Classifying videolaryngoscopy: standardise it, but describe the failure

What the paper says

Videolaryngoscopy is now the default in many departments, and a documented previous difficult intubation is one of the strongest predictors of future difficulty. What we write down after a videolaryngoscopic intubation therefore matters. Grün and colleagues from Hamburg systematically reviewed the tools proposed for classifying it.¹

They found 13 eligible studies covering seven tools:

• the Cormack–Lehane classification
• the percentage of glottic opening (POGO) score
• the intubation difficulty scale
• the Fremantle score
• the Video Classification of Intubation (VCI)
• the VIDIAC score
• the paediatric PeDiAC score

Only eight of the studies were in patients. The rest were analyses of images or videos after the event.

The central finding is that most tools have been tested only for agreement between raters, not for whether they classify anything accurately. Accuracy or discrimination was reported for only three tools:

• Cormack–Lehane performed fairly in adults (AUROC 0.75) and poorly in children (AUROC 0.68).
• VIDIAC performed well in adults, with internal and external validation (AUROC 0.92 and 0.93).
• PeDiAC performed well in children (AUROC 0.97).

POGO, the Fremantle score and the VCI had no reported accuracy data and no empirically derived thresholds. Risk of bias was high overall, particularly for the reference standard (17 of 18 assessments).

The authors make an important distinction: reliability is not accuracy. Two people agreeing on a score does not show that the score tells the next clinician anything useful. They also note that the VCI's reliability was measured between the operator and an observer at the same intubation. That says little about how well the score communicates to a different clinician months later.

What is VIDIAC?

The VIDIAC (videolaryngoscopic intubation and difficult airway classification) score comes from the same Hamburg group. It was developed to classify the difficulty of videolaryngoscopic intubation.³ The study included 374 videolaryngoscopic intubations in 320 adults with anticipated difficult airways having ear, nose and throat or maxillofacial surgery. The primary outcome was whether the anaesthetist issued a difficult airway alert afterwards, which happened after about half of the intubations. Statistical modelling selected six findings at intubation associated with an alert:

• impaired epiglottic movement
• increased lifting force
• the need to lift the epiglottis directly
• vocal cords clearly visible
• vocal cords not visible
• enlarged arytenoids

Each finding scores plus or minus one. The total is graded as easy (0 or less), moderate (1), hard (2) or severe (3 or more). It discriminated better than Cormack–Lehane (AUROC 0.92 vs 0.75) and has since been validated externally with different blade types and in children.⁴ The group has now also tested it for direct laryngoscopy.⁵

What is valuable

• It is the first rigorous attempt to assess these tools as diagnostic tests, using recognised methods (PRISMA-DTA, QUADAS-2, QAREL and PROBAST), rather than simply listing them.
• It separates reliability, concordance, accuracy, thresholds and clinical utility clearly. The definitions in its first table are worth reading on their own.
• It shows plainly how thin the evidence is for tools many of us use every day.
• It treats airway documentation as clinically important, not a box-ticking exercise, because the note we write shapes the next anaesthetist's plan.

What to bear in mind

• The tools differ in how they were built.
  • VIDIAC and PeDiAC were derived statistically against a defined outcome; for VIDIAC, whether the anaesthetist issued a difficult airway alert. That is why their accuracy and thresholds could be measured.
  • The VCI, POGO and the Fremantle score were designed as descriptions of what was seen and done, with no reference standard to test them against.
  • Judging the descriptive tools by sensitivity and specificity first requires agreement on what "difficult" means, and the review shows that this is where most studies fall down.
• VIDIAC's reference standard, the decision to issue an alert, is itself a clinical judgement. It was also derived in a selected surgical population with anticipated difficulty.
• Three of the authors co-authored included studies, and VIDIAC was developed by their group. They declare this, and the methods are robust, but readers should be aware of it.
• The review searched to December 2025, so it predates the large Spanish VCI study.

The background: what the Spanish VCI study adds

Fernández-Vaquero and colleagues applied the VCI score to 5302 videolaryngoscopic intubations across 44 Spanish hospitals.² The operator and an independent observer each scored the VCI's three domains: blade geometry, glottic view using POGO, and tube delivery. Agreement was high (Krippendorff's α 0.86). The findings confirm what many of us see every week: a good view does not mean an easy intubation.

• Even with POGO above 75%, the first strategy was difficult or failed in 252 of 4025 cases (6.3%).
• Conversely, 37 of the 151 intubations with POGO below 25% were easy.

The same data show where the information really sits. The 151 intubations with POGO below 25% made up under 3% of the cohort but accounted for 46 of the 107 failures. Yet a POGO of 0% does not distinguish between Cormack–Lehane grades 2b, 3a, 3b and 4. Nicholas Chrimes has made this point forcefully and publicly.¹² In his view, POGO precisely subclassifies easy views but loses all discriminatory value for the views we actually need to know about, and high inter-rater reliability for information no one needs is not useful. The authors concede part of the point: when the view is poor, they suggest a qualitative description of the anatomy, aligned with Cormack–Lehane, may add clinical context.

