Electronic Supplementary Materials
Table S1 Main characteristics of included studies
Study / Country/center / No. of patients(VL/DL) / Types of video laryngoscopes / Operators / Use of NMB (%)
Griesdale 2012 / Canada/single-center / 40 (20/20) / GlideScope / Medical students or non-anesthesiology residentsinexperienced in ETI / 100
Silverberg 2015 / USA/single-center / 117 (57/60) / GlideScope / Pulmonary and critical care medicine fellows / 0
Janz 2016 / USA/single-center / 150 (74/76) / McGrath MAC (98.6%)
GlideScope (1.4%) / Pulmonary and critical care medicine fellows / 96.7
Lascarrou 2017 / France/multicenter / 371 (186/185) / McGrath MAC / All physicians at the participating ICUs including experts and nonexperts / 100
Abbreviations: VL = video laryngoscopy, DL = direct laryngoscopy, ETI = endotracheal intubation, NMB = neuromuscular blocker, ICU = intensive care unit.
Table S2 Risk of bias assessment of included studies
Study / Random sequence generation / Allocation concealment / Intention-to-treat analysis / Overall risk of biasGriesdale 2012 / Random allocation tablein permutedblocks of 4 / Opaque sealed envelopes / Yes / Low
Silverberg 2015 / Even/odd numbered randomization strategy / No / Yes / High
Janz 2016 / Random permuted blocks of 4, 8, and 12 / Envelopes / Yes / Low
Lascarrou 2017 / Randomization sequence generated by a statisticianat the clinical research unitin permutedblocks of 4 / The software used to collect the data from the electronicreport form automatically allocated the patients, thereby ensuring concealment. / Yes / Low
Blinding of patients, clinicians and study staff was infeasible in these trials because of the nature of the intervention. Thus, trials with high risk of bias for any one or more key domains except blinding were considered as at high risk of bias; trials with low risk of bias for all key domains except blinding were considered as at low risk of bias; otherwise they were considered as at unclear risk of bias.
Table S3 GRADE evidence profile [1]
Quality assessment / No of patients / Effect / Quality / ImportanceNo. of trails / Design / Risk of bias / Inconsistency / Indirectness / Imprecision / Other considerations / VL / DL / Relative
(95% CI) / Absolute
First-attempt success
4 / randomised trials / no serious risk of bias / serious1 / no serious indirectness / no serious imprecision / reporting bias2 / 226/337
(67.1%) / 211/341
(61.9%) / RR 1.17 (0.89 to 1.53) / 105 more per 1000 (from 68 fewer to 328 more) /
LOW / CRITICAL
52.9% / 90 more per 1000 (from 58 fewer to 280 more)
Time for successful intubation(better indicated by lower values)
3 / randomised trials / no serious risk of bias / serious1 / no serious indirectness / serious3 / reporting bias2 / 280 / 281 / - / MD 1.37 lower (33.94 lower to 31.19 higher) /
VERY LOW / CRITICAL
Poor glottis visualization (Cormack-Lehane grade 3/4)
4 / randomised trials / no serious risk of bias / serious4 / no serious indirectness / no serious imprecision / reporting bias2
strong association5 / 26/323
(8%) / 76/326
(23.3%) / RR 0.3 (0.14 to 0.64) / 163 fewer per 1000 (from 84 fewer to 200 fewer) /
MODERATE / IMPORTANT
23% / 161 fewer per 1000 (from 83 fewer to 198 fewer)
Esophageal intubation
3 / randomised trials / no serious risk of bias / no serious inconsistency / no serious indirectness / serious6 / reporting bias2
strong association5 / 4/315
(1.3%) / 14/317
(4.4%) / RR 0.31 (0.11 to 0.9) / 30 fewer per 1000 (from 4 fewer to 39 fewer) /
MODERATE / IMPORTANT
5.3% / 37 fewer per 1000 (from 5 fewer to 47 fewer)
Severe hypoxemia (saturation less than 80 %)
