Nutritional status , survival and mortality in Alzheimer patients-a cross-sectional study

Dementia is a common health problem in elderly people, Introduction Alzheimer disease (AD) being the most prevalent. AD can be considered as a cause of death and must be registered on the death certificate of the patients. However, most of the time, the main cause of death registered is not related to AD, but as an underlying or contributing cause. For example, individuals who have AD and die from myocardium infarction. This study aimed to analyze if nutritional status was associated with survival and mortality for AD, and if AD was reported as actual cause of death on the death certificate : The study was carried out as a cross-sectional study with elderly Methods citizens of the community registered in the National Health System (SUS), with cognitive, nutritional, biochemical and hematological evaluations of 30 AD patients in Guarapuava, Paraná state, Brazil. : Significant differences were not observed between live and dead Results patients when evaluated considering the methods applied. Only 22% of the death certificates stated death due to AD. The patient’s cause of death showed a strong relation to respiratory issues; potential explanations based on immunological, biochemical and comorbidity were not confirmed on this study. : AD was not declared as the cause of death in the majority of Conclusions certificates, contributing to the underreporting and reducing the information of death due to AD in the country.


Introduction
The dementias are classified among the most frequent health problems in elderly people 1 . Alzheimer disease (AD) stands out as the most prevalent 2 , occurring in more than 50% of the cases 3,4 .
Alzheimer disease facilitates the advent of opportunistic diseases, which combined to the fragility of the person, deglutition disturbances, and malnutrition can increase the death risks 1 . People with AD live on average 8.3 years from diagnosis among 65 years old, and 3.4 with diagnosis after 90 years 4,5 .
According to the World Health Organization (WHO), AD can be considered as a cause of death and must be registered on the death certificate of the patients 5,6 . Nevertheless, in the majority of cases, the main death cause is not related to AD, but other issues are usually listed as immediate causes. According to Ministry of Health the immediate cause of death is the disease, injury or last complication that occurred immediately prior to the moment of death, having a direct correlation with the basic cause of death 7 .
The lack of national data still constitutes a gap in the knowledge of the mortality due to AD, although it has a known impact on the patient, family and caregiver.
It is known that at the present time there is no medical intervention capable of preventing or curing AD, but a protective effect can come from a good diet and quality of life, therefore, nutrition is shown to play an essential role. In this context, the present study aimed to investigate whether nutritional status was associated with AD survival and mortality, and if AD was recorded as the underlying cause of death in the death certificate.

Method
The present study was conducted in the city of Guarapuava, central-west region of Paraná state, Brazil. We performed a cross sectional study with elderly citizens of the community registered in the National Health System (SUS), through which they receive, at no charge, specific drugs for AD 8 .
The patients of this study had their diagnosis of AD performed by SUS doctors, according to the "Neurological and Communicative Disorders and Cerebral Vascular Accident National Institute and the Association of Alzheimer Disease and Related Disturbs (NINCDS -ADRDA)" 9 .

Recruitment
Data were accessed in CELEPAR System (Information and Communication Technology Company of Paraná) after research project approval. After the identification (name and address) of the 66 patients with AD, provided by a computerized system (CELEPAR System), visits were made to patients' homes. Patients / caregivers were invited to take part in the study, and a written informed consent form was signed by a person responsible/patient caregiver who agreed voluntarily to participate in the study.

