Patients were subjected to their cognitive functions evaluation through Mini Mental State Examination (MMSE) in its modified version which is adapted to the research cultural era. Patients who recorded MMSE ≤ 24/30 are likely to have dementia and subjected to DSM IV criteria [7] for dementia diagnosis. Its severity is appreciated in conformity with MMSE score. In fact,

- a score 20/30 ≤ MMSE ≤ 24/30 with a patient matching DMS IV criteria, is considered as “moderate dementia”.

- a score MMSE ≤ à 19/30 with a patient matching DMS IV criteria, is considered as “severe dementia”.

HACHINSKI score was used to determine the nature of the effect:

- >7: vascular dementia.

- Between 4 and 7: mixed dementia.

- <7 degenerative dementia.

Dependent variable: dementia.

Independent variables: they were:

Ÿ Socio-demographic: age, gender, ethnicity, profession, residence, marital status, number of children, education level.

Ÿ Clinical: Hypertension, diabetes, obesity, stroke, cardiomyopathy, sinusitis, epilepsy hyperlipidemia.

Ÿ Types of dementia: curable dementia, vascular dementia, degenerative dementia, mixed dementia.

Data processing was conducted through EPI-DATA. Data analysis was done using statistics software STATA/ IC 11.0. A descriptive analysis was completed with regard to the variables which were studied. So, for qualitative variables, frequencies and proportions were determined. Either chi 2 or FISHER test was used if only expected values are lower than 5. For quantitative values, averages together with their typical gaps, medians, minima and maxima have all been described. STUDENT test was utilized for comparisons purpose. The study of associated factors was conducted using logistical regression model in unvaried and multi-varied analyses. The multi-varied analysis was conducted by inserting into the model all variables, of which p value in unvaried analysis is less or equal to 20%, due to the exploratory nature of the study. The break-even point in terms of significance was 5% and confidence gap rated at 95%.

Each patient or his/her next of kin submitted a written notice of consent upon explanation of the research objectives and modalities. Moreover, patients’ right as well as confidentiality was highly respected.

In our research dementia prevalence was 8.8%. This result is close to 8.1% obtained by Mbelesso et al. [8] in a cross-sectional research conducted in the 3^rd district of Bangui (Central African Republic) in 2012, on dementia epidemiology in elderly population. Similarly, Molero et al. [9] recorded in 2007 a prevalence of 10.3% in a re-

^*Significant result.

search conducted on elderly population living on the Caribbean coast of Venezuela. In 2012, Coume M. et al. [7] recorded 10.8% as prevalence rate in a cross-sectional research on the estimation of cognitive prevalence within an elderly Senegalese population. Our record of 8.8% dementia prevalence is lower than 24.3% registered by Uwakwe R. et al. [10] in a cross-sectional research conducted in 2009 on dependency epidemiology in elderly Nigerian population. It is also lower than 43.2% recorded by Ndiaye et al. [11] in a research conducted from January 2004 to June 2005 on the operation of a Senegalese memory clinic in Fann teaching hospital psychiatry department (Dakar, Senegal). In fact, many reasons may justify this difference in terms of the prevalence rate. First of all, in Uwakwe R. et al. [10] research, patients were aged 65 and above and the sample was made up of 1238 patients who went through 10/66 protocol; knowing that the specificity and sensitivity of such a protocol are higher than those of MMSE we implemented. The same goes with Ndiaye et al. [11] who utilized in their research Senegal Test, which sensitivity and specificity are higher than those of MMSE. Moreover, in their re-

^*Significant result.

search, enrolled patients (132 in total) were selected among those who showed memory disorders. All this could explain the higher rate recorded in these two researches. Notwithstanding our 8.8% prevalence is higher than the figure recorded by other authors. Thus, Paraïso et al. [6] recorded 2.6% in rural area (Djidja) and 3.7% in urban area (Cotonou) in a door-to-door research conducted in Benin in 2010 on dementia epidemiology with patients aged 65 and above, whilst Guerchet et al. [12] recorded 6.7% in a cross-sectional research conducted in the year 2012 on dementia epidemiology within elderly population living in Brazzaville (Congo). Longdon et al. [13] registered 6.4% prevalence in a cross-sectional community research on dementia prevalence in rural Tanzania. El Tallawy et al. [14] also registered 2.3% in a research conducted in 2011 on dementia prevalence in Kharga (Egypte) in 2012, whilst Yusuf et al. [15] estimated dementia prevalence at 2.8% in a community research on dementia prevalence and subtypes dementia within elderly population living in northern Nigeria. These prevalence rates which are very lower than what we recorded in our research could be explained by the fact that most of these researches were conducted within a vast population with samples above 1000 patients (1198, 8173, and 1139 patients were respectively enrolled by Longdon, El Tallawy and Paraïso). In addition, the test utilized in these researches is CSID, which has a weaker specificity than MMSE.

