 |
|
 |
|
 |
 |
|
 |
 |
 |
 |
 |
 |
 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
,||
,||,||
* Departments of Surgery
Pediatrics
Health Services Research, Texas Tech University, Health Sciences Center, Lubbock, Texas
|| Covenant Children’s Hospital, Lubbock, Texas
 |
ABSTRACT |
|---|
 TOP ABSTRACT METHODSRESULTSREFERENCES |
|---|
Methods, Design, Setting, and Participants. Retrospective review of 356 children (mean age: 9.6 years; range: 1–18 years) referred to a regional pediatric surgical center for possible appendicitis from 1999 through 2001.
Interventions. Initial pediatric surgical evaluation consisted of history, physical examination, white blood cell count, differential count, and urinalysis. Children diagnosed with appendicitis underwent appendectomy without additional studies; those with equivocal findings received intravenous fluids, rest, and reevaluation after 4 to 6 hours. Imaging was used selectively by the pediatric surgeon.
Outcome Measures. Sensitivity, specificity, positive predictive value, negative predictive value, and accuracy of the protocol based on final diagnoses; rate of appendiceal perforation; and rate of negative appendectomy.
Results. Of 356 children evaluated for appendicitis, 220 (62%) had an appendectomy. Two-hundred nine (95%) had histologically proven appendicitis, and 11 (5%) had a normal appendix. Of the 209 children with appendicitis, 139 (66%) had acute appendicitis, 34 (16%) had advanced appendicitis without perforation, and 36 (17%) had advanced appendicitis with perforation. Appendectomy was performed after initial evaluation in 195 (89%) of the 220 children and after a period of supportive care and observation in 25 (11%) of 220. One hundred thirty-six children (38%) did not have an appendectomy and were discharged with other diagnoses. The sensitivity of this protocol was 99%, specificity was 92%, positive predictive value was 95%, and negative predictive value was 99%. The accuracy was 97% compared with an accuracy of 82% for ultrasound alone and 90% for CT scan alone.
Conclusions. These data show that a protocol based on clinical evaluation by a pediatric surgeon with selective use of imaging was highly accurate for the diagnosis of appendicitis in children. Low rates of negative appendectomy (5%) and perforation (17%) were achieved without the potential costs and radiation exposure of excess imaging.
Key Words: appendicitis • appendectomy • pediatric surgeon • CT • computed tomography
Abbreviations: CT, computed tomography • US, ultrasound • WBC, white blood cell
Acute appendicitis is the most common surgical emergency in children and adolescents in the United States. In 1999, an estimated 59 000 children <15 years old were diagnosed with appendicitis.1 Despite its frequency, however, the diagnosis of appendicitis in a child is sometimes difficult. Recent reports recommended imaging, particularly computed tomography (CT) with rectal contrast, as the optimal diagnostic study in both adults2 and children.3,4 One protocol used imaging (usually both ultrasound [US] and CT scan with rectal contrast) in 78.5% of children with possible appendicitis.3,5 CT scanning was calculated as cost-effective in children based on a negative appendectomy rate of 23%.6 Because in our west Texas pediatric surgical practice we rely on a clinically based strategy with selective use of imaging, and because we considered a 23% rate of negative appendectomy to be unacceptably high, we undertook the present study. We reviewed the outcomes of 356 children and adolescents referred to us for possible appendicitis over a 3-year period and calculated the accuracy of our diagnostic strategy compared with the accuracy of imaging.
|
METHODS |
|---|
|
TOPABSTRACTMETHODSRESULTSREFERENCES |
|---|
The strategy for diagnosis of appendicitis (Fig 1), agreed on by the 3 pediatric surgeons (who were associated for >20 years), was as follows: initial pediatric surgical evaluation consisted of history, physical examination, white blood cell (WBC) count, differential count, and urinalysis. This was usually done in the emergency department or occasionally on the ward or in the office. When the diagnosis of appendicitis (without perforation) was made, the child was prepared for operation. No additional laboratory or radiographic tests were done. Children with perforated appendicitis received more intensive preparation with intravenous hydration and antibiotics before operation. If a nonsurgical condition was suspected, eg, streptococcal pharyngitis or pneumonia, appropriate diagnostic studies were done, and management was continued by the pediatrician or family practitioner. If the diagnosis remained uncertain, eg, gastroenteritis or early appendicitis, we preferred that imaging (usually US in girls or CT scan in boys) be ordered by the pediatric surgeon. The child received intravenous hydration, nothing by mouth, and was allowed to rest, sometimes with sedation.
