Computed Tomography and Ultrasonography Do Not Improve and May Delay the Diagnosis and Treatment of Acute Appendicitis FREE

ALTHOUGH the treatment of acute appendicitis is simple and straightforward, its diagnosis remains a challenge, and the negative appendectomy rate in large series ranges from 15% to 33%.^1- 5 Furthermore, in the patients with either atypical history or equivocal physical examination findings, particularly in women of childbearing age, the negative appendectomy rate has been as high as 45%.⁴ With an annual rate of 250 000 cases in the US and an incidence of 86 for every 100 000 persons worldwide,^6- 7 acute appendicitis is a common acute abdomen condition in the emergency department (ED). During the last decade, studies of white blood cell scan, ultrasonography (US), and computed tomographic (CT) scan have suggested that these diagnostic imaging modalities may improve the diagnostic accuracy for acute appendicitis.^8- 11 On the other hand, there were also concerns about the appropriateness and accuracy of these modalities without a surgical evaluation.^10- 11

At our institution, history and physical examination, followed by pathologic confirmation after appendectomy, typically diagnose acute appendicitis. During the past 10 years, there has been increasing use of imaging modalities before obtaining surgical consultation in our ED and selective use of diagnostic laparoscopy in the management of acute abdomen. In this report, we reviewed the current epidemiology of acute appendicitis in a tertiary care center and assessed the impacts of US, CT scan, and laparoscopy on its management, with the hypothesis that CT scan and US do not improve the overall diagnostic accuracy for acute appendicitis.

We conducted a retrospective review of 766 consecutive patients undergoing appendectomy for suspected acute appendicitis from January 1, 1995, to December 31, 1999, at University of California Davis Medical Center, Sacramento. Patients who had incidental appendectomy as part of another procedure were excluded. We aimed to study the epidemiology of the disease, its clinical presentation, and the role of laboratory tests and imaging studies in its management. The time spent in the ED for evaluation was defined as the time from admission in the ED until the time when the surgical consultation was obtained. The timing of appendectomy, intraoperative findings, final pathologic diagnoses, and postoperative complications are also recorded. A small subgroup of our patients was treated with laparoscopic appendectomy, and we aimed to review its impact.

Data are reported as mean ± SD. Statistical analysis for significant differences between groups was done using a 2-sample t test for continuous variables and the Fisher exact test for discrete variables. Differences were considered significant at P<.05.

From January 1, 1995, to December 31, 1999, 424 men (55.4%) and 342 women (44.6%) had appendectomy for suspected acute appendicitis at University of California Davis Medical Center. Figure 1 shows the age distribution, with approximately a quarter of the patients either younger than 10 years (15.7%) or older than 50 years (8.2%). The male to female patient ratio was 1.2:1. The negative appendectomy rate was 15.7%, with a significantly higher negative rate in women than in men, 24.6% vs 8.5% (P<.001). The negative appendectomy rates were also higher in young adults (20%). Patients younger than 10 years had a negative appendectomy rate of 9.7%, and patients older than 50 years, 14.3% (Table 1).

The incidence of perforation was high in the very young and in the older age groups, 20.4% for those younger than 10 years and 28.6% for those older than 50 years (Table 1). Seventeen percent of men (72/424) and 11.4% of women (39/342) had perforated appendicitis. The perforation rate was not a function of a delay in evaluation, as the duration from the onset of pain to ED evaluation in perforated appendicitis was 2.9 ± 3.5 days compared with 2.4 ± 7.9 days for nonperforated cases (P = .28). The mean ED evaluation was 5.8 ± 9.5 hours for perforated appendicitis and 5.0 ± 3.4 hours for nonperforated appendicitis (P = .38). Perforated appendicitis resulted in a higher complication rate than did nonperforated appendicitis (19.8% vs 6.1%, P<.001).

Patients with postoperative complications had longer ED evaluations (8.0 ± 12.7 h) than did those without (4.8 ± 3.3 h) (P = .04).

