Diagnostic Peritoneal Lavage in Acute Peritonitis Frank A. Larson, Mo, Chris C. Hailer, MD, Romano Delcore, MD, James H. Thomas, MD, KansasCity, Kansas
Fifty patients were studied prospectively to determine the value of diagnostic peritoneal lavage in the evaluation of acute peritonitis. Forty-five patients had a clinical diagnosis of acute peritonitis based on physical findings, and 5 patients were normal control subjects. All lavages were performed with an open technique in the operating room prior to laparotomy. Lavage fluid was analyzed for white and red blood cell counts, the dif-
ferential, amylase, protein, bilirubin, and pH. Resuits were analyzed by multiple logistic regression. Thirty-two of the 45 patients with clinical peritonitis had the diagnosis confirmed at laparotomy. A white blood cell count in the lavage fluid greater than or equal to 2 0 0 cells/mm 3 was associated with a 99% probability of peritonitis.
examination is the standard method of diagp ) hysical nosis in patients with acute abdominal pain suggestive of peritonitis. Patients must be cooperative, have an intact nervous system, and be capable of initiating an inflammatory response for the assessment of the abdomen to be reliable. Radiologie examinations including computed tomographic (CT) scanning may be valuable diagnostic adjuncts, but these studies are limited by the inability of some patients to take oral or intravenous contrast and to be transported to the radiology suite. Norwood and Civetta [1] demonstrated that abdominal CT scans in critically ill patients in the intensive care unit rarely provide useful information. These authors also found that, as the difficulty in making a clinical diagnosis increased, the likelihood of obtaining a useful CT scan diminished. Diagnostic peritoneal lavage (DPL) has been evaluated in patients with presumed peritonitis unrelated to trauma [2-7]. DPL may be particularly helpful in patients with unreliable abdominal examinations, including those who are critically ill with suspected intra-abdominal sepsis. However, previous assessments of DPL have not provided diagnostic criteria specific for acute peritonitis. Rather, these studies used the standards originally developed by Perry and Strate [8] for DPL in abdominal trauma. Alverdy et al [9] modified the criteria, but the modifications were based on a retrospective clinical study. The current prospective study was designed to determine the usefulness of DPL in the evaluation of patients with acute peritonitis and to more precisely define the characteristics of lavage fluid in these patients.
PATIENTS AND M E T H O D S Between May 1989 and March 1992, 45 patients (30 male, 15 female) with a mean age of 52 years (range: 11 to 88 years) were studied prospectively. All patients had a presumed diagnosis of acute peritonitis based on localized or diffuse abdominal pain associated with abdominal guarding or rebound tenderness. Five additional male patients without acute peritonitis served as controls. DPL was performed in the operating room prior to laparotomy. An open technique was used employing a plastic sumpsuction catheter inserted through the site of the planned laparotomy incision. One liter of normal saline was infused into the peritoneal cavity, and the effluent was collected. The procedure required approximately 8 minutes. A minimum of 500 mL of effluent was required for patients to be included in the study. The fluid was analyzed for white blood cells (WBCs), the differential, red blood cells (RBCs), pH, and amylase, protein, and biliruFrom the Universityof Kansas Medical Center, Department of Sur- bin levels. Intraoperatively, the peritoneum was evaluatgery, Kansas City, Kansas. ed for the presence of inflammation. Since residual lavage Requests for reprints should be addressed to James H. Thomas, fluid was present in all patients, peritonitis was confirmed MD, Universityof Kansas Medical Center, 3901 RainbowBoulevard, by the presence of purulent exudate, edema, cloudiness, 4002 Murphy, Kansas City, Kansas66160-7308. or hyperemia of the parietal or visceral peritoneum. Cell Presented at the 44th Annual Meetingof the SouthwesternSurgical Congress,Scottsdale,Arizona, April 26-29, 1992. counts and chemistry levels were analyzed by multiple THE AMERICAN JOURNAL OF SURGERY VOLUME164 NOVEMBER 1992 449
LARSON ET AL
TABLE I
Operative F i n d i n g s
