Nutritional Profile and Inflammatory Status of Hemodialysis Patients

Abstract

Background. Malnutrition and abnormal inflammatory markers are prominent features of the uremic syndrome, but associations and repercussions are somewhat controversial. Objective. To determine nutritional and clinical profile of hemodialysis patients, aiming at potential diagnostic recommendations for stable subjects with elevated C-reactive protein. Material and Method. Design: Prospective observational cross-sectional clinical study in a stable chronic hemodialysis population; Setting: Renal and Nutritional Service of a mid-size charity academic hospital; Patients: Subjects (n = 44) were analyzed concerning nutritional status and C-reactive protein. Some displayed acute infections (Group I, n = 9) and others did not (Group II, n = 35). Age was 47.0 ± 16.9 years with 63.6% males. Body mass index (BMI) was 22.2 ± 3.9 kg/m2, calorie intake was 1262 ± 601 kcal/day (20.7 ± 6.7 kcal/kg/day), and protein ingestion was 74.3 ± 16.6 g protein/day (1.2 g/kg/day); Intervention: No nutritional supplement or artificial modality of alimentation was employed in this series; Main outcome measures: Subjective global assessment and C-reactive protein. Results. Malnutrition estimated by subjective global assessment (SGA) was very common (>90%), despite acceptable BMI and serum albumin. C-reactive protein was moderately elevated in 40.9% and was positively associated with SGA and negatively with plasma proteins. Comorbidities were associated positively with extracellular water and negatively with reactance (bioimpedance). When infected versus non-infected cases were analyzed, 100% of the former displayed high CRP concentrations in contrast with 22.9% of remaining patients. Conclusions. 1) Malnutrition profile was rather unique, with relatively favorable objective findings (body mass index, serum albumin) and more deranged SGA; 2) Bioimpedance analysis suggested that phase angle could be used as an indicator of nutritional status; 3) C-reactive protein was negatively associated with plasma proteins; 4) Infected subjects, although few and displaying moderate clinical troubles, were responsible for most C-reactive protein determinations above 8 mg/L; 5) Investigation of occult infectious foci is advised in these circumstances.

Keywords:

• hemodialysis
• renal failure
• C-reactive protein
• acute infection
• nutritional assessment

INTRODUCTION

Malnutrition is a prominent feature of the uremic syndrome, and nutritional markers such as serum albumin (ALB) and body mass index (BMI) are associated with mortality risk in hemodialysis (HD) patients.[1] Various anthropometric variables have also been shown to be more depressed in such individuals than in the general population, with the exception of uremic diabetics who tend to be overweight.[2]

Signs of systemic inflammation are regularly unveiled in infection-free HD cases, confirmed by elevation of acute-phase proteins and pro-inflammatory cytokines, and these are certainly endowed with long-term prognostic significance.[1],[3] The impact of asymptomatic or covert infections or contaminations on baseline nutritional status and inflammatory markers has been addressed by studies concerning periodontal diseases,[4] Helicobacter pylori gastritis,[5] Chlamydia pneumoniae foci,[6] sterility of dialytic water, and modality of equipment and membranes in use.[7] Other potential sources of aggression are oxidative stress and advanced glycosylated end-products.[8]

The role of minor but clinically relevant infection has not been adequately investigated in this context, although uremic subjects do not infrequently display acute microbial episodes of various etiologies, because most protocols prefer to list them in the exclusion criteria.

Aiming to analyze possible correlations between active infection, biochemical markers, and nutritional status, a prospective study was performed in a HD Unit.

MATERIAL AND METHODS

Population: Stable outpatient uremic adults undergoing chronic HD were recruited for this study. Criteria for inclusion were:

• Age 18–85 years, males and females

• HD for at least 3 months, 3 times/week

• Ambulatory and receiving oral diet

• Residual renal clearance <1.5 mL/min

• Hemodialysis access via arteriovenous fistula

• With or without incidental infection

• Informed consent

Exclusion criteria:

• SIRS, sepsis, shock, multiple organ failure, coma

• Clinical or surgical hospitalization in the last 30 days

• Ongoing enteral or parenteral nutrition

• Use of steroidal or non-steroidal anti-inflammatory or immune-suppressive agents

• Cardiac pacemaker

• Advanced senility or dementia interfering with application of the nutritional questionnaire

• Refusal to cooperate with the study

Forty-four consecutive patients fulfilled the listed conditions and were enrolled in the protocol. There were no exclusions.

