- Email: [email protected]
Infection in postoperative patients
Infection in Postoperative Patients
H. HARLAN STONE, M.D. Baltimore,
Maryland
The majority of surgical infections are due to multiple bacterial pathogens, usually represented by mixtures of both aerobic and anaerobic species. Contamination from endogenous sources accounts for the majority of these infections. The most virulent of all such sepsis appears to arise from a symbiosis between aerobic gram-negative rods and various anaerobes. Antibiotics have proved efficacy in both the treatment as well as the prevention of surgical infection. The choice of antimicrobial agent(s) should be based upon the drug’s spectrum of activity against known or anticipated pathogens, the biologic half-Me of the agent, which serves as a guide to the frequency of administration, and the drug’s safety. The third-generation cephalosporins have been shown to be especially useful because of their broad spectrum of activity, prolonged halflife, and limited toxicity. Sepsis that persists or is uncontrolled despite antibiotic administration often leads to failure of multiple organ systems. Only energetic surgical measures offer any real chance for patient survival when such a stage has been reached. INFECTION IN POSTOPERATIVEPATIENTS
From the Department of Surgery, University of Maryland School of Medicine, Baltimore, Maryland. Requests for reprints should be addressed to Dr. H. Harlan Stone, Department of Surgery, University of Maryland School of Medicine, 22 South Greene Street, Baltimore, Maryland 21201.
Although infection is the most common postoperative complication, it is not uniformly confined to the surgical incision or the body cavity that has been violated. Indeed, serious and life-threatening sepsis often results when there are major alterations in the normal physiology of one or more organ systems. Important factors in this regard are incisional pain, stasis in hollow visceral conduits responsible for secretion or excretion, and colonization by potential pathogens of otherwise sterile tissue planes or by new species able to supplant otherwise normal flora. Accordingly, before one can specifically incriminate the surgical incision or body cavity beneath, these other areas should be surveyed as well. Diagnosis of Postoperative Sepsis. During the postoperative period, fever usually heralds the onset of infection. However, if antibiotics have been continued after surgery, the temperature may be normal or of very low grade, i.e., seldom exceeding 37.5%. Nevertheless, leukocytosis will almost always be present, as will a significant shift to the left as manifested by many immature granulocytes. The four most common sites for the development of postoperative infection are the lungs, the urinary tract, veins that have been used for fluid therapy, and the wound itself. With respect to the latter, both the surgical incision and the cavity entered must be considered. The majority of pulmonary infections begin as atelectasis, with progression to pneumonia. Attention to proper pulmonary toilet is preventive as well as therapeutic. The diagnosis is suggested by changes on physical
July
28,1989
The American
Journal
of Medicine
Volume
81 (suppl
1A)
39
SYMPOSIUM
ON PROBLEM
INFECTIONS-STONE
examination in addition to production of purulent sputum, although chest radiography is exceedingly important in confirming the diagnosis. Nevertheless, in patients who have been receiving mechanical ventilation for long periods of time, changes of infection are indistinguishable from those of the “acute respiratory distress syndrome.” Accordingly, in this setting, purulent sputum is the most important sign. Such a sputum specimen must be obtained from deep within the tracheobronchial tree, smeared for immediate examination, and subsequently cultured to determine the pathogenic organism present. Bronchoscopy may be required to obtain the desired specimen and thus to make the diagnosis. Urinary tract infections usually occur in patients with urinary tract obstruction. These develop when stasis occurs in the upper tracts because of dehydration or when there is mechanical block to the bladder outlet because of prostatism or stricture. The diagnosis is dependent upon detecting bacteria in the urine once a brisk urine flow has been obtained. A Gram-stain preparation will give an idea as to the identity of the etiologic bacteria, but only a culture with colony count and sensitivity testing will give the final answer. Before a dilute urine has been achieved, white blood cell casts, thick pyuria, and heavy bacterial concentrations on a spun specimen can be taken as important signs that urinary tract infection is present. Bacteremia arising from an infected intravenous infusion site is relatively common. It must be kept in mind, however, that only approximately one half of these cases will manifest local phlebitis, so there may be seeding of the blood without obvious signs. Because of the latter observation, the intravenous sites should always be suspect, and therefore, bottles, tubing, and cannulas must all be changed. Material from the tip of the removed intravenous cannula should always be cultured. Obvious purulence at the area of needle or catheter entry through the skin is diagnostic, but only subsequent culture results will confirm the diagnosis when overt inflammation is absent. The most common site for infection is the surgical incision. Sepsis confined to the wound is due either to Streptococcus or Clostridium species if infection manifests between the first and second postoperative day. Infections appearing on the third to fourth postoperative day are generally due to Staphylococcus species. Mixed bacterial floras involving gram-negative bacilli and various anaerobes cause infections that are noted between the fourth and the sixth postoperative day. Mere gram-negative bacilli infection alone does not generally appear until after the seventh day. The discharge of pus or any purulent fluid from the wound, with or without a halo of cellulitis, is diagnostic. With respect to the cavity involved in the operation, infection may become diffuse and present as a fulminating form of sepsis, may loculate into one or more defined abscesses within that cavity, or may spontaneously erode
40
July
28, 1966
The American
Journal
of Medicine
Volume
through the surgical incision and discharge to the outside as a so-called subfascial abscess. Antibiotics generally lead to the confinement of the process to one or more loculi. These specific abscesses are identified by rectal examination if in the true pelvis, by mass and/or tenderness along the peritoneal gutters on abdominal examination, or by percussion tenderness along the costal margin if located in the subphrenic area. In cases in which anaerobes are involved in the infection, an overpenetrated roentgenogram of the abdomen will reveal multiple bubbles in areas where bowel should not be located. However, in situations in which confusion may arise because of the presence of intraluminal intestinal gas, or if anaerobes are not involved in the abscess, only computerized axial tomography can differentiate and identify the problem. Other modalities (i.e., sonography and galium scan, among others) are not nearly as reliable and give poorer definition. Recent clinical investigations have significantly altered our basic understanding of wound sepsis in postoperative patients. For example, more thorough processing of bacteriologic specimens has revealed that most infections are polymicrobial, usually with involvement of many different anaerobic as well as aerobic species [I]. A veritable explosion in the number and types of antibiotics available to treat such infections has demonstrated that the pharmacology and safety of the agent are as important as the antimicrobial spectrum of activity [2]. Antibiotic administration, however, must be appropriate for the indication to be