Operative Findings (n) Controls (5) Negative laparotomy (10) Perforated viscus Gastroduodenal (4) Small/large bowel (6) Appendicitis Nonperforated (6) Perforated (4) Mesenteric ischemia (8) Crohn's disease (3) Bowel obstruction ' Simple (2) Salpingitis (1) Hemobilia/gallbladder distention (t)
and L a v a g e Fluid A n a l y s i s Lavage Results (Range) RBC Protein Bilirubin (cells/mm3) (g/dL) (mg/dL)
WBC (cells/mm 3)
PMN (%)
3-20 0-196
5-10 10-92
3-447 15-20,570
0-0.5 0-0.5
0 0.1-0.5
0-6 0--30
6.0-7.0 6.27-7.6
63-19,750 220"10,900
80-100 60-100
40-12,000 930-21,000
0.1-i .0 0.5-1.0
0-1.0 0.5-1.9
1-629 1-3,532
6.61-8 6.27-7.58
88-8,250 245-24,550 410-7,000 1-220
22-95 84-97 42-96 10-80
28-1,132 350-8,694 10-200;000 1,599-5,050
0.1-1.0 0.5-6.7 0.5-4.0 0-0.5
0-0.5 0.1-1.4 0-0.3 0-0.5
2-30 3-28 3-998 2-3
6.5-6.9 6.52-7.3 6.8-8.0 6.6-7.01
0-274 0.5 0
0-0.1 O.1 0.3
5-15 2 15
6.0-8.0 6.79 7.5
31-183 2,700 9
12-33 89 t6
0-1,403 1,190 523
Amylase (IU/L)
pH
WBC = white blood cell; PMN = polymorphonuclearcell; RBC = red blood cell.
logistic regression to identify which factor(s) correlated with the presence of acute peritonitis and to determine if peritonitis could be reliably predicted. RESULTS The operative findings for control and study patients are listed in Table L Acute peritonitis was confirmed at laparotomy in 32 of the 45 patients (71%). Twenty-two of the patients with confirmed peritonitis (69%) had either a perforated viscus (n = 14) or ischemic bowel (n = 8), and the other 10 patients had inflammation of a hollow viscus (nonperforated appendicitis, Crohn's disease, and salpingitis). Lavage results are also presented in Table I. The lavage effluent from uninfiamed peritoneum in control subjects contained small numbers of WBCs (range: 3 to 20 cells/ram3), polymorphonuclear (PMN) cells (range: 5% to 10%), and RBCs (range: 3 to 447 cells/mm3), minimal protein (0 to 0.5 g/dL), a low concentration of amylase (range: 0 to 6 1U/L), and no bilirubin. In con. wast, the lavage fluid from patients with confirmed peritonitis typically had elevated numbers of WBCs (range" 1 to 24,550 cells/ram3), an increased percentage of PMN cells (range: 10% to 100%), a variable number of RBCs (range: 10 to 200,000 cells/mm3), and variable concentrations of protein (range: 0 to 6.7 g/dL), amylase (range: 0 to 3,532 IU/L), and bilirubin (range: 0 to 1.9 mg/dL). Lavage results were analyzed by multiple logistic regression, and the total number of leukocytes per cubic millimeter was the only factor predictive of peritonitis (p 200 = 99% probability of acute peritonitis
a result of the accuracy and reliability of physical examination in most patients with abdominal pain. A recent report by Adams et al [13] described a 25% negative laparotomy rate in patients with acute abdominal pain assessed clinically. In the current study, a clinical diagnosis of peritonitis was not confirmed by intraoperative findings in 13 (29%) of the 45 patients. Although DPL might improve the accuracy of physical examination in all patients with suspected peritonitis, this application is impractical. DPL may be most beneficial in patients with equivocal or unreliable physical examinations. As demonstrated by Veith et al [4] and confirmed in the current study, the DPL leukocyte count in control subjects without acute peritonitis is low, ranging from 0 to 20 cells/mm 3. This finding increases the probability that an elevated peritoneal WBC count may be diagnostic of peritoneal inflammation. At present, there are no standards that allow for the interpretation of the results of DPL in patients with acute peritonitis despite wellaccepted criteria for DPL in injured patients. Early studies used the trauma criterion of 500 W B C / m m 3 to designate a positive lavage [2,3,5,14]. In a retrospective study, Alverdy et al [9] considered positive those lavages with a WBC count of 250 to 500 cells/mm 3 if 75% or more PMN cells were present. The study described herein is the first prospective evaluation to characterize lavage fluid in patients with peritonitis. These patients typically have an increased number of WBCs in lavage fluid. A WBC count greater than or equal to 200/ram 3 was the only parameter associated with a 99% probability of acute peritonitis. Monocytes are the predominant cell type in normal peritoneal fluid. An increasing predominance of neutrophils is seen with acute peritoneal inflammation. Stewart and associates [!5] demonstrated that the percentage of ncutrophils in the leukocyte differential was of diagnostic significance. However, the percentage of PMN cells in