Objective: To determine nutritional and clinical profile of hemodialysis patients, aiming at potential diagnostic recommendations for stable subjects with elevated C-reactive protein.

Patient accrual: This was a cross-sectional study performed during a single midweek dialytic session on the population on maintenance hemodialysis in the Dialysis Unit of Santa Rita de Cassia Hospital, Vitoria, ES, Brazil. Patients were assessed at the rate of 4–6/week, and total duration of the protocol was 2 months.

Vascular access: A fistula between the radial artery and cephalic vein was constructed by a vascular surgeon with a minimum flux of 300 mL/min, and used after a 4–5 week maturation interval.

Dialytic routines: Dialysis treatment consisted of three weekly sessions using a bicarbonate buffer. Hemodynamic parameters were routinely monitored and individual adjustments (blood flow, dialysate flow) were kept stable throughout the study period. Efficacy of dialysis was confirmed by standard pre- and post-dialytic biochemical tests.

Experimental design: All data were recorded on a mid-week dialysis session. Body mass index (BMI) was obtained from height and post-dialysis body mass, while ideal body mass was calculated using the Lorentz formula. Blood was collected after an overnight fast before HD and processed by an autoanalyzer, whereas hematologic counts were also done in an automated apparatus. Questionnaires were documented at the same time.

Dietary intake: Dietary recall forms, including both midweek and weekend profiles, were processed by a specialized dietitian and analyzed by a computer program validated for Brazilian food (UNIFESP, São Paulo, Brazil). Anthropometry and BIA were investigated immediately after dialysis, according to accepted routines.[9] Conventional (single-current) whole-body BIA was assessed by the BIA-Quantum apparatus using the classic tetrapolar technique and Fluids software (RJL Systems, Clinton Twp, MI, USA).

General and demographic information: Age, gender, diagnosis, comorbidities, renal function, clinical and dialysis history, infectious and non-infectious complications.

Definition of comorbidities: diabetes mellitus requiring drugs or insulin, malignancies requiring chemo or radiotherapy, class IV congestive heart failure, unstable angina pectoris, major systemic infections such as tuberculosis or AIDS, chronic pulmonary disease requiring supplemental oxygen, and cirrhosis with encephalopathy or abnormal prothrombin time.[10]

Definition of infection: Non-critical infectious events such as vascular access infections, symptomatic periodontitis, influenza, pneumonia, diarrhea or cystitis, with or without elevations in body temperature, white blood cell count, heart rate or respiratory frequency.[11] All infections were characterized by clinical manifestations and suggestive diagnostic tests. Collection of microbial samples was enforced; however, few specific pathogens were identified, thus preventing valid analysis.

Fistula infection was registered in the presence of swelling, redness, pain, or secretion. The appearance of productive or dry cough and elevated temperature along with thoracic x-ray infiltrates reproducing lobar pattern indicated pneumonia, whereas fever and severe malaise with upper respiratory symptoms and normal X-rays was classified as influenza. Intestinal infection was reported on the basis of abdominal pain and liquid enteric discharge in the absence of chronic intestinal conditions (Crohn's disease, ulcerative rectocolitis) or diarrhea-causing drugs or nutritional preparations. Low abdominal discomfort and/or urinary urgency with abnormal urinalysis check were considered features of cystitis. Only localized derangements consistent with stable clinical conditions and full ambulation were accepted, as bedridden or hospitalized subjects, as well as those fulfilling the SIRS or sepsis definition,[12] would entail exclusion.

Nutritional variables: Weight, height, BMI, arm circumference (AC), triceps skinfold (TSF), modified Subjective Global Assessment (SGA),[13],[14] ALB, transferrin (TRANS), ferritin (FERR), and total lymphocyte count (LYMPH).

