Rapid Responses to:
|
|
Rapid Responses: Rapid Responses published:
-
![[Read Rapid Response]](/icons/misc/sectionDOWN.gif)
Methodological problems exist for RCT in diagnosing appendicitis
- Wai-Ching Leung (17 October 2000)
-
Imaging in the diagnosis of appendicitis
- Peter F Jones (19 October 2000)
-
Heads or tails?
- John Morgan (19 October 2000)
-
Authors' response
- Charles Douglas (23 October 2000)
-
When to operate in the 'intervention' group
- Reyad Al-Ghnaniem (8 November 2000)
-
Diagnostic laparoscopy
- Stephen Attwood (9 November 2000)
-
Re: Diagnostic laparoscopy
- Charles Douglas (17 November 2000)
-
Re: When to operate in the 'intervention' group
- Charles Douglas (21 November 2000)
-
Diagnosing acute appendicitis
- Alice Cutting (30 November 2000)
|
|
|||
|
Wai-Ching Leung, Senior Registrar in Public Health Medicine Epidemiology & Public Health, Newcastle General Hospital, NE4 6BE
Send response to journal:
|
I was interested the read the randomised controlled trial in diagnosing appendicitis by Douglas et al (1), but there are several methodological problems First, the project officer was involved in recruiting the subjects, obtaining informed consent and assessing the Alvarado score. However, he was also solely responsible for the patient randomisation, which he did before assessing the Alvarado score. Hence, the randomisation was not concealed and the assessment of Alvarado score was not blinded. There would be less potential bias if prior randomisation were carried out concealed in sealed envelopes. Second, he informed the admitting teams of the Alvarado score in the treatment group but not the control group. Therefore, the admitting teams in the treatment group received more initial information on the degree of urgency than the control group, which could have influenced their priority in seeing that particular patient and hence the time to therapeutic operation. This is not comparable to the usual clinical situation where the admitting team themselves assess the Alvarado score in the treatment group. Third, the study did not have sufficient power to detect clinical significant differences in the rate of non-therapeutic operations or rate of delayed treatment in association with perforation. By the authors’ own calculation, the study only has 80% chance of detecting 82% reduction in non-therapeutic operation rate (from 11% to 2%). They did not calculate the power for delayed treatment in association with perforation, but it must be very low as there were only two cases in the control group. Clearly, larger trials are needed. Reference 1. Douglas CD, Macpherson NE, Davidson PM, Gani JS. Randomised controlled trial of ultrasonography in diagnosis of acute appendicitis, incorporating the Alvarado score. BMJ 2000; 321: 919 |
|||
|
|
|||
|
Peter F Jones, Emeritus Clinical Professor of Surgery Uiversity of Aberdeen
Send response to journal:
|
Editor - The search for accuracy in the diagnosis of acute appendicitis continues, and Douglas and his colleagues have made a welcome contribution to the debate on the role of imaging (1). Reasons for using ultrasound, and perhaps computed tomography, need to be viewed against the background of the epidemiology of acute appendicitis. When all children admitted with acute abdominal pain were studied in the 1960's it was found, surprisingly, that 30-40% (mostly referred as ?acute appendicitis) settled without treatment. Thiis syndrome, which for 12-24 hours closely resembles acute appendicitis, was named acute non-specific abdominal pain (NSAP) (2), was soon found to occur equally often in adults (3), and is now widely recognised. The knowledge that as many as one-third of patients with acute abdominal pain will prove to have a self-limiting condition must have an effect on management. At the time of admission about one-third of patients immediately show a clear need for emergency surgery. Then the task among the remainder is to distinguish, with the minimum of delay, those with suspicious signs who are developing surgical or medical disease requiring treatment, from those with NSAP. When the frequency of NSAP is not allowed for there is a tendency to regard every doubtful case as a possible perforated appendix, and this still leads to as many as 15-30% of appendicectomies being unproductive. However, if a policy of Active Observation (AO) is adopted experience over 25 years in many centres has shown that repeated bedside examination is a safe method of separating the two groups, and an unproductive appendicectomy rate of 14% in the 1960's had fallen in the 1990's to 3-5% (4). Concern is expressed that delay during observation allows perforation to occur, but most patients with a perforated appendix show marked signs on admission: the few which have been recognised during AO have made a good recovery (4). The existence of imaging (especially CT with its high dose of radiation) is not a reason for using it if a simpler method is effective and safe, and experience with AO suggests that it is rarely required. Douglas and his group and Weyant et al (5) have both found that reliance on ultrasound and CT produces problems from false negative and positive results which have to be resolved at the bedside: both emphasise the central role of "more precise patient selection by clinical criteria" (5). Peter F Jones