treated, that is for cure or for prevention of infection. And finally, if the patient dies as a result of infection, it is usually a result of failure in function of various critical organ systems [3]. Bacteriology of Surgical Sepsis. A dictum that was once emphasized was that most infections arising on the surgical ward were due to a single pathogen and that aerobes were almost always responsible. However, detailed analyses of more expeditiously processed culture specimens has recently shown that surgical sepsis is usually polymicrobial [l]. The few exceptions are noted in those infections developing within the surgical incision of patients undergoing elective clean-category procedures. In these latter cases, the usual pathogen is an aerobic grampositive coccus, such as Staphylococcus aureus or Staphylococcus epidermidis [4]. Contamination has occurred primarily through an open incision during an operative procedure on the skeleton or cardiovascular system. Etiologic pathogens have almost uniformly been found to reside within the immediate environment of the incision and can be shown to be the same aerobic species as are present on the patient’s own skin or are components of some exogenous pool, such as those found on dirty instruments or in the nasopharynx of a member of the surgical team [4]. Nevertheless, in the majority of postoperative infec-
61 (suppl
1A)
SYMPOSIUM
28, 1989
INFECTIONS-STONE
death [3]. In other words, anaerobic participation correlates with morbidity, more specifically with the incidence and severity of local postoperative infection and the recurrence of such infection. The presence of aerobic bacteria, on the other hand, especially gram-negative bacilli circulating in the blood, appears to cause almost all of the deaths due to septicemia [2]. The unique synergy between aerobic and anaerobic species seems to be based upon the ability of aerobic pathogens to extract almost all of the available oxygen from the micro-environment [l]. Anaerobes can then survive, becoming significant virulence problems themselves through the elaboration of various proteolytic exotoxins. The more efficient the aerobes are at consuming oxygen radicals, the more conducive the local environment becomes for anaerobe growth. Likewise, if more potent enzymes are secreted by the anaerobes, then there is a greater destruction of host tissue and host resistance factors [7]. The incidence of bacteremia correspondingly escalates. Antibiotic administration directed against only the aerobic component can reduce both the incidence of postoperative infection and the mortality rate due to such complicating sepsis [2]. This is presumed to be accomplished by eliminating, at least in part, the aerobic support to this synergistic relationship, thereby significantly limiting anaerobic growth as a result of the persistently high oxygen tension in adjacent tissues [l]. Aerobic pathogens that have invaded the circulation are at the same time eradicated from the bloodstream. Antimicrobial activity against the anaerobes alone, however, has been shown to only decrease the postoperative incidence of sepsis. Mortality rates are not altered [2,8]. Antibiotic Therapy. The initial choice of antibiotic must always be based upon a drug’s spectrum of activity against proved or at least suspected pathogens. In cases of polymicrobial sepsis, either a single broad-spectrum agent or an antibiotic combination is indicated [2]. Both regimens appear to have equal efficacy [2]. However, persistence or recurrence of sepsis despite apparently adequate wound care and appropriate antibiotic selection suggests one or more of the following possibilities: There is active infection in some other area, such as an unrecognized pneumonia, urinary tract infection, or bacterial phlebitis, among others; There has been incomplete drainage of the original focus of infection or sepsis has recurred at that same site; The infecting bacteria are resistant to the antibiotic being administered; or Some opportunistic pathogen, such as Candida albicans, has evolved. In most cases of postoperative sepsis, a broad spectrum of antimicrobial activity should be provided by the antibiotic regimen. Ideally, it should include antimicrobial
tions, sepsis is caused by multiple species of bacteria that normally colonize one of the patient’s own organ systems [1,4]. The alimentary tract is the most common source of such endogenous pathogens. Although gram-negative rods predominate as the aerobic flora, anaerobes are involved as major pathogens in at least 80 percent of these cases [1,5]. The exact population of bacteria in a given polymicrobial sepsis is determined primarily by the level of the gastrointestinal tract entered, which serves as the source of contamination [l]. The stomach contains varying mixtures of aerobic and anaerobic species that are quite similar to those that can normally be found in the mouth. Bile and duodenal contents are generally sterile, although they may harbor aerobic gram-negative bacilli, enterococci, Clostridium species, and Bacteroides fragilis. Although the majority of bacteria found in the unobstructed small bowel are aerobic gram-negative bacilli, the very distal ileum ordinarily contains a large number of anaerobes, especially if the ileocecal valve is incompetent. In the colon, anaerobes predominate in a polymicrobial mix that contains enterococci a:- well as a large number of various aerobic gram-negative bacilli [l]. Anaerobes apparently cannot migrate across healthy, oxygenated tissue planes. Only aerobic species appear capable of such migration. Therefore, to reach the wound or peritoneal cavity, anaerobes must pass through a grossly perforated bowel or across a segment of severely ischemic or frankly gangrenous intestine [l]. Otherwise, contamination and subsequent infection will be due solely to aerobic bacteria. Similar bacteriologic results are obtained from the urinary and female genital tracts. With acute infection, cultures yield primarily a single species of aerobic pathogens. However, after repeated or chronic infection, polymicrobial sepsis usually evolves, especially in the female genital tract. Indeed, chronic pelvic infections involving the fallopian tubes and/or ovaries generally contain a mixed bacterial flora comprising anaerobes as well as aerobes [5]. Bacterial Synergy. Polymicrobial contamination routinely leads to a significantly greater incidence of postoperative sepsis than does wound inoculation by only single species. Whenever anaerobes become a part of that contaminant, the risk and severity of subsequent infection is even greater [1,5-q. Identical results have been noted with respect to the recurrence of sepsis [1,5]. Aerobes seem to play a dual role, for anaerobes alone can seldom establish an infection in otherwise healthy tissues [l]. Only in the case of clostridial sepsis does it appear that death can be caused solely by an anaerobe [B]. It is the aerobic gram-negative bacteria that are responsible for the circulation of endotoxin and other vasoactive substances in the blood. These in turn cause failure of one or more organ systems and thereby account for eventual
July
ON PROBLEM
l
l
l
l
The
American
Journal
of Mediclne
Volume
81 (suppl
1A)
41
SYMPOSIUM
ON PROBLEM
INFECTIONS-STONE
