Valerie A. Lawrence, MD; John E. Cornell, PhD; Gerald W. Smetana, MD
Disclosure: Members of the American Society of Anesthesiologists also reviewed the manuscript. Their review implies neither agreement with nor endorsement of this document.
Acknowledgments: The authors gratefully acknowledge the tremendous contribution of medical librarian Martha R. Harris, MA, for her time and expertise in searching the medical literature and managing the resulting project database. They also thank the Department of Anesthesiology, especially Christopher Jankowski, MD, of the Mayo Clinic, Rochester, Minnesota, for assistance in interpreting the anesthesiology literature.
Grant Support: By the Veterans Evidence-based Research, Dissemination, and Implementation Center (VERDICT) (Veterans Affairs Health Services Research and Development, HFP 98-002).
Potential Financial Conflicts of Interest: Stock ownership or options (other than mutual funds): G.W. Smetana (SafeMed Harvard Imaging); Other: G.W. Smetana (Novartis Pharma Schweiz).
Requests for Single Reprints: Valerie A. Lawrence, MD, Medicine/General Medicine, University of Texas Health Science Center at San Antonio, 7703 Floyd Curl Drive, Mail Code 7879, San Antonio, TX 78229-3900; e-mail, firstname.lastname@example.org.
Current Author Addresses: Drs. Lawrence and Cornell: Medicine/General Medicine, University of Texas Health Science Center at San Antonio, 7703 Floyd Curl Drive, Mail Code 7879, San Antonio, TX 78229-3900.
Dr. Smetana: Division of General Medicine and Primary Care, Beth Israel Deaconess Medical Center, 330 Brookline Avenue, Boston, MA 02215.
Lawrence VA, Cornell JE, Smetana GW. Strategies To Reduce Postoperative Pulmonary Complications after Noncardiothoracic Surgery: Systematic Review for the American College of Physicians. Ann Intern Med. 2006;144:596-608. doi: 10.7326/0003-4819-144-8-200604180-00011
Download citation file:
Published: Ann Intern Med. 2006;144(8):596-608.
Postoperative pulmonary complications are as frequent and clinically important as cardiac complications in terms of morbidity, mortality, and length of stay. However, there has been much less research and no previous systematic reviews of the evidence of interventions to prevent pulmonary complications.
To systematically review the literature on interventions to prevent postoperative pulmonary complications after noncardiothoracic surgery.
MEDLINE English-language literature search, 1 January 1980 through 30 June 2005, plus bibliographies of retrieved publications.
Randomized, controlled trials (RCTs); systematic reviews; or meta-analyses that met predefined inclusion criteria.
Using standardized forms, the authors abstracted data on study methods, quality, intervention and control groups, patient characteristics, surgery, postoperative pulmonary complications, and adverse events.
The authors qualitatively synthesized, without meta-analysis, evidence from eligible studies. Good evidence (2 systematic reviews, 5 additional RCTs) indicates that lung expansion interventions (for example, incentive spirometry, deep breathing exercises, and continuous positive airway pressure) reduce pulmonary risk. Fair evidence suggests that selective, rather than routine, use of nasogastric tubes after abdominal surgery (2 meta-analyses) and short-acting rather than long-acting intraoperative neuromuscular blocking agents (1 RCT) reduce risk. The evidence is conflicting or insufficient for preoperative smoking cessation (1 RCT), epidural anesthesia (2 meta-analyses), epidural analgesia (6 RCTs, 1 meta-analysis), and laparoscopic (vs. open) operations (1 systematic review, 1 meta-analysis, 2 additional RCTs), although laparoscopic operations reduce pain and pulmonary compromise as measured by spirometry. While malnutrition is associated with increased pulmonary risk, routine total enteral or parenteral nutrition does not reduce risk (1 meta-analysis, 3 additional RCTs). Enteral formulations designed to improve immune status (immunonutrition) may prevent postoperative pneumonia (1 meta-analysis, 1 additional RCT).
The overall quality of the literature was fair: Ten of 20 RCTs and 6 of 11 systematic reviews were good quality.
Few interventions have been shown to clearly or possibly reduce postoperative pulmonary complications.
Postoperative pulmonary complications are as common as cardiac complications for patients undergoing noncardiothoracic surgery (1-6). Further, these complications have similar mortality rates and length of stay after elective abdominal surgery or hip fracture repair (1, 2). In an accompanying systematic review (7), we identify patient, procedure, and laboratory risk factors for postoperative pulmonary complications. Our current systematic review synthesizes the evidence on preventive strategies and focuses on atelectasis, pneumonia, and respiratory failure. While we have written the review primarily for internists, this field crosses specialty disciplines.
