| JOURNAL HOME | CME HOME | THIS MONTH | PAST ISSUES | ETOC | COLLECTIONS |
| AUTHORS | REVIEWERS | EDITORIAL BOARD | FEEDBACK | RSS | HELP |
| ||||||||||||||
| ||||||||||||||
|
|
|
|
||||||||||||
|
||||||||||||||
*Department of Anesthesiology, The Mount Sinai Medical Center, New York, New York; Departments of
Anesthesiology and
Biostatistics, Duke University Medical Center, Durham, North Carolina; and
Department of Anaesthesia, University College London Hospitals, London, England
Address correspondence and reprint requests to Elliott Bennett-Guerrero, MD, Department of Anesthesiology, P&S Box 46, College of Physicians and Surgeons of Columbia University, 630 West 168th St., New York, NY 10032.
 |
Abstract |
|---|
 Top Abstract IntroductionMethodsResultsDiscussionReferences |
|---|
Implications: Little is known about the overall incidence and pattern of complications in patients with prolonged hospitalization after routine, elective surgery. We prospectively assessed these complications using a novel postoperative morbidity survey. The postoperative morbidity survey can be used in future clinical outcome trials, as well as in routine hospital-based quality assurance.
|
Introduction |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Previous efforts at characterizing postoperative morbidity have been limited to very high-risk surgical procedures (e.g., cardiac surgery) (2,3), intensive care unit patients (4), complications specific to a particular surgical procedure or study objective (5–12), or severe life-threatening complications (3,13). No previous study was designed to prospectively evaluate the overall pattern and incidence of postoperative complications in patients with prolonged hospitalization.
We prospectively observed a diverse group of surgical patients and systematically assessed them for morbidity using predefined criteria. As a secondary objective, we tested the hypothesis that intraoperative indices of tissue hypoperfusion (e.g., gastric pHi, analysis of arterial blood gas) are good predictors of postoperative morbidity.
|
Methods |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Inclusion criteria were patients undergoing the following elective surgical procedures under general anesthesia: orthopedic (revision hip arthroplasty, fusion/instrumentation of multiple lumbar or thoracic vertebrae), general (any laparotomy expected to exceed 2 h duration, including partial hepatectomy, pancreatic surgery, reoperative colon surgery), urological (radical prostatectomy, radical cystectomy, radical nephrectomy), vascular (abdominal aortic aneurysm repair), and gynecological (cancer debulking procedure, abdominal hysterectomy). These procedures were selected for several reasons: 1) they are routinely performed surgeries, 2) they represent a diverse group of procedure types, and 3) a retrospective analysis at Duke University Medical Center found these procedures to be associated with prolonged hospitalization.
Exclusion criteria included age <18 yr, emergency surgery, regional anesthesia, any religious/ethical prohibition from receiving blood products, or contraindication to the placement of a naso- or orogastric tube.
Patients were cared for by clinicians who were blinded to study measurements and were not part of the investigative team. At the conclusion of surgery, clinically available cardiovascular variables and indices of tissue perfusion were recorded. Gastric tonometry was performed in a manner consistent with the manufacturer’s guidelines and package insert. Briefly, after the induction of anesthesia and insertion of a gastric tonometer (TRIP NGS Catheter; Tonometrics, Tewksbury, MA), saline was introduced into the tonometer’s balloon and allowed to equilibrate for at least 90 min. The saline was removed at the conclusion of surgery, and PCO2 was measured using a blood gas analyzer. pHi was calculated according to the gastric tonometer’s manufacturer’s guidelines.
As a secondary study objective, we tested the hypothesis that intraoperative indices of tissue hypoperfusion (e.g., gastric pHi, arterial base excess) are good predictors of postoperative morbidity. Several indices of regional and global tissue hypoperfusion were collected at the end of surgery and compared with routinely measured and clinically available hemodynamic variables (e.g., heart rate) and indices of oxygen delivery (e.g., hematocrit, oxygen saturation). These variables were viewed as potential predictors of postoperative morbidity and were not used as an end point of how patients had fared during surgery.