Bioimpedance (BIA): Variables of body composition included resistance, reactance, phase angle, body fat, lean body mass, total water, extra and intracellular water, and body cell mass. The ratios of extracellular water (ECW) to lean body mass (LBM) and of total body water (TBW) to lean body mass (LBM)[15] were documented as well.

Biochemical tests: Total, HDL, and LDL cholesterol (CHOL, HDL, LDL), and triglycerides (TRIG).

C-reactive protein (CRP): Serum CRP concentrations were determined by nephelometry (BN II; Behring, Sâo Paulo, Brazil). The lower limit of detection of CRP was 2 mg/L. CRP reference values for healthy individuals were below 5 mg/L.

Assessment of covariates and data collection: Defined measurements were those obtained at the time of the interview, unless stated otherwise. Santa Rita de Cassia Hospital has an experienced multi-professional nutritional team, and diagnostic procedures were done in strict accordance with applicable guidelines of the American Dietetic Association.

Statistical investigation: Values are presented as mean ± SD. Parametrical (analysis of variance and Student's t test), as well as non-parametrical tests (Mann-Whitney, Chi-square analysis) were applied as appropriate, along with Parson's linear regression analysis for comparison, of CRP, body composition and nutritional indices. In all circumstances a significance of 5% (p < 0.05) was adopted.

RESULTS

Age of the group was 47.0 ± 16.9 years (18–82), with 63.6% males (28/44). HD was maintained during 34.1 ± 39.4 months (3–172), and main etiology of renal failure was systemic arterial hypertension (77.2%, 34/44), followed by diabetes mellitus (15.9%, 7/44) and miscellaneous diseases.

Comorbidities were identified in 45.5% (20/44), with a mean frequency of 1.8 ± 0.8/patient in this group, or 0.7 ± 1.0/patient for the entire population. Acute infections were recognized in nine subjects (20.5%). Most were related to the arteriovenous fistula (15.9% of the patients, 7/44), along with one pneumonia and one infectious diarrhea. Chronic microbial problems were diagnosed as well in this series, nominally one pulmonary tuberculosis and two cases of virus-C hepatitis. However, these patients were receiving adequate therapy and had become asymptomatic, therefore they were not stratified among acutely infected individuals.

BMI was 22.2 ± 3.9 kg/m2 (14.9–29.6). Triceps skinfold was 14.0 ± 5.3 mm and arm circumference was 25.0 ± 4.0 cm. Modified Subjective Global Assessment estimated 14.0 ± 4.6 points (8–33). Two patients (4.6%) were well-nourished according to these criteria, 40 (90.9%) were classified as nutritional risk or moderate malnutrition, and the remaining two (4.6%) as severe malnutrition.

BIA measurements were mostly reduced or in the lower range of normality, as usual in moderately malnourished individuals. Total body fat was relatively low (21.2% of body weight), but proportionally less affected than lean body mass and body cell mass, consistent with the relatively acceptable BW and BMI. It is worth mentioning that total body water and other fluid compartments (intra and extracellular water) were virtually normal, as patients were investigated shortly after hemodialysis, and were therefore free from excess body fluids.

Elevated body resistance is unexpected in malnutrition, but can be explained by absence of fluid retention as well as by partially conserved fat mass, both of which impair electric conduction and increase resistance (Table 1).

Table 1 Bioimpedance analysis profile of the population

		Reference range (16 + 1)
Variable	Results	Males	Females
Body weight (kg)	61.4 ± 9.2
Resistance (ohm)	582.6 ± 104.7	470 ± 67	570 ± 88
Reactance (ohm)	60.6 ± 22.3	65 ± 15	76 ± 16
Phase angle (degree)	6.0 ± 2.4	7.8 ± 1.3	7.2 ± 1.1
Total body fat (kg)	13.0 ± 7.4	15.2 ± 7.8	27.2 ± 13.6
Lean body mass (LBM) (kg)	48.4 ± 11.8	BW – Total body fat
Body cell mass (kg)	21.7 ± 5.3	25.2 ± 5.4
Total body water (TBW) (L)	34.6 ± 7.0	50–60% BW**	45–55% BW**
Intracellular water (L)	19.8 ± 4.8	Approximately 60% TBW**
Extracellular water (ECW) (L)	14.6 ± 4.9	Approximately 40% TBW**
ECW/LBM (L/kg)	0.30 ± 0.06	0.36 ± 0.08^15
TBW/LBM (L/kg)	0.71 ± 0.04	0.75 ± 0.06^15

