By Samuel Nadler, MD, PhD

Critical Care, Pulmonary Medicine, The Polyclinic Madison Center, Seattle; Clinical Instructor, University of Washington, Seattle

Dr. Nadler reports no financial relationships relevant to this field of study.

SYNOPSIS: The presence of four or more of the I-TRACH criteria (Intubation in the ICU, Tachycardia [heart rate > 110], Renal failure [blood urea nitrogen > 25], Acidosis [pH < 7.25], Creatinine > 2, HCO3 < 20) accurately predicted those patients who would require mechanical ventilation for seven or more days.

SOURCE: Clark PA, Inocencio RC, Lettieri CJ. I-TRACH: Validating a tool for predicting prolonged mechanical ventilation. J Intensive Care Med 2018;33:567-573.

The duration of mechanical ventilation carries significant implications for patient care. Many patients and their families report they do not wish to be on life support for a prolonged period and would opt not to be put on a ventilator if that were the expected outcome. However, it can be challenging to predict which patient likely will experience a lengthy course on a ventilator.

The I-TRACH trial was a prospective, observational study of 225 patients that sought to validate a prediction tool that identifies patients at greatest risk for prolonged mechanical ventilation (PMV). A previous study of 99 patients generated this prediction tool.1 In a multivariate analysis, the variables statistically associated with PMV included: intubation in the ICU, heart rate > 110 beats per minute (bpm), blood urea nitrogen > 25 mg/dL, pH < 7.25, creatinine > 2 mg/dL, and serum bicarbonate < 20 mEq/L. For this study, PMV was defined as requiring > 7 or 14 days of mechanical support. Individually, none of these values were very predictive of PMV (odds ratios ranging from 1.26 for heart rate > 110 bpm to 1.91 for intubation in ICU). However, if four of more of these criteria were met, this tool demonstrated a receiver operator curve (ROC) of 0.824 for predicting PMV. This compared favorably with other screening tools such as Acute Physiology and Chronic Health Evaluation (APACHE) III (ROC = 0.634), APACHE II (ROC = 0.652), Sequential Organ Failure Assessment (ROC = 0.608), and Acute Physiology Score (ROC = 0.575).

COMMENTARY

The results of this study validate a clinical tool using common ICU variables for predicting prolonged mechanical ventilation. The ROC of 0.824 is better than other prediction models. However, one must consider both the implications and strength of any predictive model for clinical decision-making. For PMV defined as > 14 days, the positive predictive value was 45.7%. Fewer than half of patients who met four or more of these criteria required mechanical ventilation for more than two weeks. Meanwhile, the negative predictive value was 89.8%. This implies that not meeting four of these clinical criteria can give confidence that a patient will not need PMV, but meeting these criteria is not sufficiently predictive to direct discussion of forgoing intubation or the need for tracheostomy at the time of intubation.

Additionally, there was a large portion of this cohort with neurologic conditions, such as Guillain-Barré, syndrome that represent a very different population from most ICUs. In this study, 43.8% of patients intubated had underlying neurologic disease.

Another goal of this clinical tool was to identify patients who might benefit from early tracheostomy. Again, a positive predictive value of 45.7% does not seem strong enough to consider an invasive surgical procedure early in the course of mechanical ventilation. Furthermore, the benefit of early tracheostomy remains uncertain. The TRACHMAN trial did not show significant benefit in those patients who underwent tracheostomy within four days vs. those who received a tracheostomy after 10 days.2 However, a limitation of this study was an inability to predict patients who ultimately required tracheostomy. Other researchers who use these criteria to randomize patients for early vs. late tracheostomy might be able to demonstrate benefit.

REFERENCES

  1. Clark PA, Lettieri CJ. Clinical model for predicting prolonged mechanical ventilation. J Crit Care 2013;28:880.e1-7.
  2. Young D, Harrison DA, Cuthbertson BH, Rowan K; TracMan Collaborators. Effect of early vs late tracheostomy placement on survival in patients receiving mechanical ventilation: The TracMan randomized trial. JAMA 2013;309:2121-2129.