the leukocyte differential in this study did not yield a higher probability of peritonitis than that obtained with the leukocyte count alone. There are clinical situations in which the sensitivity of DPL may be decreased. Steroids diminish the inflammatory response and may affect the number of leukocytes in the peritoneal fluid. Five patients with clinical peritonitis in this report were receiving steroids at the time of lavage. Two patients, one with focal inflammation of the mesoappendix and another with a well-localized abscess, had leukocyte counts below 200 cells/mm 3 in the lavage fluid. In contrast, each of the three patients with lavage leukocyte counts greater than or equal to 200 c e l l s / r a m 3 had massive peritoneal soilage and inflammation secondary to gastrointestinal perforations. Thus, the degree of peritoneal soilage may influence the results of DPL in patients receiving corticosteroids. Although the use of antibiotic therapy was not investigated in this report, it is conceivable that these drugs may alter the peritoneal response to infection and also decrease the number of WBCs in the peritoneal fluid. This study suggests that DPL is a rapid, accurate, and safe method of assessing the abdominal cavity in patients with suspected acute peritonitis. The leukocyte count of the lavage fluid is increased in acute peritoneal inflammation and appears to be the most accurate predictor of peritonitis. A value greater than or equal to 200 leukocytes/mm 3 is the main criterion for a positive DPL result. REFERENCES 1. Norwood SH, Civetta JM. Abdominal CT scanning in critically ill surgical patients. Ann Surg 1985; 202: 166-75. 2. Richardson JD, Flint LM, Polk HC Jr. Peritoneal lavage: a useful diagnostic adjunct for peritonitis. Surgery 1983; 94: 826-9. 3. Hoffman J, Lanng C, Shokouh-Amiri MH. Peritoneal lavage in the diagnosis of acute peritonitis. Am J Surg 1988; 155: 359-60. 4. Veith FJ, Webber WB, Karl RC, Deysine M. Diagnostic perito-
THE AMERICAN JOURNAL OF SURGERY VOLUME 164 NOVEMBER 1992 451
LARSON ET AL
toneal abscess and the accuracy of the DPL in these situations because I suspect that, as in a case of ours, the DPL is limited in detecting retroperitoneal inflammatory processes. Ernest E. Moore (Denver, CO): With the revitalized interest in gut bacterial translocation, did you measure endotoxin in the lavage fluid or use Gram's stain to identify bacteria? Second, did you examine specific markers of intestinal disruption such as alkaline phosphatase or diamine oxidase? Finally, have you assessed potential inflammatory markers in the peritoneal fluid, such as phospholipase A2 or receptor expression on neutrophils? Steve Smith (Oakland, CA): Could you comment on whether any patients in the miscellaneous group were assessed for postoperative complications of recent intraabdominal surgery? Also, would you choose not to operate under any circumstances if a patient had a clinical presentation of peritonitis and a WBC count of 200 cells/ mm a by DPL? If so, would you repeat the peritoneal 7. 12. Root HD, Keizer PJ, Perry JF. The clinicaland experimental lavage at a later time, and what other studies would you aspectsof peritonealresponseto injury.Arch Surg 1967;95:53 I-7. do to assess the patient's condition? 13. Adams ID, Chan M, Clifford PC, et aL Computer aided David Easter (San Diego, CA): Was a general anesdiagnosisof acute abdominal pain;a multiccnterstudy.B M J 1986; thetic used in all cases? What was your standard for 293: 800-4. determining positive findings (i.e., was it a visual inspec14. Barbee CL, GilsdorfRB. Diagnosticperitoneallavagein evalu- tion at laparotomy)? If so, why not add diagnostic lapaating acute abdominal pain. Ann Surg 1975; 181: 853-6. roscopy instead of lavage? 15. Stewart RJ, Holloway LJ, IsbisterW H . PeritonealneutrophiFrank A. Larson (closing): We have undertaken the lia:a potentialindicatorof the surgicalacute abdomen. Aust N Z J second phase of this study in which we are using the Surg 1984; 54: 565-8. criterion for a positive DPL developed in the first phase to test the accuracy of DPL in patients with suspected acute DISCUSSION June E. Heilman (Pocatello, ID): This study should peritonitis, We have performed lavages both in the emerhelp to more clearly define some of the indications for gency department and in the intensive care unit. Alsurgery in patients with abdominal pain who are difficult though we have just