By Naresh Kumar Veerabattini, MBBS, and Alexander Niven, MD

Dr. Veerabattini is a Research Trainee, Division of Pulmonary/Critical Care Medicine, Mayo Clinic, Rochester, MN.
Dr. Niven is Senior Associate Consultant, Division of Pulmonary/Critical Care Medicine, Mayo Clinic, Rochester, MN.

Drs. Veerabattini and Niven report no financial relationships relevant to this field of study.

SYNOPSIS: The authors of this retrospective analysis of data from two prospective randomized trials of tracheal intubation developed the AT RISK (Age, Trainee, Race, Indication, SpO2, Kg/m2) score to identify patients who may develop severe hypoxemia during this procedure. An AT RISK score of < 2 carried a high negative predictive value of procedural hypoxemia, while nearly half of patients with a score > 3 experienced an SpO2 < 80%.

SOURCE: McKown AC, Casey JD, Russell DW, et al. Risk factors for and prediction of hypoxemia during tracheal intubation of critically ill adults. Ann Am Thorac Soc 2018; Aug 15. doi: 10.1513/AnnalsATS.201802-118OC. [Epub ahead of print].

Complications from endotracheal intubation in the critically ill remain unacceptably high. Well-designed clinical trials over the past decade have revealed hypoxemia rates of 20% despite optimal patient positioning and preoxygenation.1,2 Severe hypoxemia has been associated with the need for emergent surgical airway, anoxic brain injury, prolonged ICU stay, and death. Poor prospective identification of at-risk patients, inadequate preprocedural planning, and lack of communication have been identified as risk factors for adverse airway events.3 The MACOCHA score offers the best validated tool to identify critically ill patients at risk for difficult intubation, but clinical adoption has been limited by its length and complexity.4 Investigators conducted a retrospective analysis of 442 tracheal intubations from two previously published prospective randomized trials to identify critically ill patients at risk for hypoxemia during this procedure. The first was a single-center trial of 150 intubations randomized to direct or video laryngoscopy, with further factorialized randomization to apneic vs. no apneic oxygenation. The second was a multicenter trial of 292 intubations randomized to a sniffing or modified ramped position, co-enrolled with randomization to use of usual care or a written preprocedural checklist. No intervention demonstrated a treatment effect. The database for this retrospective study included all patients > 18 years of age requiring endotracheal intubation in the ICU, with a supervised pulmonary and critical care fellow or anesthesiology trainee performing almost all initial procedures. The primary outcome was the lowest arterial oxygen saturation (SpO2) from induction medication administration until two minutes after induction, which independent observers collected in real time in all cases.

Only nine patients were excluded from the original study databases because of incomplete pulse oximetry data, leaving 433 intubation events for analysis. Potential predictors of hypoxemia were selected a priori. Investigators developed a linear regression model. To create the model, investigators used covariates of race, gender, body mass index (BMI), age, use of noninvasive ventilation (NIV) or maximum fraction of inspired O2 in the preceding six hours, APACHE II scores, indication for intubation (hypoxemic or hypercarbic respiratory failure or other), presence of sepsis, and operator experience. Then, the authors created a logistic regression model to identify risk factors for severe hypoxemia (defined as SpO2 < 80% or a decrease in SpO2 of > 10% for patients with an SpO2 at induction < 90%) with the same covariates (except for recent FiO2 or NIV use, sepsis, or APACHE II scores). The investigators performed multiple sensitivity analyses to assess for confounding variables. Finally, the authors compiled all significant variables from the severe hypoxemia logistic regression model into a point system for bedside risk-stratification using a penalized maximum likelihood estimation to improve generalizability.

Seventy-five percent of patients in the cohort had sepsis or septic shock, and 36% were intubated for hypoxemic respiratory failure. The median APACHE II score was 22 (interquartile range, 17-26). Critical care and pulmonary fellows performed almost all intubations. The authors used a linear multiple regression model to compare the 75th percentile to 25th percentile for each continuous variable. They found higher BMI, ethnicity other than black, younger age, hypoxemic respiratory failure as an indication for intubation, lower SpO2 at induction, and more limited operator intubating experience to be independently associated with a lower SpO2 during intubation. These factors persisted as independent predictors of severe hypoxemia in the simplified logistic multiple regression model, with septic shock the only other independent predictor identified in the sensitivity analysis. Of these risk factors, the presence of hypoxemic respiratory failure and SpO2 at induction were the strongest predictors of severe hypoxemia during intubation.

Simplification of a penalized maximum likelihood estimation algorithm resulted in the creation of the six-point AT RISK score. The authors assigned one point each to Age < 50 years, Trainee (operator with < 100 prior intubations), Race other than black, Indication (hypoxemic respiratory failure), SpO2 at induction < 94%, and Kg/m2 (BMI) > 35. Based on this predictive model, less than 2% of patients with an AT RISK score < 2 will develop significant desaturation during intubation (negative predictive value, 98.1%; 95% confidence interval [CI], 93.2-100). Nearly half of patients with a score > 3 will experience severe hypoxemia (positive predictive value, 47.4%; 95% CI, 37.1-59.1).