Data collection
Data collection was carried out in two phases: the first between August and October 2011, for the anthropometric, nutritional measures, and hematologic exams. "Exclusions from the study were as follows: 7 patient due to death, 11 patient because they had moved from the city, 2 were not found at the address registered in the system, and 16 the caregivers did not accept to participate due to the patient status of weakness". This left 30 patients remaining in the study.
Research was done in two phases, between August 2011 and June 2013. The first stage was based on collecting nutritional and anthropometric measurements, as well as hematological tests. One and a half years later (June 2013), the researchers contacted the patients or caregivers, verifying that 30% (n=9) of the patients died since the first stage, in 2011. Death certificates (Dc) were evaluated in accordance with International Disease Classification (CID-10) 10 . It was defined as death due to AD in cases which had AD as the main cause of the death on the death certificate. Death with AD was recorded when it was mentioned in any part of the medical report of the death certificate. It was considered as codes CID G30.0 to G30.9 10 .
With the intention of evaluating the anthropometrical measures of the patients, we performed a body mass index (BMI). The weight and the height were collected according to the methods advocated in the feed and nutritional surveillance system (SISVAN) 10 . BMI measurements were taken using a digital precision scale with kilograms scale from Plenna®. To obtain the height values, a stadiometer was used with a centimeter scale.
When it was not possible to obtain the weight and height, as some patients could not stay in the orthostatic position since the patients were bedridden, estimated values were used from the formula proposed for Chumlea (1985) 11 . BMI was calculated using the correlation between the square of total corporal weight (kilograms) and the height (meters), using the cut points for elderly people proposed for Lipschitz (1994) 12 .
The nutritional state of the participants was evaluated using the mini nutritional assessment (MAN) (13). The MAN is an instrument composed of measures and practical questions that comprise anthropometric evaluations (weight, arm and calf circumferences, height and loss of weight record), global evaluation (life style, drugs, mobility and diseases), dietary evaluation (qualitative and quantitative) and self-evaluation (nutrition perception). The sum of MAN scores allows discrimination between patients, where values below 17 characterize the patient as undernourished, between 17 and 23.5 at malnutrition risk and between 23.5 and 30 as normal 13 . The MAN was applied to the patients classified with Clinical Dementia Rating (CDR) 2 or 3. The food consumption was evaluated through the patient or caregiver report of consumed food through memory of 24 hour estimated in home measurements converted to grams 14 . This data was analysed on Avanutri 4.0® software. Each nutrient was compared to the Dietary Reference Intake (DRI's) 14,15 according to gender and age, due to the lack of specific recommendations for AD patients.
Also, for the staging of the demential process the Clinical Dementia Rating (CDR) was used. The clinical evaluation of dementia classifies the development stage of AD, where zero rating represents normal, 0.5 dementia questionable (CDR 0.5), 1 mild dementia (CDR 1), 2 moderate dementia (CDR 2) and 3 severe dementia (CDR 3) 8 .
The blood collection was performed at patients' home according to the recommendations for venous blood collection from the Brazilian Society of Clinical Pathology and Laboratory Medicine (2010) 16 . The complete blood count (CBC) was carried out in a hematological analyser Cell-dyn Ruby (Abbott, Illinois, USA) with 33 parameters that use flow cytometry for the white cells, erythrocytes and platelets analysis and the spectrophotometric methodology for the determination of hemoglobin. The hematological parameters were taken from the WHO 20 .

Statistical analysis
The statistical analysis was completed using SPSS package version 20.0. Initially, it was carried out using non-parametric analysis due to the large number of variables that do not fall in a normal distribution when evaluated for Shapiro-Wilk test. The first step of the analysis consisted of the comparison of the initial evaluation (base line) between the subjects who have died during the study, and those revaluated in a later phase according to the test proposed by Mann-Whitney. This step also analyzed the survival according to the tests proposed by Kaplan-Maier and considering CDR, IMC, clinical variables, such as age and, diagnosis data as time variable. This step was also analyzed by Qui-Square test with corrections to verify associations between death and sociodemographic and clinical variables.

Ethics and consent
The study was approved by the Research Education Committee of the State University of the Center-West -COMEP / UNICENTRO through process 026/2011. Written informed consent was obtained from a person responsible/patient caregiver for all participants.