coverage against S. aureus, the majority of aerobic gramnegative bacilli, and a significant number of anaerobes [2]. Selectivity against Pseudomonas aeruginosa is of importance only for hospital-acquired polymicrobial sepsis and for patients with deficiencies in their host defense mechanisms due to disease or medications. The anaerobic antibiotic spectrum need not be complete. Coverage against B. fragilis is also not an absolute necessity, since results appear to be just as good with as without such activity [l]. Because of their greater tolerance to atmospheric oxygen, Bacteroides species are much easier to culture. Thus, they serve as excellent “indicators” that anaerobes are present. Their greater virulence in comparison to less readily cultured anaerobes has yet to be proved. Equally as crucial as microbial susceptibility to the administered antibiotic is a dosing schedule that insures drug delivery to the site of infection at appropriate intervals and ih concentrations necessary for bacterial kill. First- and second-generation cephalosporins have good activity against aerobic gram-positive cocci, but much of this has been lost with the molecular manipulations that produced the third-generation cephalosporins [2]. The minimal inhibitory concentration required for aerobic gram-positive cocci is generally quite low, in the range of 1 or 2 pg/ml. For the aerobic gram-negative bacilli, bactericidal concentrations are somewhat higher, whereas for anaerobes an even greater concentratibn is generally required for agents with demonstrable activity [2,9]. Accordingiy, for therapy to be reliably adequate, larger doses for the treatment of aerobic gram-negative bacilli infections are required over those used for aerobic gram-positive coccal infections, and even greater doses are needed if anaerobes are present. With the newer cephalosporins, there is selectively greater antimicrobial activity against gram-negative bacilli, as well as against anaerobic organisms [2,9]. When Pseudomonas infection is suspected, an antipseudomonal penicillin (or other beta-lactam) plus an aminoglycoside should be selected. If treatment with multiple antibiotics has been chosen, then the need to modify the dose of each individual agent according to the type of infection seldom becomes a major consideration. The safety of antimicrobial therapy must also be taken into account [2]. This is not confined solely to the allergic reactions that can occur with penicillins and cephalosporins. Drug toxicity must always be kept in mind as a potential complication. Certain cephalosporins, especially in large doses, impair prothrombin production and a coagulopathy can develop [2,10]. Use of lower doses and concomitant administration of vitamin K has generally prevented this problem. On the other hand, aminoglycosides are potentially nephrotoxic [2]. Use of certain diuretics and present or prior hypovolemia significantly increase the likelihood of this complication. Overall, the incidence of
42
July
28,1988
The Amerlcen
Journal
of Medicine
Volume
aminoglycoside-induced renal failure is approximately 5 percent for surgical patients. It is a complication that significantly worsens both immediate as well as final outcomes, and, accordingly, the use of aminogiycosides should be avoided in these high-risk cases whenever possible [2]. The frequency of antibiotjc administration should be based upon the antimicrobial susceptibility of the known or anticipated pathogens responsible for infection with regards to the biologic half-life of the agent selected for therapy. Dosage intervals to treat infections due to highly sensitive bacteria need not occur more often than every sixth to eighth half-life. If less susceptible species are involved, however, dosing must be repeated every fourth half-life. For bacteria with sensitivities at only the higher concentrations of antibiotic that can be reliably achieved in blood or tissue, dosing intervals should occur even as often as every third half-life. More frequent antibiotic administration has never been shown to produce superior results [9]. For example, the twice-daily administration of an antibiotic with a prolonged half-life has given at least equal control and final cure of similar infections as when other equally effective combinations of antibiotics have been used at intervals that are appropriate to the individual agents half-life [2,9]. Antimicrobial therapy is usually continued until the acute infectious process has resolved. Exactly when that point has been reached is rarely certain. As a general rule, absence of fever indicates that sepsis has been at least controlled, although a permanent cure is not guaranteed. Probably the most reliable method is to use a combination of signs: an afebrile state, a relatively normal white blood cell count (i.e., in the range of 5,000 to 10,000/mm3 and absence of significant numbers (less than two per 100 count) of immature grahulocytes in the peripheral blood smear. Whenever these criteria have been met, the incidence of recurrent sepsis is noted to be 1 percent or less
Pll. Surgical Management. With the exception of proved streptococcal cellulitis, all wound infections require removal of sutures and opening of the incision down to fascia. Since the tissue plane of subcutaneous fat has the least resistance to infection, this is all that is required for most cases. However, if the fascia is necrotic, then debridement must be expeditiously and energetically accomplished. Necrotic tissue provides bacteria with necessary nutrients. Extensive fascial gangrene requires radical wound excision with replacement of fascia by some synthetic substitute. In the presehce of active infection, closure of the wound under tension or with transposition of tissue flaps will guarantee not only failure, but an even significant worsening of the septic process [7]. For intra-abdominal sepsis, drainage of the responsible focus is mandatory. This can be accomplished by a formal operation or by radiographically guided percutaneous
81 (suppl
1A)
SYMPOSIUM
ON PROBLEM
INFECTIONS-STONE
doses must be repeated every second half-life. As with therapy, the more ideal agents are those with a longer half-life, given an equality in antibacterial activity [15]. When infection develops despite an appropriately administered prophylactic antibiotic, the bacteria responsible for that episode of postoperative sepsis are sufficiently more resistant to the agent used for prophylaxis than would otherwise be anticipated [13]. Apparently even a brief exposure to an antibiotic guarantees the evolution of less susceptible strains. Thus, the antimicrobials selected for treatment of these complicating infections should never be the same agents as are used for prophylaxis. This observation thereby dictates that two sets of antibiotics be selected for any given type of surgical infectlous problem. For therapy, it is best to select the antimicrobial regimen that is at least equal to the best available set of antibiotics, yet has the lowest risk of drug-induced toxicity. It must be kept in mind that organ toxicity is directly related to both the total dose as well as the duration of therapy. Prophylaxis, with its short course of treatment, accordingly permits the administration of more toxic agents, since patient contact is brief. With such an approach, the complication of toxicity rarely, if ever, occurs. Multiple Organ System Failure. Persisting and uncontrolled sepsis will predictably lead to failure of one or more organ systems (31. This eventuality is noted all too frequently whenever aerobic gram-negative bacilli participate in the septic process. Endotoxin is probably the underlying mediator of these life-threatening problems. The organ systems generally involved are [3]: The lung, producing the classic adult respiratory distress syndrome; The kidney, initially presenting as oliguria, followed eventually by acute tubular necrosis; Coagulation mechanism derangements, generally beginning as hypoprothrombinemia, then platelet consumption, and eventually disseminated intravascular coagulation; and The liver, with such relatively late manifestations as jaundice and hepatic cell dysfunction. Antibiotics are useful for providing at least temporary and no more than partial control, but actual reversal of the process can be obtained only by eradication of the responsible focus of infection [l]. Removal of all nonviable tissue and revision for adequate drainage have no substitutes [7]. The only argument is how and when the definitive surgical procedure should be performed.
