The advent of understanding a possible primary role for systemic, micro-inflammatory processes in the pathobiology of renal failure likely found its root with the observation that serum albumin concentration was the strongest associate of mortal risk among dialysis patients1,2. The finding was interpreted to suggest protein-calorie malnutrition. That was supported by contemporaneous findings that high serum creatinine, not low, was associated with a risk benefit for dialysis patients1,2. The same surprising pattern was observed for serum cholesterol concentration. Serum creatinine results from a balance between its rates of removal (by dialysis and renal function) and generation (that reflects muscle mass). Higher concentration suggested greater generation. Furthermore, the concentration of albumin and creatinine is highly correlated among dialysis patients1,2,3. The finding that larger fractions of patients' ideal body weight were also favorably associated with survival when evaluated both jointly and severally with albumin, creatinine, cholesterol, and other substances3,4 bolstered the interpretation. Though consensus-based guidelines continue to regard serum albumin concentration as reflecting nutritional status among patients5, it may also decline as part of an acute phase inflammatory process (APIP)6,7 that is part of the body's normal response to diseases such as infection, osseous fracture, and so forth.
DISCOVERING THE MEANING OF ALBUMIN
Against that early background, a meeting of invited experts from selected specialties of internal medicine and surgery was sponsored by National Medical Care, Inc. (NMC, now Fresenius Medical Care or FMC, NA) in early 1995 as part of its quality assurance (CQI) program. Each expert presented a short (less then 10 min) summary of his or her speculations about possible causes of or contributing factors to malnutrition among dialysis patients with special reference to serum albumin concentration. Dr. Bruce Bistrian commented that it would be difficult to starve patients down to an albumin concentration of 3.2 gm/dL, say, and presented a physiologic summary of the APIP. Dr. Jonas Bergström commented that his group had recently noted elevated C-reactive protein (CRP) among patients presenting for dialysis at his hospital and that CRP seemed associated with future survival of patients. Those observations were subsequently presented (abstract; Bergström et al; J Am Soc Nephrol 6:573, 1995.
Representatives from NMC, then a large provider of dialysis services supported by a clinical laboratory, requested performance of a CRP (at no money charge) on a random sample of routine monthly blood samples from dialysis patients. Those were performed during June and July 1995. Dr. George Kaysen, working independently, adopted a different approach9: He measured albumin kinetics in dialysis patients, finding that its synthesis rate was reduced and was inversely associated with the concentration of acute phase proteins such as CRP. The findings implied a pathobiology characterized by relative downregulation of albumin synthesis and upregulation of the synthesis of acute phase proteins. The reports from Kaysen9, Bergström (abstract; ibid), and the CQI report to NMC physicians10 appeared in late 1995 and early 1996.
The NMC material suggested that CRP was commonly elevated11,12. Substantial correlation among concentrations of serum albumin, creatinine, blood hemoglobin, and CRP (inverse with the other three) were observed. There was strong, inverse correlation between two measures of iron metabolism: ferritin and iron binding capacity. A more refined multivariable evaluation of the many correlations observed in the data, factor analysis13, suggested that four primary domains were represented12. The first was as a nutritional or body protein domain that contained strong, direct contributions from albumin, pre-albumin, creatinine, hemoglobin, iron binding capacity, and total lymphocyte count (TLC). It contained a strong inverse contribution from CRP. Weaker inverse contributions from total neutrophil count (TNC) and reticulocyte count were also observed. The second was an iron domain that contained direct contributions from serum iron and ferritin. The third was an inflammation or blood cell domain that contained strong contributions from TNC, TLC, and platelet counts and weaker ones from reticulocyte count and CRP. The fourth domain reflected azotemia and contained strong direct contributions from blood urea nitrogen, potassium, phosphorus, and creatinine concentrations and an inverse contribution from serum bicarbonate concentration12.
The findings were interpreted to imply a biological domain reflecting body protein stores with which the APIP marker, CRP, was strongly and inversely associated. The WBC (white blood count) types appeared in their own domain but were shared with body protein such that TLC was directly, and TNC inversely, associated with it.
