A simple scoring system to reduce intraabdominal septic complications after laparoscopic appendectomy

Surg Endosc (2000) 14: 1028–1030 DOI: 10.1007/s004640000281

© Springer-Verlag New York Inc. 2000

A simple scoring system to reduce intraabdominal septic complications after laparoscopic appendectomy A. Serralta, M. Planells, J. Bueno, D. Rodero Servicio de Cirugı´a General y del Aparato Digestivo II, Hospital Universitario “La Fe,”, Avda Campanar 21, 46009 Valencia, Spain Received: 20 October 1999/Accepted: 7 March 2000/Online publication: 7 September 2000

Abstract Background: The development of intraabdominal abscess (IAA) following laparoscopic appendectomy (LA) is associated with significant morbidity. The aim of the present study was to validate an IAA risk score constructed from a previous review of 156 consecutive LA. Methods: The score was tested in 250 subsequent consecutive LA and in patients with a positive risk score. Broadspectrum antibiotics were administered in order to avoid IAA. Results: Factors related to IAA included clinically complicated appendicitis, leucocytosis >15,000/␮l, a difference of >1°C between axillary and rectal temperature, intraoperative findings such as (gangrenes and perforation), and intraoperative perforation of the appendix. In this series, broad-spectrum antibiotic therapy in patients with a positive IAA risk score reduced the incidence of IAA from 7.05% to 1.60%. Conclusion: This policy of identifying high-risk patient via the scoring system and instituting subsequent antibiotic therapy in patients at risk reduces the incidence of IAA following LA. Key words: Laparoscopic appendectomy — Postoperative complications — Scoring system — Intraabdominal abscess

Laparoscopic appendectomy (LA) remains a controversial procedure despite the recent publication of trials comparing laparoscopic and open appendectomy. One of the most troublesome issues is the incidence of postoperative complications associated with this technique. LA is associated with specific postoperative complications not typically encountered in open appendectomy. Although surgical wound infection rates are higher in open appendectomy patients [8], intraabdominal abscess (IAA) rates following LA are

Correspondence to: A. Serralta Serra

reported to be higher than those for open appendectomy [5, 10]; however, a thorough review of randomized trials statistically found no significant differences in IAA incidence rates for the two procedures [6]. The aim of this study was to validate a simple method for identifying LA patients with a significant risk of developing postoperative IAA and to evaluate the use of broadspectrum antibiotic therapy in preventing IAA after LA in high-risk patients as defined by the scoring system.

Patients and methods We prospectively analyzed 406 consecutive laparoscopic appendectomies performed in the period from 1992 to 1998 for acute appendicitis in patients ⱖ14 years older. Definition of IAA included patients who had early or late postoperative abdominal pain and fever associated with fluid collections on ultrasonography. In the preliminary study [11] conducted with the first 156 LA, data were collected prospectively, including gender, age, associated systemic diseases, obesity, white blood cell count, percentage of lymphocytes, clinical presentation (i.e., acute appendicitis vs local or diffuse peritonitis), duration of symptoms (h), axillary-rectal temperature difference of >1°C, preoperative prophylactic antibiotics, anatomical position of the appendix (ileocecal, retrocecal, paracecal, mesoileal, pelvic), inflammation status of the appendix (hyperemic, phlegmonous, gangrenous or perforated), and intraoperative complications (perforation of the appendix during manipulation). Statistical analysis was performed using the Student’s t-test for continuous variables and the chi-square test for categorical variables. Variables with statistical significance for IAA included clinically complicated appendicitis, leucocytosis >15,000/␮l, a difference of >1°C between axillary and rectal temperature, intraoperative findings of gangrenous or perforated appendicitis, and iatrogenic intraoperative rupture of the appendix. These variables were included to obtain the risk prediction system by means of the Spiegelhalter and Knill-Jones method [14], which yields Bayesian positive or negative prediction coefficients based on the presence or absence of the dependent variable (Table 1). Applying the scoring values to the series of 156 patients (group 1) results in a maximum value of 59 for patients with maximum probability of complications and a minimum value of–33 for those with no complications (cutoff value, 0). The system identified 10 of the 11 IAA patients (sensitivity, 0.91) with only 23 false positives (specificity, 0.84) [11]. We tested the scoring system prospectively in the subsequent series of 250 LA for acute appendicitis (group 2). In patients identified at risk for IAA (score value, 0–59), antibiotic therapy was indicated (imipenem+

