Appendicitis by CT Scan: Is This the Way to Go?
By John Seashore, MD, and Marc S. Keller, MD
Rao and associates, radiologists at the massachusetts General Hospital, performed appendiceal CT examinations on 100 consecutive patients seen in their emergency department who were considered, on the basis of history, physical examination, and laboratory tests, to have acute appendicitis.1 The ages of the patients ranged between 6 and 75 years. (Rao et al did not tabulate the ages of these patients.) Outcomes were made either by surgery and histological examination in 59 patients or by clinical follow-up two months later in 41 patients.
Fifty-three patients had appendicitis, and 47 patients did not. The interpretations of the appendiceal scans were 98% accurate. The CT results led to changes in management of 59 patients; unnecessary surgery was avoided in 13 patients. Hospital admissions for observation were prevented for 18 patients, and 21 patients did not need admission for observation before necessary appendectomy. Eleven patients did not require admission to the hospital for observation before the diagnosis of other conditions by CT. Rao et al estimate that the cost of CT scans in these patients was $22,800. By preventing 13 unnecessary appendectomies, they estimate a savings of $47,281. By preventing unnecessary hospitalizations for 50 patients, they estimate a savings of $20,250. Thus, the overall savings was $447 per patient.
There is no subject that generates more discussion, often heated, among general and pediatric surgeons than the diagnosis and treatment of acute appendicitis. This paper will surely add fuel to the fire. Appendicitis is the most common surgical disorder of the abdomen, the symptoms are protean and variable, and the disease can humble even experienced surgeons. A technique that might offer rapid and accurate diagnosis deserves careful scrutiny.
COMMENT BY JOHN SEASHORE, MD
This is a small study of 100 consecutive patients. The age range is listed as 6-75 years, but it is not stated how many children were included. This is important since there are significant differences between adults and children in the differential diagnosis of abdominal pain and the approach to evaluating the patient. Diverticulitis, colitis, biliary tract disease, pancreatitis, and other conditions are far more common in adults. Teenage and young adult women are among the most difficult patients to evaluate because of the variety of diseases of the pelvic organs that may mimic appendicitis. Young children may be difficult to examine because they are frightened, but experienced pediatricians and pediatric surgeons can usually overcome these children's fears. The abdominal wall is generally thin, which makes examination more reliable.
Rao et al state that the diagnosis of appendicitis is not correct in 20% of patients, and that 15-40% of patients operated on for presumed appendicitis have a normal appendix. These statistics widely quoted in the literature do not apply to children. A study from Johns Hopkins in 1975 reported that the error in diagnosis rate in appendicitis was reduced to 2% by careful examination and, when necessary, in hospital observation.2 An unpublished study (an epidemiology thesis) from Yale University evaluated various criteria from the history, the physical examination, and laboratory tests and found that the best correlation (93%) with appendicitis was the identification of right lower quadrant muscle spasm by the attending surgeon or senior surgical resident. We have evaluated our results with appendicitis on numerous occasions and consistently find 90-95% accuracy. In Rao et al's, it is not clear which patients actually had surgery and when. If we assume that, without the CT scan, those patients who had a clinical assessment of "definite" or "probable" appendicitis would have had immediate operation, then the negative appendectomy rate would have been 35% (20 of 57), which is far higher than most published reports. Sixteen patients who ultimately were found to have appendicitis were admitted for observation for up to a day before having surgery; this is almost one-third of the entire appendicitis group and, again, seems an unusually high figure. We have a policy of admitting for observation in equivocal cases, but only about 10% of the patients who prove to have appendicitis fall into this category. It is true that significant numbers of patients who do not have appendicitis are admitted for observation, usually for 24 hours or less, but many of those patients are acutely ill with pain, vomiting, and dehydration and could not comfortably go home even if appendicitis is ruled out by CT scan.
The cost savings in this study are predicated on an accuracy rate for CT scans, which is somewhat higher than previous reports, and on a surgical inaccuracy rate, which is far higher than other published data and our experience. It is particularly questionable for children. Clinical evaluation by an experienced surgeon, CT scan, and ultrasound have diagnostic accuracy rates in the range of 90-95%. It is possible that CT scans may be marginally better, but further studies are needed to document this, especially in children.