undertaken this second phase, the to examine and are fairly ill. This is the lowest number of early results are encouraging. DPL was performed in all white blood cell (WBC) counts of any of the studies using of the initial patients in the operating room since this was diagnostic peritoneal lavage (DPL) to indicate peritoni- part of the study protocol. Dr. Schauer, there were two tis. Perhaps it will help to more precisely identify patients patients who had retroperitoneal abscesses--one had a with peritonitis. I don't think that the four false-negative positive lavage with a WBC count greater than 200 cells/ results are surprising, considering that these were all lo- mm a, and the second patient had a negative study, in that calized processes. It was interesting to note that of the 10 his lavage WBC count was less than 200 cells/mm3: This cases of appendicitis, 6 were nonperforated. DPL does was the patient who had been receiving antibiotics postidentify some localized processes. Thirteen of your pa- operatively. Regarding Dr. Moore's question about endotients with negative laparotomy studies had WBC counts toxin, Gram's stains, and other markers, we did examine ranging from 0 to 196/mm 3 in the effluent. Did the the Gram's stains of the lavage fluid in just over half of majority of these patients have localized signs or diffuse these patients, and we found that the patients who had tenderness? Will these data alter your care in the future? positive Gram's stains all had WBC counts greater than Could you identify any patients who would not need or equal to 200 celis/mm 3 in the lavage fluid. No other laparotomy by relying on the results of the DPL? Patients markers in the lavage fluid were studied. We do not who have had previous laparotomies, especially potential- believe that other markers would be as readily obtainable ly septic patients in an intensive care unit, are difficult to in the community hospital setting, and we hope that this evaluate. Have you used DPL in any of these patients? will be a technique that can be employed relatively quickly at the bedside. Regarding the question on the patients Ray Ryan (Wichita, KS): In your study, all D P L s were performed in the operating room. Could this proce- in the other category, this included three patients with Crohn's disease, one patient with acute gallbladder disdure be done in the intensive care unit? Philip R. Schauer (San Antonio, TX): Dr. Larson, I tention secondary to hemobilia, and one patient with would like to know if any of your patients had a retropcri- acute salpingitis.
neal lavagein acute abdominal disease:normal findingsand evaluation in I00 patients.Ann Surg 1967; 166: 290-5. 5. BaileyRL, Laws HL. Diagnosticperitoneallavagein evaluation of acute abdominal disease.South Mcd J 1990; 83: 422-4. 6. Lobbato V, Cioroiu M, LaRaja RD, Rothenberg RE, Strand M. Peritoneallavage as an aid to diagnosisof peritonitisin debilitated and elderlypatients.A m Surg 1985; 51: 508-10. 7. Evans C, Rashid A, Rosenbcrg [L, PollackAV. An appraisalof peritoneallavage in the diagnosisof the acute abdomen. Br J Surg 1975; 62: 119-20. 8. Perry JF, Strate RG. Diagnostic peritoneal lavage in blunt abdominal trauma: indications and results.Surgery 1972; 71: 898-901. 9. Alverdy JC, Saunders J, Chambcrlin W H , Moss GS. Diagnostic peritoneal lavage in intra-abdorninalsepsis.A m Surg 1988; 54: 456-9. 10. Root HD, Hauser CW, McKinley CR, LaFave JW, Mendola RP. Diagnostic peritoneallavage. Surgery 1965; 57: 633-7. I I. Advanced trauma lifesupport student manual. Chicago, IL: Committee on Trauma, American Collegeof Surgeons, 1989:113-
4,52 THE AMERICAN JOURNAL OF SURGERY VOLUME164 NOVEMBER1992
Fifty patients were studied prospectively to determine the value of diagnostic peritoneal lavage in the evaluation of acute peritonitis. Forty-five patients had a clinical diagnosis of acute peritonitis based on physical findings, and 5 patients were normal control subjects. All lavages were performed with an open technique in the operating room prior to laparotomy. Lavage fluid was analyzed for white and red blood cell counts, the dif-
ferential, amylase, protein, bilirubin, and pH. Resuits were analyzed by multiple logistic regression. Thirty-two of the 45 patients with clinical peritonitis had the diagnosis confirmed at laparotomy. A white blood cell count in the lavage fluid greater than or equal to 2 0 0 cells/mm 3 was associated with a 99% probability of peritonitis.
