Oxygen desaturation is a potentially dangerous complication that commonly occurs during endotracheal intubation in the intensive care unit (ICU), emergency department (ED), and out-of-hospital settings.1,2,3,4,5 Desaturation during intubation places patients at risk for cardiac dysrhythmias, hemodynamic decompensation, and cerebral ischemia and may ultimately result in death.6 Despite advances in intubation sequences and techniques,7 emergency intubation is associated with a reported oxygen desaturation rate of 10.9-33.5%.1,2,3,4,5 Severe hypoxemia, defined as an oxygen desaturation of < 80%, can occur in up to 25% of cases.8 Prevention of hypoxemia is critical as a UK national review of emergency intubation identified hypoxemia as a cause of death in 50% of ICU and 27% of ED patients requiring intubation.9 Therefore, additional methods to prevent this life-threatening complication of hypoxemia are essential.

Recently, a technique known as apneic oxygenation (AO) has been described to determine whether its use can reduce the rate of desaturation during an intubation.10 Apneic oxygenation is the passive movement of oxygen to the alveoli without spontaneous or artificial ventilation. This technique provides continuous oxygen delivery using nasal cannulae,11 thereby extending the safe apnea time. Apneic oxygenation has also been shown to prevent hypoxia during intubations performed in the operating room,10,12,13 prolonging the time to desaturation from 3.5-5.3 min using 5 L·min−1 of flow13 and up to 17 min using high flow (60 L·min−1) humidified nasal oxygenation.10

Studies examining the benefit of AO in ED and ICU settings have yielded conflicting results.14,15,16,17 It is challenging to estimate the true rate of desaturation, as studies often utilize inconsistent definitions of hypoxemia.1,2,3,4,5 The primary objective of this systematic review was to examine studies evaluating patients undergoing an emergency intubation (defined as intubation occurring in the ICU, ED, or out of hospital) to determine whether AO (using nasal cannulae) compared with conventional therapy (i.e., no AO and no initial bag valve mask [BVM] therapy) would prevent or minimize oxygen desaturation (as defined in each individual paper). The secondary objective of this review was to establish whether the use of AO results in a lower incidence of severe oxygen desaturation as defined by an incidence of oxygen saturation (SpO2) < 80%.

Methods

Data sources and search strategy

A systematic review was performed using the Preferred Reporting Items for Systematic reviews and Meta-Analysis (PRISMA) guidelines.18 Potential studies were identified by searching three electronic databases (MEDLINE [1950-7 August 2017], EMBASE [1947-7 August 2017], and CINAHL [1961-7 August 2017]). The search strategy was developed in collaboration with an experienced librarian and modified as needed for each database (see Appendix). Additional search methods included hand searching key journals and reviewing the bibliographies of retrieved articles. The final search outcome was collated in Refworks (Proquest Software, Ann Arbor, MI, USA) where duplicate references were removed.

Study selection

All case-controlled, cohort and randomized-control trials (RCTs) were included in the analysis, providing that they met the following inclusion criteria: 1) design, primary study of original data involving human participants; 2) population, adult patients (as individually defined in each paper) undergoing emergency intubation (i.e., in the ICU, ED, or out of hospital); 3) exposure, patients receiving AO via nasal cannulae (either conventional nasal or high-flow nasal cannulae [HFNC]); 4) outcome, oxygen desaturation (as defined by the individual studies), as well as the incidence of severe oxygen desaturation (SpO2 < 80%). Studies were excluded if they did not provide a comparison group to the intervention or if initial BVM ventilation was used during the apneic period or in conjunction with AO. Articles that were identified through the search for which only an abstract was available were excluded if no primary data were included (though an attempt was made to contact the corresponding authors for primary data).

The title and abstract of each included article were screened for relevance by two reviewers (E.T. and O.L.); any studies not meeting the inclusion criteria were excluded. The full texts of remaining studies were assessed by both reviewers. Any disagreements were resolved through consensus; if consensus could not be reached, a third reviewer (N.K.) was consulted to resolve the disagreement.