Patients were prospectively observed from the day of surgery until either discharge from the hospital or death. Postoperative length of stay was defined as the number of days from the day of operation (Day 0) until hospital discharge or death. Postoperative complication was the primary outcome of this study and was defined as either in-hospital death or postoperative length of stay >7 days. Seven days was chosen as a reasonable cutoff because a pilot retrospective analysis of these same surgical procedures at Duke University Medical Center revealed that the median postoperative duration of hospitalization for each of the individual procedures was 
7 days. Furthermore, this same analysis revealed that most patients whose postoperative hospitalization exceeded 7 days had evidence of persistent organ dysfunction.
Hospitalized patients were evaluated on Postoperative Days 5, 8, and 15 using the predefined criteria listed in Table 1. Each assessment was conducted by speaking with the patient, examining the patient, reviewing the medical record, and/or consulting with the patient’s caregivers. The criteria used in this postoperative morbidity survey (POMS) were selected to identify both severe complications (e.g., pulmonary failure/acute respiratory distress syndrome) and more subtle morbidity that could delay discharge from the hospital (e.g., moderate pulmonary dysfunction necessitating supplemental oxygen therapy). These criteria were not intended to represent the "gold standard" for assessing a given complication. Our goal was to design a survey that was not cumbersome, not complicated, and could be used routinely to screen thousands of patients for multiple complication types. In addition, other possible reasons for delay in hospital discharge were recorded as a free text entry. We included a free text entry space on the data sheet to avoid a bias toward recording only predefined complications.
|
= 0.05. Given an estimated incidence of 25% for the primary end point from pilot data, a sample size of at least 400 patients was estimated to generate enough events (100) to allow for relatively narrow (approximately 10%) 95% CIs for the specific complication types observed. Relationships among continuous variables were tested by using linear regression, and associations among categorical variables were tested using the contingency table 
2 test. This study was not designed to rigorously assess all of the possible perioperative predictors of postoperative complications; therefore, a full multivariate analysis was not deemed appropriate. Nevertheless, as a secondary study objective, we tested for univariate associations between clinically available predictors and the primary outcome, i.e., postoperative complications, using a series of univariate (unadjusted) logistic regression models. |
Results |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
|
|
|
Discussion |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Several previous studies have focused on specific "anesthetic-related" complications occurring immediately after surgery; for example, hypoxemia in the recovery room (14) and postoperative nausea and vomiting (15). Other studies have focused on selected high risk surgeries (e.g., cardiac surgery) (2,3), or focused on a specific type of complication, e.g., infection (10–12). Studies of complications have also been retrospective, have limited their scope to life-threatening severe complications, or have not specifically examined complications in patients with prolonged hospitalization (9,13,16,17). In one example, "severe perioperative adverse outcomes" were studied in 17,201 patients undergoing minor or intermediate-risk elective surgery (13). In that study, several complications that are known to occur after surgery, such as renal dysfunction and wound infection, were not mentioned. In a separate retrospective study (n = 8,126), only two types of noncardiac surgery were included, and the definitions for most adverse events were limited to organ failure (16).
We observed a more frequent incidence of morbidity than has been reported in previous studies. This difference arose because we did not limit our scope to severe life-threatening complications such as organ failure. Instead, we included less severe forms of morbidity that may not result in death but can nevertheless increase hospital length of stay and healthcare costs.
Our study was not designed or powered to determine the exact incidence of complications. For example, it was not designed to prove that pulmonary complications (25%, 95% CI 17–33) are more common than renal complications (21%, 95% CI 14–28). Our study was instead designed to elucidate the overall incidence and pattern of postoperative morbidity. The high incidence of organ dysfunction observed in the present study contradicts anecdotal claims that prolonged hospitalization is due to errors in surgical technique or to the underlying surgical pathology. Several of the complications observed warrant discussion.
Gastrointestinal dysfunction was the most common type of postoperative complication in all surgery types. This finding is consistent with those of previous studies that suggest that the gastrointestinal tract is extremely sensitive to hypovolemia, catecholamines, and other insults (18–20). Patients undergoing procedures involving a laparotomy had approximately twice the incidence of postoperative gastrointestinal dysfunction, which confirms the notion that mechanical trauma is an important determinant of this complication. However, our findings demonstrate that mechanical trauma cannot be the only cause of gastrointestinal dysfunction because 32% of patients with prolonged hospitalization exhibited this complication after extraabdominal procedures (e.g., revision hip arthroplasty).