examination is the standard method of diagp ) hysical nosis in patients with acute abdominal pain suggestive of peritonitis. Patients must be cooperative, have an intact nervous system, and be capable of initiating an inflammatory response for the assessment of the abdomen to be reliable. Radiologie examinations including computed tomographic (CT) scanning may be valuable diagnostic adjuncts, but these studies are limited by the inability of some patients to take oral or intravenous contrast and to be transported to the radiology suite. Norwood and Civetta [1] demonstrated that abdominal CT scans in critically ill patients in the intensive care unit rarely provide useful information. These authors also found that, as the difficulty in making a clinical diagnosis increased, the likelihood of obtaining a useful CT scan diminished. Diagnostic peritoneal lavage (DPL) has been evaluated in patients with presumed peritonitis unrelated to trauma [2-7]. DPL may be particularly helpful in patients with unreliable abdominal examinations, including those who are critically ill with suspected intra-abdominal sepsis. However, previous assessments of DPL have not provided diagnostic criteria specific for acute peritonitis. Rather, these studies used the standards originally developed by Perry and Strate [8] for DPL in abdominal trauma. Alverdy et al [9] modified the criteria, but the modifications were based on a retrospective clinical study. The current prospective study was designed to determine the usefulness of DPL in the evaluation of patients with acute peritonitis and to more precisely define the characteristics of lavage fluid in these patients.
PATIENTS AND M E T H O D S Between May 1989 and March 1992, 45 patients (30 male, 15 female) with a mean age of 52 years (range: 11 to 88 years) were studied prospectively. All patients had a presumed diagnosis of acute peritonitis based on localized or diffuse abdominal pain associated with abdominal guarding or rebound tenderness. Five additional male patients without acute peritonitis served as controls. DPL was performed in the operating room prior to laparotomy. An open technique was used employing a plastic sumpsuction catheter inserted through the site of the planned laparotomy incision. One liter of normal saline was infused into the peritoneal cavity, and the effluent was collected. The procedure required approximately 8 minutes. A minimum of 500 mL of effluent was required for patients to be included in the study. The fluid was analyzed for white blood cells (WBCs), the differential, red blood cells (RBCs), pH, and amylase, protein, and biliruFrom the Universityof Kansas Medical Center, Department of Sur- bin levels. Intraoperatively, the peritoneum was evaluatgery, Kansas City, Kansas. ed for the presence of inflammation. Since residual lavage Requests for reprints should be addressed to James H. Thomas, fluid was present in all patients, peritonitis was confirmed MD, Universityof Kansas Medical Center, 3901 RainbowBoulevard, by the presence of purulent exudate, edema, cloudiness, 4002 Murphy, Kansas City, Kansas66160-7308. or hyperemia of the parietal or visceral peritoneum. Cell Presented at the 44th Annual Meetingof the SouthwesternSurgical Congress,Scottsdale,Arizona, April 26-29, 1992. counts and chemistry levels were analyzed by multiple THE AMERICAN JOURNAL OF SURGERY VOLUME164 NOVEMBER 1992 449
LARSON ET AL
TABLE I
Operative F i n d i n g s
Operative Findings (n) Controls (5) Negative laparotomy (10) Perforated viscus Gastroduodenal (4) Small/large bowel (6) Appendicitis Nonperforated (6) Perforated (4) Mesenteric ischemia (8) Crohn's disease (3) Bowel obstruction ' Simple (2) Salpingitis (1) Hemobilia/gallbladder distention (t)
and L a v a g e Fluid A n a l y s i s Lavage Results (Range) RBC Protein Bilirubin (cells/mm3) (g/dL) (mg/dL)
WBC (cells/mm 3)
PMN (%)
3-20 0-196
5-10 10-92
3-447 15-20,570
0-0.5 0-0.5
0 0.1-0.5
0-6 0--30
6.0-7.0 6.27-7.6
63-19,750 220"10,900
80-100 60-100
40-12,000 930-21,000
0.1-i .0 0.5-1.0