In our research dementia patients average age was 66 years ± 9.1. This result is close to 67 years ±17 observed by Ndiaye et al. [11] as well as 67 years ± 7.5% recorded by Coume et al. [7] . The same result is lower to 76 years ± 7.1 recorded by Guerchet et al. [12] . This result is explained by the fact that the patients enrolled in the research were 65 years and above against 50 years and above in our research. It comes out that age is the sole risk factor which was corroborated. In our research, 62% of dementia patients were aged beyond 60 years. Moreover, dementia prevalence increased with age. The figure varied from 5.3% below 60 years to 12.7% beyond 60 years. El Tallawy [14] recorded that dementia prevalence of 2.3% increased drastically to 18.5% with subjects aged beyond 80 years. Libre et al. [16] confirmed that beyond 65 years, dementia prevalence doubles anytime age increases by 5 years. Other authors such as Longdon, Coume, Yusuf et Paraïso [6] [7] [13] [15] recorded similar results. Out of 22 dementia patients 11 were males; meaning a sex ratio of 1. This result is dissimilar to the ratio recorded by most researches which revealed a female predominance. Thus, Mbeleso et al. [8] recorded 82.35% female predominance whilst Stewart et al. [17] recorded 60.8%. Letenneur et al. [18] as well as Nitrini et al. [19] corroborated this female predominance which according to Stewart et al. [17] might be explained by the biological and hormonal differences between the two sexes, particularly estrogens. Our result which is quite different from others’ could be explained by the fact that the population was male biased and not large enough. Indeed 18 (81.8%) of dementia patients were married. This result is similar to that of Ndiaye et al. (76.5) and Coume et al. (79%). However, it is different from most authors’ findings. Thus, Fratiglioni [20] maintains that being a widower or a divorcee implies an under-performance regarding MMSE; such under-performance is not recorded with elderly married persons. We could then bring up the hypothesis of marriage playing a role of social balance. Van Gelder et al. [21] in a research conducted in 2006 in Finland noted that men who lost their partner, those who weren’t married and those who live single life or lived single life for at least 5 years were much more exposed to cognitive decline, unlike married men or men who lived with a partner. This predominance of married persons has its explanation in our socio-cultural realities. In fact, it is rare to see persons aged 50 and above who are not married. Out of the 22 dementia patients, 21 (95.5%) lived in a family. Fratiglioni et al. [20] proved that single patients were twice exposed to dementia comparing to those living in a family. They also found out that a closed social network increases by 60% the risk. Dementia risk is therefore correlated with the significance of social network. Baiyewu et al. [3] exposed that in Africa elderly persons live in extended family made up of several generations. 6.7% live alone; which contrast the prevalence rate in Europe (35% - 39%). Our results reflect the importance of social network in Africa particularly in Benin. Such a rich social network may explain the low dementia prevalence in our societies. Unfortunately, this social network tends to ebb with a progressive urbanized lifestyle leading to individualism. However, we should not forget that these maladies are under-diagnosed. 5 (22.7%) of the dementia patients were uneducated. Among the 17 (77.3%) who were educated, 6 (27.3%) had primary education only. In Egypt, El Tallawy et al. [14] conducted a research which revealed that dementia prevalence rate was significantly higher among uneducated persons (6.35%) than literate persons (0.6%). They asserted that the low education level played a predominant role in dementia occurrence in elderly population especially when they did not reach elementary school. This assertion is supported by Guerchet et al. who demonstrated that, not acquiring elementary education was significantly associated with dementia. On the contrary, a research conducted in Kenya revealed that there was no proof of little education being associated with dementia occurrence [22] .