We performed a systematic MEDLINE English-language literature search from 1 January 1980 to 30 June 2005. The search strategy and inclusion and exclusion criteria are described in the accompanying review of risk factors and in further detail in its Appendix (7). The search strategy used 1) the Medical Subject Heading (MeSH) terms preoperative care, intraoperative care, postoperative care, intraoperative complications, and postoperative complications as a focus of the article; 2) the MeSH text term perioperative complications as a text term in the title or abstract; and 3) additional MeSH and text terms for pulmonary, respiratory, or cardiopulmonary conditions, complications, or care. In addition, we performed additional focused searches for preoperative chest radiography and spirometry, laparoscopic versus open major abdominal operations, general versus spinal or epidural anesthesia, intraoperative neuromuscular blockade, postoperative pain management, and postoperative lung expansion techniques. Eligible studies were randomized, controlled trials; systematic reviews; or meta-analyses. We excluded studies with less than 25 participants per group; studies from developing countries (because of potential differences in respiratory and intensive care technology); studies that used physiologic (for example, lung volumes and flow, oximetry) rather than clinical outcome measures; studies of gastric pH manipulation; studies of complications that are unique to the surgery (for example, upper airway obstruction after uvulectomy); studies of cardiopulmonary, pediatric, or organ transplantation surgery (because of profoundly immunosuppressive drugs); and studies that used only administrative data to identify postoperative complications (for example, International Classification of Diseases, Ninth Revision, Clinical Modification [ICD-9-CM], codes) because of recent evidence that administrative data have poor validity for postoperative complications (8, 9).
We used the Quality of Reporting of Meta-analyses (QUOROM) statement for reporting meta-analyses and the U.S. Preventive Services Task Force criteria for hierarchy of research design to assess internal validity and study quality (good, fair, or poor) and to make conclusions about strength of the evidence (10, 11).
We used simple means and chi-square tests to calculate CIs and P values when they were not provided in publications. We did not perform quantitative pooling because multiple meta-analyses were beyond the scope of a broad review of multiple potential interventions. We report pooled results from previous meta-analyses when applicable.
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.
The search and inclusion criteria identified 20 randomized clinical trials and 11 systematic reviews or meta-analyses (12-42). Figure 1 in the accompanying review (7) of risk factors for postoperative pulmonary complications details the search results. Appendix Tables 1 through 7 provide detailed characteristics of the eligible randomized trials and systematic reviews.
In the only trial of preoperative smoking cessation (12), 108 older, relatively healthy men undergoing hip or knee replacement were randomly assigned to usual care or weekly meetings with a nurse for advice about smoking cessation and nicotine withdrawal plus individualized nicotine replacement for 6 to 8 weeks before surgery until 10 days after surgery. The mean age of the men was 65 years, and 95% were American Society of Anesthesiologists (ASA) physical status class I or II. Of 56 patients in the intervention group, 36 stopped smoking and 14 reduced smoking before surgery. Overall complications rates were lower in the intervention group (18% vs. 52%; P < 0.001), primarily due to fewer wound complications and urinary tract infections. The only pulmonary outcome, postoperative ventilator support, occurred in 1 patient in each group. Non–statistically significant trends favored shorter mean hospital stay (11 days vs. 13 days; P = 0.41) and fewer cardiac complications (0% vs. 10%; P = 0.08) in the intervention group.
Although the trial was of good quality, several factors limit its ability to demonstrate decreased risk for postoperative pulmonary complications. Pulmonary risk is inherently low with hip and knee replacement. Furthermore, the timing of smoking cessation seems important. A previous cohort study showed paradoxically higher postoperative pulmonary complication rates for smokers who stopped or reduced smoking within 2 months before noncardiothoracic surgery (43). Smoking cessation may increase short-term risk because of transiently increased mucus production due to improved mucociliary activity and reduced coughing due to less bronchial irritation.
Anesthetics disrupt central regulation of breathing and result in uncoordinated neural messaging. Due to resulting hypoventilation plus positional dependence, regional atelectasis occurs shortly after induction. It persists postoperatively and is compounded by ongoing disruption of respiratory muscles, limited respiratory excursion due to pain, and disruption of neurally mediated diaphragmatic functions after manipulation of abdominal viscera (43).
One good-quality trial found no difference in rates of postoperative pulmonary complications between intermediate-acting (atracurium, vecuronium) and long-acting (pancuronium) neuromuscular blocking agents among 691 patients undergoing elective abdominal, gynecologic, or orthopedic surgery (13). However, the incidence of residual neuromuscular block was higher among patients receiving pancuronium (26% vs. 5%; P < 0.001). Patients with residual blockade after pancuronium were 3 times more likely to develop postoperative pulmonary complications than those without residual block (17% vs. 5%; P < 0.02). In contrast, among patients receiving intermediate-acting agents, postoperative pulmonary complication rates did not differ between those with (4%) and without (5%) prolonged blockade. Therefore, pancuronium may directly lead to higher rates of prolonged neuromuscular blockade and indirectly to increased pulmonary risk compared with shorter-acting agents.