0-1.0 0.5-1.9
1-629 1-3,532
6.61-8 6.27-7.58
88-8,250 245-24,550 410-7,000 1-220
22-95 84-97 42-96 10-80
28-1,132 350-8,694 10-200;000 1,599-5,050
0.1-1.0 0.5-6.7 0.5-4.0 0-0.5
0-0.5 0.1-1.4 0-0.3 0-0.5
2-30 3-28 3-998 2-3
6.5-6.9 6.52-7.3 6.8-8.0 6.6-7.01
0-274 0.5 0
0-0.1 O.1 0.3
5-15 2 15
6.0-8.0 6.79 7.5
31-183 2,700 9
12-33 89 t6
0-1,403 1,190 523
Amylase (IU/L)
pH
WBC = white blood cell; PMN = polymorphonuclearcell; RBC = red blood cell.
logistic regression to identify which factor(s) correlated with the presence of acute peritonitis and to determine if peritonitis could be reliably predicted. RESULTS The operative findings for control and study patients are listed in Table L Acute peritonitis was confirmed at laparotomy in 32 of the 45 patients (71%). Twenty-two of the patients with confirmed peritonitis (69%) had either a perforated viscus (n = 14) or ischemic bowel (n = 8), and the other 10 patients had inflammation of a hollow viscus (nonperforated appendicitis, Crohn's disease, and salpingitis). Lavage results are also presented in Table I. The lavage effluent from uninfiamed peritoneum in control subjects contained small numbers of WBCs (range: 3 to 20 cells/ram3), polymorphonuclear (PMN) cells (range: 5% to 10%), and RBCs (range: 3 to 447 cells/mm3), minimal protein (0 to 0.5 g/dL), a low concentration of amylase (range: 0 to 6 1U/L), and no bilirubin. In con. wast, the lavage fluid from patients with confirmed peritonitis typically had elevated numbers of WBCs (range" 1 to 24,550 cells/ram3), an increased percentage of PMN cells (range: 10% to 100%), a variable number of RBCs (range: 10 to 200,000 cells/mm3), and variable concentrations of protein (range: 0 to 6.7 g/dL), amylase (range: 0 to 3,532 IU/L), and bilirubin (range: 0 to 1.9 mg/dL). Lavage results were analyzed by multiple logistic regression, and the total number of leukocytes per cubic millimeter was the only factor predictive of peritonitis (p 200 = 99% probability of acute peritonitis
a result of the accuracy and reliability of physical examination in most patients with abdominal pain. A recent report by Adams et al [13] described a 25% negative laparotomy rate in patients with acute abdominal pain assessed clinically. In the current study, a clinical diagnosis of peritonitis was not confirmed by intraoperative findings in 13 (29%) of the 45 patients. Although DPL might improve the accuracy of physical examination in all patients with suspected peritonitis, this application is impractical. DPL may be most beneficial in patients with equivocal or unreliable physical examinations. As demonstrated by Veith et al [4] and confirmed in the current study, the DPL leukocyte count in control subjects without acute peritonitis is low, ranging from 0 to 20 cells/mm 3. This finding increases the probability that an elevated peritoneal WBC count may be diagnostic of peritoneal inflammation. At present, there are no standards that allow for the interpretation of the results of DPL in patients with acute peritonitis despite wellaccepted criteria for DPL in injured patients. Early studies used the trauma criterion of 500 W B C / m m 3 to designate a positive lavage [2,3,5,14]. In a retrospective study, Alverdy et al [9] considered positive those lavages with a WBC count of 250 to 500 cells/mm 3 if 75% or more PMN cells were present. The study described herein is the first prospective evaluation to characterize lavage fluid in patients with peritonitis. These patients typically have an increased number of WBCs in lavage fluid. A WBC count greater than or equal to 200/ram 3 was the only parameter associated with a 99% probability of acute peritonitis. Monocytes are the predominant cell type in normal peritoneal fluid. An increasing predominance of neutrophils is seen with acute peritoneal inflammation. Stewart and associates [!5] demonstrated that the percentage of ncutrophils in the leukocyte differential was of diagnostic significance. However, the percentage of PMN cells in
