Rapid Responses to:
|
|
Rapid Responses: Rapid Responses published:
-
![[Read Rapid Response]](/icons/misc/sectionDOWN.gif)
Wallflowers
- Diane-Marie Campbell (14 January 2005)
-
How to use oral antibiotics in cellulitis & don't forget to look between the toes.
- Steve J Searle (14 January 2005)
-
Re: How to use oral antibiotics in cellulitis & don't forget to look between the toes.
- Dr John Rumbold (15 January 2005)
-
Congratulations!!useful study.
- Vaishali MOna Verma, nil (16 January 2005)
-
Cellulitis in the home and hospital treated differently - is this real equivalence?
- Peter Leman (20 January 2005)
-
Reply to Peter Leman's Rapid Response
- Paul A Corwin, Les Toop, J Elisabeth Wells (27 January 2005)
|
|
|||
|
Diane-Marie Campbell, locum emergency physician Bunbury, West Australia 6230
Send response to journal:
|
It is quite common for the strict criteria necessary for entry into controlled trials to be waived and then ignored once the treatment concerned becomes available outside these studies. One criterion clinicians seem to cling to is the need for a "caregiver" at home or nearby when Hospital in the Home (HITH) is considered. Is there any evidence that this is necessary, given that a telephone and reasonable proximity is also required? Mothers and other patients with dependents may benefit from the respite of hospital admission, and the 12 - 14 hours a day when visitors are discouraged. Those of us who live alone may value solitude and find it particularly trying to be admitted to a ward containing not only other patients, but their extended households, for much of the day. HITH is spreading, and a good thing too, embracing those whose co- morbidities initially excluded them. Singles are still excluded, even when there are no hospital beds available. The absurd result is that it is these unaccompanied patients who are asked to attend the ED daily or more often for their antibiotics or heparin. Competing interests: Spinster. Locum Emergency Physician, no HITH scheme at current place of employment. |
|||
|
|
|||
|
Steve J Searle, GP and Emergency Department Doctor Emergency Department, Dunedin Hospital, New Zealand 9006
Send response to journal:
|
I work in both GP and ED, and the outcome data in your study fit with my clinical exeperience - namely 1 to 4 days to no advancement on IV ABs, and 2.5 to 11 days of IV antibiotics being required, and oral antibiotics for 10 to nearly 30 days being just as commonly needed as just 5 to 10days worth. The standard medical texts for years have which oral antibiotic to give for cellulitis but almost never give information on duration of oral antibiotics. Did you in your study, or related reading, come across answers to any of the following questions?; 1) any way to predict likely duration of oral antibiotics (patient or
localised cellulitis variables)
As a parting comment I think the studies on tinea between the toes suggest this breaks down the skin barrier and allows bacteria in as a source of infection for leg cellulitis (Ref 1.) (even well proximal to the toes). This seems to be recent information and most doctors are not looking for this with cellulitis or treating it or taking swabs from between the toes before starting therapy (which can culture the bacteria causing the cellulitis). I think it is worth having this as a passing comment in most published studies of cellulitis over the next few years to raise awareness as it can also lead to prevention. The risk of cellulitis from tinea is for example much greater than the risk from having diabetes. 1) Alain Dupuy, Hakima Benchikhi, Jean-Claude Roujeau, Philippe Bernard, Loïc Vaillant, Olivier Chosidow, Bruno Sassolas, Jean-Claude Guillaume, Jean-Jacques Grob, and Sylvie Bastuji-Garin Risk factors for erysipelas of the leg (cellulitis): case-control study BMJ, Jun 1999; 318: 1591 - 1594 Competing interests: None declared |
|||
|
|
|||
|
Dr John Rumbold, n/a West Midlands
Send response to journal:
|
I do regularly check for tinea in cases of lower limb cellulitis but culturing the interdigital clefts is new to me - do you culture all 4 or depending on the skin condition? I have had a patient with recurrent cellulitis in the same leg take extreme umbrage at being examined for tinea though! Competing interests: None declared |
|||
|
|
|||
|
Vaishali MOna Verma, GP Christchurch, New Xealand 8005, nil
Send response to journal:
|
Dear Ed, I take this opportunity to congratulate my colleagues Simon, Paul and Lees Top, for this very useful study.Since under the Pegasus Health Banner we have launched this I/V antibiotic program, it is essential that individual local GPs, be given the list of medical conditions which are unsafe for this program, to avoid medical error. Thanks, regards, Dr Mona Verma GP Christchurch New Zealand Competing interests: None declared |
|||
|
|
|||
|
Peter Leman, Clinical Senior Lecturer Royal Perth Hospital, Perth, WA, Australia, 6001
Send response to journal:
|
Editor, I read with interest Corwin et als study on cellulitis treatment in Christchurch, NZ and am interested in their main outcome measure, time to ‘no advancement of the cellulitis’ and their finding that there were no significant differences between the 2 groups. There are several causes for concern in the study design that may affect interpretation of the data. It seems that the power calculation was designed to show a significant difference of 2 days, the fact that they did not do so may be because the study isn’t designed as a non-inferiority study and thus no difference found may be due to a lack of power (a type 2 error). As nearly 2/3 of screened patients were deemed unsuitable for study entry they have chosen a very select group to study and the conclusion that home treatment is safe must take this very large selection bias into account. They conclude that there may have been some patients in the non- randomised group that were also suitable for home treatment but present no data to support this. Other factors affecting recruitment make it difficult to interpret the study, e.g. why would a white cell count of > 12x10^9/l be an exclusion criteria at all (especially when blood tests weren’t even required) indeed how many actually had a white cell count performed? Why choose a heart rate > 90/min, and how do they define ‘a very large affected area’ to mandate exclusion? Determining illness severity from the methodology is difficult and makes generalisation of the findings less meaningful. The fact that the junior in ED thought they might need something intravenously is pragmatic, but not reproducible. The authors