aspiration. The directive as to which is preferable for a given locus of infection within the abdomen has yet to be determined. However, whenever there is doubt, the surgical approach is probably more reliable for providing immediate reversal of the septic process. Antibiotics are useful and can at least temporarily control the attendant bacteremia. However, unless the primary focus has been corrected by debridement or drainage, bacteremia will persist and death is a likely outcome [2]. This is equally true in cases of necrotizing infection [7]. Truly, the “infection war is won or lost in the surgical wound.” Antibiotic Prophylaxis. The incidence of serious postoperative infection can be decreased significantly by the timely administration of an appropriate antibiotic [12]. The types of situations in which prophylaxis is indicated are relatively straightforward [I 31 and can be summarized as follows: Patients undergoing clean surgery have exceedingly low infection rates. Therefore, prophylaxis is indicated if infection of the operative area is predictably followed by a high mortality or permanent morbidity rate (i.e., cardiac, vascular, and orthopedic procedures); Prophylaxis is indicated in patients who have undergone clean-contaminated surgical procedures in which the risk of postoperative infection is great but seldom leads to death (i.e., alimentary, biliary, and gynecologic surgery); and Prophylaxis is indicated in patients with a profound impairment of their host defense mechanisms (i.e., leukemia, purposeful immunosuppression) who have undergone invasive surgical procedures. For prophylaxis to be effective, the antibiotic selected should be active against the majority of anticipated aerobic bacterial pathogens [13]. Susceptibility of anaerobic components within a polymicrobial inoculum has never been proved to be absolutely necessary. A maximal benefit is obtained only if effective concentrations of the antibiotic are continuously in the tissues at risk from the moment of contamination until the time of final wound closure [12-141. Thus, the antibiotic must be started prior to the operation, so as to guarantee desired tissue levels, and it must be maintained throughout the procedure until the incision has been sutured [14]. If the bacterial pathogens are believed to be highly sensitive to the antimicrobial agent chosen, then doses repeated at every third or fourth half-life are probably sufficient. However, if more resistant pathogens are expected, then l
l
l
l
l
l
l
REFERENCES 1.
2.
3.
Stone HH, Kolb LD, Geheber CE: Incidence and significance of intraperitoneal anaerobic bacteria. Ann Surg 1975; 181: 705710. Stone HH, Strom PR, Fabian TC, Dunlop WE: Third-generation cephalosporins for polymicrobial surgical sepsis. Arch Surg 1983; 118: 193-199. Polk HC, Shields CL: Remote organ failure: a valid sign of occult
July
28, 1989
4. 5. 6.
The
intra-abdominal infection. Surgery 1977; 81: 31 O-31 7. Cruse PJE, Foord R: A five-year prospective study of 23,649 surgical wounds. Arch Surg 1973; 107: 206-209. Gorbach SL, Bartlett JG: Anaerobic infections. N Engl J Med 1974; 290: 1177-1164, 1237-1245, 1269-1294. Meleney FL: Haemolytic streptococcus gangrene. Arch Surg 1924; 9: 317-323.
American
Journal
of Medicine
Volume
81 (suppl
1A)
43
SYMPOSIUM
7. 8. 9.
10.
11.
44
ON PROBLEM
INFECTIONS-STONE
Stone HH, Martin JD Jr: Synergistic necrotizing cellulitis. Ann Surg 1972; 175: 702-706. Fry DE, Garrison RN, Polk HC Jr: Clinical implications in Bacteroides bacteremia. Surg Gynecol Obstet 1979; 149: 189-l 92. Stone HH, Mullins RJ, Strom PR, Bourneuf AA, Geheber CE: Ceftriaxone versus combined gentamicin and clindamycin for polymicrobial surgical sepsis. Am J Surg 1984; 148: 30-34. Hooper CA, Haney BB, Stone HH: Gastrointestinal bleeding due to vitamin K deficiency in patients on parenteral cefamandole. Lancet 1980; I: 39. Stone HH, Bourneuf AA, Stinson LD: Reliability of criteria for predicting persistent or recurrent surgical sepsis. Arch Surg 1985; 120: 17-20.
July
28, 1986
The American
Journal
of Medicine
Volume
12.
13.
14.
15.
81 (suppl
Burke JF: The effective period of preventive action in experimental incisions and dermal lesions. Surgery 1961; 50: 161165. Stone HH, Hooper CA, Kolb LD, Geheber CE, Dawkins EJ: Antibiotic prophylaxis in gastric, biliary, and colonic surgery. Ann Surg 1976; 184: 443-451. Polk JC Jr, Lopez-Major JF: Postoperative wound infection: a prospective study of determinant factors and prevention, Surgery 1969; 66: 97-103. Condon RE, Bartlett JG, Nichols RL, Schulte WJ, Gorbach SL, Ochi S: Preoperative prophylactic cephalothin fails to control septic complications of colorectal operations: results of controlled clinical trial. Am J Surg 1973; 137: 68-74.