Abdominal pain was the most common presentation of acute appendicitis, reported by 99% of our patients. A history of classic migratory pain (initial periumbilical or epigastric pain localizing to the right lower quadrant) provided the highest positive predictive value (91%) for patients clinically suspected of having acute appendicitis . In patients found to have perforated appendicitis, 23.4% presented with diffuse or bilateral lower abdominal pain, compared with only 11.9% of those with nonperforated acute appendicitis (P = .002). Nausea occurred in 81.7% of the patients, anorexia in 72.4%, and emesis in 67.7%. Fever was reported in 44.9% and chills in 27.3% of patients with acute appendicitis, with similar incidence in all stages of appendicitis and in alternative diagnoses, such as mesenteric adenitis, gastroenteritis, pelvic inflammatory disease, and gynecologic disorders. Leukocytosis (>12 × 10⁹/L) provided the second-best positive predictive value (90.1%). Patients with perforated acute appendicitis had a higher degree of leukocytosis (16.4 ± 5 × 10⁹/L) than did those with nonperforated acute appendicitis (14.7 ± 4.6 × 10⁹/L, P = .002). Many patients with acute appendicitis had no leukocytosis, and 65.3% of these patients had perforation. Conversely, patients with mesenteric adenitis had slightly higher degrees of leukocytosis (16.5 ± 6.5 × 10⁹/L) than did those patients with nonperforated acute appendicitis. Patients with gastroenteritis had degrees of leukocytosis similar to those of patients with acute appendicitis (14.9 ± 6.4 × 10⁹/L). In 84.9% of patients, repeated white blood cell counts were lower than the initial values obtained in the ED, suggesting that repeated counts may not be clinically useful and may be misleading. Evaluation by surgeons correctly diagnosed 536 of 646 patients with acute appendicitis and correctly ruled out the condition in 38 of 120 patients who did not have acute appendicitis. Overall, clinical assessment yielded an accuracy of 74.9% (Table 2).

Abdominal roentgenograms were not helpful in the diagnosis of acute appendicitis, with 96.3% of all abdominal films showing either normal or nonspecific findings. Ileus and small-bowel obstruction were the 2 most common abnormalities detected, 12 patients had fecaliths, and 1 had free air. Of the 47 patients who had undergone CT imaging, there were 6 false-positive test results and 6 false-negative test results, giving CT scan an overall accuracy of 74.5%. Ultrasonography had a false-negative rate of 76.1% (118 of 155 patients), correctly identified 65 of 183 patients with acute appendicitis, and ruled out 37 of 52 patients with normal findings, having an accuracy of only 43.4%. In patients clinically suspected of having acute appendicitis, neither CT scan nor US improved the diagnostic accuracy compared with clinical assessment alone (Table 2).

Most patients (93.5%) were initially evaluated in the ED. Primary care physicians or outlying hospitals directly referred the rest. Seventy-four patients (9.7%) had been previously evaluated and sent home by their ED or primary care physicians. Of these, 11 patients had been seen twice and 3 patients, 3 times. Most of these patients returned to the ED within 48 hours of their previous evaluation, 67.6% within 24 hours and 21.6% within 48 hours. Emergency department evaluation took a mean ± SD of 5.2 ± 5.4 hours for all patients. It was significantly longer if either US or CT scan was obtained (6.4 ± 7.4 h and 7.8 ± 10.8 h, respectively). When no imaging studies were obtained, ED evaluation took only 4.7 ± 5.2 hours. The addition of abdominal plain x-ray films added minimal time required for ED evaluation (4.9 ± 2.9 h, Table 3). Of the 148 patients who were hospitalized and observed instead of having immediate appendectomy, 30 patients (20.3%) had US, 18 (12.2%) had CT scan, and 108 (73%) had repeated laboratory tests. Twenty-one patients (14.2%) eventually had laparoscopic appendectomy, and the rest had open appendectomy. Despite observation and additional diagnostic tests, the negative appendectomy rate in the whole observation group was 25.7%.

Exploration was carried out through a standard right lower quadrant incision in 86.2% of the patients, a midline incision in 3.8%, and laparoscopy in 10%. All patients received preoperative antibiotics; however, postoperative antibiotic treatment regimens varied and were based on the attending physician's preference. Table 4 lists the final pathologic conditions, with a 1% incidence of malignancy.

One hundred thirty patients had a normal-appearing appendix at the time of exploration. Of these, 10 (7.7%) had early appendicitis, 10 (7.7%) had mesenteric adenitis, and 40 (30.8%) had no other pathologic condition. Twenty women (15.4%) had pelvic inflammatory disease, and 19 (14.6%) had other gynecologic disorders. Seventy-two percent of men and 30% of women with a normal-appearing appendix at the time of exploration had gastroenteritis or other gastrointestinal disorders.

There was 1 death (0.1%) in our series, a 67-year-old woman who was seen once for abdominal pain and sent home (Table 5). When she was reevaluated in the ED, she was admitted and observed for another 4 days before surgical consultation. At exploration, she was found to have perforated appendicitis. Thus, in the group of patients older than 50 years, the adjusted mortality for acute appendicitis was 1.7%.

