Young age and symptoms should be included in predictive algorithms of urethral infection; they are better predictors of urethral infection in men than numbers of urethral polymorphonuclear leucocytes (PMNs), an evaluation study of non-gonococcal urethritis (NGU) has disclosed.

The evaluation substituted the ligase chain reaction for PMN count as the diagnostic test for C trachomatis and calculated sensitivity and specificity of the cut offs in PMN count in stained urethral smears as well as the likelihood of PMN count, symptoms, and age predicting C trachomatis urethritis. Over 300 men being screened for sexually transmitted infection at one clinic over four weeks were tested; none had gonococcal urethritis.

Age below 30 versus 40 and above increased likelihood of infection with C trachomatis 13-fold; PMN count, at ⩾20 per high power field (hpf), sixfold and 5-9 per hpf, threefold; and dysuria threefold. However, the sensitivity of PMN count at ⩾5 per hpf was just 63% and specificity 77%. Over a third of the men with C trachomatis by ligase chain reaction had PMN counts <5 per hpf—too many to justify the accepted diagnostic cut off at ⩾5. The proportion of men with NGU on a stained urethral smear was 26%, and it was associated with C trachomatis in 8%, according to the ligase chain reaction.

With nucleic acid amplification becoming available for C trachomatis it made sense to re-evaluate apparent relations between urethral infection and clinical picture, PMN count, and routinely available information like age and symptoms.