1A)
H. HARLAN STONE, M.D. Baltimore,
Maryland
The majority of surgical infections are due to multiple bacterial pathogens, usually represented by mixtures of both aerobic and anaerobic species. Contamination from endogenous sources accounts for the majority of these infections. The most virulent of all such sepsis appears to arise from a symbiosis between aerobic gram-negative rods and various anaerobes. Antibiotics have proved efficacy in both the treatment as well as the prevention of surgical infection. The choice of antimicrobial agent(s) should be based upon the drug’s spectrum of activity against known or anticipated pathogens, the biologic half-Me of the agent, which serves as a guide to the frequency of administration, and the drug’s safety. The third-generation cephalosporins have been shown to be especially useful because of their broad spectrum of activity, prolonged halflife, and limited toxicity. Sepsis that persists or is uncontrolled despite antibiotic administration often leads to failure of multiple organ systems. Only energetic surgical measures offer any real chance for patient survival when such a stage has been reached. INFECTION IN POSTOPERATIVEPATIENTS
From the Department of Surgery, University of Maryland School of Medicine, Baltimore, Maryland. Requests for reprints should be addressed to Dr. H. Harlan Stone, Department of Surgery, University of Maryland School of Medicine, 22 South Greene Street, Baltimore, Maryland 21201.
Although infection is the most common postoperative complication, it is not uniformly confined to the surgical incision or the body cavity that has been violated. Indeed, serious and life-threatening sepsis often results when there are major alterations in the normal physiology of one or more organ systems. Important factors in this regard are incisional pain, stasis in hollow visceral conduits responsible for secretion or excretion, and colonization by potential pathogens of otherwise sterile tissue planes or by new species able to supplant otherwise normal flora. Accordingly, before one can specifically incriminate the surgical incision or body cavity beneath, these other areas should be surveyed as well. Diagnosis of Postoperative Sepsis. During the postoperative period, fever usually heralds the onset of infection. However, if antibiotics have been continued after surgery, the temperature may be normal or of very low grade, i.e., seldom exceeding 37.5%. Nevertheless, leukocytosis will almost always be present, as will a significant shift to the left as manifested by many immature granulocytes. The four most common sites for the development of postoperative infection are the lungs, the urinary tract, veins that have been used for fluid therapy, and the wound itself. With respect to the latter, both the surgical incision and the cavity entered must be considered. The majority of pulmonary infections begin as atelectasis, with progression to pneumonia. Attention to proper pulmonary toilet is preventive as well as therapeutic. The diagnosis is suggested by changes on physical
July
28,1989
The American
Journal
of Medicine
Volume
81 (suppl
1A)
39
SYMPOSIUM
ON PROBLEM
INFECTIONS-STONE
examination in addition to production of purulent sputum, although chest radiography is exceedingly important in confirming the diagnosis. Nevertheless, in patients who have been receiving mechanical ventilation for long periods of time, changes of infection are indistinguishable from those of the “acute respiratory distress syndrome.” Accordingly, in this setting, purulent sputum is the most important sign. Such a sputum specimen must be obtained from deep within the tracheobronchial tree, smeared for immediate examination, and subsequently cultured to determine the pathogenic organism present. Bronchoscopy may be required to obtain the desired specimen and thus to make the diagnosis. Urinary tract infections usually occur in patients with urinary tract obstruction. These develop when stasis occurs in the upper tracts because of dehydration or when there is mechanical block to the bladder outlet because of prostatism or stricture. The diagnosis is dependent upon detecting bacteria in the urine once a brisk urine flow has been obtained. A Gram-stain preparation will give an idea as to the identity of the etiologic bacteria, but only a culture with colony count and sensitivity testing will give the final answer. Before a dilute urine has been achieved, white blood cell casts, thick pyuria, and heavy bacterial concentrations on a spun specimen can be taken as important signs that urinary tract infection is present. Bacteremia arising from an infected intravenous infusion site is relatively common. It must be kept in mind, however, that only approximately one half of these cases will manifest local phlebitis, so there may be seeding of the blood without obvious signs. Because of the latter observation, the intravenous sites should always be suspect, and therefore, bottles, tubing, and cannulas must all be changed. Material from the tip of the removed intravenous cannula should always be cultured. Obvious purulence at the area of needle or catheter entry through the skin is diagnostic, but only subsequent culture results will confirm the diagnosis when overt inflammation is absent. The most common site for infection is the surgical incision. Sepsis confined to the wound is due either to Streptococcus or Clostridium species if infection manifests between the first and second postoperative day. Infections appearing on the third to fourth postoperative day are generally due to Staphylococcus species. Mixed bacterial floras involving gram-negative bacilli and various anaerobes cause infections that are noted between the fourth and the sixth postoperative day. Mere gram-negative bacilli infection alone does not generally appear until after the seventh day. The discharge of pus or any purulent fluid from the wound, with or without a halo of cellulitis, is diagnostic. With respect to the cavity involved in the operation, infection may become diffuse and present as a fulminating form of sepsis, may loculate into one or more defined abscesses within that cavity, or may spontaneously erode
40
July
28, 1966
The American
Journal
of Medicine
Volume
through the surgical incision and discharge to the outside as a so-called subfascial abscess. Antibiotics generally lead to the confinement of the process to one or more loculi. These specific abscesses are identified by rectal examination if in the true pelvis, by mass and/or tenderness along the peritoneal gutters on abdominal examination, or by percussion tenderness along the costal margin if located in the subphrenic area. In cases in which anaerobes are involved in the infection, an overpenetrated roentgenogram of the abdomen will reveal multiple bubbles in areas where bowel should not be located. However, in situations in which confusion may arise because of the presence of intraluminal intestinal gas, or if anaerobes are not involved in the abscess, only computerized axial tomography can differentiate and identify the problem. Other modalities (i.e., sonography and galium scan, among others) are not nearly as reliable and give poorer definition. Recent clinical investigations have significantly altered our basic understanding of wound sepsis in postoperative patients. For example, more thorough processing of bacteriologic specimens has revealed that most infections are polymicrobial, usually with involvement of many different anaerobic as well as aerobic species [I]. A veritable explosion in the number and types of antibiotics available to treat such infections has demonstrated that the pharmacology and safety of the agent are as important as the antimicrobial spectrum of activity [2]. Antibiotic administration, however, must be appropriate for the indication to be treated, that is for cure or for prevention of infection. And