Original Contribution

Appendicitis outcomes with increasing computed tomographic scanning

Steven P. Frei MD^a,?, William F. Bond MD^a, Robert K. Bazuro DO^b, David M. Richardson MD^a, Gina M. Sierzega MA^c, James F. Reed PhD^d

^aEmergency Department, Lehigh Valley Hospital–Muhlenberg, Bethlehem, PA 18017, USA

^bEmergency Department, Danbury Hospital, Danbury, CT, USA formerly Emergency Medicine Residency,

Lehigh Valley Hospital–Muhlenberg, Bethlehem, PA 18017, USA ^cHealthcare Research, Lehigh Valley Hospital, Allentown, PA 18105, USA ^dHealth Sciences, Lehigh Valley Hospital, Allentown, PA 18105, USA

Received 13 June 2007; accepted 19 June 2007

Abstract

Purpose: The purpose of the study was to examine appendicitis outcomes over time as computed tomographic scanning was incorporated into practice.

Basic Procedures: Using chart review, appendectomy cases from 1998 to 2004 were analyzed by year for CT scanning rate, delay in treatment, complications, Negative appendectomies, and time to surgery. Delay in treatment was defined as discharge from the ED at First visit or more than 20 hours from examination until surgery.

Main Findings: Computed tomographic scanning increased from 12.3% in 1998 to 84.4% in 2004. Delay in treatment decreased from 7.8% in 1998 to 3.0% in 2004. Complications decreased from 33.3% in 1998 to 21.3% in 2004. Negative appendectomy rate did not change significantly over time. There was a slight decrease that may have resulted from chances, variation (p=.087) for the line trend. Median time to surgery increased from 250 minutes in 1998 to 426 minutes in 2002, decreasing to 370 minutes by 2004.

Conclusion: During the period when CT scanning increased dramatically, delays in treatment and complications decreased significantly, but negative appendectomy rates decreased only slightly, if at all. Median time to surgery increased.

(C) 2008

Introduction

Recent years have seen changes in the Diagnostic approach to appendicitis. Perhaps the most dramatic has

A part of this study was presented at the Society of Academic Emergency Medicine annual meeting as a moderated poster presentation on May 25, 2005.

* Corresponding author. Fax: +1 610 861 7783.

E-mail address: [email protected] (S.P. Frei).

0735-6757/$ – see front matter (C) 2008 doi:10.1016/j.ajem.2007.06.027

been the use of computed tomographic (CT) scanning. The accuracy of the CT scan [1-7] has led to increased use of this modality in cases of suspected appendicitis. However, the effect on actual outcomes has been controversial. Some previous research studies showed lower rates of negative operations and lower rates of perforation [8-12], whereas other studies revealed no change [13-16]. One author stated that CT use simply delays treatment in many cases and does not improve outcomes [17]. However, Menes et al [18] reported a shorter time to surgery with increased use of CT.

Disadvantages of CT scanning may include patient incon- venience, crowding of emergency departments (EDs) while patients await CT, radiation exposure, and cost.

Our premise was that if recent changes, and in particular, CT scanning, are beneficial, then we should see improve- ments in patient outcomes. If there are no improvements in patient outcomes, then we must question our present approach, including CT scanning. If there is some improve- ment, then we must consider risk vs benefit.

This study examined the time during which the use of CT scanning for appendicitis had increased substantially in our practice. We wished to determine if there was improvement in the outcomes of delayed treatment, negative laparotomy rate, complications (of perforation, gangrene, or abscess), and time to surgery in patients with suspected appendicitis.

Methods

We performed a retrospective chart review for all cases of appendicitis and appendectomy from 1998 through 2004. The study can be considered a comparison of cohorts of appendectomy patients, each based on a given period (1 year), and then looking at the outcomes of delayed treatment, complications, negative appendectomy, and time to surgery. It was during these years that CT scan use had increased dramatically for appendicitis. Our institutional review board granted exempt status and a waiver of requirement for patient authorization because the study incorporated measures to maintain security of records and to delete patient identifiers. The study was performed in a 3-hospital system, including 2 community hospitals (urban and suburban) teaching ED residents and 1 tertiary care suburban teaching hospital with multiple residencies. Yearly ED visits for the system averaged about 100000 during this period. Patients were initially identified by final Diagnosis codes for appendicitis and procedure codes for appendectomy as coded at hospital discharge. This study included all such patients aged 12 to 65 years who were seen in the ED and had an appendectomy. Patients were excluded if they left against medical advice, had missing medical records, or had an incidental or elective appendectomy (predefined by a 3-part rule). If appendicitis was not confirmed, 2 other groups that did not represent true negative laparotomy were excluded: those with other pathology of the appendix (such as carcinoid or cancer) and those in whom another surgical procedure was

done (laparotomy was still therapeutic).

