**5. Malnutrition-inflammation complex**

302 Progress in Hemodialysis – From Emergent Biotechnology to Clinical Practice

Evaluation, and Treatment of High Blood Cholesterol in Adults (Adult Treatment Panel III) Final Report, 2002) have established that lipids play a significant role in the progression of CVD and with reductions in cholesterol levels there is a similar and graded reduction in risk (Kalantar-Zadeh, et al., 2003). Dyslipidemia characterized in HP is primarily associated with hypertriglyceridemia, low HDL concentrations, elevated levels of LDL, elevated LDL particle numbers, a higher propensity of smaller and denser LDL particles which contain high levels of residual triglycerides, and elevated levels of lipoprotein (a) (Vaziri, 2009). Cholesterol metabolism in HP can be significantly altered by dialysis protocol, lipid controlling medication, malnutrition, and inflammation (Montazerifar, Hashemi, Karajibani, & Dikshit, 2010). HDL can be significantly reduced in HP due to a reduced plasma concentration of ApoA-I and ApoA-II (Bowden, Hebert, Wilson, Gentile, & Lanning, 2007; Vaziri, 2009), reduced transporter proteins such as ABCA-1 (Vaziri, 2009), and the downregulation of enzymatic process associated with HDL maturation (Malgorzewicz et al., 2010). Down-regulation also occurs with HDL paraoxonase which is associated with impaired protection against oxidative stress. Hypertriglyceridemia also occurs in HP, is normally associated with impaired VLDL and is the most prominent dyslipidemic abmormality reported in 70% of HP (Eisenhardt, Habersberger, & Peter, 2009). Additionally, hypertriglyceridemia is associated with increases in IDL, chylomicrons, and chylomicron remnants postprandial (Chmielewski, et al., 2008). Primarily a reduction in lipoprotein lipase, an enzyme associated with binding to VLDL and the release of the corresponding triglycerides, is associated with changes that cause a concomitant increase in plasma triglycerides that seems to correspond with a decrease in VLDL receptors. Finally, LDL in HP is normally more atherogenic as particles are usually smaller, denser and are more likely to be oxidized and engulfed by macrophages leading to more unstable arterial plaques (Kaysen, 2009). Oxidation can further cause LDL to no longer be recognizable to LDL receptors on the cell causing less deposition of LDL in intracellular pools. Less deposition can lead to the same amount of circulating LDL cholesterol being associated with smaller and denser LDL particles which can carry more risk for CVD. Excess LDL cholesterol is then removed from circulation primarily by being engulfed by macrophages which lead to more atherogenic foam cells leading to risk acceleration. Lipoprotein (a) is a LDL-like lipid with higher levels associated with increased risk of CVD and is associated with overproduction of apolipoproteins, specifically apoB. Apolipoprotein B is bound with LDL and is associated

Though cholesterol metabolism in HP is associated with dyslipidemia and many CVD related deaths are associated with elevated levels of lipoproteins, recent evidence suggests that many HP have normal or reduced plasma cholesterol levels that are associated with a higher rate of mortality (Bowden & Wilson, 2010). CVD related mortality accounts for 40- 45% of all-cause mortality in HP suggesting factors other than lipids may be associated with CVD in this patient population (Dungan, et al., 2007; Tsirpanlis, et al., 2009). Counterintuitive outcomes regarding the role cholesterol may play in disease progression in HP may be affected by chronic inflammation and malnutrition with much of the published literature concerned with MIC (Bowden & Wilson, 2010; Chavalitdhamrong, et al., 2007). The term, though controversial, that has been used in the literature associated with this counter-intuitive finding is reverse epidemiology (Dungan, et al., 2007). Experimental

with increased numbers of LDL particles.

**4.1 Reverse epidemiology** 

Cano et al. (2009) has reported that between 20-60% of HP may have MIC and that patients also have lower body weights, lower BMI, lower albumin levels, lower blood pressure and elevated CRP (Diepeveen, et al., 2008). Albumin has been specifically mentioned as a marker of nutritional status and identified as one of the criteria used to measure PEM which is associated with MIC. The association between hypoalbuminemia (< 3.9 mg/dL) and mortality has been well-established and albumin has been identified as a strong predictor of cardiovascular disease (Kaysen, 2009). Supporting this theory is the fact that most studies reporting on reverse epidemiology report no counter-intuitive findings with HDL. The cholesterol esters in HDL are primarily received from albumin (Vaziri, 2009) and therefore hypoalbuminemia would be associated with low HDL. This is also one possible explanation for why other lipoproteins would be low, yet still associated with a significantly higher mortality rate in HP. Challenging our understanding of how this might impact mortality and cholesterol is that albumin is also used as a measure of inflammation and may not necessarily be related a nutritional etiology (Trivedi, et al., 2009). Finally, CRP along with IL-6 are inflammatory markers that have been associated with higher rates of mortality as well (Bowden & Wilson, 2010).

