Inflammation and Overweight in Peritoneal Dialysis: Is There an Association?

Abstract

More than 30% of the patients on peritoneal dialysis show chronic systemic inflammatory activity with high levels of C-reactive protein. The purpose of this cross-sectional study was to investigate the influence of the inflammatory state on clinical and nutritional markers in patients on peritoneal dialysis. Twenty-seven patients were included: mean age was 57.6 ± 19 years, 48% were male, and median time on peritoneal dialysis was 16.0 (8.3; 35.8) months. Clinical, dialytic, laboratory, anthropometric and electric bioimpedance data were collected with the sample stratified for C-reactive protein. In patients, the levels of Interleukin-6 and tumor necrosis factor-α were higher, while adiponectin levels were lower than in healthy individuals (p ≤ 0.001), indicating the presence of inflammatory activity in the sample. When compared to patients with C-reactive protein <1 mg/dL, those with ≥1mg/dL showed higher body mass index (29.4 ± 6.1 vs. 24.4 ± 4.5 kg/m²; p = 0.009), percent of standard body weight (124.5 ± 25.4 vs. 106.8 ± 17.9 %; p = 0.012), and percent of body fat as assessed by both anthropometry (31.3 ± 9.9 vs. 23.9 ± 9.1%; p = 0.056) and bioimpedance (38.9 ± 6.3 vs. 26.2 ± 12.6 %; p < 0.001). Patients with C-reactive protein ≥1mg/dL also exhibited higher levels of ferritin (701 ± 568 vs. 532 ± 356 ng/mL; p =0.054) and lower total lymphocyte count (median 1838 vs. 1638 mm³; p = 0.001). In conclusion, higher body mass index and body fat markers were associated with C-reactive protein ≥ 1mg/dL, and higher C-reactive protein was associated with immunocompetence impairment evidenced by the lower total lymphocyte count. Our findings confirm the relationship between inflammation, body fat, and immunocompetence, which may be superimposed potentializing the inflammatory status.

Keywords:

• dialysis
• nutritional state
• inflammation
• body fat
• overweight
• lymphocytes count

INTRODUCTION

More than 30% of the patients on peritoneal dialysis (PD) show chronic systemic inflammatory activity with high levels of C-reactive protein (CRP).[1,2] Attributed causes include reduced renal cytokine clearance and residual renal function, accumulation of advanced glycosylation end products, chronic heart failure, atherosclerosis per se, peritonitis, persistent infections, fluid overload, and incompatibility with the peritoneal dialysis solution.[3,4] Furthermore, an increased adipose tissue mass may induce adipocyte hypoxia stimulating angiogenesis and inflammation. This can be observed through the elevation of tumor necrosis factor-alpha (TNFα), Interleukin (IL)-6, IL-8, IL-10, alpha-1-acid glycoprotein (α1AG), and CRP, which characterize a subclinic inflammatory state in obesity.[5]

In a study by Pennell et al.,[6] excessive weight gain (greater than 10%) was observed in 20% of the cases after PD was started. Energy intake secondary to dialysate glucose absorption contributes for the occurrence of obesity in PD patients. However, this nutritional condition can be also influenced by other factors, such as peritoneal membrane transport characteristics and genetic characters that can alter metabolic rate.[7]

Some studies have suggested that overweight is associated with improved survival in patients on dialysis, characterizing reverse epidemiology. However, this has been more frequently observed in patients on hemodialysis rather than those on PD, and this benefit effect can be limited only to those with high fat-free mass.[8–12]

Despite the large body of literature on inflammation in dialysis, the relationship between nutritional status and inflammation in chronic PD remains unclear. Considering that inflammation and overweight are associated, the purpose of this study was to further investigate the influence of CRP on nutritional markers.

MATERIALS AND METHODS

Twenty-seven patients with chronic renal failure were included in this cross-sectional study. Initial evaluation started with 45 patients attending the Dialysis Center of Botucatu Medical School Hospital, São Paulo State University, UNESP, São Paulo, Brazil, with 18 of them excluded for the following reasons: 2 refused to participate, 1 was diagnosed with cancer, 2 were hepatitis C-positive, 3 had been on dialysis for less than four months, 7 showed acute or chronic infection, and 3 were younger than 18 years. None of the patients presented positive serology for HIV or Hepatitis B. This study was approved by the local Ethics Review Board.

