Presentation Authors: John Onofrey, New Haven, CT, Jose Murillo Netto, new Haven , CT, Israel Franco*, New Haven, CT
Introduction: Uroflow measurements are subject to sberrant spikes that can alter Qmax and inaccurate capturing of the start of urination can have serious effects on Voiding Time , Qavg, acceleration, and time to Qmax (TTQmax). The aim of this study is to define if filtering the flow data and initiation of the measurement of voiding is set to a small volume will this alter these same parameters in the same unfiltered set.
Methods: 236 uroflow studies were identified that were known to have a bell shape within the 1 SD parameters used to define a bell curve based on flow indexes. All studies were reviewed personally by one of the authors to confirm that there was no evidence of EMG activity (IAEMG) during voiding. Raw uroflow data was utilized in a digital format and volume data was pre-processed with a Gaussian smoothing kernel to eliminate noise. We then calculated Flow by calculating the derivative of this Volume curve (using second order finite differences), and down sampled the data to have 0.25 second resolution. Acceleration curves were similarly calculated. All subjects were aligned to a common reference time by shifting the curves so that 4% value of the maximum volume was achieved at time zero, t=0. Standard uroflow parameters were compared between the new data obtained and the original data utilizing non parametric methods.
Results: We analyzed flows of girls between 12 - 21 years who were nulliparous ( mean age 14.8Â± 2.9) We digitized 84 uroflows in which the data was adequate for analysis. We compared Qmax from original data group consisting of the normalized flow (NF) (20.2Â±6.2) vs IAEMG (20.3Â±4.9) and found no differences. In the TTQmax analysis there were differences between NF (4.5Â±2.1) and IAEMG (7.1Â±6.9) that were significant (pâ‰¤0.0001). Qavg also differed between NF (10.1Â±4.1) and IAEMG (11.7Â±4.0)(p=0.008). Qmax FI was compared between NF (0.872Â±0.171) and quiet EMG (0.973Â±0.138) and found that they varied significantly p < 0.001. We also looked to define if the percentage of time to achieving Qmax (%TTQmax) was different between the groups. The percentage varied significantly between NF (27.6Â±11%) compared to IAEMG (40.6Â±31%) p < 0.0001.
Conclusions: We have shown that a system of normalizing, smoothing and standardizing flows has a significant impact on the Qmax, and Qmax flow index aside from the timed based measures that were routinely prone to errors in doing the calculations necessary to obtain uroflow results. These time-based errors further erode the quality and confidence of any data we obtain that is determined by uroflowmetry across different investigators.