|
The medical records of 356 children and adolescents referred for possible appendicitis from January 1999 through December 2001 were reviewed. Patients were identified from our practice database by using the diagnoses of appendicitis, perforated appendicitis, and abdominal pain/possible appendicitis. (Children with abdominal pain who did not have possible appendicitis were not referred and thus were excluded from this patient population.) Incidental appendectomies performed as part of another procedure were excluded. A standardized data collection tool was used that included age, gender, duration of symptoms, county of residence, imaging (US or CT scan), physician ordering imaging studies, results of imaging studies (positive, negative, or equivocal), interval (hours) from arrival to pediatric surgical consultation, interval (hours) from arrival to appendectomy, operative diagnosis, and pathologic diagnosis. Pathologic criteria for acute appendicitis were mucosal and intramural inflammation. The presence of advanced appendicitis, eg, right lower quadrant peritonitis with or without gross appendiceal perforation was based on the surgeon’s operative note. The presence of perforation was based on the pathologist’s report. In children who did not have appendicitis, the discharge diagnosis was recorded. Children who improved under observation were discharged; those who did not return to the hospital were presumed not to have appendicitis. Two reviewers (including A.M.K.) performed >95% of chart reviews, and 2 individuals (including A.M.K.) performed all the data entry. Outliers were double-checked by a second review of the original record. The data were entered into a computer program (Epi Info 2002, Centers for Disease Control and Prevention, Atlanta, GA) for analysis. Significance tests were performed to compare differences between groups. Means were tested by using the Student t or the Kruskal-Wallace test. 
2 tests were used for comparison of categorical variables. Sensitivity, specificity, positive predictive value, and negative predictive value were calculated by standard epidemiologic methods.7 Accuracy was calculated by number of patients with correct diagnoses/total number of patients. Reports from US or CT which were diagnostically equivocal were not included in the calculation of accuracy. The study did not attempt to analyze separate aspects of the pediatric surgical evaluation (history, physical examination, WBC count, differential count, or urinalysis) for determination of rank of importance in the diagnostic process.
|
RESULTS |
|---|
|
TOPABSTRACTMETHODSRESULTSREFERENCES |
|---|
Disposition of the patients after initial pediatric surgical consultation was: operation in 195 (55%) of the 356 patients; observation in 152 patients (43%); and discharge home in 9 patients (2%). Of the 152 children who were observed, 25 were subsequently operated on for appendicitis, and 127 improved under observation (with hydration and supportive therapy) and were discharged from the hospital with a diagnosis other than appendicitis. One of these 127 children returned to the hospital 6 days later with perforated appendicitis and was considered a case of missed appendicitis. The diagnoses of 146 children who did not have appendicitis are listed in Table 1.
|
|
.0001). Median time from arrival at the hospital to appendectomy in the 195 children diagnosed with appendicitis at the initial pediatric surgical evaluation was 5 hours. Median time to appendectomy in the 25 children who had appendectomy after a period of observation was 18 hours. Despite our preference that pediatric surgical evaluation be conducted before imaging, 117 imaging studies (67 US and 50 CT scans) were ordered by the referring physician in the 356 children (33%) before pediatric surgical evaluation. An additional 60 studies (17%) were ordered by the pediatric surgeon. (The physician ordering 5 other imaging studies could not be determined.) US alone was performed in 96 patients; CT scan alone was performed in 54 patients, and both US and CT scan were performed in 16 patients. The sensitivity, specificity, positive predictive value, negative predictive value, and accuracy of the pediatric surgical protocol for US alone and CT scan alone for diagnosis of appendicitis are included in Table 2. We did not calculate accuracy of sequential US and CT scan because only 16 (4.5%) of 356 children had both studies. Fourteen (12.5%) of a total of 112 US examinations and 8 (11.4%) of a total of 70 CT examinations were equivocal and omitted from the calculations. The accuracy of diagnosis was similar in the children with prereferral diagnostic imaging studies (102 [94%] of 109) and children who did not receive such studies (235 [95%] of 247; P = .5452). Table 3 compares the results of 3 recent reports for diagnosis of appendicitis based on imaging3,8,9 with our series based on clinical evaluation and selective imaging.
|
Nearly 40% (39.6%) of children were referred from rural counties outside of Lubbock County. Although mean duration of symptoms was greater in patients from rural counties than in those from Lubbock County (46.0 vs 35.1 hours; P = .0993, using a 1-tailed test), the rate of advanced/perforated appendicitis was not different in rural children than in those from Lubbock County (35.8% vs 29.5%; P = .3328). Complications (wound infection, intraabdominal abscess, and prolonged ileus) occurred in 8 (11%) of 70 children with advanced/perforated appendicitis and in 1 child (wound infection) of 150 (0.7%) with a nonperforated appendix. All children survived.