1. Douglas CD, Macpherson NE, Davidson PM, Gani JS. Randomised controlled trial of ultrasonography in diagnosis of acute appendicitis, incorporating the Alvarado score. BMJ 2000; 321: 919-22. (14 October). 2. Jones PF. Acute abdominal pain in childhood, with special reference to cases not due to acute appendicitis. BMJ 1969; 1: 284-6. 3. de Dombal FT, Leaper DJ, Staniland JR, McCann AP, Horrocks JC. Computer-aided diagnosis of acute abdominal pain. BMJ 1972; 2: 9-13. 4. Jones PF, Bagley FH. Acute appendicitis. In: Jones PF, Krukowski ZH, Youngson GG,eds. Emergency Abdominal Surgery. 3rd ed. London: Chapman and Hall, 1998: 48-52.. 5. Weyant MJ, Eachempati SR, Maluccio MA, Rivadeneira DE, Grobmyer SR, Hydo LJ et al. Interpretation of computed tomography does not correlate with laboratory or pathologic findings in surgically confirmed acute appendicitis. Surgery 2000; 128: 145-52. |
|||
|
|
|||
|
John Morgan, 4th year medical student Department of Epidemiology and Public Health, University of Newcastle Medical school
Send response to journal:
|
Editor, Investigations into the rapid diagnosis of acute appendicitis are welcomed. In their paper,(1) Douglas et al contribute to this field of research with their study of the use of diagnostic ultrasonography. Although we applaud the use of a randomised control trial, we would highlight several drawbacks in their methodology. Firstly, we question the use of coin tossing as a means of fair randomisation. A different method might have been more appropriate to eliminate possible bias. An example would be the use of a computer generated number system. Secondly, the authors state that randomisation was performed by the Project Officer. This single individual, aware of the group allocation, proceeded to score the patients on the basis of clinical signs and symptoms (Alvarado score). Does this not introduce the potential for bias? The lack of blinding means that the fair treatment of both control and intervention groups cannot be assured. Furthermore, the admitting team (including the ultrasonographers) was informed of the Alvarado scores of patients in the intervention group, but not for the controls. Might this affect the interpretation of the ultrasound results and subsequent time to operation? Thirdly, due to ethical issues surrounding delayed treatment of suspected appendicitis, the surgical team used discretion in the use of ultrasound in both groups. We commend the use of the principle of ‘intention to treat’ to reflect real-life clinical situations. However, in this case, the 18% of patients in the intervention group not receiving the intervention will undoubtedly skew the results. The authors realistically conclude that ultrasound-aided diagnosis does not improve outcome in acute appendicitis. This suggests that clinical judgement is still the key to correct diagnosis. We look forward to further advances that may be of benefit in the rapid diagnosis of this common surgical emergency. John Morgan, 4th year medical student
Department of Epidemiology and Public Health,
The Medical School,
University of Newcastle upon Tyne NE2 4HH
1 Douglas, C.D., McPherson, N.E. et al. Randomised control trial of ultrasonography in diagnosis of acute appendicitis, incorporating the Alvarado score. BMJ 2000; 321:919 |
|||
|
|
|||
|
Charles Douglas, Surgical registrar John Hunter Hospital, Newcastle, NSW, Australia
Send response to journal:
|
The Editor I would like to thank those who made comments on our paper. The following response is made on behalf of my co-authors. We acknowledge that randomisation by coin toss is not the theoretical ideal, but apart from a small difference in age that is clinically insignificant, the groups in this study are well matched. There is no evidence of systematic bias that would have favoured either group. Randomisation was performed in every case by the project officer and co- author, NEM, whose role was as a researcher, not a clinician. We are confident that it was done rigorously. The Alvarado score was used in this study primarily as a means to help select patients in the intervention group for imaging. Nonetheless it did constitute additional information that was (indeed had to be) provided to the admitting team for patients in the intervention group, and was clearly part of the intervention, as we described in the methods section of the paper. It hardly constitutes bias that this score was available to the intervention group, but not to controls. Nonetheless, Wai-Ching Leung raises a legitimate question regarding the role of the Alvarado score in reducing times to therapeutic operation in the trial group (for ‘urgent’cases who did not require ultrasonography). In fact such effect appears to have been littl, if any - the mean times to therapeutic operation for patients with Alvarado scores of 9 and 10 were: intervention 6.5 hours, control 7.0 hours (p=0.74). It is incorrect that 18% of the intervention group did not receive the intervention, as John Morgan and colleagues claim. In fact, only 6 patients in the intervention group (3.8%) failed to follow the diagnostic protocol (see Figure 1 in the paper) and a secondary analysis with these patients excluded still showed no difference between groups. We presume that Mr. Morgan is referring to