Abstract
Background
Regarding general anesthesia, postoperative reintubation after planned extubation (RAP) is needed when circumstances compel us to do so, irrespective of prolongation of hospital stay and increase of medical expenditure. We describe here our implementation of a case-controlled model to identify the risk factors of RAP.
Methods
Patients who saw RAP in the space from January 1, 2005 to December 31, 2007 were retrospectively sorted out from the Quality Assurance (QA) database of the Department of Anesthesiology. We compared RAP cases with the control group and analyzed the factors using descriptive statistics and logistic regression. Reintubation was defined as intubation after the extubation for the initial endotracheal intubation, for general anesthesia, at the time period before departure from the post-anesthesia care unit.
Results
Of the 137,866 patients who underwent endotracheal intubation for general anesthesia, 83 (0.06%) sustained RAP. The control group included 249 patients randomly selected for endotracheal intubation without RAP. Twenty-two variables, including demographic, operative and anesthetic data, were analyzed. We found that patients with preoperative COPD (odds ratio: 7.17, 95% CI: 1.98–26.00), pneumonia (odds ratio: 7.94, 95% CI: 1.93–32.78), ascites (odds ratio: 13.76, 95% CI: 1.08–174.74) and systemic inflammatory response syndrome (SIRS) (odds ratio: 11.90, 95% CI: 2.63–53.86) were more likely to be subjected to RAP. Airway surgery and head–neck surgery also predisposed patients to reintubation. However, administration of both an extra dose of opioid and neuromuscular blocker at the end of surgery proved irrelevant to RAP.
Conclusions
Risk factors for RAP are clear and unambiguous. This study will prompt further studies on preventative measures or evaluation of how to improve outcome.
Keywords
reintubation; risk factors; quality assurance, health care;
1. Introduction
General anesthesia with endotracheal intubation and controlled ventilation is the most common anesthetic technique currently in use. Patients are often extubated at the end of surgery, while those who need postoperative ventilator support are not and are thus transferred to the intensive care unit (ICU) until circumstances allow extubation. Although extubation immediately in the wake of surgery is generally safe and practicable, postoperative reintubation after planned extubation (RAP) is mandatory when patient’s conditions deteriorate, such as occurrence of hypoxia, hypercapnia or respiratory failure. However, RAP is associated with increased complication morbidity, hospital cost and late mortality.1, 2 The incidence of RAP ranges from 0.17 to 0.27%.3, 4 A retrospective study at the University of Michigan concluded that respiratory problems were the most common causes of reintubation.3 A multi-center study in Thailand stated that the three main triggering matters are residual neuromuscular blocking and anesthetic effects, upper airway obstruction, and unstable hemodynamics.4 However, both of these studies lacked controlled evidence.
The present study was conducted through case-controlled model in an attempt to determine the factors that might lead to RAP, by making use of the Quality Assurance (QA) database of the Department of Anesthesiology for collecting data.
2. Materials and methods
After obtaining approval from the Review Board of Chang Gung Memorial Hospital, we retrospectively analyzed all perioperative reintubation cases from January 1, 2005, through December 31, 2007, from our QA anesthesiology database. The QA database is composed of a number of a standardized forms (database cards) containing demographic data, patient comorbidities, anesthetic technique used, and perioperative anesthetic management details such as monitoring, drugs and fluids administered, and minor or major perioperative complications. In our practice, once patients have received anesthesia for surgery, these forms must be filled out by anesthesiologists and registered in database by nurses at the time of care.
All major complications are reviewed and analyzed by the department’s quality control group each month. The said group is composed of the chairman of the anesthesiology department, one resident and five attending anesthesiologists, and five certified nurse anesthetists. They discuss the causes of mischiefs, management techniques, adverse outcomes, and suggest preventive measures for recommitments. After monthly analysis, details are presented in a morbidity and mortality conference where all personnel are made certain the outcomes of quality assessment to initiate a positive impact on their practice.
