1026 — Preoperative Pulmonary Risk Stratification for Noncardiopulmonary Surgery: A Systematic Review

Author List:
Smetana GW (Harvard Medical School)
Lawrence VA (VERDICT, South Texas Veterans Health Care System)
Cornell JE (VERDICT, South Texas Veterans Health Care System)

Objectives:
To systematically review evidence regarding preoperative risk for postoperative pulmonary complications (PPC) after noncardiopulmonary surgery.

Methods:
Systematic literature search,1980 – 2002, and abstraction of eligible studies with standardized forms. Inclusion criteria: primary data analysis, explicit definitions of pulmonary complications. Exclusion criteria: non-English; <25 subjects/group; physiologic outcomes only, use of administrative data to determine PPC; studies from developing countries; cardiopulmonary surgery; and pediatric or immunosuppressed patients (pts). Meta-analytic methods included: adjustment for bias in unadjusted analyses; assessment of publication bias, heterogeneity and outliers; and random effects models.

Results:
Of 15,499 citations, 120 were eligible for detailed review; 88 studies provided unadjusted data on predictors for 11,851 PPC among 173,500 pts and 32 studies reported multivariable analyses of predictors for 10, 960 PPC among 321,819 pts. Among studies with multivariable analyses, the 3 largest used patient subsets from the Veterans’ Affairs National Surgical Quality Improvement Project and accounted for 91% of pts and 85% of PPC. PPC rates varied systematically with study size; studies with <500 pts had highly variable and higher PPC rates (median rate 14%, 95% CI 8.7 - 25) than studies of >500 pts (median rate 4%, 95% CI 2.6 - 6.3). Significant risk factors varied with sample size and surgical populations. Good evidence supports the following risk factors for PPC: COPD; advanced age; poor functional status; ASA class 3-5; prolonged, abdominal, aortic, noncardiac/nonpulmonary thoracic, or emergency surgery; and albumin <3 gm/dl. Fair evidence suggests the following may be important: obstructive sleep apnea; impaired sensorium; congestive heart failure; abnormal chest exam; poor exercise capacity; cigarette use; general anesthesia (versus spinal/epidural); neck, esophageal, and neurosurgery; pancuronium (versus shorter-acting agents); significant blood loss; increased BUN or creatinine; and abnormal chest radiograph. Good evidence indicates these are not important risk factors for PPC: obesity; asthma; and hip or genitourinary/gynecologic surgery. Evidence is insufficient regarding: corticosteroid or alcohol use; arrhythmia; weight loss; diabetes; perioperative transfusion; vascular or laparoscopic surgery; transverse abdominal incisions; and preoperative spirometry.

Implications:
Incidence of PPC and important risk factors vary with study size and surgical population. Studies with >500 patients provide the most stable estimates of PPC incidence. The importance of several clinical risk factors for PPC is confirmed while other, traditionally held, factors do not appear important.

Impacts:
For clinicians, this systematic review summarizes the evidence regarding preoperative pulmonary risk stratification before noncardiopulmonary surgery. For clinical and health services researchers, it provides clear direction regarding appropriate sample size and the minimum set of covariates to be considered in designing trials of interventions to prevent PPC.