1029 Table 1. Predictive properties and Bayesian coefficents of significant variables Variable

Sensitivity

Specificity

Coefficient (+)

Coefficient (−)

Complicated appendicitisa Leucocytes >15,000/␮l Axillary-rectal difference T >1°C Advanced appendicitisb Intraoperative perforationc

0.36 0.73 0.55 0.64 0.55

0.87 0.66 0.45 0.36 0.45

10 8 10 12 19

−3 −9 −6 −8 −7

a b c

Clinical findings of generalized peritonitis as opposed to findings of acute appendicitis Intraoperative findings of gangrenous or perforated appendicitis Iatrogenic perforation secondary to inflammation of appendix

cilastatin, 4 g IV/day until 48 h apyrexia, followed by amoxicillin+ clavulanate 1.5 g/day for 7 days). Ultrasonographic monitoring and clinical evaluation were used as final endpoints to assess the efficacy of the system. The Student’s t-test was used to compare continuous data, operative time, and hospital stay. Pearson’s chi-square test with Fisher’s correction when necessary was used to compare operative findings, postoperative complications, and the development of IAA in the two series. The chi-square test with Fisher’s correction was used to compare the percentage of positive risk patients among patients with long hospital stays in both series. A p value of 0.05 was considered significant.

Table 2. Clinical and analytical values for two groups of patients undergoing laparoscopic appendectomy

Age (yr) Leucocytes/␮l Neutrophils (%) Operative time (min)

Results We found 15 IAA cases in a total of 406 laparoscopic appendectomies (overall incidence, 3.7%). There were no cases of postoperative IAA in patients converted to open surgery (26 patients, 6.4% conversion rate). The mean hospital stay for IAA patients was 17 days (mean postoperative stay for the series, 3.88 ± 0.38 days). Data for presenting symptoms, operative time, and hospital stay in groups 1 and 2 are shown in Table 2. Intraoperative findings in both groups are shown in Table 3. There were no statistically significant differences between the two series related to patient distribution, presenting symptoms, or intraoperative findings. There were 11 cases of IAA in group 1 (7.05%). In group 2, 62 patients had a score between 0 and 59 after applying the risk prediction system, and only one case of IAA developed in spite of antibiotic therapy. There were also three IAA cases among the patients who were considered a priori to have a low likelihood of developing IAA (score, 0 to–33). These cases must be considered false negatives. This means that applying the scoring system and instituting antibiotic therapy reduced the development of IAA to four cases (1.60%); by comparison, in group 1 the percentage was 7.05% (p ⳱ 0.004) (Table 4). The 15 patients who developed IAA were given broadspectrum antibiotics (imipenem+cilastatin IV 4 g/day), which resolved the IAA in nine of 15 patients. Two cases required percutaneous US-guided drainage, and four required reoperation. Microbiological analysis showed polymicrobial growth in all IAA that required percutaneous drainage or reoperation. E. coli was the most common pathogen, isolated, followed by streptococcus and bacteroides, P. aeruginosa, and Enterococcus faecalis. Discussion LA has proven advantages over open appendectomy. The risk of wound infection is reduced (with an odds ratio of 2:6

Postoperative hospital stay (days)

Group 1

Group 2

pa

29.7 (24.7–34.7) 15,274.6 (13,804–16,548) 81.4 (76.8–84.1) 56 (49.2–63.8) 4.3 (2.7–5.9)

31.8 (27.9–35.8) 15,699.3 (14,685–16,713) 80.5 (72.2–82.8) 52 (46.6–57.3) 3.62 (2.9–4.3)