Further analysis of the NMC information suggested that the concentrations of albumin, creatinine, pre-albumin, iron binding capacity, and TLC were all favorably associated with the survival of patients, whereas TNC was unfavorably associated12,13. The association for CRP was unfavorable but only marginally so. Albumin, creatinine, and TLC were favorably associated with survival in fully saturated statistical models, but CRP was not. However, reports suggesting otherwise, that CRP was important in multivariable survival models with or without significant contributions from serum albumin, were available8,14.
NUTRITION VERSUS INFLAMMATION
There has evolved some debate about which was more important to the survival of patients, an APIP (i.e., CRP), or a nutritional process (i.e., albumin, creatinine, etc.). Kaysen has recently evaluated both serum albumin and creatinine concentrations in terms of proxies for inflammation (CRP) and nutritional intake (the normalized protein catabolic rate, nPCR), concluding that both were important15. nPCR was favorably associated with the concentration of both substances. CRP was also unfavorably associated with both.
Substantial evidence appears suggesting that activation of the APIP unfavorably influences the nutritional status of patients and their survival. All phases of the process are mediated by cellular mechanisms7. The unfavorable association of WBC with survival was observed a decade ago3, but more recent studies suggest that the direction of association may be different for different cell types11,12. Therefore, we evaluated the association of WBC types with mortal risk among patients.
ASSOCIATIONS OF WBC TYPES WITH BODY PROTEIN STORES AND MORTAL RISK
The information was taken from the FMC (NA) data system that has been described before2,3,4. Complete data including differential count of the WBC types were available for 21,970 patients during the last three months of 1998. Average values for those and other measures were calculated for each patient during that period. Survival analyses (Poisson regression) were performed as functions of them for the 1999 calendar year.
Table 1 shows the distributions of WBC types and evaluates the correlation among them and with selected other measures. The distribution of WBC (mean = 6.9, SD = 2.6, 1st Decile = 4.4, Median = 6.6, 9th Decile = 9.9) revealed that most patients were well within the normal range (4.5 - 10.8 
103/
L). In general, there was a high degree of correlation among the granulocytes and between them and monocytes. TLC was less strongly associated with TNC but nonetheless was directly so. TNC and monocyte count were inversely associated with albumin, creatinine, and hemoglobin. TLC, however, was directly associated with those measures but less strongly so.
Table 1 - Distribution of WBC (103 cells/
L) types and correlation among types and with selected measures.
![Table 1 - Distribution of WBC (103 cells/[mgr]L) types and correlation among types and with selected measures - Unfortunately we are unable to provide accessible alternative text for this. If you require assistance to access this image, please contact help@nature.com or the author](/common/images/table_thumb.gif)
Full table Table 2 shows the association of WBC and its cell types with death risk after progressive adjustment for case mix attributes and other contemporaneous measures. TNC and monocyte count were strongly and unfavorably associated with mortal risk in the unadjusted models. Adjustment for other measures such as albumin, creatinine, and hemoglobin that are also associated with risk1,2,3 reduced but did not extinguish the associations. TLC was favorably associated with risk in all models. The strength of association was lower in the fully adjusted than in the unadjusted model, but the magnitude of improvement was increased from a 6.6% per 103 cells/L to nearly 20%.
Table 2 - Survival analyses for WBC and the WBC types with progressive adjustment for case mix and other measures.
Full table Figure 1 illustrates the risk profiles for TNC and TLC. All analyses of TNC suggested monotonic or near-monotonic deterioration of risk with increasing count. All analyses of TLC suggested monotonic improvement of risk associated with increasing count up to the highest count group. Thereafter, risk may have deteriorated marginally suggesting a possible reversed J-shape to the profile (null probabilities between the>2.00 and 1.75–2.00 groups were>0.10, =0.05, and>0.10 for the None, CM, and CM & Other analyses, respectively).
Figure 1.
Risk profiles of total neutrophil count and total lymphocyte count. Adjustments: none (black); case mix (dark hading); CM & other (light shading).