*Average values according to manufacturer's guidelines and references. **Some non-validated normal ranges are included as indicative proportions only.

Total energy ingestion was low (1261.7 ± 601.4 kcal/day or 20.7 ± 6.7 kcal/kg/day), but protein in the diet was quite acceptable (1.2 ± 0.6 g/kg/day).

Profile of serum albumin, transferrin, ferritin, and total lymphocyte count is displayed in Table 2. Lipid fractions, total cholesterol (CHOL), HDL, LDL, and triglycerides (TRIG), respectively, are shown in Table 3.

Table 2 General biochemical determinations

Variable	Mean (range)	Normal	Moderate change	Severe change
ALB g/L	36 ± 7 (13–50)	>40: 20.5%	30–40: 65.9%	<30: 13.6%
(g/dL)	3.6 ± 0.7 (1.3–5.0)	>4.0: 20.5%	3.0–4.0: 65.9%	< 3.0: 13.6%
TRANS umol/L	25.2 ± 6.8 (12.7–41.3)	45–68: 0	27–45: 36.4%	<27: 63.6%
(ug/dL)	141 ± 38 (71–231)	250–380: 0	150–250: 36.4 %	<150: 63.6%
FERR ug/L	770 ± 874 (5–5280)	200–2000: 79.5%	<200: 18.2%	>2000: 2.3%
LYMPH/mm3	1471 ± 550 (393–3374)	>1500: 38.6%	1500–900: 54.5%	<900: 6.8%
CRP mg/L	2.5 ± 5.2 (0.1–27.5)	<0.8: 59.1%	0.8–5.0: 27.3%	>5.0: 13.6%

ALB: Serum albumin; TRANS: Transferrin; FERR: Ferritin; LYMPH: Total lymphocyte count; CRP: C-reactive protein.

For FERR, <200 was considered compatible with iron deficiency and >2000 with iron overload.

Table 3 Plasma lipids

Variable	Mean (range)	Normal	Increased	Reduced
CHOL mMol/L	4.6 ± 1.3 (2.6–8.4)	(4.2–5.2): 36.4%	20.5%	43.1%
mg/dL	175 ± 50 (99–319)	(160–200): 36.4%	20.5%	43.1%
HDL mMol/L	0.9 ± 0.3 (0.02–2.5)	>1.1: 18.2%	0	81.8%
mg/dL	34.9 ± 11.6 (7.7 ± 97.0)	>40: 18.2%	0	81.8%
LDL mMol/L	2.9 ± 0.9 (1.3–5.3)	<3.4: 56.8%	43.2%	0
mg/dL	110.7 ± 34.4 (49.7–202.5)	<130: 56.8%	43.2%	0
TRIG mMol/L	2.0 ± 1.2 (0.7–5.5)	(1.1–1.7): 31.8%	18.2%	50.0%
mg/dL	178 ± 107 (62–444)	(100–150): 31.8%	18.2%	50.0%

CHOL: Total cholesterol; HDL: HDL-cholesterol; LDL: LDL-cholesterol; TRIG: Triglycerides.

The variables exhibited result halfway between those of BMI, with mostly normal individuals, some overweight, and just one-quarter below the acceptable range, and those of SGA, in which virtually all were abnormal. Between 20–40% of the group maintained adequate serum albumin and lymphocyte counts, however, transferrin was unexpectedly low in all tests. Plasma lipids were below expectations only when total cholesterol and especially HDL were examined. The other fractions were either normal or increased.

