Abstract
Objective: To evaluate the predictive value of the Braden scale and laboratory indicators for oral mucosal pressure injury (OMPI) in patients admitted to the Emergency Intensive Care Unit (ICU). Methods: A retrospective analysis was conducted on 238 intubated patients admitted to the First Affiliated Hospital of Anhui Medical University. Patients were divided into a training set (n = 166) and a validation set (n = 72). The training set was further classified into OMPI (n = 67) and non-OMPI (n = 99) groups. Clinical data were compared between the two groups, and a logistic regression model was constructed to develop a predictive nomogram. Model performance was assessed using discrimination and calibration metrics, and internal validation was performed with the validation cohort. Results: The training and validation sets were comparable. Significant predictors of OMPI included Braden scale score (P < 0.001), ICU length of stay (P < 0.001), intubation duration (P = 0.039), and hemoglobin (P < 0.001). Logistic regression identified Braden scale, intubation duration, hemoglobin, and hematocrit as independent risk factors. Conclusion: The combination of Braden scale score, hemoglobin, and hematocrit demonstrated good predictive value for OMPI in EICU patients.
Keywords: Tracheal intubation, oral mucosal pressure injury, Braden scale, hemoglobin, hematocrit
Introduction
With advances in intensive care and airway management, mechanical ventilation has become a standard respiratory support technique for critically ill patients in intensive care units (ICUs) [1]. Transoral intubation is the most commonly used method, accounting for over 96% of all tracheal intubations due to its ease of learning and operational simplicity [2-4].
Mucosal pressure injury (MPI) is a type of device-related pressure injury (DRPI) characterized by localized mucosal tissue damage caused by sustained pressure or shear forces from medical devices such as tracheal tubes, nasogastric tubes, and oxygen catheters [5]. Among these, oral MPI (OMPI) due to transoral intubation is the most prevalent [6,7].
Patients undergoing transoral intubation are particularly vulnerable to OMPI due to impaired swallowing and coughing reflexes, disruption of the oral and respiratory tract defenses, altered oral microbiota, and mechanical irritation from tracheal tubes, bite blocks, fixation devices, and suctioning procedures [8,9]. The reported incidence of OMPI ranges from 0.16% to 55.6%, typically occurring within 2 to 13 days post-intubation [7,10]. Critically ill ICU patients, due to their compromised condition, are prone to recurrent and slow-healing OMPIs, which may lead to infection, adhesion, or ulceration. These complications exacerbate patient discomfort, cause dysphagia, increase medical costs, prolong hospitalization, and may strain the physician-patient relationship [11]. Therefore, early identification and prevention of OMPI are essential to improving care quality and nursing outcomes.
Identifying high-risk individuals is crucial for OMPI prevention; however, no standardized risk assessment tool currently exists for MPI or OMPI. The Braden Pressure Ulcer Risk Assessment Scale, widely used in clinical practice, was originally designed for assessing pressure injuries over bony prominences [12]. Given the structural and functional differences between skin and mucosa, its applicability to OMPI remains uncertain. This study thus aimed to investigate the predictive value of the Braden scale combined with laboratory parameters for OMPI in ICU patients, with the goal of informing the development of targeted risk prediction models and assessment tools for OMPI associated with transoral intubation.
Materials and methods
Study design
This retrospective study included patients admitted to the First Affiliated Hospital of Anhui Medical University between June 2022 and March 2024. Inclusion criteria were as follows: transoral tracheal intubation, age ≥ 18 years, ICU stay ≥ 24 hours, and the presence of an oral indwelling endotracheal tube. Exclusion criteria included a history of oral mucosal disease (which could interfere with observation), preexisting OMPI prior to ICU admission, and incomplete clinical data. A total of 238 eligible patients were enrolled and randomly assigned to a training set (n = 166) and a validation set (n = 72). In the training set, 67 patients developed OMPI, while 99 did not; in the validation set, 36 patients developed OMPI, and 36 did not. The study was approved by the Clinical Medical Research Ethics Committee of the First Affiliated Hospital of Anhui Medical University.
Data acquisition
Data were collected using the hospital’s electronic medical records and an observation checklist. Variables included demographic information (sex, age, BMI), clinical characteristics (ICU length of stay, history of diabetes, fever, use of vasoactive agents, sedatives/analgesics, antibiotics, use of continuous renal replacement therapy (CRRT) and extracorporeal membrane oxygenation (ECMO), duration of tracheal intubation, etc.). Hematological indices collected included white blood cell count, albumin, hemoglobin, hematocrit, platelet count, and arterial oxygen partial pressure. Severity scores such as the Braden Scale and acute physiology and chronic health evaluation II scale (APACHE II) score were also recorded.