finally, if the patient dies as a result of infection, it is usually a result of failure in function of various critical organ systems [3]. Bacteriology of Surgical Sepsis. A dictum that was once emphasized was that most infections arising on the surgical ward were due to a single pathogen and that aerobes were almost always responsible. However, detailed analyses of more expeditiously processed culture specimens has recently shown that surgical sepsis is usually polymicrobial [l]. The few exceptions are noted in those infections developing within the surgical incision of patients undergoing elective clean-category procedures. In these latter cases, the usual pathogen is an aerobic grampositive coccus, such as Staphylococcus aureus or Staphylococcus epidermidis [4]. Contamination has occurred primarily through an open incision during an operative procedure on the skeleton or cardiovascular system. Etiologic pathogens have almost uniformly been found to reside within the immediate environment of the incision and can be shown to be the same aerobic species as are present on the patient’s own skin or are components of some exogenous pool, such as those found on dirty instruments or in the nasopharynx of a member of the surgical team [4]. Nevertheless, in the majority of postoperative infec-
61 (suppl
1A)
SYMPOSIUM
28, 1989
INFECTIONS-STONE
death [3]. In other words, anaerobic participation correlates with morbidity, more specifically with the incidence and severity of local postoperative infection and the recurrence of such infection. The presence of aerobic bacteria, on the other hand, especially gram-negative bacilli circulating in the blood, appears to cause almost all of the deaths due to septicemia [2]. The unique synergy between aerobic and anaerobic species seems to be based upon the ability of aerobic pathogens to extract almost all of the available oxygen from the micro-environment [l]. Anaerobes can then survive, becoming significant virulence problems themselves through the elaboration of various proteolytic exotoxins. The more efficient the aerobes are at consuming oxygen radicals, the more conducive the local environment becomes for anaerobe growth. Likewise, if more potent enzymes are secreted by the anaerobes, then there is a greater destruction of host tissue and host resistance factors [7]. The incidence of bacteremia correspondingly escalates. Antibiotic administration directed against only the aerobic component can reduce both the incidence of postoperative infection and the mortality rate due to such complicating sepsis [2]. This is presumed to be accomplished by eliminating, at least in part, the aerobic support to this synergistic relationship, thereby significantly limiting anaerobic growth as a result of the persistently high oxygen tension in adjacent tissues [l]. Aerobic pathogens that have invaded the circulation are at the same time eradicated from the bloodstream. Antimicrobial activity against the anaerobes alone, however, has been shown to only decrease the postoperative incidence of sepsis. Mortality rates are not altered [2,8]. Antibiotic Therapy. The initial choice of antibiotic must always be based upon a drug’s spectrum of activity against proved or at least suspected pathogens. In cases of polymicrobial sepsis, either a single broad-spectrum agent or an antibiotic combination is indicated [2]. Both regimens appear to have equal efficacy [2]. However, persistence or recurrence of sepsis despite apparently adequate wound care and appropriate antibiotic selection suggests one or more of the following possibilities: There is active infection in some other area, such as an unrecognized pneumonia, urinary tract infection, or bacterial phlebitis, among others; There has been incomplete drainage of the original focus of infection or sepsis has recurred at that same site; The infecting bacteria are resistant to the antibiotic being administered; or Some opportunistic pathogen, such as Candida albicans, has evolved. In most cases of postoperative sepsis, a broad spectrum of antimicrobial activity should be provided by the antibiotic regimen. Ideally, it should include antimicrobial
tions, sepsis is caused by multiple species of bacteria that normally colonize one of the patient’s own organ systems [1,4]. The alimentary tract is the most common source of such endogenous pathogens. Although gram-negative rods predominate as the aerobic flora, anaerobes are involved as major pathogens in at least 80 percent of these cases [1,5]. The exact population of bacteria in a given polymicrobial sepsis is determined primarily by the level of the gastrointestinal tract entered, which serves as the source of contamination [l]. The stomach contains varying mixtures of aerobic and anaerobic species that are quite similar to those that can normally be found in the mouth. Bile and duodenal contents are generally sterile, although they may harbor aerobic gram-negative bacilli, enterococci, Clostridium species, and Bacteroides fragilis. Although the majority of bacteria found in the unobstructed small bowel are aerobic gram-negative bacilli, the very distal ileum ordinarily contains a large number of anaerobes, especially if the ileocecal valve is incompetent. In the colon, anaerobes predominate in a polymicrobial mix that contains enterococci a:- well as a large number of various aerobic gram-negative bacilli [l]. Anaerobes apparently cannot migrate across healthy, oxygenated tissue planes. Only aerobic species appear capable of such migration. Therefore, to reach the wound or peritoneal cavity, anaerobes must pass through a grossly perforated bowel or across a segment of severely ischemic or frankly gangrenous intestine [l]. Otherwise, contamination and subsequent infection will be due solely to aerobic bacteria. Similar bacteriologic results are obtained from the urinary and female genital tracts. With acute infection, cultures yield primarily a single species of aerobic pathogens. However, after repeated or chronic infection, polymicrobial sepsis usually evolves, especially in the female genital tract. Indeed, chronic pelvic infections involving the fallopian tubes and/or ovaries generally contain a mixed bacterial flora comprising anaerobes as well as aerobes [5]. Bacterial Synergy. Polymicrobial contamination routinely leads to a significantly greater incidence of postoperative sepsis than does wound inoculation by only single species. Whenever anaerobes become a part of that contaminant, the risk and severity of subsequent infection is even greater [1,5-q. Identical results have been noted with respect to the recurrence of sepsis [1,5]. Aerobes seem to play a dual role, for anaerobes alone can seldom establish an infection in otherwise healthy tissues [l]. Only in the case of clostridial sepsis does it appear that death can be caused solely by an anaerobe [B]. It is the aerobic gram-negative bacteria that are responsible for the circulation of endotoxin and other vasoactive substances in the blood. These in turn cause failure of one or more organ systems and thereby account for eventual
July
ON PROBLEM
l
l
l
l
The
American
Journal
of Mediclne
Volume
81 (suppl
1A)
41
SYMPOSIUM
ON PROBLEM
INFECTIONS-STONE