Data extraction

A standardized data extraction form was created and data were extracted for the following elements: authorship, year of publication, study design, number of patients, patient population, age, body mass index (BMI), gender, definition and incidence of oxygen desaturation, time to intubation, control of the upper airway (defined as an attempt to maintain a patent upper airway during the apneic phase [i.e., a protocol that described the use of a jaw thrust, chin lift, or use of an oral/nasal airway during the apneic period]), incidence of patients experiencing SpO2 < 80%, and the partial pressure arterial oxygen/fraction of the inspired oxygen (PaO2/FiO2) ratio.

Outcomes

The primary outcome was the incidence of oxygen desaturation in patients undergoing emergency intubation with or without AO. This outcome was based on the definition or measurement of oxygen desaturation described by the individual studies. As a secondary outcome, we assessed the incidence of severe oxygen desaturation (defined as SpO2 < 80%), as previously described in studies of critically ill patients.19,20 A planned meta-analysis was performed to combine the results of a subgroup of independent studies based on the common outcome of severe oxygen desaturation < 80%.

Quality assessment

The methodologic quality of each study was assessed by one investigator (E.T.), using the Newcastle-Ottawa Scale (NOS) for non-randomized studies and the Cochrane Collaboration Risk of Bias Tool for RCTs.21,22 Potential for publication bias was assessed by visually inspecting a funnel plot of the risk ratio vs standard error.23

Statistical analysis

Descriptive statistics for categorical variables are presented as frequencies and percentages, and continuous variables are reported as mean (standard deviation [SD]) or median [interquartile range (IQR)]. A meta-analysis was performed on the primary outcome (oxygen desaturation) and the results of a subgroup of independent studies based on the secondary outcome. Continuous variables were pooled with a random-effects generic-inverse-variance method, providing the summary point estimates and associated 95% confidence interval (CI). Dichotomous outcomes were pooled with a random-effect generic-inverse-variance method, providing summary point estimates and associated 95% CI.

Pooled results were reported as relative risk (RR) and a P < 0.05 was considered significant. Between-study heterogeneity of effect size was quantified using the I2 statistic.24 Review Manager (RevMan Version 5.3. Copenhagen: The Nordic Cochrane Centre, The Cochrane Collaboration, 2014) was used to pool and analyze results from the individual studies.

Results

A total of 544 articles were identified from the described search. After removal of 109 duplicates, we screened the titles and abstracts of the remaining 435 articles for eligibility. We identified 34 potentially eligible articles and reviewed the full text of each. Ten studies met all of the inclusion and exclusion criteria. A flow diagram outlining the search strategy and final selection is shown in Fig. 1.

Fig. 1
figure 1

Flow diagram of studies selected for inclusion in the review

Full size image

Study characteristics

Data were extracted from the ten studies involving 2,322 participants. There were four prospective cohort studies,14,15,25,26 four RCTs,16,17,27,28 and two retrospective cohort studies.29,30 Table 1 summarizes the characteristics of each included study. The patient populations in the included studies were diverse, with three studies examining intubation in the ED,15,25,28 two studies describing out-of-hospital intubation,29,30 four studies investigating intubation in the ICU,14,16,17,28 and one in a mixed ICU and ED population.28 Patients were predominately male (55-71%), and mean (SD) age ranged from 45 (21) to 62 (14) yr. Three studies reported mean (SD) BMI16,17,28 that ranged from 26.1 (6.2) to 28.6 (SD not reported), and the mean (SD) severity of hypoxia based on the PaO2/FiO2 was described in two studies16,27 as 118 (59) and 203 (58). Regarding controlling for confounding factors, all trials used neuromuscular blockers for the intubation, with seven using a rapid sequence induction (RSI) as part of their protocol.14,15,25,27,28,29,30 Two studies reported controlling for patency of the upper airway during the intubation procedure, which was described as maintenance of a jaw thrust or chin lift during the apneic phase of the intubation.14,28 Five of the studies recorded the mean (SD) time to intubation14,16,17,27,28 in each population group, which ranged from 34 (22) sec to 150 (SD not reported) sec.

