Autocycling is a form of patient-ventilator asynchrony that can occur during mechanical ventilation. In the case described, autocycling during noninvasive ventilation led to severe hyperventilation and subsequently produced a prolonged central apnea that resulted in syncope. This case represents the first description of a severe adverse event from autocycling during noninvasive ventilation.
Mu S, Rautela L, Howard ME, Hannan L. Autocycling during noninvasive positive pressure ventilation producing a prolonged severe apnea and syncope. J Clin Sleep Med. 2019;15(4):663–665.
Autocycling is a form of patient-ventilator asynchrony (PVA) that can occur during mechanical ventilation.1,2 It can be recognised by the rapid delivery of multiple cycles of ventilator breaths above the preset backup respiratory rate, without evidence of user effort,1–3 and is differentiated from autotriggering by the presence of multiple consecutive breaths.3 It is thought to occur due to the presence of aberrant flow signals within the ventilator circuit that trigger the ventilator.1–3
Various forms of PVA have previously been described in users of noninvasive positive pressure ventilation (NIV),4 in both acute and long-term use. PVA has been associated with worse nocturnal gas exchange, more frequent arousals from sleep, as well as reductions in total sleep time and percentage rapid eye movement sleep.
In the following case, autocycling was observed during nocturnal polysomnography and produced gross iatrogenic hyperventilation resulting in a prolonged central apnea, severe hypoxaemia and syncope.
REPORT OF CASE
The patient was a 68-year-old female with a history of chronic hypoventilation secondary to post-tuberculous restrictive chest wall disease who was admitted electively for titration of long-term NIV. Respiratory function testing was consistent with severe restrictive chest wall disease with FEV1 0.85 L (41% predicted) and FVC 1.16 L (44% predicted) producing a forced expiratory ratio of 0.73. The carbon monoxide transfer factor was 58% of predicted value while KCO was 111% predicted. Arterial blood gases at rest on room air prior to polysomnography demonstrated; pH 7.40, PaCO2 60 mmHg, PaO2 56 mmHg, HCO3 36 mmol/L. Prior to admission, the patient had been prescribed CPAP therapy at a pressure of 17 cmH2O. Compliance was poor with 3 out of 30 days usage recorded in the preceding month. Initial NIV settings for the subsequent study night were determined during a daytime trial by an experienced physiotherapist.
At the commencement of polysomnography, prior to commencing NIV, SpO2 measured on room air was 77% to 80% and PtcCO2 was 66 mmHg with a respiratory rate of 20 breaths/ min. NIV was commenced and delivered using an oronasal mask and VPAP S9 Tx (ResMed, Bella Vista, Australia) in spontaneous/timed mode with the following initial settings: inspiratory pressure (IPAP) 24 cmH2O, expiratory pressure (EPAP) 8 cmH2O and back-up respiratory rate 14 breaths/min. The minimum inspiratory time was 1.0 second and the maximum inspiratory time was 1.6 seconds. Soon after commencing NIV, brief respiratory events were observed during non-rapid eye movement sleep that were consistent with partial or complete upper airway obstruction associated with a loss of respiratory drive (falls in the amplitude of respiratory movements and device derived flow signals with minimal diaphragm EMG activity and associated oxygen desaturations). The leak signal was also elevated (at 35–40 L/min). In response, the attending sleep scientist increased EPAP which resulted in the abolition of these events with SpO2 improved to 88% to 93% and PtcCO2 falling to 33 mmHg.
