Introduction
Securing a patent airway is one of the most critical skills required of any anaesthesiologist. Failure to do so in a timely and safe manner can rapidly lead to life-threatening complications including hypoxia, aspiration, hypoxic brain injury, and death [1]. Despite continuous technological advancement—including video laryngoscopes, flexible bronchoscopes, and supraglottic airway devices—difficult laryngoscopy and failed tracheal intubation remain significant challenges in contemporary anaesthetic practice [2][3].
Epidemiological data from the American Society of Anesthesiologists (ASA) Closed Claims Project indicate that airway-related incidents continue to account for approximately 17% of all anaesthesia-related respiratory injuries, with difficult tracheal intubation identified as a leading cause of anaesthesia-related mortality [4]. An estimated 34% of anaesthesia-related deaths are directly attributable to ineffective management of a difficult airway [5]. The reported incidence of difficult intubation in the general surgical population ranges from 1% to 3.6%, and failed intubation occurs in approximately 0.05%–0.3% of cases [6]. While seemingly low, these figures carry devastating consequences when unanticipated or inadequately managed.
Diabetes mellitus (DM) represents a particularly important risk factor for difficult airway management. It is the world's most prevalent endocrine disorder, and approximately half of all diabetic individuals will require surgical intervention at some point in their lifetime [7]. Chronic hyperglycaemia facilitates non-enzymatic glycosylation of collagen and elastin fibres, leading to cross-linking and fibrosis of peri-articular connective tissues. This progressive process—clinically recognised as Limited Joint Mobility Syndrome (LJMS) or diabetic cheiroarthropathy, first described by Rosenbloom and Frias in 1982—restricts the mobility of key laryngoscopic joints including the temporomandibular joint (TMJ) and atlanto-occipital (AO) articulation [8-10].
Several landmark studies have confirmed this predisposition. Hogan et al. (1988) reported a difficult laryngoscopy rate of 32% among diabetic patients undergoing renal transplantation, compared to only 2.7% in non-diabetic controls, attributing this disparity to glycosylation-induced stiff-joint changes [11]. Reissell et al. (1990) and McLennan et al. (1986) subsequently demonstrated that the same pathological alterations affecting the small joints of the hands also involve the atlanto-occipital and temporomandibular articulations, directly compromising head extension and mouth opening during laryngoscopy [8][10].
Over the years, various bedside airway assessment tools have been developed to preoperatively identify patients at risk for difficult laryngoscopy. These include the Modified Mallampati Classification (MMC), thyromental distance (TMD), Upper Lip Bite Test (ULBT), sternomental distance (SMD), and neck extension assessment [12,13]. Diabetes-specific tests such as the Prayer Sign and Palm Print Sign have also been established as surrogate markers of LJMS, and several studies have demonstrated their superior predictive value in diabetic populations [14,15].
Despite this body of evidence, a fundamental challenge persists: no single airway assessment tool demonstrates sufficient sensitivity and specificity to reliably predict difficult laryngoscopy across all patient populations [6]. This limitation is further amplified in diabetic individuals, where the underlying mechanism of airway difficulty differs substantially from that of the general population. While oropharyngeal crowding predominantly drives difficulty in non-diabetic patients, soft-tissue stiffness due to collagen glycosylation represents the principal factor in diabetic individuals—a distinction that renders standard non-diabetes-specific indices potentially less effective in this group [8][10][11].
George and Jacob (2003), in a landmark Indian study comparing 71 diabetic and 70 non-diabetic surgical patients, observed that the Palm Print Sign was the most sensitive predictor of difficult intubation in diabetics, whereas the MMC exhibited superior predictive value among non-diabetics [11]. Subsequent research by has further refined the comparative diagnostic performance of these indices in both populations [16,17].
