The size of the macrophage aggregates also increased with longer operations. To demonstrate this observation, we have divided the data into three categories of duration: short (<2 h; n = 13), intermediate (2-6 h; n = 17), and long (>6 h; n = 11). Average percentages of alveolar macrophages in aggregates found before and after operation (Pre and Post) are shown for each category in Table 3. On average, 8%-10% of the cells in the Pre samples in all categories are found in 2-9 cell aggregates. The Pre samples contained virtually no 10-cell or larger aggregates. For short procedures, aggregation was unchanged at the end of operation. For intermediate procedures lasting 2-6 h, the percentage of macrophages in 2-9 cell aggregates increased in the Post BAL fluid, while the number of aggregates having 10 or more cells was essentially unchanged. When procedures exceeded 6 h, the percentage of macrophages in aggregates of 10 or more cells increased significantly, as did the number of smaller aggregates Table 3. The larger aggregates ranged in size from 10 to approximately 100 macrophages, representing a significant fraction of the macrophages present. Although more than 95% of aggregates consisted of only macrophages, several aggregations contained lymphocytes with other lymphocytes or with macrophages.
(Figure 2) shows that cell viability decreased as the duration of operation and anesthesia increased. The increase in nonviable cells found in Post samples is strongly correlated with procedure length (R = 0.82, R2 = 0.67) and approaches 25% for the longest procedures. Most of the nonviable cells were aggregated macrophages. Nonviable lymphocytes were rare. Normalization of the results to 1% isoflurane had no effect on the correlation between procedure duration and loss of viability (data not shown).
Of the 41 patients, 19 had received steroid therapy before surgery and were given perioperative stress doses of hydrocortisone. These patients were distributed randomly throughout the parent sample with respect to duration of surgery. Analysis of aggregation and loss of viability as a function of the duration of surgery was repeated independently for each group and showed no significant differences between groups, indicating that steroid exposure was not a major independent influence on alveolar macrophage survival (data not shown).
Inspired gas was humidified by a low flow anesthetic technique in 18 patients, who were distributed throughout the groups. The remaining 23 patients received higher gas flows. No differences between the two subgroups could be found with respect to aggregation or viability of alveolar macrophages (data not shown).
We have demonstrated that aggregation of alveolar macrophages increases and viability decreases with prolonged anesthesia and surgery. These changes may be due to the disruptive effects of anesthesia or surgery, or both, on the normal milieu of pulmonary macrophages, or they may reflect toxic effects on the macrophages. Anesthesia-related effects can be divided into those caused by manipulation of the airway and pulmonary tissues (i.e., intubation and positive pressure mechanical ventilation) and those caused by exposure to anesthetic gases and increased oxygen tension. Effects related to surgery include fluid shifts due to inflammation and insensible losses from exposed gut, release of inflammatory mediators or gut bacteria (or both) into the bloodstream with prolonged handling of the gut, as well as changes in immune response related to the stress response to surgery. Because all of these factors varied independently in our study, it was impossible to identify which, if any, was the primary cause of the reduced macrophage viability we observed.
During anesthesia, the lung is exposed to multiple potential insults. Intubation and mechanical ventilation bypass anatomic barriers to infection. Inflation of the endotracheal cuff can inhibit the outward flow of mucus around the cuff. Constant tidal volumes provided by mechanical ventilation may not provide optimal protection from infection. Alveolar macrophages are uniquely dependent on oxidative metabolism for phagocytosis and oxygen radical formation , suggesting that they may be less effective in atelectatic areas. In anesthetized rats, mechanical ventilation with constant tidal volumes reduced the clearance of exogenously introduced Staphylococcus aureus from the lung as compared to unanesthetized controls or anesthetized animals given periodic sighs (large breaths exceeding tidal volume) . The salutary effect of sighing was attributed to the reexpansion of small atelectatic regions, which maintained the normal air-liquid-surfactant interface and reduced areas of hypoxia . Although it is possible that bacterial clearance, and by extension, pulmonary macrophage function, is depressed by the development of atelectasis during prolonged mechanical ventilation, it is unlikely that this mechanism is solely responsible for the decrease in viability we observed.