Neuraxial blockade (either spinal or epidural anesthesia) blocks a constellation of stress responses to surgery (neuroendocrine, cytokine, and pain threshold) and may improve recovery and prevent complications (44). Postoperative epidural analgesia may reduce respiratory muscle dysfunction and pain-related hypoventilation. The epidural approach involves either a single injection or an infusion and can be used for both intraoperative anesthesia and postoperative analgesia. Spinal anesthesia has a faster onset (5 to 10 minutes vs. 15 to 20 minutes), produces denser sensory and motor block, and is technically easier than epidural anesthesia. However, spinal anesthesia is administered only as a single injection because of practical constraints of indwelling intrathecal catheters. The possible benefit of neuraxial blockade has generated studies of general versus neuraxial blockade anesthesia, followed by trials comparing epidural analgesia to other modes of analgesic delivery (for example, oral, intramuscular, intravenous, patient-controlled analgesia) and, more recently, trials of combined epidural intraoperative anesthesia and epidural postoperative analgesia.
A recent good-quality meta-analysis combined 141 trials (n = 9559) comparing general anesthesia and neuraxial blockade in patients undergoing a variety of operations (32). The authors compared patients receiving neuraxial blockade (with or without concomitant general anesthesia) with those receiving only general anesthesia. Neuraxial blockade reduced overall mortality (2% vs. 3%; odds ratio, 0.70 [95% CI, 0.54 to 0.90]), pneumonia (3% vs. 5%; odds ratio, 0.61 [CI, 0.48 to 0.76]), and respiratory failure (0.5% vs. 0.8%; odds ratio, 0.41 [CI, 0.23 to 0.73]). In a subgroup analysis of trials of neuraxial blockade alone versus general anesthesia alone, results were similar (odds ratio, 0.63 [CI, 0.46 to 0.87] for pneumonia; odds ratio, 0.37 [CI, 0.11 to 1.21] for respiratory failure).
Potential sources of bias in the meta-analysis include 1) clinically heterogeneous studies; 2) unusually high mortality rates in several trials; 3) older literature (82% of included studies were published before 1990); 4) small studies (81% of included studies had ≤ 50 patients); and 5) statistically significant benefit only for orthopedic surgery in subgroup analyses (45-47).
A smaller good-quality systematic review identified 15 randomized or quasi-randomized trials of 2162 patients undergoing hip fracture repair (33). Postoperative pneumonia rates were almost identical: 5.1% of 529 patients having neuraxial blockade and 5.5% of 567 patients having general anesthesia (odds ratio, 0.92 [CI, 0.53 to 1.59]). Twelve of the 15 trials were also included in the larger meta-analysis (32), which included 44 trials of orthopedic surgery (n = 3617). Why the results for pneumonia differ between the 2 meta-analyses is not clear, but important variables may include type or duration of procedure (hip fracture repair is inherently low risk for postoperative pulmonary complications), intraoperative fluids, and postoperative pain and rehabilitative management.
Table 1 summarizes 6 eligible trials that compared various regimens of intraoperative anesthesia and postoperative analgesia. In a double-blind, good-quality efficacy trial of patients undergoing abdominal aortic surgery (14), investigators randomly assigned patients to 1 of 4 combined anesthetic and analgesic protocols. The trial standardized the entire episode of anesthesia and pain management to optimize efficacy in all 4 groups. The primary outcome measure was length of stay; secondary outcomes included postoperative pulmonary complications. Sample sizes were small (37, 38, 39, and 46 participants, respectively), and median length of stay (7 to 8 days for all groups) or postoperative pulmonary complication rates did not differ among the groups.
Strengths of the trial include the double-blind design and equally highly standardized protocols for both anesthesia and analgesia (48, 49). Unequally optimized regimens can introduce bias that systematically favors one type of intervention. Potential weaknesses, regarding prevention of postoperative pulmonary complication, include length of stay as the primary outcome measure and small sample size (50-52).
In a subsequent fair-quality effectiveness trial (15), 915 patients undergoing major abdominal surgery were randomly assigned to general anesthesia and 1) postoperative intravenous opioid or 2) intraoperative epidural local anesthetic plus postoperative epidural analgesia. Overall infections did not differ, and the authors did not report results for pneumonia. Statistically significantly less pain and respiratory failure occurred with epidural anesthesia, but only 225 of 447 patients in the epidural group completed the protocol.
In a subgroup analysis of high-risk patients (respiratory insufficiency by arterial blood gas analysis, severe obstructive or restrictive lung disease, acute respiratory failure within the past 2 years, or morbid obesity), rates of pneumonia (11% vs. 12%; P = 0.71) or mechanical ventilation for more than 24 hours (8% for both groups) did not differ. Respiratory failure (ventilation > 24 hours, reintubation, Pao2 ≥ 50 mm Hg, or Paco2 ≥ 50 mm Hg on room air) occurred significantly less often with epidural (45% vs. 29%; odds ratio, 0.5 [CI, 0.29 to 0.88]) (53).
In a large fair-quality trial (16), 1021 patients undergoing abdominal surgery were randomly assigned to general anesthesia and 1) postoperative systemic opioid or 2) intraoperative epidural anesthesia plus postoperative epidural morphine. Mortality rates did not differ, and non–statistically significant trends favored the epidural approach for pneumonia and respiratory failure.
In 1 fair-quality trial (17), 217 patients undergoing elective abdominal aortic surgery were randomly assigned to general anesthesia alone or general anesthesia plus intraoperative epidural opioid. Both groups received the same postoperative pain management. A non–statistically significant trend favored intraoperative epidural for overall postoperative pulmonary complications. Rates of individual types of postoperative pulmonary complications did not differ, but statistical power was low.