3 / randomised trials / no serious risk of bias / serious4 / no serious indirectness / serious6 / reporting bias2 / 22/307
(7.2%) / 22/317
(6.9%) / RR 1.07 (0.34 to 3.32) / 5 more per 1000 (from 46 fewer to 161 more) /
VERY LOW / CRITICAL
8.3% / 6 more per 1000 (from 55 fewer to 193 more)
Hypotension
3 / randomised trials / no serious risk of bias / no serious inconsistency / no serious indirectness / no serious imprecision / reporting bias2 / 22/311
(7.1%) / 19/315
(6%) / RR 1.19 (0.66 to 2.14) / 11 more per 1000 (from 21 fewer to 69 more) /
MODERATE / CRITICAL
9.2% / 17 more per 1000 (from 31 fewer to 105 more)
Mechanical ventilation duration(better indicated by lower values)
3 / randomised trials / no serious risk of bias / no serious inconsistency / no serious indirectness / serious3 / reporting bias2 / 280 / 281 / - / MD 0.85 lower (1.81 lower to 0.11 higher) /
LOW / IMPORTANT
ICU mortality
3 / randomised trials / no serious risk of bias / no serious inconsistency / no serious indirectness / no serious imprecision / reporting bias2 / 91/280
(32.5%) / 95/281
(33.8%) / RR 0.96 (0.76 to 1.22) / 14 fewer per 1000 (from 81 fewer to 74 more) /
MODERATE / CRITICAL
31.4% / 13 fewer per 1000 (from 75 fewer to 69 more)
1Substantial heterogeneity (I2 > 60%)
2Low number of randomized trials and patients
3Assumption that data are normally distributed
4Moderate heterogeneity (I2between 30% and 60%)
5Large effect (RR < 0.5)
6Wide confidence interval
Fig. S1 Forest plot of glottisvisualization
Forest plot of video laryngoscopy versus direct laryngoscopy on poor glottis visualization (Cormack-Lehane grade 3/4 on first attempt).Heterogeneity was assessed using both the I2statisticsandtheQ test with associatedPvalue. Random-effects analysis was used to estimate the summarized relative risk.
Fig. S2 Forest plot of esophageal intubation
Forest plot of video laryngoscopy versus direct laryngoscopy on esophageal intubation.Heterogeneity was assessed using both the I2statisticsandtheQ test with associatedPvalue. Fixed-effects analysis was used to estimate the summarized relative risk.
Fig. S3 Forest plot of time for successful intubation
Forest plot of video laryngoscopy versus direct laryngoscopy on time for successful intubation (in seconds).Heterogeneity was assessed using both the I2statisticsandtheQ test with associatedPvalue. Random-effects analysis was used to estimate the summarized mean difference.
Fig. S4 Forest plot of severe hypoxemia
Forest plot of video laryngoscopy versus direct laryngoscopy on severe hypoxemia.Heterogeneity was assessed using both the I2statisticsandtheQ test with associatedPvalue. Random-effects analysis was used to estimate the summarized relative risk.
Fig. S5 Forest plot of hypotension
Forest plot of video laryngoscopy versus direct laryngoscopy on hypotension.Heterogeneity was assessed using both the I2statisticsandtheQ test with associatedPvalue. Fixed-effects analysis was used to estimate the summarized relative risk.
Fig. S6 Forest plot of mechanical ventilation duration
Forest plot of video laryngoscopy versus direct laryngoscopy on mechanical ventilation duration (in days).Heterogeneity was assessed using both the I2statisticsandtheQ test with associatedPvalue. Fixed-effects analysis was used to estimate the summarized mean difference.
Fig. S7 Forest plot of ICU mortality
Forest plot of video laryngoscopy versus direct laryngoscopy on ICU mortality.Heterogeneity was assessed using both the I2statisticsandtheQ test with associatedPvalue. Fixed-effects analysis was used to estimate the summarized relative risk.