In dogs, the state of anesthesia itself impaired mucociliary clearance in a dose-dependent fashion . Both halothane and thiopental caused this effect in intubated, mechanically ventilated animals . A similar result was obtained in humans undergoing abdominal vascular surgery or lower extremity orthopedic procedures. After the relatively short orthopedic cases, radiopaque particles insufflated into the lung at the end of operation were cleared within 48 h. In the abdominal cases, which lasted approximately three times longer, a large fraction of the particles was retained even after 100 h. Because the two groups of patients differed in anesthetic regimen, the site and duration of operation, and underlying health, the exact cause of the observed differences could not be determined . Thus, although it is difficult to separate their effects, it appears that general anesthesia and endotracheal intubation undermine a major anatomic defense against pulmonary infection.
During mechanical ventilation, alveolar macrophages in our patients were exposed to volatile anesthetics, N2 O, supernormal oxygen tensions, and dry inspired gases, all of which are potentially toxic. Exposure to supernormal oxygen tension, 60% O2 for up to 24 h, was well tolerated by alveolar macrophages from guinea pigs , suggesting that oxygen toxicity was not the cause of alveolar macrophage death and aggregation in our patients.
Although the humidity of inspired gas was not measured in this study, approximately half of our patients received low flow anesthesia to provide humidification. Humidification did not appear to have a major influence on alveolar macrophage viability or aggregation, but the small number of patients in our study precludes detection of differences in aggregation and viability between the two groups.
The toxic effects of inhaled and intravenous anesthetics on alveolar macrophages are incompletely understood [7-9]. Volatile and barbiturate anesthetics appear to suppress phagocytosis by circulating and tissue macrophages in vivo . Halothane (4% in air) caused dose-dependent inhibition of protein synthesis in alveolar macrophages in vitro. Halothane did not appear to kill the cells directly, as assessed by release of cellular lactate dehydrogenase, nor did it deplete cellular adenosine triphosphate (ATP) over 30 min . At clinically relevant concentrations, halothane, isoflurane, and enflurane reversibly inhibited the microbicidal oxidative activity of alveolar macrophages [12,13]. Administration of 5% halothane or 5% enflurane in 60% O2 over 24 h decreased the concentration of ATP in alveolar macrophages exposed to the gas phase . At clinically relevant concentrations in 50:50 N2 O: O2 these drugs reduced cellular ATP after 6 h. The contribution of the halogenated drugs could not be separated from the effect of the carrier gas because 50:50 N2 O: O2 depleted cellular ATP and 60% O2 alone did not deplete intracellular ATP after 24 h . These studies (conducted in vitro with animal cells) suggest that volatile anesthetics can suppress the cytotoxic or phagocytic response of alveolar macrophages.
Thus, high isoflurane concentration during the course of anesthesia may have decreased alveolar macrophage viability and increased aggregation in our study. The studies cited above suggest that the effect of anesthetics on alveolar macrophages is reversible. In our study, duration of anesthesia and percent-hours of isoflurane were highly correlated (i.e., patients received relatively constant doses of isoflurane for the duration of their procedures), so an independent effect of the concentration of isoflurane could not be assessed. Isoflurane concentration was measured qualitatively so that the normalization of the data to isoflurane dose was not precise. A group of similar patients not exposed to isoflurane would have been required to detect an independent effect of isoflurane on alveolar macrophage aggregation and viability.
Steroid exposure is another possible contributor to the loss of alveolar macrophages we observed. Almost half of the patients in the study had received steroids in the immediate preoperative period, most for management of inflammatory bowel disease. No significant differences were found when aggregation and loss of viability as a function of the duration of surgery were compared for patients who had received steroids and those who had not. The analyses were limited in their power to detect small differences in macrophage survival owing to steroid exposure, but they were sufficient to exclude steroids as the sole cause of our observations.
