Noninvasive mechanical ventilation in high-risk pulmonary infections: a clinical review

NIV in SARS

SARS, caused by the novel coronavirus (SARS-CoV), emerged in November 2002 and was first identified by the WHO in mid-March 2003. SARS is a severe and progressive disease with ∼25% of patients developing respiratory failure [9].

In order to reduce endotracheal intubation (ETI) and its associated complications, NIV was commonly employed in the SARS outbreak in China, including Hong Kong. The use of NIV has become part of the standard treatment protocol for SARS [10]. A summary of clinical studies and corresponding recommendations are described in table 1.

NIV does not demonstrate absolute benefit among non-chronic obstructive pulmonary disease (COPD) patients with acute hypoxaemic respiratory failure or acute respiratory distress syndrome (ARDS), e.g. in SARS patients [8, 9, 11–17]. It is well known that ETI is associated with higher risk of disease transmission and associated complications. The use of NIV as initial ventilator support for respiratory failure in the presence of SARS appears to be a reasonable option, albeit under strict infection control measures.

NIV in influenza A H1N1 infection

Pandemic influenza A (pH1N1) is a new strain of influenza virus that was first identified in Mexico and the USA in the early part of 2009 [18–30]. The pH1N1 virus originated from the swine influenza (H1N1) virus circulating in North American pigs [23].

Animal studies have shown that the novel influenza virus caused increased morbidity and replicated to higher titers in lung tissue; thus, explaining its pathogenicity and ability to invade the lower respiratory tract in humans, resulting in rapid and fulminant respiratory failure [19]. Such clinical deterioration is characterised by sudden and rapidly progressive respiratory failure with persistent and refractory hypoxia, bilateral diffuse pulmonary infiltrates and a low arterial oxygen tension (P_aO2)/inspiratory oxygen fraction (F_IO2) ratio reaching ARDS criteria [20–22, 25, 26, 31, 32]. Severe respiratory failure is common during the first week and the incidence decreases over time. Refractory hypoxia was the major cause of death, followed by multi-organ failure and shock [26–28, 32].

About 10–30% of hospitalised patients required admission to the intensive care unit (ICU) during the pH1N1 epidemic [27, 28, 32]. Comorbidities were noted in 32–84% of cases admitted to the ICU. These included obesity, COPD, diabetes mellitus, asthma, immunosuppression, chronic kidney disease and heart failure [27–30, 32]. The overall ICU mortality rate for critically ill cases of pH1N1 was close to 17% [19–33]. Factors that were independently associated with mortality included requirement for invasive mechanical ventilation (IMV) and a low P_aO2/F_IO2 ratio at ICU admission, the presence of comorbidities and older age [19, 20, 22, 23]. Autopsy findings showed three distinct pulmonary pathologies: diffuse alveolar damage, necrotising bronchiolitis, and diffuse alveolar damage with alveolar haemorrhage [21, 23, 24, 31]. If we analyse ventilatory management, IMV with lung protective ventilatory strategy and fluid restriction is recommended as the initial approach for managing patients with pH1N1 infection complicated by ARDS [21, 24, 26, 27, 32, 33].

Based on the guidelines from the European Respiratory Society (ERS)/European Society of Intensive Care Medicine (ESICM), WHO, the UK National Health Service, Hong Kong Lung Foundation and the American Association for Respiratory Care (AARC), NIV is not to be used as first-line therapy in pH1N1-associated respiratory failure for the following reasons. 1) Poor clinical efficacy in severe respiratory failure that rapidly progresses to refractory hypoxaemia and ARDS. 2) The prevalent pattern of hypoxaemic failure instead of hypercapnic respiratory failure in patients with pH1N1. 3) Concern about aerosol droplet particle dispersion and spread of infection [10, 19, 34, 35].

A summary of published studies on NIV in pH1N1 is shown in table 2 [19–33, 36–48]. Of the 22 studies included in table 2, the majority were case series reports. The contribution to pH1N1 case series according to country is shown in table 3. There is only one international study [37] that has shown a favourable result with the use of NIV in pH1N1. Five of the studies, most of which were case series applying a face mask for NIV, yielded a failure rate of 10–15%. A full face mask was the NIV interface choice in all but one study; in which the NIV helmet and face mask were used as an alternative. The results regarding the success of NIV were highly variable. Among the causes of NIV failure are refractory hypoxaemia and shock, with multi-organ failure being the most unfavourable. As shown in table 2, mortality is clearly lower in the group treated with NIV versus the IMV group. While the Chinese and European studies appear to argue in favour of NIV use, the Canadian study did not support its use. These studies do not report on significant complications for patients and healthcare workers and in none of these studies do the authors report cases of infection transmission to healthcare workers.