In a smaller poor-quality trial (18), 70 elderly patients undergoing major abdominal surgery were randomly assigned to general anesthesia and 1) postoperative patient-controlled morphine or 2) intraoperative neuraxial blockade plus postoperative patient-controlled epidural analgesia. Rates of atelectasis or major pulmonary complications did not differ. In an additional, small poor-quality trial (19), 75 patients undergoing elective cholecystectomy were randomly assigned to general anesthesia and 1) postoperative intramuscular morphine, 2) continuous intravenous morphine, or 3) intraoperative epidural local anesthesia plus postoperative epidural local anesthetic for 12 hours. Postoperative pneumonia occurred less often with epidural (4%) than with either intramuscular morphine (24%; P = 0.05) or intravenous morphine (20%; P = 0.11).
A fair-quality meta-analysis examined the evidence for 3 epidural techniques: intercostal nerve block, systemic opioids, and wound infiltration with local anesthetic (34). The number of trials that compared any 2 strategies varied from 2 to 11. Compared with systemic opioids, epidural opioids reduced atelectasis (relative risk, 0.53 [CI, 0.33 to 0.85]; 11 studies) but not pneumonia (relative risk, 0.53 [CI, 0.18 to 1.53]; 5 studies). Compared with systemic opioids, epidural local anesthetic reduced “pulmonary infection” (relative risk, 0.36 [CI, 0.21 to 0.65]; 5 studies) but not atelectasis (relative risk, 0.74 [CI, 0.50 to 1.11]; 4 studies). However, the authors pooled studies of both on-demand (that is, as requested) and patient-controlled intravenous analgesia, which could bias the results of the meta-analysis in favor of epidural analgesia.
In contrast, a good-quality meta-analysis identified 32 trials (n = 1029) of patient-controlled opioid analgesia versus the same drug given intravenously, intramuscularly, or subcutaneously (35). Opioid consumption, pain scores, length of stay, or adverse effects did not differ. In the 2 trials reporting postoperative pulmonary complications, fewer complications occurred in the patient-controlled analgesia group (odds ratio, 0.93 [CI, 0.86 to 0.99]; number needed to treat, 15 [CI, 8 to 98]).
Evidence from 1 good-quality trial suggests that shorter-acting neuromuscular blocking drugs may prevent postoperative pulmonary complications. Intraoperative neuraxial blockade, either alone or in combination with general anesthesia, may prevent postoperative pulmonary complications, but the evidence is conflicting. Several meta-analyses (which included small unblinded studies) suggest that epidural anesthesia may reduce pulmonary risk, but recent large randomized trials do not confirm benefit. Randomized trials of combined intraoperative and postoperative anesthetic or analgesic regimens do not clearly indicate that a combined epidural approach prevents postoperative pulmonary complications. Two meta-analyses of postoperative analgesic regimens suggest that part of the variability may be due to on-demand analgesia (intravenous, intramuscular, or subcutaneous) versus patient-controlled analgesia (intravenous or epidural). Postoperative epidural and patient-controlled intravenous analgesia both seem superior to on-demand delivery of opioids in preventing postoperative pulmonary complications. Epidural analgesia may further reduce postoperative pulmonary complications. More good-quality efficacy trials with standardized optimal regimens for all groups and sufficient size to examine pulmonary complication rates are needed (14). The risk for epidural bleeding due to postoperative epidural catheters in patients receiving heparin (especially low-molecular-weight heparin) makes timing of catheter placement important and may influence decisions about modalities for pain control and thromboembolism prophylaxis (54-56).
Our search identified many trials comparing laparoscopic and open procedures, but few reported postoperative pulmonary complication rates. Those that did focused on cholecystectomy and colorectal surgery. Downs and colleagues (36) performed a good-quality systematic review through March 1995 of open and laparoscopic cholecystectomy. They identified 18 trials (n = 1645) that were published with sufficient detail to judge methodologic quality. Twelve trials had at least 40 patients per study group, and the largest study had 150 participants per group. Since the authors did not quantitatively pool data because of clinical heterogeneity and methodologic problems, we examined the studies individually. None met the criteria for inclusion in our review (<25 participants per group [8 studies] or no clinical postoperative pulmonary complications reported [10 studies]).
Among the 4 highest-quality trials reporting spirometric outcomes, unblinded outcome assessment found statistically significantly greater compromise in FVC and FEV1 at 24 hours and 48 hours postoperatively with open cholecystectomy. In 1 study that followed patients until pulmonary function recovered to within 10% to 15% of preoperative levels, pulmonary function recovered 4 to 10 days earlier with laparoscopic cholecystectomy. Only 1 very small (n = 40) blinded trial assessed whether reduced pulmonary dysfunction translated into clinically important differences in postoperative pulmonary complication rates. On postoperative chest radiography, atelectasis occurred significantly less often with laparoscopic operations (40% vs. 90%; P = 0.001).