coverage against S. aureus, the majority of aerobic gramnegative bacilli, and a significant number of anaerobes [2]. Selectivity against Pseudomonas aeruginosa is of importance only for hospital-acquired polymicrobial sepsis and for patients with deficiencies in their host defense mechanisms due to disease or medications. The anaerobic antibiotic spectrum need not be complete. Coverage against B. fragilis is also not an absolute necessity, since results appear to be just as good with as without such activity [l]. Because of their greater tolerance to atmospheric oxygen, Bacteroides species are much easier to culture. Thus, they serve as excellent “indicators” that anaerobes are present. Their greater virulence in comparison to less readily cultured anaerobes has yet to be proved. Equally as crucial as microbial susceptibility to the administered antibiotic is a dosing schedule that insures drug delivery to the site of infection at appropriate intervals and ih concentrations necessary for bacterial kill. First- and second-generation cephalosporins have good activity against aerobic gram-positive cocci, but much of this has been lost with the molecular manipulations that produced the third-generation cephalosporins [2]. The minimal inhibitory concentration required for aerobic gram-positive cocci is generally quite low, in the range of 1 or 2 pg/ml. For the aerobic gram-negative bacilli, bactericidal concentrations are somewhat higher, whereas for anaerobes an even greater concentratibn is generally required for agents with demonstrable activity [2,9]. Accordingiy, for therapy to be reliably adequate, larger doses for the treatment of aerobic gram-negative bacilli infections are required over those used for aerobic gram-positive coccal infections, and even greater doses are needed if anaerobes are present. With the newer cephalosporins, there is selectively greater antimicrobial activity against gram-negative bacilli, as well as against anaerobic organisms [2,9]. When Pseudomonas infection is suspected, an antipseudomonal penicillin (or other beta-lactam) plus an aminoglycoside should be selected. If treatment with multiple antibiotics has been chosen, then the need to modify the dose of each individual agent according to the type of infection seldom becomes a major consideration. The safety of antimicrobial therapy must also be taken into account [2]. This is not confined solely to the allergic reactions that can occur with penicillins and cephalosporins. Drug toxicity must always be kept in mind as a potential complication. Certain cephalosporins, especially in large doses, impair prothrombin production and a coagulopathy can develop [2,10]. Use of lower doses and concomitant administration of vitamin K has generally prevented this problem. On the other hand, aminoglycosides are potentially nephrotoxic [2]. Use of certain diuretics and present or prior hypovolemia significantly increase the likelihood of this complication. Overall, the incidence of
42
July
28,1988
The Amerlcen
Journal
of Medicine
Volume
aminoglycoside-induced renal failure is approximately 5 percent for surgical patients. It is a complication that significantly worsens both immediate as well as final outcomes, and, accordingly, the use of aminogiycosides should be avoided in these high-risk cases whenever possible [2]. The frequency of antibiotjc administration should be based upon the antimicrobial susceptibility of the known or anticipated pathogens responsible for infection with regards to the biologic half-life of the agent selected for therapy. Dosage intervals to treat infections due to highly sensitive bacteria need not occur more often than every sixth to eighth half-life. If less susceptible species are involved, however, dosing must be repeated every fourth half-life. For bacteria with sensitivities at only the higher concentrations of antibiotic that can be reliably achieved in blood or tissue, dosing intervals should occur even as often as every third half-life. More frequent antibiotic administration has never been shown to produce superior results [9]. For example, the twice-daily administration of an antibiotic with a prolonged half-life has given at least equal control and final cure of similar infections as when other equally effective combinations of antibiotics have been used at intervals that are appropriate to the individual agents half-life [2,9]. Antimicrobial therapy is usually continued until the acute infectious process has resolved. Exactly when that point has been reached is rarely certain. As a general rule, absence of fever indicates that sepsis has been at least controlled, although a permanent cure is not guaranteed. Probably the most reliable method is to use a combination of signs: an afebrile state, a relatively normal white blood cell count (i.e., in the range of 5,000 to 10,000/mm3 and absence of significant numbers (less than two per 100 count) of immature grahulocytes in the peripheral blood smear. Whenever these criteria have been met, the incidence of recurrent sepsis is noted to be 1 percent or less
Pll. Surgical Management. With the exception of proved streptococcal cellulitis, all wound infections require removal of sutures and opening of the incision down to fascia. Since the tissue plane of subcutaneous fat has the least resistance to infection, this is all that is required for most cases. However, if the fascia is necrotic, then debridement must be expeditiously and energetically accomplished. Necrotic tissue provides bacteria with necessary nutrients. Extensive fascial gangrene requires radical wound excision with replacement of fascia by some synthetic substitute. In the presehce of active infection, closure of the wound under tension or with transposition of tissue flaps will guarantee not only failure, but an even significant worsening of the septic process [7]. For intra-abdominal sepsis, drainage of the responsible focus is mandatory. This can be accomplished by a formal operation or by radiographically guided percutaneous
81 (suppl
1A)
SYMPOSIUM
ON PROBLEM
INFECTIONS-STONE
doses must be repeated every second half-life. As with therapy, the more ideal agents are those with a longer half-life, given an equality in antibacterial activity [15]. When infection develops despite an appropriately administered prophylactic antibiotic, the bacteria responsible for that episode of postoperative sepsis are sufficiently more resistant to the agent used for prophylaxis than would otherwise be anticipated [13]. Apparently even a brief exposure to an antibiotic guarantees the evolution of less susceptible strains. Thus, the antimicrobials selected for treatment of these complicating infections should never be the same agents as are used for prophylaxis. This observation thereby dictates that two sets of antibiotics be selected for any given type of surgical infectlous problem. For therapy, it is best to select the antimicrobial regimen that is at least equal to the best available set of antibiotics, yet has the lowest risk of drug-induced toxicity. It must be kept in mind that organ toxicity is directly related to both the total dose as well as the duration of therapy. Prophylaxis, with its short course of treatment, accordingly permits the administration of more toxic agents, since patient contact is brief. With such an approach, the complication of toxicity rarely, if ever, occurs. Multiple Organ System Failure. Persisting and uncontrolled sepsis will predictably lead to failure of one or more organ systems (31. This eventuality is noted all too frequently whenever aerobic gram-negative bacilli participate in the septic process. Endotoxin is probably the underlying mediator of these life-threatening problems. The organ systems generally involved are [3]: The lung, producing the classic adult respiratory distress syndrome; The kidney, initially presenting as oliguria, followed eventually by acute tubular necrosis; Coagulation mechanism derangements, generally beginning as hypoprothrombinemia, then platelet consumption, and eventually disseminated intravascular coagulation; and The liver, with such relatively late manifestations as jaundice and hepatic cell dysfunction. Antibiotics are useful for providing at least temporary and no more than partial control, but actual reversal of the process can be obtained only by eradication of the responsible focus of infection [l]. Removal of all nonviable tissue and revision for adequate drainage have no substitutes [7]. The only argument is how and when the definitive surgical procedure should be performed.