Linear Regression Analysis

Correlations between clinical and biochemical features were investigated concerning phase angle, nutritional assessment, comorbidities, and other parameters (Tables 46). Phase angle was positively associated with several co-variates of healthy body composition and nutritional status, such as transferrin and body cell mass, with correspondingly elevated negative association concerning extracellular water, a traditional indicator of fluid accumulation and malnutrition. Nevertheless, correlation with comorbidities, which may also increase extracellular water, or CRP, did not occur.

Table 4 Correlations of phase-angle

Variable	Pearson index (“r”)	Significance
TRANS	0.362	p < 0.05
TRIG	0.383	p < 0.05
BCM	0.458	p < 0.01
ICW	0.457	p < 0.01
ECW	−0.637	p < 0.0001

TRANS: Transferrin; TRIG: Triglycerides; BCM: Body cell mass; ICW: Intracellular water; ECW: Extracellular water.

Table 5 Correlations of C-reactive protein

Variable	Pearson index (“r”)	Significance
COMORB	0.367	p < 0.05
TRANS	−0.398	p < 0.01
ALB	−0.565	p < 0.001
SGA	0.545	p < 0.001

COMORB: Comorbidities; TRANS: Transferrin; ALB: Serum albumin; SGA: Subjective global assessment.

Table 6 Principal associations of comorbidities

Variable	Pearson index (“r”)	Significance
CRP	0.367	p < 0.05
SGA	0.333	p < 0.05
Resistance	−0.406	p < 0.01
ECW	0.323	p < 0.05
ECW/TBW	0.318	p < 0.05

CRP: C-reactive protein; SGA: Subjective global assessment; ECW: Extracellular water; TBW: Total body water.

CRP increased in parallel with comorbidities and SGA, but had opposite behavior concerning ALB and TRANS. Comorbidities followed changes in CRP and SGA, but exhibited links with bioimpedance profile as well.

Interestingly, time on hemodialysis directly correlated to FERR and indirectly to CHOL, but not to other indices of nutritional status or body composition.

Infectious Events

Patients with microbial troubles (Group I) were compared to the remaining cases (Group II) regarding general as well as specific results. The two populations were similar regarding age, gender, BMI, and clinical profile. Time on hemodialysis was numerically different, but could not be statistically confirmed. The only exceptions were comorbidities and C-reactive protein.

In fact, 97.1% (34/35) of infection-free subjects had CRP under 30 mg/L, whereas 77.8% (7/9) of the patients in Group I displayed values above this threshold. When the normal range is considered (<8 mg/L), then the proportions were, respectively, 22.9% (8/35) and 100% (9/9).

General Correlations in Infected Subjects

Despite the small number of patients in this group, regression analysis confirmed a few correlations in this population that were different from those shown in infection-free cases. Correlation between phase angle and CRP is not included because it failed to reach statistical significance. Results can be found in Table 7.

Table 7 Correlations of infected patients

Variables	Pearson index (“r”)	Significance
CRP X TSF	0.642	p < 0.05
CRP X lean body mass	−0.666	p < 0.05
Comorb. X reactance	−0.672	p < 0.05
Comorb X ECW/LBM ratio	0.740	p < 0.01
Comorb X ECW/TBW ratio	0.724	p < 0.01
Comorb. X time on dialysis	−0.672	p < 0.05
Time on dialysis X	−0.813	p < 0.001
Extracellular water

CRP: C-reactive protein; TSF: Triceps skinfold; ECW: Extracellular water; LBM: Lean body mass; TBW: Total body water.