0.20 0.32 0.30 0.48 0.07

Results are expressed as mean (95% confidence interval) a Student’s t-test Table 3. Distribution of appendiceal inflammation status

Hyperemic Phlegmonous Gangrenous Perforated

Group 1

Group 2

pa

28 90 25 13

40 (16.0%) 148 (59.2%) 38 (15.2%) 24 (9.6%)

0.589 0.821 0.802 0.523

(17.9%) (57.7%) (16.0%) (8.3%)

Results are expressed as number of patients (percentage) a Pearson’s chi-square test

vs open appendectomy) [6, 8, 10], postoperative comfort is increased, and hospital stay is shorter [1, 6, 8, 10], with early resumption of work or normal activity [3, 4, 6, 10]. However, higher postappendectomy IAA rates have been reported for LA than for open appendectomy [5, 10]. Thus, although the overall morbidity rate appears to be similar following open and laparoscopic procedures in the majority of trials, each of these approaches seems to entail specific potential complications. Wound infections are apparently more frequent following the open approach, whereas IAA is more frequent following the laparoscopic approach. This evidence may lead some surgeons [13] to reject LA as the procedure of choice in acute appendicitis because it appears to involve the potential of a more serious complication than wound infection (according to the classification of postoperative complications [2], wound infections are grade IIa complications whereas IAA are grade IIb). The reported incidence of IAA in conventional appendectomy ranges from zero to 2.8% [9, 12], reaching 6.6% in patients with advanced acute appendicitis (gangrenous or perforated). IAA rates in LA increase to 11% in patients with perforated appendicitis [15].

1030 Table 4. Incidence of septic complications (IAA and postoperative fever) and risk score predictions Group 1

Long hospital stayc IAA Wound infection Feverd Conversions

Group 2

n ⳱ 156

Score >0

n ⳱ 250

Score >0

pa

pb

33 11 3 3 16

20/33 (61%) 10/11 (91%) — 1/3 (33%) —

40 4 1 20 10

27/40 (67%) 1/4 (25%) — 12/20 (60%) —

0.180 0.004 0.561 0.013 0.012

0.540 0.033 — 0.560 —

(21.1%) (7.05%) (1.9%) (1.9%) (10.3%)

(16.0%) (1.6%) (0.4%) (8.0%) (4.0%)

a

Comparison of groups 1 and 2; chi-square test with Fisher’s correction if necessary Comparison of patients with score >0; chi-square test with Fisher’s correction c Patients with hospital stay >5 days d Patients with hospital stay >5 days for fever or abdominal pain but IAA not detected b

In our series, as has also been documented in other studies [12, 15], patients with advanced acute appendicitis (gangrenous and perforated) had the highest risk of IAA. Therefore, our scoring system used the following predictive factors: advanced symptoms, marked difference between axillary and rectal temperature, gangrenous or perforated appendicitis, and intraoperative perforation of the appendix; these are all synonymous with advanced appendicitis. Using a simple mathematical calculation based on variables that are easy to collect, this system allows patients to be classified quickly into either the high-risk or the low-risk group for IAA soon after LA. This classification enables high-risk patients to receive early and evidence-based treatment with broad-spectrum antibiotics. Therefore, since the incidence of advanced forms of appendicitis was similar in both groups, two findings confirm that the scoring system identifies patients at risk and that antibiotic therapy aborts the development of IAA. First, the incidence of septic complications—that is, IAA and postoperative fever—in group 1 (9.0%) was similar to the incidence of septic complications in group 2 (9.6%), indicating the resolution of potentially developing abscesses in group 2. Second, the predicted frequency of patients at risk for IAA was similar for both groups, but the patients in group 1 developed abscess whereas the patients in group 2 did not develop abscess and evinced a clinical pattern of postoperative fever without US collection (Table 4). Consequently, patients identified by the scoring system as at risk for IAA and treated with antibiotics did not develop abscess, and the infectious process was limited to a brief period of postoperative fever. This outcome is reflected in the reduction in the number of long hospital stays and shorter median postoperative hospital stay seen in group 2, results that were also influenced by the lower conversion base. The use of broad-spectrum antibiotics is warranted by the presence of anaerobic and gram-negative microorganisms with polymicrobial growth in IAA cultures and by the fact that postlaparoscopic IAA are often small fluid collections that resolve appropriately with antibiotic therapy and do not require invasive procedures [7, 9]. The power of this risk-prediction system (sensitivity, 0.91 and specificity, 0.84) [11] suggests that an error in data collection may be the source of the three false negative results for the prediction of IAA, especially considering that the scores in two of these cases were very near the cutoff limit. We can therefore conclude that the selective use of