Later in the study, a break was requested by the patient where NIV was ceased for a short period. On resuming NIV, the device leak signal was noted to be significantly lower and autocycling began to be observed (Figure 1) at a rate of approximately 33 breaths/min. The patient did not resume sleep during this period and PtcCO2 fell rapidly down to 14 mmHg while SpO2 increased significantly above the baseline level to 97% to 98%. IPAP was reduced in response to these observations however PtcCO2 remained low at 16 mmHg. Eventually the patient was evaluated clinically by the sleep scientist. Upon removing NIV, the patient was observed clinically to make no spontaneous efforts to breathe, become unconscious and demonstrate some tonic posturing but without evidence of myoclonus. Immediately a medical emergency response was initiated. Spontaneous respiratory efforts were observed after 3 minutes although oxygen saturation and stabilization of respiratory efforts did not occur until 5 minutes. During this time the patient regained consciousness and did not have specific recollection of the event but was not confused or disoriented. An arterial blood gas that was performed 15–20 minutes after the episode of syncope demonstrated an acute respiratory acidosis pH 7.28 and PaCO2 of 65 mmHg, HCO3 30 mmol/L (Figure 1), with the rise in PaCO2 reflected in the transcutaneous PtcCO2 signal. Subsequent review of electroencephalography recordings demonstrated no evidence of ictal or interictal epileptiform discharges. No cardiac arrhythmias were observed. The event was associated with a prolonged central apnea and profound oxygen desaturation to < 50%.
Polysomnography recording of autocycling leading to syncope.
Polysomnographic parameters during wakefulness (A); polysomnographic parameters during NREM sleep (B); polysomnographic parameters demonstrating autocycling during wakefulness (C); and polysomnographic parameters just prior to the syncopal event (D). Abdo = abdominal movement, Flow = oronasal airflow, Leak = amount of leak L/min from mask, NREM = non-rapid eye movement, Pmask = mask pressure, PtcCO2 = transcutaneous carbon dioxide, SpO2 = peripheral capillary oxygen saturation, Thor = thoracic movement.
Polysomnography recording of autocycling leading to syncope.
Syncope is characterized by the sudden and temporary loss of consciousness (with spontaneous recovery) caused by insufficient oxygen delivery to the brain.5 The lack of oxygen delivery is most commonly secondary to hypotension.5 In this case however, the proposed precipitant was iatrogenically induced hyperventilation leading to a severe and prolonged central apnea followed by syncope.
This unusual form of syncope occurred due to the confluence of a number of factors that potentially predisposed the patient to the development of a prolonged central apnea. These include baseline elevations in PaCO2 and HCO3 and a high plant gain—where small changes in ventilation produce large changes in PaCO2. Thus, when acutely exposed to excessive ventilatory support from autocycling combined with an inadvertent reduction in mask leak, the patient experienced a rapid and precipitous fall in PaCO2 resulting in a severe and prolonged central apnea and syncope when NIV was removed. It is postulated that the elevated mask leak observed early in the study may have mitigated hyperventilation from NIV during sleep and provided false reassurance to the attending sleep scientist that the level of ventilatory support was appropriate.
This case represents the first description of a severe adverse event resulting from autocycling during NIV. Autocycling has previously been described as occurring due to flow disturbance within the ventilator circuit, often as a result of condensation within the mask or tubing or in the presence of excessive leak. Other potential contributors include patients with low respiratory drive and effort (where random noise is more likely to trigger the ventilator) or in those with vigorous cardiogenic oscillations.2 Although autotriggering and autocycling during invasive ventilation have resulted in the delayed recognition of brain death in patients in the intensive care unit, to our knowledge no reports exist that describe adverse outcomes from autocycling during NIV.6 Autocycling during NIV may contribute to sleep fragmentation, however this case highlights a significant and novel complication arising from this type of PVA event. The use of complex nocturnal monitoring (such as polysomnography and polygraphy) to titrate NIV provides the clinician with the ability to identify and reduce the frequency of PVA events such as autocycling,3 and reduce complications that may result. For example, adjusting trigger and cycle sensitivity or altering inspiratory time can often resolve autocycling.3
This case highlights the need for laboratory protocols to adjust ventilator support and settings in the event that significant hyperventilation is observed and the potential for autocycling to inadvertently cause severe hyperventilation. Directly observed nocturnal monitoring with trained staff that perform these adjustments when required may have advantages over simpler titration strategies, particularly in individuals with complex respiratory mechanics and abnormal physiology.
Work for this study was performed at Austin Hospital, Heidelberg, Victoria. All authors listed have seen and approved the manuscript. The authors report no conflicts of interest.