Neuraxial anesthesia is most often administered to parturients for Cesarean delivery, but general anesthesia is still needed in approximately 6% of cases,1 the majority of these being emergent.2 If general anesthesia is required, supraglottic airway devices are not recommended because of the increased risk of aspiration and secondary aspiration pneumonitis,3 and hence tracheal intubation remains the mainstay of airway management.

Compared with the general population, the incidence of difficult or failed intubation is higher in obstetrical patients.4,5,6,7 This is attributed, in part, to anatomic and physiologic changes during pregnancy, including laryngopharyngeal edema, mucosal engorgement, decreased functional residual capacity, as well as environmental and human factors.8 Should difficult or failed intubation be encountered, pulmonary aspiration can still occur with an incidence of 8% in obstetrical patients relative to 1% in matched controls.4,9 Further, difficult and failed intubation has been associated with other complications such as hypoxia, hypertension, the need for an emergency surgical airway, unanticipated intensive care admission, and death.9,10 The possible consequences of a failed intubation in the parturient cannot only affect the mother but also the fetus.4

In the Obstetric Anaesthetists’ Association and Difficult Airway Society guidelines for the management of difficult and failed tracheal intubation, it has been recommended that videolaryngoscopes should be immediately available for all obstetric general anesthetics, and others have suggested that videolaryngoscopy should become the first-line standard of care.11,12 Comparison of videolaryngoscopy with direct laryngoscopy in a meta-analysis showed that videolaryngoscopy reduced the incidence of failed intubation and decreased the occurrence of complications such as airway trauma and hoarseness.13 Only one of the included trials, however, was performed in the setting of obstetrics, probably owing to the difficulties of research in this context where most general anesthetics are for an emergency Cesarean delivery,14 and airway device performance may vary with the circumstances of the difficult airway.

In view of this, we aimed to undertake a mixed-methods systematic review and meta-analysis to examine the efficacy, efficiency, and safety of videolaryngoscopy compared with direct laryngoscopy in parturients. Our primary objectives were to assess the efficacy through the first-attempt success rate of tracheal intubation and the efficiency via the time taken for tracheal intubation.

Methods

The study was registered in the PROSPERO database (CRD42020189521) and followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines.15

We performed a systematic search of the Central, CINAHL, Embase, MEDLINE, and Web of Science Core Collection electronic databases from inception to 27 May 2020, with no restrictions on language or publication type. Controlled vocabulary terms and text words, relating to the main components of the review were chosen, including obstetrics and videolaryngoscopy. The complete search strategy is provided in Appendix 1.

All retrieved article citations were entered into the reference management software program, Rayyan (Qatar Computing Research Institute, 2016, Doha, Qatar), where duplicates were removed and the remainder were screened for eligibility. Case reports and series, prospective and retrospective observational studies, and randomized-controlled trials (RCTs) that reported on the application of videolaryngoscopy to intubate the trachea during general anesthesia in pregnant patients were considered for inclusion. Search results were independently screened by two authors (R.H. and N.D.) using the title and the abstract. The full texts of potentially eligible citations were subsequently evaluated for inclusion. Discrepancies were resolved by discussion and disagreements by a third author (D.O.). Moreover, the reference lists of all included articles were manually reviewed for additional citations by one author (S.H.).

Once all included articles had been identified, two authors (S.H. and D.O.) independently assessed the methodological quality in each non-randomized study with the Risk of Bias in Non-randomised Studies of Interventions (ROBINS-I)16 and in each RCT with the revised Cochrane risk of bias tool for randomized trials (RoB 2).17 Data were extracted by two authors (R.H. and N.D.), discrepancies were resolved by discussion and disagreements settled by a third author (D.O.).

