European Journal of Cardiovascular Medicine

Unanticipated difficult laryngoscopy and endotracheal intubation remain among the most feared perioperative events because they can rapidly lead to hypoxemia, aspiration, airway trauma, and escalation to emergency rescue techniques despite otherwise routine anaesthesia. The clinical importance of anticipating a difficult airway is reflected in major airway practice guidelines, which emphasize structured preoperative airway assessment, preparation of alternative devices, and a clear plan for failed intubation scenarios to minimize complications[1].

Among bedside screening tools, the Mallampati classification is one of the most widely used because it is simple, quick, and requires no equipment. The original Mallampati test was described as a clinical sign to predict difficult tracheal intubation based on the visibility of oropharyngeal structures during maximal mouth opening and tongue protrusion. [2] Samsoon and Young later modified the system by adding a fourth class (Modified Mallampati Test; MMT), and this four-grade version remains the commonly documented form in anaesthesia practice. [3] In parallel, intraoperative laryngoscopic difficulty has been traditionally described using the Cormack–Lehane grading system, which links the quality of glottic visualization at laryngoscopy to increasing technical difficulty and potential intubation failure. [4]

Despite widespread use, the predictive performance of the Mallampati test is inconsistent across populations and study designs, with many reports showing that Mallampati alone provides only modest sensitivity and may miss a substantial fraction of difficult laryngoscopies or intubations when used as a single predictor. Meta-analytic evidence indicates that common bedside airway tests—including Mallampati—generally demonstrate poor-to-moderate sensitivity with moderate-to-fair specificity, and their clinical value improves when combined with other assessments rather than used in isolation. [5] This limitation is clinically relevant because low sensitivity increases the risk of “unanticipated” difficulty—precisely the scenario most associated with adverse airway outcomes—while low positive predictive value can lead to unnecessary escalation of equipment and staffing in low-prevalence elective cohorts. Large-scale syntheses focusing specifically on the modified Mallampati score have similarly concluded that it has limited prognostic value as a stand-alone predictor and that variability in test performance is influenced by inconsistent execution and interpretation. [6]

A key procedural variation that may influence Mallampati grading is when the patient is asked to phonate (e.g., saying “ah”) during the assessment. Phonation can alter soft palate position and the apparent oropharyngeal view, potentially shifting Mallampati class and changing diagnostic characteristics. In a prospective comparison evaluating Mallampati testing in different positions and with phonation versus without phonation,  was shown to affect grading and prediction of difficult laryngoscopy and intubation. [7] However, clinical practice remains heterogeneous, and there is no universally standardized recommendation regarding phonation during MMT, creating uncertainty about which technique provides better screening utility. Therefore, the present study was undertaken to compare modified Mallampati assessment performed with phonation versus without phonation for predicting difficult visualization of the larynx and difficult laryngoscopy, using accepted intraoperative reference standards and clinically relevant definitions of difficulty. [8]

AIMS AND OBJECTIVES

Aim

This study compared MMT performed with phonation and without phonation to predict difficult intubation among adults undergoing general anaesthesia.

Objectives

1.Primary: Compare MMT with phonation versus without phonation for predicting difficult laryngoscopy (CL grade III–IV).

2.Secondary: Compare PPV and NPV of both MMT approaches for predicting difficult intubation (IDS-defined).