In the present study, reintubation is defined as intubation after removal of the endotracheal tube for the initial intubation for general anesthesia. The time period covered the commencement of extubation in the operating room (OR) until the entire stay in the post-anesthesia care unit (PACU). Criteria for appropriateness of tracheal extubation in the OR and timing of transfer to the PACU were at the discretion of the anesthesiologist in charge. Neostigmine in proper dose was routinely administered before extubation. Patients who underwent intubation for general anesthesia and were successfully extubated served as the control group, whose chart numbers of above-mentioned candidates were entered in SPSS (SPSS Inc., Troy, NY).
Variables were assessed for both RAP cases and controls; medical records were reviewed if additional data were required. Preoperative variables included age, gender, body weight, ASA physical status, emergency, and specific comorbidities [CNS disorder, conscious disturbance, congestive heart failure (CHF), angina or myocardial ischemia, COPD, preoperative pneumonia or pleural effusion, chronic renal failure, ascities/cirrhosis, and SIRS (systemic inflammatory response syndrome)]. The criteria for SIRS in this study were as follows: body temperature >38°C or <36°C, heart rate >90/min, respiratory rate >20/min or arterial CO2 <32 mmHg, and WBC >12,000 mm3 or <4000 mm3 or >10% band form. Patient height as with body mass index (BMI) was not always noted, rendering it a limitation of this study.
Operative data included type of surgery (orthopedic, head and neck, airway, neurosurgery, abdominal, others), preoperative oxygen saturation breathing room air, surgical duration, and fluid balance. The Surgical sites of the airway ranged from the oropharynx (tongue base), the hypopharynx to the larynx. Other head and neck surgical sites, exclusive of airway or neurosurgery, were attributed to head and neck surgery. Fluid balance was defined as the periodic total input (crystalloids, synthetic colloids and blood products) minus total output (basic maintenance, insensible loss, urine and blood loss) from the time of endotracheal intubation to the end of surgery.
Balanced anesthesia, defined as one dose of opioid and/or non-depolarizing neuromuscular blocking agent (NMB) given at the end of surgery, had been considered as a probable cause of reintubation by our anesthesia faculty; thus, we also collected anesthesic-related data concerning whether or not the last dose of opioid or muscle relaxant was given within 45 min prior to extubation. The limitation here is that multiple drugs, including opioids (fentanyl, morphine, meperidine and nalbuphine) and NMBs (cisatracurium and rocuronium), in a multi-dose regimen were administrated.
2.1. Statistical analysis
All data were entered into a Microsoft Excel (Microsoft Corp., Redmond, WA) spreadsheet and statistically analyzed using SPSS version 17.0 (SPSS Inc., Troy, NY). Differences in the distribution of preoperative, operative, and anesthetic-related variables between RAP cases and controls were assessed using descriptive statistics for mean and standard deviation of continuous variables; frequency tables for categorical variables and logistic regression for p-values, odds ratios and 95% confidence intervals (CIs). Predictive validity, sensitivity and specificity of the predictive model were estimated through receiver–operator-characteristic curve analysis. To estimate the simultaneous effects of multiple variables on reintubation, we used multivariate logistic regression for significant variables (p < 0.05) after univariate logistic regression.
3. Results
Of the 137,866 patients who underwent general anesthesia with endotracheal tube intubation from January 1, 2005, through December 31, 2007, 83 (0.06%) sustained RAP. Two hundred and forty-nine controls were randomly selected in the same period. Among the RAP cases, 65 (78.3%) saw the event in the OR while 18 (21.7%) saw it in the PACU. Reasons for RAP were prolonged uneasy breathing (33 patients; 39.8%), airway obstruction (26; 31.2%), surgical complication (17; 20.4%), and hypotension (7; 8.4%). Most RAPs (65; 78.3%) were performed within 10 min after extubation.
