Hyperkalaemia: physicians were poor at diagnosing hyperkalaemia from an ECG.

Clinical bottom line (level 2b)

  1. In patients with renal insufficiency or hyperkalaemia, emergency room physicians could not reliably detect or exclude hyperkalaemia nor grade its severity from the ECG alone.
Wrenn et al: Annals of Emergency Medicine 1991; 20: 1229-1232
Expires February 2004

The study

Setting: emergency department, university hospital, USA

220 patients (aged mean 59 years, 61% male) admitted to hospital with a diagnosis of hyperkalaemia or renal insufficiency

Excluded if
  • <16 years old
  • haemolysed blood sample
  • therapy instituted between the performance of the ECG and blood drawing



    • Independent blinded reference standard, applied in all patients from a consecutive inappropriate spectrum.
    Reference standard:
    • serum potassium drawn within 1 hour of the ECG
    Diagnostic test: ECG readings by clinicians- criteria for hyperkalaemia included any of: increased T wave amplitude with a narrow base; broadening and decreased amplitude of the P wave; absence of the P wave; widened QRS complex; a sinusoidal appearance of the QRS-T complexes

    The evidence

    pre-test probability of 37.0: 95%, (95% CI: 30.0% to 43.0%)

    diagnostic test hyperkalaemia no hyperkalaemia LR+
    (95% CI)    		 post-test probability LR-
    (95% CI)    		 post-test probability
    clinician 1 diagnosis with ECG 37 113 2.83
    (1.76 to 4.53)     		65.0% 0.68
    (0.56 to 0.82)     		31.0%
    clinician 2 diagnosed with ECG 30 115 2.55
    (1.52 to 4.28)     		63.0% 0.76
    (0.64 to 0.90)     		33.0%
    total 87 133

    interobserver agreement kappa = 0.73

    Comments

    1. ECGs were read by two American-board-certified physicians (emergency medicine, critical care medicine) blinded to the blood results, to the specific diagnosis of the patient, and to each other's reading.
    2. When patients with potassium >6.5 mmol/L were studied in subgroup analysis, the sensitivity of the physicians improved but was still poor (55% for one, 62% for the other).
    3. A large number of patients had baseline ECG abnormalities (left ventricular hypertrophy, bundle branch block).
    4. Although they did not know the specific diagnosis of the patients, the physicians did know that only patients with a diagnosis of renal insufficiency or hyperkalaemia were included. In a more representative patients population, the physicians' specificity for hyperkalaemia from the ECG would be expected to lower.

    Citation

    1. Wrenn KD, Slovis CM, Slovis BS, et al: The ability of physicians to predict hyperkalemia from the ECG. Annals of Emergency Medicine 1991; 20: 1229-1232
    Search Terms: Medline: MeSH term- hyperkalemia
    Contributor: Warren Lee, Chris Ball and Clare Wotton, February 2000
    Reviewer:

    Clinical Question.
    Patient hyperkalaemia
    Intervention or Exposure ECG
    Comparison blood specimens
    Outcome diagnosis