Considering CRP ≥ 1mg/dL as indicative of inflammatory activity, cases were allocated into two groups: CRP ≥ 1mg/dL (n = 12) and CRP < 1mg/dL (n = 15). Data collected at enrollment included information on age; gender; race; underlying renal disease; time on PD; PD mode; smoking; systolic and diastolic blood pressure; and presence of diabetes mellitus (DM), arterial hypertension, and systolic dysfunction by echocardiography (ejection fraction <30%).

Nutritional assessment was performed under fasting conditions with the abdominal cavity empty of dialysis solution. A Lange® adipometer (Beta Technology®, Santa Cruz, California, USA), an inextensible tape, and Filizola Personal® digital scales with a stadiometer (Filizola®, Campo Grande, Brazil) were used to measure weight, height, and skinfold thicknesses and circumferences on the non-dominant side of the body. Body mass index (BMI) was calculated as weight (Kg) divided by height (m)²; percent body fat (%BF) was estimated from the sum of thickness of the biceps, triceps, and subscapular and iliac skinfolds[13,14]; percent of standard body weight in relation to ideal body weight (%SBW) was calculated considering the Metropolitan Life Insurance tables adapted by Grant et al.[15]; percent of standard mid arm circumference (%MAC), mid arm muscle circumference (%MAMC), and tricipital skinfold thickness (%TST) were calculated based on the National Health and Nutrition Examination Survey (NHANES) percentile distribution tables adapted by Frisancho.[16] MAMC was calculated using the equation, MAC-TSF × 0.314.

Bioimpedance (BIA) measurements were made using a Biodynamics® device (model 450, Biodynamics Corporation®, USA). A single frequency current (800 μA, and 50 Khz) was carried out to assess resistance, reactance, phase angle, %BF, total body water (TBW), and extracellular water (ECW).

The assessment of energy-protein intake was based on three-day dietary records using the software Nutwin®- UNIFESP, taking into account the energy secondary to dialysate glucose absorption. Evaluating protein intake also considered the protein nitrogen appearance (PNA) by the following equation[17]:

Patients reporting regular, little, or very little appetite during the previous seven days were considered inappetent.[18] Subjective global assessment (SGA) with a seven-point scale was done, considering a score ≤ 5 as malnutrition.

Laboratory investigations included albumin, cholesterol, creatinine, urea, total lymphocytes, and transferring and venous bicarbonate. As inflammatory indicators, CRP (immunochemistry), ferritin (immunoenzyme), and α1AG (immunoturbidimetry) were measured considering the reference values used in our laboratory; TNFα, IL-6, and adiponectin (ELISA) were also measured, considering as normal the levels measured in 20 volunteers healthy adult individuals, once this cytokines did not have a standard value of normality. Renal function, dialysis dose, and peritoneal ultrafiltration were assessed by collecting the total fluid output (urine and ultrafiltrate) during the 24 hours preceding collection.

Data were expressed as percent, mean and standard deviation, or median and quartiles. Patients with CRP ≥1 mg/dL were compared with those with CRP<1 mg/dL by covariance analysis (ANCOVA), adjusted for presence of DM, systolic dysfunction, and numerical variables previously correlated by Spearman or Pearson or Mann-Whitney's test. Proportions were compared by the chi-square test or the exact test of Fisher. p ≤ 0.05 was considered significant level.

RESULTS

Table 1 shows the general characteristics of the sample. Mean age was 57.6 years, and 92.6% were white. Median time on PD was 16 months; 55.5% of patients were on automated peritoneal dialysis and had fractional plasma urea clearance of 2.2 (1.9; 2,5). Regarding primary diagnoses, 26% presented diabetic nephropathy, 33.3% glomerulonephritis, and 22.2% hypertensive nephropathy. Around 15% smoked. DM was present in 33.3%, and arterial hypertension in 74% of the patients, while systolic dysfunction, diagnosed by echocardiography, was observed in 14.8%. Mean TBW was 29.7 (50.1% in the extracellular compartment). Almost 70% of the patients presented albumin lower than 3.5 g/ dL, and 52% showed BMI higher than 25 Kg/m².