Based on the review of these reports our conclusions are as follows. NIV has a role in the management of early respiratory failure due to pH1N1 infection; however, in a strictly controlled environment with close monitoring of healthcare workers and with adequate precautions for prevention of infection transmission [20, 22, 24–26, 29–33, 36–48]. NIV has no role in severe respiratory failure and ARDS related to severe pH1N1 infection. These patients must undergo ETI and IMV [26, 30, 32, 33, 42–44, 46].

NIV in pulmonary TB

NIV remains an option for ventilatory support in both acute and chronic respiratory failure secondary to pulmonary TB [49–54]. In acute situations, judicious use of NIV in carefully selected patients can potentially obviate the need for IMV [49–53, 55–58]. NIV can be regarded as a simple and suitable solution not only for the provision of long-term and domiciliary-assisted ventilation in patients with sequelae of pulmonary TB, but also for acute exacerbations of pulmonary TB [50, 54, 59, 60]. The concern with the use of NIV in such acute situations is linked with the potential risk of the transmission of TB [49, 56]. Whether or not NIV should be considered a high-risk procedure in infectious diseases like TB continues to be a controversial issue [49, 56]. However, the limited availability of invasive ventilators and/or ICU beds, especially in developing countries where TB incidence remains high, is an important constraint in managing patients with TB and acute respiratory failure by means of NIV.

The medical literature offers very few published reports concerning the use of NIV in contagious TB patients. This is due to the fact that there is a limited number of published articles where NIV was used in exacerbations of chronic respiratory failure in patients with TB sequelae [50, 54, 59, 60], and even fewer reports were available about the use of mechanical ventilation in acute respiratory failure due to active pulmonary TB [56, 57].

In the seven published studies dealing with patients suffering acute exacerbations of pulmonary TB sequelae, nasal mask was the most frequently used interface for delivery of NIV (used in four studies). These studies suggest no apparent complications for patients and workers and report no cases of death in treated patients. Continued applications at home following the acute phase yielded similar results. A summary of published data on NIV in acute exacerbation of pulmonary TB is shown in table 4.

In summary, the practitioner who has to decide whether to initiate NIV or IMV in acute respiratory failure due to pulmonary TB has to take into account several factors, such as: the prolonged period of infectiousness of these patients; the lag time for noticeable clinical improvement following initiation of treatment; and the risks related to positive pressure ventilation, such as pneumothorax and haemoptysis [49].

Risk of infection spread with NIV in pandemics

During the SARS pandemic, nearly half of all SARS victims were healthcare workers, highlighting the occupational risks associated with the care of infectious pneumonias and pandemic infections [11–14]. Evidence related to airborne transmission of the SARS virus and the potential risk of aerosol generation with the use of NIV in SARS still remain controversial [8, 9, 11–17].

Fowler et al. [13] examined transmission rates in healthcare workers caring for SARS patients who required ETI and IMV. This study showed a greater risk of developing SARS in physicians and nurses performing ETI (relative risk 13.29 (95% CI 2.99–59.04); p=0.03). Although nurses caring for infected SARS patients who are on NIV may have been at increased risk of droplet infections (RR 2.3 (95% CI 0.25–21.76)), this theoretical health-related risk was not translated into statistically significant clinical events (p=0.5). Two other reports analysed cases of SARS in Hong Kong, China, and there were no reported cases of SARS among healthcare workers caring for patients treated by NIV. However, it is unclear how this surveillance was carried out [8, 12, 17, 34].

Yu et al. [17] conducted a case–control study to investigate nosocomial spread of SARS by identifying “super spreading events” in multiple hospitals in China, including Hong Kong [8, 12, 16, 17]. The results were of importance since 71.1% and 74.8% of the infections were attributable to super spreading events in Hong Kong and Singapore (Malaysia), respectively. In the final multiple-logistic model, the use of NIV was one of the six factors found to be associated with the risk of dissemination [20, 21]. However, the significance of NIV as a contributing factor to a rapidly spreading healthcare event could not be established [8, 12–14, 15–17].