Characteristics extracted from the observational studies, RCTs, and case reports/series where applicable included the following: sample size and number of patients in each study arm; indication for general anesthesia; predicted difficulty of the intubation; type of videolaryngoscope; experience of intubator; and the definition of failed intubation. Case reports/series were qualitatively summarized to provide information on the advantages, disadvantages, and the technical aspects of videolaryngoscopes. With respect to observational studies and RCTs, our co-primary outcomes were the first-attempt success rate at tracheal intubation and the time to tracheal intubation (defined as the total time for insertion of the laryngoscope, glottic visualization, and the placement of the endotracheal tube [ETT] until obtainment of end-tidal carbon dioxide on capnography). Secondary outcomes included the number of attempts at tracheal intubation, time to optimal laryngeal view (defined as the time taken for insertion of the laryngoscope and glottic visualization), time to place the ETT (defined as the time taken for the placement of the ETT until obtainment of capnography), perceived difficulty of tracheal intubation, reported visualization of the larynx with Cormack and Lehane grade or percentage of glottic opening (POGO),18 change in heart rate and mean arterial pressure after induction of general anesthesia, incidence of failed intubation and serious airway complications, as well as the rate of airway or laryngeal trauma, hoarseness, and sore throat.

Continuous outcomes were extracted as the mean and standard deviation (SD). If the mean and the SD had not been reported, we followed the recommendations from the Cochrane Collaboration, approximating the mean to be equivalent to the median, and the SD to be the interquartile range/1.35 or the range/4.19 Dichotomous outcomes were converted to the overall numbers of incidence. Data only presented in graphical format were extracted with a plot-digitizing software program, Plot Digitizer (Version 2.1, Free Software Foundation, 2015, Boston, MA, USA). If we needed to clarify the details of a study’s methodology or request missing data, we contacted the authors of the respective articles up to three times to request these data.

For the RCTs, data were inputted from a standardized data collection form in Microsoft® Excel (Microsoft Corp, Redmond, WA, USA) to Review Manager (Version 5.3, The Nordic Cochrane Centre, 2014, Copenhagen, Denmark) by one author (N.D.), the accuracy of which was confirmed by another author (R.H.). It was our intention to conduct meta-analysis for an outcome of interest if it was reported by two or more RCTs. Statistical heterogeneity (I2) was calculated for each outcome with predetermined thresholds for low (25–49%), moderate (50–74%), and high (more than or equal to 75%) levels.20 If low heterogeneity was found, it was assumed that the true effect of the intervention was the same in every trial and a fixed-effect model was chosen to represent the best estimate of the intervention effect. In the event that moderate or high heterogeneity was present, it was assumed that the effect of the intervention was not the same in every trial and the DerSimonian and Laird random-effects model was chosen to represent the average intervention effect. For continuous outcomes, data were subjected to the inverse-variance method, where the weight attributed to each trial is the inverse of the variance of the effect estimate, resulting in the calculation of a weighted mean difference with its 95% confidence interval (CI). The mean difference was the absolute difference in the mean between the two groups. For dichotomous outcomes, data were subjected to the Mantel–Haenzel method, resulting in the calculation of a risk ratio (95% CI). The risk ratio is the ratio of the risk of an event in the two groups. All tests were two-tailed and performed at 5% significance level.

To evaluate for the risk of publication bias in relation to all outcomes, a funnel plot was drawn and visually examined for symmetry. Our results were verified by Duval and Tweedie’s trim and fill test, in which the smaller studies producing funnel plot asymmetry are removed and the omitted trials and their missing counterparts are replaced, and by Egger’s linear regression test using Comprehensive Meta-Analysis (Version 3.3, Biostat, 2014, NJ, USA).

