Table 3

Baseline ECM biomarkers concentrations between responders versus non-responders and their predictive value within included studies

ECMStudy IDBaselineModelPredicting response
PINPUmar et al10~TC: 35.4±5.0 ug/l; R: 32.9±2.2 ug/L; NR: 41.9±4.3 ug/L, p=0.04(Stepwise) multiple logistic regression‡Univariate: OR 0.99, 95% CI 0.93 to1.00, p=0.05
Multivariable: OR 0.96, 95% CI 0.93 to 0.99, p=0.03
Lopez-Andres et al8*CRT-p: 33.0 ug/L (24.6–49.4); OMT: 33.1 ug/L (23.0–49.3), p=NS(Stepwise) Multiple logistic regression†No association with response
PICPGarcia-Bolao et al9TC: 74.3±29.9 ug/l; R: 85.6±29.4 ug/L; NR: 57.8±22.2 ug/L, p≤0.001ROC:PICP:CITPAUC 0.71, 95% CI 0.57 to 0.85; Cut-off 14.4, 95% CI 9.8 to 17.7; Sensitivity 63% (51–80); Specificity 70% (50–85); OR 2.07, 95% CI 0.98 to 4.39
PIIINPDong et al26TC: 0.88±0.21 ug/l; R: 0.80±0.20 ug/L; NR: 0.96±0.19 ug/L, p=0.03(Stepwise) multiple logistic regression‡Univariate: OR 0.77, 95% CI 0.62 to 0.97, p=0.03
Multivariable: OR 0.20, 95% CI 0.03 to 1.17, p=0.07
Umar et al10~R: 4.59±0.24 ug/L; NR: <responders, p=NS(Stepwise) multiple logistic regression‡Univariate: OR 1.23, 95% CI 0.86 to 1.76, p=0.23
Multivariable: OR 1.35, 95% CI 0.94 to 1.93, p=0.1
Lopez-Andres et al8*CRT-p: 4.6 ug/L (3.8–6.8); OMT: 4.7 ug/L (3.8–6.5), p=NS(Stepwise) multiple logistic regression†No association with response
ICTPUmar et al10~TC: 3.1±0.8 ug/l; R: 3.5±0.6 ug/L; NR: 2.1±0.4 ug/L, p=ns(Stepwise) multiple logistic regression‡Univariate: OR 1.24, 95% CI 0.93 to 1.66, p=0.13
Multivariable: No association with response
Lopez-Andres et al8*CRT-p: 4.1 ug/L (2.6–6.0); OMT: 3.4 ug/L (2.7–5.0), p=NS(Stepwise) multiple logistic regression†No association with response
CITPGarcia-Bolao et al9TC: 5.1±2.5 ug/l; R: 4.90±2.5 ug/L; NR: 5.3±2.5 ug/L, p=0.51ROC:PICP:CITPAUC 0.71, 95% CI 0.57 to 0.85; Cut-off value 14.4, 95% CI 9.8 to 17.7; Sensitivity 63% (51–80); Specificity 70% (50–85); OR 2.07, 95% CI 0.98to 4.39
ProMMP-1Umar et al10~TC: 7.7±0.8 ug/l; R: 7.6±0.7 ug/L; NR: 8.0±1.1 ug/L, p=0.71(Stepwise) multiple logistic regression‡Univariate: OR 0.97, 95% CI 0.87 to 1.09, p=0.71
Multivariable: No association demonstrated
MMP-1Garcia-Bolao et al9TC: 8.9±11.4 ug/l; R: 7.3±10.5 ug/L; NR: 11.3±12.5 ug/L, p=0.17ROCNot performed as no difference Baseline MMP-1:TIMP-1 ratio: 0.005±0.001, R: 0.004±0.0007 versus NR: 0.0063±0.0008, p=0.297
Lopez-Andres et al8*CRT-p: 2.7 ug/L (2.1–3.5); OMT: 2.7 ug/L (2.0–3.9), p=NS(Stepwise) multiple logistic regression†Univariate ≤3 ug/L: OR 2.42, 95% CI 1.23 to 4.76, p=0.011
Multivariable ≤3 ug/L: OR 3.04, 95% CI 1.37 to 6.71, p<0.01
MMP-2Tolosana et al24TC:295±70 ug/L;
R:258±56 ug/L; NR:325±116 ug/L, p=0.02
Cox Regression Model‡Univariate: difference already noted (p=0.02)
Multivariable: No association demonstrated
Garcia-Bolao et al9TC: 1434±401.5 ug/l; R: 1393.8±374.5 ug/L; NR: 1496.6±438.9 ug/L, p=0.36ROCNot performed as no difference demonstrated
MMP-9Garcia-Bolao et al9TC: 44.7±23.2 ug/l; R: 41.1±22.8 ug/L; NR: 49.9±23.3 ug/L, p=0.17ROCNot performed as no difference demonstrated
TIMP-1Tolosana et al24TC:242±61 ug/l; R:216±50 ug/L; NR:277±59 ug/L, p=0.001Cox Regression‡; ROCMultivariate: OR 0.97, 95% CI 0.96 to 0.99, p=0.005, ROC: ≥248 ug/L, Sensitivity 71%, Specificity 72%, OR 6.8, 95% CI 1.5 to 31
Umar et al10~TC: 120.3±8.2 ug/l; R: 124±5.2 ug/L; NR: 111±7.1 ug/L, p=0.16(Stepwise) multiple logistic regression‡Univariate: OR 1.01, 95% CI 0.99 to 1.03, p=0.16
Multivariable: No association demonstrated
Garcia-Bolao et al9TC: 488.9±249.5 ug/l; R: 437.5±136.5 ug/L; NR: 563.8±345.7 ug/L, p=0.135ROCNot performed as no difference. Baseline MMP-1:TIMP-1:TC: 0.005±0.001, R: 0.004±0.0007 versus NR: 0.0063±0.0008, p=0.297
Gal-3Truong et al25*TC: 18.1 ug/L (14.0–23.0), Positive result >25.9 ug/L2×2 table;
McNemar test
Peripheral: Sensitivity 15% (5–32), Specificity 80% (64–91), PPV: 38% (14–68), NPV: 53% (40–66)
Lopez-Andres et al8*CRTp: 25.7 ug/L (20.6–31.4); OMT: 25.1 ug/L (19.6–30.9), p=NS(Stepwise) multiple logistic regression†No association with response
  • *Median (IQR) given, data represented as CRT-p and OMT groups.

  • †Statistical model predicts ‘non-response’.

  • ‡Statistical model predicts ‘response’.

  •  ~ mean±standard error.

  • ≠biomarker mean±SD given in logarithmic form

  • AUC, area under the curve, CITP, carboxy-terminal telopeptide of collagen type-I; CRT-p, CRT-pacemaker; ECM, extracellular cardiac matrix; Gal-3, galectin-3; MMP, metalloproteinases; NPV, negative predictive value; NS=not significant; NR=non-responder; OMT, optimal medical therapy; PICP, carboxy-terminal propeptide of procollagen type I; PINP, N-terminal propeptides of type I procollagens; PIIINP, N-terminal propeptides of type III procollagens; PPV, positive predictive value; R, responder; ROC, receiver operator curve; TC, total cohort; TIMP, tissue inhibitors of MMPs.