Benjamin Franklin’s adage that “an ounce of prevention is worth a pound of cure” is especially true of tracheal intubation in the critically ill. Often, intubation in the ICU is emergent, limiting opportunities for systematic airway assessment and using anatomic features with limited predictive power to identify a difficult airway.5 Patients often present with significant cardiopulmonary disease, hemodynamic instability, and a full stomach, which makes preoxygenation difficult and the risk of desaturation and aspiration during intubation attempts high. Unfortunately, these factors are not under the practicing intensivist’s control. The 4th National Audit Project of the Royal College of Anaesthetists and The Difficult Airway Society was the largest of a series of studies that have consistently identified an association between airway management complications and the lack of systematic preoxygenation, airway equipment and training, inconsistent communication and teamwork, and the absence of a clearly articulated backup plan if initial intubation attempts fail.3 The intubation team can control these factors and offer an important opportunity to improve outcomes from ICU intubation.

The McKown et al study is an important step toward more effective, early identification of patients at increased risk of severe hypoxemia who can benefit most from deliberate preoxygenation efforts, teamwork, and planning to prevent this potentially life-threatening complication. Strengths of the AT RISK score include the high-quality methods used to collect the data employed for this study, the systematic data analysis, and the simplicity and performance characteristics of this tool. The similarity between many elements of the AT RISK and MACOCHA scores also strengthens the external validity of these results.4

As with any retrospective study, implementation of the AT RISK tool without prospective validation using a new patient cohort should be undertaken with caution. The reasons that younger patients were at greater risk for severe hypoxemia are difficult to discern from this data analysis and underline the importance of a systematic and deliberate approach to intubation planning and preparation, regardless of the patient and perceived physiologic reserve. The authors went to great lengths to discuss the effect of intubator experience on severe hypoxemia. It is difficult to draw meaningful conclusions regarding training or practice based on these data alone. There is significant literature suggesting that even experienced airway managers receive limited systematic training in emergent airway management in the critically ill. The difference in lowest oxygen saturation seen between intubators with 50 and 85 prior procedures in the McKown et al study was clinically insignificant. Efforts to use a systematic curriculum that employed a blend of didactics, simulation, and supervised clinical experience in pulmonary and critical care fellows with lower intubation volume have yielded equivalent outcomes and complication rates. This suggests procedure numbers alone should not be used as a sole criterion for competence, credentialing, and procedural scope of practice.6

These findings underline the fact that obese patients with hypoxemic respiratory failure are perhaps the greatest at-risk population of critically ill patients for airway management. Deliberate preoxygenation, preprocedural planning, and preparation in this setting also are vital. Unfortunately, the incidence of severe hypoxemia seen in even these high-quality clinical trials using best practices to minimize patient hypoxemia during intubation remains unacceptably high. Given the conflicting data with previous commonly held practices of apneic oxygenation, the “ramped” position, and the use of intubation checklists in ICU patients, we look forward to other researchers using the AT RISK tool in future studies to identify better strategies to reduce the risk of intubation in the critically ill.7-12


  1. Griesdale DE, Bosma TL, Kurth T, et al. Complications of endotracheal intubation in the critically ill. Intensive Care Med 2008;34:1835-1842.
  2. Janz DR, Semler MW, Lentz RJ, et al. Randomized trial of video laryngoscopy for endotracheal intubation of critically ill adults. Crit Care Med 2016;44:1980-1987.
  3. Cook TM, Woodall N, Harper J, Benger J; Fourth National Audit Project. Major complications of airway management in the UK: Results of the Fourth National Audit Project of the Royal College of Anaesthetists and the Difficult Airway Society. Part 2: Intensive care and emergency departments. Br J Anaesth 2011;106:632-642.
  4. De Jong A, Molinari N, Terzi N, et al. Early identification of patients at risk for difficult intubation in the intensive care unit: Development and validation of the MACOCHA score in a multicenter cohort study. Am J Respir Crit Care Med 2013;187:832-839.
  5. Naguib M, Scamman FL, O’Sullivan C, et al. Predictive performance of three multivariate difficult intubation models: A double-blind, case-controlled study. Anesth Analg 2006;102:818-824.
  6. Mosier JM, Malo J, Sakles JC, et al. The impact of a comprehensive airway management training program for pulmonary and critical care fellows. Ann Am Thorac Soc 2015;12:539-548.
  7. Oliveira J E Silva L, Cabrera D, Barrionuevo P, et al. Effectiveness of apneic oxygenation during intubation: A systematic review and meta-analysis. Ann Emerg Med 2017;70:483-494.e11.
  8. Semler MW, Janz DR, Lentz RJ, et al. Randomized trial of apneic oxygenation during endotracheal intubation of the critically ill. Am J Respir Crit Care Med 2016;193:273-280.
  9. Collins JS, Lemmens HJ, Brodsky JB, et al. Laryngoscopy and morbid obesity: A comparison of the “sniff” and “ramped” positions. Obes Surg 2004;14:1171-1175.
  10. Semler MW, Janz DR, Russell DW, et al. A multicenter, randomized trial of ramped position vs sniffing position during endotracheal intubation of critically ill adults. Chest 2017;152:712-722.
  11. Janz DR, Semler MW, Joffe AM, et al. A multicenter randomized trial of a checklist for endotracheal intubation of critically ill adults. Chest 2018;153:816-824.
  12. Doerschug KC, Niven AS. Checklists for safety during ICU intubations: The details matter. Chest 2018;153:1505-1506.