The overall morbidity was 9.1%. Patients with perforated acute appendicitis had a 19.8% morbidity and patients without perforation, 6.4%. Infectious complications accounted for 76.8% of the perioperative morbidity. The overall wound infection rate was 3.5%, higher in perforated cases (9%) than in nonperforated cases (2.9%) (P = .006). The incidence of intra-abdominal abscess was 2.1%, again higher in perforated cases (6.3%) than in nonperforated cases (1.7%) (P = .01). There were 12 respiratory complications, including 4 patients with pneumonia and 2 with pulmonary embolism. Two of the patients with pneumonia had a normal-appearing appendix. Atelectasis that resolved with incentive spirometry was not considered a complication in this study. There was no case of postoperative myocardial infarction or congestive heart failure. Enterotomies were made in 2 patients; 1 of them had a normal-appearing appendix. Two patients developed postoperative small-bowel obstruction. Two others had prolonged postoperative ileus; both had a normal-appearing appendix.

Laparoscopic appendectomy was performed in 76 patients (10%), 62 women and 14 men. In this subgroup, 9 (11.8%) had had a CT scan, 41 (53.9%) had had a US, and 21 (27.6%) were clinically observed before laparoscopic appendectomy. The additional imaging or observation led to a significantly longer time from admission to appendectomy (laparoscopic, 10.4 ± 14.8 h; open, 5.7 ± 8.4 h; P = .008). The negative appendectomy rate was significantly higher in the laparoscopic subgroup (laparoscopic, 42.1%; open, 15.4%; P<.001). The use of laparoscopy, however, provided valuable additional diagnostic information in patients with a normal-appearing appendix. When a normal-appearing appendix was encountered, the underlying abnormality was correctly identified in 81.5% of patients undergoing laparoscopic appendectomy: 9, pelvic inflammatory disease; 8, ruptured or hemorrhaged ovarian cysts; 4, adhesion; and 5, mesenteric adenitis. Of the 6 patients without an intraoperatively identified pathologic source, 3 were found to have early acute appendicitis. Thus, when a normal-appearing appendix was found during laparoscopy, the underlying abnormality was identified in 91.6% of patients. With respect to open appendectomy, the abnormality was correctly identified in only 74% of patients having a normal-appearing appendix. The morbidity of laparoscopic appendectomy was low (1.3%), a single patient with superficial wound infection. The low morbidity may be accounted for by the lower perforation rate (10.3%) in these patients.

There has been a slight change in the epidemiology of acute appendicitis during the past 25 years. In the 1960s and 1970s, age distribution showed 9.1% of the patients with acute appendicitis to be younger than 10 years and 9.9% to be older than 50 years.⁴ At the end of the 20th century, our data showed that approximately a quarter of the patients were either younger than 10 years (15.7%) or older than 50 years (8.2%). The disease has also become more prevalent in women, with the ratio of male to female patients decreasing from 2:1 a quarter of a century ago to 1.2:1 today.

The overall negative appendectomy rate in this study was 15.7%, which is on the low end of previously reported rates of 15% to 33%.^1- 5 Looking at specific subgroups, there was no excessively high negative appendectomy rate for women aged 20 to 40 years, as was reported 25 years ago (24.4% vs 45%).⁴ The low negative appendectomy rate at our institution was not at the expense of a higher perforation rate. The perforation rate in our series was 14.6%, which was also lower than the previously reported rates of 17% to 39%.^1- 4,12 Today, the perforation rate remained higher in patients younger than 10 years (20.4%) or older than 50 years (28.6%), but it was not because of a delay in presentation or diagnosis, as previously suggested.^4,13- 14 Furthermore, in both age groups, the perforation rates were significantly lower than those reported during the last several decades, which ranged from 17% to 59% for the very young and from 37% to 69% for older persons.^{1- 4,13- 14}

History of migratory pain, right lower quadrant tenderness, and leukocytosis remain reliable and accurate diagnostic clues in the patients suspected of having acute appendicitis. History of migratory pain and leukocytosis (>12 × 10⁹/L) had positive predictive values higher than 90%, and nearly all patients with acute appendicitis had localized right lower quadrant tenderness. When acute appendicitis presents in this classic form, it is easily diagnosed and should be treated without the aid of further imaging modalities. In prospective studies, US has had an excellent performance, with a mean sensitivity of 86% and a median specificity of 96% in diagnosing appendicitis.¹⁰ However, US has the limitation of variable reliability and its well-known operator dependency. Although US is noninvasive, the test adds expense and may delay surgical consultation. In our experience, ED evaluation was significantly longer when an abdominal or pelvic US was obtained. Furthermore, US missed the disease in 118 patients who eventually were found to have acute appendicitis despite normal US findings, yielding a false-negative rate of 76.1%. Thus, it appears that US may be most useful in excluding potential pelvic abnormality in equivocal cases, but there are little data available to unequivocally support the benefit of US in the patients with classic clinical signs and symptoms of acute appendicitis.