DISCUSSION

Hemodialysis patients face a 25% annual mortality rate, with 50% of reported deaths attributed to cardiovascular disease.[18] The second cause of demise in this condition is classically infection, but malnutrition underlies many of these complications.[19–21]

Bioimpedance analysis in HD is not endorsed by all groups because major fluid shifts, and also lack of appropriate software in first-generation devices, tend to interfere with the reliability of the method. Multi-frequency BIA or spectroscopy has been suggested instead.[15],[16] Nevertheless, our experience with standard whole-body BIA in HD and other critical subjects, as well as that of others[22–24] in the last 10 years, have supported its routine use, with valuable insight gained on body composition changes even in highly susceptible populations. In the present context (Table 1), findings were broadly consistent with published experience in well-maintained HD patients, including ECW/LBM, TBW/LBM, and ECW/TBW ratios.[15],[16],[22],[23]

Recent investigations have called attention to the prognostic properties of phase angle as a clinically relevant marker of malnutrition positively associated with pre-albumin and albumin and negatively with comorbidities. More importantly, it is a strong signal for long-term mortality, including cardiovascular deaths.[24],[25] It was possible to demonstrate that phase angle is a surrogate measurement of nutritional status, positively correlating with transferrin, triglycerides, intracellular water, and body cell mass, all of which tend to be well-conserved in healthy subjects. High negative correlation with extracellular water, that indicates the opposite situation, further confirms the pattern.

Given the fact that malnutrition is an independent predictor of mortality in hemodialysis patients, the reported prognostic ability of phase angle is perfectly conceivable, but cannot be affirmed on the basis of this experience, as mortality was not documented. Still, given the various significant correlations, phase angle may be used as an alternative marker of nutritional status in HD populations.

SGA revealed only 36% of normal findings, with 51% displaying mild malnutrition and 13% advanced nutritional deficit. Older patients and those with more comorbidities were the most affected, but reduced body weight, hypoalbuminemia, and increased CRP were also significant. These observations are in agreement with those of a multicenter clinical trial[2] involving 1,000 patients, in which all anthropometrics were reduced in chronic HD cases, with the exception of diabetics who were overweight. Duration of HD and comorbidities were demonstrated to have an impact on these findings.

In the present study, BMI was recorded as too low in slightly less than one-quarter of the population, but another quarter was overweight. When subjective nutritional assessment was employed, nearly all patients were outside the normal range, confirming that both objectively and subjectively inappropriate nutritional status is a major problem in this population.

When serum albumin, transferrin, and total lymphocyte count were jointly employed, as little as 20–40% of the values corresponded to the acceptable range.

Hepatic CRP synthesis is upregulated by inflammation. Nevertheless, elevated CRP values are frequently found in the absence of apparent infection or inflammation.[17–19] In contrast, it has been recorded that massively increased CRP (>20 mg/L) mainly reflected the presence of infection/inflammation.[21] In the current protocol, 100% of infected individuals displayed high CRP concentrations in contrast with 22.9% of patients free from septic abnormalities, however, a somewhat lower threshold was observed. As alluded to in Table 7, their mean concentration was just 7 ± 11 mg/L.

One remarkable clinical message was that when acute infection was ruled out, CRP became normal or nearly normal. In other words, visible microbial problems accounted for the majority of aberrant findings of that variable, even though most never reached the cited 20 mg/L cut-off point[21] or 25 mg/L as recommended by others.[26]

It may therefore be stated that a permanent screening effort toward inconspicuous or inaccurately diagnosed infectious foci seems warranted in chronic renal failure. Individuals with elevated C-reactive protein are the prime candidates, especially addressing arteriovenous fistulas or grafts. These were the principal sources of inflammatory/infectious stress in this experience. Comparable considerations have already been forwarded by others,[26],[27] albeit in the form of selected patients, not in a consecutive series.

A tentative recommendation for daily practice could thus be advanced in the case of stable chronic hemodialysis patients, based on current findings (Figure 1).

Figure 1 Occult infection screening for hemodialysis patients.

CONCLUSIONS

1. Malnutrition profile was rather unique, with relatively favorable objective findings (body mass index, plasma proteins) and more altered subjective global assessment.

2. Bioimpedance analysis suggested that phase angle could be used as an indicator of nutritional status.

3. C-reactive protein was negatively associated with plasma proteins.

4. Infected subjects, although small in number, were responsible for most C-reactive protein determinations above 8 mg/L.

5. Investigation of occult infectious foci is advised in these circumstances.