Results

Of the initial 526 unique article citations identified by the search strategy, 35 case reports and series,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,51,52,53,54,55 seven cross-sectional studies,56,57,58,59,60,61,62 one cohort study,63 one crossover study,64 and four RCTs fulfilled our inclusion criteria.65,66,67,68 Details of the screening process are illustrated in Fig. 1. In the 36 instances where we needed to clarify details of trial methodology or request missing data, nine authors responded with the required information.24,38,40,41,42,51,55

Fig. 1
figure 1

PRISMA flow diagram summarizing the retrieved, included, and excluded randomized-controlled trials. PRISMA = Preferred Reporting Items for Systematic Reviews and Meta-Analyses

Full size image

Characteristics of the case reports and series are detailed in Table 1. The case reports/series, describing the use of videolaryngoscopes in 100 patients, provided an indication of the potential value of videolaryngoscopy as a first choice and rescue modality in patients with and without predictors of difficult intubation. Compared with direct laryngoscopy, videolaryngoscopy has been used to improve laryngeal visualization and success at tracheal intubation in patients with morbid obesity and unfavourable airway anatomy (e.g., restricted mouth opening). Videolaryngoscopy has further been utilized in the setting of awake tracheal intubation.24,48 The experience of the intubator was not reported in most cases. Further, our search uncovered some limitations of videolaryngoscopy, including the inability to fully insert a channelled videolaryngoscope (e.g., Airtraq™; Prodol Meditec S.A., Vizcaya, Spain that tends to be bulkier to accommodate the channel for the ETT) in the presence of restricted mouth opening, and the development of complications such as split ETT T-shaped connectors following insertion of the Gliderite® stylet which accompanies the Glidescope® (Verathon; Bothell, Australia), palatal perforation, and postoperative hoarseness.27,28,42

Table 1 Characteristics of included case reports and series
Full size table

In five cross-sectional studies that examined the incidence of difficult and/or failed intubation in obstetrics,56,57,58,59,60 only one defined difficult intubation (i.e., Cormack and Lehane grade of 3 or 4).57 Kimura showed a difficult intubation incidence of 2.8% (eight in 286 patients), including two patients who were eventually intubated with an unspecified videolaryngoscope (after a first-attempt with direct laryngoscopy), and none of the 286 patients had a failed intubation.57 John et al. found the rate of difficult and failed intubation to be 1.9% (one in 52 patients) and 0.48% (one in 210 patients), respectively.60 Of the four patients who were difficult to intubate, three occurred out of normal working hours, one took place in an elective Cesarean delivery, and all four had subsequent successful tracheal intubation with the McGrath™ videolaryngoscope (Medtronic; Watford, UK). In the one patient where tracheal intubation failed, intubation of the trachea was attempted three times with a videolaryngoscope. Fennessy et al. reported the incidence of difficult intubation to be 0.36% (four in 1,113), and none of the 1,113 patients had a failed intubation.58 Of the four cases who were difficult to intubate, the most senior anesthesiologist present was a consultant in two cases and a speciality registrar in the other two cases. In one patient, the McGrath videolaryngoscope was selected from the start as the patient had retrognathia, and the remaining three had subsequent successful tracheal intubation following use of direct laryngoscopy. Rajagopalan et al. showed that the C-MAC® (KARL STORZ; Tuttlingen, Germany) or the Glidescope were used for the first intubation attempt in ten cases where the intubation was predicted to be difficult.59 Freedman et al. reported six failed intubations in their local hospital over a 20-year period, one of which was subsequently intubated successfully with an unspecified videolaryngoscope by a second anesthesiologist.56

Characteristics of the cohort and crossover studies and RCTs are presented in Table 2. The risk of bias was assessed with ROBINS-I for non-randomized studies (Appendix 2) and RoB 2 for RCTs (Fig. 2). Aziz et al. was the only non-randomized study where ROBINS-I could be conducted, and it was evaluated to be at serious risk of bias due to the possibility of confounding in the study design.63 Kurvey et al. did not present sufficient information in their published abstract to fairly judge the risk of bias.64 Some concerns were uncovered in all RCTs for the potential of bias in the measurement of the outcome as assessors were not blinded to the allocated interventions.

