Mojoli F, Orlando A, Bianchi I, Torriglia F, Bianzina S, Pozzi M, Iotti GA, Braschi A, PLUG Working Group

Intensive Care Medicine Experimental 2016, 4(Suppl 1):27;597

Introduction: Patient-ventilator asynchronies are associated with

poor outcome. It was suggested that bedside analysis of ventilator

waveforms may help detecting different types of asynchrony and

setting properly the ventilator [1].

Objectives: To test accuracy of a “waveform” method, based on specific

signs on airway pressure (Paw) and flow curves, in detecting

spontaneous respiratory activity and asynchronies in patients under

Pressure Support Ventilation (PSV).

Methods: 16 recordings (12 min each) of esophageal pressure

(Pes), Paw and flow were obtained in obstructive (75 %) and restrictive

(25 %) patients under PSV with clinical evidence of poor

patient-ventilator interaction. Tracings of 4426 breaths were visually

analyzed for detection of spontaneous respiratory activity

both with Pes (reference method) and without Pes (waveform

method) by different operators. Breaths were defined as assisted,

unassisted or autotriggered, and assisted breaths as delayed triggered,

early cycled or delayed cycled. The waveforms method

was applied in a selection of tracings (20 min, 544 breaths) by 4

different operators for assessment of inter-rater agreement.

Results: The reference method detected 6 autotriggered (0.1 %),

976 unassisted (22.1 %) and 3444 assisted (77.8 %) breaths;

among assisted breaths, 897 delayed triggered (26.0 %), 1231 delayed

cycled (35.7 %) and 439 early cycled (12.7 %). Table 102

shows sensitivities and specificities (95 % CI) of the waveform

method in evaluating patient-ventilator interaction. The waveform

method detected the start of patient’s inspiration and expiration

with a bias of −23 and −32 ms and a precision (±1.98 SD) of 184

and 202 ms respectively. Absolute agreement among operators

was almost perfect for unassisted breaths, strong for delayed triggered,

delayed cycled and early cycled breaths, and weak for

autotriggered breaths.

Conclusions: The waveforms method is a reliable, accurate and reproducible

method to assess patient-ventilator interaction and could

help optimal setting of the ventilator. Automation of this method

may allow continuous monitoring of ventilated patients and/or improved

breath triggering and cycling.

References

1. Mojoli F et al. Intensive Care Med. 2015 [Epub ahead of print]

From Springer Open Access ICMX