Also of note was our finding that cardiac morbidity was less common than other types of morbidity, such as pulmonary, renal, and infectious complications. It was previously suggested that cardiac complications are the most relevant type of morbidity after noncardiac surgery (1,17). Studies reporting a high incidence of cardiac complications included predominantly high-risk patients, e.g., Veterans Administration hospitalized patients undergoing vascular surgery (17). In contrast, our study enrolled a diverse group of patients undergoing varying surgical procedures, which is a likely reason for the differences observed. Because studies reporting a high incidence of cardiac complications also did not rigorously assess patients for noncardiac morbidity, it is possible that patients in those studies exhibited a high degree of noncardiac morbidity that went unreported.
Several plausible theories could account for our finding of diffuse, multisystem complications unrelated to the type of surgery. It is clear from the critical care literature that a massive injury or insult often results in multiple organ dysfunction syndrome that is often not directly related to the site of the initial injury (21–23). Organ dysfunction in these patients is thought to be due to a systemic process, perhaps excessive systemic inflammation induced by endotoxin, cytokines, and/or other mediators (20–24). Our study’s findings support the theory that a large proportion of postoperative morbidity is similar in pattern, and perhaps pathophysiology, to complications observed in intensive care unit patients (18), albeit of a less severe form.
As a secondary study objective, we tested the hypothesis that intraoperative indices of tissue hypoperfusion (e.g., gastric pHi, arterial base excess) are good predictors of postoperative morbidity. We chose to exclude most of the preoperative "predictors" previously studied for the following reasons. It is already known that patients who have a major preoperative comorbidity (e.g., left ventricular ejection fraction of 10%) are more likely to have postoperative complications. In general, very little can be done to alter these preoperative variables (e.g., age) so as to alter outcome.
Several indices of regional and global tissue hypoperfusion were collected at the end of surgery and compared with routinely measured and clinically available hemodynamic variables (e.g., heart rate) and indices of oxygen delivery (e.g., hematocrit, oxygen saturation). This study was not designed to rigorously assess all of the possible perioperative predictors of postoperative complications; therefore, a full multivariate analysis was not deemed appropriate. Nevertheless, as a secondary study objective, we tested for univariate associations on clinically available predictors. The strongest predictors of complications were characteristics reflective of a more traumatic operation (e.g., greater blood loss and longer duration of surgery) and indices of inadequate resuscitation and tissue hypoperfusion (e.g., acute global metabolic acidosis). This finding is consistent with those of numerous studies that have demonstrated an association between organ hypoperfusion and organ dysfunction (4,25–28). In fact, several randomized clinical trials in surgical patients have shown reductions in postoperative complications by using prophylactic strategies aimed at optimizing tissue perfusion in the immediate perioperative period (29–32). Although not conclusive, results from these trials suggest that some complications observed in surgical patients may be preventable by using prophylactic strategies that optimize tissue perfusion.
In our study, patients who had been discharged home were not observed by our study team. This limitation in no way invalidates our findings with respect to the pattern and high incidence of complications in hospitalized patients. The complications we observed are the likely cause for the prolonged length of hospitalization in these patients for two reasons. First, the complications were defined by the need for some specific acute medical therapy or intervention. For example, pulmonary complication was defined as the need for supplemental oxygen. It is unlikely that a physician would discharge a patient home on supplemental oxygen if the patient did not require home oxygen therapy preoperatively. Second, we included a free text entry in the trial data sheet to record other potential reasons for delay of discharge. Fewer than 2% of patients had a reason for delay of discharge not covered by our POMS.
Our results can be used in several ways. Our nine-point POMS can be easily used by healthcare providers and payers (e.g., clinicians, hospitals, health maintenance organizations) to characterize postoperative morbidity in their respective settings. The high incidence and cost of these complications have significant medical and economic consequences, given the millions of patients undergoing moderate-risk surgery in the United States annually (1). Our findings and the use of our POMS may be useful in future studies involving surgical patients. In fact, we believe that surgical patients represent an ideal group to show effectiveness for interventions that may also be useful in critically ill patients. An intervention that is not effective when administered prophylactically to surgical patients is unlikely to be effective when administered to critically ill patients who usually present to the intensive care unit many hours or days after the inciting injury.