The effects of surgery on alveolar macrophage function are also poorly understood. In mice, simple midline laparotomy during ether anesthesia reduced antigen presentation and interleukin-1 synthesis by peritoneal macrophages, while ether alone had no effect . Alveolar macrophages were recovered between 6 and 24 h after cecal ligation and puncture in mice (inducing sepsis); macrophages did not increase in number or decrease in viability. Over time, the macrophages showed signs of activation, including increased spreading and oxygen radical production as the mice developed sepsis .
Prolonged handling of the gut during operation in our patients may have led to the release of gut bacteria, bacterial endotoxin, inflammatory mediators such as interleukin-1 or tumor necrosis factor, or any combination of these, into the bloodstream. These agents, if deposited in the lung, may have activated alveolar macrophages , although it is unclear how activation by contaminants from the gut would then lead directly to the death of alveolar macrophages. If what we observed was the result of exposure to substances released from the gut during operation, then it is possible that the complexity of the procedure (i.e., the amount of manipulation sustained by the gut) is the factor most closely associated with the loss of alveolar macrophages, and the correlation arises between procedure duration and loss of macrophages because more complex procedures tend to be longer.
The loss of alveolar macrophage viability we observed during anesthesia and operation may have been entirely due to a single factor, but it is also possible that multiple factors contributed. For example, in a rabbit model of inflammatory lung injury, intravenous administration of rabbit C5a, a complement-derived chemotactic factor, induced neutrophil migration to the pulmonary microvasculature but not translocation into the alveolar spaces. However, when this maneuver was accompanied by anesthesia with intubation, mechanical ventilation, and operation (femoral artery cutdown), neutrophils increased significantly in the alveolar air-spaces . The aforementioned study illustrates the cumulative effects of multiple insults in the development of lung injury.
The fact that no postoperative pulmonary complications occurred during our study might at first suggest that the loss of alveolar macrophages has little impact on clinical outcome. However, prolonged operation and anesthesia are associated with increased incidence of such complications [1-3]. The overall incidence of postoperative pulmonary complications related to anesthesia, mostly atelectasis and pneumonia, varies from 3% to 76%. These figures depend on the criteria used to define pulmonary complications and on the diligence of those searching for the problem . Thus, it is likely that our patient group (especially those undergoing long procedures with the highest likelihood of pulmonary complication) was too small to permit detection of differences in the incidence of postoperative pulmonary complications.
The clinical importance of a 25% reduction in the numbers of viable alveolar macrophages or a 40% increase in aggregation with the longest cases is not clear. Some speculations are possible. Reductions of 25% in viable macrophages may not have disastrous consequences because most patients do not develop pulmonary infections postoperatively. Like most organ systems, the lung possesses considerable reserve capacity, and it is reasonable to assume that this reserve applies to the pulmonary phagocytic system as well. Reserve capacity may explain why reductions in viable macrophages are tolerated without pulmonary complications. Alternatively, the aggregated and nonviable macrophages noted in our study may represent only the most severely injured cells. Many alveolar macrophages may be injured without being killed outright by anesthesia and surgery, and the function of the injured cells may have been compromised to the point that true phagocytic reserve capacity was quite small. Such injury need not be clinically evident, since the loss of reserve capacity in most physiologic processes is usually clinically silent. One may then speculate that pulmonary complications occur when additional insults are delivered to lungs that have suffered serious but clinically undetected reductions in phagocytic capacity. A larger study with strict criteria to define pulmonary complications is needed to determine whether loss of alveolar macrophages correlates with the development of pulmonary complications.
Although our study cannot detect the cause of alveolar macrophage aggregation and death, there is a clear correlation between the duration (or perhaps the complexity) of the procedure and macrophage death. This result may help to explain why longer cases tend to have higher pulmonary morbidity and may serve as a reminder to clinicians to be especially vigilant for pulmonary complications after long cases. Because procedure length is a relatively inflexible quantity, efforts to minimize decreased viability of alveolar macrophages may be directed toward minimizing those insults that damage them in vitro. However, it is unclear what steps will prove efficacious because the mechanism underlying the loss of macrophage viability in vivo is still unknown. Hopefully, a series of interventional studies in animals will elucidate the mechanism of alveolar macrophage death during surgery and anesthesia as well as the clinical significance of this phenomenon.