Table 2 Characteristics of included cohort studies and randomized-controlled trials
Full size table
Fig. 2
figure 2

Risk of bias assessment of included randomized-controlled trials using the revised Cochrane risk of bias tool. ! = unclear risk; − = high risk; + = low risk

Full size image

Overall, the RCTs comprised a total of 428 patients, in whom direct laryngoscopy was used in 184 and videolaryngoscopy in 244. All of the RCTs, unlike in the cohort studies, excluded patients with features suggestive of predicted difficulty in intubation. Further, the intubators were trainees in the cohort studies,63,64 while they were consultants, one trainee or unspecified in two, one, and one RCT, respectively.65,66,67,68

In a crossover simulation study, when a mannequin was prepared to simulate an easy obstetric airway with a corresponding Cormack and Lehane grade 1 laryngeal view, all 30 anesthesiologists were able to successfully intubate the trachea, and no differences were shown between direct laryngoscopy and use of the GlideScope in the average time to tracheal intubation (32.2 vs 35.1 sec; P = 0.26).64 If, however, the mannequin was set up to simulate a difficult obstetric airway (with a corresponding Cormack and Lehane grade 3 laryngeal view), the GlideScope was superior to direct laryngoscopy in the rate of successful tracheal intubation (96.7% vs 66.7%; P = 0.01) but not the time to tracheal intubation (33.1 vs 39.8 sec; P = 0.06). Of 180 intubations in an obstetric unit over a three year period, direct laryngoscopy was chosen as the first-line technique in 163, six of which failed and were rescued with a videolaryngoscope.63 The 18 patients where a videolaryngoscope was selected from the outset were more likely to have predictors of difficult intubation and were all intubated successfully on the first-attempt.

In the RCTs, our first co-primary outcome, the first-attempt success rate at tracheal intubation, was reported in 417 patients by three trials.65,67,68 No difference was shown between videolaryngoscopy and direct laryngoscopy (risk ratio, 1.02; 95% CI, 0.98 to 1.06; P = 0.29; I2 = 0%) (Fig. 3). Duval and Tweedie’s trim and fill test indicated the presence of publication bias (Appendix 3). Our second co-primary outcome, the time to tracheal intubation, was reported in 417 patients by three trials.65,67,68 No difference was found between videolaryngoscopy and direct laryngoscopy (mean difference, 1.20 sec; 95% CI, -6.63 to 9.04; P = 0.76; I2 = 95%) (Fig. 4). Neither the Duval and Tweedie’s trim and fill test nor the Egger’s test suggested the presence of publication bias. Regarding the secondary outcomes, the results of the meta-analyses are presented in Table 3. Sensitivity analysis showed that the use of a fixed-effect model to analyze outcomes when the statistical heterogeneity was less than 50% did not markedly influence the direction, magnitude, or statistical significance of the results. Data were insufficient to facilitate the meta-analysis of the other outcomes. No difference was found between direct laryngoscopy and videolaryngoscopy in the perceived difficulty of tracheal intubation or the change in heart rate and mean arterial pressure.67 Failed intubation or serious airway complications did not occur in any of the trials.

Fig. 3
figure 3

Forest plot of the first-pass success rate at tracheal intubation. For each trial, the square depicts the risk ratio and the horizontal lines either side of it represent the 95% CI. The summary result is presented as a diamond. CI = confidence interval; DL = direct laryngoscopy; SD = standard deviation; VLS = videolaryngoscope

Full size image
Fig. 4
figure 4

Forest plot of the time to tracheal intubation. For each trial, the square depicts the mean difference and the horizontal lines either side of it represent the 95% CI. The summary result is presented as a diamond. CI = confidence interval; DL = direct laryngoscopy; SD = standard deviation; VLS = videolaryngoscope

Full size image
Table 3 Meta-analysis of the secondary outcomes. Values are mean difference or risk ratio
Full size table