Study design, setting, and duration A cross-sectional diagnostic accuracy study was conducted over 3 months in the operating theatres of the Department of Anaesthesiology at Sri Devaraj Urs Medical College. Participants Adults aged 18–65 years scheduled for elective surgery under general anaesthesia requiring endotracheal intubation were recruited. Inclusion criteria: adults (18–65 years), patients of either sex, undergoing elective surgery requiring general anaesthesia with endotracheal intubation. Exclusion criteria: patients with peritonsillar/submandibular abscess, airway burns or trauma, restricted neck movement (including cervical spine pathology), temporomandibular joint limitation, oral cavity/airway malignancy, congenital airway anomalies and pregnant females. Sample size The sample size was 95, calculated using kappa-based reliability methodology assuming κ₁=0.598 and κ₀=0.40 with α=0.01 (two-sided) and 90% power. Ethics Ethics committee approval was obtained. Written informed consent was obtained from all participants. Index tests (MMT without and with phonation) In the preoperative area, a first assessor recorded demographics and performed MMT in the sitting position under two conditions: 1. MMT without phonation: maximal mouth opening and tongue protrusion without vocalization. 2. MMT with phonation: maximal mouth opening and tongue protrusion while producing an “ah” sound. MMT grades were recorded as I–IV and dichotomized as easy (I–II) and difficult (III–IV). Reference standards A second assessor, blinded to MMT findings, recorded: • Cormack–Lehane grade during direct laryngoscopy using Macintosh blade size 3 or 4. Difficult laryngoscopy was defined as CL III–IV. • Intubation Difficulty Scale (IDS) as a validated quantitative measure of intubation complexity. Difficult intubation was defined as IDS >5. • Fig A : INTUBATION DIFFICULTY SCORE Statistical analysis Data were analyzed using SPSS v22. Continuous variables were summarized as mean±SD and categorical variables as frequency (%). For diagnostic accuracy, 2×2 tables were constructed for each MMT method against CL III–IV and IDS >5. Sensitivity, specificity, PPV, NPV, and accuracy were calculated. Agreement between MMT methods was assessed using Cohen’s kappa. A two-sided p<0.05 was considered statistically significant.

Participant profile and outcome frequency

Illustrative simulated dataset (N=95): The mean age was 41.6±12.3 years and 58 (52.4%) were male. Difficult laryngoscopy (CL III–IV) occurred in 12/105 (13.3%), while difficult intubation (IDS >5) occurred in 9/105 (9.5%). Most participants had CL grade I–II (86.7%), consistent with elective surgical populations where severe laryngoscopic difficulty is relatively infrequent.

Distribution of MMT grades with and without phonation

MMT without phonation classified 18/105 (19.0%) as Mallampati III–IV, whereas MMT with phonation classified 25/105 (26.7%) as Mallampati III–IV, indicating that phonation shifted a proportion of patients into higher-risk classes. Agreement between the two MMT methods (dichotomized) was moderate (κ≈0.60), suggesting meaningful but incomplete overlap between the two approaches.

Diagnostic performance for predicting difficult laryngoscopy (CL III–IV)

Against CL grade III–IV, MMT with phonation demonstrated higher sensitivity than MMT without phonation (85.7% vs 64.3%), indicating fewer false negatives when phonation was used. Specificity was slightly lower with phonation (82.4% vs 87.9%), consistent with a modest increase in false positives. Both approaches maintained high NPV (>94%), reflecting the low prevalence of difficult laryngoscopy in elective cohorts and the usefulness of MMT primarily for ruling out difficulty rather than confirming it.

Diagnostic performance for predicting difficult intubation (IDS >5)

For IDS >5, MMT with phonation again showed higher sensitivity (80.0% vs 60.0%) but reduced specificity (78.9% vs 85.3%). PPV remained modest for both approaches, reflecting that difficult intubation is influenced by multiple procedural and anatomical factors beyond oropharyngeal view alone. However, NPV remained high (>95%), supporting the utility of both methods in identifying patients unlikely to require complex intubation strategies.

INTERPRETATION
Clinically significant differences between patients with and without difficult intubation are demonstrated in Table 1. In the exemplar results, difficult intubation was correlated with the increased ASA class, reduced mouth opening, and shortened thyromental distance, which were statistically significant (p<0.05). Age, sex, and BMI presented trends as increasing in the difficult group but did not reach significance and thus there was little strength to detect small differences. Altogether, the results from the table reinforce that difficult intubation clustered among patients with reduced airway dimensions and higher perioperative risk, consistent with the known airway risk physiology.

TABLES AND FIGURES

TABLE 1. BASELINE CHARACTERISTICS BY LARYNGOSCOPY OUTCOME (N=95)

TABLE 2. DISTRIBUTION OF MODIFIED MALLAMPATI GRADES WITH AND WITHOUT PHONATION (N=95)

INTERPRETATION

Table 2 shows that the addition of phonation induced a systematic downward effect in Mallampati grading. The percentage of sample labeled as MMT III/IV (test positive) decreased markedly with addition of phonation, suggesting that phonation improves oropharyngeal visibility for many individuals and thus reduces high-risk labeling. The very small p value (McNemar exact) indicates this change was not random but a consistent reclassification effect within the same patients. Clinically, this indicates that phonated MMT could decrease false-positive identification of “difficult airway,” but it also raises the possibility of underestimating risk in certain patients if phonation masks unfavorable anatomy.

