TY - JOUR
T1 - Molecular and Structural Biomarkers of Inflammation at Two Years After Acute Anterior Cruciate Ligament Injury Do Not Predict Structural Knee Osteoarthritis at Five Years
AU - Roemer, Frank W.
AU - Englund, Martin
AU - Turkiewicz, Aleksandra
AU - Struglics, André
AU - Guermazi, Ali
AU - Lohmander, L. Stefan
AU - Larsson, Staffan
AU - Frobell, Richard
PY - 2019
Y1 - 2019
N2 - Objective: To determine the role of inflammatory biomarkers at 2 years post–anterior cruciate ligament (ACL) injury to predict radiographic knee osteoarthritis (OA) and magnetic resonance imaging (MRI)–defined knee OA at 5 years postinjury, with a secondary aim of estimating the concordance of inflammatory biomarkers assessed by MRI and synovial fluid (SF) analysis. Methods: We studied 113 patients with acute ACL injury. Knee scans using 1.5T MRIs were read for Hoffa- and effusion-synovitis. Biomarkers of inflammation that we assessed included interleukin-6 (IL-6), IL-8, IL-10, tumor necrosis factor, and interferon-ɣ in serum and SF, and IL-12p70 in serum. We defined the outcome as radiographic knee OA (ROA) or MRI-defined OA (MROA) at 5 years. The area under the receiver operating characteristic curve (AUC), sensitivity, and specificity were evaluated in models that included MRI features only (model 1), inflammation biomarkers only (serum [model 2a] or SF [model 2b]), both MRI features and serum biomarkers (model 3a), or both MRI features and SF (model 3b) biomarkers. Linear regression analysis was used to evaluate the association between MRI features and SF biomarkers. Results: At 5 years postinjury, ROA was present in 26% of the injured knees, and MROA was present in 32%. The AUCs for ROA in each model were 0.44 (95% confidence interval [95% CI] 0.42, 0.47) for model 1, 0.62 (95% CI 0.59, 0.65) for model 2a, 0.53 (95% CI 0.50, 0.56) for model 2b, 0.58 (95% CI 0.55, 0.61) for model 3a, and 0.50 (95% CI 0.46, 0.53) for model 3b. The AUCs for MROA in each model were 0.67 (95% CI 0.64, 0.70) for model 1, 0.49 (95% CI 0.47, 0.52) for model 2a, 0.56 (95% CI 0.52, 0.59) for model 2b, 0.65 (95% CI 0.61, 0.68) for model 3a, and 0.69 (95% CI 0.66, 0.72) for model 3b. The concordance between MRI and SF biomarkers was statistically significant only for effusion-synovitis and IL-8. Conclusion: Neither MRI-detected inflammation nor selected SF/serum inflammation biomarkers at 2 years postinjury predicted ROA or MROA at 5 years postinjury. Concordance between MRI and SF inflammatory biomarkers was weak.
AB - Objective: To determine the role of inflammatory biomarkers at 2 years post–anterior cruciate ligament (ACL) injury to predict radiographic knee osteoarthritis (OA) and magnetic resonance imaging (MRI)–defined knee OA at 5 years postinjury, with a secondary aim of estimating the concordance of inflammatory biomarkers assessed by MRI and synovial fluid (SF) analysis. Methods: We studied 113 patients with acute ACL injury. Knee scans using 1.5T MRIs were read for Hoffa- and effusion-synovitis. Biomarkers of inflammation that we assessed included interleukin-6 (IL-6), IL-8, IL-10, tumor necrosis factor, and interferon-ɣ in serum and SF, and IL-12p70 in serum. We defined the outcome as radiographic knee OA (ROA) or MRI-defined OA (MROA) at 5 years. The area under the receiver operating characteristic curve (AUC), sensitivity, and specificity were evaluated in models that included MRI features only (model 1), inflammation biomarkers only (serum [model 2a] or SF [model 2b]), both MRI features and serum biomarkers (model 3a), or both MRI features and SF (model 3b) biomarkers. Linear regression analysis was used to evaluate the association between MRI features and SF biomarkers. Results: At 5 years postinjury, ROA was present in 26% of the injured knees, and MROA was present in 32%. The AUCs for ROA in each model were 0.44 (95% confidence interval [95% CI] 0.42, 0.47) for model 1, 0.62 (95% CI 0.59, 0.65) for model 2a, 0.53 (95% CI 0.50, 0.56) for model 2b, 0.58 (95% CI 0.55, 0.61) for model 3a, and 0.50 (95% CI 0.46, 0.53) for model 3b. The AUCs for MROA in each model were 0.67 (95% CI 0.64, 0.70) for model 1, 0.49 (95% CI 0.47, 0.52) for model 2a, 0.56 (95% CI 0.52, 0.59) for model 2b, 0.65 (95% CI 0.61, 0.68) for model 3a, and 0.69 (95% CI 0.66, 0.72) for model 3b. The concordance between MRI and SF biomarkers was statistically significant only for effusion-synovitis and IL-8. Conclusion: Neither MRI-detected inflammation nor selected SF/serum inflammation biomarkers at 2 years postinjury predicted ROA or MROA at 5 years postinjury. Concordance between MRI and SF inflammatory biomarkers was weak.
U2 - 10.1002/art.40687
DO - 10.1002/art.40687
M3 - Article
C2 - 30079991
AN - SCOPUS:85053297731
SN - 2326-5191
VL - 71
SP - 238
EP - 243
JO - Arthritis and Rheumatology
JF - Arthritis and Rheumatology
IS - 2
ER -