Aim Long QTc is associated with poor outcomes on CCU and acute stroke units. Its incidence and outcome in unselected medical admissions is unknown. A range of drug classes and electrolyte abnormalities may lead to acquired long QTc. Given the frequency with which these are encountered in medical admissions, the incidence of long QTc requires study. We have calculated the incidence of long QTc encountered in acute medical admissions. As inaccuracies in the correction formula may give misleading data, this study compared five formulas commonly used to correct QT interval duration by heart rate (HR): Bazett (B), Fridericia (Fa), Framingham (Fm), Hodges (H), Van der Water’s (VdW) as a well as the automated QTc correction provided by the Welch Allyn (WA) ECG recorder.

Methods 1000 records of consecutive patients admitted to acute medical wards at a tertiary teaching hospital were reviewed by two physicians. Admission ECGs were examined. QTc was calculated using the five formulas specified. The Pearson correlation coefficient (R²) was used to estimate the relationship between QTc and HR. Patient age and gender were recorded. QTc of ≥500 ms was considered significantly prolonged in both genders. ECGs with missing patient identifiers or poor trace quality were excluded. Cohen’s kappa statistic, for inter-observer agreement, was made from a subgroup of 200 randomly chosen ECGs evaluated by a third observer.

Results 715 ECGs were included. 356 males, 359 females. Age 61 ± 20 years (mean ± SD). Kappa inter-observer agreement was 0.65 indicating substantial agreement. Incidence of QTc ≥500 ms was 1.7% (B), 0.4% (Fa), 0.4% (Fm), 0.3% (H), 0.4% (VdW) but was 6.2% with WA. Pearson correlation between QTc and HR was 0.02 (B), 0.06 (Fa), 0.5 (Fm), 0.001 (H), 0.5 (VdW) 0.009 (WA).

Conclusion The incidence of significant QT prolongation in unselected medical admissions varied by a factor of ten depending on the method used. Automated QTc software gave a far higher incidence of long QTc than manual interpretation. Framingham and Van der Water’s equations show bias by HR. The clinical sequelae of long QT in unselected medical admissions needs further study.