The Braden Scale, assessed by nurses within 24 hours of ICU admission, evaluates pressure injury risk across six subscales. Each subscale is scored from 1 to 4 (except friction/shear, scored 1-3), with total scores ranging from 6 to 23; lower scores indicate higher risk [13].
The APACHE II score, with a theoretical maximum of 71, reflects disease severity and complexity. Higher scores indicate worse prognosis and greater clinical complexity [14].
Statistical analysis
Continuous variables were presented as mean ± standard deviation. Group comparisons were performed using independent samples t-tests or nonparametric tests based on distribution normality. Categorical variables were reported as frequencies and percentages, and analyzed using the chi-square test. Univariate and multivariate logistic regression analyses were conducted to identify potential risk factors for OMPI in ICU patients.
A nomogram model was constructed using the training set to predict OMPI risk in intubated patients. Model performance was evaluated in terms of discrimination (ROC curves), calibration, and clinical utility via decision curve analysis (DCA). Predictive power of the Braden Scale, hemoglobin, and hematocrit was also assessed. A significance level of α = 0.05 was used. Statistical analyses were performed using SPSS version 27.0, with two-sided P-values < 0.05 considered statistically significant.
Results
Comparison of general data between the training set and validation set
As shown in Table 1, the incidence of OMPI was 40.36% in the training set and 50.00% in the validation set. No statistically significant differences in baseline clinical characteristics were observed between the two sets (all P > 0.05) (Table 1).
Table 1.
Comparison of general data between training and validation sets
| Validation set (n = 72) | Training set (n = 166) | t/Z/χ2 | P value | |
|---|---|---|---|---|
| Sex, n (%) | 0.98 | 0.323 | ||
| Male | 42 (58.33) | 108 (65.06) | ||
| Female | 30 (41.67) | 58 (34.94) | ||
| Age, n (%) | 0.55 | 0.761 | ||
| < 45 | 8 (11.11) | 21 (12.65) | ||
| 45-64 | 26 (36.11) | 66 (39.76) | ||
| ≥ 65 | 38 (52.78) | 79 (47.59) | ||
| BMI, n (%) | 5.75 | 0.056 | ||
| < 18.5 | 27 (37.50) | 45 (27.11) | ||
| 18.5-24 | 42 (58.33) | 99 (59.64) | ||
| ≥ 24 | 3 (4.17) | 22 (13.25) | ||
| Fever, n (%) | 1.68 | 0.195 | ||
| No | 29 (40.28) | 82 (49.40) | ||
| Yes | 43 (59.72) | 84 (50.60) | ||
| Diabetes, n (%) | 1.69 | 0.194 | ||
| No | 57 (79.17) | 118 (71.08) | ||
| Yes | 15 (20.83) | 48 (28.92) | ||
| Use of vasoactive drugs, n (%) | 0.01 | 0.925 | ||
| No | 23 (31.94) | 52 (31.33) | ||
| Yes | 49 (68.06) | 114 (68.67) | ||
| Use of sedative and analgesic drugs, n (%) | 0.01 | 0.940 | ||
| No | 8 (11.11) | 19 (11.45) | ||
| Yes | 64 (88.89) | 147 (88.55) | ||
| Use of antibiotics, n (%) | 0.52 | 0.471 | ||
| No | 1 (1.39) | 7 (4.22) | ||
| Yes | 71 (98.61) | 159 (95.78) | ||
| Use of CRRT, n (%) | 0.03 | 0.855 | ||
| No | 58 (80.56) | 132 (79.52) | ||
| Yes | 14 (19.44) | 34 (20.48) | ||
| Use of ECMO, n (%) | 0.04 | 0.848 | ||
| No | 68 (94.44) | 154 (92.77) | ||
| Yes | 4 (5.56) | 12 (7.23) | ||
| OMPI, n (%) | 1.90 | 0.168 | ||
| No | 36 (50.00) | 99 (59.64) | ||
| Yes | 36 (50.00) | 67 (40.36) | ||
| Hematocrit, Mean ± SD | 30.82 ± 9.28 | 32.85 ± 7.74 | -1.75 | 0.081 |
| Braden scale, M (Q1, Q3) | 12.00 (11.00, 13.00) | 12.00 (12.00, 13.00) | -0.06 | 0.950 |
| Length of ICU stay (d), M (Q1, Q3) | 7.00 (5.00, 11.00) | 7.00 (4.25, 10.75) | -0.73 | 0.465 |
| APACHE II score, M (Q1, Q3) | 20.50 (17.00, 24.00) | 20.00 (17.00, 25.00) | -0.03 | 0.976 |
| Tracheal intubation duration (h), M (Q1, Q3) | 127.00 (66.50, 187.25) | 120.00 (71.25, 168.00) | -0.33 | 0.744 |
| White blood cell, M (Q1, Q3) | 11.01 (8.12, 14.28) | 10.07 (7.19, 14.25) | -1.01 | 0.311 |
| Albumin, M (Q1, Q3) | 31.85 (28.25, 35.35) | 31.30 (27.60, 34.95) | -0.64 | 0.520 |
| Hemoglobin, M (Q1, Q3) | 101.00 (87.50, 120.00) | 107.00 (90.00, 120.00) | -0.98 | 0.325 |
| Platelet, M (Q1, Q3) | 151.00 (71.75, 218.00) | 136.50 (91.25, 206.50) | -0.13 | 0.897 |
| Partial pressure of oxygen, M (Q1, Q3) | 82.50 (68.30, 106.25) | 88.00 (67.47, 115.28) | -0.59 | 0.558 |
BMI: body mass index, CRRT: continuous renal replacement therapy, ECMO: extracorporeal membrane oxygenation, OMPI: oral mucosal pressure injury.