In one comparative study, after discussions and staff training, a decision was made to change from direct laryngoscopy to first-line universal videolaryngoscopy with the GlideScope in all operating rooms, including obstetrics.61 Support from the anesthetic department for this paradigm shift in intubation technique increased from 33% prior to the trial phase to 100% 30 months later. The benefits reported included better training for anesthetic assistants and trainees, improved communication and human factors and, importantly, no reported requirement for cricothyroidotomy since 2014. In a tertiary obstetric unit where the McGrath was introduced, it was utilized in 66 of 100 cases and, although predictors of difficult intubation were present in ten of these, all tracheal intubations were successful.62

Discussion

In our meta-analysis of RCTs, no difference was shown between direct laryngoscopy and videolaryngoscopy in either the first-attempt success rate at tracheal intubation, despite improved laryngeal visualization with the latter, or the time to tracheal intubation in obstetrics. The synthesis of observational studies identified the usefulness of videolaryngoscopes as a first-line device for patients with characteristics indicative of a difficult airway and as a rescue tool after failed direct laryngoscopy.

It is likely that the optimal conditions afforded by the exclusion of parturients with characteristics indicative of a possible difficult airway, nature of elective Cesarean delivery, and the experience of intubators contributed towards the first-attempt success rate of 96% with direct laryngoscopy. If videolaryngoscopy were to increase this first-attempt success rate at tracheal intubation to 99%, then more than 800 parturients (as opposed to the 417 included in our analysis) would be needed to attain statistical significance. Consistent with this, under simulation conditions, differences were not found between direct laryngoscopy and videolaryngoscopy in the rate of successful tracheal intubation in a mannequin that had an easy obstetric airway.64 In contrast, however, videolaryngoscopy was superior to direct laryngoscopy in the rate of successful tracheal intubation when the mannequin was prepared to simulate a difficult obstetric airway. In the face of a decreased functional residual capacity and the increased oxygen consumption in a pregnant patient that shortens the safe apneic time,69 it is important to acknowledge that, although videolaryngoscopes did not reduce the time to tracheal intubation, they did not increase it. It has been previously reported that videolaryngoscopy requires the application of less force to the tongue base than direct laryngoscopy does and, as a surrogate measure of the stress response, this could potentially translate into reduced sympathetic stimulation,70,71 something that was not shown in our systematic review.

In order to evaluate the benefits and the drawbacks of videolaryngoscopy vs direct laryngoscopy in obstetrics, we must study a broader cohort indicative of the characteristics of patients we encounter in clinical practice, including those with predictors of difficult intubation having an emergency Cesarean delivery under the care of anesthesiologists with varying experience. Evidence from heterogenous obstetric populations in cross-sectional and comparative studies uncovers the potential role that videolaryngoscopy may have. There are many examples of patients who had direct laryngoscopy that failed and was rescued with videolaryngoscopy,56,57,58,59,60,62,63 as well as those where a videolaryngoscope was chosen as a first-line device when a difficult airway was possible.58,59,60,62,63 Only one study selected a videolaryngoscope as a first-line modality even in the absence of predictors of difficult intubation.62 The success rate for tracheal intubation with videolaryngoscopy across these studies was high, with one failure in a patient who was predicted to have a difficult airway.60 Given that the incidence of difficult intubation may be over 3%,59 immediate access to a videolaryngoscope in the labour ward might be appropriate, especially in view of the emergent nature of many of these Cesarean deliveries. In the UK, however, just over half of obstetric units in 2015 reported the presence of an available videolaryngoscope,72 and in a more recent prospective and multicentre observational study, videolaryngoscopy was used for only 1.9% of parturients who required general anesthesia.73