Routine contrast-enhanced CT scan has also been described as an accurate diagnostic imaging modality for patients with suspected acute appendicitis and equivocal findings. In clinical trials, CT scan has been superior to US in terms of diagnostic accuracy and reliability in atypical presentation of appendicitis, and it can be specific and sensitive in diagnosing classic acute appendicitis.⁹ However, most authors do not enthusiastically endorse the use of CT imaging alone, but instead recommend that CT results be correlated with clinical findings.¹¹ In our study, the positive predictive value for CT scan in patients with clinical findings consistent with acute appendicitis was only 83.8%, a rate that can be achieved solely with good history taking and physical examination. Our data suggest that the next step in managing the patients with suspected acute appendicitis who have a history of migratory pain or leukocytosis is to obtain surgical consultation for appendectomy. Further imaging studies in such patients do not help, and they may significantly prolong ED evaluation and delay the eventually needed appendectomy.

In reality, the classic migratory pain occurred in only one half of the patients with acute appendicitis in our study, 396 patients (51.7%). Thus, accurate and timely diagnosis of acute appendicitis remains challenging, and acute appendicitis still is one of the most commonly misdiagnosed conditions in the ED.¹⁵ Although immediate appendectomy is indicated in the patients with a classic presentation of acute appendicitis, those with an atypical presentation should be hospitalized for observation, serial examination, and additional laboratory or radiographic studies. Although the risk of perforation during this observation is small, an accurate preoperative diagnosis is often not accomplished. We observed and reevaluated 148 patients with suspected acute appendicitis before appendectomy, with additional CT scan or US performed in 49 patients and subsequent laboratory data in 108 patients, but the negative appendectomy rate remained high at 25.7%. In fact, successive white blood cell counts offered little diagnostic value and should be avoided. In such patients, diagnostic laparoscopy may have a role. In our series, 10% of the patients underwent laparoscopic appendectomy. Although a high negative appendectomy rate (42.1%) was encountered, the pathologic source of the patient's symptoms was identified and effectively treated in 91.6% of patients undergoing this diagnostic and treatment modality. A possible factor affecting the high negative appendectomy rate is that laparoscopy was often used in patients in whom diagnostic dilemmas remained despite additional imaging studies, laboratory data, and clinical observation. Given its excellent diagnostic yield, zero mortality, low morbidity (1.3%), and easy availability, one should consider the early use of diagnostic laparoscopy instead of CT scan, US, or observation in the selected patients suspected of having acute appendicitis but with atypical presentation.

Mortality was 0.1% in this study, consistent with low mortalities of 0% to 1.4% reported during the 1960s and 1970s.^1- 5,12 Unlike the morbidities of 21% in the 1960s and 23% in the 1970s,^2,4,11 the total complication rate in our series was only 9.1%. Infectious complications continued to account for most postoperative morbidity following appendectomy, with wound infection accounting for 3.5% and intra-abdominal abscess, 2.1%. Although the overall complication rate has decreased, perforated appendicitis continues to be associated with a significantly higher complication rate than that associated with nonperforated appendicitis.^2,12- 13

In conclusion, the demographics of acute appendicitis in the late 1990s include an increased incidence of patients younger than 10 years and older than 50 years. These 2 age groups continue to have a significantly higher perforation rate and an increased morbidity. Clinical acumen remains the most reliable diagnostic asset for evaluating a patient with suspected acute appendicitis in the ED. In such patients, a history of migratory pain or leukocytosis greater than 12 × 10⁹/L provides high positive predictive values above 90%. These patients should have immediate appendectomy; imaging modalities such as CT scan or US do not improve the overall accuracy, but may delay surgical consultation and the eventually needed appendectomy. The lack of prospective data in our study prevents us from concluding that the early use of diagnostic laparoscopy is a more cost-effective, accurate, and rapid method of diagnosing atypical cases of acute appendicitis. However, its low morbidity and high yield of alternative diagnoses merit prospective, randomized clinical trials of the selective use of laparoscopic appendectomy in the management of atypical acute appendicitis. Such an approach may reduce further the perforation rate and subsequently the mortality and morbidity, especially in women of childbearing age or in the older age group.