Results
Respiratory issues accounted for the majority of deaths (Table 1). Alzheimer disease was specified in 22% of cases (n=2) and the patients without a specific cause of death on the death certificate was 22% (n=2). No patient had ICD G30.0, early-onset Alzheimer's disease (Table 1) 9 .
As shown in Table 2, the anthropometric variables and macronutrients ingestion are compared between live and patients who died up until two years after the study evaluation. None of the subjects showed values greater than zero on basophils, atypical lymphocytes, blasts, myelocytes and metamyelocytes variables; values greater than zero indicate pathologies. Table 3 shows the differences on micronutrients consumption between the two groups investigated. The only significant difference found between live and deceased patients, was the phosphorus consumption. Figure 1 shows that the phosphorus consumption was lower those who died compared to the live    group, although the phosphorus consumption was sufficient for both groups.
In addition, Table 4 shows that there were no deaths in patients classified as eutrophic (MAN specification), while all the undernourished patients and at risk of it had died.
According to the CDR classification of the patients who died, 55% were in mild cognitive failure. Of these, 44% were male, which represents half of the male group of the sample. The health conditions showed that 23% (n=7) of the patients had some type of cancer, and from them, 42% (n=3) have died. Figure 2 shows the AD patient survival curve as a function of age. The survival rate is modified for the disease stage and the results demonstrate a lower survival rate for the patients in the early stages of the disease.

