Nutritional and epidemiological aspects of patients with chronic renal failure undergoing hemodialysis from Brazil , 2010

INTRODUCTION
The Nutrition Committee of the Brazilian Society of Nephrology (SBN) held in 2010 the first Brazilian Nutrition Census in hemodialysis patients. Multicenter data contribute to clinical development and nutritional intervention.


OBJECTIVE
To describe epidemiological and nutritional aspects of hemodialysis patients.


METHOD
Cross-sectional study in 36 dialysis clinics and 2,622 randomly selected participants. Socio-demographical, clinical, biochemical and anthropometric records were collected.


RESULTS
60.45% of the patients lived in the Brazilian Southeast. 13.53% came from Northeast region, while 12.81% from South, 10.33% from Midwest and 2.86% from North regions. Approximately 58% were male and 63.1% were below 60 years old. 58.5% of patients were married or in cohabitation. Around 80% of them depended on the government Unified Health System. Smoking showed a difference between gender and age. Presumptive etiologies were Hypertensive Nephrosclerosis (26.4%), Diabetic Nephropathy (24.6%), unknown/undiagnosed causes (19.9%), Glomerulopathies (13.6%) and others (11.2%). Both Hypertension and Diabetes Mellitus affect approximately 30% of patients, especially over 60 years. Body Mass Index did not differ between genders, although it differed between age groups and when used different evaluation criteria. Men and women average waist circumference were respectively 90.5 and 88.0 cm. Lipid profile did not differ between age groups, but it did between genders. Albumin values were lower in women and in patients older than 60 years.


CONCLUSION
This study characterized Brazilian hemodialysis patients in 2010, and may support further studies to monitor nutrition and epidemiological transitions of the population.


IntroductIon
The Nutrition Committee of the Brazilian Society of Nephrology conducted the first Brazilian Nutrition Census in hemodialysis patients in 2010, by voluntary participation. 1 Several difficulties were encountered in collecting and obtaining data, which depended on the commitment of several professionals (nephrologists and nutritional physicians) from dialysis centers across Brazil.Nevertheless, we obtained an impressive number of contributions, which allowed the analysis of approximately 2500 patients under a chronic hemodialysis program.
4][5][6] This census is the first study to present the nutritional aspects of this population covering all of the Brazilian territory.Our goal is to describe the nutritional and epidemiological characteristics of patients with CKD undergoing hemodialysis in Brazil in the year 2010.