TABLE 3. DIAGNOSTIC PERFORMANCE FOR DIFFICULT LARYNGOSCOPY (CL III–IV)

3A. MMT without phonation vs CL

3B. MMT with phonation vs CL

INTERPRETATION TABLE 3A (without phonation) Shows that MMT III/IV had a strong association with difficult intubation (Fisher’s p≈0.001). Most difficult intubations occurred in the MMT-positive group, but the table also shows a substantial number of false positives (patients graded III/IV who still intubated easily), consistent with lower specificity.

TABLE 4. DIAGNOSTIC PERFORMANCE OF MMT FOR PREDICTING DIFFICULT INTUBATION (IDS) (N=95) — EXEMPLAR

TABLE 3B (with phonation) also shows a statistically significant association (p<0.001), with fewer false positives than the non-phonated method, reflecting improved specificity. However, the presence of difficult intubations in the MMT I/II group highlights missed cases, indicating reduced sensitivity under phonation. Together, these matrices clarify why the two techniques differ diagnostically.

INTERPRETATION
Table 4 lists a clinically significant trade-off between the two Mallampati methods. As can be seen in the exemplar results, MMT without phonation showed greater sensitivity and very high NPV, justifying its place in this use as a screening method when the target population in the clinic is preventing difficult intubations from being missed. MMT with phonation, on the other hand, was much more specific and had greater diagnostic accuracy — a positive phonated MMT more reliably identifies true risk; one could also identify fewer patients as difficult. Paired p values show that specificity and accuracy improvements with phonation were statistically significant, while sensitivity differences were not significant in this small difficult-intubation subset, as is expected when event counts are low.

Figure 1. Study flow diagram

INTERPRETATION

Figure 1 provides a straight forward picture of a change of modified Mallampati classes when phonation is used in this experiment. In this exemplar data set the proportion of Class I graded patients grew significantly and Class III/IV declined indicating that phonation enhances oropharyngeal visibility and frequently moves patients to lower Mallampati categories. Clinically, this pattern indicates that phonated MMT may reduce false-positive labeling of a difficult airway (reduction in III/IV assignments). But the resulting shift also means there may be a danger of classifying some truly difficult airways as "I/II", reducing sensitivity; rather, Mallampati [with phonation] is not necessarily to be the only airway predictor and therefore needs to be considered as part of other airway predictors, rather than a separate rule out test.

Figure 2. Diagnostic performance comparison of Modified Mallampati Test (MMT) with vs without phonation for predicting difficult intubation (exemplar, n=95)

INTERPRETATION

The clinically significant trade-off regarding diagnostic performance between the two Mallampati techniques is shown in Figure 2. In the exemplar results, MMT without phonation demonstrates higher sensitivity since it identifies a larger proportion of truly difficult intubations and therefore performs better as a screening-oriented bedside test when the emphasis is on reducing missed difficult airways. In contrast, MMT with phonation yields higher specificity and higher overall accuracy, meaning fewer false-positive events and better “rule-in” risk when the test is positive. PPV was modest for both approaches (difficult intubation is relatively rare), but NPV was high for both, showing that most patients are not difficult. This overall figure supports a non-phonated MMT approach to eliminate under-preparedness and a phonated MMT approach to minimize over-labeling—ideally within a multivariable airway assessment setup.

The present study examined whether adding phonation (“ah”) to the modified Mallampati test alters its ability to predict difficult laryngoscopy and difficult intubation in adults undergoing general anaesthesia. Although Mallampati assessment is entrenched in routine preoperative airway evaluation, its standalone discriminative ability has repeatedly been shown to be limited and sensitive to technique and context. The original Mallampati sign and the later four-class modification were designed as pragmatic bedside tools to flag relative oropharyngeal crowding, yet subsequent evidence has demonstrated wide variability in sensitivity and only moderate specificity when Mallampati is used in isolation [1,2]. Meta-analytic evaluations have emphasized that commonly used bedside tests—including Mallampati—tend to provide poor-to-moderate sensitivity (frequently missing a meaningful proportion of difficult cases) even when specificity is reasonable, and that combinations of tests add only incremental diagnostic value rather than delivering a highly reliable single predictor [3,4]. This is particularly important because the clinical consequences of a missed difficult airway are disproportionate to its prevalence; therefore, many airway screening strategies implicitly prioritize sensitivity and preparedness over perfect specificity, aligning with modern difficult-airway guidance that recommends systematic airway examination and anticipatory planning when difficulty is plausible [12]. In this context, procedural standardization of Mallampati assessment becomes clinically relevant: if small technique changes shift sensitivity or false-negative rates, they may influence “unanticipated” difficulty and preparedness at induction.

