ASN's Mission

To create a world without kidney diseases, the ASN Alliance for Kidney Health elevates care by educating and informing, driving breakthroughs and innovation, and advocating for policies that create transformative changes in kidney medicine throughout the world.

learn more

Contact ASN

1401 H St, NW, Ste 900, Washington, DC 20005


The Latest on Twitter

Kidney Week

Abstract: SA-PO433

Development of a Predictive Model for Hyperkalemia

Session Information

Category: Chronic Kidney Disease (Non-Dialysis)

  • 304 CKD: Epidemiology, Outcomes - Non-Cardiovascular


  • Betts, Keith, Analysis Group, Inc., Boston, Massachusetts, United States
  • Woolley, J. Michael, ZS Pharma Inc., San Mateo, California, United States
  • Mu, Fan, Analysis Group, Inc., Boston, Massachusetts, United States
  • Bocharova, Iryna, Analysis Group, Inc., Boston, Massachusetts, United States
  • Bensimon, Arielle G, Analysis Group, Inc., Boston, Massachusetts, United States
  • Wu, Eric, Analysis Group, Inc., Boston, Massachusetts, United States

The objective of this study was to develop and validate a predictive model for the risk of hyperkalemia (HK) in US adults.


Adults were selected from a large US commercial claims database (2013-2014) if they were continuously enrolled from 7/1/13-12/31/13 (baseline) and 1/1/14-12/31/14 (follow-up) and had at least one serum potassium (K+) lab result during follow-up. The resulting sample was partitioned into two subsamples to train (60%) and validate (40%) the model. HK was defined as having in 2014: two elevated K+ values (>5.0 mEq/L); or one diagnosis for HK (ICD-9=276.7); or one prescription fill of sodium polystyrene sulfonate. In the training sample, multivariate logistic regression was used to develop a model estimating the 1-year probability of HK as a function of baseline covariates. Receiver operating characteristic (ROC) curve analysis of the validation sample was used to assess the predictive performance of the model.


HK was identified in 4,815 (1.6%) of 295,511 adult patients in the training sample. Some important baseline predictors of HK included: CKD stages 3-5 (e.g., odds ratio [OR] for stage 5 without dialysis= 8.11; 95% CI: 6.51-10.11); each additional year in age (OR=1.028; 1.025-1.030); history of HK (e.g., number of HK-related hospitalizations OR=1.65; 1.20-2.26); type II diabetes (OR=1.65; 1.53-1.77); and use of renin-angiotensin-aldosterone system inhibitors (RAASi, OR=1.38; 1.29-1.47). ROC curve analysis in the validation sample showed good predictive accuracy (area under the curve=0.78). The figure shows the probability of HK for a RAASi patient as a function of CKD stage, age, and history of hyperkalemia.


This study developed a HK prediction model with the most important predictors being CKD stage, age, and history of HK. More frequent K+ monitoring may be warranted for patients at elevated risk.