Interpretation of uroflowmetry curves by urologists.

PURPOSE

Uroflowmetry has become a routine investigation in patients with symptoms of the lower urinary tract. Little is known about the variation in the use of uroflowmetry and in the interpretation of its outcomes. We investigated the diagnostic value of uroflowmetry as a freestanding test, and examined the interobserver and intra-observer variation in the interpretation of uroflowmetry curves.

MATERIALS AND METHODS

A representative panel of 58 urologists was questioned about the relevance of visual inspection and flow parameters for interpretation. In addition, they individually assessed 25 randomly selected uroflowmetry curves (from patients with no abnormalities and those with various lower urinary tract symptoms) regarding normal findings and the most likely diagnosis. To investigate intra-observer agreement 4 of these curves were studied twice.

RESULTS

Voided volume (81%), visual inspection (77%) and maximum flow rate (77%) were most frequently mentioned as relevant for interpretation. Large differences existed between panel opinions and actual case information. For 43% of the normal cases the panel members considered the curves as abnormal. Of the abnormal cases 6% of the curves were regarded as normal. The urologists predicted correctly the actual diagnosis in 36% of all cases. Interobserver agreement was moderate for normalcy (kappa 0.46, standard error 0.087) and poor for the most likely diagnosis (kappa 0.30, standard error 0.043). Intra-observer agreement was also not satisfactory. On average, for the 4 cases studied twice 29% of the panel members chose another option for normalcy, while 41% mentioned another diagnosis the second time.

CONCLUSIONS

These results necessitate reconsideration of the diagnostic use of uroflowmetry in daily urological practice.