Preoperative Pulmonary Risk Stratification for Noncardiothoracic Surgery: Systematic Review for the American College of Physicians

By G. Smetana, V. Lawrence, J. Cornell. Annals of Internal Medicine, 2006

987 citations recorded by the source corpus.

Postoperative pulmonary complications contribute importantly to the risk for surgery and anesthesia. The most important and morbid postoperative pulmonary complications are atelectasis, pneumonia, respiratory failure, and exacerbation of underlying chronic lung disease. Since the publication of the first cardiac risk index in 1977 (1), clinicians have been aware of the importance of, and the risk factors for, cardiac complications. Clinicians who care for patients in the perioperative period may be surprised to learn that postoperative pulmonary complications are equally prevalent and contribute similarly to morbidity, mortality, and length of stay. For example, in a large retrospective cohort study of 8930 patients undergoing hip fracture repair, 1737 (19%) patients had postoperative medical complications (2). Serious pulmonary complications occurred in 229 (2.6%) patients and serious cardiac complications occurred in 178 (2.0%) patients. Similarly, in a study of 2964 patients undergoing elective noncardiac surgery, postoperative pulmonary and cardiac complications occurred in 53 patients and 64 patients, respectively (3). Rates of postoperative cardiac and pulmonary complications are similar in other large cohort studies of patients undergoing noncardiac surgery (4-6). Pulmonary complications may also be more likely than cardiac complications to predict long-term mortality after surgery. For example, among postoperative complications in a recent study of patients older than 70 years of age who were undergoing noncardiac surgery, only pulmonary and renal complications predicted long-term mortality (7). In another report of patients undergoing esophagectomy for cancer, postoperative pneumonia was second only to tumor stage in predicting long-term survival after surgery and predicted long-term mortality to a greater degree than postoperative cardiac, renal, or hepatic complications (8). Office and hospital consultation for patients preparing for surgery is an important activity for internists. While guidelines and consensus statements for perioperative cardiac evaluation have been published (9, 10), no similar guideline is available to assist in perioperative pulmonary evaluation. The quality and number of studies that estimate perioperative pulmonary risk have increased in the past 2 decades, and this is no longer a neglected area of inquiry. We prepared this 2-part systematic review 1) to guide clinicians on clinical and laboratory predictors of perioperative pulmonary risk before noncardiothoracic surgery and 2) to evaluate the efficacy of strategies to reduce the risk for postoperative pulmonary complications (11). Risk factors for postoperative venous thromboembolism differ substantially from those for postoperative pulmonary complications, and they are not the subject of our review. Methods Literature Search and Selection Criteria The Appendix contains a detailed description of our methodology. We performed a MEDLINE search to identify relevant publications from 1 January 1980 through 30 June 2005. We used the following Medical Subject Heading (MeSH) terms and specified that they be the article's primary focus: intraoperative complications, postoperative complications, preoperative care, intraoperative care, and postoperative care, plus the text term perioperative complications in the title or abstract. We identified additional MeSH and text terms by a review of the MEDLINE indexing for the retrieved articles. These included terms for pulmonary, respiratory, or cardiopulmonary diseases, conditions, or complications and terms for oxygenation and chest roentgenography. We performed additional searches specific to preoperative chest radiography and preoperative spirometry. We identified additional references by reviewing bibliographies of retrieved studies. We included only English-language publications and excluded publication types without primary data (that is, letters, editorials, case reports, conference proceedings, and narrative reviews). We excluded 1) studies with fewer than 25 participants per study group; 2) studies that used only administrative data (for example, International Classification of Diseases, Ninth Revision, Clinical Modification [ICD-9-CM] codes) or lacked explicit criteria or definitions for pulmonary complications; 3) studies from developing countries (because of potential differences in respiratory and intensive care technology); 4) studies of ambulatory surgery; 5) studies of physiologic (for example, lung volumes and flow, oximetry) rather than clinical outcomes; 6) studies of gastric pH manipulation; 7) studies of complications unique to a particular type of surgery (for example, upper airway obstruction after uvulectomy); 8) studies of cardiopulmonary or pediatric surgery; and 9) studies of organ transplantation surgery (because of profoundly immunosuppressive drugs). Of 16959 citations identified by the search, 1223 citations were duplicates and 14793 citations were not relevant by title and abstract