Given the rising prevalence of diabetes mellitus in India and its recognised impact on airway management, the present study was undertaken to prospectively compare the predictive value of commonly used airway assessment indices in diabetic versus non-diabetic patients undergoing elective surgery under general anaesthesia with tracheal intubation. The findings are intended to guide anaesthesiologists in selecting the most appropriate and reliable combination of assessment tools for preoperative airway evaluation in this high-risk population.
Methods
This was a prospective, randomised comparative study conducted at the Department of Anaesthesiology, SRN Hospital, Motilal Nehru Medical College, Prayagraj, Uttar Pradesh, India, over a period of one year. Ethical approval was obtained from the Institutional Ethics Committee prior to commencement of the study, and written informed consent was obtained from all participants.
Consecutive sampling was employed to enrol eligible patients meeting the inclusion criteria. Patients aged 18–65 years of either sex, classified as ASA Physical Status I–III, and scheduled for elective surgery requiring general anaesthesia with endotracheal intubation were included. Patients were categorised into two groups: Group D (Diabetic, n = 46) and Group ND (Non-Diabetic, n = 46). Patients with known airway abnormalities or deformities, history of head and neck surgery or trauma, mechanical restriction of mouth opening, emergency surgical status, or pregnancy and lactation were excluded.
Sample size was calculated based on the expected incidence of difficult intubation in diabetic patients (approximately 25%) versus non-diabetic patients (approximately 5%), as reported in prior literature [10][15]. Using Cochran's formula for comparison of two proportions, with a significance level of α = 0.05 (95% confidence interval) and power of 80% (β = 0.20), a minimum of 46 patients per group was required, yielding a total sample size of 92. This calculation accounted for an expected effect size of approximately 20%.
All preoperative airway assessments were performed by a single trained investigator blinded to the planned laryngoscopy. The following indices were recorded:
-
Modified Mallampati Classification (MMC): Assessed in the sitting position with full mouth opening and tongue protrusion without phonation. Classified as Grades I–IV.
-
Thyromental Distance (TMD): Measured in centimetres from the mentum to the thyroid notch with full head extension and closed mouth.
-
Upper Lip Bite Test (ULBT): Assessed ability of lower incisors to bite the upper lip, classified as Class I, II, or III.
-
Prayer Sign (PS): Assessed ability to fully appose palmar surfaces in the 'namaste' position. Recorded as normal or abnormal.
-
Palm Print Sign (PPS): Evaluated by ink impression of the dominant hand's palm, graded 0–3 based on inter-phalangeal area contact.
All patients received standardised general anaesthesia. Following pre-oxygenation, induction was achieved with intravenous fentanyl, propofol, and vecuronium. Direct laryngoscopy was performed using a standard Macintosh blade by a senior anaesthesiologist blinded to the preoperative assessment findings. The Cormack–Lehane (CL) grading was recorded as the primary outcome indicator of laryngoscopic difficulty, with CL Grade III–IV defined as difficult laryngoscopy. Additional data collected included number of intubation attempts, use of adjuncts (bougie, stylet, external laryngeal manipulation), and occurrence of desaturation (SpO₂ < 90%).
Statistical analysis
Statistical analysis was performed using SPSS version 24. Categorical variables were compared using chi-square tests. Continuous variables were compared using Student's independent t-test. Sensitivity, specificity, PPV, and NPV were calculated for each airway index against the CL grading standard. A p-value of < 0.05 was considered statistically significant.
Results
A total of 92 patients were enrolled and equally distributed between the diabetic group (n = 46) and non-diabetic group (n = 46). The results of all assessments are detailed in Tables 1 through 6 below.
| Parameter | Diabetic (n=46) | Non-Diabetic (n=46) | t / χ² | p-Value |
| Age (years) — Mean ± SD | 53.83 ± 8.35 | 46.11 ± 7.07 | 4.78 | < 0.001* |
| Male Gender — n (%) | 28 (60.87%) | 25 (54.35%) | 0.40 | 0.527 |
| BMI (kg/m²) — Mean ± SD | 26.10 ± 2.06 | 23.28 ± 1.51 | 7.41 | < 0.001* |
| ASA Grade III — n (%) | 15 (32.61%) | 8 (17.39%) | 7.33 | 0.026* |
*Statistically significant (p < 0.05). Diabetic patients were significantly older, had higher BMI, and a higher prevalence of ASA Grade III status, reflecting the greater systemic disease burden associated with diabetes mellitus. Gender distribution between the two groups was comparable.