Comparison of general data OMPI and non-OMPI patients in the training set
As shown in Table 2, among the 166 patients in the training set, 67 developed OMPI. In this group, 67.16% were male, and the age distribution was < 45 years (10.45%), 45-64 years (35.82%), and ≥ 65 years (53.73%). Corresponding figures in the non-OMPI group were 63.64% male and age distribution of < 45 (14.14%), 45-64 (42.42%), and ≥ 65 (43.43%). These differences were not statistically significant (all P > 0.05). However, patients with OMPI had significantly longer ICU stays (P < 0.001) and longer tracheal intubation duration (P = 0.039). Additionally, Braden scale scores (P < 0.001) and hemoglobin levels (P < 0.001) were significantly lower in the OMPI group. No significant differences were found between the groups in hematocrit, APACHE II score, white blood cell count, albumin, platelet count, or arterial oxygen partial pressure (all P > 0.05).
Table 2.
Comparison of general data OMPI and non-OMPI patients in the training set
| Non-OMPI group (n = 99) | OMPI group (n = 67) | t/Z/χ2 | P value | |
|---|---|---|---|---|
| Sex, n (%) | 0.22 | 0.640 | ||
| Male | 63 (63.64) | 45 (67.16) | ||
| Female | 36 (36.36) | 22 (32.84) | ||
| Age, n (%) | 1.76 | 0.415 | ||
| < 45 | 14 (14.14) | 7 (10.45) | ||
| 45-64 | 42 (42.42) | 24 (35.82) | ||
| ≥ 65 | 43 (43.43) | 36 (53.73) | ||
| BMI, n (%) | 0.45 | 0.800 | ||
| < 18.5 | 28 (28.28) | 17 (25.37) | ||
| 18.5-24 | 57 (57.58) | 42 (62.69) | ||
| ≥ 24 | 14 (14.14) | 8 (11.94) | ||
| Fever, n (%) | 0.84 | 0.358 | ||
| No | 46 (46.46) | 36 (53.73) | ||
| Yes | 53 (53.54) | 31 (46.27) | ||
| Diabetes, n (%) | 2.61 | 0.106 | ||
| No | 75 (75.76) | 43 (64.18) | ||
| Yes | 24 (24.24) | 24 (35.82) | ||
| Use of vasoactive drugs, n (%) | 1.04 | 0.308 | ||
| No | 34 (34.34) | 18 (26.87) | ||
| Yes | 65 (65.66) | 49 (73.13) | ||
| Use of sedative and analgesic drugs, n (%) | 0.44 | 0.508 | ||
| No | 10 (10.10) | 9 (13.43) | ||
| Yes | 89 (89.90) | 58 (86.57) | ||
| Use of antibiotics, n (%) | 0.07 | 0.798 | ||
| No | 5 (5.05) | 2 (2.99) | ||
| Yes | 94 (94.95) | 65 (97.01) | ||
| Use of CRRT, n (%) | 1.65 | 0.199 | ||
| No | 82 (82.83) | 50 (74.63) | ||
| Yes | 17 (17.17) | 17 (25.37) | ||
| Use of ECMO, n (%) | 0.00 | 1.000 | ||
| No | 92 (92.93) | 62 (92.54) | ||
| Yes | 7 (7.07) | 5 (7.46) | ||
| Hematocrit, Mean ± SD | 31.99 ± 7.31 | 34.12 ± 8.23 | -1.75 | 0.081 |