We identified 1 subsequent, poor-quality trial of laparoscopic versus open cholecystectomy (20). Among 82 patients, the frequency and severity of atelectasis, assessed by radiologists who were blinded to type of procedure, were significantly less among patients randomly assigned to laparoscopic cholecystectomy (frequency, 29% vs. 63% [P < 0.05]; severity, chi-square for trend P < 0.05). However, the analysis was not intention-to-treat: Patients who converted from laparoscopic to open operations were excluded from analysis.
Table 2 summarizes the results of a good-quality meta-analysis of laparoscopic versus open resection of colorectal cancer (37). Overall mortality did not differ. Risk was consistently less with laparoscopic operations for several complications but was not statistically significantly less with the more conservative statistical approach of random-effects modeling. The reduced risk for overall complications was primarily due to fewer wound complications, especially wound infection. Regarding pulmonary complications, a non–statistically significant trend favored laparoscopic resection. Three studies that evaluated postoperative pulmonary complications reported that respiratory recovery (defined by spirometry) was statistically significantly faster. Nine studies reporting data confirmed shorter length of hospital stay (mean, 21% shorter [range, 14% to 38%]) after laparoscopic operations (37).
We identified 1 additional good-quality trial of laparoscopic versus open colorectal resection in 384 patients; 269 were in an earlier publication that was included in the previous meta-analysis discussed (21). Again, a nonsignificant trend favored lower rates of pneumonia after laparoscopic operations (3 of 190 [1.8%] patients and 6 of 194 [3.5%] patients; P = 0.52) (21).
Two large retrospective cohort studies highlight the problems with using designs other than a randomized trial and ICD-9-CM codes to identify postoperative pulmonary complications to compare open and laparoscopic operations (57, 58). Atelectasis (the only postoperative pulmonary complication studied) occurred significantly less often with laparoscopic (n = 19 662) compared with open (n = 23 771) cholecystectomy (4% vs. 2%; P < 0.001) (57). Overall postoperative pulmonary complications were significantly less frequent after laparoscopic (n = 709) compared with open (n = 17 735) sigmoid resection (2.5% vs. 6%; P < 0.001) (58). The results of these 2 studies may be unreliable because of lack of systematic prospective surveillance. Clinicians may have been biased to order more postoperative chest radiographs (and therefore identify more atelectasis) after open procedures. Furthermore, in a recent study comparing discharge ICD-9-CM codes and systematic chart review for complications (8), specificity of ICD-9 codes was high but sensitivity was low: 35% (CI, 30% to 41%) for all complications and 32% (CI, 19% to 45%) for all infectious complications.
In summary, while supported by spirometric, postoperative pain, and length of stay data, whether laparoscopic procedures reduce the risk for clinically important pulmonary complications is not clear. The literature did not systematically assess or report pulmonary complications, and most studies did not have sufficient statistical power to detect differences in postoperative pulmonary complication rates.
Selective use of nasogastric decompression, or tubes, refers to use only for postoperative nausea or vomiting, inability to tolerate oral intake, or symptomatic abdominal distension. Routine decompression (that is, standard use until bowel function returns) has been thought to speed bowel recovery and decrease risk for aspiration. We identified 2 meta-analyses of studies of routine versus selective nasogastric decompression (38, 39, 59). One or both meta-analyses included all the trials that we identified.
The first meta-analysis was of good methodologic quality up to quantitative analyses, which pooled data from randomized trials, nonrandomized trials, “uncontrolled” trials, and case–control studies (38). For the overall group of 26 studies (which comprised 15 RCTs, 3 nonrandomized trials, and 8 case–control studies [n = 3964]), patients receiving selective decompression had significantly lower rates of pneumonia (odds ratio, 0.49; P < 0.001) and atelectasis (odds ratio, 0.46; P = 0.001) and shorter time to oral intake (3.5 days vs. 4.6 days; P = 0.04). Aspiration rates (odds ratio, 0.61; P = 0.88) did not differ. Selective decompression did not statistically significantly increase nausea, vomiting, or abdominal distension. For 20 higher-quality studies (15 trials and 5 case–control studies), patients receiving selective decompression also had lower rates of pneumonia (odds ratio, 0.59; P = 0.01) and atelectasis (odds ratio, 0.52; P = 0.002), a trend toward shorter time to oral intake (3.5 days vs. 4.5 days; P = 0.07), no difference in aspiration rates, and significantly higher rates of vomiting (odds ratio, 1.45; P = 0.005) and abdominal distension (odds ratio, 1.34; P = 0.02)
The recent meta-analysis was of good quality, and it identified 28 eligible trials (n = 4194) of routine versus selective nasogastric decompression after open laparotomy (39, 59). It included 15 of the 17 RCTs in the previous meta-analysis. Of 19 trials (n = 2892) reporting postoperative pulmonary complication, 18 included only elective operations (39, 59). Patients who were randomly assigned to selective decompression had fewer postoperative pulmonary complications, and the benefit approached statistical significance (relative benefit increase, 1.35 [CI, 0.98 to 1.86]; P = 0.07; calculated relative risk reduction, 0.74 [CI, 0.54 to 1.02]). Selective decompression also resulted in earlier bowel recovery (8 studies; n = 862; 0.46 day [CI, 0.28 day to 0.64 day]; P < 0.001).