My commentary

This review is a welcome piece of discipline. We have spent more than a decade proposing scores for videolaryngoscopy, and I have been part of that conversation. Not long after the Fremantle score appeared in Anaesthesia and Intensive Care,⁹ colleagues and I wrote in the same journal on scoring systems for videolaryngoscopes.¹⁰ More recently I corresponded with Chaggar and colleagues in the European Journal of Anaesthesiology about their VCI pilot.^7,8 Grün and colleagues now show how little of that effort has been tested against what matters: whether the next clinician, reading our note, makes a better decision.

I like the VCI, and the Spanish study is a substantial piece of work. Its strength is standardisation. It asks three questions in the same order every time: what was used, what was seen and what was done. That alone is an improvement on a free-text note that says "grade 1 view, easy" after a hyperangulated blade, a stylet and two attempts. A shared structure also makes airway data usable for audit, registries and departmental dashboards, which free text never will be.

The weakness is at the point that matters most. When intubation is easy, the VCI code says all that is needed. When it is difficult or fails, a letter in brackets is not enough. The next anaesthetist needs to know:

• whether the problem was the view, tube delivery or something else
• what was tried, how many attempts were made, and what worked
• which adjuncts were used, and whether they were planned or a rescue
• the anatomy actually seen, in plain words, when the view was poor
• whether oxygenation was maintained
• what I would do differently next time

It is telling that the tool with the best evidence in this review, VIDIAC, is built from exactly this kind of description: how the epiglottis behaved, how much force was needed, whether it had to be lifted directly, and what the arytenoids looked like. Those are the findings that led experienced anaesthetists to issue an alert, and they are the observations the next anaesthetist would want handed over. The lesson is not that we need a cleverer number. It is that the description is where the information lies.

That is where I agree with Chrimes. Effort, outcome and a subjective appraisal of difficulty cannot be reduced to a number. Where I differ is that I do not think this means abandoning structure. The answer is a standardised framework such as the VCI that captures every intubation consistently, with a mandatory structured description whenever delivery is difficult or fails. This is also where POGO's floor effect needs to be replaced by a description of the anatomy.

Grün and colleagues end by calling for more high-quality research. I would add a practical aim. The test of any classification is not inter-rater agreement at the time of intubation, but whether a colleague reading it a year later chooses a better plan. That is the study we have not yet done.

References

1. Grün C, Dankert A, Wünsch VA, et al. Classification of videolaryngoscopy: a systematic review. Anaesthesia 2026; 81: 1731–41. doi:10.1111/anae.70358
2. Fernández-Vaquero MÁ, Charco-Mora P, Gómez-Ríos MÁ, et al. Association between glottic view and tracheal intubation during videolaryngoscopy: a prospective observational study. Anaesthesia 2026. doi:10.1111/anae.70417
3. Kohse EK, Siebert HK, Sasu PB, et al. A model to predict difficult airway alerts after videolaryngoscopy in adults with anticipated difficult airways – the VIDIAC score. Anaesthesia 2022; 77: 1089–96. doi:10.1111/anae.15841
4. Köhl V, Wünsch VA, Dohrmann T, Krause L, Petzoldt M. Classification of difficult videolaryngoscopic tracheal intubation with different blade types: a prospective external validation study of the VIDIAC score. Anaesthesia 2025; 80: 1137–9. doi:10.1111/anae.16678
5. Dohrmann T, Schäfersküpper V, Grün C, Wünsch VA, Petzoldt M. Classification of difficult direct laryngoscopy using the VIDIAC score: a prospective external validation study. Anaesthesia 2026; 81: 1758–60. doi:10.1111/anae.70382
6. Halliday C, Berry M, Shah SV, et al. The Video Classification of Intubation (VCI) score for videolaryngoscopy: a multicentre international feasibility study. Br J Anaesth 2025; 136: 1616–24. doi:10.1016/j.bja.2025.11.007
7. Chaggar RS, Shah SV, Berry M, et al. The Video Classification of Intubation (VCI) score: a new description tool for tracheal intubation using videolaryngoscopy: a pilot study. Eur J Anaesthesiol 2021; 38: 324–6. doi:10.1097/EJA.0000000000001374
8. Bradley P. Reply to: The Video Classification of Intubation score: a new description tool for tracheal intubation using videolaryngoscopy. Eur J Anaesthesiol 2022; 39: 182–3. doi:10.1097/EJA.0000000000001563
9. Swann AD, English JD, O'Loughlin EJ. The development and preliminary evaluation of a proposed new scoring system for videolaryngoscopy. Anaesth Intensive Care 2012; 40: 697–701. doi:10.1177/0310057X1204000417
10. Bradley WP, Bain C, Mehra R, Symons J. Scoring systems for videolaryngoscopes. Anaesth Intensive Care 2013; 41: 122.
11. O'Loughlin EJ, Swann AD, English JD, Ramadas R. Accuracy, intra- and inter-rater reliability of three scoring systems for the glottic view at videolaryngoscopy. Anaesthesia 2017; 72: 835–9. doi:10.1111/anae.13837
12. Chrimes N. Posts on X (@NicholasChrimes), March–October 2026.

Related commentary: Measuring mask ventilation: capnography first, effort second

This commentary is my own professional opinion, written for anaesthetists and other professional readers. It does not represent Alfred Health, ANZCA or any other organisation with which I am associated, and it is not individual medical advice. The original articles are the work of their authors and are linked above.