Fig. S8 Trial sequential analysis plot of first-attempt success
In meta-analyses, random errors because of sparse data and repetitive testing of accumulating data can increase the risks of false positive and false negative results. Trial sequential analysishas been introduced to minimize random errors by calculating the required information size and an adjusted threshold for statistical significance in cumulative meta-analysis [2-4]. To examine the reliability and conclusiveness of the available evidence, we conducted trial sequential analysis for trials that compared video laryngoscopy versus direct laryngoscopy for the primary outcome, success of first-attempt intubation.
Our assumptions included two-sided testing, type I error of 5%, and power of 80%. Diversity-adjusted information size was calculated based on an average event proportion of 61.9% (211/341) in the direct laryngoscopy groups, and an anticipated relative benefit increment of 20% in the video laryngoscopy groups.
A diversity-adjusted required information size of 2,643 patients was calculated. Only 25.7% (678/2,643) of the required information size has been accrued. The cumulative z curve did not cross the conventional boundary. The information size was too small to produce the inner wedge futility area, indicating the current evidence of video laryngoscopy on first-attempt success is inconclusive.
Supplementary discussion
In this meta-analysis of randomised controlled trials in ICU patients, video laryngoscopy did not improvethe frequency of successful first-attempt orotracheal intubation compared with direct laryngoscopy. Video laryngoscopy did generate better visualization of the glottis, but this didnot translate into a higher first-attempt success rate. Thus, it was unsurprising that intubation-related morbidity and mortality rates were unchanged.
The explanationsof the non-significant results may be various. Looking at the reported reasons for intubation failures we found that poor glottis visualization accounted for most cases of intubation failure with direct laryngoscopywhile difficulty in inserting the endotracheal tube appeared the commonest reason for failed intubation with video laryngoscopy. This is shown in two recent trials [5, 6], and is supported by previous data [7]. There are concerns that the use of video laryngoscopy can lead to the creation of visual and cognitive blind spots [8]. The pharynx and hypopharynx are invisible during video laryngoscopy because the camera is located at the tip of the blades. Manipulating the endotracheal tubeinto view canbe difficult because it occurs within this blind spot. This phenomenon has been linked to higher rates of pharyngeal soft tissue injury [9]. Cognitiveblind spot means that video laryngoscopy operators may fail toabort an attempt in a timely manner becausethey have such a clear view of the larynx. This may prolong the patient’s apnea time, leading to severe hypoxemia.
Our findings should be interpreted in light ofthe limitations of current body of evidence. First, the heterogeneity of trial results is substantial for many outcomes, including the primary outcome. Of the four trials included, only one trial observed significantly higher first-attempt success rate in the video laryngoscopy group while three did not. Among several factors that may explain the discrepancy, an obvious one is the routine use of neuromuscular blockers in the three trials with negative findings. Neuromuscular blockade is associated with improved glottic view and reduced attempts in endotracheal intubation [10, 11], and the use of neuromuscular blockers may improve the success rates of direct laryngoscopy over video laryngoscopy. Second, the number of available trials is so small that firm conclusions cannot be drawn for most outcomes. The trial sequential analysis found that a diversity-adjusted required information size of 2,643 patients was need, but only 25.7% of that has been accrued. Third, the intubation experiences of operators vary across trials, but are generally insufficient, especially in video laryngoscopy. For example, in Janz’s trial, the median number of times operator has previously used the video laryngoscope at the time of intubation was only 10 [6]. However, the use of video laryngoscope for intubation requires lots of skills and experiences, and the learning curve can be long. The potential advantage of video laryngoscope may not be easily demonstrated in trials of inexperienced users.
To answer the question of video laryngoscopy versus direct laryngoscopy for ICU intubations, more high-quality randomised controlled trials with adequate sample sizes should be conducted, including trials in more experienced users of video laryngoscopy. Besides, given the weakness of video laryngoscopy in tracheal catheterization, several supplementary devices have been introduced, such as video laryngoscope with an intubation channel, gum elastic bougie, and intubation stylet [12-14]. Evidence on the efficacy of these devices is limited, and larger trials are needed.