| Braden scale, M (Q1, Q3) | 13.00 (12.00, 14.00) | 12.00 (11.00, 12.50) | -3.50 | < 0.001 |
| Length of ICU stay (d), M (Q1, Q3) | 6.00 (3.00, 9.00) | 8.00 (6.00, 15.00) | -3.47 | < 0.001 |
| APACHE II score, M (Q1, Q3) | 20.00 (16.00, 23.50) | 21.00 (17.00, 25.00) | -1.17 | 0.242 |
| Tracheal intubation duration (h), M (Q1, Q3) | 110.00 (58.00, 167.50) | 137.00 (88.50, 178.50) | -2.06 | 0.039 |
| White blood cell, M (Q1, Q3) | 10.08 (7.06, 14.80) | 10.06 (7.42, 13.33) | -0.27 | 0.790 |
| Albumin, M (Q1, Q3) | 31.10 (27.00, 34.85) | 31.80 (28.00, 35.40) | -0.96 | 0.337 |
| Hemoglobin, M (Q1, Q3) | 115.00 (102.00, 130.00) | 90.00 (80.00, 109.00) | -6.28 | < 0.001 |
| Platelet, M (Q1, Q3) | 151.00 (94.00, 214.50) | 129.00 (72.50, 188.00) | -1.34 | 0.180 |
| Partial pressure of oxygen, M (Q1, Q3) | 95.00 (72.55, 118.00) | 81.00 (65.80, 109.50) | -1.46 | 0.145 |
BMI: body mass index, CRRT: continuous renal replacement therapy, ECMO: extracorporeal membrane oxygenation, OMPI: oral mucosal pressure injury.
Univariate logistic regression analysis of OMIP in the training set
To further identify potential predictors of OMPI, univariate logistic regression analysis was conducted using OMPI occurrence as the dependent variable and all relevant patient data as independent variables. Variable coding and assignments are detailed in Table 3. As shown in Table 4, lower Braden scale scores (OR = 0.731, P = 0.002), lower hemoglobin levels (OR = 0.942, P < 0.001), longer intubation duration (OR = 1.005, P = 0.015), and longer ICU stay (OR = 1.051, P = 0.022) were significantly associated with an increased risk of OMPI.
Table 3.
Variable assignments for logistic regression analysis
| Type | Assignment | |
|---|---|---|
| OMPI (Outcome) | Binary (Dependent) | Non-OMPI = 0; OMPI = 1 |
| Sex | Binary | Male = 1; Female = 2 |
| Age | Categorical | < 45 = 1; 45-64 = 2; ≥ 65 = 3 |
| BMI | Categorical | < 18.5 = 1; 18.5-24 = 2; ≥ 24 = 3 |
| Fever | Binary | No = 0; Yes = 1 |
| Diabetes | Binary | No = 0; Yes = 1 |
| Vasoactive medication | Binary | No = 0; Yes = 1 |
| Sedative or analgesic medication | Binary | No = 0; Yes = 1 |
| Antibiotics | Binary | No = 0; Yes = 1 |
| CRRT | Binary | No = 0; Yes = 1 |
| ECMO | Binary | No = 0; Yes = 1 |
BMI: body mass index, CRRT: continuous renal replacement therapy, ECMO: extracorporeal membrane oxygenation, OMPI: oral mucosal pressure injury.
Table 4.
