ISRNM PROCEEDINGS Accelerated or Out of Control: The Final Months on Dialysis Peter Kotanko, MD,* Jeroen Kooman, MD, PhD,† Frank van der Sande, MD, PhD,† Franz Kappel, PhD,‡ and Len Usvyat, PhD* , § Mortality rates in dialysis patients are as high as 20% per year. Recent epidemiologic evidence indicates common patterns in biological and clinical indicators before death. Blood pressure, serum albumin levels, C-reactive protein, body weight, and other indicators change, at a population level, at an accelerated rate of change months before death. The escalation of inflammation (as indicated by a surge of C-reactive protein) appears to be a central event. The etiology of blood pressure decline before death is unclear. Although progressive heart failure is well documented in dialysis patients and a potential cause for blood pressure decline, adaptive functional changes need to be considered. In the general population, an inverse relationship between body weight and blood pressure is well described, and it is reasonable to hypothesize that the blood pressure decline before death is an adaptive response to a decline in body weight. Comparable trajectories of biological indicators before death are also documented in nonrenal illnesses, such as chronic obstructive pulmonary disease and certain malignancies. Descriptive analysis of clinical and laboratory variables clearly indicates an accelerated time course in most instances, pointing toward a loss of regulatory functions in a number of physiological subsystems. One can think of 2 reasons for that, (a) important control loops are damaged by the long-term effects of the disease, so that the system cannot react in a proper fashion to perturbations; and (b) the regulatory functions are still properly working, but are confronted with sit- uations, which were of no relevance in the evolution of the system. More extensive research into qualitative and quantitative aspects of physiological control mechanisms is required to answer the question if the observed changes before death are indicative of failed control mechanisms and/or adaptive processes. Ó 2014 by the National Kidney Foundation, Inc. All rights reserved. Observations in Dialysis Patients F IFTEEN TO 20% of chronic dialysis patients die within 1 year, a mortality rate comparable with malignancies such as colon and breast cancers. The leading causes of death are cardiovascular disease and infections, with sudden cardiac death being the single most frequent cause. 1 Recent data from the United States suggested that death in chronic hemodialysis (HD) patients is preceded by a series of characteristic events. 2 To further explore these findings on an international scale, the Monitoring Dialysis Outcomes (MONDO) initiative was founded. 3 Currently, MONDO encompasses data from dialysis patients from more than 40 countries. 4 The findings reported in US patients were subsequently corroborated in the diverse international MONDO cohorts. 5 Now, what exactly do these findings show? First, there is a gradual deterioration in hemodynamic, nutritional, and in- flammatory indicators (Figs. 1-3). Second, indicators start to deteriorate as long as 18 months before death. Third, the deterioration of indicators accelerates with the time of death approaching; in other words, the rate of change over time (one can think about the rate of change of a variable as its first derivative) is not constant and shows for most regions of the world and most biological variables studied so far a pattern of acceleration (in other words, the a nonzero second derivative of the variable over time). Although the absolute levels of indicators may differ sub- stantially between regions of the world, it is important to note that the dynamics of change are comparable. This point is illustrated by the fact that predialysis systolic blood pressure (preSBP) a year before death is around 150 mm Hg in the United States and 130 to 135 mm Hg in Latin Amer- ica and Europe; however, the rate of change of preSBP in these 2 regions is almost identical when compared with the United States (Fig. 2). Similar observations were made for serum albumin and interdialytic weight gain normalized to postdialysis body weight. Of note, these parameters show somewhat different dynamics in HD patients from South East Asia, in who the dynamics of preSBP and serum albu- min before death show a less pronounced decline. 5 Is there a common underlying etiology, which would explain these dynamics changes? The escalation of inflamma- tion appears to be a central event. It is instructive to study the * Renal Research Institute, New York, New York. † Department of Nephrology, University Hospital Maastricht, Maastricht, The Netherlands. ‡ Institute for Mathematics and Scientific Computing, University of Graz, Graz, Austria. § Fresenius Medical Care North America, Waltham, Massachusetts. Financial Disclosure: P.K. holds stock in Fresenius Medical Care, a leading dialysis provider. Address correspondence to Peter Kotanko, MD, Renal Research Institute, 315 East 62nd Street, 4th Floor, New York, NY 10065. E-mail: pkotanko@rriny.com Ó 2014 by the National Kidney Foundation, Inc. All rights reserved. 1051-2276/$36.00 http://dx.doi.org/10.1053/j.jrn.2014.06.011 Journal of Renal Nutrition, Vol 24, No 6 (November), 2014: pp 357-363 357