review (Figure). Of the remaining 943 potentially relevant citations, we excluded 626 citations after review of the full publication and abstracted 145 citations in detail. Figure 1. Flow chart for article selection process. Assessing Study Quality We used the U.S. Preventive Services Task Force (USPSTF) criteria for assigning hierarchy of research design, grading a study's internal validity as our basis for assessing study quality, and assigning summary strength of recommendations for each risk factor and laboratory test (12). Statistical Analysis Our literature search yielded primarily unadjusted estimates for most laboratory factors of interest. Limited multivariable, adjusted studies were available for serum albumin level less than 30 g/L and elevated blood urea nitrogen level. However, rather than attempt to compute potentially biased summary estimates, we provided narrative descriptions of the pattern of results for these potential risk factors. The eligible multivariable risk factor studies varied considerably in the number and type of competing risks and confounders included in the analyses. Extensive use of prescreening methods and variable selection algorithms often limited reporting to the subset of risk factors that were determined to be statistically significant in a given sample. The result is a subtle form of publication bias, which we verified by examination of the funnel plots and trim-and-fill estimates for each risk factor. We extracted odds ratios from each study, along with their respective SEs, 95% confidence limits, or both. We used the I 2 statistic (13) and the Cochran Q statistic (14) to assess study heterogeneity. We also recomputed pooled estimates with and without studies that produced extreme results. An I 2 statistic of 50% or more indicates substantial heterogeneity among study estimates. We used the DerSimonianLaird method to compute random-effects estimates when the set of studies was heterogeneous (15). In cases where 3 or more studies contributed estimates for a risk factor, we used the trim-and-fill method to adjust pooled estimates of a risk factor's effect on postoperative pulmonary complications for publication bias (16). Trim-and-fill estimates check the sensitivity of pooled estimates to potential publication bias (17). We used meta-analysis procedures available in Stata software, version 8 (Stata Corp., College Station, Texas), to conduct these analyses (18). Role of the Funding Source The Veterans Evidence-based Research, Dissemination, and Implementation Center (VERDICT) (Veterans Affairs Health Services Research and Development, HFP 98-002) provided the research librarian and administrative support for the study. The funding source had no role in the design, conduct, or reporting of the study or in the decision to submit the manuscript for publication. Results Eighty-three publications provided univariate data on clinical predictors of postoperative pulmonary complications. Appendix Table 1 summarizes the characteristics of these studies (2, 3, 19-99). Seventy-three (88.0%) publications were cohort studies; 3 (3.6%) were randomized, controlled trials; 2 (2.4%) were casecontrol studies; and the remaining 5 (6.0%) were case-series studies. Slightly less than half (45.8%) of the cohort studies used a prospective design. Ten studies were of good quality, 18 studies were of fair quality, and 55 studies were of poor quality. Eligible studies included 11851 postoperative pulmonary complication events among 173500 patients. Appendix Table 1. Study Characteristics of Univariate Studies of Clinical Risk Factors Twenty-seven studies reporting multivariable analyses (10960 postoperative pulmonary complication events among 324648 patients) met our inclusion criteria (Appendix Table 2) (100-126). These studies form the principal basis of our review. Most studies (96%) were prospective cohort studies, and only 1 report was a casecontrol study. The 3 largest studies (118, 120, 123) used subsets of patients from the Veterans Affairs National Surgical Quality Improvement Project (NSQIP) (127). These 3 studies accounted for 89.8% of all patients included in the multivariable studies and 82.3% of the observed postoperative pulmonary complications. The crude postoperative pulmonary complication rate among the cohort studies was 3.4%. Appendix Table 2. Study Characteristics of Multivariate Studies of Clinical Predictors The studies were heterogeneous with respect to study objectives, study samples, and criteria for defining a postoperative pulmonary complication. Seventeen of the 27 (63.0%) studies aimed to identify potential risk factors for postoperative pulmonary complications. The objective in 3 studies was to develop a risk index for postoperative pulmonary complications (113, 118, 120). The remaining studies focused on high-risk subgroups, such as patients undergoing aortic surgery (104, 125), patients with smoking histories (114), elderly patients

Subfields: Cardiology and Cardiovascular Medicine, Pulmonary and Respiratory Medicine, Surgery

View Paper: https://www.semanticscholar.org/paper/6901561b864b7fc820510755ffcea4866a162a93

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Surgery

  1. V. Lawrence
  2. G. Smetana

Citations in the corpus