In summary, the evidence suggests that selective nasogastric decompression (that is, for specific indications rather than routine decompression) improves return of bowel function and may reduce the incidence of postoperative pulmonary complications after elective abdominal operations.
Decreased lung volumes and atelectasis due to surgery-related shallow breathing, bed rest, diaphragmatic dysfunction, pain, and impaired mucociliary clearance may be the first events in a cascade leading to postoperative pulmonary complication. However, the evidence on prophylactic lung expansion is limited by variable techniques, inconsistent definitions of postoperative pulmonary complications, and poor-quality trials. Techniques include incentive spirometry, deep breathing exercises, chest physical therapy (which may include variable combinations of deep breathing, cough, postural drainage, percussion and vibration, suctioning, and ambulation), intermittent positive-pressure breathing, and continuous positive airway pressure. Table 3 summarizes the evidence.
The first of 2 poor-quality systematic reviews focused on upper abdominal surgery and identified 14 randomized trials (sample size, 17 to 200 participants) (40). Across all lung expansion modalities, a trend favored fewer postoperative pulmonary complications compared with controls (odds ratio, 0.85 [CI, 0.59 to 1.2)], but heterogeneity was statistically significant. In 2 studies, postoperative pulmonary complications occurred less often in patients receiving incentive spirometry compared with control (odds ratio, 0.44 [CI, 0.18 to 0.99]). In 4 studies, postoperative pulmonary complications occurred less often in patients who were randomly assigned to deep breathing exercises (odds ratio, 0.43 [CI, 0.27 to 0.63]), but heterogeneity was again statistically significant. Across other studies, no single modality was clearly superior. The second systematic review identified 4 randomized trials of patients undergoing abdominal surgery (41). The authors did not report raw or pooled postoperative pulmonary complication rates. In the only trial in our systematic review that met our sample size criteria, incentive spirometry, deep breathing exercises, and intermittent positive-pressure breathing equally prevented postoperative pulmonary complications compared with no intervention.
We identified 5 other trials in patients undergoing major abdominal surgery. The first 4 trials were of poor quality. Two trials compared chest physiotherapy with no prophylaxis (22, 23). In the first study, patients who were randomly assigned to chest expansion, maximum inspiration exercises, cough, and early ambulation had fewer abnormalities on postoperative chest radiography and a non–statistically significant trend toward fewer postoperative pulmonary complications (22). In the second trial, patients who were randomly assigned to cough and deep breathing exercises had significantly lower rates of pneumonia (0.6% vs. 7%; P < 0.05) (23).
The third trial compared “conventional chest physiotherapy” with incentive spirometry in 876 patients undergoing abdominal surgery and found no difference in rates of overall postoperative pulmonary complication, abnormal postoperative chest radiography, or Pao2 less than 60 mm Hg (24). The fourth poor-quality study compared 1) incentive spirometry and deep breathing exercises in 155 low-risk patients and 2) incentive spirometry versus incentive spirometry plus chest physiotherapy in 301 high-risk patients (ASA class > I or age ≥ 60 years) undergoing abdominal surgery (25). Among high- or low-risk patients, postoperative pulmonary complication rates did not differ with any intervention.
In the fifth and only good-quality trial (26), 204 patients undergoing intra-abdominal vascular surgery were randomly assigned to supplemental oxygen to maintain arterial oxygen saturation greater than 95% or to nasal continuous positive airway pressure for 12 hours after surgery. Severe hypoxemia (Pao2 < 70 mm Hg at fraction of inspired oxygen ≥ 0.70%) occurred statistically significantly less often (5% vs. 16%; P = 0.01), and non–statistically significant trends favored less pneumonia and reintubation with continuous positive airway pressure.
For patients having abdominal surgery, the evidence suggests that any type of lung expansion intervention is better than no prophylaxis. No modality seems superior, and combined modalities do not seem to provide additional risk reduction. Incentive spirometry may be the least labor-intensive, while continuous positive airway pressure may be particularly beneficial for patients who cannot participate in incentive spirometry or deep breathing exercises.
A fair-quality meta-analysis of 14 randomized or quasi-randomized trials of total parenteral nutrition (TPN) versus no TPN through August 1986 concluded that routine TPN in major surgery was not beneficial, except perhaps for severe malnutrition or for extended periods (10 days to 14 days) of inadequate enteral nutrition (60). The meta-analysis did not report results specific to pulmonary complications and was therefore ineligible for our review.
Subsequently, a good-quality multisite trial randomly assigned 395 patients undergoing laparotomy or noncardiac thoracotomy to perioperative TPN or no TPN (27). Overall rates of major complications (26% vs. 25%) and 90-day mortality (13% vs. 11%) were similar between the groups. Total parenteral nutrition was associated with non–statistically significant trends toward higher rates of pneumonia and empyema but significantly lower rates of noninfectious complications (5.3% vs. 42.9%; P = 0.03).