Full figure and legend (58K)These analyses are consistent with the earlier factor analyses11. Inflammatory cells like TNC are inversely associated with proxies for the body protein content and are directly associated with death risk. The proxies are inversely associated with risk. Statistical adjustment for the proxies reduced but did not extinguish the direct association between TNC and death risk. TLC, on the other hand, is inversely associated with risk and directly, if weakly, associated with proxies for body protein content. Therefore, both TNC and TLC are jointly and independently associated with death risk among dialysis patients, but the risk vectors for them are in opposite directions.
THE ROLES OF THE WBC TYPES
The unfavorable association of low TLC with future coronary events has been reported16. Owen12 has reported the favorable association of higher TLC with survival among dialysis patients. The findings could result simply form the known suppressive effect of glucocorticoids on TLC17, often elevated in an acute phase reaction7. TLC is often used as a proxy for visceral protein mass, and lower mass could be associated with both lower counts and greater risk. Albumin and TLC were directly associated in this analysis, but the strength of association was weak. Finally, lower TLC could be associated with defects of cellular immunity, humoral immunity, or both quite apart from the APIP. These data did not permit classification of lymphocytes into subtypes that would facilitate evaluation of such possibilities.
WBC, like CRP, appears associated with mortality due to coronary artery disease in the general population18, mortality following myocardial infarction19, and in patients with unstable angina20. Thus, systemic, micro-inflammatory processes appear to have pathobiologic roles in both ESRD and CHD patients.
The nutritional compromise associated with chronic inflammatory process among dialysis patients has received considerable attention5. The clinical dynamics, reduced skeletal mass, reduced albumin concentration, and so forth resemble cachetic syndromes21 seen with cancer, arthritis, inflammatory gastrointestinal disease, and chronic infection11,22 more than they do simple starvation21. The cytokine-mediated acute phase process seems active in all such syndromes7, as it is in dialysis23,24 and uremia24,25.
This phylogenetically old and non-specific process is the first level of host defense against injury from bacterial invasion or other tissue destruction7. Its biological purpose seems to shift the body's priorities from growth and maintenance to defense. It is mediated by a complex cellular signaling system6,7 at the core of which is the monocyte-derived tissue macrophage7. The APIP is usually self-limited, shutting down when the threat is neutralized and the injury repaired. It may be sustained and become chronic in such conditions as cancer, chronic infection, and uremia.
Professional phagocytes (granulocytes and monocyte/macrophages) are the effector cells in this system that have among their functions the destruction of microorganisms26. They release highly toxic agents outside of the cell that can attack malignant cells, normal tissue, and organisms too large to be ingested. Superoxide, hydrogen peroxide, and a number of halogenated species (e.g., hypochlorus acid and chloramines) are products of a peroxidase-H2O2-halide system that can kill, injure, or inhibit a variety of micro-organisms, mammalian cells (e.g., tumor, red blood cell, granulocytes), soluble mediators like bacterial toxins, and normal proteins27 such as albumin28 and lipoprotein14.
POSSIBLE CAUSES
Increased activity of inflammatory cytokines roughly proportional to the reduction of kidney function has been demonstrated in patients with chronic renal failure even before they start dialysis24,25. Thus it seems clear that inflammatory pathobiology is a part of uremia. The cause(s) are not clear and deserve substantial research. It is tempting to believe that some retained, toxic substance causes the syndrome, but minor, local tissue inflammation may contribute29 as well.
Clinicians have known for years that exposure of peripheral blood to biomaterials in dialysis systems activates leukocytes and complement30. Exposure to trace amounts of endotoxin can activate inflammatory mechanisms31. Thus, treatment by dialysis may have potential to cause damage by sustaining the activity of this systemic inflammatory state even as the treatment sustains life.
Recent studies at FMC (NA) have shown that patients treated using vascular access catheters and grafts have respectively 50% and 300% greater death risk than patients treated using fistulae (unpublished observations). The differences persisted even after adjustment for the case mix and other measures that included albumin, creatinine, and WBC (OR = 1.47 and 1.96, respectively, for grafts and catheters compared with fistulae). TNC was significantly higher and TLC significantly lower in patients with catheters than either grafts or fistulae. Furthermore, albumin and creatinine were significantly lower in persons with catheters than the other access types. Thus, recent changes of medical practice may have led to deleterious effects for patients in the United States because the use of catheters increased by more than 75% between 1996 and 199932. The increase may be contributing to greater prevalence of chronic inflammatory processes, reduced body protein content, and finally to higher mortal risk.