Supplementary references
1. Guyatt GH, Oxman AD, Vist GE, et al (2008) GRADE: an emerging consensus on rating quality of evidence and strength of recommendations. BMJ 336:924-926. Doi:10.1136/bmj.39489.470347.AD.
2. Wetterslev J, Thorlund K, Brok J, Gluud C (2008) Trial sequential analysis may establishwhen firm evidence is reached in cumulative meta-analysis. J Clin Epidemiol61:64-75.Doi:10.1016/j.jclinepi.2007.03.013
3. Brok J, Thorlund K, Wetterslev J, Gluud C (2009) Apparently conclusive meta-analysesmay be inconclusive - Trial sequential analysis adjustment of random error risk due to repetitive testing of accumulating data in apparently conclusive neonatal meta-analyses. Int J Epidemiol38:287-298.Doi:10.1093/ije/dyn188
4. Brok J, Thorlund K, Gluud C, Wetterslev J (2008) Trial sequential analysis revealsinsufficient information size and potentially falsepositive results in manymeta-analyses. J Clin Epidemiol61:763-769.Doi:10.1016/j.jclinepi.2007.10.007
5. Lascarrou JB, Boisrame-Helms J, Bailly A,et al (2017) Video laryngoscopy vs direct laryngoscopy on successful first-pass orotracheal intubation among ICU patients: a randomized clinical trial. JAMA317:483-493. Doi:10.1001/jama.2016.20603
6. Janz DR, Semler MW, Lentz RJ,et al (2016) Randomized trial of video laryngoscopy for endotracheal intubation of critically ill adults. Crit Care Med44:1980–1987. Doi:10.1097/CCM.0000000000001841
7. Griesdale DE, Chau A, Isac G, et al (2012) Video-laryngoscopyversusdirect laryngoscopy in critically illpatients: a pilot randomized trial. Can J Anaesth59:1032–1039. Doi:10.1007/s12630-012-9775-8
8. O'Gara B, Brown S, Talmor D (2017) Video laryngoscopy in the intensive care unit. Seeing is believing, but that does not mean it’s true. JAMA317:479-480. Doi:10.1001/jama.2016.21036
9. Greer D, Marshall KE, Bevans S, Standlee A, McAdams P, Harsha W (2017) Review of videolaryngoscopy pharyngeal wall injuries. Laryngoscope 127:349–353.Doi:10.1002/lary.26134
10. Mosier JM,Sakles JC,StolzU,etal (2015) Neuromuscularblockadeimprovesfirst-attemptsuccessforintubationintheintensivecareunit.Apropensitymatchedanalysis.Ann Am Thorac Soc12:734–741. Doi:10.1513/AnnalsATS.201411-517OC
11. WilcoxSR,BittnerEA,ElmerJ,etal (2012) Neuromuscularblocking agentadministrationforemergenttrachealintubationisassociatedwithdecreasedprevalenceofprocedure-relatedcomplications.Crit CareMed40:1808–1813. Doi:10.1097/CCM.0b013e31824e0e67
12. Kleine-BrueggeneyM, Greif R, Schoettker P, et al (2016) Evaluation ofsix videolaryngoscopes in 720patients with asimulated difficult airway: amulticentrerandomized controlled trial.Br J Anaesth116:670–679. Doi:10.1093/bja/aew058
13. Noguchi T, Koga K, Shiga Y, Shigematsu A (2003) The gumelastic bougie eases tracheal intubationwhile applying cricoid pressure compared to astylet.Can J Anaesth50:712–717.Doi:10.1007/BF03018715
14. Batuwitage B, McDonald A, Nishikawa K,Lythgoe D, Mercer S, Charters P (2015) Comparisonbetween bougies and stylets for simulated trachealintubationwith the C-MACD-bladevideolaryngoscope.Eur J Anaesthesiol32:400–405. Doi:10.1097/EJA.0000000000000070