Incidence of and risk factors for severe cardiovascular collapse after endotracheal intubation in the ICU: a multicenter observational study

This was a secondary analysis of patients in the FRIDAREA study database [7]. Briefly, FRIDAREA was a prospective, observational, multicenter study conducted in 42 ICUs to develop a model predictive of difficult intubation (original cohort), and in 18 ICUs to validate the model (validation cohort) [7]. All adult patients consecutively intubated in involved ICUs were included. Exclusion criteria were pregnancy, refusal to participate after information was provided or age <18 years. The primary endpoint was the incidence of severe CVC after ETI in the ICU and secondary endpoints were risk factors predictive of severe CVC in these critically ill patients and evaluated mortality at 28 days.

Because of the observational, noninvasive design of this study, the need for written consent was waived. The local ethics committee, the Comité de Protection des Personnes Sud-Mediterranée III, approved the study design (code UF 8819, register 2011-A001122-39).

Clinical parameters were prospectively assessed before, during, and after intubation procedures, with an independent observer collecting variables during and after intubation. Data assessed before intubation included: age; body mass index; severity score (modified Simplified Acute Physiologic Score (SAPS) II at admission, with age eliminated to avoid colinearity with age in the multivariate analysis, as described previously [8, 9]); Sequential Organ Failure Assessment (SOFA) score on the day of the procedure; type of admission (medical versus surgical); co-morbidities such as alcoholism, smoking, cirrhosis, and chronic obstructive pulmonary disease (COPD); cause of admission; cause of intubation (coma was defined as a Glasgow score <8); date and hour of intubation (daytime procedures were those performed from 8 am to 7 pm, with all others defined as on-call procedures); intubation during the previous 2 weeks; nature and number of operators; fluid loading; vasopressor use; noninvasive ventilation (NIV); and emergency characteristics of the intubation (with a real emergency defined as the requirement for immediate ETI, relative emergency as ETI required within 1 hour, and deferred emergency as ETI required in >1 hour). Just before intubation preoxygenation and method of preoxygenation (standard versus NIV) were recorded.

Minimal and maximal heart rate, arterial pressure and saturation were measured before intubation, during intubation (between induction of anesthesia and tube insertion) and within 30 minutes after intubation. Drugs used for intubation were recorded. The MACOCHA score, a seven-item (Mallampati score III or IV, obstructive sleep apnea syndrome, reduced mobility of cervical spine, limited mouth opening, severe hypoxia, coma, nonanesthesiologist as operator) predictor of difficult intubation with a cut-off of 3 points [7], was assessed.

Finally, the percentages of patients undergoing capnography (end-tidal CO₂ curve) and esophageal intubation; the rates of agitation, aspiration, cardiac arrest, and arrhythmias; ventilation parameters; and 28-day mortality rate were evaluated.

Arterial blood pressure was monitored continuously in patients carrying an intra-arterial catheter for 5 of 30 minutes after intubation in patients with cuff measurements. Severe CVC was defined as systolic blood pressure ≤65 mmHg recorded at least once and/or ≤90 mmHg lasting ≥30 minutes despite vascular loading with 500–1000 mL crystalloid and/or colloid solutions and/or a requirement for vasoactive support [1‐4]. To avoid confounding factors of hemodynamic variability, only patients who were hemodynamically stable before intubation (defined as mean arterial blood pressure >65 mmHg without vasoactive drugs [10, 11]) were included. Therefore, patients intubated for cardiac arrest and shock were secondarily excluded.

All statistical analyses were performed using Stata software, version 12 (StataCorp, College Station, Tx, USA). The tests were two-sided, with a type I error set at α = 0.05. Mean and standard deviation (SD) or median and interquartile range were calculated for continuous variables, and number of patients and associated percentages were calculated for categorical parameters. Categorical variables were compared between independent groups using the Chi-squared test or Fisher’s exact test, and continuous variables were compared using Student's t-test or the Mann-Whitney test, with normality verified by the Shapiro-Wilk test and homoscedasticity by the Fisher-Snedecor test. Factors significant in the univariate analysis (p < 0.1 [12, 13]) and parameters deemed clinically relevant [6, 14, 15], such as use of ketamine or etomidate, COPD and fluid challenge >500 mL (adjustment factors), were included in backward and forward stepwise multivariate logistic regression analyses to determine risk factors independently associated with CVC. The interactions between possible predictive factors were also tested. Results were expressed as odds ratios (ORs) and 95 % confidence intervals (CIs). Multivariate analysis consisted of three steps: 1) patient characteristics (first model), 2) intubation procedures in the ICU (second model), and 3) parameters statistically significant in these two models (model). Univariate analyses identified a cut-off of inspired oxygen concentration (FiO₂) of 70 % and age in three stages (<60, 60–75 and >75 years) for testing in the multivariate models. The Hosmer-Lemeshow test was used to assess the goodness of fit of the logistic model. A cross-validation process was considered to assess the goodness of fit of the final models obtained. Following these multivariate analyses, a receiver-operating characteristic (ROC) curve associated with the occurrence of CVC (model) was plotted for each proposed model. The statistical power of the final was tested according to works proposed by Tosteson and Demidenko [16, 17].