Electronic Letters to:

Knee: B. Fink, C. Makowiak, M. Fuerst, I. Berger, P. Schäfer, and L. Frommelt The value of synovial biopsy, joint aspiration and C-reactive protein in the diagnosis of late peri-prosthetic infection of total knee replacements J Bone Joint Surg Br 2008; 90-B: 874-878 [Abstract] [Full text] [PDF]

eLetters: Submit a response to this article

Electronic letters published:

Which test to choose for the diagnosis of late peri-prosthetic infection of total knee replacements?

Yu-Min Lin, Chien-Chou Pan, Tu-Sheng Lee   (21 August 2008)

Which test to choose for the diagnosis of late peri-prosthetic infection of total knee replacements? 21 August 2008
Yu-Min Lin, Orthopaedic surgeon Taichung Veterans General Hospital, Taichung, Taiwan, Chien-Chou Pan, Tu-Sheng Lee Send letter to journal: Re: Which test to choose for the diagnosis of late peri-prosthetic infection of total knee replacements? ymlin{at}vghtc.gov.tw Yu-Min Lin, et al.	Sir, We read this paper with interest in which the authors have suggested that biopsy is superior to joint aspiration and C-reactive protein (CRP) in the diagnosis of late peri-prosthetic infection of total knee replacements. We have the following points to make. When comparing two or more diagnostic tests, it is convenient to summarise the test’s accuracies with a single summary measure.1 Several such summary measures are used in the literature. One is Youden’s index,2 defined as sensitivity + specificity - 1. Youden indexes for CRP, aspiration, bacteriological examination, histological examination and biopsy are 0.534, 0.677, 0.756, 0.852, and 0.981, respectively. Another commonly used is the probability of a correct diagnosis, often referred to simply as ”accuracy” in the literature, defined as (prevalence of disease in the sample) x sensitivity + (1 - prevalence of disease in the sample) x specificity.1 The ”accuracy” for CRP, aspiration, bacteriological examination, histological examination and biopsy are 78.6%, 88.9%, 92.4%, 93.7%, and 98.6%, respectively. On the other hand, however, the authors calculated the accuracy of these different diagnostic modalities from the sum of the true positives and the true negatives divided by the number of tests carried out. The authors report their findings by using a decision matrix table.3 We cannot construct different 2 x 2 matched sample tables3 because of a lack of original dataset, so we cannot use the McNemar test for comparing different diagnostic tests. Instead, we compared Youden’s indexes of these diagnostic tests by means of the t-test.2 Because ten comparisons were performed, a Bonferroni adjusted α value of 0.005 (0.05/10) was used to preserve the overall 0.05 error rate. The t- and p-values for these comparisons are as follows: CRP versus aspiration (1.311, 0.192), CRP versus bacteriological examination (2.112, 0.036), CRP versus histological examination (3.326, 0.001), CRP versus biopsy (5.482, < 0.001), aspiration versus bacteriological examination (0.788, 0.432), aspiration versus histological examination (1.944, 0.054), aspiration versus biopsy (4.059, < 0.001), bacteriological examination versus histological examination (1.135, 0.258), bacteriological examination versus biopsy (3.277, 0.001), and histological examination versus biopsy (2.404, 0.017), respectively. We found biopsy is superior to CRP, aspiration and bacteriological examination, but not to histological examination. Histological examination is superior to CRP, but not to aspiration and bacteriological examination. C.-C. Pan, MD, T.-S. Lee, MD, PhD, Y.-M. Lin, MD, MMS, Taichung Veterans General Hospital, Taichung, Taiwan. 1. Obuchowski NA. ROC analysis. AJR Am J Roentgenol 2005;184:364-72. 2. Youden WJ. Index for rating diagnostic tests. Cancer 1950;3:32-5. 3. Hawass NE. Comparing the sensitivities and specificities of two diagnostic procedures performed on the same group of patients. Br J Radiol 1997;70:360-6.