Discussion
This study has observed that the respiratory issues may appear as one of the main death cause in AD patients, which may be related to age, due to the advanced age of the majority of the patients. The limitation of the respiratory functional system on AD, due to the depletion of the cough reflex and the accumulation of discharge, reduce the pulmonary ventilation and the elderly immunological system is more affected, facilitating the occurrence of infection, especially respiratory ones 21,22 .
The elderly with dementia can exhibit dysphagia, making necessary the use of enteral probes for feeding and the elimination of the discharge from upper airways, increasing the respiratory infections. Pneumonia is one of the most common causes of hospitalization of AD patients and its pathology is directly related to its mortality [21][22][23][24] .
Furthermore, seniors are more likely to present with secondary drug effects, increasing the death risks. The elevated chance of pneumonia in AD patients has to be remembered when treating these people 25 . A study in Finland compared the risk of pneumonia associated with acetylcholinesterase inhibitors (AChEls) (donepezil, Transdermal rivastigmine, galantamine and memantine) in people with AD diagnosis and concluded that people who used rivastigmine and memantine in comparison with donepezil had an increased risk of pneumonia 26 . This risk occurred, not only from the drug characteristics, but also for the severity of the AD. The same study considered the use of pneumococcal vaccination for elevated risk of pneumonia with great efficiency 26 . In Brazil, memantine is used only in moderate and severe stages of the disease.
Other medicines that deserve special attention are the neuroleptics, or tranquilizers, used in AD patients to control the agitation, hostility, delusions and hallucinations 26 . These medicines are antagonists of dopamine, decreasing its transmission and compromising the deglutination function, facilitating aspiration pneumonia 26 . The effects of the neuroleptics on the dopaminergic system usually occurs in severe AD, so it is used in lower doses.
On the other hand, the benzodiazepines that can also be used as tranquilizers in AD have slight effects on the dysphagia, decreasing aspiration pneumonia 26 .
For cause of death analysis, the death certificates must be filled in a correct form. The death certificate is the only source of mortality data and depends directly on the quality of the information provided by the doctor 27-29 .
The underlying death cause was defined by the World Health Organization as a disease or lesion that had initiated the pathological events chain and brought the patient directly to death 6 .
In line with the methodology proposed for this study, it was recorded as AD with the codes CID G30.0 to G30.9. Underlying cause of death being AD, the code G30.0 did not appear in any certificate, reinforcing the idea that the death certificates frequently register the immediate death causes instead of to register AD as a cause 28,29 .
AD generally causes undernourishment to the elderly and it can be associated with physiological alterations, dental issues, physical limitations, and diseases that decrease the appetite, food absorption and, drug interactions 7,30 .
Although there is no difference between the groups in this study, it could be observed that values are below those recommended by the Dietary Reference Intakes -DRIS (15) for almost all micronutrients. Figure 1 shows that phosphorus was the only micronutrient with significant differences found between the live and dead group 31 . However, this difference cannot explain the observed mortality rates, since the percentage of this micronutrient adequacy was inside the normal values in those who died and above the recommended level in the live group 7,31 . In addition, the most significant symptoms of phosphorus deficiency include bone weaknesses, discomfort in many joints, weakness, dental decay and rickets 31,32 . In this study, there was no biochemical analysis of phosphorus dosage in the study groups due to the methodology limitations.
Therefore, the early diagnosis of malnutrition is important due the association with the greater mortality in AD. The data from this study supports this, since all the patients who have died were undernourished according to MAN classification 7,32 .
The consequences of the weight loss and malnutrition in people with dementia occurs due to the dependency of care, difficulties in mobility, food preparation, chewing and deglutition. Moreover, the nutritional status of the elderly AD patients has a significant impact on the disease course, affecting cognitive parameters and functional symptoms 7,30,31 .
Additionally, the malnutrition causes a decrease in white blood cells, leading to alterations in antibodies production and immune resistance, decreasing the survival rate in undernourished patients 33 . Thus, to reduce or at least delay the AD prevalence, maintaining a proper diet from childhood is crucial and reinforced due the fact that the cardiovascular risks are directly related with the development of dementia.
In relation to the AD associated factor, our study reveals that 33% of those who died had higher levels of cholesterol and 44% had diabetes. An epidemiological study has demonstrated the directly relation between high cholesterol levels and the increased risk of cognitive issues, vascular alterations and AD 34 . In terms of diabetes, in this study its presence was not associated with the mortality rates; the survival curve did not show differences between subjects with and without diabetes (p=0.572), although diabetes mellitus is a disease which has risk factors common to AD 34 .
A prospective study from 2010 has evaluated the associated factors of evolutionary differences and mortality due to AD between males and females. The males had more comorbidities due the differences in exposure, work, alcohol consumption, tobacco consumption and differences on the attitudes related to the disease, increasing the chance of death 35 , reinforcing our study, where 44% of the those who died during the study were males. In some cases, the male mortality occurs due to external causes, given that females have higher chances of developing AD in advanced ages, correlating with our sample composition 27,35 . On the other hand, a study conducted by Teixeira et al. (2015) shows a higher mortality in AD females, which may be explained by the prevalence of cases in this gender in Brazil and other countries.
Among the possible comorbidities occurring in this population, cancer stands out, decreasing survival in the elderly. This effect was not completely observed in this study, where 76.7% of the patients were not cancer patients. From the 23.3% that were cancer patients, 57% have died since the beginning of the study. Still according to Teixeira et al. (2015), the neoplasms were more frequent in patients without AD 29 .
In the study performed by Mussico et al. (2013), elderly patients with cancer have a reduced risk of dementia due to AD and vice-versa 36 . Although demonstrating common pathophysiological mechanisms such as mitochondrial dysfunction, oxidative stress and DNA damage that are propitious to the genetic and metabolic connections between cancer and neurodegeneration, there is an inverse association between the pathologies 36,37 . Romero (2014), in a prospective study, analyzed the cancer notification proportion on death certificates of dementia patients and concluded that AD patients have a lower risk of mortality due to malignant neoplasia 38 . In our study, the survival curve has not shown significant differences when the subjects with and without cancer were compared (p=0.998).
A decrease in the survival rate was observed in Figure 2 showed a difference between the distinct disease stages. However, the lower survival values were observed in patients who were classified in the mild dementia stage, an unexpected result that is not in line with the literature 6,23,28 . Nevertheless, the death of patients of the study is strongly related to respiratory issues and to the progression of the disease. The risk factors for pneumonia on initial AD stage are not sufficiently understood, but one possibility is that a silent aspiration of saliva or bolus without causing cough, respiratory difficulty or other external sign can cause pneumonia due to aspiration 24,39 . Also, because the symptoms and effects caused by the AD are less evident in mild disease stage patients, the preoccupation with immunological care may be reduced 33 .
Furthermore, while several studies use the diagnostic time as the survival time, this study has not used it; first due to the difficulties, such as lack of specialized doctor in SUS to assist the great demand and the lack of information about AD, confusing its symptoms with simple senility or not searching for an appropriate diagnostic. Second, the correlation with age is associated with the disease and the survival.
The obtained results were discussed and interpreted considering the limitation of a cross sectional study with a limited number of patients. In addition, the information was not collected before the patient's death due to the improbability of each case. Nonetheless, the longest past period since the evaluation and the death was 18 months. Thus, the number and quality of the information fortify the results with social relevance, since they were obtained from a non-institutionalized dementia, where caregivers are relatives, a common situation among people with dementia in Brazil.
The nutritional status, biochemical and basic comorbidities were not confirmed in this study to explain this relation. Therefore, the AD patients profile characterization in the study and the observed reduction on the survival probability may indicate the absence of proper attention, suggesting new searches for its confirmation. At the same time, the quality of the information in the death certificates that frequently do not register dementia as an implicit cause generates underreporting of the long-term chronic degenerative diseases.
Other limitation is the quality of information given in the death certificate, which does not record AD as an implied reason, giving no notice of chronic degenerative diseases and hiding results. This suggests there is a need for more research in this area, because there is still a shortage of studies on this subject.