the leukocyte differential in this study did not yield a higher probability of peritonitis than that obtained with the leukocyte count alone. There are clinical situations in which the sensitivity of DPL may be decreased. Steroids diminish the inflammatory response and may affect the number of leukocytes in the peritoneal fluid. Five patients with clinical peritonitis in this report were receiving steroids at the time of lavage. Two patients, one with focal inflammation of the mesoappendix and another with a well-localized abscess, had leukocyte counts below 200 cells/mm 3 in the lavage fluid. In contrast, each of the three patients with lavage leukocyte counts greater than or equal to 200 c e l l s / r a m 3 had massive peritoneal soilage and inflammation secondary to gastrointestinal perforations. Thus, the degree of peritoneal soilage may influence the results of DPL in patients receiving corticosteroids. Although the use of antibiotic therapy was not investigated in this report, it is conceivable that these drugs may alter the peritoneal response to infection and also decrease the number of WBCs in the peritoneal fluid. This study suggests that DPL is a rapid, accurate, and safe method of assessing the abdominal cavity in patients with suspected acute peritonitis. The leukocyte count of the lavage fluid is increased in acute peritoneal inflammation and appears to be the most accurate predictor of peritonitis. A value greater than or equal to 200 leukocytes/mm 3 is the main criterion for a positive DPL result. REFERENCES 1. Norwood SH, Civetta JM. Abdominal CT scanning in critically ill surgical patients. Ann Surg 1985; 202: 166-75. 2. Richardson JD, Flint LM, Polk HC Jr. Peritoneal lavage: a useful diagnostic adjunct for peritonitis. Surgery 1983; 94: 826-9. 3. Hoffman J, Lanng C, Shokouh-Amiri MH. Peritoneal lavage in the diagnosis of acute peritonitis. Am J Surg 1988; 155: 359-60. 4. Veith FJ, Webber WB, Karl RC, Deysine M. Diagnostic perito-
THE AMERICAN JOURNAL OF SURGERY VOLUME 164 NOVEMBER 1992 451
LARSON ET AL
toneal abscess and the accuracy of the DPL in these situations because I suspect that, as in a case of ours, the DPL is limited in detecting retroperitoneal inflammatory processes. Ernest E. Moore (Denver, CO): With the revitalized interest in gut bacterial translocation, did you measure endotoxin in the lavage fluid or use Gram's stain to identify bacteria? Second, did you examine specific markers of intestinal disruption such as alkaline phosphatase or diamine oxidase? Finally, have you assessed potential inflammatory markers in the peritoneal fluid, such as phospholipase A2 or receptor expression on neutrophils? Steve Smith (Oakland, CA): Could you comment on whether any patients in the miscellaneous group were assessed for postoperative complications of recent intraabdominal surgery? Also, would you choose not to operate under any circumstances if a patient had a clinical presentation of peritonitis and a WBC count of 200 cells/ mm a by DPL? If so, would you repeat the peritoneal 7. 12. Root HD, Keizer PJ, Perry JF. The clinicaland experimental lavage at a later time, and what other studies would you aspectsof peritonealresponseto injury.Arch Surg 1967;95:53 I-7. do to assess the patient's condition? 13. Adams ID, Chan M, Clifford PC, et aL Computer aided David Easter (San Diego, CA): Was a general anesdiagnosisof acute abdominal pain;a multiccnterstudy.B M J 1986; thetic used in all cases? What was your standard for 293: 800-4. determining positive findings (i.e., was it a visual inspec14. Barbee CL, GilsdorfRB. Diagnosticperitoneallavagein evalu- tion at laparotomy)? If so, why not add diagnostic lapaating acute abdominal pain. Ann Surg 1975; 181: 853-6. roscopy instead of lavage? 