| MMC Grade | Diabetic n | Diabetic % | Non-Diabetic n | Non-Diabetic % | p-Value |
| Grade I | 3 | 6.52% | 8 | 17.39% | 0.032* |
| Grade II | 15 | 32.61% | 23 | 50.00% | |
| Grade III | 23 | 50.00% | 14 | 30.43% | |
| Grade IV | 5 | 10.87% | 1 | 2.17% | |
| Grades III–IV (total) | 28 | 60.87% | 15 | 32.60% | 0.032* |
Higher Mallampati grades (III–IV), indicative of potential airway difficulty, were significantly more prevalent among diabetic patients. Grades III–IV were recorded in 60.87% of diabetic patients versus 32.60% of non-diabetic patients (χ² = 8.81, p = 0.032), suggesting that diabetes-related connective tissue changes adversely affect oropharyngeal visualisation.
| Parameter | Category | Diabetic n | Diabetic % | ND n | ND % | p-Value |
| Prayer Sign | Abnormal | 32 | 69.57% | 19 | 41.30% | 0.012* |
| Normal | 14 | 30.43% | 27 | 58.70% | ||
| Palm Print Sign | Positive | 32 | 69.57% | 21 | 45.65% | 0.035* |
| Negative | 14 | 30.43% | 25 | 54.35% |
ND = Non-Diabetic. Both diabetes-specific indices were significantly more prevalent in the diabetic group. A positive Prayer Sign was found in 69.57% of diabetics versus 41.30% of non-diabetics (p = 0.012), and a positive Palm Print Sign in 69.57% of diabetics versus 45.65% of non-diabetics (p = 0.035), reflecting the systemic connective tissue involvement of long-standing diabetes.
| Parameter | Category | Diabetic n | Diabetic % | ND n | ND % | p-Value |
| CL Grade | Grade I | 9 | 19.57% | 20 | 43.48% | 0.032* |
| Grade II | 13 | 28.26% | 13 | 28.26% | ||
| Grade III | 14 | 30.43% | 9 | 19.57% | ||
| Grade IV | 10 | 21.74% | 4 | 8.70% | ||
| Difficult Laryngoscopy (III–IV) | Combined | 24 | 52.17% | 13 | 28.26% | 0.032* |
| Intubation Attempts | 1 attempt | 18 (39.13%) | 33 (71.70%) | 0.007* | ||
| ≥2 attempts | 28 (60.87%) | 13 (28.26%) | ||||
| Adjunct Use | Yes | 32 (69.57%) | 16 (34.78%) | 0.002* |
CL = Cormack–Lehane; ND = Non-Diabetic. Difficult laryngoscopy (CL Grade III–IV) was recorded in 52.17% of diabetic patients and 28.26% of non-diabetics. Diabetic patients also required significantly more intubation attempts and more frequent use of airway adjuncts, consistent with the higher incidence of difficult laryngoscopy in this population.
| Airway Test | Sensitivity (%) | Specificity (%) | PPV (%) | NPV (%) |
| Modified Mallampati Classification | 95.65 | 92.59 | 90.00 | 94.34 |
| Prayer Sign | 89.47 | 44.44 | 53.13 | 85.71 |
| Palm Print Sign | 68.42 | 92.59 | 83.87 | 82.35 |
| Upper Lip Bite Test | 21.10 | 87.00 | 66.00 | 54.10 |
PPV = Positive Predictive Value; NPV = Negative Predictive Value. The Modified Mallampati Classification demonstrated the highest sensitivity (95.65%) and specificity (92.59%) in the diabetic group, establishing it as the most reliable overall screening tool. The Palm Print Sign offered excellent specificity (92.59%) with moderate sensitivity, making it valuable as a confirmatory test. The Prayer Sign provided high sensitivity but low specificity, functioning best as a screening tool. The Upper Lip Bite Test showed limited sensitivity in this group.
