At times, the 0.40 lps level has been erroneously referenced in posts as the leak level at which sensors and algorithmic function will begin to fail. That this cannot be true is clear when we remember that ResMed leak reporting excludes expected venting; the machine must be designed to work at least up to 0.40 lps above the highest expected venting levels for any supported mask at the highest operating pressures. Further, documentation warns of potential "machine failure" only above the 0.70 lps level.ResMed wrote:Generally, a leak rate of more than 0.4 L/s (24 L/min) is associated with patient discomfort, disturbed sleep, and reduced efficacy of treatment.
Notes:
a. If the leak exceeded 0.4 L/s (24 L/min) for more than 30% of the previous session, the following message will appear on the LCD: High leak in last session.
b. The Leak Alert feature, if enabled, will alert the patient when mask leak exceeds 0.7 L/s (42 L/min) for more than 20 seconds. When the leak exceeds this level, the autotitrating algorithm ceases to perform optimally.
Nonetheless, I have just stumbled upon what appears to be a relevant ResMed data management practice which gives increased importance to the 0.40 lps leak level. Last night I tried a new mask (which I will decline to identify) with unsatisfactory results. (For the very first time, I learned what the ResMed "Leak Alert" audible alarm sounded like.)
The ResScan pressure and leak graphs appeared as follows:
(I would opine that the first two pressure spikes "caused" the increased leaking, rather than the reverse. The final high leak was probably positional. In all three cases, I appear to have been awakened by the leaks.) Clearly, maximum pressure rose to well over 16 cmH20 (in fact, the maximum was 16.6cm). However, examining the statistical report, a maximum pressure of only 14.8 was indicated:
I am able to utilize the older AutoScan software, which includes the ability to export detailed data into a spreadsheet. Examining this data, sorted by pressure, revealed the following:
Note that the reported maximum pressure of 14.8 was the highest pressure encountered while the leak rate was under 0.40 lps. Higher pressures, occurring while the leak rate was over 0.40 lps, appear to be ignored by the "maximum pressure" statistics.
(In addition, the ResScan leak statistics report a maximum leak of 0.6 lps - which was actually 0.56 lps according to AutoScan. This level occurred twice, at 2:42 and at 6:35am. Information about the "Leak Alert" feature, which was sounding as I awoke and shut down the blower at 8:37am, stipulates a leak of 0.7 lps for at least 20 seconds. During my last 15 minutes of sleep the maximum leak reported in any 60-second-reporting-window was 0.48 lps, and the last full minute before awakening showed only 0.38 lps. This does not rule out a leak of 0.7 lps, but it does suggest that the reported (and graphed) leak rate is not the maximum leak for the period, but is more likely the median leak over the 60-second interval.)
My conclusion is that ResMed data presentations of maximum pressure ignore any pressures which may occur in the presence of leak over 0.40 lps. Whether such pressures are also ignored in the calculation of 95th centile and of median pressures is less clear. Further, maximum leak reporting does not reflect an instantaneous maximum, but more likely the maximum median leak for each 60-second reporting period.