Univariate analysis revealed that patients who tended to require RAP were older, males and those with higher scores of ASA physical status (Table 1). Comorbidities among those at risk for RAP included preoperative pulmonary conditions (COPD and pneumonia/pleural effusion), heart disease (congestive heart failure and angina history), altered consciousness, ascites, chronic renal failure, and SIRS (Table 2). Patients with preoperative oxygen desaturation (Sat O2 < 95%) while breathing room air were more likely subjected to reintubation. Reintubation occurred most frequently after airway surgery and neurosurgery. However, a lengthy surgical duration might not increase the RAP incidence though an intraoperative fluid balance over 1000 ml showed significance between RAP cases and controls (odds ratio: 7.92, 95% CI: 1.51–41.61, p = 0.015) and two RAP patients had extremely positive fluid balance (+3400 ml and +3480 ml). As a matter of fact cumulative fluid balance had little influence on the likelihood of RAP (Table 3). As to the balanced anesthesia at the end of surgery, the last dose of opioid was less used in the reintubation group than in the control group. However, administration of the last dose of muscle relaxant did not significantly differ in both groups.
For multivariate logistic regression (Table 4), after adjustment for age, gender, ASA classification, preoperative oxygen desaturated breathing room air, type of surgery and administration of last dose of opioid, comorbidity with COPD (odds ratio: 7.17), pneumonia (odds ratio: 7.94), ascites (odds ratio: 13.76), and SIRS (odds ratio: 11.90) were more likely to render RAP. In the same model, airway surgery (odds ratio: 5.98) was consistent with the univariate regression outcome; however, neurosurgery lost its significance and was replaced by head and neck surgery (odds ratio: 3.43). Concerning anesthesic-related factors, administration of the last dose of opioid became irrelevant to RAP after multivariate regression. Area-under-the-receiver–operator-characteristic-curve (AUC) was 0.823 (p < 0.001), which indicated good predictive ability (Fig. 1).
Download full-size image
A logistic regression of a binary response variable (presence or absence of reintubation) on a binary covariant (COPD) with a sample size of 332 observations (of which 5.42% were in the COPD group) achieved a 95% power at a 0.05 significance level for detecting differences in reintubation between COPD (0.722) and non-COPD (0.223). This difference corresponded to an odds ratio of 9.06.
4. Discussion
In this retrospective, QA data- and chart-reviewed, case-controlled study, we aimed to recognize the risk factors for RAP. By identifying them, it might be possible to elaborate extubation planning and reduce the need for reintubation. The RAP incidence in our study was 0.06%. Though this figure was somewhat lower than those in previous reports (0.17–0.27%), we did have an ample sample size (achieving 95% power) and interpreted it in a more methodological (case-controlled) way.3, 4
The risk factors identified previously for adverse respiratory events (desaturation, hypoventilation and reintubation) include advanced age, masculine gender, higher scores of ASA status, COPD, CHF and obesity; these were ascertained from ICU or PACU data.1, 5 In our study, age, masculine gender and higher ASA classification were all eliminated as risk factors as multivariate regression proved them insignificant. The narrower range of outcome (reintubation) in our study and the relatively shorter care time, compared with those in the ICU, might exclude these factors from the precipitating factor model.
Pneumonia and COPD, both of which manifested narrowing of the small airways and increase of breathing effort would exacerbate respiratory muscle (diaphragm) fatigue.6 They were ranked third and fourth in terms of strength as precipitating factors (odds ratios: 7.94 and 7.17), respectively, in our study. Concomitantly, oxygen desaturated (<95%) breathing room air, used as an indicator of impaired cardiopulmonary capacity in our clinical practice, though proved to be insignificant after multivariate regression, had a minor odds ratio (2.99) toward RAP. These findings suggest that judicious evaluation of preoperative pulmonary conditions for general anesthesia may predict the likelihood of reintubation.
Epstein and Ciubotaru, in their study consisting of 74 ICU patients who required reintubation within 72 h after extubation, stated that CHF was the cause of extubation failure in 23% of patients.7 Cardiac-related extubation failure seems more likely when a partial support mode is used to determine readiness for extubation or when patients cannot handle the volume overload because of poor cardiovascular reserves.8, 9 In our study, CHF showed significance after univariate analysis but became insignificant (odds ratio: 1.88, 95% CI: 0.50–7.10) after multivariate regression. Intraoperative fluid balances, in our study, were homogenously equated in both the RAP and control groups (Table 3), except in two extremely positive cases in the reintubation group. CHF may, therefore, not be an independent risk factor for RAP and hypervolemia in a typically short surgical procedure, should be avoided.