Presented at the 108th Scientific Session of the Western Surgical Association, Dana Point, Calif, November 14, 2000.

Corresponding author and reprints: Hung S. Ho, MD, Department of Surgery, University of California Davis Medical Center, 2221 Stockton Blvd, 3rd Floor, Sacramento, CA 95817 (e-mail: hsho@ucdavis.edu).

Wayne H. Schwesinger, MD, San Antonio, Tex: The results that have been projected over 5 years are excellent: one death and a morbidity rate of less than 10%. Still, the 15% rate of negative appendectomies deserves special attention. It seems to be the premise of this paper that we can do better, and I tend to agree that we should not accept the same results that have been in existence for the last 50 years. In fact, technology is available to suggest that we can do better. Unfortunately, this retrospective study found that very technology, CT scans and US, was lacking in 2 major respects. The first is that neither study was very accurate. The second is that both studies delayed surgical intervention. The authors concluded from these data that diagnostic laparoscopy, in fact, may be the preferable study.

The data that they provide clearly document the limitations of US and CT scan in their clinical environment, with accuracies of 43% for US and 75% for CT scan were identified. But I would like to paraphrase one of our still-waiting national presidential candidates: we shouldn't rush to judgment. The indictment of these 2 technologies may be somewhat premature. I say that for one specific reason. The technology is continuing to improve. This perspective leads me to several questions that perhaps will add some light to this issue. First, it appears that there is no radiologist included in the authorship of this paper. Certainly, I don't mean this as a criticism, but it does raise an issue, and that is how active or involved are your radiologists in this particular area? Do they attempt to optimize the studies that you have identified?

Second, and more specifically, how standardized is the CT scan that you are describing? There are numerous authors now who are suggesting that, with the use of thin-section CT scanning isolated to the right lower quadrant, that is, with 5-mm collimation, one can achieve very high levels of accuracy in the diagnosis of appendicitis. In another regard, who was actually performing the US studies? Is it the radiologist, a radiology technician, or someone in the emergency center, and is graded compression used? In the case of women who are at risk for gynecologic diseases, is endovaginal US also available?

Finally, in relation to the delays that seem to impact on how fast patients can get to the operating room, can you not control or minimize delays by using a clinical pathway that defines when you are called to see the patient?

This manuscript raises an appropriate caution about the use of technology, and it directs us to reexamine what the proper role of radiologic imaging is vs laparoscopy.

Jay L. Grosfeld, MD, Indianapolis, Ind: I couldn't judge from the abstract whether the patients who underwent laparoscopy were mainly teenage girls. Can the authors streamline their comments regarding abdominal pain evaluation in teenage girls, a group in whom the negative appendectomy rate traditionally has been the highest? Would they recommend more liberal use of laparoscopy prior to considering some of the other studies (ie, US, CT)? Similarly, I would like the authors to further define what they called the "very young." For instance, pediatric surgeons refer to the very young with appendicitis as patients less than 3 to 4 years of age. The perforation rate in this subset of children may be as high as 75%. This is a much higher rate than noted in older children. What would the authors do for a patient who is evaluated for abdominal pain after a 4-day illness who has a mass in the right lower quadrant? Would you promptly operate on that patient? Would you acquire a CT scan first? Would you consider draining an appendiceal abscess percutaneously with US or CT guidance? We didn't hear much about whether US or CT studies are of value in patients with a delay in diagnosis.

Joe Kuhn, MD, Dallas, Tex: The authors are trying to say what many of us feel when the presentation is typical and unequivocal. In these cases, additional confirmative tests may, in fact, delay the trip to surgery. My question deals with the workup of patients with atypical presentations. Our own recently published experience at Baylor in Dallas involving over 100 patients with uncertain or equivocal findings of appendicitis showed that there was a clear benefit in CT scan, with a resultant 8% negative appendectomy rate. In your experience, since CT scan was used in 47 patients, or about 5%, was CT scan restricted to patients with atypical presentations? Do you feel that the delay in the ED in patients with CT scans was due to their atypical presentations? Then, finally, do you feel that CT scans in your experience may be beneficial in patients with atypical presentations?