Table 1 Characteristics of included studies
Full size table

Table 2 summarizes the oxygen flow rate of AO through the nasal cannulae used in each study. Two studies compared a variable flow of AO,15,25 five studies 15 L·min−117,26,28,29,30 and three studies compared AO using humidified HFNC at 50-60 L·min−1 with conventional therapy.14,16,27

Table 2 Type of apneic oxygenation and measurement of desaturation
Full size table

Preoxygenation

Table 3 summarizes the preoxygenation techniques described in each of the studies. Overall, preoxygenation using a non-rebreather mask was used in 1,882 (81.5%) patients, BVM in 225 (9.7%) patients, HFNC in 133 (5.7%) patients, non-invasive ventilation in 102 (4.4%) patients, and other methods of preoxygenation in 15 (0.6%) patients. In all study groups that used nasal cannulae for AO, preoxygenation with nasal cannulae was performed in combination with the described preoxygenation technique.

Table 3 Preoxygenation technique prior to intubation as described in the studies
Full size table

Oxygen desaturation

Table 4 summarizes the level of oxygen desaturation experienced by patients in the included studies. In the two studies that defined oxygen desaturation as SpO2 < 93%,26,29 the incidence of desaturation was 13-17% in patients that received AO compared with 20-23% in patients that did not receive the intervention. Five studies defined oxygen desaturation as SpO2 < 90%15,17,25,28,30; in these patients, the incidence of oxygen desaturation was 7-45% in those that received AO and 15-47% in the control group. Three studies defined oxygen desaturation as SpO2 < 80%14,16,27; in these patients the incidence of oxygen desaturation was 2-26% in those that received AO and 14-25% in the control group. A meta-analysis was performed comparing AO vs conventional therapy using the definition of oxygen desaturation as defined in each of the ten included studies. Overall, there was a significant reduction in the incidence of desaturation (RR, 0.76; 95% CI, 0.61 to 0.95; P = 0.02) with low heterogeneity (I2 = 37%, P = 0.11), which favoured AO as an intervention to prevent oxygen desaturation.

Table 4 Definition of desaturation across included studies
Full size table

Severe oxygen desaturation

We performed a sub-group meta-analysis of six studies (four RCTs16,17,27,28 and two observational studies),14,15 which examined the incidence of severe desaturation (SpO2 < 80%). In the AO group, the incidence of severe desaturation was 2-26%, whereas the incidence of severe desaturation was 4-25% in the control group. A meta-analysis was performed for severe oxygen desaturation in the six identified articles. Overall there was no significant reduction in the incidence of severe oxygen desaturation (RR, 0.65; 95% CI, 0.38 to 1.11; P = 0.12) with low heterogeneity (I2 = 39%, P = 0.15) (Figs. 2 and 3).

Fig. 2
figure 2

Incidence of oxygen desaturation as defined in each study. Forest plot comparing apneic oxygenation vs no apneic oxygenation during emergency intubations in ten studies under the random effects meta-analysis. The sizes of the boxes in the figure are in proportion to the weight assigned to each study, whereas the horizontal bar is the reported 95% confidence interval for each risk ratio.

Full size image
Fig. 3
figure 3

Incidence of severe desaturation (SpO2 < 80%). Forest plot comparing apneic oxygenation vs no apneic oxygenation during emergency intubations in six studies under the random effects meta-analysis. The sizes of the boxes in the figure are in proportion to the weight assigned to each study, whereas the horizontal bar is the reported 95% confidence interval for each risk ratio.

Full size image

Risk of bias

The quality assessment of each study is presented in Tables 5 and 6. The risk of bias of the four RCTs was graded as low risk overall (Table 5). For the six non-randomized trials, the NOS scores for the included studies ranged from 7-10 (Table 6). Visual inspection of the funnel plot (Fig. 4) showed some asymmetry indicating that potential for publication bias may exist.