MAterIAls And Methods
This is a cross-sectional study comprising adults and elderly individuals undergoing hemodialysis in 36 dialysis clinics spread throughout Brazil, who voluntarily accepted (December 2009) the invitation from the Nutrition Committee of the Brazilian Society of Nephrology to participate in the study, after signing and sending an agreement.Data collection took place between January and July 2010.Most patients in the participating clinics were undergoing the classic hemodialysis treatment: 4-hour sessions, 3 times a week.The participating centers received an e-mail with a form containing specific questions developed by the committee team.After the data collection, the databases were e-mailed to the Nutrition Committee.The nutritionists at the clinics, in charge of collecting the data, were instructed about the procedures and protocols when the form was sent, but the committee did not interfere with the process.
The inclusion criteria for patients were as follows: aged ≥ 18 years, undergoing a classic hemodialysis program; the exclusion criteria were as follows: aged < 18 years, or undergoing another type of renal replacement therapy.A total of 2622 patients undergoing a chronic hemodialysis program comprise the study sample (50% of the total number of patients from the dialysis unit).The selection was performed at random by choosing alternating patients in the alphabetical list, and all the selected patients agreed to participate in the study.
The sociodemographic information (age, sex, race, civil status, and smoking status) and clinical data (hemodialysis time, dry weight, interdialytic weight gain [average value at the month the data were collected], access type for the procedure, number of absences, comorbidities [Diabetes Mellitus -DM, systemic arterial hypertension -SAH], etiology of the CKD [diabetic nephropathy, hypertensive nephrosclerosis, glomerulopathies, polycystic kidney, unknown causes, other]) were obtained through a clinical records survey.
We followed the consensus of the World Health Organization (WHO) that establishes "elderly individuals" in developing countries as those ≥ 60 years of age.We also analyzed serum biochemical parameters (albumin, triglycerides, total cholesterol, and its fractions) from the results of the last examinations at pre-dialysis, available in medical records (varying from 0 to 6 months, as some of these examinations are performed quarterly or biannually in accordance to RDC154). 7The efficiency of the dialysis was estimated using Kt/V indices (according to the Daugirdas II formula). 8Waist circumference (WC) and body mass index (BMI) were used as anthropometric measurements; BMI is the ratio between the dry weight (kg) and the square of the height (m 2 ).
To avoid potential influences of the hydration state in this assessment, all of the researchers were nutritionists and they were instructed that the anthropometric measurements collected in the nutritional assessment protocol of each dialysis center be within the following standards: a stadiometer should be used for the measurement of height; at the time of measurement, the patient should be barefoot, with heels together, back straight, and arms extended alongside the body.For the WC, the standard used was the umbilicus to minimize error among researchers.
Descriptive statistical analysis of the variables was conducted by calculating the frequencies, medians, 25% values, and 75% values.The sample standard errors or 95% confidence intervals of the variables expressed in proportion form were calculated and expressed along with the percentage values.In relation to variables such as smoking, BMI, WC, and Kt/V, the possible interactions between sex and age of the patients were verified by the odds ratio of Mantel-Haenszel or by the Friedman test.The nonparametric verification of the quantitative variables was conducted using the Shapiro-Wilk "W" test, before or after logarithmic and radical transformations of the data.
The comparisons were performed for the country as a whole and, hence, potential characteristic influences, such as region of the country or racial groups interviewed, were disregarded.Comparisons with different parameters for patient groups of different sexes and different ages (< 60 years and ≥ 60 years) were conducted.Comparisons between groups of nondichotomous and nonnormalized variables were conducted using the Mann Whitney test. 9The comparisons between the proportions for the groups were conducted using the Pearson Chi-squared test, 9 with a posteriori comparisons of dummy variables for the vascular access type and race data (white, black/mixed race, and others).For Nutritional characterization of hemodialysis patients in Brazil the analyses, the Stata program, version 10.0, was used, considering differences with significance level equal to or below 1%.
Most patients were males (58.4% ± 1.9%) aged ≤ 60 years (63.1% ± 1.86%).We verified a higher number of Caucasian patients in the south region and those of African descent in the north and northeast regions (Table 1).
With regard to civil status, 58.5% ± 1.8% of patients were married or in cohabitation.With regard to education, we verified a high rate of illiteracy, especially in the northeast region, and a predominance of < 8 years of education in this population (Table 1).Most patients were homeowners (80%).However, sanitary conditions varied according to region.While more than 90% of the homes in the north and southeast regions, and 85% in the south region had a sewage network, this percentage was reduced to 60% and 70% in homes in the northeast and midwest regions, respectively.
Most patients (80%) were on welfare and dependent on the Universal Healthcare System (SUS).However, we observed a great disparity in the percentage of patients undergoing treatment through  1).Smoking is a cardiovascular risk factor.We verified a higher ratio of smokers among men aged ≤ 60 years (18%), when compared with men > 60 years (15.5%).Such ratio difference did not occur significantly among women of different ages (Table 2).
The arteriovenous fistula is the preferred type of vascular access for the hemodialysis procedure across the country, present in approximately 90% of the patients.However, patients ≥ 60 years of age showed a significantly higher number of vascular accesses with a permanent catheter.
We verified good treatment attendance in most of the assessed clinics, as only 3.5% (2.8%-4.9%) of the patients missed the sessions more than once a month. 1 BMI: body mass index, according to WHO for adults and Lipschitz for the elderly. 2 EBPG Nutrition Guideline. 3Waist circumference according to WHO. 4 Albumin according to the European consensus. 5IDWG: interdialytic weight gain. 6DW: dry weight.
With regard to the nutritional characteristics characterized by region, the northeast region showed higher indices of elderly patients with low weight, when compared with the other regions (Table 3).The north region showed a lower ratio of patients with WC values above the references adopted by WHO, in both men and women (Table 3).
However, the north region showed a higher ratio of patients with albumin < 3.8 g/dl and a higher ratio of weight gain above 4.5% of the dry weight.
When compared according to sex (Table 4), female patients gained less interdialytic weight than male patients did, and they presented higher Kt/V indices, or better dialytic compliance.We compared  2), we verified a higher frequency of DM in patients aged ≥ 60 years.We also observed, according to race, that Caucasians have a higher ratio of patients ≥ 60 years old than African descendants.
In relation to anthropometric characteristics, we verified that the BMI calculation did not differ between sexes.However, there were significant differences between age groups (Table 5), and both median values were within the reference values used for the general population, according to WHO (18.5-24.9kg/m 2 ). 10 With regard to the WC, women showed median values that were higher than those recommended by WHO (≤ 80.0 cm). 11As for men, more than half of the patients in the sample were within the recommended values (≤ 94.0 cm). 11he lipid profile of the patients did not differ in relation to age; however, we observed differences between the sexes.The values of the medians were within the reference values according to the IV Brazilian Guideline for Dyslipidemia. 12he median values of albumin were lower in women as well as in patients aged ≥ 60 years; however, the difference was small and insignificant.

