IntroductionTo evaluate an updated algorithm in the detection of urinary tract infection (UTI) prior to high-dose corticosteroid treatment in acute relapses in multiple sclerosis (MS). This updated algorithm aimed to decrease the unnecessary use of antibiotics, whilst maintaining accuracy and safety. MethodsProspective cohort study of 471 consecutive patients with MS relapses in a hospital-based outpatient acute relapse clinic. 172 patients met exclusion criteria, leaving 299 patients for analysis. Patients underwent urine dipstick and were treated for UTI if 2 or more of: nitrites, leukocyte esterase and cloudy urine were positive. Patients with confirmed acute MS relapse were treated with high dose intravenous or oral methylprednisolone. ResultsSignificant bacteriuria (>105 colony forming units/mL) was present in 33 (11%, 95% CI 8-15) patients. The algorithm sensitivity and specificity was 24% and 94% respectively; the negative predictive value was 91%. The overall accuracy of the algorithm was 87%. No adverse sequelae were identified in 25 patients who received high dose methylprednisolone in the presence of an untreated UTI. ConclusionWith an improved specificity, this updated algorithm addresses previous issues concerning the unnecessary prescription of antibiotics, whilst improving accuracy and maintaining safety.