patients in the intervention group who did not have an ultrasound. However there is already very good evidence from non-randomised studies that ultrasonography is unlikely to benefit those with a very high clinical probability of appendicitis, and we would have been justifiably criticised if we had required ultrasonography for these patients. A study on diagnostic imaging must incorporate some means of selecting patients with appropriate pre-test probabilities. Finally, Dr. Leung is quite right that a larger study is required if small differences in adverse outcomes are to be detected. (In fact, to detect a 25% reduction in adverse outcomes (from 12% to 9%) with 80% power would require recruitment of more than 3000 patients, which would take 10 years at our institution). However it should not be implied that lack of statistical power is the main reason that we produced a negative result – in our study there was not even a small difference between groups with respect to total adverse outcomes, and the non-therapeutic operation rate hardly dropped at all in the trial group. This appears to be largely because surgeons cannot safely trust a negative ultrasound result. |
|||
|
|
|||
|
Reyad Al-Ghnaniem
Send response to journal:
|
Dear Sir - We read with interest the article by Douglas et al.(1)
The article does not describe the algorithm of decision-making with regard
to when to operate in the 'intervention' group. Six patients in the
intervention group were not operated upon despite a positive or equivocal
ultrasound scan (USS) and 19 underwent an operation despite a negative
USS. What was the ultimate criterion for decision-making? If it were
clinical judgement, it would then be difficult to attribute observed
differences (or lack of them) to the intervention protocol.
This study has shown that preoperative USS does not significantly reduce
the NAR or the patients' stay in hospital. The only difference between the
two groups seems to be the cost of routine USS, which was not discussed.
This amounts to US$32,526 (139 scans @ $234) in this particular study and
to a hospital that admits 300 patients per year with suspected
appendicitis the cost is about $70,000 per year.(2) Does routine USS in
acute appendicitis make any clinical or economic sense?
References
1. Douglas CD, Macpherson NE, Davidson PM, Gani JS. Randomised
controlled trial of ultrasonography in diagnosis of acute appendicitis,
incorporating the Alvorado score. BMJ 2000;321:1-7.
2. Axelrod DA, Sonnad SS, Hirschl RB. An economic evaluation of
sonographic examination of children with suspected appendicitis. J Pediatr
Surg 2000;35:1236-41
Reyad Al-Ghnaniem MB, BS; MRCS
|
|||
|
|
|||
|
Stephen Attwood
Send response to journal:
|
Dear Sir - The randomised controlled trial of ultrasonography in the diagnosis of acute appendicitis by Douglas et al (1) and the accompanying editorial (2) correctly highlight the importance of an accurate diagnosis in acute right iliac fossa abdominal pain, and the need to avoid unnecessary appendicectomy. The results of their study indicate that there is really little practical value in the technique due to the combination of false positives, false negatives and the inability to identify alternate diagnoses. It is surprising that neither article mentioned the use of diagnostic laparoscopy in this circumstance. This is an increasingly used approach that makes an accurate diagnosis clearly, especially useful in females of any age and in elderly males, where diagnostic doubt is common. As well as avoiding inappropriate appendicectomy this approach defines the correct operative intervention if an alternative diagnosis requires surgery. If surgery is not required, a definitive management plan is usually clear. A further advantage of laparoscopy is that where surgeons have appropriate training and experience the appendix can be removed laparoscopically with advantages in patient recovery (3). If the abdominal signs are sufficiently clear to indicate focal right iliac fossa peritonism in elderly patients or females of reproductive age there is little to be gained from ultrasonography. Laparoscopy should be undertaken based on the clinical judgement of an experienced surgeon and subsequent surgical intervention proceeded as appropriate. References: 1. Douglas C.D., Macpherson N.E., Davidson P.M., Gani J.S. Randomised controlled trial of ultrasonography in diagnosis of acute appendicitis, incorporating the Alvarado score. BMJ 2000;321:919-922 (14 October). 2. Beasley S.W. Can we improve diagnosis of acute appendicitis? BMJ 2000;321:907-908. 3. Attwood S.E.A., Hill A.D.K., Murphy P.G., Thornton J., Stephens R.B. A prospective randomised trial of laparoscopic versus open appendectomy. Surgery 1992;112:497-501 Mr Stephen Attwood
I have no competing interests or conflict of interest in this work. Authors current appointment: Consultant Surgeon, Hope Hospital, Salford Royal Hospitals NHS Trust, Stott Lane, Salford M6 8HD |
|||
|
|
|||
|
Charles Douglas, Surgical Registrar John Hunter Hospital, NSW, Australia
Send response to journal:
|