Appendicitis was defined as a diagnosis of acute appendicitis on the pathology report. Statements of “neu- trophils on the surface epithelium” or “secondary appendi- citis” or “periappendicitis” were not considered to be appendicitis. Delay in treatment was defined as discharge from the ED at the first visit (a visit for Abdominal complaints within the 2 weeks before surgery) or a time from initial examination to surgery of 20 hours or more, as suggested by

Graff et al [19]. Complications were defined as perforation, gangrene, or abscess. Complications of perforation or gangrene were considered present only if noted on the pathology report. Abscess was considered present only if noted in the operative report. Negative appendectomy was defined as not having acute appendicitis on pathology and not meeting the exclusion criteria. Time to surgery was measured from the initial physician-level examination until the start of anesthesia for surgery. Computed tomographic scan was recorded as done if it occurred before surgery. Missing times were estimated from the record by interpola- tion from available times.

The standard for CT scanning for appendicitis in our hospital system was the use of both oral and intravenous contrast, with scanning of the entire abdomen and pelvis. The oral contrast used varied between barium and diatrizoate meglumine/diatrizoate sodium and was not recorded. At least 6 different CT scanners were in use at varioUS times during the study period, all helical scanners but varying from single- slice type to 4-slice multidetector type. The exact scanner used for each patient was not recorded.

In our hospital system, data and electronic records are stored in a central database. Initial patient identification was done through the available electronic databases. Charts were reviewed from an electronic record or on paper or microfilm for the time predating the electronic system (before approximately mid-1999). Team members reviewed charts retrospectively and recorded data on standardized forms according to the prespecified Explicit criteria. For each member of the team, the first 25 charts were rechecked by one of the authors, and any necessary retraining was done. All data were entered into a computerized database. Cases were divided into 1-year segments by visit date. This division was arbitrarily chosen to provide an adequate number of patients per segment. Sixty charts were examined by a second reviewer, and a reliability analysis was conducted. Cohen ? statistics were generated for the primary outcome measures. ? statistics ranged from 0.49 to 1.00 for each of the primary outcome measures.

The primary outcomes were delay in treatment, complica- tions (gangrene, perforation, or abscess), negative laparot- omy rate, and time to surgery. This was measured for categorical outcomes by determining the percentage of patients for each year who had these outcomes and for time to surgery by determining the median and mean time to surgery for each year.

Rate data were first analyzed using Pearson ?2. Linear trend analysis was completed using a binomial linear association analysis for CT scanning rate, delay in treatment, complication rate (perforation, gangrene, or abscess), and negative appendectomy rates. Negative cases were used for calculating the negative appendectomy rate but were not included for calculating other outcomes. Analysis of variance with Student-Neuman-Kuels post hoc analysis for time to surgery and a linear trend analysis was also completed. P values of less than .05 were considered significant.

