Protective face masks add significant dead space

To the Editor:

Lately, protective face masks have become part of everyday life and many patients and healthcare workers complain of reduced well-being and performance due to the permanent wearing of such masks. Reportedly, the main symptoms are headaches and difficult breathing, especially in combination with stress and physical exertion [1–3]. It is suspected that protective masks impede oxygen supply to the lungs and increase carbon dioxide rebreathing [4]. Theoretical estimations suggest that indeed the dead space increases for mask wearers [5] and this would translate into increased arterial CO₂ levels with concomitant increase in work of breathing through control-of-breathing mechanisms. Added flow resistance of protective face masks, as well as comfort and general physiological effects, have been described in the literature [6–8]. However, the effect on gas exchange is poorly investigated, basically because standard procedures involve using a mouthpiece and a nose clip, or a dedicated face mask (European Respiratory Society recommendation [9]), which is obviously not possible in subjects wearing a protective face mask. To investigate this question, the use of a lung simulator seemed to be an obvious approach.

A spontaneous breathing lung simulator (TestChest, Organis GmbH, Landquart, Switzerland) fitted with an intubation training head (Airway Larry Trainer, Nasco, Fort Atkinson, WI, USA) was used to measure indices of gas exchange when breathing through five different masks (surgical mask: Xiantao XinHengrun Plastic Products Co. Ltd; FFP2 Aura 1862+: 3M Germany; FFP3 Aura 1873V+: 3M Germany; KN95: Changshu Heng Yun Textile Co. Ltd; and a fabric mask from a general public store: Manor, Switzerland).

The parameters were measured when each mask was fitted normally or was taped to the face and was combined in each case with the setting of an open and closed mouth. Each of the measurements was preceded by a reference measurement (no mask on the simulation head). Performing the measurements during quiet breathing and during moderate exercise, a total of 40 measurements with masks were taken.

The breathing lung simulator was set to match the predicted values for a 175 cm male subject [10, 11] in supine position (total compliance of 54 mL·mbar⁻¹, airway resistance of 5 mbar·min·L⁻¹, functional residual capacity of 1823 mL). To simulate quiet breathing, the simulator was programmed to inhale with a sinusoidal pattern of −14 mbar for 1.4 s at a rate of 14 per minute (tidal volume (V_T) of 637±2 mL, minute ventilation of 9±0.01 L·min⁻¹). CO₂ production was simulated by infusion of pure CO₂ at 200 mL·min⁻¹ STPD using a mass flow controller (MFC220, Axetris, Switzerland). To simulate moderate exercise, the lung simulator was set to inhale with a sinusoidal pattern of −15 mbar for 1.2 s at a rate of 28 per minute, yielding a V_T of 792±3 mL and a minute ventilation of 22 L·min⁻¹. CO₂ production was set at 500 mL·min⁻¹ STPD to simulate increased metabolic demand during exercise.

The lung model was used factory calibrated and provided real-time volume and airway pressure data, which was exported and analysed using Microsoft Excel. V_T was measured as the difference between end-inspiratory and end-expiratory lung volume. Inspiratory and expiratory flows were plotted against alveolar pressure and linear regression was used to measure added airways resistance in inspiration and expiration (mbar·s·L⁻¹). An oxygen cell (Fuel Cell R-22A, Teledyne) and a side stream capnometer (Microstream Extension, Philips X3 Monitor) were inserted between the intubation head and the breathing lung model. Measurements were taken in steady state and included end-tidal CO₂ and end-inspiratory CO₂.

Bohr dead space (V_dBohr) was calculated using the end-tidal CO₂ tension (P_etCO2), the applied CO₂ production (V′_CO2) and V_T and respiratory rate (RR) as measured by the lung simulator: All measurements were taken in steady state after allowing 3 min for equilibration and results are reported as mean values over 1 min.

The results are provided in table 1.

Near “normal” P_etCO2 resulted in quiet breathing as well as in moderate exercise when wearing no mask. Interestingly, end-inspiratory CO₂ remained above zero, even without mask. Breathing through the nose alone apparently decreases V_dBohr, even when wearing no mask. All other values obtained in the reference measurements show no clinically relevant differences.

All measurements with masks showed significant increase in P_etCO2 (+17.4 mmHg on average), even more pronounced in the moderate exercise setting (+25.9 mmHg on average). The extent of the increase of P_etCO2 varied with different mask types and with the mouth open or closed. The V_dBohr nearly doubled when wearing a mask.

Intratracheal oxygen concentration decreased marginally when wearing masks (−0.6 vol% on average). The inspiratory and expiratory resistance increased marginally, on average by 0.4  mbar·s·L⁻¹ and 0.3 mbar·s·L⁻¹, respectively.

This bench study reports findings on gas exchange indices when adding face masks to an intubation training head connected to a breathing lung model.

All masks significantly increased V_dBohr between 89–204 mL which resulted in a significantly increased P_etCO2. These results are in line with those reported by Xu et al. [5], who calculated the added dead space of protective face masks to be between 107.5–167.5 mL, and of Bourassa et al. [12], who studied the effect of gas masks, albeit the latter did not measure dead space. Interestingly, dead space increase was lowest when the mouth of the intubation head was closed.

In contrast, the effect on inspiratory oxygen concentration and resistance to breathing was only marginal even in moderate exercise and even when the masks were tightly taped onto the face.

Perceived difficulties of breathing by mask wearers are therefore unlikely caused by lack of oxygen. It is more likely that the increased concentrations of CO₂ are responsible for reported symptoms of mask wearers such as headache, dizziness, sweating, exhaustion and dyspnoea [7]. Indeed, these very symptoms are described in the literature as the first signs of CO₂ intoxication [13, 14].

In this study, the breathing pattern was kept constant by design, even when P_etCO2 and thus hypothetical arterial CO₂ tension (P_aCO2) significantly increased. The advantage of this design is to study gas exchange without interference. The disadvantage is that responses of the central respiratory control centre are eliminated. However, it is plausible to assume that an increase in P_aCO2 would force an increase in ventilation and, thus, an increase in work of breathing. It is likely that this would be perceived as an increase in workload even though the resistive load was essentially unchanged.

This in vitro study has various limitations. First, factors potentially influencing the conditions of the respiratory system, such as age, sex, body mass and previous illnesses, were not considered. Second, only one sample per type of mask was measured. The data therefore represents a range of masks in general but is insufficient to report on the quality of individual mask types even though differences between masks are likely. Third, data was measured in a lung model and the absolute values represent the physical reality of this model only. Fourth, the lung model was set to represent a “healthy” subject. People with pre-existing medical conditions have more difficulties to cope with additional CO₂ exposure, which is why further studies are necessary for this patient group.

The data show that protective face masks do not reduce inspiratory oxygen concentration but add significant dead space. Reduction of dead space by nasal breathing and tightly fitting masks can reduce the accumulation of CO₂ and contribute to the well-being of the mask wearers.

Shareable PDF