state that hospital in the home is equivalent in efficacy, yet there may well be significant crossover effect in this study, all patients were admitted overnight to the observation ward. Spending how many hours/days in hospital before going home, and how many doses of antibiotics, or perhaps more importantly hours of bed rest and limb elevation? Also, despite the reported safety, over 10% of patients required subsequent admission. A major concern is to do with equivalence of treatment between the home group and the hospital group. In hospital only 55% of patients actually received the study drug, the rest something else – how can any meaningful comparisons of treatment outcomes be made here? In addition the outcome measure is open to significant misinterpretation. When was the line marked, did regression from the line equal the outcome measure or just non advancement (and what amount of variation was allowed), what was the relationship of the line measurement to the initial marking and how was observer error taken into account? Let alone the fact that the study was unblinded. Using such a simple non-continuous measure leads to potential misclassification bias – no effect measured isn’t equivalent to no effect. The hazard ratios look at outcomes such as time on intravenous or oral antibiotics, yet the study isn’t designed to assess these, is unblinded, has no defined end points and as there are vastly different treatments provided between groups comparison is of little value. The patients were happier at home than in the hospital, in this study location (a finding which probably is generalisable) that doesn’t seem in doubt. Otherwise the findings of the paper add little to ‘what is already known’ on the subject of home delivered antibiotics in cellulitis. [1] Corwin P, Toop L, McGeoch G, Than M, Wynn-Thomas S, Wells JE, Dawson R, Abernethy P, Pithie A, Chambers S, Fletcher L, Richards D. Randomised controlled trial of intravenous antibiotic treatment for cellulitis at home compared with hospital. BMJ. 2005 Jan 15;330(7483):129 Competing interests: None declared |
|||
|
|
|||
|
Paul A Corwin, Senior Lecturer Christchurch School of Medicine and Health Sciences, PO Box 4345, Christchurch, NZ, Les Toop, J Elisabeth Wells
Send response to journal:
|
We would like to respond to the points raised by Peter Leman regarding our RCT of home versus hospital treatment of cellulitis. Dr Leman criticizes the study for possible lack of power. This criticism can be re-expressed as insufficient precision in estimation of the treatment effect. As stated in the clinical outcomes section of the results the 95% confidence interval around the difference in days to no advancement between the home and hospital groups was -0.30 to 0.28 days, well within the difference of two days which our clinicians regarded as clinically important. It is true that no effect detected, using significance testing , is not the same as no effect. This is why the British Medical Journal insists on confidence intervals around estimates, not just p values. There was discussion within the research team about analysis of this study as a non-inferiority trial, since it was originally planned as such. Using a proportional hazards model with Ą1=0.05 (as is standard for an equivalence trial) and a margin of equivalence of two days the hazard ratio was 0.98 (90% CI 0.77, 1.26) which was significantly similar with p=0.01. Patients were NOT all admitted to hospital overnight. Only patients presenting to the emergency department between 10 pm and 8 am were looked after in the observation ward. This applied to only a few patients in the trial and was done as we did not have the funding to have trial and home care staff on call around the clock. Dr. Leman criticises our outcome measures. There are no accepted outcome measures in monitoring cellulitis. We did an extensive literature search prior to our study and also did a pilot trial of different methods of trying to record the progression of cellulitis. Most cellulitis trials have used very vague outcome measures such as clinical improvement, treatment failure or febrifuge. In our pilot we tried recording the area of cellulitis, both the advancement of the cellulitis margin and regression of cellulitis margin and digital photographs. The most useful measurement in terms of intraobserver variation was failure of the cellulitis to advance and we therefore used this as our primary outcome measure. Our other outcome measures of time on oral and intravenous antibiotics and time to discharge were of course influenced by individual clinician decisions that we had no control over and were therefore secondary outcomes. Dr. Leman refers several times to the fact that our study was not blinded. Clearly it is impossible to perform a study such as this and blind patients and researchers as to the site of treatment. We had no control over the antibiotic regimes chosen by the many different hospital teams that looked after cellulitis patients who were admitted. Nevertheless most hospital patients were treated with the same cephazolin as the home patients. One third of hospital patients did receive intravenous flucloxacillin. We are unaware of any data to suggest these two antibiotics differ in terms of efficacy in the treatment of cellulitis and are unclear why he has stated that the treatment regimes were "vastly different." Dr Leman also takes issue with our inclusion and exclusion criteria. The white cell count and heart rate criteria are standard criteria in defining systemic sepsis. We erred on the side of caution and no doubt excluded patients who could have been safely treated at home successfully. As the paper states the decision that someone needed intravenous antibiotics was always made by a senior consultant, never a junior doctor. Three quarters of our patients had failed to improve on oral antibiotics prior to trial entry. The fact that our rate of transfer from home to hospital treatment was similar to that found in other cellulitis trials would suggest that our criteria were appropriate. Dr. Leman concludes that our study adds little to what is already known regarding home treatment of cellulitis. Given that the only other RCT of home versus hospital treatment for cellulitis we have been able to find only involved 37 patients and the only clinical outcome reported was transfer to hospital, we think our study provides the best current evidence that home treatment for cellulitis is a safe and effective option for many patients. Yours, Paul Corwin, Les Toop, J Elisabeth Wells Competing interests: None declared |
|||