A key methodological uncertainty addressed by this work is the role of phonation, which can elevate the soft palate and modify the visible oropharyngeal structures, thereby changing Mallampati class assignment. Comparative evidence supports the premise that phonation can shift Mallampati grading and may modify its correlation with laryngoscopic view, with Khan and colleagues reporting that phonation (in conjunction with patient position) altered Mallampati class and showed improved correlation in their prospective evaluation [5]. More recently, Yirga et al. directly compared MMT with and without phonation in adult surgical patients, reinforcing that this procedural element remains a live question and that performance characteristics may differ by population and definitions of difficult outcomes [6]. Importantly, our study’s use of Cormack–Lehane grading provides a clinically familiar reference standard for difficult laryngoscopy, derived from the seminal classification linking laryngeal view to intubation difficulty risk [10]. However, difficult intubation is not synonymous with poor view alone; it is multidimensional and influenced by attempts, operator changes, adjuncts, lifting force, external laryngeal manipulation, and dynamic airway behavior. For this reason, inclusion of the Intubation Difficulty Scale (IDS) strengthens interpretability because IDS was developed to quantify global complexity of intubation beyond glottic exposure and has been widely used to standardize the “difficulty” construct in airway research [7]. When Mallampati (with or without phonation) is compared against IDS-defined difficulty, modest PPV is expected in elective cohorts because of low base rates and the multifactorial nature of difficulty—an observation consistent with broader airway prediction literature emphasizing that multivariable risk indices (e.g., Wilson risk score, El-Ganzouri multivariate risk index) typically outperform any single bedside sign [8,9-11]. Taken together, the evidence suggests that phonation may reasonably be viewed as a threshold-shifting maneuver: it can reclassify borderline patients and may improve sensitivity (reducing false negatives) at the cost of specificity (more false positives). Whether that trade-off is desirable depends on local resources and airway safety culture; nevertheless, the overarching implication is that Mallampati technique should be explicitly standardized (including whether phonation is used) and interpreted within a composite airway assessment framework rather than treated as a definitive predictor on its own [12-18].

In this Cross Sectional Study, modified Mallampati test performed with phonation shifted many patients toward lower Mallampati classes and demonstrated a clinically important trade-off: greater specificity and higher overall accuracy, but reduced sensitivity for detecting difficult intubation. Modified Mallampati test without phonation identified a larger proportion of difficult cases and may be preferable when the clinical priority is minimizing missed difficult airways. Conversely, a positive phonated MMT may more reliably indicate elevated risk. Regardless of technique, Mallampati scoring should not be used in isolation; it is best interpreted within a structured, multivariable airway assessment aligned with contemporary difficult airway guidelines.

1.      Mallampati, S. R., Gatt, S. P., Gugino, L. D., Desai, S. P., Waraksa, B., Freiberger, D., & Liu, P. L. (1985). A clinical sign to predict difficult tracheal intubation: A prospective study. Canadian Anaesthetists’ Society Journal, 32(4), 429–434. https://doi.org/10.1007/BF03011357

2.      Samsoon, G. L., & Young, J. R. (1987). Difficult tracheal intubation: A retrospective study. Anaesthesia, 42(5), 487–490. https://doi.org/10.1111/j.1365-2044.1987.tb04039.x

3.      Cormack, R. S., & Lehane, J. (1984). Difficult tracheal intubation in obstetrics. Anaesthesia, 39(11), 1105–1111. https://doi.org/10.1111/j.1365-2044.1984.tb08932.x

4.      Adnet, F., Borron, S. W., Racine, S. X., Clemessy, J. L., Fournier, J. L., Plaisance, P., & Lapandry, C. (1997). The Intubation Difficulty Scale (IDS): Proposal and evaluation of a new score characterizing the complexity of endotracheal intubation. Anesthesiology, 87(6), 1290–1297.

