Sep 082018

Jon-Emile S. Kenny MD [@heart_lung]

WANTED: Somebody to go back in time with me.  This is not a joke …

-John Silveira


A 78 year old man with known moderate-to-severe emphysema is extubated in the ICU; he was initially admitted with hypercapneic respiratory failure secondary to influenza pneumonia and pulmonary edema from the medical floor.  On first arrival to the general medical floor, he had been diagnosed with sepsis and, within one hour, received aggressive crystalloid resuscitation.  Then, following 4 days of mechanical ventilation, diuresis and bronchodilation, his gas exchange and respiratory mechanics improved in the ICU.  Ensuing extubation, he is immediately placed on high-flow oxygen at 60 litres/min – the patient reports severe claustrophobia from BiPAP.  The on-call ICU resident asks the respiratory therapist if high-flow oxygen is beneficial for this particular patient – he voices concern about carbon dioxide retention.


As a fussy physiologist, I frequently find myself deciphering literature on hemodynamics and respiratory mechanics – not because of their complexity, but rather their inconsistent nomenclature.  A recent report by Di mussi et al. multiplies my frustrations.

I have previously posted on the mechanisms of carbon dioxide excretion with oxygen delivered by nasal high flow [NHF], as well as on the important distinction between respiratory work and power.  A fresh publication addressing the beneficial effects of NHF in patients with chronic obstructive pulmonary disease [COPD] focuses these erstwhile ruminations and further informs on the subtle disparities between respiratory work, effort and efficiency.  Without jest, I invite you to travel back in physiological time with me … you must bring your own weapons of critical appraisal; your safety is not guaranteed.

What They Did

14 patients with moderate-to-severe COPD and hypercapneic respiratory failure [PaCO2 level not defined] were studied.  Importantly, patients with tracheostomy or those planned for prophylactic non-invasive positive pressure ventilation were excluded.  In those enrolled, a nasogastric tube was placed – prior to extubation – to monitor the electrical activity of the diaphragm [EAdi in microvolts – uv].  Of the 20 eligible patients screened, 4 declined to participate and 2 had EAdi data that was not adequate for analysis.

The EAdi was used as a surrogate for respiratory muscle effort.  How?  Just before extubation, an end-expiratory hold maneuver was undertaken on the ventilator and the peak fall in airway occlusion pressure on inspiration was matched to the peak rise in the electrical activity of the diaphragm.  The investigators then related the measured EAdi to the fall in airway pressure and created an index relating the two – that is, continuously measured electrical activity of the diaphragm was converted to pressure in the airway and this was used as a surrogate for the pressure generated by the muscles of respiration [Pmusc].

The effort exerted by the muscles of respiration [as estimated by the EAdi] was measured after extubation in 3 consecutive 30 minute periods – firstly on NHF, then on conventional oxygen therapy followed again by NHF.  Tidal volume was not measured throughout.

What They Found

There was no clinically-significant difference between respiratory rate, pH or PaCO2 between the NHF periods and the conventional oxygen therapy period.  The average respiratory rate was about 20-21 breaths per minute while the pH and PaCO2 ranged between 7.44 to 7.46 and 50 to 52 mmHg, respectively.  Oxygen tension was slightly higher during the second phase of NHF by about 10 mmHg.

While on NHF, the pressure time product [PTP] estimated from the electrical activity of the diaphragm was, on average, in the low 130s cmH2O ⦁ s/min while this increased to 211 cmH2O ⦁ s/min while on conventional oxygen therapy.


The authors sought to translate the electrical activity of the diaphragm into the muscular activity of the respiratory system.  They based this on a paper by Bellani and colleagues which found a good relationship between the EAdi and the muscular effort of the respiratory system.  Importantly, however, the paper by Bellani did not study the effort of the diaphragm – as incorrectly reported in the paper by Di mussi et al.  There is a difference between the pressure time product of the respiratory system [i.e. the PTPes] – which is derived from esophageal pressure relative to the passive chest wall recoil pressure – and the pressure time product of the diaphgram [PTPdi].  The latter is derived from the difference between gastric pressure and esophageal pressure.  The PTPes is thought to reflect the muscular activity of all the muscles of respiration while the PTPdi focuses on the activity of the diaphragm.