Thus, it seems clear that many events, processes, and substances may contribute to a chronic, systemic, micro-inflammatory state among dialysis patients. Those include possible "uremic" substances, local inflammation from renal or other tissue, exposure to biomaterials or environmental substances, and clinical interventions or practices. Furthermore, a sustained APIP could itself cause leukocyte-mediated tissue injury, further invoking the APIP in a self-enhancing cycle. It remains the task of the ESRD community to untangle the causative and effector chains of the syndrome and to fashion cures. In the meantime, clinicians should be mindful of the oath to do no harm by being sure that existing and evolving clinical practice does not make matters worse.
References
| 1. | Lowrie EG & Lew NL. Death risk in hemodialysis patients: The predictive value of commonly measured variables and an evaluation of death rate differences between facilities. Am J Kidney Dis 1990; 15: 458−482. | PubMed | ISI | ChemPort | |
| 2. | Lowrie EG & Lew NL. Commonly measured laboratory variables in hemodialysis patients: Relationships among them and to death risk. Semin Nephrol 1992; 12: 276−283. | PubMed | ISI | ChemPort | |
| 3. | Lowrie EG, Huang WH, Lew NL & Liu Y. The relative contribution of measured variables to death risk among hemodialysis patients. inDeath on Hemodialysis: Preventable or Inevitable? 1994; edited by Friedman E Hingham, MA, Kluwer Academic Publishers pp 121−141. |
| 4. | Kopple JD, Zhu X, Lew NL & Lowrie EG. Body weight-for-height relationships predict mortality in maintenance hemodialysis patients. Kidney Int 1999; 56: 1136−1148. | Article | PubMed | ISI | ChemPort | |
| 5. | Kidney Disease Outcome Quality Initiative/National Kidney Foundation. Clinical practice guidelines for nutrition in chronic renal failure. Am J Kidney Dis 2000; 35. |
| 6. | Gabay C & Kushner I. Acute phase proteins and other systemic responses to inflammation. N Engl J Med 1999; 360: 448−454. |
| 7. | Janeway CA & Travers T. Immunobiology 1994; New York, Garland Publishing. |
| 8. | Deleted in proof.. |
| 9. | Kaysen GA, Rathore V, Shearer GC & Depner TA. Mechanisms of hypoalbuminemia in hemodialysis patients. Kidney Int 1995; 48: 510−516. | PubMed | ISI | ChemPort | |
| 10. | LOWRIE EG, MA L, ZHANG J & LEW NL. Thoughts About Anemia, Iron, Proteins and a "Chronic Acute Phase" State Among Hemodialysis Patients: Memorandum to NMC Medical Directors. FMC (NA) ref no. 96−02−14 1996;. |
| 11. | Lowrie EG. Conceptual model for a core pathobiology of uremia with special reference to anemia, malnourishment, and mortality among dialysis patients. Semin Dial 1997; 10: 115−129. | ISI | |
| 12. | Owen WF & Lowrie EG. C-reactive protein as an outcome predictor for maintenance hemodialysis patients. Kidney Int 1998; 54: 627−636. | Article | PubMed | ISI | ChemPort | |
| 13. | TABACHNICK BG & FIDELL LS. Principle components and factor analysis. inUsing Multivariate Statistics 1989; ed 2 New York, Harper Collins. |
| 14. | Zimmerman J, Herrlinger S & Pruy A. et al Inflammation enhances cardiovascular risk and mortality in hemodialysis patients. Kidney Int 1999; 55: 648−658 10.1046/j.1523-1755.1999.00273.x. | PubMed | |
| 15. | Kaysen GA, Chertow GM & Adhikarla R. et al Inflammation and dietary protein intake exert competing effects on serum albumin and creatinine in hemodialysis patients. Kidney Int 2001; 60: 333−340. | Article | PubMed | ISI | ChemPort | |