Data availability
All raw data was freely available on the Open Science Framework site: Doi: 10.17605/OSF.IO/PE38C. 1.

5.
causes of death. Conditions such as respiratory disease are common causes of death in any older population and the study does not evaluate whether they are any higher in people with dementia than the general population, so it is misleading to be discussing, for example, elevated risks of pneumonia or cancer or whatever.
Although I don't object to the authors displaying descriptive data on their baseline covariates stratified according to subsequent death, a study of this size would only be able to identify very large differences at statistical significance; therefore I don't feel that it is appropriate to be discussing the findings in any more detail than a simple presentation of descriptive data. The difference in phosphorus levels is highly likely to represent type 1 statistical error because of the very large number of comparisons made, so it is reasonable not to draw any inferences from this. The rest of the discussion about risk factors for mortality is not justified because the study does not have statistical power to detect these associations and, indeed, no significant associations are identified.
It is reasonable to comment on levels of survival in the sample, but I don't think that anything can be concluded about stage of dementia and survival because the numbers are simply not large enough.
The repeated description of the study as cross-sectional is misleading. The study took a group of 30 people and followed them up for mortality. It is therefore surely a cohort study? Similarly, I don't understand the sentence in the third-to-last paragraph: "In addition, the information was not collected before the patient's death due to the improbability of each case." Surely all the information was collected before death (apart from the death certification)?
As far as I can tell, the participants were people who were recorded as receiving medications for AD and who were then approached and visited at home. There does not appear to have been any formal evaluation of the diagnosis, apart from the completion of a CDR. Is this correct? It would be helpful if the sample identification and recruitment process was a little more clear.

Is the work clearly and accurately presented and does it cite the current literature? Partly
Is the study design appropriate and is the work technically sound? Partly

If applicable, is the statistical analysis and its interpretation appropriate? Partly
Are all the source data underlying the results available to ensure full reproducibility? Yes

Are the conclusions drawn adequately supported by the results? No
No competing interests were disclosed.

Competing Interests:
Reviewer Expertise: Dementia epidemiology I confirm that I have read this submission and believe that I have an appropriate level of expertise to confirm that it is of an acceptable scientific standard, however I have significant reservations, as outlined above.
The benefits of publishing with F1000Research: Your article is published within days, with no editorial bias You can publish traditional articles, null/negative results, case reports, data notes and more The peer review process is transparent and collaborative Your article is indexed in PubMed after passing peer review Dedicated customer support at every stage For pre-submission enquiries, contact research@f1000.com