### **5.1 Reverse epidemiology and cholesterol**

In a subset of HP a decrease in baseline cholesterol levels, excluding HDL, has been associated with decreases in CVD mortality and all-cause mortality when patients have chronic levels of inflammation and poor nutritional status (Krane et al., 2009). Recent study authors have suggested the need to consider both malnutrition and inflammation simultaneously as albumin has been shown to be affected by inflammation levels. Inflammatory cytokines and acute phase reactants have been reported to decrease appetite, reduced albumin, increase catabolism (Liu et al., 2004) and quite possibly cause lower levels of cholesterol suggesting that low albumin levels may simply be a reflection of inflammation rather than malnutrition. Specifically most lipoproteins seem to be affected by MIC with the exception of HDL. In most studies where HDL was either low or unaffected most patients experience hypoalbuminemia and low protein intake in the diet (Kilpatrick, et al., 2007). Krane et al. (2009) in a study of 1,229 Type II diabetics who were HP reported that high levels of CRP strongly predicted all-cause mortality, sudden death and myocardial infarction when cholesterol levels were both low and high. Yet, when CRP levels were high and cholesterol levels were low, relative risks for CVD related deaths were even greater. It was further reported that CRP level and not LDL level was the variable more likely to predict risk for mortality and cardiovascular events. Another study confirmed the existence of reverse epidemiology in the presence of high inflammation suggesting the inflammation better predicted both morbidity and mortality (Tsirpanlis, et al., 2009). The study authors also reported that nutritional cachexia along with low cholesterol were good nutritional indices for malnutrition and may be the reason low cholesterol levels are associated with

Malnutrition, Inflammation and Reverse Epidemiology in Hemodialysis Patients 305

term dialysis vintage, making LDL smaller and denser and is associated with more endothelial dysfunction, inflammation, stenosis and intima media thickness. Therefore, the oxidative stress hypothesis may help to further explain reverse epidemiology. Oxidative stress in combination with MIC could cause lower levels of LDL and LDL particle numbers, but more atherogenic particles that can accelerate mortality and morbidity. Yet, those associations do exist, a causal relationship between oxidative stress, CVD, and reverse

Chmielewski (2008) reports on an endotoxinlipoprotein hypothesis that suggests higher levels of cholesterol may be beneficial in HP. The theory suggests that higher levels of lipoproteins more readily bind bacterial lipopolysaccharide or endotoxins and modulate inflammatory immune responses. The authors continues to state that cholesterol levels and their ability to predict CVD fall on a continuum with lower and higher levels associated with mortality (Kalantar-Zadeh, 2007). Liu, et al. (2004) may help support this theory as they have reported a U-shaped curve regarding cholesterol levels and moratlity in HP. This

Finally, as mentioned previously, HP patients have been reported to have smaller, dense LDL particles that are more oxidized which may play a stronger role than LDL. Small LDL particle size has been identified as an emerging risk factor for CVD (*Third Report of the National Cholesterol Education Program Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults (Adult Treatment Panel III) Final Report*, 2002). A previous study (Bowden, Griggs, Wilson, & Gentile, 2009) has demonstrated that LDL particle number and LDL size can classify more HP at risk when compared to LDL and triglycerides. Since more HP have small, dense LDL particles there may be mortality associated with smaller LDL particles even when LDL particle number is lower. Therefore, though HP may have less LDL cholesterol and less of the smaller LDL particles, because smaller LDL particles are more atherogenic, more patients have increased morality with decrease levels

Reverse epidemiology has been associated with cholesterol but also body mass index (BMI), systolic and diastolic blood pressure, homocysteine, and creatinine levels (Balakrishnan & Rao, 2007) in HP suggesting the need for novel treatments. Since the counter-intuitive findings are prevalent in HP additional approaches are necessary to help identify novel treatments and to discover if new therapies are warranted. Though counter-intuitive findings are present in HP as well as advanced age, congestive heart failure, malignancies, and AIDS little data exists to support changes in non-traditional risk factors as a means to control CVD. To help increase albumin levels it has been proposed to increase protein content in the diet. But levels of protein augmentation with HP patients on chronic hemodialysis has not been sufficiently elucidated simply because dietary protein can be a significant source of uremic toxins and increase phosphate levels which can have deleterious effects on the health of HP patients (Stolic, 2010). It should be noted that hypercaloric consumption may take several years and even decades for serious health effects to occur, yet hypocaloric consumption, measured by albumin levels in HP patients, normally causes a more rapid deterioration in health. Combined with high inflammation levels, a decreased appetite and hypocaloric levels may cause a significant decrease in cholesterol but accelerated risk for mortality (Kalantar-Zadeh, et al., 2001). This short-term effect of

epidemiology has yet to be established suggesting the need for further study.

theory is highly speculative and needs further study.

of cholesterol.