Results from our research revealed that degenerative dementia cases (50%) were mostly prevalent, followed by vascular dementia (31.8%) then curable dementia (27.3%) and finally mixed dementia (18.2%). This result is similar to that of Shelley et al. [23] who recorded 57.1% and 26.9% respectively for degenerative and vascular dementia. Alzheimer disease (AD) is the commonest neurodegenerative dementia; it is cortically predominant and represents dementia major cause in elderly population. AD is fundamentally caused by two types of cerebral lesions: senile plaques and fibrillary neurodegeneration. Senile plaques are extracellular deposits of amyloidal substance, mostly made up of peptide polymer Aβ (or β-amyloidal). This substance settles down progressively in the brain predominantly in the gray substance. These insoluble deposits are not metabolized by the body and destroy progressively adjacent nervous fibers. Parallel to this, in the fibrillary degeneration process, pathologic filaments made of protein build up around the cell body of the neurons. These neurons depolymerize then aggregate in the cytoplasm. Generally, this degeneration begins from the hippocampus and spreads across temporal regions. Other researches also revealed a high prevalence of AD. Thus, Yusuf et al. [15] recorded 66.7% AD while Alladi et al. [24] and Shelley et al. [23] respectively recorded 38.39% and 52.6%. Vascular dementia (VD) was the second etiology of dementia syndromes after degenerative dementia of Alzheimer type. Vascular dementia encompasses the whole dementia states secondary to cerebrovascular lesions. El Tallawy et al. [14] recorded 28.7% of vascular dementia cases. They are caused by the consequences of cerebrovascular affection (blockage of vessels or cerebral hemorrhage). In general they worsen discretely (during a new stroke occurrence). Neuro-imaging plays a predominant role in diagnosing by showing lesions of vascular origin. Curable dementia cases are those of which the etiology was identified. The most frequent causes are neurosurgical: cerebral tumors, adult hydrocephaly, and subdural hematoma. These causes are easily revealed by the neuro-imaging. The systematic results of research about infectious dementia (mainly syphilis, HIV-Aids or hardly ever Whipple disease), endocrine or deficient dementia are rarely productive because these types of dementia can totally be reversible. The discovery of curable cause for dementia does not always guarantee a complete recovery from the disease. However, systematic results from the research allow finding and treating reversible concomitant affections and disorders which can contribute to worsening dementia.

In our research two factors were revealed to be associated with dementia. There were stroke past records and smoking. 63.6% had stroke. This result is similar to the record of the University of Oxford in the United Kingdom [25] , following a research conducted by the department of research on diseases prevention with stroke inpatients. The prevalence of dementia with these patients was comparable to the rate of prevalence recorded with stroke free patients who were 10 years older. Only one stroke occurrence provokes no diffuse decline of cognitive functions leading to dementia. It is the repetitive occurrence of such incidents with the same patient which can provoke cognitive decline. We then refer to multiple infarction dementia or multi-infarction dementia. In a research conducted by Altieri et al. [26] , about 30% of stroke survivors suffered from dementia thereafter. The risk is high after stroke occurrence, but it is higher even after the embolic thrombosis occurrence. Post stroke dementia risk factors are not well known. However, three theoretical reasons could explain post stroke dementia occurrence. First of all, dementia could be the direct consequence of cerebral lesions of vascular origin. Then it could be due to the combination of degenerative lesions of Alzheimer type in a pre-clinical stage; some post stroke dementia states are progressive. Finally, white substance anomalies could also contribute to the decline in the sense that, they are associated with a high risk of vascular relapses and could stimulate neuropsychological disorders. Vascular lesions, white substance and lesions of Alzheimer type anomalies all put together, could stimulate post stroke dementia occurrence [27] . 18% were addicted to smoking. This result is similar to Juan et al. [28] findings, as they maintained that smoking revealed to be a risk factor. Likewise, Rusanen et al. [29] in 2011 established that smoking increases Alzheimer disease risk occurrence by 2. These Finnish scientists conducted a broad research between 1978 and 2008. In this research, 21,123 patients aged 50 to 60 years were thoroughly followed during 23 years. They diagnosed 5367 senile dementia cases of which 1136 Alzheimer diseases. Having reviewed different risk factors associated with dementia, researchers discovered that smoking played an important role. In fact, dementia risk increased by hundred percent (100%) concerning heavy smokers consuming over two packets of cigarette a day. Globally, smokers presented a risk of vascular dementia multiplied by 2.72. The same researchers showed that smoking affected the brain directly and increased both oxidative stress and inflammation which were two risk factors related to Alzheimer disease [29] . Nonetheless, the effect of smoking in dementia occurrence is not well explained. We only know that smoking has cerebrovascular effects and amplifies the cholinergic metabolism by dysfunctioning brain nicotine receptacles. A Chinese research [30] proved that second-hand smoking constitutes a risk factor which could increase by 29% the occurrence of severe dementia.