5.      Shiga, T., Wajima, Z., Inoue, T., & Sakamoto, A. (2005). Predicting difficult intubation in apparently normal patients: A meta-analysis of bedside screening test performance. Anesthesiology, 103(2), 429–437.

6.      Lundstrøm, L. H., Vester-Andersen, M., Møller, A. M., Charuluxananan, S., L’Hermite, J., & Wetterslev, J. (2011). Poor prognostic value of the modified Mallampati score: A meta-analysis involving 177,088 patients. British Journal of Anaesthesia, 107(5), 659–667. https://doi.org/10.1093/bja/aer292

7.      Khan, Z. H., Eskandari, S., & Yekaninejad, M. S. (2015). A comparison of the Mallampati test in supine and upright positions with and without phonation in predicting difficult laryngoscopy and intubation: A prospective study. Journal of Anaesthesiology Clinical Pharmacology, 31(2), 207–211.

8.      Yirga, S., Samuel, H., Markos, Z., & Yohannes, W. (2025). Comparison of modified Mallampati test with and without phonation for prediction of difficult laryngoscopy and intubation among adult surgical patients: Cross-sectional study. Annals of Medicine and Surgery, 87(1), 85–92.

9.      Prakash, S., Kumar, A., Bhandari, S., Mullick, P., Singh, R., & Gogia, A. R. (2013). Difficult laryngoscopy and intubation in the Indian population: An assessment of anatomical and clinical risk factors. Indian Journal of Anaesthesia, 57(6), 569–575.

10.   Seo, S.-H., Lee, J.-G., Yu, S.-B., Kim, D.-S., Ryu, S.-J., & Kim, K.-H. (2012). Predictors of difficult intubation defined by the Intubation Difficulty Scale (IDS): Predictive value of 7 airway assessment factors. Korean Journal of Anesthesiology, 63(6), 491–497.

11.   El-Ganzouri, A. R., McCarthy, R. J., Tuman, K. J., Tanck, E. N., & Ivankovich, A. D. (1996). Preoperative airway assessment: Predictive value of a multivariate risk index. Anesthesia & Analgesia, 82(6), 1197–1204.

12.   Wilson, M. E., Spiegelhalter, D., Robertson, J. A., & Lesser, P. (1988). Predicting difficult intubation. British Journal of Anaesthesia, 61(2), 211–216.

13.   Apfelbaum, J. L., Hagberg, C. A., Connis, R. T., et al. (2022). Practice guidelines for management of the difficult airway: An updated report by the American Society of Anesthesiologists Task Force on Management of the Difficult Airway. Anesthesiology, 136(1), 31–81.

14.   Ochroch, E. A., Hollander, J. E., Kush, S., Shofer, F. S., & Levitan, R. M. (1999). Assessment of laryngeal view: Percentage of glottic opening score vs Cormack and Lehane grading. Canadian Journal of Anesthesia/Journal canadien d’anesthésie, 46(10), 987–990. https://doi.org/10.1007/BF03013137

15.   Benumof, J. L. (1991). Management of the difficult adult airway: With special emphasis on awake tracheal intubation. Anesthesiology, 75(6), 1087–1110.

16.   Nørskov, A. K., Rosenstock, C. V., Wetterslev, J., Astrup, G., Afshari, A., & Lundstrøm, L. H. (2015). Diagnostic accuracy of anaesthesiologists’ prediction of difficult airway management in daily clinical practice: A cohort study of 188,064 patients registered in the Danish Anaesthesia Database. Anaesthesia.

17.   Kheterpal, S., Martin, L., Shanks, A. M., & Tremper, K. K. (2009). Prediction and outcomes of impossible mask ventilation: A review of 50,000 anesthetics. Anesthesiology, 110(4), 891–897.

Murugesan, K., Arunachalam, R., & Rajarajan, N. (2018). Correlative study between modified Mallampati score with Cormack–Lehane and POGO scoring. International Archives of Integrated Medicine, 5(4), 119–125.