Table 1 General characteristics of the sample

Type	Variables	n = 27
Demographic	Age (years)	57.6 ± 19.0
	Gender (% male)^‡	48.1
	Race (% white)‡	92.6
Nutritional	Malnutrition (SGA) (%)	18.5
	BMI (Kg/m^2)	26.7 ± 5.7
	%BF (anthropometry)	27.2 ± 10.0
	%BF (BIA)	31.8 ± 12.0
	Phase angle (°)	4.9 ± 1.1
	PNA (g/Kg)^†	0.8 (0.6; 1.3)
	Proteín/W (g/Kg)	1.1 ± 0.4
	TEI/W (Kcal/Kg)	31.2 ± 6.8
	Albumin (g/dL)	3.3 ± 0.5
Inflammatory	CRP (mg/dL)^†	0.4 (0.1; 2.3)
	IL-6 (pg/mL)^†	409.0 (356.0; 506.0)
	TNF-α (pg/mL)^†	304.0 (249.0; 349.0)
	Ferritin (ng/mL)^†	494.0 (312; 755.5)
	α1GA (mg/dL)	99 ± 39.4
	Adiponectin (ng/mL)	205.3 ± 66.5

*Data expressed as mean ± standard deviation.

^†Data expressed as median (quartiles).

^‡Data expressed as percentage of the total sample.

Abbreviations: SGA = subjective global assessment, BMI = body mass index, %BF = body fat percent, PNA = protein nitrogen appearance, Protein/W= protein intake/weight, TEI/W = total energy intake (including glucose absorbed from dialysate)/weight, CRP = C-reactive protein, IL-6 = Interleukin-6, TNF-α = tumor necrosis factor α, α1GA = α1 acid glycoprotein.

Eleven of the 27 patients showed CRP higher than 1 mg/ dL. Inflammatory cytokine levels in dialysis patients significantly differed from those in healthy individuals (p ≤ 0.001), with IL-6 and TNFα being 2-fold higher and adiponectin 2.2-fold lower.

Table 2 shows the anthropometric variables compared with those obtained by BIA when the sample was stratified for CRP. Patients with CRP ≥ 1mg/dL had higher BMI (29.4 ± 6.1 vs. 24.4 ± 4.5 kg/m²; p = 0.01), %SBW (124.5 ± 25.4 vs. 106.8 ± 17.9%; p = 0.01), and resistance (582.1 ± 110.2 vs. 541.5 ± 130.3 Ω; p = 0.01), which might have resulted from the higher %BF observed by both anthropometry (31.3 ± 9.9 vs. 23.9 ± 9.1%; p = 0.05) and BIA (38.9 ± 6.3 vs. 26.2 ± 12.6%; p < 0.001).

Table 2 Comparison of nutritional variables according to inflammatory state

	Inflammatory state
Variables	CRP < 1 mg/dl(n = 15)	CRP ≥ 1mg/dl(n = 12)	p value
Anthropometric
Weight (kg)	61.7 ± 14.2	69.1 ± 15.2	0.37
BMI (Kg/m^2)	24.4 ± 4.5	29.4 ± 6.1	0.01
%SBW	106.8 ± 17.9	124.5 ± 25.4	0.01
%MAC^†	88.3 (82.8; 102.5)	97.4 (88.3; 122.9)	0.21
%MAMC	95.2 ± 15.5	108.2 ± 25.1	0.31
%TST	79.9 ± 51.6	89.1 ± 45.3	0.24
%BF	23.9 ± 9.1	31.3 ± 9.9	0.05
Bioimpedance
Resistance (Ω)	541.5 ± 130.3	582.1 ± 110.2	0.01
Reactance (Ω)	43.6 ± 12.3	51.5 ± 12.6	0.13
Phase angle (°)	4.7 ± 1.2	5.1 ± 0.9	0.65
%BF	26.2 ± 12.6	38.9 ± 6.3	<0.001
TBW (L)	32.8 ± 9.1	28.9 ± 6.5	0.43
ECW (% TBW)	50.1 ± 6.0	50.0 ± 3.7	0.88
Protein-energy intake
PNA (g/Kg)	1.02 ± 0.6	1.07 ± 0.5	0.68
Protein/W (g/Kg)	1.16 ± 0.4	1.03 ± 0.26	0.61
TEI/W (Kcal/Kg)	31 ± 7.6	31.4 ± 5.8	0.25
Inappetence (%)^‡	20	16.7	1.00^§
SGA (% Malnutrition)^‡	26.7	8.3	0.34^§

*Data expressed as mean ± standard deviation.