15. Stewart RJ, Holloway LJ, IsbisterW H . PeritonealneutrophiFrank A. Larson (closing): We have undertaken the lia:a potentialindicatorof the surgicalacute abdomen. Aust N Z J second phase of this study in which we are using the Surg 1984; 54: 565-8. criterion for a positive DPL developed in the first phase to test the accuracy of DPL in patients with suspected acute DISCUSSION June E. Heilman (Pocatello, ID): This study should peritonitis, We have performed lavages both in the emerhelp to more clearly define some of the indications for gency department and in the intensive care unit. Alsurgery in patients with abdominal pain who are difficult though we have just undertaken this second phase, the to examine and are fairly ill. This is the lowest number of early results are encouraging. DPL was performed in all white blood cell (WBC) counts of any of the studies using of the initial patients in the operating room since this was diagnostic peritoneal lavage (DPL) to indicate peritoni- part of the study protocol. Dr. Schauer, there were two tis. Perhaps it will help to more precisely identify patients patients who had retroperitoneal abscesses--one had a with peritonitis. I don't think that the four false-negative positive lavage with a WBC count greater than 200 cells/ results are surprising, considering that these were all lo- mm a, and the second patient had a negative study, in that calized processes. It was interesting to note that of the 10 his lavage WBC count was less than 200 cells/mm3: This cases of appendicitis, 6 were nonperforated. DPL does was the patient who had been receiving antibiotics postidentify some localized processes. Thirteen of your pa- operatively. Regarding Dr. Moore's question about endotients with negative laparotomy studies had WBC counts toxin, Gram's stains, and other markers, we did examine ranging from 0 to 196/mm 3 in the effluent. Did the the Gram's stains of the lavage fluid in just over half of majority of these patients have localized signs or diffuse these patients, and we found that the patients who had tenderness? Will these data alter your care in the future? positive Gram's stains all had WBC counts greater than Could you identify any patients who would not need or equal to 200 celis/mm 3 in the lavage fluid. No other laparotomy by relying on the results of the DPL? Patients markers in the lavage fluid were studied. We do not who have had previous laparotomies, especially potential- believe that other markers would be as readily obtainable ly septic patients in an intensive care unit, are difficult to in the community hospital setting, and we hope that this evaluate. Have you used DPL in any of these patients? will be a technique that can be employed relatively quickly at the bedside. Regarding the question on the patients Ray Ryan (Wichita, KS): In your study, all D P L s were performed in the operating room. Could this proce- in the other category, this included three patients with Crohn's disease, one patient with acute gallbladder disdure be done in the intensive care unit? Philip R. Schauer (San Antonio, TX): Dr. Larson, I tention secondary to hemobilia, and one patient with would like to know if any of your patients had a retropcri- acute salpingitis.
neal lavagein acute abdominal disease:normal findingsand evaluation in I00 patients.Ann Surg 1967; 166: 290-5. 5. BaileyRL, Laws HL. Diagnosticperitoneallavagein evaluation of acute abdominal disease.South Mcd J 1990; 83: 422-4. 6. Lobbato V, Cioroiu M, LaRaja RD, Rothenberg RE, Strand M. Peritoneallavage as an aid to diagnosisof peritonitisin debilitated and elderlypatients.A m Surg 1985; 51: 508-10. 7. Evans C, Rashid A, Rosenbcrg [L, PollackAV. An appraisalof peritoneallavage in the diagnosisof the acute abdomen. Br J Surg 1975; 62: 119-20. 8. Perry JF, Strate RG. Diagnostic peritoneal lavage in blunt abdominal trauma: indications and results.Surgery 1972; 71: 898-901. 9. Alverdy JC, Saunders J, Chambcrlin W H , Moss GS. Diagnostic peritoneal lavage in intra-abdorninalsepsis.A m Surg 1988; 54: 456-9. 10. Root HD, Hauser CW, McKinley CR, LaFave JW, Mendola RP. Diagnostic peritoneallavage. Surgery 1965; 57: 633-7. I I. Advanced trauma lifesupport student manual. Chicago, IL: Committee on Trauma, American Collegeof Surgeons, 1989:113-
4,52 THE AMERICAN JOURNAL OF SURGERY VOLUME164 NOVEMBER1992