Table 1 Outcome data by year
	1998	1999	2000	2001	2002	2003	2004	P
	(n = 268)	(n = 292)	(n = 272)	(n = 274)	(n = 279)	(n = 331)	(n = 302)
CT scan rate (%)	12.3	21.9	43.8	57.7	66.7	76.1	84.4	.001
95% CI	(8.6-16.9)	(17.3-27.1)	(37.8-49.9)	(51.6-63.6)	(60.8-72.2)	(71.2-80.6)	(79.9-88.3)
Delay ratea (%)	7.8	7.9	8.5	6.6	4.3	3.3	3.0	.008
95% CI	(4.9-11.7)	(5.1-11.6)	(5.4-12.4)	(3.9-10.2)	(2.2-7.4)	(1.7-5.9)	(1.4-5.7)
Negative appendectomy (%)	16.0	10.6	12.1	9.1	12.2	9.7	8.3	.072
95% CI	(11.9-21.0)	(7.3-14.7)	(8.5-16.6)	(6.0-13.2)	(8.6-16.6)	(6.7-13.4)	(5.4-12.0)
Complication rate (+pathology), %	33.3	27.3	25.4	24.9	18.0	20.1	21.3	.002
95% CI	(27.2-39.9)	(20.7-31.4)	(20.0-31.4)	(19.7-30.8)	(13.4-23.4)	(15.7-25.1)	(16.6-26.6)
Time to surgeryb (mean +- SD)	354 +- 305	360 +- 281	385 +- 279	405 +- 251	454 +- 267	434 +- 267	393 +- 233	.001
Median (IQR)	236 (332)	259 (310)	313 (305)	360 (290)	410 (331)	393 (301)	365 (266)	.001
Time to surgeryc (mean +- SD)	527 +- 814	592 +- 1512	568 +- 868	635 +- 1247	740 +- 1916	545 +- 744	485 +- 897	.294
Median (IQR)	250 (509)	293 (401)	352 (395)	385 (335)	426 (355)	405 (320)	370 (286)	.001

Rates are reported as percentage with Clopper-Pearson 95% confidence intervals. Time to surgery is reported as median with interquartile range. SPSS software, version 11.5 (SPSS, Inc, Chicago, Ill), was used for all data analysis.

CI indicates confidence interval; IQR, interquartile range.

^a Delay rate= revisit or >=1200 minutes.

^b Time to surgery when there was no Delay in diagnosis (very delayed cases skew the mean).

^c Time to surgery (includes all cases).

Results

There were a total of 2714 appendectomies performed in our system from 1998 to 2004 for patients aged 12 to 65 years. Records were unavailable for 4 patients. There was 1 patient who left against medical advice and still returned later for appendectomy. Another 610 cases either were not seen in the ED (direct admissions) and/or were incidental or elective appendectomies. This included 12 with elective appendect- omy several weeks after percutaneous drainage of presumed appendiceal abscess. Other pathology such as cancer of the appendix was found in 22 cases. There were 59 cases of negative appendix pathology that required other surgery.

We then had 2018 remaining patients, who comprised our study group. Overall, these patients were 51.6% male, and the mean age was 33 +- 14.2 years. Primary outcome results are summarized in Table 1. The overall CT rate increased

during the study period from 12.3% in 1998 to 84.4% in 2004 (P = .001). The incremental increase was also significant and linear (P = .001; Fig. 1).

The rate of delay in treatment remained constant from 1998 through 2000 and then decreased linearly from 8.5% in 2000 to 3.0% in 2004 (linear trend, P = .001; Fig. 2). The negative appendectomy rate for 1998 through 2004 is shown in Fig. 3. The negative appendectomy rate did not differ by year (P = .072), nor was there a linear trend noted (P = .087). The overall complication rate (perforation, gangrene, or abscess) was 33.3% in 1998 and decreased to 21.3% in 2004 (P = .002). From 1998 through 2001, the complication rates did not reach a statistical difference (P = .054). However, complication rates from 2002 to 2004 were significantly lower than the previous years (P = .001) (Fig. 4).

The mean time to surgery for those who did not meet the definition of delayed treatment increased from 354 +- 305 minutes in 1998 to a high of 454 +- 267 minutes in 2002. The median time to surgery for all cases also increased from 250 minutes in 1998 to 426 minutes in 2002. Both the mean and median time to surgery decreased during the last 2 years of the study. Mean and median time to surgery did not differ statistically in years 1998, 1999, and 2000. However,

Fig. 1 Computed tomographic scanning rate over time. Fig. 2 Rate of delay in treatment.

Fig. 3 Negative appendectomy rate.

2001, 2002, 2003, and 2004 mean times to surgery were higher than the first 3 years (P = .001).

An analysis by sex for overall CT scan rate, delayed treatment rate, negative appendectomy rate, complication rate, and time to surgery shows similar patterns as the overall analysis. Note, however, that the negative appendectomy rate in females is typically twice the rate of males and that the complication rate among males is consistently higher than in females (Table 2).