SIRS, manifesting fever, tachypnea, tachycardia or leukocytosis before general anesthesia proved negligible as a strong predictor (odds ratio: 11.9, 95% CI: 2.63–53.86) for extubation failure. Bien et al., reported that among 78 SIRS patients who were mechanically-ventilated in surgical ICU after undergoing abdominal surgery, 26.9% failed to be weaned and 15.4% required reintubation.10 Incomplete recovery of respiratory regulatory mechanisms, injured by endotoxins, would explain tachypnea after extubation in SIRS patients, leading to consequent RAP.11
Ascites with severe sepsis or liver cirrhosis compromises gas exchange and potentiates RAP.1, 12 Though ascites had the highest odds ratio (13.76) over any variable in our study, its 95% CI (1.08–174.74) range was wide due to small case number.
Regarding non-surgical or non-comorbid factors, body weight had surprisingly little impact on reintubation. In this study we used body weight rather than BMI because the records of patient heights were incomplete. Furthermore, unlike Rose’s study, the body weight range in our study (58.9 ± 17.55 kg) was far less risky, as defined in that study (>120 kg in males, >100 kg in females).5
The sites of airway surgery, in our study, involved the oropharynx, hypopharynx or larynx; most surgeries were for tumor resection. The major cause of extubation failure in this group was assumed to be postoperative laryngeal edema, leading to partial airway obstruction. Hemorrhage at the surgical site as a complication of extubation might also occur. In Eikermann’s study, even at the time deemed suitable for extubation, postoperative upper airway obstruction was still observed in most patients after non-airway surgery.13 This may imply that the best way to avoid reintubation after airway surgery is to postpone extubation irrespective of full regainment of muscle power, until resolution of laryngeal edema has been achieved. Good cooperation and communication with the surgeon would be needed in such instances.
Some studies have suggested that consciousness level may influence extubation outcome, especially in patients who sustain neurosurgery.14, 15 In the present study, reintubation caused by conscious disturbance did not significantly differ between studied and control groups.
To achieve smooth emergence from anesthesia, a number of techniques has been used in our anesthesiology department, including removal of the endotracheal tube while the patient is still deeply anesthetized or administration of a last dose of opioid and/or muscle relaxants to prevent involuntary movement. However, some studies have demonstrated that residual opioid effects do influence critical postoperative respiratory events.5, 16 In our study, after multivariate adjustment for administration of a last dose of opioid, this factor seemed not to affect extubation outcome (odds ratio: 0.53, 95% CI: 0.22–1.27). Various reports concerning fast-track anesthesia have shown that a tailored opioid dose given during emergence of anesthesia does not prolong the wake-up period, but can maintain hemodynamic stability along with postoperative respiratory functions.17, 18 A limitation of our study is that it involved multimodal analgesia. Therefore, studies on dose adjustment may be required to determine the optimal dose for use during emergence from anesthesia.
Incomplete neuromuscular recovery is known to be an important contributing factor to adverse respiratory events in the PACU.19 Murphy’s study had a small number of reintubation cases (5 out of 61 cases) and only 72% sensitivity for correctly classifying patients with TOF <0.7 to the outcome of interest. Our study demonstrated that both surgical duration and administration of the last dose of NMBs have no significant differences between the RAP and control groups. Administration rates for both groups were 31.3%, which implies the concordance of our anasthesiologists’ prescriptions of NMBs. Hence, the cumulative effect of NMBs could still be reversed appropriately and has little influence on RAP. If neostigmine administration is timed at a TOF count of 4, then clinically significant postoperative residual neuromuscular blockade should be a rare event.20
In summary, an extubation strategy should be thoroughly planned before administration of general anesthesia. Through this retrospective, case-controlled study, risk factors for RAP have indeed been demonstrated. We have found that patients with preoperative COPD, pneumonia/pleural effusion, ascities, or SIRS are indicators of a higher incidence of reintubation. The higher risk of reintubation following airway surgery warrants delayed extubation and good communication with surgeons. Administration of a last dose of opioid and/or NMBs is not associated with RAP. Further studies are necessary to assess whether or not these variables can work out a scheme for the preventative measures or evaluation of outcome improvements.