Table 5 Cochrane risk of bias for randomized control trials
Full size table
Table 6 Newcastle-Ottawa scale for risk of bias and quality assessment of included studies
Full size table
Fig. 4
figure 4

Funnel plot of the risk ratio (RR) of included studies against the standard error (SE) of the log RR. Individual study results are represented by open circles. The pooled effect size is represented by the vertical line in the plot. Asymmetry of the funnel plot suggests publication bias.

Full size image

Discussion

This systematic review identified ten studies that compared AO via nasal cannulae with conventional therapy during an emergency intubation. Based upon the definitions used in each study, AO was associated with an overall reduction of oxygen desaturation compared with conventional therapy (RR, 0.76; 95% CI, 0.61 to 0.95; P = 0.02). In our subgroup meta-analysis examining severe oxygen desaturation, AO was not associated with a reduction in the incidence of severe oxygen desaturation compared with conventional therapy (RR, 0.65; 95% CI, 0.38 to 1.11; P = 0.12). Our findings suggest that AO may be more effective than conventional therapy for preventing oxygen desaturation during an emergency intubation.

Life-threatening desaturation is a common complication of emergency intubation, occurring in up to 25% of all intubations in critically ill patients.8 Preoxygenation prior to intubation is required to prevent desaturation during the apneic period after induction and paralysis and during intubation. Non-rebreather masks and bag valve devices are routinely used for preoxygenation in emergency situations.28 Although oxygen desaturation is often a transient event, several studies have determined that desaturation during the intubation procedure is a significant risk factor for cardiovascular collapse during the intubation process.9,31,32 Nevertheless, to date no study has correlated this transient hypoxemia with worse long-term outcomes such as mortality or time on ventilator. Despite this paucity of data to implicate worse outcomes from transient hypoxemia, the high incidence of oxygen desaturation in this patient population highlights the need to develop improved methods to minimize any complications that may arise.

Apneic oxygenation via nasal cannulae is a recently described airway technique, which prevents oxygen desaturation during intubation.11,12,13 First described in 1959,33 AO is a method of providing alveolar oxygen delivery despite the lack of active respiration (apnea). Alveolar oxygen delivery is thought to occur through the generation of a pressure gradient between the lungs and oropharynx. During apnea, there is a discrepancy in the absorption of oxygen and the generation of CO2 in the lung alveoli.34,35 The subsequent gas deficit results in sub-atmospheric pressures in the alveolar space, generating a pressure gradient between the oropharynx (higher pressure) and the lungs (lower pressure), causing bulk flow of air from the upper airway into the alveoli.

Only recently have studies explored the benefits of using AO via nasal cannulae during endotracheal intubation. Results from case control and observational studies suggest that AO may reduce desaturation events in critically ill patients.14,15,29 To date, there have been no documented adverse effects from the use of AO via nasal cannulae including nasal trauma or interference of laryngoscopy.36 Effective AO relies on the availability of an oxygen source, nasal prongs, and a patent upper airway. Patency of the upper airway is a critical factor in the success of AO; an obstructed upper airway prevents oxygen diffusion from the nasal pharynx to the trachea.36

The few studies carried out to date in elective surgery patients have showed a clear benefit for the extension of safe apnea time when using AO.10,13 Conversely, AO literature in critically ill patients has shown mixed results for prolongation of oxygen desaturation during emergency intubation.14,16 Possible explanations for the difference in these findings are the severity of illness in the patient population and variation of airway technique. Patients requiring emergency intubations are generally unstable compared with those seen in the operating room. Sepsis, respiratory distress, and respiratory failure increase the cardiac output and metabolic demand in critically ill patients, which results in a higher rate of oxygen consumption to maintain homeostasis.37 In addition, pulmonary pathology (e.g., infection and pulmonary edema) may lead to pulmonary shunting, which may contribute to the lack of significant benefit of AO in preventing oxygen desaturation in emergency intubation compared with during induction of anesthesia for elective surgery.