Discussion

Our study was based on the simple premise that if diagnosis and treatment are improving over time, patient outcomes should improve also. If a test is useful, outcomes should improve as the test is put into practice. For this reason, we examined the time frame when CT scanning was increasingly used for suspected appendicitis. We did not rigidly define what we felt would be a clinically significant improvement but left that for the reader’s judgment.

Our decrease in delayed treatment was rather dramatic, with an absolute decrease of approximately 5%. We are not aware of any previous study of this outcome in association with CT scanning. Because this is likely to be the main area of malpractice risk (erroneously discharging patients with appendicitis), ED physicians may have intuitively used CT scanning more because of this. Our data appear to confirm this improvement in treatment, although we cannot defini- tively say it results from increased CT scanning.

For perforation or complication rates, it is not clear why CT should improve these, unless delayed treatment cases are decreased or unless surgery is done more promptly. We did find a decrease in delayed cases but no decrease in time to surgery. Because of its retrospective nature, our study cannot show if the decrease in complication rate was due totally to CT scanning or to other factors. It could be that patients are simply coming to the hospital sooner in more recent years.

Negative appendectomy rate is the outcome where it seems intuitive that CT would bring improvement. Our data did not show a statistically significant decrease overall, but this could be because of a lack of power because there did seem to be a statistical trend toward a decrease. But even if

there is a small decrease, why is it so minimal? Possible explanations could include some false-positive CT scans (which we had) and occasional pressure on surgeons to operate despite negative CT scans based on clinical course; or perhaps clinical judgment has worsened on cases taken directly to surgery without CT.

Of note, our definition of negative appendectomy may have differed from previous work. We excluded cases where another therapeutic procedure was performed. Because the patient still required surgery with the attendant risks, this was not considered an adverse outcome and was not counted as a negative laparotomy. We believe that including such cases might unfairly favor CT scanning. If an alternate diagnosis is known ahead because of CT, surgery may be planned differently and might not include removal of the appendix. But if CT was not done ahead and the surgery was planned as an appendectomy, the surgeon may be more likely to remove the appendix also in addition to the other required surgery. Because we identified only appendectomy cases, we would then identify more “negative” cases where CT was not done, even if the same number of surgeries was being done using CT. Regarding time to surgery, from our data, it appears that although time to surgery has increased, it is in the range of, at most, 2 to 3 hours extra and may even be lessening in the most recent years. It seems likely that the extra time to perform CT is partially balanced by less observation periods before surgery. Emergency department physicians may also be ordering CT scans more quickly in recent years, so less

time is spent before a CT result is available.

Previous studies have looked at outcomes in various ways. There have been 2 randomized trials of CT scanning vs clinical assessment alone [16,20]. Both were small and came to opposite conclusions. Other studies have looked at planned disposition before CT compared with actual disposition after CT [1,21-23]. These studies showed changes after CT that appeared favorable, but using planned disposition may not be the same as observing real disposition. A third type of study has examined various outcomes in a before-and-after comparison, typically Perforation rate and negative appendectomy rate [8-15]. The results have been divergent. Based on our finding of considerable year-to-year variation, the exact periods selected may have been critical. Flum et al [24] performed

Fig. 4 Complication rate (abscess, gangrene, or perforation).