Univariate Logistic regression analysis of OMIP in the training set
| β | S.E | P value | OR (95% CI) | |
|---|---|---|---|---|
| Sex | ||||
| Male | Reference | |||
| Female | -0.156 | 0.334 | 0.640 | 0.856 (0.445-1.645) |
| Age | ||||
| < 45 | Reference | |||
| 45-64 | 0.134 | 0.529 | 0.801 | 1.143 (0.405-3.223) |
| ≥ 65 | 0.515 | 0.515 | 0.317 | 1.674 (0.610-4.595) |
| BMI | ||||
| < 18.5 | Reference | |||
| 18.5-24 | 0.194 | 0.369 | 0.599 | 1.214 (0.589-2.500) |
| ≥ 24 | -0.061 | 0.539 | 0.911 | 0.941 (0.327-2.709) |
| Fever | ||||
| No | Reference | |||
| Yes | -0.291 | 0.317 | 0.359 | 0.747 (0.401-1.392) |
| Diabetes | ||||
| No | Reference | |||
| Yes | 0.556 | 0.346 | 0.108 | 1.744 (0.885-3.438) |
| Use of vasoactive drugs | ||||
| No | Reference | |||
| Yes | 0.353 | 0.348 | 0.309 | 1.424 (0.721-2.814) |
| Use of sedative and analgesic drugs | ||||
| No | Reference | |||
| Yes | -0.323 | 0.489 | 0.510 | 0.724 (0.277-1.890) |
| Use of antibiotics | ||||
| No | Reference | |||
| Yes | 0.547 | 0.852 | 0.521 | 1.729 (0.325-9.183) |
| Use of CRRT | ||||
| No | Reference | |||
| Yes | 0.495 | 0.387 | 0.201 | 1.640 (0.768-3.502) |
| Use of ECMO | ||||
| No | Reference | |||
| Yes | 0.058 | 0.608 | 0.924 | 1.060 (0.322-3.491) |
| Braden scale | -0.313 | 0.100 | 0.002 | 0.731 (0.601-0.890) |
| APACHE II score | 0.034 | 0.026 | 0.196 | 1.034 (0.983-1.089) |
| White blood cell | 0.007 | 0.029 | 0.818 | 1.007 (0.951-1.066) |
| Albumin | 0.023 | 0.025 | 0.366 | 1.023 (0.973-1.076) |
| Hemoglobin | -0.060 | 0.011 | < 0.001 | 0.942 (0.922-0.962) |
| Hematocrit | 0.037 | 0.021 | 0.083 | 1.037 (0.995-1.081) |
| Platelet | -0.002 | 0.002 | 0.167 | 0.998 (0.994-1.001) |
| Partial pressure of oxygen | -0.004 | 0.004 | 0.295 | 0.996 (0.988-1.004) |
| Tracheal intubation duration | 0.005 | 0.002 | 0.015 | 1.005 (1.001-1.009) |
| Length of ICU stay (d) | 0.049 | 0.022 | 0.022 | 1.051 (1.007-1.096) |
BMI: body mass index, CRRT: continuous renal replacement therapy, ECMO: extracorporeal membrane oxygenation, OMPI: oral mucosal pressure injury.
Multivariate logistic regression analysis of OMIP in the training set
To identify independent risk factors for OMPI, variables with P < 0.1 from the univariate analysis were entered into a multivariate logistic regression model [15]. As shown in Table 5, independent predictors of OMPI included lower Braden scale scores (OR = 0.770, P = 0.046), lower hemoglobin levels (OR = 0.915, P < 0.001), higher hematocrit levels (OR = 1.176, P < 0.001), and longer intubation duration (OR = 1.008, P = 0.016).
Table 5.
Multivariate Logistic regression analysis of OMIP in the training set
| β | S.E | P value | OR (95% CI) | |
|---|---|---|---|---|
| Braden scale | -0.262 | 0.131 | 0.046 | 0.770 (0.595-0.996) |
| Hemoglobin | -0.089 | 0.015 | < 0.001 | 0.915 (0.889-0.943) |
| Hematocrit | 0.162 | 0.045 | < 0.001 | 1.176 (1.076-1.285) |
| Tracheal intubation duration | 0.008 | 0.003 | 0.016 | 1.008 (1.001-1.014) |
OMPI: oral mucosal pressure injury.
Nomogram development for OMPI prediction
Based on the independent risk factors identified - lower Braden scale scores, lower hemoglobin, higher hematocrit, and longer intubation duration - a nomogram model was constructed to predict OMPI risk in tracheally intubated EICU patients (Figure 1). Since all patients in this study underwent oral intubation and a strong association was established between intubation duration and OMPI, the predictive contribution of intubation duration was not separately validated [6]. In the nomogram, each variable is assigned a point value proportional to its effect size, with the total score corresponding to the predicted risk of OMPI.
Figure 1.
Nomogram prediction model.
Evaluation of nomogram model
The predictive performance of the nomogram was assessed using multiple metrics. As shown in Figure 2A, the AUC was 0.874 (95% CI: 0.816-0.933) in the training set and 0.875 (95% CI: 0.785-0.965) in the validation set, indicating excellent discrimination. Calibration curves (Figure 2B) demonstrated good agreement between predicted and observed probabilities. DCA (Figure 2C) showed that the nomogram provided a high net clinical benefit across a wide range of threshold probabilities, confirming its potential utility in clinical practice.
Figure 2.
ROC, Calibration curves, and DCA curves for training and validation sets. A: ROC curves; B: Calibration curves; C: DCA curves. ROC: receiver operating characteristic, DCA: decision curve analysis.