The authors report the PTPdi [which, as above, is actually the PTPes] as ‘work of breathing.’  This too is not exactly correct.  Work requires displacement [a force deployed over a distance, in joules]; however, the pressure time product [PTP] can rise when lung volume – displacement – does not change.  An extreme example is a Muller maneuver [not to be mistaken with indicting presidential campaign members] which is when the muscles of inspiration are contracted against a closed glottis.  This leads to great effort and energy expenditure of the respiratory muscles, but there is no volume change and, therefore, no work done.  The important point here is that PTP tracks energy consumption by the muscles of respiration more closely than breathing work.

Accordingly, the reason that PTP is often used as an index of respiratory effort – rather than work – is because it accounts for respiratory energy consumption that occurs during isometric contractions; for example, when overcoming auto-PEEP.  PTP also accounts for the time spent during an inspiratory effort which is also directly related to energy use.  Because PTP is the integral of the respiratory muscle pressure change over time, its units are cmH2O ⦁ s/min and not cmH2O/s/min as listed by the authors.

An interesting paper from 2014 highlights the difference between breathing work and breathing effort.  In patients with COPD and hyperinflation with auto-PEEP, EAdi, esophageal pressure and tidal volume were measured in response to the addition of external PEEP.  The addition of external PEEP from 2 cm H2O to 8 cm H2O while on pressure support increased tidal volume by about 20% and decreased both Pmusc and the EAdi by a similar amount.  Consequently, if volume increases and pressure falls, then work [the area defined by the pressure-volume loop] may actually change minimally; importantly, however, respiratory effort [e.g. the PTP] falls because the effort required to overcome auto-PEEP is diminished by external PEEP.  Accordingly, if work is unchanged at a lower respiratory effort/energy consumption, then the efficiency of breathing has improved.  The concept of work [in joules] relative to its efficiency is important because this value must stay at-or-below the amount of energy delivered to the muscles of respiration [i.e. oxygen delivery] to maintain adequate ventilation energetics.

Interestingly, in the 2014 paper, the amount of external PEEP applied to lower EAdi to a value similar to the EAdi seen in Di mussi et al. was 8 cm H2O [EAdi fell to 15 uV in both cohorts], somewhat consistent with the level of high-flow oxygen applied [60 L/min is expected to supply about 5 cmH2O of external PEEP with the mouth closed].  Appropriately, it was felt that the improvement in PTP during NHF was the mitigation of the isometric energy needed to overcome auto-PEEP.

Because PTP is correlated with oxygen consumption [VO2], its fall during NHF is expected to lower carbon dioxide production by the muscles of respiration as well.  As described previously, carbon dioxide production [VCO2] is directly proportional to the arterial tension of carbon dioxide [PaCO2].  While PTP fell in the paper by Di mussi et al. [from 211 cmH2O ⦁ s/min to roughly 130 cmH2O ⦁ s/min – normal values are considered to be between 50 and 150 cmH2O ⦁ s/min] this is expected to lower the PaCO2.  A lower PaCO2 was not seen during NHF which may have been due to contemporaneous fall in minute ventilation or, perhaps, increased dead space from hyperoxygenation.

In conclusion, Di mussi et al. did not measure tidal volume, so nothing may be definitively said about the work of breathing.  Nevertheless, the fall in EAdi and, therefore, estimated PTPes [the pressure time product of the muscles of respiration – not the diaphragm] did fall, proving diminished respiratory effort and, likely, respiratory muscle energy consumption.

Return to Case

The respiratory therapist explains the physiology of nasal high flow oxygen to the resident and the patient tolerates 60 L/min.  His respiratory rate slowly falls and so too does his PaCO2.  His oxygen saturation is maintained in the low 90s and he is transferred to the general medical service for additional care in the respiratory step-down unit.    



Dr. Kenny is the cofounder and Chief Medical Officer of Flosonics Medical; he is also the creator and author of a free hemodynamic curriculum at

Get our weekly email update, and explore our library of practice updates and review articles.

PulmCCM is an independent publication not affiliated with or endorsed by any organization, society or journal referenced on the website. (Terms of Use | Privacy Policy)


High Flow Oxygen in Chronic Obstructive Pulmonary Disease: improved work of breathing or respiratory effort?