broad-spectrum antibiotic therapy based on the scoring system can reduce the incidence of IAA following laparoscopic appendectomy.

References 1. Attwood SEA, Hill ADK, Murphy PG, Thornton J, Stephens RB (1992) A prospective randomized trial comparing open versus laparoscopic appendicectomy. Surgery 112: 497–501 2. Clavien PA, Sanabria JR, Strasberg SM (1992) Proposed classification of complications of surgery with examples of utility in cholecystectomy. Surgery 111: 518–526 3. Cox MR, McCall JL, Toouli J, Padbury RTA, Wilson TG, Wattchow DA, Langcake M (1996) Prospective randomized comparison of open versus laparoscopic appendicectomy in men. World J Surg 20: 263– 266 4. Frazee RC, Roberts JW, Symmonds RE, Snyder SK, Hendricks JC, Smith RW, Custer 3rd MD, Harrison JB (1994) A prospective randomized trial comparing open versus laparoscopic appendicectomy. Ann Surg 219: 725–731 5. Frizelle FA, Hanna GB (1996) Pelvic abscess following laparoscopic appendectomy. Surg Endosc 10: 947–948 6. Golub R, Siddiqui F, Pohl D (1998) Laparoscopic versus open appendectomy: a metaanalysis. J Am Coll Surg 186: 545–553 7. Kocher HM, Tekkis PP, Payne JG (1999) Abscess after appendicectomy in children: the role of conservative management. Br J Surg 86: 138–139 8. McCall JL, Sharples K, Jadallah F (1997) Systematic review of randomized controlled trials comparing laparoscopic with open appendicectomy. Br J Surg 84: 1045–1050 9. Okoye BO, Rampersad B, Marantos A, Abernethy LJ, Losty PD, Lloyd DA (1998) Abscess after appendicectomy in children: the role of conservative management. Br J Surg 85: 1111–1113 10. Ortega AE, Hunter JG, Peters JH, Swanstrom LL, Schirmer B (1995) A prospective randomized comparison of laparoscopic appendectomy with open appendectomy. Laparoscopic Appendectomy Study Group. Am J Surg 169: 208–213 11. Planells MV, Garcı´a Espinosa R, Lo´pez Martı´nez C, Anaya P, Serralta A, Ballester C, Rodero D (1997) Factores predictivos de complicaciones se´pticas en la apendicectomı´a laparosco´pica. Cir Esp 62: 380– 385 12. Schmit PJ, Hiyama DT, Swisher SG, Bennion RS, Thompson JE (1994) Analysis of risk factors of postappendectomy intra-abdominal abscess. J Am Coll Surg 179: 721–726 13. Slim K, Pezet D, Chipponi J (1998) Laparoscopic or open appendectomy? Critical review of randomized, controled trials. Dis Colon Rectum 41: 398–403 14. Spiegelhalter DJ, Knill-Jones R (1984) Statistical and knowledge based approachs to clinical decision support systems with an application in gastroenterology. J R Stat Soc (A) 147: 35–77 15. Tang E, Ortega AE, Anthone GJ, Beart Jr RW (1996) Intraabdominal abscesses following laparoscopic and open appendectomies. Surg Endosc 10: 327–328 DOI: 10.1007/s004649910052