dIscussIon
The assessment of the obtained data enabled us to determine the epidemiological and nutritional profile of Brazilian patients under chronic hemodialysis programs.However, the study has some limitations.Among them, the low responsiveness of the forms was a particularity taken into consideration during the characterizations conducted, as approximately one-twentieth of all dialysis centers participated in the study-approximately 5% of the population of patients undergoing dialysis treatment in Brazil, with most records being from patients receiving treatment in the southeast region.
Moreover, nutritional profile comparisons of the Brazilian regions should be conducted in new studies, taking into consideration several characteristics of these geographical spaces, to control the presence of confusing variables and to discuss in depth all of the diversities of the target population.Biochemical serum values, such as bicarbonate and phosphorus, would be useful in assessing the effect of metabolic acidosis on the lean mass of these patients; however, these were not analyzed in this study because of unavailability of data.
Another limitation worth mentioning is that laboratory data were collected from patient charts.This means that different biochemical variables may have been measured using different techniques.Additionally, the examinations were not performed on the same date in all clinics, and they were not collected for the study; those from routine examinations were used.The northeast region showed the least number of homes with a sewage network, the lowest education, the highest illiteracy rate, and the highest dependency on public health aid.Adding to these characteristics, we also verified a higher ratio of patients with low weight (WHO 10 and European Consensus on Nutrition). 13These results signal the need for new studies to verify the effects of housing infrastructure quality and educational level on the nutritional indicators covered by this study.
The data on sociodemographic characterization, CKD causes, and the presence of comorbidity are similar to the results obtained in the last dialysis census conducted by the Brazilian Society of Nephrology. 2,14We observed a lower number of patients on welfare (80%) in relation to the dialysis patient census conducted by the Brazilian Society of Nephrology (85.8%), perhaps due to the difficulty in organizing nutrition services in clinics that are more dependent on the SUS.
We verified that the most frequent presumptive etiology of CKD was hypertensive nephrosclerosis (26.4%) followed by diabetic nephropathy (24.6%).These data are in compliance with the national data, which, however, show a more accentuated dominance of hypertensive nephrosclerosis (35.2%), followed by 27.5% of cases of diabetic nephropathy.
In relation to the comorbidities, SAH was frequent in this population, regardless of sex or age, being a little more present in male patients (59.68% versus 40.32%).On the other hand, DM was more present in elderly patients (46.32% versus 21.22%), in compliance with the data from VIGITEL 2010, which verified a growing number of diabetic and hypertensive patients with increasing age in Brazil. 15moking is not frequent in the population assessed; the overall frequency found was low.Nevertheless, smoking is more common in elderly men, which deserves additional attention.
With regard to dialytic compliance, most patients showed regular attendance to the sessions, with Kt/V above the recommended values (Kt/V > 1.2) and a small gain of interdialytic weight (a gain below 4% of the interdialytic dry weight was considered as a low weight gain).Other obtained data (20.7%)were similar to those obtained by the dialysis census from the Brazilian Society of Nephrology, which detected 19.2% with Kt/V values below the recommended values.We verified that the Kt/V value was better in female patients, possibly because the dialysis dosage was higher in this group with less body surface, as most dialysis clinics standardize the size of the capillary filter for the overall population in the program.