1998 (n = 268)	1999 (n = 292)	2000 (n = 272)	2001 (n = 274)	2002 (n = 279)	2003 (n = 331)	2004 (n = 302)	P
CT scan rate (%)
Male 15.2	25.2	43.8	56.5	62.2	70.4	77.9	.001
Female 8.9	18.8	43.7	59.1	70.8	83.8	90.4	.001
Delay ratea (%)
Male 8.3	9.1	3.6	7.5	3.7	3.2	0.0	.003
Female 7.3	6.7	13.3	5.5	4.9	3.5	5.7	.039
Negative appendectomy rate (%)
Male 10.3	5.6	9.5	6.8	5.9	6.3	7.6	.022
Female 22.8	15.4	14.8	11.8	18.1	14.1	7.6	.602
Complication rate (%)
Male 38.5	32.1	26.4	24.8	19.7	20.9	24.2	.006
Female 26.3	22.2	24.3	25.0	16.1	18.9	18.6	.425
Time to surgeryb (male)
Mean +- SD 330 +- 291	319 +- 271	343 +- 223	397 +- 259	424 +- 268	390 +- 237	355 +- 218	.007
Median (IQR) 218 (280)	223 (262)	290 (285)	360 (299)	350 (349)	348 (295)	324 (288)	.001
Time to surgeryc (male)
Mean +- SD 469 +- 595	513 +- 984	467 +- 881	629 +- 1371	712 +- 2235	503 +- 770	355 +- 218	.234
Median (IQR) 235 (486)	245 (326)	310 (313)	385 (320)	357 (365)	352 (311)	324 (288)	.001
Time to surgeryb (female)
Mean +- SD 386 +- 323	403 +- 285	437 +- 329	414 +- 242	487 +- 262	499 +- 295	430 +- 242	.027
Median (IQR) 264 (389)	300 (335)	352 (382)	362 (308)	440 (268)	452 (292)	395 (275)	.001
Time to surgeryc (female)
Mean +- SD 685 +- 1257	664 +- 1792	746 +- 1150	731 +- 1576	729 +- 1379	604 +- 705	604 +- 1213	.955
Median (IQR) 475 (570)	330 (469)	390 (665)	375 (344)	445 (322)	467 (368)	409 (339)	.010
a Delay rate = revisit or >=1200 minutes. b Time to surgery when there was no delay in diagnosis. c Time to surgery (includes those cases where there was a delay in diagnosis).

a study similar to ours, looking for improved outcomes (decreased negative appendectomy or perforation rates) during the time CT scanning was introduced. They applied this to a large population and did not find any improvements in outcomes. They did note problems with their data, however [25]. We were able to obtain more detailed and possibly more accurate information by taking data directly from records instead of using administrative data (hospital coding) as they did.

Table 2 Outcome data by sex by year

For the future, we must decide whether the modest improvements in outcomes we found are worth the risks of radiation exposure and ED crowding, as well as the cost. We may have to expose several suspected cases to CT for each case of appendicitis, which multiplies these risks. Perhaps further research can define subgroups where CT scanning is not helpful or where observation is the preferable option.

Our study presents several limitations. The retrospective collection of data is usually less complete and reliable than prospective collection. Chart reviews can present difficulties of interpretation and legibility as well as abstracting error. Also, we did not evaluate other possible outcomes that may have changed over time, such as admissions for observation in patients who did not end up requiring surgery and number of Surgical consultations overall. Several different CT scanners were used during the time of the study, as well as

various radiologists, and these could not be tracked. This could have led to some variations over time. We were unable to determine the accuracy of CT scanning at our institution because of the nature of our study, which did not include all suspected cases.

Changes over time were not limited to CT scanning, and there could have been many confounding variables. We could not control for these, and we did not make an attempt to do so. Our purpose was not to control for every possible variable but merely to determine if outcomes did actually improve during this period. Improvements in outcomes cannot therefore be definitely attributed to CT scanning. Lack of improvement over time cannot definitely be attributed to lack of benefit for CT scanning, either, but this is less likely because it requires us to postulate a confounding variable that would worsen outcomes over time.

We believe that these limitations should not have prevented us from finding clinically significant improve- ments in outcomes. We would expect that advances in treatment should result in improvements in our outcomes over time. We would expect CT scanners and radiologist performance to improve over time, which should result in increased improvement in outcomes. Limitations in data collection should be similar over time and not cause any apparent changes in outcomes.

In summary, we found improvements in outcomes of delayed treatment and complications but minimal changes in negative appendectomy rate and modestly increased time to surgery over the period when CT scanning increased for appendicitis. It may be appropriate to consider more carefully the risks and costs of CT scanning for appendicitis compared with the benefit.

Acknowledgments

The authors would like to acknowledge the following for their assistance: Kathleen Moser (manuscript preparation) and Kristen Koenig and Brian Belmont (data collection).