DISCUSSION
In many centers, imaging for possible appendicitis has become routine. Reports advocating the CT scan with rectal contrast as the gold standard for diagnosis of appendicitis have appeared in the medical,2,3 pediatric,4–6 radiologic,8 emergency medical,10 and even the surgical11 literature and have had an enormous influence on practice. Few reports have questioned the accuracy or wisdom of CT scanning for appendicitis.9,12 Parents of a child with possible appendicitis may request a CT scan because they have read about it in the lay press as the definitive test.13 Evaluation by a pediatric surgeon early in the course of a child with possible appendicitis has rarely been emphasized.
Our data, however, support a diagnostic strategy based primarily on the clinical acumen of a pediatric surgeon rather than imaging. Our sensitivity (99%), specificity (93%), diagnostic accuracy (97%), and negative appendectomy rate (5%) compare favorably with recent reports of imaging-based strategies in both adults and children (Table 3). Our rate of perforated appendicitis (17%) compares favorably to other large pediatric series since 1995, the rates of which have ranged from 15.5% to 47%.4,14–18 Many different factors are associated with the perforation rate, which in general varies inversely with age and directly with duration of illness. Surprisingly, our large proportion of children (40%) from rural counties did not have an increased rate of perforation compared with local children, despite their longer duration of symptoms.
Because appendicitis is an evolving pathologic process, and because early appendicitis may be impossible to differentiate from other causes of abdominal pain in children, clinical reevaluation after a period of observation and supportive care may be necessary. Cost analyses, however, may be biased toward testing to make the diagnosis at first encounter; if appendicitis can be ruled out (by testing), the patient may be sent home from the emergency department. In an urban setting, this strategy may be successful, especially if intravenous fluid resuscitation has been completed during the period of testing. In our rural west Texas population, however, discharge home from the emergency department was rarely an option.
The 25 children who underwent appendectomy after a period of observation represented a small proportion of children who were observed (25 [16%] of 152) and of children who underwent appendectomy (25 [11%] of 220). Although their rate of advanced or perforated appendicitis was higher than that of our 195 children with appendectomy after first evaluation (56% vs 29%; P = .0070), it could not be concluded that the period of observation was responsible for the higher rate of advanced disease. Many confounding factors could have influenced the outcome in these 25 children. For example, the 25 children may have represented a subgroup selected by greater complexity of their clinical presentations. This hypothesis is supported by their high rates of imaging studies, of incorrect or equivocal interpretation of imaging studies, and of negative appendectomy. Theoretically, the outcome of the observed patients might have been improved by more timely surgical decision making and more accurate interpretation of imaging studies. Such refinements might shorten the period of observation and lower the rate of advanced/perforated disease.
Because a missed diagnosis often leads to perforation and complications, rates of negative appendectomy of 12% to 18% are considered acceptable in children.15,19 A recent nationwide study of >261 000 appendectomies in both adults and children reported a 15.3% rate of negative appendectomy20 and emphasized the potential for enormous cost savings by a decrease in this rate. Our clinically based approach, with its low rate of negative appendectomy (5%), may be more cost-effective than other diagnostic strategies. We did not perform a cost-effectiveness analysis in this study because we did not have a comparison group of children who did not receive pediatric surgical evaluation early in their management for possible appendicitis.
Protocols from tertiary medical centers may not be generalizable. For example, an imaging protocol for childhood appendicitis (US followed by CT with rectal contrast) that originated from a large, urban, university children’s hospital (Children’s Hospital, Boston, MA) depended on a coterie of pediatric radiologists with special interest and expertise in the radiographic diagnosis of pediatric appendicitis.3–6 The cost of imaging (US + CT, in 1997 dollars) was reported as $907 per patient.6 Other tertiary centers have chosen opposite strategies with successful outcomes. One large, urban, university children’s hospital (Children’s Hospital Medical Center, Cincinnati, OH), for example, used a clinical, evidence-based pathway for appendicitis in which pediatric surgical evaluation was conducted before any tests were ordered. Imaging was done in doubtful cases. Quality of care, using as indicators the rates of appendiceal perforation (25%) and negative appendectomy (12%), was unchanged with this pathway, and hospital costs were reduced significantly.19 A diagnostic strategy that depends on the clinical acumen of a pediatric surgeon may be more generalizable than one that requires the technologic skill and expertise unique to pediatric radiologists.
Improved technology does not always translate into improved diagnosis and patient outcomes. Weyant et al9 studied 625 patients with appendicitis but found no correlation between CT findings and pathologically proven appendiceal disease. A population-based study from the state of Washington analyzed 63 707 appendectomies performed during a 12-year period (1987–1998), during which great improvements in CT, US, and laparoscopy occurred. Contrary to expectation, however, the incidence of negative appendectomies (15.5%) and perforation (25.8%) did not change with the availability of advanced diagnostic testing.21
A limitation of our investigation is its retrospective format. Despite our preference that imaging studies be ordered by the pediatric surgeon, a CT scan, which some authors now consider as the definitive imaging study,4 was ordered in 50 (14%) of 356 children by the referring physician before pediatric surgical evaluation. We could not determine retrospectively whether such prereferral imaging was helpful or superfluous in making the diagnosis of appendicitis; however, the data suggest that it may have been superfluous, because the accuracy of diagnosis of appendicitis was no better with prereferral imaging (94%) than without it (95%).
