology 2016; 87:1681–7. [8] Matioli MNPDS, Suemoto CK, Rodriguez RD, Farias DS, da Silva MM, Leite REP, et al. Association between diabetes and causes RESEARCH IN CONTEXT of dementia: Evidence from a clinicopathological study. Dement Neu- ropsychol 2017;11:406–12. [9] Hendrie HC, Zheng M, Lane KA, Ambuehl R, Purnell C, Li S, et al. Changes of glucose levels precede dementia in African-Americans 1. Systematic review: Literature review, using PubMed with diabetes but not in Caucasians. Alzheimers Dement 2018; and International Diabetes Federation site, focused 14:1572–9. on crosstalk between diabetes and dementia and so- [10] Wang KC, Woung LC, Tsai MT, Liu CC, Su YH, Li CY. Risk of Alz- ciodemographic and biophysical risk factors. The heimer’s disease in relation to diabetes: a population-based cohort search indicated dramatic rising of diabetes and study. Neuroepidemiology 2012;38:237–44. [11] Ogbera AO, Ekpebegh C. Diabetes mellitus in Nigeria: The past, pre- risk of cognitive decline/dementia in sub-Saharan sent and future. World J Diabetes 2014;5:905–11. African low-to-middle–income countries, with [12] Ronquillo JG, Baer MR, Lester WT. Sex-specific patterns and differ- Nigeria and women being in the eye of storm. The ences in dementia and Alzheimer’s disease using informatics ap- relevant citations are appropriately cited. proaches. J Women Aging 2015;28:403–11. [13] Paddick S-M, Longdon A, Gray WK, Dotchin C, Kisoli A, Chaote P, 2. Interpretation: The distinct sex/gender scenario in et al. The association between educational level and dementia. Dement the rising diabetes mellitus type 2 population in Neuropsychol 2014;8:117–25. Nigeria warns about the impact of present findings [14] Murray AD, Staff RT, McNeil CJ, Salarirad S, Ahearn TS, Mustafa N, et al. The balance between cognitive reserve and brain imaging bio- on the hallmarks short stature and poor education markers of cerebrovascular and Alzheimer’s diseases. Brain 2011; on the mental health of women with diabetes mellitus 134:3687–96. type 2 in Nigeria. [15] Exalto LG, Biessels GJ, Karter AJ, Huang ES, Kayton WJ, Minkoff JR, et al. Risk score for prediction of 10-year dementia risk 3. Future directions: We propose that height, an easy- in individuals with type 2 diabetes: a cohort study. Lancet Diabetes En- to-measure parameter in most clinical settings in docrinol 2013;1:183–90. Nigeria, could unveil increased risk factor for de- [16] West RK, Ravona-Springer R, Heyman A, Schmeidler J, Leroith D, Kofman K, et al. Shorter adult height is associated with poorer cogni- mentia among poorly educated Nigerian diabetic tive performance in elderly men with type 2 diabetes. J Alzheimers Dis females. Carried out routinely in Nigeria, it will help 2015;44:927–35. to increase detection of cognitive impairment in [17] Prince MJ, Acosta D, Guerra M, Huang Y, Jimenez-Velazquez IZ, susceptible patients and improve preventive and Rodriguez JJL, et al. Leg length, skull circumference and the incidence therapeutic interventions in the country. of dementia in Latin America and China: A 10/66 Population-based study. PLoS One 2018;13:e0195133. [18] de Boo HA, Harding JE. The developmental origins of adult disease (Barker) hypothesis. Aust NZ J Obstet Gynaecol 2006;46:4–14. E.K. Oghagbon and L. Gimenez-Llort / Alzheimer’s & Dementia: Diagnosis, Assessment & Disease Monitoring 11 (2019) 493-499 499 [19] Straus RS. Adult functional outcome of those born small for gesta- Morbidity and Mortality in a Tertiary Healthcare Facility in the South tional age: twenty-six year follow-up of the 1970 British Birth Cohort. Eastern Nigeria. Ann Med Health Sci Res 2018;8:4–10. JAMA 2000;283:625–32. [21] Isaac EB, Gadian DG, Sabatini S, Chong WK, Quinn BT, Fischl BR, [20] Chidiebere ODI, Uchenna E, Ikenna NK, Christian I, Nwabueze AI, et al. The effect of early human diet on caudate volumes and IQ. Pe- Ogechukwu AF, et al. The Low-birth weight Infants: Pattern of diatr Res 2008;63:308–14.