Prediction Tool for Mortality Risk among Older Adults Includes Seven Variables

Maintenance dialysis is the most commonly utilized modality for renal replacement therapy among older adults. In Canada, the proportion of incident dialysis patients ≥65 years of age rose from 41.8% in 1994 to 53.5% in 2013. Older adults on dialysis are at increased risk for poor outcomes, including death; the mortality rate in that patient population is as much as 37.0 deaths per 100 person-years within the first 6 months of initiation of dialysis therapy.

The frequent occurrence of comorbidities in older adults with kidney failure contribute to the complication of decision making regarding dialysis modality choice. Clinical risk prediction tools may be of use to patients and clinicians in the decision-making process by providing comparisons of the risk for mortality with that of other similar patients.

There are few tools designed to provide clinical risk prediction specifically in older adults. Results of earlier studies have suggested differences in outcomes and their predictors for older adults, including initiation of dialysis therapy at higher estimated glomerular filtration rates (eGFRs), increased comorbid conditions, and greater risk for hospitalization.

James P. Wick, MSc, and colleagues in the departments of community health services, family medicine, and medicine at the Cumming School of Medicine, University of Calgary, Calgary, Canada, derived and validated a clinical risk prediction tool for all-cause mortality among older adults during the first 6 months of treatment with maintenance dialysis treatment. They reported results in the American Journal of Kidney Diseases [2017;69(5):568-575].

The researchers identified a cohort of all adults ≥65 years of age who initiated maintenance hemodialysis or peritoneal dialysis therapy in Alberta, Canada, from May 1, 2003, through March31, 2012. Data from the Northern and Southern Alberta Renal Program (NARP/SARP) registries. The registries include all individuals initiating maintenance therapy in Alberta. The study outcome of interest was all-cause mortality within 6 months of initiation of dialysis therapy.

The cohort included 2199 individuals ≥65 years of age who initiated maintenance therapy in Alberta during the study period. Most of the 2199 individuals initiated renal replacement therapy on hemodialysis. Mean age at initiation of dialysis therapy was 75.2 years, 60.8% were men, 96.5% were non-First Nations, 75.4% were urban, and the majority had a below-average median neighborhood income (50.4% < 3rd quintile).

In the 6 months prior to dialysis initiation, 63.7% of study participants had visited an emergency department (ED); 43.5% had been hospitalized. Death occurred in 17.1% (n=375) during the first 6 months of dialysis therapy. Those who died within the first 6 months of dialysis therapy were more likely to be older, male, and First Nations, and live in a rural location. They were also more likely to have visited the ED and/or had a hospital stay in the 6 months prior to initiation of dialysis, more likely to have been referred to a nephrologist within the final 90 days before initiation of dialysis therapy, and have higher kidney function and higher burden of comorbid conditions, particular atrial fibrillation, congestive heart failure, chronic obstructive pulmonary disease, and stroke.

There were seven predictors in the final model for 6-month mortality: age, eGFR, hospitalization in the 6 months prior to dialysis initiation, atrial fibrillation, congestive heart failure, metastatic cancer, and lymphoma.

The C statistic for the final model for 6-month mortality was 0.72, representing reasonable model discrimination. For each category of each predictor in the final model, equivalent integer points based on the strength of the category and relative strength in relation to other predictors in the model were assigned. The combination of these point categories created a single point score of 19. A score <5 equated to a <25% 6-month mortality; a score >12 resulted in more than half the individuals dying in that time. In general, as the score increased, so did the percentage of individuals with that number of points who died within 6 months.

The researchers noted that the results should be interpreted in light of the study’s limitations, including the tool not being externally validated, limiting the generalizability of the findings, as well as the use of retrospective administrative data in the study.

“We derived and internally validated a clinical risk prediction tool for 6-month mortality subsequent to maintenance dialysis therapy initiation for older adults using a large population-based data source. Age, eGFR, hospitalization in the prior 6 months, atrial fibrillation, congestive heart failure, metastatic cancer, and lymphoma were included within the final tool, and a 19-point score was developed to reflect individual prognosis for 6-month mortality. The measures of performance and validation suggest promising potential for future external validation and clinical testing prior to its clinical application. Tools of this nature may be a useful decision aid for patients who are considering dialysis therapy initiation and the clinicians caring for them,” the researchers said.

Takeaway Points

  1. There are clinical risk prediction tools available to predict the risk of 6-month mortality in patients initiating renal replacement therapy; however, few of those tools focus specifically on older adults.
  2. Canadian researchers linked research and administrative data to develop and internally validate a prediction model that could be used to predict all-cause mortality among older adults during their first 6 months of maintenance dialysis therapy.
  3. The final tool included seven variables: age, eGFR, hospitalization in the prior 6 months, atrial fibrillation, congestive heart failure, metastatic cancer, and lymphoma.