The Editor, While laparoscopy may be the best operation for some patients with right iliac fossa pain, it is an operation nonetheless, with well-recognised short and long term morbidity and even mortality. The overall benefits remain unproven. It would take a very large trial to show that 'laparoscopy and proceed' is superior to 'immediate appendicectomy', given the low rate of complications with both operations. However the first dilemma in suspected appendicitis is whether to operate at all. Dr. Charles Douglas
|
|||
|
|
|||
|
Charles Douglas, Surgical Registrar John Hunter Hospital, Newcastle, NSW, Australia
Send response to journal:
|
The Editor, In reply to the question from Reyad Al-Ghnaniem: Surgeons in our study were encouraged to act in the best interests of the patient, taking into account the results of the ultrasound as well as clinical factors. Had we insisted that management be dictated by ultrasound findings alone, at least 3 patients would have suffered serious consequences from non- operative management of gangrenous or perforated appendicitis. In clinical practice, no single diagnostic test determines management on its own. Nor should clinical trials be designed so as to commit the intervention group to potentially harmful management decisions in a way that is completely artificial. If the untreated disease is potentially fatal (as is appendicitis) and the test is known to be less than 100% sensitive (as is ultrasonography) then clinicians must sometimes treat a patient who has had a negative test. “What would have happened if the test result had always dictated management?” is a legitimate retrospective question, but in most cases it is not the ultimate measure of the utility of the test. The value of a diagnostic test depends not only on its accuracy, but also on the consequences of unnecessary or delayed treatment, and whether clinicians can decide when to override a possibly incorrect test result (in the name of safety) without overriding too many correct results. While many studies have shown that ultrasonography is more accurate than unaided clinical diagnosis in acute appendicitis, it has also been established, long before our study, that there is a false negative rate in acute appendicitis of at least 5%. In advocating the use of ultrasonography, many authors seem to have assumed that it is possible to gain the benefit of true negatives (thus reducing the non-therapeutic operation rate) without incurring any harm from the false negatives. Our study suggests that this assumption is incorrect. Dr. Charles Douglas, John Hunter Hospital, Newcastle, NSW, Australia Dr. Jon Gani, John Hunter Hospital and University of Newcastle, NSW, Australia |
|||
|
|
|||
|
Alice Cutting, House Officer Pontefract General Infirmary
Send response to journal:
|
The Editor We read with interest the paper by Douglas et al which described a prospective randomised study of the use of a combination of graded compression ultrasonography and the Alvarado Score in the diagnosis of acute appendicitis. We would concur with the view that the diagnosis of appendicitis is essentially clinical, and the results of this study would appear to support this. The aim of this study seems to have been to provide a "one stop" diagnosis soon after admission. We would advocate the benefits of a good history and repeated review of the patient, with prompt recognition of changes in clinical signs. There is no place for routine use of USS in the diagnosis of appendicitis. USS should be used to diagnose other pelvic pathology and in equivocal cases. Such use of non-selective investigations represents another way in which the role of clinical skills is being all too rapidly eroded. In this study the only statistically significant difference found was the increase in time to operation in the control group (10 vs 7 hours) although 70/129 (54%) if USS were delayed until 8 am. Even if a diagnosis of early appendicitis is made out of hours, outcome will not alter as surgery is still likely to take place in daylight hours. However, statistical significance is not the same as clinical significance. The important outcome measures are surgical complications and hospital stay, and these were no different. In addition USS did not reduce the number of non-theraputic operations. We would argue that reported accuracy of US (sensitivity 95%, specificity 89%) in this and other papers is not reproducible in most centres. Availability of ultrasound is a further problem, especially out of working hours and when there is a nationwide shortage of radiologists. The Alvarado Score has proved to be of little value in diagnosing appendicitis. It's role in this study seemed superfluous as intervention often occurred regardless of a low Alvarado Score. In conclusion, we cannot see that this study would influence our current clinical practice of using USS in equivocal cases and using laprascopy to avoid unnecessary appendectomy. ALICE CUTTING
REF: 1. Charles D Douglas et al Randomised Controlled Trial of Ultrasonography in diagnosis of acute appendicitis, Incorporating the Alvarado Score. BMJ 2000; 321:919-22 2. Galindo Gallego M et al. Evaluation of Ultrasonography and Clinical diagnostic scoring in suspected appendicitis. BJS 85 (1):37-40 1998 Jan. 3. Alvarado A A practical score for the early diagnosis of acute Appendicitis. Ann.Emerg.Med.1986;15:557-64 4. Macklin CP et al. A prospective evaluation of the modified Alvarado score for acute appendicitis in children. Ann R Coll Surg.Eng 1. 1997;79:203-5 |
|||