We identified 1 poor-quality meta-analysis (230 patients) and 2 additional, good-quality trials of TPN versus total enteral nutrition (TEN) (28, 29, 42). In the meta-analysis, infections were twice as common among patients receiving TPN (35% vs. 16%; P = 0.01) even after excluding patients with catheter sepsis from analysis (29% vs. 16%; P = 0.03) (42). There was a nonstatistically significant trend toward more frequent pneumonia in patients receiving TPN. In a trial of 241 patients, there was a non–statistically significant trend toward more postoperative pulmonary complications with TEN (7% vs. 13%; P = 0.12) (28). In a second, larger trial, 317 malnourished patients (>10% weight loss in previous 6 months) were randomly assigned to TPN or TEN (29). Rates of overall complications and infectious complications were statistically significantly lower with TEN, but rates of pneumonia (9 of 159 patients vs. 14 of 158 patients; P = 0.39) or the combined outcome of pneumonia and respiratory failure (13 of 159 patients vs. 20 of 158 patients; P = 0.27) did not differ.
Immunonutrition refers to enteral feedings with additional ingredients (variable combinations of arginine, Ω-3 fatty acids, or ribonucleic acids) to enhance the immune system and to possibly prevent infection. A good-quality meta-analysis of trials found that for patients undergoing elective surgery, enteral immunonutrition had no mortality benefit but resulted in significantly fewer overall infectious complications (relative risk, 0.53 [CI, 0.42 to 0.68]) (61). The authors did not report results for respiratory infections; therefore, the study was not eligible for our review.
In a subsequent good-quality trial of enteral immunonutrition (30), 305 patients undergoing elective resection of gastrointestinal cancer were randomly assigned to an enteral solution enriched with arginine, Ω-3 fatty acids, and ribonucleic acids preoperatively (5 days before surgery; n = 102) or perioperatively (5 days before surgery plus jejunal tube feeding begun within 12 hours of surgery and continued until oral intake was resumed; n = 101) or to a control group (n = 102) of postoperative intravenous glucose and electrolytes. Overall infection rates were significantly lower with immunonutrition (14% and 16% vs. 30%; P = 0.006 and 0.02, respectively), but rates of pneumonia (3 of 102 patients and 6 of 101 patients vs. 8 of 102 patients) did not differ (30).
In summary, while hypoalbuminemia and malnutrition increase postoperative complications, including pneumonia, routine TPN has no benefit over either TEN or no hyperalimentation, except perhaps for patients with severe malnutrition or for long periods of inadequate oral nutrition. More research is needed on enteral formulations that may enhance immune status. Prompt resumption of oral intake after surgery is important because atrophy of intestinal villi occurs quickly with inadequate intake and increases the risk for bacterial translocation across gut mucosa and subsequent sepsis (62).
After observational data suggested higher rates of respiratory failure and pneumonia in patients receiving right- heart catheterization for noncardiac surgery (63), Sandham and colleagues (31) performed a good-quality RCT. High-risk patients (n = 1994; age ≥ 60 years; ASA class III or IV) undergoing urgent or elective major noncardiac operations were randomly assigned to usual care or treatment guided by perioperative pulmonary artery catheter. Pulmonary artery catheterization did not reduce the primary outcome of in-hospital all-cause mortality (7.8% vs. 7.7%) or the rate of postoperative pneumonia, a secondary outcome (6.7% vs. 7.3%; P = 0.70).
Recent evidence has shown that postoperative pulmonary and cardiac complications are equally prevalent and clinically important in morbidity, mortality, and length of stay. However, compared with postoperative cardiac complications, much less research on prevention of pulmonary complications has been published. Table 4 summarizes the strength of available evidence, based on our systematic review, on interventions to reduce the risk for postoperative pulmonary complications.
Good evidence suggests that lung expansion therapy (for example, incentive spirometry, deep breathing exercises, and continuous positive airway pressure) reduces postoperative pulmonary risk after abdominal surgery. Well-designed trials are needed to clarify the magnitude of benefit and the comparative effectiveness of different modalities.
Fair evidence suggests that selective nasogastric tube decompression after abdominal surgery reduces risk. Fair evidence also suggests that short-acting neuromuscular blocking agents result in lower rates of residual neuromuscular blockade and may reduce risk for pulmonary complications.
Laparoscopic, compared with open, abdominal operations reduce pain and pulmonary compromise as measured by spirometry and oxygenation. However, the evidence is insufficient to determine whether laparoscopic operations prevent clinically important pulmonary complications. Given the benefits of laparoscopic procedures in pain control and length of stay, future trials to adequately assess clinical pulmonary outcomes are unlikely.
Evidence is insufficient to judge the potential benefit of preoperative smoking cessation in reducing risk. Risk may actually increase transiently after stopping or reducing smoking within 2 months of surgery due to increased secretions. We need trials of preoperative smoking cessation before higher-risk surgeries that adequately address optimal duration of cessation.
Evidence on intraoperative epidural anesthesia and postoperative epidural analgesia is insufficient. More good-quality efficacy trials of sufficient size (in which all groups receive equally standardized and optimized regimens) are needed to accurately examine complication rates.