References

1. Rao PM, Rhea JT, Novelline RA, et al. Effect of computed tomography of the appendix on treatment of patients and use of Hospital resources. N Engl J Med 1998;338:141-6.
2. Stroman DL, Bayouth CV, Kuhn JA, et al. The role of computed tomography in the diagnosis of acute appendicitis. Am J Surg 1999; 178:485-9.
3. Ege G, Akman H, Sahin A, et al. Diagnostic value of unenhanced helical CT in adult patients with suspected acute appendicitis. Br J Radiol 2002;75:721-5.
4. Wijetunga R, Tan BS, Rouse JC, et al. Diagnostic accuracy of focused appendiceal CT in clinically equivocal cases of acute appendicitis. Radiology 2001;221:747-53.
5. Poh AC, Lin M, Teh HS, et al. The role of computed tomography in clinically-suspected but equivocal acute appendicitis. Singapore Med J 2004;45:379-84.
6. Raman SS, Lu DS, Kadell BM, et al. Accuracy of nonfocused helical CT for the diagnosis of acute appendicitis: a 5-year review. AJR Am J Roentgenol 2002;178:1319-25.
7. Ujiki MB, Murayama KM, Cribbins AJ, et al. CT scan in the Management of acute appendicitis. J Surg Res 2002;105:119-22.
8. Harswick C, Uvenishi AA, Kordick MF, et al. Clinical guidelines, computed tomography scan, and negative appendectomies: a case series. Am J Emerg Med 2006;24(1):68-72.
9. Fuchs JR, Schlamberg JS, Shortsleeve MJ, et al. Impact of abdominal CT imaging on the management of appendicitis: an update. J Surg Res 2002;106:131-6.
10. DeArmond GM, Dent DL, Myers JG, et al. Appendicitis: selective use of abdominal CT reduces negative appendectomy rate. Surg Infect 2003;4:213-8 [Larchmt].
11. Regimbeau JM, Panis Y, Le P, et al. Should routine Spiral computed tomography be performed in adult patients with acute abdominal pain of the right lower quadrant? Int J Colorectal Dis 2003;18:126-30.
12. Rao PM, Rhea JT, Rattner DW, et al. Introduction of appendiceal CT: impact on negative appendectomy and appendiceal perforation rates. Ann Surg 1999;229:344-9.
13. Vadeboncoeur TF, Heister RR, Behling CA, et al. Impact of helical computed tomography on negative appendicitis. Am J Emerg Med 2006;24(1):43-7.
14. Perez J, Barone JE, Wilbanks TO, et al. Liberal use of Computed tomography scanning does not improve diagnostic accuracy in appendicitis. Am J Surg 2003;185:194-7.
15. Hershko DD, Sroka G, Bahouth H, et al. The role of selective computed tomography in the diagnosis and management of suspected acute appendicitis. Am Surg 2002;68:1003-7.
16. Hong JJ, Cohn SM, Ekeh AP, et al. A prospective randomized study of clinical assessment versus computed tomography for the diagnosis of acute appendicitis. Surg Infect 2003;4:231-9 [Larchmt].
17. Lee SL, Walsh AJ, Ho HS. Computed tomography and ultrasono- graphy do not improve and may delay the diagnosis and treatment of acute appendicitis. Arch Surg 2001;136:556-62.
18. Menes TS, Aufses Jr AH, Rojas M, et al. Increased use of computed tomography does not harm patients with acute appendicitis. Am Surg 2006;72(4):326-9.
19. Graff L, Russell J, Seashore J, et al. False-negative and false-positive errors in abdominal pain evaluation: failure to diagnose acute appen- dicitis and unnecessary surgery. Acad Emerg Med 2000;7:1244-55.
20. Walker S, Haun W, Clark J, et al. The value of limited computed tomography with rectal contrast in the diagnosis of acute appendicitis. Am J Surg 2000;180:450-4 [discussion 454-5].
21. Rosen MP, Siewert B, Sands DZ, et al. Value of abdominal CT in the emergency department for patients with abdominal pain. Eur Radiol 2003;13:418-24.
22. Rosen MP, Sands DZ, Longmaid III HE, et al. Impact of abdominal CT on the management of patients presenting to the emergency department with acute abdominal pain. AJR Am J Roentgenol 2000; 174:1391-6.
23. Christopher FL, Lane MJ, Ward JA, et al. Unenhanced helical CT scanning of the abdomen and pelvis changes disposition of patients presenting to the emergency department with possible acute appendicitis. J Emerg Med 2002;23:1-7.
24. Flum DR, Morris A, Koepsell T, et al. Has misdiagnosis of appendicitis decreased over time? A population-based analysis. JAMA 2001;286: 1748-53.
25. Flum DR, Koepsell TD. Evaluating diagnostic accuracy in appendicitis using administrative data. J Surg Res 2005;123(2):257-61.