Written by Clay Smith
Spoon Feed
ECG can’t be used to rule out hyperkalemia. It may be present and electrocardiographically silent. But if you see the changes, ECG can be used to rule it in. Sensitivity was 19%; specificity 97%.
Why does this matter?
Corey Slovis teaches hyper-K = ECG. And this is true. There are ECG changes which portend bad things in hyperkalemic patients. But is the ECG a good screening tool to detect hyperkalemia?
Hyper-K still = ECG
This was a retrospective review of emergency physician evaluation of ECGs in patients with ESRD on hemodialysis. In this cohort with poor access to dialysis, prevalence of hyperkalemia was 60%, with half of those over 6.5meq/L. They had a prior dataset with paired ECGs and electrolytes. Bundle branch blocks, LVH, poor tracings, and paced ECGs were excluded. They selected 66 ECGs, 60% of which were in patients with hyperkalemia but spanned a range of potassium values, from normal to very severe. These were assessed by eight ED attending physicians at this single center, with a range of clinical experience from 1- 25 years. Overall, sensitivity was poor for detecting hyperkalemia on ECG, 19%; specificity was 97%. For the subgroup with K > 6.5, sensitivity was 29%, specificity 95%. The most common way kyperkalemia was detected was, “T-wave amplitude in 97.5%…, QRS length in 14%, prolonged PR duration in 6%, and P-wave changes in 15%.” My take away is the ECG is a poor screening test, but if you see the tell tale changes, you’re probably right. In other words, you can’t rule out hyperkalemia with ECG, but you can rule it in. Like Corey Slovis says, hyper-K still = ECG.
Source
Can physicians detect hyperkalemia based on the electrocardiogram? Am J Emerg Med. 2019 Apr 22. pii: S0735-6757(19)30260-8. doi: 10.1016/j.ajem.2019.04.036. [Epub ahead of print]
Open in Read by QxMD
Reviewed by Thomas Davis
Note that the study is only from ESRD patients on hemodialysis. These patients have myocytes that may respond differently to hyperkalemia (they are more used to it) than patients with acute hyperk. And one of the 8 evaluators had a sensitivity of 0,51 with 0,92 specificity. This is more in line with previous studies. Still not good enough, but better.