^†Data expressed as median (quartiles).

^‡Data expressed as percentage of the stratified sample.

^§Fischer's exact test; all other variables were compared by ANCOVA.

Abbreviations: CRP = C-reactive protein; BMI = body mass index; %SBW = percent of standard body weight in relation to ideal body weight; %MAC, %MAMC, %TST = percentage of standard mid-arm circumference, mid-arm muscle circumference, and tricipital skinfold thickness, respectively; %BF = percent body fat; TBW = total body water; ECW = extracellular water; PNA = protein nitrogen appearance; Protein/W= protein intake/weight; TEI/W = total energy intake/weight.

Comparison of laboratory variables is shown in Table 3. In patients with CRP ≥ 1 mg/dL total lymphocyte count was lower, with a median of 1638 (1287; 2227) vs. 1838.5 (1650; 2362)/mm³, p < 0.001), while ferritin (701.1 ± 568.0 vs. 532.6 ± 356.3 ng/dL; p = 0.05), α1GA (108.4 ± 50.6 vs. 91.5 ± 26.5 mg/dL; 0.08), hematocrit (37.7 ± 6.1 vs. 34.2 ± 5.2 %; p = 0.02), and hemoglobin (12.7 ± 1.3 vs. 11.5 ± 1.9 g/dL; p = 0.02) were higher.

Table 3 Comparison of laboratory, nutritional, and inflammatory markers according to inflammatory state

	Inflammatory state
Variables	CRP<1 mg/dl(n = 15)	CRP≥1mg/dl(n =12)	p value
Nutritional markers
Albumin (g/dL)	3.3 ± 0.7	3.3 ± 0.4	0.99
Cholesterol (mg/dL)	170.2 ± 35.3	150.7 ± 23.9	0.22
Creatinine (mg/dL)	8.9 ± 3.7	6.6 ± 3.1	0.54
Urea (mg/dL)	82.7 ± 30.1	74.4 ± 26.9	0.81
Venous bicarbonate (mmol/L)	26.5 ± 3.5	28.1 ± 3.3	0.24
Iron (μg/dL)	75.1 ± 40.7	74.6 ± 26.5	0.77
Transferrin (g/L)	1.75 ± 0.4	1.59 ± 0.4	0.39
Hematocrit (%)	34.2 ± 5.2	37.7 ± 6.1	0.02
Hemoglobin (g/dL)	11.5 ± 1.9	12.7 ± 1.3	0.02
Total lymphocytes count (/mm^3)^†	1838.5 (1650.3; 2362)	1638 (1287; 2227)	<0.001^‡
Inflammatory markers
TNF-α (pg/mL)	312.4 ± 90.9	302.5 ± 110.7	0.90
Ferritin (ng/dL)	532.6 ± 356.3	701.1 ± 568.0	0.05
Adiponectin(ng/mL)	192.5 ± 65.6	221.3 ± 66.9	0.85
IL-6 (pg/mL)	454.7 ± 139.1	429.8 ± 148.6	0.82
α1GA (mg/dL)	91.5 ± 26.5	108.4 ± 50.6	0.08

*Data expressed as mean ± standard deviation.

^†Data expressed as median (quartiles).

^‡Mann-Whitney; all other variables were compared by ANCOVA.

Abbreviations: CRP = C-reactive protein, TNF-α = tumor necrosis factor α, IL-6= Interleukin-6, α1GA= α1 acid glycoprotein.

No other differences related to CRP were observed between variables.

DISCUSSION

In this study, patients with CRP ≥ 1mg/dL showed higher BMI, %SBW, and %BF, parameters that may be pointing to adipose tissue role in the pathogenesis of inflammation.