aspiration. The directive as to which is preferable for a given locus of infection within the abdomen has yet to be determined. However, whenever there is doubt, the surgical approach is probably more reliable for providing immediate reversal of the septic process. Antibiotics are useful and can at least temporarily control the attendant bacteremia. However, unless the primary focus has been corrected by debridement or drainage, bacteremia will persist and death is a likely outcome [2]. This is equally true in cases of necrotizing infection [7]. Truly, the “infection war is won or lost in the surgical wound.” Antibiotic Prophylaxis. The incidence of serious postoperative infection can be decreased significantly by the timely administration of an appropriate antibiotic [12]. The types of situations in which prophylaxis is indicated are relatively straightforward [I 31 and can be summarized as follows: Patients undergoing clean surgery have exceedingly low infection rates. Therefore, prophylaxis is indicated if infection of the operative area is predictably followed by a high mortality or permanent morbidity rate (i.e., cardiac, vascular, and orthopedic procedures); Prophylaxis is indicated in patients who have undergone clean-contaminated surgical procedures in which the risk of postoperative infection is great but seldom leads to death (i.e., alimentary, biliary, and gynecologic surgery); and Prophylaxis is indicated in patients with a profound impairment of their host defense mechanisms (i.e., leukemia, purposeful immunosuppression) who have undergone invasive surgical procedures. For prophylaxis to be effective, the antibiotic selected should be active against the majority of anticipated aerobic bacterial pathogens [13]. Susceptibility of anaerobic components within a polymicrobial inoculum has never been proved to be absolutely necessary. A maximal benefit is obtained only if effective concentrations of the antibiotic are continuously in the tissues at risk from the moment of contamination until the time of final wound closure [12-141. Thus, the antibiotic must be started prior to the operation, so as to guarantee desired tissue levels, and it must be maintained throughout the procedure until the incision has been sutured [14]. If the bacterial pathogens are believed to be highly sensitive to the antimicrobial agent chosen, then doses repeated at every third or fourth half-life are probably sufficient. However, if more resistant pathogens are expected, then l
l
l
l
l
l
l
REFERENCES 1.
2.
3.
Stone HH, Kolb LD, Geheber CE: Incidence and significance of intraperitoneal anaerobic bacteria. Ann Surg 1975; 181: 705710. Stone HH, Strom PR, Fabian TC, Dunlop WE: Third-generation cephalosporins for polymicrobial surgical sepsis. Arch Surg 1983; 118: 193-199. Polk HC, Shields CL: Remote organ failure: a valid sign of occult
July
28, 1989
4. 5. 6.
The
intra-abdominal infection. Surgery 1977; 81: 31 O-31 7. Cruse PJE, Foord R: A five-year prospective study of 23,649 surgical wounds. Arch Surg 1973; 107: 206-209. Gorbach SL, Bartlett JG: Anaerobic infections. N Engl J Med 1974; 290: 1177-1164, 1237-1245, 1269-1294. Meleney FL: Haemolytic streptococcus gangrene. Arch Surg 1924; 9: 317-323.
American
Journal
of Medicine
Volume
81 (suppl
1A)
43
SYMPOSIUM
7. 8. 9.
10.
11.
44
ON PROBLEM
INFECTIONS-STONE
Stone HH, Martin JD Jr: Synergistic necrotizing cellulitis. Ann Surg 1972; 175: 702-706. Fry DE, Garrison RN, Polk HC Jr: Clinical implications in Bacteroides bacteremia. Surg Gynecol Obstet 1979; 149: 189-l 92. Stone HH, Mullins RJ, Strom PR, Bourneuf AA, Geheber CE: Ceftriaxone versus combined gentamicin and clindamycin for polymicrobial surgical sepsis. Am J Surg 1984; 148: 30-34. Hooper CA, Haney BB, Stone HH: Gastrointestinal bleeding due to vitamin K deficiency in patients on parenteral cefamandole. Lancet 1980; I: 39. Stone HH, Bourneuf AA, Stinson LD: Reliability of criteria for predicting persistent or recurrent surgical sepsis. Arch Surg 1985; 120: 17-20.
July
28, 1986
The American
Journal
of Medicine
Volume
12.
13.
14.
15.
81 (suppl
Burke JF: The effective period of preventive action in experimental incisions and dermal lesions. Surgery 1961; 50: 161165. Stone HH, Hooper CA, Kolb LD, Geheber CE, Dawkins EJ: Antibiotic prophylaxis in gastric, biliary, and colonic surgery. Ann Surg 1976; 184: 443-451. Polk JC Jr, Lopez-Major JF: Postoperative wound infection: a prospective study of determinant factors and prevention, Surgery 1969; 66: 97-103. Condon RE, Bartlett JG, Nichols RL, Schulte WJ, Gorbach SL, Ochi S: Preoperative prophylactic cephalothin fails to control septic complications of colorectal operations: results of controlled clinical trial. Am J Surg 1973; 137: 68-74.
1A)