LAUSR

To the Editor We read with interest the paper authored by Jinxia Zhao et al. [1] published in Clinical Rheumatology. The aim of the study was measuring the prevalence of serum connective tissue growth factor (CTGF) and investigating the associations between CTGF and the clinical and laboratory features of rheumatoid arthritis (RA). Serum samples were obtained from 348 patients. Sixty-one of the 180 patients with RA were found to be positive for CTGF. Correlations were determined by computing Spearman rank correlation coefficients. Also, the diagnostic performance of CTGF was assessed by sensitivity, specificity, positive predictive value and negative predictive value, and receiver operating characteristic curve (ROC curve). At the cut-off value of 263.30 pg/ml, the sensitivity, specificity, positive predictive value (PPV), negative predictive (NPV), and ROC curve value of serum CTGF for RAwere 33.9%, 96.5%, 88.4%, 55.4%, and 0.71 respectively. As conclusion, the authors claimed that serum CTGF, as a novel biomarker, has certain diagnostic value for RA. In our opinion, however, there are some important methodological issues regarding the diagnostic value of serum CTGF. First, there is a difference between test research and diagnostic research. Diagnostic knowledge is the information needed to answer the question, “What is the probability of the presence or absence of a specific disease given these test results. In fact, diagnostic research focuses on the test’s added contribution to estimate the diagnostic probability of disease presence or absence [2]. So, other diagnostic tests and markers should be included to estimate the added value of serum CTGF. Second, sensitivity and specificity are appropriate indexes in public health and both PPV and NPV are influenced by prevalence of RA [3]. Therefore, we suggest the authors to calculate the positive and negative likelihood ratios (LR), since they are not influenced by prevalence of RA. In fact, LR+ = 9.6 and LR= 0.68 indicate that serum CTGF is not an accurate biomarker to use in the clinics [3]. Third, it is essential to consider both accuracy and precision of diagnostic research. Without determination of reliability, any judgment would be wrong. Weighted or Fleiss kappa and Intraclass Correlation Coefficient (ICC) are among measures of reliability that can be applied for qualitative and quantitative variables, respectively [4, 5]. Finally, the gold standard for selecting population in diagnostic study is the patients presenting with symptoms and signs indicative of the RA. Rather, a group of patients with evident disease is selected and compared to a group of non-diseased patients. Such selection of study subjects, however, will lead to biased estimates of the test’s performance [6].