Measuring Temperature Postoperatively Appears to Be a Waste of Time
Measuring Temperature Postoperatively Appears to Be a Waste of Time
Abstract & Commentary
J. Peter Donnelly, PhD, Clinical Microbiologist, University Hospital Nijmegen, The Netherlands, Section Editor, Microbiology, is Associate Editor for Infectious Disease Alert
Synopsis: The positive predictive value for infection of an aural temperature > 38°C measured post-operatively was found to be only 12%, indicating the practice is of limited value.
Source: Vermeulen H, et al. Diagnostic Accuracy of Routine Postoperative Body Temperature Measurements. Clin Infect Dis. 2005;40:1404-1410.
This was a prospective study involving 308 consecutive patients who had surgery, and whose body temperature was measured twice-daily for up to 14 days after surgery. A temperature of > 38°C was considered a positive test result and postoperative infection was diagnosed microbiologically or on clinical grounds as defined by the CDC. The physician, nurses, and patient were all blinded to the results of the temperature measurements. A total of 2282 measurements for 284 patients were analysed and 19 [7%] of them had infection. Using temperature curves showed that a temperature ³ 38°C exhibited a sensitivity of 37%, a specificity of 80% and a likelihood ratio of 1.8 for a positive test results and 0.8 for a negative test result. Using each measurement as an independent result yielded positive and negative predictive values of 8% and 90%, for a cut-off value of ³ 38°C. Nineteen patients had an infection and, of the 8 patients with severe infection, the temperature remained below 38°C before diagnosis. On the basis of these results, Vermeulen and colleagues advise abandoning measuring the temperature routinely after surgery except when there is good reason to expect infection.
Comment by J. Peter Donnelly, PhD
Measuring temperature to monitor patients is embedded in hospital practice and has a long history mainly because of the not unreasonable belief that infection is accompanied by inflammation, which, in turn, is manifest by an elevated body temperature. Indeed, as Vermeulen et al state, information about body temperature is considered essential to support clinical judgment and confirm signs of infection. It is considered nothing short of heresy to even doubt the value of this time-honoured practice. Aside from the issues around how, when, and where the body temperature should be measured, few question the validity of the practice, let alone conduct a study to estimate the performance characteristics. Vermeulen et al are therefore to be commended for attempting this study at all, and to be congratulated for persuading their ethics committee to approve it and their colleagues for going along with it. That the conviction that taking the temperature is a good thing was apparently shared by almost a fifth of the patients who declined to participate in the study. The attrition rate also shows how difficult and disheartening it can be to test long held convictions afresh (See figure 1), since in order to recruit sufficient numbers of patients, almost twice as many patients had to be approached. We are not told the fate of those who were not eligible, but is seems reasonably fair to assume that they fared no worse than those who participated lending further credence to both the low infection rate and the poor performance of monitoring the temperature.
Like any other surrogate test, the utility of measuring body temperature is dependent upon the prevalence of the event it corresponds to, in this case, infection. Experience shows that surrogate tests are generally poor when prevalence is low, as in this case. Indeed, reliance on an elevated temperature may well have meant a delay in diagnosing infections and, hence, starting treatment soon enough, since 6 of the 8 patients with severe infection did not develop a temperature ³ 38°C until afterwards. Whether these results would be replicated in other settings remains to be seen. Nevertheless, it would seem worthwhile to look critically at the practice of taking the temperature in other areas of clinical practice to at least determine its utility in the interests of evidence-based medicine, if not of the individual patient at risk of developing infection.
This was a prospective study involving 308 consecutive patients who had surgery, and whose body temperature was measured twice-daily for up to 14 days after surgery. A temperature of > 38°C was considered a positive test result and postoperative infection was diagnosed microbiologically or on clinical grounds as defined by the CDC.Subscribe Now for Access
You have reached your article limit for the month. We hope you found our articles both enjoyable and insightful. For information on new subscriptions, product trials, alternative billing arrangements or group and site discounts please call 800-688-2421. We look forward to having you as a long-term member of the Relias Media community.