Nutritional characterization of hemodialysis patients in Brazil
With regard to vascular access, we verified that temporary catheters were more frequent in female and elderly patients, approximately twice that of permanent catheters.In contrast, we verified a total percentage (8.5%)below the national average of 13.6%.
On the basis of the WHO classification for identifying the elderly population ( ≥ 60 years), we verified that 36.9% of patients in hemodialysis programs are elderly-data similar to that obtained by the Brazilian Society of Nephrology 2009 census (39.9%).
With regard to the nutritional characteristics, approximately 45% of the individuals in all regions were classified as eutrophic.Approximately 7% of the adults (18-60 years old) and 25% of the elderly (> 60 years old) showed low weight (WHO 10 and Lipschitz). 16However, according to the EBPG Guideline on Nutrition, 13 approximately 42% of the patients were classified as being under nutritional risk.The ideal cutoff point for BMI is controversial, as lower values for this index are associated with a higher mortality rate. 17,18It is also worth noting that BMI is not considered sensitive for detecting protein depletion and visceral fat increase, and it may also be influenced by water retention, which is relatively frequent in patients with chronic renal disease. 19C is considered the best predictor of visceral fat 20,21 and, at present, the recommendations proposed are for young adults, without considering changes in fat distribution that commonly occur with aging. 22,23s in our study most patients were in the age range below 60 years, we can infer that the WC value is a parameter that can be used in this population.
Lipid metabolism presents changes from the initial phases of CKD, and the presence of coronary disease is common, even in the presence of normal levels of low-density lipoproteins.Although in this population, the expected plasma concentrations of triglyceride are high, we did not verify any difference between the triglyceride values across groups.We verified high-density lipoprotein levels 24 below what is expected to protect the cardiovascular system in both sexes.However, these findings should be interpreted with caution, as the present study did not survey which patients made use of hypolipemiant drugs, which may directly interfere with the results found.
Cardiovascular mortality is increased in patients with CKD undergoing a hemodialysis program.In the United States, data analysis performed between 1994 and 1996 for this population observed a risk of death of approximately 10 to 100 times greater in patients undergoing hemodialysis than for the overall population corrected by age, 25 and cardiac disease was the cause of death in approximately 45% of patients undergoing hemodialysis. 26n Brazil, according to the Brazilian Society of Nephrology dialysis census of 2009, 14 the mortality rate in this population is also high and the main causes of death are cardiovascular (34.9%) and cerebrovascular (8.8%) diseases, totaling 43.7% of the cases, followed by infection (24.3%).
Serum levels of albumin may be considered late markers of malnutrition, broadly used in individualized clinical practice for CKD patients, as hypoalbuminemia values suggest future risk of disease and death. 13Results of albumin serum level measurements may be influenced by inadequate dietary consumption of calories or proteins, hydration status, inflammation, anabolic or catabolic processes, age, urinary loss, hepatopathies, and iron deficiency anemia. 13,27However, values below 3.8 g/dl, as proposed by the EBPG, were present in 37.4% of the sample, which shows the need for more studies with this cutoff point, to investigate the adequate value to be validated for CKD in Brazil.The study was performed nationwide and may differ from the findings of other countries.

conclusIon
This study covered nutritional variables used routinely in clinical practice by nutritionists who assist CKD patients undergoing hemodialysis.The results showed that age must be taken into consideration in the nutritional assessment, in addition to the differences found in lipid profile, dialysis efficiency, and ratio of hypertensive patients.
Similarly, the patient's sex is also relevant and should be given attention, as it presented a difference with regard to dialysis quality, interdialytic weight gain, WC, lipid profile, and albumin.It should not be assessed along with comorbidities, race, and BMI.
This study showed the importance of describing the nutritional epidemiological profile in Brazil, and may in the future subsidize new comparative

tAble 2 ratiO
Of sOciOdemOgraPhic and clinical data Of the Patients, accOrding tO sex and age fOr Patients with chrOnic Kidney disease undergOing hemOdialysis, brazil, 2010 * significance 0.01; DM: Diabetes Mellitus; SAH: systemic arterial hypertension; Perm.Cat.: permanent catheter; Temp.Cat.: temporary catheter.n: number of sampled patients.