| Airway Test | Sensitivity (%) | Specificity (%) | PPV (%) | NPV (%) |
| Modified Mallampati Classification | 90.48 | 80.00 | 82.61 | 89.47 |
| Prayer Sign | 78.95 | 66.67 | — | 74.07 |
| Palm Print Sign | 44.44 | 95.24 | — | 80.00 |
| Upper Lip Bite Test | 21.05 | 92.59 | — | 62.50 |
PPV = Positive Predictive Value; NPV = Negative Predictive Value; — = Data not separately reported. In non-diabetic patients, the Modified Mallampati Classification remained the best overall predictor. The Palm Print Sign showed the highest specificity (95.24%) in this group, reinforcing its role as a highly reliable confirmatory test when positive. The Upper Lip Bite Test again showed very low sensitivity, limiting its standalone utility.
Discussion
The findings of this study confirm that diabetic patients face a substantially higher risk of difficult laryngoscopy during general anaesthesia compared to their non-diabetic counterparts. The observed incidence of difficult laryngoscopy (Cormack–Lehane Grade III–IV) was 52.17% in the diabetic group versus 28.26% in the non-diabetic group, a difference that was statistically significant (p = 0.032). These results align with the established understanding that diabetes mellitus fundamentally alters airway-related anatomy and tissue compliance through the mechanism of collagen glycosylation [8-11].
The significant difference in mean age between the two groups (53.83 ± 8.35 years in diabetics vs 46.11 ± 7.07 years in non-diabetics, p < 0.001) was expected, given that type 2 diabetes mellitus predominantly affects middle-aged to elderly individuals. Older diabetic patients typically have a longer duration of disease and correspondingly greater cumulative collagen glycosylation, resulting in more pronounced limited joint mobility. A similar age disparity was documented by Rao et al. (2019) (52.50 ± 8.518 vs 47.2 ± 8.925 years, p = 0.008), whereas Kute et al. (2019) reported a non-significant age difference, possibly reflecting the older overall age profile of their sample [16]. Gondane et al. (2017) similarly found a trend toward older age in diabetic patients within their surgical cohort [18].
The higher BMI observed in the diabetic group (26.10 ± 2.06 vs 23.28 ± 1.51 kg/m², p < 0.001) is consistent with the well-established pathophysiological relationship between central adiposity and insulin resistance in type 2 diabetes. From an airway management perspective, elevated BMI independently contributes to difficult laryngoscopy through increased soft tissue deposition in the neck and reduced cervical mobility. Mahalakshmi et al. (2023) specifically identified BMI as a significant correlate of intubation difficulty, while Corrente et al. (2020) highlighted super-morbid obesity as a significant independent risk factor for difficult laryngoscopy [19,20]. Kute et al. (2019) similarly noted that BMI exceeding 30 kg/m² was significantly more common in their diabetic group (p = 0.042) [16].
The significantly higher prevalence of ASA Grade III status in our diabetic patients (32.61% vs 17.39%, p = 0.026) reflects the greater systemic disease burden associated with long-standing diabetes mellitus, including cardiovascular, renal, and neurological complications. The clinical importance of this finding lies in the compounding effect of multi-system comorbidities on airway management complexity. Srivastava et al. (2024) noted significant associations between ASA grade and intubation difficulty [21], and Tiwari et al. (2024) acknowledged that higher systemic disease burden contributes to the challenges of airway assessment [22].