We suggest that the lack of difference shown between direct laryngoscopy and videolaryngoscopy in our meta-analysis for patients not predicted to have a difficult airway and scheduled for elective surgery should encourage an overall increased adoption of videolaryngoscopy in obstetrics. In doing so, we would obviate the need for a risk stratification to determine which laryngoscope is chosen as first-line, particularly as the positive predictive value and the sensitivity of airway examination is low,74 and have the potential to decrease the incidence of rescue intubations in what is otherwise a stressful and time-critical scenario. Further, this strategy should not increase either the number of attempts at tracheal intubation or the time to tracheal intubation in those patients predicted to have an easy airway. The use of videolaryngoscopy in these situations would contribute to the maintenence of proficiency needed to facilitate optimal and skilled use in difficult scenarios. Such implementation has already been achieved in some obstetric units. Indeed, one hospital that introduced videolaryngoscopes in all clinical areas reported that the requirement for emergency front of neck access reduced from one to two patients per year to none over five years.61 Interestingly, in a decision analysis, rapid sequence induction in conjunction with videolaryngoscopy in the context of a category one emergency Cesarean delivery for a patient with anticipated difficult tracheal intubation was shown to be associated with a failure rate of 21 per 100,000 and a shorter time to successful readiness to surgery relative to awake tracheal intubation.75

Most of our findings are consistent with those of prior systematic reviews including either exclusively or mainly non-obstetric patients.13,76,77,78,79 In a meta-analysis of 7,044 patients, including simulation studies, videolaryngoscopy decreased the failure rate of tracheal intubation with predicted difficult airways but not predicted non-difficult airways and with experienced operators unlike inexperienced operators.13 It did not, however, increase the first-attempt success rate at tracheal intubation. Hoarseness and laryngeal trauma were reduced with the use of a videolaryngoscope. In a further systematic review of 1,329 patients with known difficult airways, videolaryngoscopy increased the first-pass success rate at tracheal intubation.78 In all of these meta-analyses, the laryngeal view was improved with videolaryngoscopy,13,76,77,78,79 and in two of them, no difference between direct laryngoscopy and videolaryngoscopy in the time to tracheal intubation was revealed,77,79 with one showing a shorter time to tracheal intubation in those patients who presented with a difficult intubation.79

Strengths and limitations

Some limitations of our systematic review serve to restrict our conclusions. First, the number of RCTs in obstetrics is limited, and it is recognized that there are inherent problems to performing meta-analysis on a small number of trials.80 Despite this, the number of patients in the meta-analysis was disproportionately large for the number of included trials, which is why we decided to proceed with quantification. The statistical analysis for one of the primary outcomes of first-pass success rate was, however, still underpowered. Second, the assessors in the RCTs were not blinded to the allocated interventions resulting in the potential of bias for the measurement of the outcome. One observational study was at serious risk of bias secondary to the possibility of confounding in the study design. It is not possible, however, to blind the assessor and intubator to the nature of the airway device in these RCTs, and so it should be recognized that the introduction of such bias cannot be obviated. Third, the definition of outcomes was not standardized in the RCTs and this was likely responsible, at least in part, for the high statistical heterogeneity associated with the time to tracheal intubation. In one trial, the time to tracheal intubation was defined from the point at which the videolaryngoscope was held by the anesthetist65 and in two trials, it was specified from the time at which the blade of videolaryngoscope was inserted into the mouth.67,68 Fourth, many of the case reports and observational studies were published as abstracts in conference proceedings and thus had not undergone peer review. Lastly, in the absence of sufficient data, we were unable to determine which videolaryngoscope performs best in the setting of obstetrics.

Conclusions

Research into the use of videolaryngoscopy in obstetrics continues to evolve. In parturients without predictors of difficult intubation, videolaryngoscopy compared with direct laryngoscopy in the hands of experienced operators does not affect the first-pass success rate at tracheal intubation with no difference in the time to tracheal intubation. In a mixed cohort of pregnant patients, nevertheless, the videolaryngoscope has shown its utility both as a first-line device for patients with features suggestive of a difficult airway and as a rescue tool after failed direct laryngoscopy. This supports its increased adoption in obstetrics where videolaryngoscopes should be immediately available for use as the first-line device.