Keith W. Millikan, MD, Chicago, Ill: I would like to disagree with the authors for the reason being that I think we are looking at the wrong denominator. The denominator, as they said, is that the patients get their CT scan and US for abdominal pain long before the surgeon is consulted. So, when we are talking about how good we are with appendectomies, the patients who have been diagnosed with gynecologic disease or sigmoid diverticulitis in the elderly by getting a CT scan first are making us look better because our ED physicians have ordered the test and excluded the patients with other diagnoses. The denominator in this study should be the patients who come in with abdominal pain and how have these tests brought the surgeon a fewer number of patients to evaluate for appendicitis.

Edward T. Peter, MD, Redbluff, Calif: I got interested in appendicitis when I was in Fresno. This was prompted by a study by Dr Blaisdell and his colleagues when he was at San Francisco General, where they studied almost 1000 patients. We reviewed a 5-year period of patients with appendicitis, including approximately 700 patients, mostly performed by house staff. We had a negative rate of approximately 5%. I have been appalled at these high negative rates, 15% to 20%, persisting through the years. In my personal experience, for 8 years I have not had a negative appendectomy. You might say that I missed some cases of appendicitis, but I am in a small town and there is only one emergency room. I am certain I would see them again. Therefore, I can't believe I have not been missing cases of appendicitis. As far as the CT scan, I think it can be helpful when you use it selectively. It is most helpful in the female patient who has a retrocecal appendicitis.

Dr Ho: With regard to the caution of not rushing to judgment, it's not the surgeon who rushes to judgment; it's our ED colleagues who might. They decided that they should do these diagnostic procedures before getting our opinion on the patient with acute appendicitis. It is difficult for me to understand the concept that is advocated by some to replace the surgeon's diagnostic skill on a patient with acute appendicitis with imaging tests alone.

Dr Schwesinger also brings up the issue of whether or not the CT scan or US was read by the radiologist. I am very much sure that it was read by the radiologist, perhaps after the fact. We probably did the appendectomy by then. That brings up another point. You don't expect to have dedicated CT scan radiologists in the ED at the time that you will examine the patient. So even if radiologists read the CT scan, they still come out with a 75% accuracy; therefore, you might as well go with your clinical impression and operate on the patient.

Second, who did the US? That is the main problem with US. It is totally operator dependent and, worse yet, it is not performed by a physician. So, the question is: do you trust a surgeon or do you trust the technician in the ED that night? I certainly have my bias.

Last, as far as clinical pathways are concerned, that is absolutely an excellent idea, and the purpose of the paper was to raise some caution about rushing to judgment on using CT scan or US without a clearly defined clinical pathway.

Dr Grosfeld mentioned the atypical group of patients with laparoscopic appendectomy, specifically the teenage girls. We did not look specifically at those subgroups the way you suggest, but that is absolutely one of the indications for using diagnostic laparoscopy. There are many studies about laparoscopic appendectomy vs open appendectomy, and usually most of the authors miss the main point, which is, if you use laparoscopic appendectomy selectively, you will see the most benefit out of that procedure. The rest of the patients, most of the patients, probably should have an open appendectomy and have the problem solved at that time. At UC Davis, we would consider laparoscopic appendectomy, or diagnostic laparoscopy for that matter, if the patient happened to be in the childbearing age group. We do not look specifically at the age group younger than 3 years, but the data can certainly be analyzed.

Dr Kuhn also asked about atypical findings. In the atypical group of patients, CT scan does have a role. According to our data, it is probably best that the patients have a surgical consultation to decide which ones need the CT scan. With that approach, you not only save time but also reduce cost, and you improve the usefulness of CT scan, which is meant for the atypical cases.

Another way to do this is to use diagnostic laparoscopy, but that requires a prospective randomized study comparing CT scan vs laparoscopy in atypical cases before we can have any recommendation.

Dr Millikan disagreed with us on the wrong denominator of our analysis. The point that we wanted to bring up is that, if you use CT scans for all abdominal pain in the ED and then call the surgeon, whereas you know for sure that it is typical of acute appendicitis, not only are you wasting your time, you are wasting the health care resources. Although such practice might make a surgeon look better because they preselected the nonsurgical cases for us, what about the patient they selected out incorrectly and has a missed acute appendicitis?

For the last question from Dr Peter, I congratulate you for a very good negative appendectomy rate, but do you have the data on the perforation rate? If one has ever seen a patient who died or suffered from perforated appendicitis, one will change his or her mind the next time in the ED. In a small town, I agree with you that the patient will come back to you because yours is the only emergency room. In a large urban area, when you send them out, you may not see them again.