In early November, the United States Renal Data System released its 2017 Annual Data Report, which coincided with the beginning of the American Society of Nephrology’s Kidney Week.

The annual data report comprises two volumes: one with data about chronic kidney disease; and another with data about end-stage renal disease (ESRD). The latter includes a wealth of analyses about the incidence of ESRD, mortality and hospitalization among both dialysis and transplant patients, Medicare expenditures and prescription drug utilization.

Inside this year’s annual data report are numerous trends that merit further analysis and discussion. From a purely clinical perspective, trends in mortality and hospitalization are key markers of population health. Unfortunately, trends are developing in both domains that all in the kidney community should find alarming.

Patient survival: New vs old

The mortality rate among patients on contemporary dialysis in the United States is lower than that among historical patients. In fact, relative to patients on dialysis in 2001, the adjusted mortality rate among patients in 2015 was 28% lower. However, this long-term improvement obscures a more recent development: Between 2014 and 2015, the adjusted mortality rate among patients on dialysis increased slightly from 165.1 to 166.1 deaths per 100 patient-years. This is the first year-over-year increase in mortality in this century. Year-over-year increases were apparent in both patients on hemodialysis at 0.5% and patients on peritoneal dialysis at 1.8% (Click figure below to enlarge). In addition, while the adjusted mortality rate in patients on dialysis with vintage of more than 2 years fell by 1.1% between 2014 and 2015, rates in patients with vintage of 2 to 4 years and at least 5 years rose by 1.4% and 3.2%, respectively. In the latter group, the increase in the mortality rate sets the population back to the mortality of comparable patients in 2011 to 2012.

Decline in hospitalization rates

The hospitalization rate among contemporary patients on dialysis in the United States is also lower than among historical patients. Relative to patients on dialysis in 2004, the adjusted rate among patients in 2015 was 18% lower. However, after rather impressive year-over-year declines of 4%, 4.6% and 2.7% from 2011 to 2012, 2012 to 2013 and 2013 to 2014, respectively, the rate fell only 0.9% between 2014 and 2015. The story was similar in the large subset of patients on hemodialysis, in whom the 1% decrease in the admission rate between 2014 and 2015 was the smallest improvement since an uptick between 2008 and 2009. In patients on peritoneal dialysis, the admission rate increased 0.7% between 2014 and 2015.

In both patients on hemodialysis and peritoneal dialysis, rates of cardiovascular-related hospitalization fell from 1% to 2% between 2014 and 2015. However, rates of infection-related hospitalization fell trivially in patients on hemodialysis and, worryingly, increased more than 5% in patients on peritoneal dialysis.

Potential factors of change

Clearly, long-term, downward trends in rates of death and hospitalization are decelerating in the U.S. dialysis population, and in the case of peritoneal dialysis, trends are reversing themselves for the first time in at least 15 years.

Why is this happening? There are no simple answers, so more research needs to be done to prove or disprove any hypothesis. For now, we can speculate that:

  • Because hemodialysis is utilized by 90% of U.S. patients, hemodialysis population outcomes are dialysis population outcomes. Across the landscape of hemodialysis, there is little evolution in the delivery of the therapy. This lack of movement can be readily seen in the Dialysis Outcomes and Practice Pattern Study (DOPPS) Practice Monitor. In the last 7 years, mean session length has increased less than 5 minutes. Today, only 27% of patients dialyze for at least 4 hours, despite the 2014 consensus statement that “fluid removal should be gradual and dialysis treatment duration should not routinely be less than 4 hours without justification based on individual patient factors.” Although the practice monitor covers a relatively small sample of patients, data about mean session length in that sample are largely congruent with corresponding data in the recently reported TIME trial. Of course, treatment frequency in the dialysis facility is nearly always three sessions per week. Unsurprisingly, markers of fluid volume and clearance are also nearly fixed. Mean pre-dialysis systolic blood pressure oscillates between 145 mm Hg and 148 mm Hg, and nearly 30% of patients present pre-dialysis systolic blood pressure of 160 mm Hg or greater. Mean serum phosphorus circles around 5.3 mg/dL.
  • The vintage of the peritoneal dialysis population is increasing. In patients with prevalent peritoneal dialysis in 2010, mean numbers of years already on continuous ambulatory peritoneal dialysis and automated peritoneal dialysis were 1.2 and 1.1 years, respectively. In comparison, corresponding means in 2015 were 1.8 and 1.6 years, respectively. This shift in vintage translates to more patients who are in their third, fourth and fifth years of peritoneal dialysis, when complications related to ultrafiltration failure and fluid overload are more likely. The shift can be traced to population dynamics: The advent of bundled payment in Medicare spurred aggressive growth in peritoneal dialysis utilization, but subsequent shortages in solution curtailed that growth.
  • We lack high-quality evidence from large randomized clinical trials to prove that widely used cardioprotective medications are efficacious in patients on dialysis. Nevertheless, trends in the use of beta blockers and renin-angiotensin system (RAS) inhibitors are troubling. According to the DOPPS Practice Monitor, beta blocker use in 2014 dipped as low as 67% after exceeding 70% in 2012 and 2013. According to the ADR, use in patients with a history of myocardial infarction and in patients with heart failure is not greater than in the dialysis population. Use of RAS inhibitors is sharply declining, with less than 39% of patients in the DOPPS Practice Monitor on medication at the end of 2014 compared with more than 48% in the middle of 2010. The use of newer medications is limited or uncertain. The use of direct oral anticoagulants in patients on dialysis is low, and applications of sacubitril/valsartan are unknown. Thus, there appears to be no growth in use of either old or new pharmacologic therapies for cardiovascular disease.
  • Infection remains a major challenge. Its incidence results in expensive hospitalizations and likely confers additional cardiovascular risk to patients on dialysis. In patients with peritoneal dialysis, in whom peritonitis is the foremost cause of infection-related hospitalization, any upward trend in burden is alarming. Nephrologists Transforming Dialysis Safety (NDTS), a program funded by CDC and led by the American Society of Nephrology, has shone a bright light on infection control in the dialysis facility. Are we also maintaining focus on infection control in the home setting?
  • In October 2012, the federal government began to reduce payments to hospitals with excess readmissions within 30 days of discharge. The 30-day readmission rate in the dialysis population is high compared to the broader Medicare population. Of course, patients on dialysis are frequently admitted for myocardial infarction, heart failure and pneumonia. Is it possible that hospital initiatives to limit 30-day readmissions have actually harmed patients on dialysis, in part by substituting emergency and observation care for hospitalization? New data suggests this hypothesis is not far-fetched. In a study published in JAMA Cardiology, the 30-day risk-adjusted readmission rate after heart failure hospitalization (in the broader Medicare population) declined from 20% before payment penalties to 18.4% in the payment penalty era. Meanwhile, the 30-day risk-adjusted mortality rate increased from 7.2% to 8.6%. Data from the Peer Kidney Care Initiative has vividly shown upward trends in utilization of observation care in the dialysis population. There is an urgent need to understand whether such shifts in care from inpatient to outpatient settings are helping or harming the dialysis population.

The challenges implicit in these hypotheses are not new to the kidney community. The limitations of volume management with conventional hemodialysis (and long-term use of peritoneal dialysis), resistance to use of cardioprotective medications, persistent infections, and the interplay between dialysis facilities and hospitals are long-standing obstacles. With myriad signs of neutral and negative trends in clinical outcomes among patients on dialysis, now is the time to face these challenges with renewed dedication and ingenuity, not only in the clinical arena, but also in the political arena, where policy is crafted. How will the kidney community respond?

References:
Dember LM, et al. Primary results of the Time to Reduce Mortality in End-Stage Renal Disease Trial: A pagmatic trial demonstration project of the NIH Health Care Systems Research Collaboratory. Presented at: Kidney Week; Oct. 31-Nov. 5, 2017; New Orleans.

Gupta A, et al. JAMA Cardiol. 2017; doi:10.1001/jamacardio.2017.4265.

Weiner DE, et al. Am J Kidney Dis. 2014;doi:10.1053/j.ajkd.2014.07.003.

Wetmore JB. Insights from the 2016 Peer Kidney Care Initiative Report: Still a ways to go to improve care for dialysis patients. Am J Kidney Dis. doi:10.1053/j.ajkd.2017.08.023.

For more information:

Eric Weinhandl, PhD, MS, is a clinical epidemiologist and biostatistician at NxStage Medical and an adjunct assistant professor in the department of pharmaceutical care and health systems at the University of Minnesota. He has studied the epidemiology of the dialysis patient population for 13 years. Allan Collins, MD, FACP, is professor of medicine at the University of Minnesota, Hennepin County Medical Center. He serves as the senior medical advisor to NxStage Medical, based in Lawrence, Massachusetts. Disclosures: Weinhandl and Collins report no relevant financial disclosures.