Regarding the Modified Mallampati Classification, our finding that higher MMC grades were significantly more prevalent among diabetic patients (p = 0.032) is supported by an extensive body of literature. Kute et al. (2019) reported MMC class III–IV in 63.6% of diabetic versus 22.7% of non-diabetic patients (p = 0.001), with the MMC demonstrating the highest sensitivity (78.7%) among all tests in the diabetic group [16]. Gondane et al. (2017) found a sensitivity of 80.77% for the MMC in predicting difficult laryngoscopy in diabetic patients [18]. Rao et al. (2019) documented a high MMC specificity of 96.3% in their diabetic cohort, while Vakilian et al. (2023) reported 100% sensitivity in type 2 diabetic patients [9]. Wang et al. (2024), in a systematic review and meta-analysis, confirmed the MMC as a widely applicable and clinically relevant tool, while acknowledging that a combined approach surpasses any single predictor [23]. In our study, the MMC demonstrated the highest sensitivity (95.65%) and specificity (92.59%) in diabetic patients, establishing it as the most reliable overall screening test in this population.
The Upper Lip Bite Test showed statistically significant differences between groups in terms of grade distribution (p = 0.0025), with higher ULBT grades significantly more prevalent in diabetic patients. This is consistent with the pathophysiology of glycosylation-induced temporomandibular joint restriction, which directly impairs mandibular protrusion. Nirmala et al. (2025) reported that ULBT Class 3 was found exclusively in diabetic individuals and was absent in non-diabetic patients (p = 0.000) [24]. However, Kute et al. (2019) found no statistically significant group difference in ULBT grade (p = 0.11) [16]. In terms of diagnostic performance, the ULBT demonstrated low sensitivity (21.10% in diabetics, 21.05% in non-diabetics) but reasonable specificity in both groups—a pattern consistent with findings by Zachariah et al. (2023) (sensitivity 27.41%, specificity 96.29%) [25] and Wang et al. (2024), who reported a pooled sensitivity of 0.52 and specificity of 0.84 in their meta-analysis [23]. Nishana et al. (2023) found no significant agreement between ULBT and CL grading [26]. These data collectively suggest that the ULBT, while not sufficiently sensitive as a standalone predictor, contributes meaningfully to a multiparameter airway assessment framework.
Thyromental distance did not differ significantly between groups in our study (p = 0.239), which is consistent with findings from Kute et al. (2019) (p = 0.825) and Rao et al. (2019) (p = 1.000) [16]. While Hashim et al. (2014) documented poor diagnostic performance for TMD alone (sensitivity 15.4%, specificity 68.1%) [27], Gupta et al. (2025) also identified the RHTMD as the most significant predictor of difficult laryngoscopy in their paediatric cohort [28]. Sinha et al. (2022) reported the maximum area under the ROC curve for a combination of MMC and TMD [29], reinforcing the principle that TMD, while of limited standalone value in diabetic versus non-diabetic comparisons, retains clinical relevance when incorporated into composite assessment strategies.
The Prayer Sign was significantly more prevalent among diabetic patients in our study (69.57% abnormal vs 41.30% in non-diabetics, p = 0.012). The Prayer Sign reflects systemic peri-articular fibrosis due to collagen glycosylation—the same mechanism that restricts the temporomandibular and atlanto-occipital joints during laryngoscopy. Kute et al. (2019) similarly reported a significant difference (p = 0.001) with the Prayer Sign demonstrating sensitivity 72.3% and specificity 84.1% in their diabetic cohort [18]. Erden et al. (2003) validated the Prayer Sign as a diagnostic marker of systemic connective tissue stiffening in diabetic patients [45]. However, Baig and Khan (2014) found that the Prayer Sign had significantly lower sensitivity (29.6%) than the MMC (79.3%) and concluded that it is not suitable as a standalone predictor [30]. Hashim et al. (2014) also found limited significance for the Prayer Sign in their smaller cohort [27], and Nirmala et al. (2025) reported no significant difference between groups in their study [24]. In our study, the Prayer Sign showed high sensitivity (89.47%) but low specificity (44.44%), confirming its utility as a screening tool rather than a confirmatory test, and emphasising the need for its use alongside other indices.