Because the children in our study population were prescreened by another physician before referral, they were more likely to have appendicitis than an unscreened population of children with the initial presentation of abdominal pain. Selection thus would account for our relatively high proportion of children with confirmed appendicitis (62%). Selection by prescreening, however, does not affect the study outcomes, ie, the perforation rate or negative appendectomy rate, because the denominator of these outcomes is the number of children who actually undergo operation. Further, selection by prescreening does not affect the epidemiologic measures (sensitivity, specificity, positive predictive value, negative predictive value, and accuracy), because these measures are based on correct diagnoses, not on the proportion of subjects with appendicitis. A population that is prescreened to exclude patients with abdominal pain who have no suspicion of appendicitis is therefore appropriate and was used by Peña et al3–6 in several recent studies on diagnosis of pediatric appendicitis. Selection bias, which typically occurs when nonrandomized groups with different characteristics are compared (incorrectly), is not a factor if an entire population is selected for an observational study such as ours and those of Peña et al.
Childhood CT is not innocuous; recent reports in the radiologic literature have warned of a significant increase in lifetime radiation risk. Investigators from Columbia University, alarmed by a rapidly increasing number of pediatric CT examinations, estimated that 500 individuals currently <15 years old might ultimately die from cancer attributable to the CT radiation.22 A study from Belfast documented that 
50% of pediatric CT examinations at general hospitals failed to adjust the technique for patient age, thus exposing children to an unnecessarily high radiation dose.23 Donnelly et al,24 from Cincinnati, outlined a strategy for adjustment of standard adult CT protocols to control the radiation dose in children. The risk is not theoretical but is based on Japanese data on actual excess cancer in those who were irradiated as children in 1945.25 Because children are 10 times more sensitive than adults to the induction of cancer, Hall26 estimated that an abdominal helical CT scan in a young girl results in a risk of fatal cancer later in life that amounts to about 1 in 1000. The public health problem becomes significant when the small individual risk is multiplied by the 2.7 million of such procedures performed annually.26 Thus a strategy that relies on careful history and physical examination and minimizes radiation exposure may be inherently safer than one that relies on routine imaging. Future investigations should consider the risk of radiation exposure against possible benefits in cost effectiveness or diagnostic accuracy for this common pediatric condition.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
Reprint requests to (A.M.K.) 1273 Par View Drive, Sanibel, FL 33957. E-mail: akosloske{at}hotmail.com
|
REFERENCES |
|---|
|
TOPABSTRACTMETHODSRESULTSREFERENCES |
|---|
This article has been cited by other articles:
 |
|
 |
|
  A. A. Miller CT Scans on Children: Is This a Problem? Clinical Pediatrics, April 1, 2008; 47(3): 220 - 223. [PDF]
|
|
|
|
 |
|
 R. H. Gottlieb Imaging for Whom: Patient or Physician? Am. J. Roentgenol., December 1, 2005; 185(6): 1399 - 1403. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. Swarr and R. Keren Comparison of Alternative Diagnostic Approaches for Managing Appendicitis in Children: The Effect of Disease Prevalence and Spectrum Pediatrics, August 1, 2004; 114(2): 513 - 514. [Full Text] [PDF]
|
|
|
|
 |
|
 Clinical Assessment, Limited Imaging Diagnose Appendicitis Journal Watch Dermatology, May 25, 2004; 2004(525): 8 - 8. [Full Text]
|
|
|
|
|
|
Diagnosing Appendicitis in Children: What's the Best Strategy? Journal Watch Dermatology, March 3, 2004; 2004(303): 9 - 9. [Full Text]
|
|
|
|
 |
|
 Clinical Assessment, Limited Imaging Diagnose Appendicitis Journal Watch Pediatrics and Adolescent Medicine, March 1, 2004; 2004(301): 1 - 1. [Full Text]
|
|
|
|
 |
|
 Diagnosing Appendicitis in Children: What's the Best Strategy? Journal Watch (General), February 6, 2004; 2004(206): 4 - 4. [Full Text]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
 |
||
|
||
Home | My Pediatrics | Journal Information | Current Issue | Past Issues | Subscriptions & Services | Contact Us Click here for faster international access © 2004 American Academy of Pediatrics. All rights reserved. |
||
|
||