Although malnutrition is associated with increased risk for postoperative pulmonary complications, good evidence indicates that routine total parenteral or enteral hyperalimentation nutrition does not reduce risk, except perhaps for patients with severe malnutrition or for those undergoing long periods with inadequate oral intake. Enteral formulations that are tailored to enhance immune status and reduce postoperative infections may be promising.
Evidence from 1 well-done randomized trial indicates that invasive perioperative monitoring with pulmonary artery catheterization does not reduce risk of pulmonary complications.
A limitation of our review is the overall quality of the literature. Only 10 of 20 RCTs and 6 of 11 systematic reviews or meta-analyses were of good quality.
Future studies of interventions to reduce postoperative pulmonary complications should be randomized trials that are designed to overcome methodologic problems in earlier literature. Cohort studies using secondary analyses of administrative databases should use measures for pulmonary complications that are proven valid and reliable by direct clinical assessment or medical chart audit. Studies should be large enough to adjust for known potential risk factors and confounding variables (as synthesized in the accompanying systematic review of preoperative risk stratification ) and should go beyond surrogate or intermediate physiologic or spirometric outcomes to detect clinically meaningful differences in clinical pulmonary complications. This is important for 2 reasons: to base patient care on clinically meaningful evidence and to determine when it is appropriate to substitute intermediate outcomes to shorten study timelines and reduce study cost. Researchers should define postoperative pulmonary complications a priori according to explicit criteria and, whenever possible, use outcome assessment that is masked, or blinded, to intervention assignment.
The In the Clinic® slide sets are owned and copyrighted by the American College of Physicians (ACP). All text, graphics, trademarks, and other intellectual property incorporated into the slide sets remain the sole and exclusive property of the ACP. The slide sets may be used only by the person who downloads or purchases them and only for the purpose of presenting them during not-for-profit educational activities. Users may incorporate the entire slide set or selected individual slides into their own teaching presentations but may not alter the content of the slides in any way or remove the ACP copyright notice. Users may make print copies for use as hand-outs for the audience the user is personally addressing but may not otherwise reproduce or distribute the slides by any means or media, including but not limited to sending them as e-mail attachments, posting them on Internet or Intranet sites, publishing them in meeting proceedings, or making them available for sale or distribution in any unauthorized form, without the express written permission of the ACP. Unauthorized use of the In the Clinic slide sets will constitute copyright infringement.
Department of Anesthesiology and Reanimation. Hospital General Universitario de Elche (Spain)
April 22, 2006
Strategies to Reduce Postoperative Pulmonary Complications after Noncardiothoracic Surgery
Dr. Lawrence and colleagues (1) write that "Evidence on intraoperative epidural anesthesia and postoperative epidural analgesia is insufficient". However, they affirm that "Good evidence suggests that lung expansion therapy reduces postoperative pulmonary risk after abdominal surgery".
The use of postoperative lung expansion therapy (incentive spirometry, deep breathing exercises, and continuous positive airway pressure) may be very difficult because of pain. Epidural analgesia has been shown to reduce postoperative pain, improve gastrointestinal function and reduce the incidence of serious complications (2,3).
The authors state that the risk for epidural bleeding due to postoperative epidural catheters may influence decisions about modalities for pain control and thromboembolism prophylaxis. They don't consider that epidural analgesia permits early mobilization programmes, avoiding immobility.
In our hospital anesthesiologists provide critical care for surgical patients; the use of epidural analgesia provides them safe analgesia, avoiding opioids which are known to produce ileus and ease development of hyperalgesia (4).
If the authors haven't found evidence that epidural analgesia reduces postoperative pulmonary complications, they must admit that epidural analgesia does reduce pain, allowing lung expansion therapies and early mobilization programmes which have shown to reduce length of hospital stay (5).
1. Lawrence VA, Cornell JE, Smetana GW. Strategies To Reduce Postoperative Pulmonary Complications after Noncardiothoracic Surgery: Systematic Review for the American College of Physicians. Ann Intern Med. 2006;144:596-608. [PMID:16618957]
2. Carli F, Mayo N, Klubien K, Schricker T, Trudel J, Belliveau P. Epidural analgesia enhances functional exercise capacity and health- related quality of life after colonic surgery: results of a randomized trial. Anesthesiology 2002;97:540"“549. [PMID:12218518]
3. Rodgers A, Walker N, Schug S,McKee A, KehletH, van Zundert A et al. Reduction of postoperative mortality and morbidity with epidural or spinal anaesthesia: results from overview of randomised trials. BMJ 2000;321:1493. [PMID: 11118174]
4. Angst MS, Clark JD. Opioid-induced hyperalgesia: a qualitative systematic review. Anesthesiology. 2006;104:570-87. [PMID:16508405]
5. Anderson AD, McNaught CE, MacFie J, Tring I, Barker P, Mitchell CJ. Randomized clinical trial of multimodal optimization and standard perioperative surgical care. Br J Surg. 2003;90:1497-504. [PMID:14648727]
Results provided by:
Copyright © 2016 American College of Physicians. All Rights Reserved.
Print ISSN: 0003-4819 | Online ISSN: 1539-3704
Conditions of Use
This PDF is available to Subscribers Only