The aim of this study was to evaluate the sensitivity of the long exercise test (LET) in the diagnosis of periodic paralysis (PP) and assess correlations with clinical phenotypes and genotypes.

From an unselected cohort of 335 patients who had an LET we analyzed 67 patients with genetic confirmation of PP and/or a positive LET.

32/45 patients with genetically confirmed PP had a significant decrement after exercise (sensitivity of 71%). Performing the short exercise test before the LET in the same hand confounded results in four patients. Sensitivity was highest in patients with frequent (daily or weekly) attacks (8/8, 100%), intermediate with up to monthly attacks (15/21, 71%) and lowest in those with rare attacks (9/16, 56%) (p = .035, Mann–Whitney U‐test). Patients with a positive LET without confirmed PP mutation comprised those with typical PP phenotype and a group with atypical features.

In our cohort, the LET is strongly correlated with the frequency of paralytic attacks suggesting a role as a functional marker. A negative test in the context of frequent attacks makes a diagnosis of PP unlikely but it does not rule out the condition in less severely affected patients.