Discussion
As the most common form of MPI in the ICU, OMPI shares both similarities and differences with skin pressure injuries, warranting increased clinical attention. Currently, there are no dedicated risk assessment tools for OMPI either domestically or internationally. This study analyzed predictive factors for OMPI associated with transoral intubation in 238 EICU patients and evaluated their predictive value for OMPI risk. Our findings indicate that Braden score, hemoglobin, and hematocrit are significant predictors of OMPI, and their combined use demonstrates strong predictive efficacy, underscoring their importance in early prevention and intervention strategies.
Previous studies have shown that the Braden scale exhibits variable reliability, validity, and cut-off values when used in ICU settings, and should therefore be applied with caution [12,16]. Despite these limitations, its utility lies in identifying patients at high risk for pressure injuries and guiding preventive care [12]. Adibelli et al. [17] confirmed the Braden scale as a reliable and valid tool for pressure injury risk assessment in ICU patients. Consistent with these findings, our study supports the Braden scale’s predictive value in identifying OMPI risk among transorally intubated patients in the EICU.
Hemoglobin, a key protein responsible for oxygen transport in red blood cells [18], plays a vital role in maintaining tissue oxygenation. Low hemoglobin levels can result in tissue hypoxia, increasing the risk of pressure injuries [19]. Song et al. [20] highlighted hemoglobin as a marker of hemodynamic status and an important risk factor for pressure injury, suggesting its potential inclusion in future predictive models. In our study, low hemoglobin was independently associated with an increased risk of OMPI, demonstrating good predictive value.
Hematocrit, the proportion of red blood cells in whole blood, typically comprises about 40% of total blood volume [21]. When hemoglobin levels fall below normal, tissue oxygenation may be compromised [22]. Additionally, inflammatory mediators released after mucosal injury can trigger systemic inflammation [23]. Hematocrit may also serve as a marker of oxidative stress, as inflammation can impair erythropoiesis and reduce red cell survival [24,25]. Choi et al. [11] identified hematocrit as a risk factor for upper OMPI. Interestingly, our study found that OMPI patients had both lower hemoglobin and higher hematocrit levels - an apparently contradictory finding. We hypothesize that this may be attributed to hypoxia and dehydration frequently observed in EICU patients [21].
Although some studies have identified the APACHE II score as a predictor of MPI and OMPI in ICU patients [26,27], our results did not show a significant difference in APACHE II scores between the OMPI and non-OMPI groups. APACHE II, a revised version of the original APACHE system proposed by Knaus et al. [28], is widely used for evaluating disease severity and predicting ICU outcomes. While some studies [29,30] support a positive association between APACHE II scores and pressure injury development, others, including a study on nasal MPI [31], report no such correlation. Our findings align with the latter, suggesting that APACHE II may not be a reliable predictor for OMPI in this specific context.
This study has several limitations. First, it is a retrospective study with a relatively small sample size, which may limit generalizability. Future studies with larger, multicenter cohorts are needed to enhance predictive accuracy. Second, the effect of hematocrit on OMPI risk remains complex and requires further investigation. Third, while the nomogram was internally validated, external validation using independent cohorts or public databases is essential to confirm its robustness and applicability.
In conclusion, Braden score, hemoglobin, and hematocrit were identified as significant predictors of OMPI in ICU patients. Their combined application yields high predictive accuracy, offering valuable insight for early prevention and intervention strategies aimed at reducing patient discomfort and improving clinical outcomes.
Acknowledgements
This work was supported by Anhui Provincial Medical and Healthcare Key Specialty Construction Project (No. [2021] 273-3).
Disclosure of conflict of interest
None.
