Studies of Unexplained Inflammation in Haemodialysis and the Role of Endotoxin

Abstract

Clinical outcomes for patients with end-stage kidney disease on haemodialysis remain poor and have similar mortality risk to some cancers. A significant proportion of haemodialysis patients have chronic unexplained inflammation – the pathophysiology of this syndrome is unclear but is associated with poor quality of life and survival. A large number of studies have proposed that circulating endotoxin, fragments of gram negative bacterial cell wall, plays a key role in driving chronic inflammation in dialysis patients in the absence of clinical infection. However, interpretation of published literature is difficult for several reasons. Firstly, reports have mainly been cross-sectional in nature. Secondly, various different types of endotoxin detection assays have been used and reported blood endotoxin levels in clinically stable non-infected patients were exceedingly high which is incompatible with the clinical state. Additionally, nearly all studies used endotoxin detection assays that could be also be activated by (1-3)-β-D glucan, a cell wall constituent of cereals, yeast and fungi. This is a particularly important given that materials used to construct some dialyser membranes are known to leach (1-3)-β-D glucan. The doctoral research project had two central hypotheses requiring investigation. Firstly, that endotoxin and other exogenous molecules in the gut including (1-3)-β-D glucan are contributors to unexplained inflammation in patients with chronic kidney disease. Second, that their presence in blood is a result of translocation from the gut due to increased gastrointestinal permeability in the uraemic state. The clinical importance of these hypotheses is that if the relationship between presence of endotoxin and (1-3)-β-D glucan in blood of patients with chronic kidney disease and inflammation were clarified, and a positive association was confirmed, this would justify embarking on endotoxin or (1-3)-β-D glucan lowering interventional studies such as extracorporeal endotoxin adsorption therapy or exploring therapies to modify gut flora. In order to address these hypotheses systematically, the studies were designed with the following specific aims: 1. To determine the prevalence of chronic unexplained inflammation in patients receiving haemodialysis to determine the scope and extent of the clinical problem. 2. To determine the optimum endotoxin assay, in terms of accuracy and precision, for measurement of endotoxin in uraemic human blood and to limit interference of other factors in the assay such as presence of (1-3)-β-D glucan. 3. To determine the relationship of blood levels of endotoxin measured with an optimised assay and (1-3)-β-D glucan in patients with chronic kidney disease with patient symptoms and inflammation both cross-sectionally and longitudinally. 4. To determine factors influencing the presence of endotoxin and (1-3)-β-D glucan in the blood of patients requiring haemodialysis to determine whether increased gut permeability, potentially induced by intravascular volume depletion during dialysis, is a potential contributor. To determine the extent of unexplained inflammation, a retrospective study was performed of 444 prevalent haemodialysis patients at the East and North Herts NHS Trusts in which medical records and serial CRP measurements over 3 months were reviewed. Specific potential causes of inflammation were sought for each patient. 64.8% of patients had evidence of elevated CRP >5mg/L on at least two occasions one month apart and 23.4% of the study population had chronic unexplained inflammation. To determine the optimum blood endotoxin detection assay, a literature review was conducted to identify candidate assays for testing. All published endotoxin studies in dialysis patients used variations of assays based on the Limulus Amoebocyte Lysate (LAL) assay. The kinetic turbidimetric, kinetic chromogenic LAL and Endotoxin Scattering Photometry assay were compared in a direct head-to-head study in terms of accuracy and precision. The kinetic turbidimetric LAL assay was found to have the best performance and used for blood endotoxin detection in clinical studies. In a study of 40 haemodialysis patients, blood samples were measured simultaneously for endotoxin and (1-3)-β-D glucan. Endotoxin signals correlated significantly with (1-3)-β-D glucan. However, on modifying the kinetic turbidimetric LAL assay with a blocker which prevents assay activation by (1-3)-β-D glucan, all previously detected signals were extinguished and true endotoxemia in haemodialysis patients in this study was low, contrary to findings from previous studies. ‘Endotoxin’ signals detected using LAL assays suffer from significant false positive interference from elevated (1-3)-β-D glucan in haemodialysis patients. Given the low prevalence of endotoxemia detected, haemodialysis patients were further studied during clinical states that were hypothesised to increase endotoxin influx including during exercise and after haemodialysis. Blood samples were collected from 10 haemodialysis patients before and after haemodialysis and intra-dialytic exercise. Similarly, after modifying the LAL assay with a blocker to prevent false activation from (1-3)-β-D glucan, only one sample tested positive for endotoxin whereas 43% of samples had elevated (1-3)-β-D glucan. The relationship between (1-3)-β-D glucan with kidney function, inflammation and endotoxin levels were further studied in a study of 60 haemodialysis patients (30 with high-risk features including chronic unexplained inflammation, hypotension during haemodialysis or high ultrafiltration requirements and 30 low-risk patient [no inflammation, hypotension and low ultrafiltration requirement], 15 peritoneal dialysis patients, 20 patients with CKD 4-5, 20 patients with CKD 1-3 and 20 healthy controls. Similarly, while endotoxin signals were highest in haemodialysis patients, most signals disappeared after modifying the LAL assay with a (1-3)-β-D glucan blocking buffer. Low level endotoxemia was found in only 3 haemodialysis patients. BG levels increased with worsening CKD stage and were highest in haemodialysis patients, 22% having significantly elevated levels. (1-3)-β-D glucan correlated strongly with endotoxin signals (r=0.545, p<0.001) suggesting that previous reports of elevated endotoxin may be artefactual due to raised blood (1-3)-β-D glucan in renal impairment. (1-3)-β-D glucan correlated strongly with markers of inflammation and were associated with higher depression and kidney disease related symptoms scores. Intestinal barrier dysfunction and increased intestinal permeability has been proposed as a source of inflammation in haemodialysis patients and the haemodialysis procedure has been suggested to exacerbate intestinal permeability due to the large blood volume changes that occur during treatment. Intestinal permeability has not been studied directly in haemodialysis patients since conventional assays rely on measurement of probes in urine but this method cannot be applied to anuric dialysis patients. Using a novel intestinal permeability assay which measures probes in plasma instead, intestinal permeability was measured in 10 haemodialysis patients before and after haemodialysis treatment and compared with 5 healthy controls. Direct comparison of intestinal permeability between haemodialysis patients and healthy controls was difficult since concentration of probes in plasma was influenced by level of kidney function, although there was a suggestion that small intestinal permeability was increased in haemodialysis patients. However, intestinal permeability was not significantly affected by the haemodialysis procedure. The main finding of this work is that endotoxemia is relatively uncommon in dialysis patients and hence unlikely to be the sole cause of unexplained chronic inflammation in these patients. Previous reports of apparent high levels of blood endotoxin in dialysis patients were due, in part at least, to false positive interference of LAL assays by elevated (1-3)-β-D glucan present in the blood of patients with kidney disease. The work presented in this thesis supports this. Furthermore, the clinical studies carried out did not demonstrate a convincing association between endotoxin and inflammation. In contrast, this work has demonstrated a progressive increase in levels of (1-3)-β-D glucan with worsening degree of kidney disease and an association of these levels with inflammation, higher depression and kidney-disease related symptom scores. The cause of elevated (1-3)-β-D glucan in dialysis patients is unclear and warrants further study, though an intestinal source is a possibility. The direct intestinal permeability studies carried out in dialysis patients suggest impaired intestinal barrier function, although until an accurate method of assessing intestinal permeability in vivo in uraemic patients is available, this hypothesis remains to be proven.