We are grateful to Elisabeth Lanzl and Pauline Snider for editorial review of the manuscript.
1. Otto CW. Respiratory morbidity and mortality. Int Anesthesiol Clin 1980;18:85-106.
2. Modell JH, Moya F. Postoperative pulmonary complications. Incidence and management. Anesth Analg 1966;45:432-9.
3. Wightman JAK. A prospective survey of the incidence of postoperative pulmonary complications. Br J Surg 1968;55:85-91.
4. The BAL Cooperative Group Steering Committee. Bronchoalveolar lavage constituents in healthy individuals, idiopathic pulmonary fibrosis, and selected comparison groups. Am Rev Respir Dis 1990;141 (Suppl):S169-95.
5. Fels AOS, Cohn ZA. The alveolar macrophage. J Appl Physiol 1986;60:353-69.
6. Kelley J. Cytokines of the lung. Am Rev Respir Dis 1990;141:765-88.
7. Duncan PG, Cullen BF. Anesthesia and immunology. Anesthesiology 1976;45:522-38.
8. Graham CW. Immunological and carcinogenic side effects of anesthetics. Int Anesthesiol Clin 1980;18:173-86.
9. Walton B. Effects of anaesthesia and surgery on immune status. Br J Anaesth 1979;51:37-43.
10. Moudgil GC, Forrest JB. Comparative effects of volatile anaesthetic agents and nitrous oxide on human leukocyte chemotaxis in vitro. Can Anaesth Soc J 1984;31:631-7.
11. Hammer JA III, Rannels DE. Effects of halothane on protein synthesis and degradation in rabbit pulmonary macrophages. Am Rev Respir Dis 1981;124:50-5.
12. Welch WD. Halothane inhibits the microbicidal oxidative activity of pulmonary alveolar macrophages. Anesthesiology 1983;58:456-9.
13. Welch WD. Enflurane and isoflurane inhibit the oxidative activity of pulmonary alveolar macrophages. Respiration 1985;47:24-9.
14. Voisin C, Scherpereel PA, Aerts C, Lepot D. In vitro toxicity of halogenated anaesthetics on guineapig alveolar macrophages, surviving in gas phase. Br J Anaesth 1984;56:415-20.
15. Laplante C, Lemaire I. Interactions between alveolar macrophage subpopulations modulate their migratory function. Am J Pathol 1990;136:199-206.
16. Frederick D, Pesanti E. Intrapulmonary growth of Staphylococcus aureus in rats during induced atelectasis. Infect Immun 1987;55:2747-53.
17. Forbes AR, Horrigan RW. Mucociliary flow in the trachea during anesthesia with enflurane, ether, nitrous oxide, and morphine. Anesthesiology 1977;46:319-21.
18. Forbes AR, Gamsu G. Depression of lung mucocilliary clearance by thiopental and halothane. Anesth Analg 1979;58:387-9.
19. Gamsu G, Singer MM, Vincent HH, et al. Postoperative impairment of mucous transport in the lung. Am Rev Respir Dis 1976;114:673-9.
20. Stephan RN, Saizawa M, Conrad PJ, et al. Depressed antigen presentation function and membrane interleukin-1 activity of peritoneal macrophages after laparotomy. Surgery 1987;102:147-54.
21. Goya T, Abe M, Shimura H, Torisu M. Characteristics of alveolar macrophages in experimental septic lung. J Leukoc Biol 1992;52:236-43.
22. Said SI, Foda HD. Pharmacologic modulation of lung injury. Am Rev Respir Dis 1989;139:1553-64.
© 1995 International Anesthesia Research Society
23. Webster RO, Larsen GL, Mitchell BC, et al. Absence of inflammatory lung injury in rabbits challenged intravascularly with complement-derived chemotactic factors. Am Rev Respir Dis 1982;125:335-40.