To date, there are no reports demonstrating an increased risk of NIV-related dissemination in healthcare workers treating infected patients [10]. During the SARS outbreak, a study performed in Hong Kong by Cheung et al. [12] examined the efficacy of NIV in early ARDS and also evaluated the infection risk among healthcare workers who had direct contact with patients on NIV. A total of 22 (25%) patients needed NIV and a total of 155 healthcare workers, including doctors, nurses and healthcare assistants, were exposed to these patients on NIV therapy and were regularly screened for signs of infection. Coronavirus serology was obtained in 97% of healthcare workers. NIV equipped with expiratory bacterial and viral filters was provided in isolated cubicles in the ward or ICU, which were centrally air-conditioned, and fitted with exhaust ventilation fans to achieve negative pressure flow. This study concluded that NIV was not only effective in preventing IMV in 70% of patients with acute respiratory failure due to SARS but effectively reduced the ICU length of stay or avoided ICU admission. Moreover, no infection was noted in any of the 155 healthcare workers. Their serology tests for coronavirus were negative. The potential risk of particle dispersion and spread of infection due to NIV must not be overlooked, even though, at present, the data remains inconclusive.

In pulmonary TB, the majority of patients have a relatively long period of infectiousness (weeks to months), until the diagnosis is established and an effective anti-TB therapy is initiated [49]. There are certain characteristics in TB patients that increase the risk for infectiousness: presence of cough, cavitation on chest radiograph, sputum smear-positive for acid-fast bacilli, and failure to cover the mouth and/or nose during coughing or with cough-inducing and aerosol-generating procedures [49, 50, 56, 58].

Mycobacterium tuberculosis is carried in airborne particles as droplet nuclei that can be generated when persons with pulmonary or laryngeal TB cough, sneeze, shout or sing. The infective particles are ∼1–5 μm in diameter and are maintained in the air and spread throughout the room/building by air currents. Therefore, TB infection is transmitted by air, not by surface contact. Contact with oculonasal secretions, saliva or other fluids from the patient do not represent a mode of transmission of M. tuberculosis [58, 60].

High-risk infections and healthcare worker protocols

The AARC recently issued a statement concerning ventilation in mass casualty situations, including pandemics [63]. The recommendation is against the use of NIV in pandemic influenza as the disease often progresses to ARDS, for which NIV is not the standard of care [63]. The development of infection control protocols for healthcare workers is essential to avoid the risk of acquiring contagious diseases [10, 18, 19, 34, 35, 49, 63, 65]. Protocols should be established in advance, before the onset of a crisis. Protocol should include handling procedures pertaining to aerosol-generating devices such as NIV, environment, equipment disinfection, and transport of infected patients. The designed protocol should be multidisciplinary in nature and involve doctors, nurses, physical therapists and technical personnel [10, 18, 19, 34, 35, 49, 63, 65]. However, these protocols are not intended to supplant physician judgment with respect to special clinical situations or individual patients. NIV must be managed under strict isolation measures in patients suspected of or diagnosed with highly contagious infections, with adequate protection of the healthcare workers who attend to the patient [10, 18, 19, 34, 35, 49, 63, 65].

Infected patients should be isolated in rooms with negative pressure, and visits should be restricted. The isolation time will depend on: the period of replication of the microorganism; comorbidity of the patient, such as immunosuppression; and availability of scientific information [10, 18, 19, 34, 35, 49, 63, 65]. Healthcare workers must wear personal protective equipment, as recommended by the Centers for Disease Control and Prevention’s guidelines. These include N95 masks, gloves, gowns, caps, and face shields or goggles, complemented by strict hand hygiene before and after entering the room and after managing the patient [10, 18, 19, 34, 35, 49, 63, 65]. It is essential that the healthcare workers are familiar with NIV and should have received adequate training for the use of NIV in contagious patients. They should also adhere strictly to the use of personal protective equipment to avoid the risk of acquiring contagious diseases. Recommendations for NIV in SARS, pH1N1 and TB are summarised in table 5.