References
- 1.Rubulotta F, Blanch Torra L, Naidoo KD, Aboumarie HS, Mathivha LR, Asiri AY, Sarlabous Uranga L, Soussi S. Mechanical ventilation, past, present, and future. Anesth Analg. 2024;138:308–325. doi: 10.1213/ANE.0000000000006701. [DOI] [PubMed] [Google Scholar]
- 2.Walter K. Mechanical ventilation. JAMA. 2021;326:1452. doi: 10.1001/jama.2021.13084. [DOI] [PubMed] [Google Scholar]
- 3.Taboada M, Almeida X, Cariñena A, Costa J, Carmona-Monge J, Agilda A, Barreiro L, Castillo J, Williams K, Segurola J, Álvarez J, Seoane-Pillado T. Complications and degree of difficulty of orotracheal intubation in the Intensive Care Unit before and after the establishment of an intubation protocol for critically ill patients: a prospective, observational study. Rev Esp Anestesiol Reanim (Engl Ed) 2024;71:17–27. doi: 10.1016/j.redare.2023.12.004. [DOI] [PubMed] [Google Scholar]
- 4.TEAM Study Investigators and the ANZICS Clinical Trials Group. Hodgson CL, Bailey M, Bellomo R, Brickell K, Broadley T, Buhr H, Gabbe BJ, Gould DW, Harrold M, Higgins AM, Hurford S, Iwashyna TJ, Serpa Neto A, Nichol AD, Presneill JJ, Schaller SJ, Sivasuthan J, Tipping CJ, Webb S, Young PJ. Early active mobilization during mechanical ventilation in the ICU. N Engl J Med. 2022;387:1747–1758. doi: 10.1056/NEJMoa2209083. [DOI] [PubMed] [Google Scholar]
- 5.Erbay Dallı Ö, Ceylan İ, Kelebek Girgin N. Incidence, characteristics and risk factors of medical device-related pressure injuries: an observational cohort study. Intensive Crit Care Nurs. 2022;69:103180. doi: 10.1016/j.iccn.2021.103180. [DOI] [PubMed] [Google Scholar]
- 6.Zhao Y, Guo J, Ma J, Ge Y, Wang J, Li C, Shi C. Characteristics of oral mucosal pressure injuries in children with orotracheal intubation in intensive care units: an observational study. Nurs Crit Care. 2025;30:e13174. doi: 10.1111/nicc.13174. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Chen G, Li X, Li X, Liu S, Xie J. Mucosal membrane pressure injury in intensive care units: a scoping review. Intensive Crit Care Nurs. 2024;80:103560. doi: 10.1016/j.iccn.2023.103560. [DOI] [PubMed] [Google Scholar]
- 8.Hampson J, Green C, Stewart J, Armitstead L, Degan G, Aubrey A, Paul E, Tiruvoipati R. Impact of the introduction of an endotracheal tube attachment device on the incidence and severity of oral pressure injuries in the intensive care unit: a retrospective observational study. BMC Nurs. 2018;17:4. doi: 10.1186/s12912-018-0274-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Kim CH, Kim MS, Kang MJ, Kim HH, Park NJ, Jung HK. Oral mucosa pressure ulcers in intensive care unit patients: a preliminary observational study of incidence and risk factors. J Tissue Viability. 2019;28:27–34. doi: 10.1016/j.jtv.2018.11.002. [DOI] [PubMed] [Google Scholar]
- 10.Amrani G, Gefen A. Which endotracheal tube location minimises the device-related pressure ulcer risk: the centre or a corner of the mouth? Int Wound J. 2020;17:268–276. doi: 10.1111/iwj.13267. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Choi BK, Kim MS, Kim SH. Risk prediction models for the development of oral-mucosal pressure injuries in intubated patients in intensive care units: a prospective observational study. J Tissue Viability. 2020;29:252–257. doi: 10.1016/j.jtv.2020.06.002. [DOI] [PubMed] [Google Scholar]
- 12.Mehicic A, Burston A, Fulbrook P. Psychometric properties of the Braden scale to assess pressure injury risk in intensive care: a systematic review. Intensive Crit Care Nurs. 2024;83:103686. doi: 10.1016/j.iccn.2024.103686. [DOI] [PubMed] [Google Scholar]
- 13.Ding Y, Yan Y, Niu J, Zhang Y, Gu Z, Tang P, Liu Y. Braden scale for assessing pneumonia after acute ischaemic stroke. BMC Geriatr. 2019;19:259. doi: 10.1186/s12877-019-1269-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Tang H, Jin Z, Deng J, She Y, Zhong Y, Sun W, Ren Y, Cao N, Chen C. Development and validation of a deep learning model to predict the survival of patients in ICU. J Am Med Inform Assoc. 2022;29:1567–1576. doi: 10.1093/jamia/ocac098. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Ye S, Pan H, Li W, Wang J, Zhang H. Development and validation of a clinical nomogram for differentiating hemorrhagic and ischemic stroke prehospital. BMC Neurol. 2023;23:95. doi: 10.1186/s12883-023-03138-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Lovegrove J, Fulbrook P, Yuan C, Lin F, Liu XL. The Chinese mandarin COMHON index and braden scale to assess pressure injury risk in intensive care: an inter-rater reliability and convergent validity study. Aust Crit Care. 2025;38:101093. doi: 10.1016/j.aucc.2024.05.015. [DOI] [PubMed] [Google Scholar]
