Mar 242018

Jon-Emile S. Kenny MD [@heart_lung]

I grow old ... I grow old ... I shall wear the bottoms of my trousers rolled.”

-T. S. Eliot


While it is tempting to isolate nasal high flow [NHF] into one’s cognitive schema for hypoxemia, NHF rightly deserves an esteemed position within one’s cerebral scaffolds for both hypercapnia and work of breathing.  This is especially important given the emerging indications for NHF as an intervention with a high benefit – to – risk ratio.  Indeed, we may be quickly in need of a new hashtag – #FOANHF – free and open access nasal high flow?

Having both recently summarized an excellent lecture on respiratory load, and posted a brief missive surrounding the effects NHF on carbon dioxide excretion, it is an important time to reflect upon a new and small investigation concerning the effect of NHF on respiratory effort.  As a grouchy gentleman of growing vintage, I will make fussy distinctions between energy, work and power.  Nevertheless, I invite you to join me on a stroll down this physiological street that follows “… like a tedious argument … Of insidious intent.”

What They Did

12 patients were studied, 6 of whom were in the ICU and the other 6 on a respiratory ward.  Importantly 3 patients received NHF prior to inclusion in the study.  5 of the patients had hypercapnia [defined as a PaCO2 above 45 mmHg and acidemia] and the other 7 had primary hypoxemic respiratory failure.

The patients were studied with oxygen-only as a baseline and then randomly received 3 levels of NHF [20, 40 and 60 L/min].  Using an esophageal balloon to measure pressure and respiratory inductance plethysmography to measure tidal volume, ‘work of breathing’ was obtained [given as Joules per minute].  Additionally, the esophageal pressure-time-product and esophageal pressure variation were measured as indices of effort.  Dynamic lung compliance and airway resistance were also determined as were basic blood gas data.

What They Found

Compared to baseline, none of respiratory rate, tidal volume or minute ventilation changed with NHF up to, and including, 60 L/min.  Despite these steady values, there was a clinically-significant reduction in PaCO2 from a baseline of 52 mmHg to 44 mmHg at 60 L/min; while this was a positive clinical trend, it was not statistically-significant.

Compared to baseline, each of esophageal pressure variation, esophageal pressure-time product and work of breathing all decreased at 60 L/min to a clinical and statistically-significant degree.  For instance, the ‘work of breathing’ fell by 50% from baseline to 60 L/min; there was no clear dose-response.


Of note, this small study demonstrated, like previous investigations on NHF, a reduction in PaCO2 without an increase in minute ventilation.  Admittedly, the fall in PaCO2 was not statistically-significant but clearly followed a strong clinical trend [52 to 44 mmHg].

By the alveolar ventilation equation, a fall in PaCO2 without an increase in minute ventilation necessitates either a decrease in dead space and/or reduced carbon dioxide production [VCO2].  Diminution of dead space may occur secondary to a fall in anatomical dead space – which has been observed with NHF via nasopharyngeal wash-out – and/or mitigated physiological dead space fraction [Vd/Vt].  As described in this previous post, a large Vt – or deeper breathing – can reduce dead space!  Yet, as above, larger Vt were not noted in this study.  There are many reasons why greater Vt may not have been observed in response to NHF and these may range from measurement error – by relying upon inductance plethysmography – to studying less-sick patients – the respiratory rates at baseline ranged between 18 and 25 breaths per minute.

Yet, if we accept that Vt did not truly increase, it is still entirely possible that the reduction in breathing work lessened VCO2 and therefore assisted the fall in PaCO2.  On average, ‘work of breathing’, esophageal pressure-time product and esophageal pressure variation all fell by roughly 50%.  This isn’t surprising as they are conceptually and mathematically coupled.  If we consider a normal work of breathing [WOB] to be 0.3 to 0.6 Joules/Litre, the patients at hand demonstrated a WOB at baseline of 0.8 J/L which fell to 0.5 J/L at 60 L/min NHF*

[*note, I calculated Joules/L by multiplying the patients’ ‘WOB’ in J/min by the inverse of their minute ventilation; technically, ‘work’ is the amount of energy required to move something over a distance. Importantly, the authors reported 'work of breathing' in Joules per time, which is actually ‘power’ – measured in Watts].

Clearly here, the patients in this trial had a relatively small increased work of breathing at baseline, nonetheless, there was a significant reduction with NHF.  If we consider true power applied to the lung, a porcine model demonstrated a ventilation induced lung injury [VILI] threshold of 12 J/min.  By contrast, the patients in this study had a baseline power of 4.3 J/min that fell to 2.1 J/min with 60 L/min NHF.  One must be careful directly comparing human and pig lungs, however, as their specific elastances are different.

The reduction in WOB seemed to stem from attenuated elastic and resistive breathing work as the authors calculated both increased dynamic lung compliance and decreased airways resistance with the application of NHF.  However, given the use of inductance plethysmography to measure lung volume and their assumptions about PEEP and airflow, I think these numbers should be viewed as hypothesis generating.  Previous authors have found reduced airways resistance with NHF, however, it may have been entirely within the nasopharynx.

Lastly, the authors wisely mention the potential for NHF to mitigate P-SILI – patient self-induced lung injury.  This phenomenon has been described as a consequence of the high trans-pulmonary pressures generated when vigorously breathing spontaneously, over time.  Indeed, prolonged, deep, spontaneous inspirations will increase the power applied to the lung parenchyma, thereby increasing the risk of VILI.  As above, while reported as ‘work of breathing’ what the authors actually relay in this small study is that the power applied to the lung [in Joules/min] fell by 50% from unassisted breathing to the application of NHF at 60 L/min.

Happy Birthday EASK,


Dr. Kenny is the cofounder and Chief Medical Officer of Flosonics Medical; he 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 other organization, society or journal referenced on the website. (Terms of Use | Privacy Policy)


High-Flow Nasal Cannula, Work of Breathing & Mechanical Power: is there benefit?