ICS-standard cystometry is performed with a fluid filled tubing system (FL) with external pressure sensors. The installing and measuring procedures are standardized, however there is operator dependency in the process and the system is intrinsically sensitive to patient movements, as a consequence of the external pressure sensors and tubes required for the test. An air filled catheter –balloon- pressure recording system (AF) was developed to circumvent (erroneous) external pressure reference and to avoid artefacts resulting from (patient &) tubes movements. Contrary to the FL system, the AF system measures the urodynamic pressures at the site of the catheter tip (inside the patient). This also requires careful operating, but patient movements cause lesser measurement artefacts. 
We have compared how both systems perform when used head to head during otherwise ICS standard cystometry.

Published and or conference-presented other studies have demonstrated differences between the two systems at hallmark moments during the urodynamic study, e.g. pressures at sensations and or during detrusor overactive contractions or Valsalva strain peak pressures. For our study we compared not only hallmark (peak) pressures, but all (digitally recorded) samples of the complete measurements. Both channels; pves and pabd and both systems; FL and AF have been sampled with 20Hz. 
On the basis of the collective evidence from earlier reports we have concluded that the observed differences are a consequence of the differences in zeroing technique between both systems; differences in the levelling of the external pressure sensors in the FL system, differences in charging of the AF system and differences in (air) balloon position in the body cavity may potentially have caused systematic base-line differences, represented in the –seemingly random- differences as are reported in the event -peak -pressure -reporting -studies.
The earlier published in vitro comparative technical validation studies are not biased by these, clinically almost unavoidable, baseline pressure differences and report that the AF system is responding damped, when compared to FL; steep pressure increments are recorded lower and slower in vitro with the AF system than with its comparator.
Thirty-eight patients with signs and or symptoms of LUT dysfunction were included, after IRB approval of the protocol and individual written informed consent. Men with intact saddle region sensation were excluded because of the double (2x7F side by side) catheterisation. Also women (or men) with a flowrate <15mL/s or with other (pre-urodynamic) signs of voiding dysfunction were not included. 
Transurethral medium fill-rate simultaneous double system (fluid (FL) and air (AF)) cystometry was performed in seated position when possible and supine when needed. Twenty seven of the head to head comparisons were complete for both abdominal pressure (pabd) as well as for intravesical pressure (pves). We excluded patients/measurements where technical problems, because the ‘double -urodynamics’ caused specific artefacts (usually: (partially) missing (the recording of) one of the four pressure channels). 
Our full-traces (all data-samples) urodynamic measurement head to head comparing study also reports detailed in vivo cough- responses of air filled versus water filled system in pves and pabd, to unravel the clinical relevance of the in vitro observations as well as to study the consequences for detrusor pressure (pdet).
For this analysis pressures of all four channels were levelled –equalized from the first sample after flushing of the fluid filled system. This equalizing (violating ICS standards in this post processing analysis) fades away the absolute pressures in both systems but allows better understanding of the specific responses of both systems to pressure variations of every origin during the measurements.

Figure 1 is an example of one equalized study with overlay of FL and AF. 
Curve fitting analysis was done with matlab® to observe time shift and or pressure damping: The average time shift (measured over the full urodynamic studies) in pves 0s [-0.05 - 0s] and in pabd -0.05s [-0.05 - 0s] was almost zero; less than the inter -sample time of the system. The cross correlation of the pressure channels (per location pabd or pves) of both systems (FL and AF) was extremely high: 0.995 [0.981 - 0.998] for pves and 0.993 [0.982 - 0.998] for pabd. Peak absolute -baseline equalized- pressure increment differences per location during the entire cystometry were never >10 cmH2O.
Figure 2 shows an example of one, and two coughs, in detail (baselines 3,5 and 5 seconds). We have analysed 68 cough curves. Within the FL system the pabd is delayed to pves, causing a sinusoid in the pdet. The AF signals are (more) synchronous for pabd and pves, but both delayed compared to the FL intravesical pressure. In the pdet the deviation from the baseline is the highest in the water filled catheter and the AF signal is smoother. Overall (68 coughs), the amplitude difference of the normalized signals, has a median of 6.8 [3.8-8.9] cmH2O for pves, 6.0 [1.7-8.0] cmH2O for pabd and 4.9 [1.2-9.7] cmH2O for pdet. The cross-correlation for the coughs was 0.996 [0.990-0.997] in the FL and 0.998 [0.994-0.999] in AF. The time delay was on average 0 [0-0] s in both systems. The average amplitude difference is 3.8 [1-7.0] cmH2O in FL and 2.9 [0.6-6.1] cmH2O in AF. There was no statistical significant difference between both systems with regard to area under the couch curve (pves 25.2 cmh2O*s WF versus 24.7 cmh2O*s AF (t-test p .30) and pabd 25.0 cmh2O*s versus 26.1cmh2O*s; p.031).
We analysed the damping of both systems with the averaged bode diagram (not shown).  Both systems record similar in the most relevant (<2-3Hz) frequency range, with little damping in the AF. The highest frequencies, around 8 to 10 Hz, are not well recorded in both systems, however in the intermediate frequencies the AF signal reacts as smoothed, filtered version of the FL signal.

The high cross-correlation of datasamples of both systems per location indicates that the pves and pabd signals are around 99% similar after equalizing the two systems at the first relevant sample. The crosscorrelation in the calculated (pdet) detrusor pressure is however less, with median of 72%. This can be explained by the poor precision of the two measured signals, pves and pabd in both systems. When these signals are subtracted, an error of 10% in pves can potentially become and error of 1000% in pdet. This explains that FL and AF pdet (subtracted detrusor) signals differ more than their originating signals. 
The time shift was observed between 0s and -0.05s between the systems, which indicates that the AF signal is potentially one data-sample slower than the FL signal. We consider that a delay of (less then) 5 milliseconds is clinically not relevant. The detailed cough pressures analysis demonstrated identical (absence of relevant) time delay and lower peak pressures in the AF comparison with the ICS-standard FL, however the area under the curve did not significantly differ between FL and AF.

There are pressure response differences between fluid filled and air filled urodynamic measurements. These differences exist apart from the already reported differences that result from baseline setting variations. Although the conclusion may also be that the fluid filled system responds erroneously underdamped, the air filled system responds, also in vitro, with damping however not leading to clinically relevant differences that are systematically larger than those resulting from baseline urodynamic installing (set-up) and measurement pinciple variations.