Obesity, particularly visceral adiposity, has been associated with subclinic inflammation, given that adipose tissues produce inflammatory cytokines and CRP.[19] The relationship between BMI and inflammation has been demonstrated. According to Wang et al.,[20] BMI was higher in PD patients with higher CRP values. Other studies verified the influence of body composition on mortality by observing that death risk was lower in patients with high BMI, although this protective effect was limited to those patients with normal or high muscle mass.[9,12]

Visser et al.,[21] studying 16,616 healthy individuals, found that obesity (BMI > 30 kg/m²) elevated CRP levels 2.1 times in men and 6.2 times in women as compared to their eutrophic counterparts (BMI <25 Kg/m²), and that abdominal visceral fat, assessed by the waist/hip ratio, was positively associated with systemic inflammation in both genders. Similarly, Pannacciulli et al.[22] suggested that waist circumference, as an indicator of body fat central distribution, can predict CRP levels more strongly than total fat determined by BIA or BMI. According to Axelsson et al.,[23] the presence of inflammation (CRP ≥ 1 mg/dL) was associated with higher total fat mass and truncal fat mass estimated by DEXA in patients with chronic renal failure. In this study, abdominal visceral fat was not assessed. The waist/hip ratio was not used because we believe that the constant infusion of dialysis solution into the peritoneal cavity promotes abdominal distension, super-estimating these values, even when assessment is performed with the peritoneal cavity empty of dialysis solution.

Assessing the impact of obesity on PD outcomes has yielded conflicting results. Some authors, such as Snyder et al.,[24] found that survival was longer among overweight patients. Others encountered no difference in survival between obese and non-obese patients,[25,26] while others reported that obesity was independently associated with death and technique failure during PD treatment.[8]

In our sample, patients with CRP ≥ 1mg/dL showed lower total lymphocyte values. Carvounis et al.[27] observed that total lymphocytes was significantly lower in nonsurvivors than in survivors (1277 ± 143 mm³ vs. 2249 ± 236 mm³; p = 0.0036), demonstrating a correlation between total lymphocyte count and mortality. Ates et al.[28] suggested that although total lymphocyte count can be used as a prognostic tool in PD patients, other factors, such as hydration status, should as well be taken into account. The association between CRP ≥ 1mg/dL and lower total lymphocyte count observed here confirms the relationship of this immunological marker with inflammation, in agreement with other studies associating it with nutritional status impairment.[29,30]

In this study, plasma CRP was considered an inflammation marker, as it has been widely used in research and because of its relationship with nutritional parameters as well as morbidity and mortality in patients with chronic renal failure.[31] However, no relationship between CRP and IL-6 or TNFα was observed. This might be due to differences in half-life between CRP and the cytokines, or to the fact that a single CRP measurement was performed (variability ranges from 10.6–63%), or even to the CRP cutoff point employed.[24] Moreover, the collection site and the detection method used are key points for the determination of inflammatory activity in the analysis of inflammatory mediators.[32]

The reduced levels of adiponectin observed in our PD patients are not in line with those reported in some other studies,[33,34] but agree with those found by Guebre-Egziabher et al.,[35] who suggested that increased levels of TNFα and IL-6 inhibit adiponectin production and favor inflammation and atherosclerosis, contradicting an antiinflammatory role for this cytokine but supporting the idea that the cardiovascular risk in these patients is elevated.

Our patients showed elevated ferritin levels, which were higher in those with CRP ≥ 1 mg/dL. This finding associated with the α1GA levels observed support the relationship of these markers with inflammation, leading us to agree with Kalantar-Zadeh et al.,[36] who suggested that high levels of ferritin and CRP point to inflammatory activity regardless of iron status.

Causes of inflammation in PD patients are varied, and there is not much information is this regard for Latin-American patients. In conclusion, despite the small size of the sample (which might have limited data interpretation), our findings confirm the relationship between body fat, immunocompetence, and inflammation, considered as prognostic markers in other studies.

ACKNOWLEDGMENT

The authors are thankful to FAPESP and CAPES for their financial support and Mariza Branco for English support.