| 16. | Zouridakis EG, Garcia-Moll X & Kaski JC. Usefulness of the blood lymphocyte count in predicting recurrent instability and death in patients with unstable angina pectoris. Am J Cardiol 2000; 86: 449−451. | Article | PubMed | ISI | ChemPort | |
| 17. | Omson SP, McMahon LJ & Nugent CA. Endogenous cortisol: A regulator of the number of lymphocytes in peripheral blood. Clin Immunol Immunopahtol 1980; 17: 506−514. |
| 18. | Brown DW, Giles WH & Croft JB. White blood cell count: An independent predictor of coronary heart disease mortality among a national cohort. J Clin Epidemiol 2001; 54: 316−322. | Article | PubMed | ISI | ChemPort | |
| 19. | Furman MI, Becker RC & Yarzebski J. et al Effects of elevated leukocyte count on in-hospital mortality following acute myocardial infarction. Am J Cardiol 1996; 78: 945−948. | Article | PubMed | ISI | ChemPort | |
| 20. | Cannon CP, McCabe CH & Wilcox RG. et al Association of white blood count with increased mortality in acute myocardial infarction and unstable angina patients. Am J Cardiol 2001; 87: 636−639. | Article | PubMed | ISI | ChemPort | |
| 21. | Kotler DP. Cachexia. Ann Intern Med 2000; 133: 622−634. | PubMed | ISI | ChemPort | |
| 22. | Beisel WR. Herman award lecture, 1995: Infection-induced malnutrition: From cholera to cytokines. Am J Clin Nutr 1995; 62: 813−819. | PubMed | ISI | ChemPort | |
| 23. | Kimmel PL, Phillips TM & Simmens SJ. et al Immunologic function and survival in hemodialysis patients. Kidney Int 1998; 54: 236−244. | Article | PubMed | ISI | ChemPort | |
| 24. | Nakanishi I, Moutabarrik A & Okada N. et al Interlukin-8 in chronic renal failure and dialysis patients. Nephrol Dial Transplant 1994; 9: 1435−1442. | PubMed | ISI | ChemPort | |
| 25. | Descamps-Latscha B, Herbelin A & Nguyen AT. et al Balance between IL-1 , TNF- , and their specific inhibitors in chronic renal failure and maintenance dialysis. J Imunol 1995; 154: 882−892. | ChemPort | |
| 26. | Klebanoff SJ. Oxygen metabolites from phagocytes. inInflammation: Principles and Clinical Correlates 1999; edited by Gallin JI, Snyderman R Philadelphia, Lippencott Williams & Wilkins. |
| 27. | Deschamps-Latsha B & Witko-Sarsat V. Importance of oxidatively modified proteins in chronic renal failure. Kidney Int 2001; 59 Suppl 78: S108−S113. |
| 28. | Himmelfarb J & McMonagle E. Albumin is the major plasma protein target of oxidant stress in uremia. Kidney Int 2001; 60: 358−63 10.1046/j.1523-1755.2001.00807.x. | Article | PubMed | ISI | ChemPort | |
| 29. | Ymalik N, Delilbasi L & Gulay H. et al The histological investigation of gingival biopsies from patients with chronic renal failure, renal transplants, and periodontitis: A light and electron microscopic study. J Periodontol 1991; 62: 737−44. | PubMed | |
| 30. | Toren M, Goffinet J & Kaplow L. Pulmonary bed sequestration of neutrophils during hemodialysis. Blood 1970; 36: 337−340. | PubMed | ISI | ChemPort | |
| 31. | Wratten ML, Tetta C, Ursini F & Sevanian A. Oxidant stress in hemodialysis: Prevention and treatment strategies. Kidney Int 2000; 58 Suppl 76: S126−S132. | Article | ISI | |
| 32. | U.S. Renal Data System. USRDS 2001 Annual Data Report: Atlas of End-Stage Renal Diseases in the United States Bethesda, National Institutes of Health National Institutes of Diabetes and Digestive and Kidney Diseases, pp 87−89. |
Acknowledgments
The author must acknowledge and thank Dr. Ming Teng and Ms. Nancy Lew for their contributions to the data management and statistical analysis of FMC (NA) information about the contributions of WBC types to mortal risk among dialysis patients.