Our study demonstrates that, despite a low mortality rate, complications involving multiple organ systems are common after routine, moderate-risk, elective surgery. Many of these complications are not directly related to the type/site of surgery, which supports the theory that a systemic inflammatory response may be responsible for much of this morbidity.
|
References |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  K. M. Harboe, K. Anthonsen, and L. Bardram Validation of data and indicators in the Danish Cholecystectomy Database Int. J. Qual. Health Care, June 1, 2009; 21(3): 160 - 168. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 E. Ashby, M. P. W. Grocott, and F. S. Haddad Outcome measures for orthopaedic interventions on the hip J Bone Joint Surg Br, May 1, 2008; 90-B(5): 545 - 549. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. G. Hahn Fluid Therapy Might Be More Difficult Than You Think Anesth. Analg., August 1, 2007; 105(2): 304 - 305. [Full Text] [PDF]
|
|
|
|
 |
|
 H. G. Wakeling, M. R. McFall, C. S. Jenkins, W. G. A. Woods, W. F. A. Miles, G. R. Barclay, and S. C. Fleming Intraoperative oesophageal Doppler guided fluid management shortens postoperative hospital stay after major bowel surgery Br. J. Anaesth., November 1, 2005; 95(5): 634 - 642. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. J. Martin, E. FitzSullivan, A. Salim, T. V. Berne, and S. Towfigh Use of Serum Bicarbonate Measurement in Place of Arterial Base Deficit in the Surgical Intensive Care Unit Arch Surg, August 1, 2005; 140(8): 745 - 751. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. P. W. Grocott, M. G. Mythen, and T. J. Gan Perioperative Fluid Management and Clinical Outcomes in Adults Anesth. Analg., April 1, 2005; 100(4): 1093 - 1106. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. G. Mythen Postoperative Gastrointestinal Tract Dysfunction Anesth. Analg., January 1, 2005; 100(1): 196 - 204. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
G. Lebuffe, B. Vallet, J. Takala, G. Hartstein, M. Lamy, M. Mythen, J. Bakker, D. Bennett, O. Boyd, and A. Webb A European, Multicenter, Observational Study to Assess the Value of Gastric-to-End Tidal PCO2 Difference in Predicting Postoperative Complications Anesth. Analg., July 1, 2004; 99(1): 166 - 172. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. D. Stone, R. J. T. Wilson, J. Cross, and B. T. Williams Effect of adding dopexamine to intraoperative volume expansion in patients undergoing major elective abdominal surgery{dagger} Br. J. Anaesth., November 1, 2003; 91(5): 619 - 624. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
E. W. Moretti, K. M. Robertson, H. El-Moalem, and T. J. Gan Intraoperative Colloid Administration Reduces Postoperative Nausea and Vomiting and Improves Postoperative Outcomes Compared with Crystalloid Administration Anesth. Analg., February 1, 2003; 96(2): 611 - 617. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. L. Reich, E. Bennett-Guerrero, C. A. Bodian, S. Hossain, W. Winfree, and M. Krol Intraoperative Tachycardia and Hypertension Are Independently Associated with Adverse Outcome in Noncardiac Surgery of Long Duration Anesth. Analg., August 1, 2002; 95(2): 273 - 277. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 N. W. Knudsen, M. W. Sebastian, and D. A. Lubarsky Cost Containment in Vascular Surgery Seminars in Cardiothoracic and Vascular Anesthesia, November 1, 2000; 4(4): 256 - 264. [Abstract] [PDF]
|
|
|
|
|
|
A. Nikolic, K. Lampl, C. Klasen, C. Weinstabl, and C. G. Krenn The Use of a Tracheostomy Tube for Enteral Stomal Control Anesth. Analg., July 1, 2000; 91(1): 170 - 171. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
E. Bennett-Guerrero, G. R. Barclay, M. E. Youssef, S. Hossain, F. Vela-Cantos, L. A. Andres, and I. R. Poxton Exposure to Bacteroides fragilis Endotoxin During Cardiac Surgery Anesth. Analg., April 1, 2000; 90(4): 819 - 823. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