The Palm Print Sign was significantly more prevalent in the diabetic group (69.57% positive vs 45.65% non-diabetic, p = 0.035). This finding has strong mechanistic backing—the collagen glycosylation that reduces inter-phalangeal apposition in the hand simultaneously restricts airway-critical joints. Hashim et al. (2014) identified the Palm Print Sign as the most reliable predictor of difficult laryngoscopy in diabetic patients, with sensitivity of 76.9% and specificity of 89.4% [27]. Kute et al. (2019) reported the highest specificity (93.7%) for the Palm Print Sign in their diabetic group [16], and Gondane et al. (2017) found specificity of 97.06%, with the highest positive predictive value of all indices tested [18]. George and Jacob (2003) reported the Palm Print Sign as the most sensitive predictor in their diabetic cohort [11]. Mandal et al. (2025) recently confirmed sensitivity of 78.9% and specificity of 85.2%, identifying it as an independent predictor of difficult airway with an odds ratio of 4.91 [31]. Vakilian et al. (2023) reported 100% sensitivity for the Palm Print Sign in type 2 diabetics [32], while Nirmala et al. (2025) confirmed that higher Palm Print grades were significantly more frequent in diabetic patients, with a highly significant correlation with the Prayer Sign [24]. In our study, the Palm Print Sign showed excellent specificity (92.59%) in diabetic patients, validating its role as the most effective confirmatory test in this population.
The higher number of intubation attempts and significantly greater adjunct utilisation in diabetic patients (69.57% vs 34.78%, p = 0.002) directly mirror the elevated incidence of difficult laryngoscopy in this group. These findings reinforce the clinical necessity of advance airway planning, availability of adjuncts such as bougies and video laryngoscopes, and presence of experienced personnel when managing diabetic patients under general anaesthesia. Bergesio et al. (2016) highlighted the critical role of adjunct devices in managing unanticipated difficult intubation [33], and Tiwari et al. (2024) demonstrated a difficult intubation prevalence of 17.1%, underscoring the importance of robust preoperative prediction tools [21]. Shiga et al. (2005) and Wang et al. (2024) have both concluded that no single test has ideal diagnostic performance, and that combining multiple assessment methods significantly improves predictive accuracy [6,23]—a principle strongly endorsed by our findings.
Limitations
The study has several limitations that should be acknowledged. The sample size, though adequately powered, was relatively small (n = 92), and single-centre recruitment may limit the generalisability of findings to broader populations. Observer bias cannot be entirely excluded, as airway assessment and laryngoscopy grading were performed by clinical personnel, despite blinding efforts. Duration and severity of diabetes were not systematically recorded, and these variables may have modified the degree of airway-related tissue changes. Larger, multicentre studies with comprehensive characterisation of diabetic duration and complications are warranted to validate and extend these findings.
Conclusion
Diabetes mellitus is an important independent risk factor for difficult laryngoscopy during general anaesthesia, with diabetic patients demonstrating significantly more adverse airway predictors and higher rates of difficult intubation. Modified Mallampati Classification showed the best overall predictive performance, while Palm Print Sign provided high specificity and Prayer Sign offered useful sensitivity. Since no single test was adequate alone, a multiparameter airway assessment incorporating Mallampati, Palm Print Sign, Prayer Sign, and Upper Lip Bite Test is recommended to facilitate early identification, preparedness, and safer airway management.
Declarations
Patient Consent
Written informed consent was obtained from the patient for the publication of this case report and any accompanying clinical images.
Conflict of Interest
The authors declare no conflict of interest.
Funding
This case report received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.
Ethical Approval
Ethical approval was obtained from the Institutional Ethics Committee, MLN Medical College, Prayagraj. Written informed consent was obtained from all study participants.