COMMENT BY MARC S. KELLER, MD
Rao et al note in the introduction that failure to diagnose appendicitis is the most common successful malpractice claim against emergency department physicians. This study could lead to the widespread and unnecessary use of CT scans for the evaluation of all or many patients with abdominal pain before surgical intervention is requested. Rao et al do not report how many patients were seen by the surgical consultant and discharged without having a CT scan. Surgeons are almost never sued for missed appendicitis, probably because it is unusual for a patient to progress to a ruptured appendicitis while under observation by the surgical team.
Appendicitis is a surgical disease. The evaluation and treatment should be in the hands of surgeons, who have the most experience with this common problem. Plain radiographs, ultrasound examinations, and CT scans are all valuable tools to assist the surgeon in evaluating selected patients and may occasionally prompt earlier operation or avoid negative exploration. The decision to use these tools should be made by the surgeon after careful examination of the patient.
Rao et al have a demonstrated wealth of experience in the CT assessment of a general hospital mix of patients with suspected appendicitis. As one who cares only for children, I feel that specific comments must be made directing attention to certain parts of this study, especially if expansion of this method into pediatrics is contemplated.
With regard to the age distribution, the youngest in this series was 6 years. In practice, at times, the occurrence of abdominal pain from appendicitis might uncommonly occur in children as young as 2 years. Persuading the toddler and pre-school child to lie still enough in the perceived threatening CT scanner environment for the first time is unusual; therefore, their CT imaging might require sedation, monitoring, and additional use of physician and nursing resources.
This study reported that appendicitis was the final diagnosis in 53% of patients with abdominal pain. The tabulation and distribution of final outcomes would likely be different in a pediatric population.
Breadth of diagnosis might include enteritis, inflammatory bowel disease, intussusception, ovarian tumors and cysts with or without torsion, Henoch-Schonlein Purpura, hemolytic-uremic syndrome, iliac vein thrombosis, or others. Focusing attention by CT only on the pelvis without the use of intravenous contrast might miss some of these other conditions. Furthermore, the use of enema contrast in every child presenting with abdominal pain seems, as a children's doctor, to be an overly aggressive approach.
One of the other issues in CT imaging of the appendix, well depicted in the illustrations in the article, is the intrinsic tissue contrast created by fat to outline the normal and abnormal appendix. Young children lack much intraabdominal fat, and their scans are certainly harder to read for this purpose.
As a fellow radiologist endeavoring to use imaging to help colleagues and their patients, I am not ready or convinced by the approach proposed by Rao et al. In some instances, pediatric surgeons are quite certain that a child with abdominal pain needs an operation, and, whether the final diagnosis is appendicitis or a mimicking condition, the direct surgical approach is sound and warranted. At other times, the use of imaging consultation may be useful in pediatric care. The approach I have chosen in children is to use lower abdominal sonography as the primary imaging approach when clinicians need assistance. In girls, the internal genitalia are well imaged through the filled bladder and always examined. Ovarian perfusion can be readily assessed by Doppler sonography. Children, who tend to be smaller, thinner, and better subjects for sonography than adults, yield accurate results when examined for appendicitis.
Ramachandran and colleagues reported a 95% accuracy in a four-year experience of 452 children in the sonographic evaluation of abdominal pain for appendicitis.3 Additionally, other conditions, such as those listed above, can be diagnosed without ionizing radiation, use of enemas or drugs, and in a less threatening environment than the CT scanner.
Finally, if I am baffled by a sonographic finding, detect a large intraabdominal process not completely imaged, or feel that the large size of the child is ill suited for ultrasound examination, I would certainly proceed to CT, and this article convinces me further that the reasoning behind that decision is sound. (Dr. Seashore is Professor of Pediatric Surgery, Yale University School of Medicine, and Dr. Keller is Professor of Diagnostic Radiology and Pediatrics, Yale University School of Medicine.)