**6. Treatment** 

higher levels of mortality in HP. A more recent study reported on the cross-sectional association with MIC and the effects on cholesterol in HP (Bowden & Wilson, 2010). When comparing cholesterol levels, those classified as having hypoalbuminemia had a reverse epidemiological effect with LDL particle number. In the same study when HP were classified as having high inflammation, based on CRP, LDL, VLDL, and LDL particle number reported counter-intuitive findings. But, in patients classified as having both hypoalbuminemia and high inflammation, all lipid variables, with the exception of HDL, reported a reverse epidemiological effect. This suggests that though inflammation may play a more prominent role with reverse epidemiology, both inflammation and malnutrition can have a pronounced effect on lipids causing low cholesterol levels to be associated with more mortality. Additionally, other study authors have reported inverse associations with cholesterol and all-cause mortality, but also a U-shaped relationship with cholesterol in the presence of malnutrition and inflammation (Liu, et al., 2004). When malnutrition and inflammation were controlled for in the analysis, a strong, graded and positive association existed between high cholesterol levels and mortality further supporting the thought that reverse epidemiological associations with cholesterol in HP is associated with MIC. An additional review paper has suggested that previous studies that have controlled statistically for factors associated with MIC in HP, risk associated with cholesterol was the same as the general population and further speculates that cholesterol may be an additional marker of malnutrition (Chmielewski, et al., 2008). It should be noted that not all studies agree with these findings.

Another theory behind why low cholesterol in HP patients can be associated with high mortality has been called survival selection. It is well-established that most chronic kidney disease patients will not survive to kidney failure and End-Stage Renal Disease. Presently, it is suggested that only 10% (Trivedi, et al., 2009) of patients will live long enough to initiate dialysis. Therefore, since so few survive it is thought that only those who have a strong genetic predisposition for survival may be more likely to live but may also have a poorer risk profile. Additionally, simply stated, HP may not live long enough to die of the consequences of traditional CVD risk factors. Though this theory is not very new, it has yet to be confirmed with well controlled studies that take into consideration the age of the patients. Kalantar-Zadeh et al. (2003) also suggests that reverse epidemiology may be normal and that over-nutrition is primarily a 20th and 21st century phenomenon and that our understanding of traditional Framingham risk factors may in fact be new and the exception. Though over-nutrition is a problem in many countries it has been associated with longer living populations in Western nations. As one would expect this idea is highly controversial among scientist and nephrologists.

Additional study authors (Nanayakkara & Gaillard, 2010) have reported reverse epidemiology in HP concerning lipids could possibly be due oxidative stress. Reactive oxygen species (ROS) production that is not balanced by antioxidant control is associated with oxidative stress. Furthermore, oxidative injury has been reported to alter lipids in both the general population and HP and is involved in CVD acceleration (Diepeveen, et al., 2008). Though HP are normally supplemented with B vitamins and folic acid, many still have deficiencies in antioxidants with many patients shifting to a more pro-oxidative stress profile. High oxidized LDL is a well established comorbidity in both HP and CKD patients and is associated with small, dense LDL particles. Moreover, modification of LDL through oxidation is thought to be the first step in the development of CVD and specifically atherosclerosis. Oxidative stress occurs routinely in HP, especially in patients with a longterm dialysis vintage, making LDL smaller and denser and is associated with more endothelial dysfunction, inflammation, stenosis and intima media thickness. Therefore, the oxidative stress hypothesis may help to further explain reverse epidemiology. Oxidative stress in combination with MIC could cause lower levels of LDL and LDL particle numbers, but more atherogenic particles that can accelerate mortality and morbidity. Yet, those associations do exist, a causal relationship between oxidative stress, CVD, and reverse epidemiology has yet to be established suggesting the need for further study.

Chmielewski (2008) reports on an endotoxinlipoprotein hypothesis that suggests higher levels of cholesterol may be beneficial in HP. The theory suggests that higher levels of lipoproteins more readily bind bacterial lipopolysaccharide or endotoxins and modulate inflammatory immune responses. The authors continues to state that cholesterol levels and their ability to predict CVD fall on a continuum with lower and higher levels associated with mortality (Kalantar-Zadeh, 2007). Liu, et al. (2004) may help support this theory as they have reported a U-shaped curve regarding cholesterol levels and moratlity in HP. This theory is highly speculative and needs further study.

Finally, as mentioned previously, HP patients have been reported to have smaller, dense LDL particles that are more oxidized which may play a stronger role than LDL. Small LDL particle size has been identified as an emerging risk factor for CVD (*Third Report of the National Cholesterol Education Program Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults (Adult Treatment Panel III) Final Report*, 2002). A previous study (Bowden, Griggs, Wilson, & Gentile, 2009) has demonstrated that LDL particle number and LDL size can classify more HP at risk when compared to LDL and triglycerides. Since more HP have small, dense LDL particles there may be mortality associated with smaller LDL particles even when LDL particle number is lower. Therefore, though HP may have less LDL cholesterol and less of the smaller LDL particles, because smaller LDL particles are more atherogenic, more patients have increased morality with decrease levels of cholesterol.