- 17.Adibelli S, Korkmaz F. Pressure injury risk assessment in intensive care units: comparison of the reliability and predictive validity of the Braden and Jackson/Cubbin scales. J Clin Nurs. 2019;28:4595–4605. doi: 10.1111/jocn.15054. [DOI] [PubMed] [Google Scholar]
- 18.Ciaccio C, Coletta A, Coletta M. Role of hemoglobin structural-functional relationships in oxygen transport. Mol Aspects Med. 2022;84:101022. doi: 10.1016/j.mam.2021.101022. [DOI] [PubMed] [Google Scholar]
- 19.Estep TN. Issues in the development of hemoglobin based oxygen carriers. Semin Hematol. 2019;56:257–261. doi: 10.1053/j.seminhematol.2019.11.006. [DOI] [PubMed] [Google Scholar]
- 20.Song W, Kang MJ, Zhang L, Jung W, Song J, Bates DW, Dykes PC. Predicting pressure injury using nursing assessment phenotypes and machine learning methods. J Am Med Inform Assoc. 2021;28:759–765. doi: 10.1093/jamia/ocaa336. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Ekanayake P, Mudaliar S. Increase in hematocrit with SGLT-2 inhibitors - hemoconcentration from diuresis or increased erythropoiesis after amelioration of hypoxia? Diabetes Metab Syndr. 2023;17:102702. doi: 10.1016/j.dsx.2022.102702. [DOI] [PubMed] [Google Scholar]
- 22.Sitina M, Stark H, Schuster S. Optimal hematocrit theory: a review. J Appl Physiol (1985) 2024;137:494–509. doi: 10.1152/japplphysiol.00034.2024. [DOI] [PubMed] [Google Scholar]
- 23.Estornut C, Rinaldi G, Carceller MC, Estornut S, Pérez-Leal M. Systemic and local effect of oxidative stress on recurrent aphthous stomatitis: systematic review. J Mol Med. 2024;102:453–463. doi: 10.1007/s00109-024-02434-8. [DOI] [PubMed] [Google Scholar]
- 24.Tijjani A, Yusuf AA, Kuliya-Gwarzo A. Lower hematocrit is associated with high oxidative stress and low antioxidant status in men occupationally exposed to petrol. Niger J Basic Clin Sci. 2024;21:65–72. [Google Scholar]
- 25.Hanson AL, Mulè MP, Ruffieux H, Mescia F, Bergamaschi L, Pelly VS, Turner L, Kotagiri P Cambridge Institute of Therapeutic Immunology and Infectious Disease-National Institute for Health Research (CITIID-NIHR) COVID BioResource Collaboration. Göttgens B, Hess C, Gleadall N, Bradley JR, Nathan JA, Lyons PA, Drakesmith H, Smith KGC. Iron dysregulation and inflammatory stress erythropoiesis associates with long-term outcome of COVID-19. Nat Immunol. 2024;25:471–482. doi: 10.1038/s41590-024-01754-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Tang W, Zha ML, Zhang WQ, Hu SQ, Chen HL. APACHE scoring system and pressure injury risk for intensive care patients: a systematic review and meta-analysis. Wound Repair Regen. 2022;30:498–508. doi: 10.1111/wrr.13021. [DOI] [PubMed] [Google Scholar]
- 27.Lu W, Zhang J, Qiu Y, Fei N, Yin L. Correlations between APACHE-II score and pressure parameters of mechanical ventilation in patients with ARDS and their value in prognostic evaluation. Pak J Med Sci. 2023;39:1584–1588. doi: 10.12669/pjms.39.6.7190. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Knaus WA, Draper EA, Wagner DP, Zimmerman JE. APACHE II: a severity of disease classification system. Crit Care Med. 1985;13:818–829. [PubMed] [Google Scholar]
- 29.Kumta N, Coyer F, David M. Perioperative factors and pressure ulcer development in postoperative ICU patients: a retrospective review. J Wound Care. 2018;27:475–485. doi: 10.12968/jowc.2018.27.8.475. [DOI] [PubMed] [Google Scholar]
- 30.Sayan HE, Girgin NK, Asan A. Prevalence of pressure ulcers in hospitalized adult patients in Bursa, Turkey: a multicentre, point prevalence study. J Eval Clin Pract. 2020;26:1669–1676. doi: 10.1111/jep.13354. [DOI] [PubMed] [Google Scholar]
- 31.Shapira-Galitz Y, Karp G, Cohen O, Halperin D, Lahav Y, Adi N. Evaluation and predictors for nasogastric tube associated pressure ulcers in critically Ill patients. Isr Med Assoc J. 2018;20:731–736. [PubMed] [Google Scholar]