Studies of AO during anesthesia may have also benefited from the use of airway protocols. All operating room studies were protocolized in their airway set-up including maintenance of the upper airway with a jaw thrust, chin lift, oral airway, or suspension laryngoscopy.10,12,13 This is in contrast to the majority of studies included in this review, where airway management was the responsibility of each individual physician performing the intubation and the study methodology did not specifically account for patency of the upper airway. Maintenance of the upper airway is critical for AO to work and, as currently described, the maintenance of the upper airway patency is not part of a rapid sequence intubation, though it does recommend cricoid pressure, which may exacerbate an upper airway obstruction.38 As the majority of the studies utilized an RSI for their intubations, it is possible that many of the patients in the reviewed studies did not have a patent upper airway during the intubation procedure, which may impact the demonstrated effectiveness of AO as reported in more stable patients.

It is challenging to determine whether the flow of AO impacted the effect on oxygen desaturation because of the differences between studies. Higher AO flows (i.e., HFNC) have been shown to be effective at prolonging the time to oxygen desaturation time in anesthesia10 as it can help to ensure a higher concentration of oxygen in the nasal/oral pharynx as well as provide some positive end-expiratory pressure during the apneic period.11 The most common flow of AO used in our studies was 15 L·min−1; however, flow rates among studies varied from 5-60 L·min−1. Due to this heterogeneity, it was not possible to perform subgroup analysis to test whether higher flow was superior in AO. When examining studies that used HFNC, we found mixed outcomes—two studies showing no difference,16,27 while a third study showed a possible benefit.14 Therefore, based on the findings of this review, it is unclear whether a higher flow of AO results in greater benefit.

The major limitations of the review stem from the wide heterogeneity of patient factors and measurements among the included studies; therefore, the results should be interpreted with caution. Many studies did not account for conditions known to increase the risk of oxygen desaturation during intubation such as increased BMI and severity of hypoxia.39 In addition, significant heterogeneity existed in the definition of oxygen desaturation—i.e., there were three different definitions (SpO2 < 93, 90, and 80%) used in the studies included. Anticipating that there would be a heterogeneous definition of oxygen desaturation in the literature, we elected to use a predefined and more clinically significant measurement (SpO2 < 80%) of desaturation8 as our secondary outcome and combined the results from six studies in a subgroup meta-analysis. Agreement on a standard definition of oxygen desaturation would improve our ability to compare and generalize the findings of future studies. Another potential limitation was the inclusion of both randomized and non-randomized studies in the meta-analysis. Based upon a growing body of evidence recommending the inclusion of non-randomized studies into meta-analyses, a decision was made to include both study groups in the analysis.40,41,42

The strength of this review lies in the patient selection for emergency intubations. As highlighted above, intubations in the ICU, out of hospital, and in the ED have a larger incidence of severe complications and examining the benefit of AO in this vulnerable patient population is critical. The results from this review are in keeping with results found with other systematic reviews suggesting that AO via nasal cannulae may reduce the overall incidence of oxygen desaturation.43,44,45,46,47 When examining the effects of AO on severe oxygen desaturation there are conflicting results.43,44,45,46,47 Some reviews suggest that AO resulted in a decreased incidence of severe oxygen desaturation,43,44,45 while others have concluded that there was no difference in AO compared with no AO.46,47 The inconsistent results, as well as the significant clinical heterogeneity of the published studies, suggest that more well-designed clinical trials are required before standard implementation of AO during an emergency intubation should be recommended. Future studies would benefit from a standard airway protocol for the intubation process (standard preoxygenation, positioning, and attempts to maintain patency of the upper airway with a jaw thrust, chin lift, or nasal/oral airway). Furthermore, having a consistent definition of oxygen desaturation and accurate documentation of important airway variables, including the time to intubation and severity of hypoxemia for intubation (PaO2/FiO2 ratio), are recommended.

Conclusion

The results of this review suggest that AO via nasal cannulae is associated with a decreased incidence of oxygen desaturation in emergency intubations. To ascertain whether any patient- or system-level factors affect the outcome of AO on oxygen desaturation in emergency intubations, we suggest that further clinical trials are required.