Author + information
- Ignacio García-Bolao, MD, PhD,
- Alfonso Macías, MD, PhD,
- Begoña López, PhD,
- Arantxa González, PhD,
- Juan J. Gavira, MD,
- Pedro Azcárate, MD,
- Eduardo Alegría, MD, PhD and
- Javier Díez, MD, PhD⁎ ()
- ↵⁎Área de Ciencias Cardiovasculares, CIMA and University Clinic, University of Navarra, Pío XII, 55, 31008 Pamplona, Spain
To the Editor:Serum concentrations of the carboxy-terminal propeptide of procollagen type I (PICP), a peptide that is cleaved from procollagen type I during the synthesis of fibril-forming collagen type I (CTI), may provide indirect diagnostic information on both the synthesis of fibril-forming CTI molecules and the extent of myocardial deposition of CTI fibers (1). We investigated whether PICP is related to the clinical long-term response to cardiac resynchronization therapy (CRT) in heart failure (HF) patients.
Thirty-eight consecutive patients were prospectively studied. All patients received CRT for New York Heart Association (NYHA) functional class III/IV HF, left ventricular (LV) ejection fraction (EF) ≤35%, and QRS ≥130 ms. Twenty healthy subjects were included as control subjects.
Patients were evaluated at baseline and at the 1-year follow-up. Evaluation included NYHA functional class, 6-min walk test, blinded echocardiographic study with measurement of interventricular and intraventricular dyssynchrony parameters (septal-to-posterior-wall motion delay [SPWMD] and septal-to-lateral-wall motion delay [SLWMD]), and obtaining of blood samples. At 1 year, patients were categorized as nonresponders if they died of HF, were scheduled for heart transplantation, or did not increase the distance walked in 6 minutes by >10%.
Serum PICP was determined by a sandwich enzyme-linked immunosorbent assay (ELISA) (inter- and intra-assay variations were 6.3% and 6.4%, respectively). The minimum analytical detection limit was 1 μg/l. Serum amino-terminal propeptide of brain natriuretic peptide (NT-proBNP) was measured by ELISA (inter- and intra-assay coefficients of variation lower than 2%).
Differences between baseline values and between final values in the two groups of patients were analysed with the Mann-Whitney Utest for SLWMD, interventricular dyssynchrony, and NT-proBNP and with the Student ttest for unpaired data for the rest of the quantitative variables. Intragroup comparisons between baseline and final values were analyzed with the Wilcoxon test (SLWMD, interventricular dyssynchrony, and NT-proBNP) and with the Student ttest for paired data for the rest of quantitative variables. Categoric variables were analyzed by chi-square test. Significant variables in univariate analysis were used in logistic regression analysis to predict the probability of positive response to CRT. A model was constructed using stepwise variable selection, verified with the Hosmer-Lemeshow test. Receiver-operating characteristic (ROC) curves allowed determination of the overall performance for predicting a positive response to CRT. The results are expressed as mean ± SD.
At 1 year, 26 patients (68%) were considered responders to CRT (Table 1).At baseline, nonresponders exhibited higher left ventricular end-diastolic diameter (LVEDD) and lower SLWMD than responders (p < 0.05). Baseline PICP was higher in responders than in controls (p < 0.01) and nonresponders (p < 0.05). Baseline NT-proBNP was higher (p < 0.001) in the two groups of patients than in controls (36 ± 5 pg/ml). Baseline NT-proBNP was increased (p < 0.05) in responders compared with nonresponders.
Whereas EF, LV diameters, and the dyssynchrony parameters decreased (p < 0.05) at 1 year in responders, they remained unchanged in nonresponders. At 1 year, PICP decreased (p < 0.005) in responders and increased (p < 0.005) in nonresponders. Whereas in responders PICP decreased in 21 patients (81%) and increased in 5 patients (19%), in nonresponders PICP decreased in 1 patient (8%) and increased in 11 patients (92%); these differences were significant (p < 0.05).
The NT-proBNP level tended to decrease in responders at 1 year, although the difference was not significant; NT-proBNP was unchanged in nonresponders.
Significant associations were observed between positive response to CRT and high baseline values of PICP (>73 μg/l) (chi square = 16.29; p < 0.001) and low baseline values of LVEDD (<67.5 mm) (chi square = 5.37; p < 0.05). The only independent predictor of a positive response to CRT was PICP (odds ratio 13.9, 95% confidence interval 7 to 97; p < 0.001). The accuracy of the model was confirmed by the nonsignificant Hosmer-Lemeshow goodness-of-fit-test (p = 0.796)
As shown in Figure 1,PICP exhibited the larger area under the ROC curve. In addition, only the area under the ROC curve for PICP and LVEDD was higher (p < 0.001 and p < 0.05, respectively) than 0.50. The cutoff value of PICP showed better sensitivity and specificity than the others parameters analyzed. The odds ratio of presenting a favorable response to CRT was higher for patients with PICP >73 μg/l than for patients with NT-proBNP >815 pg/ml, LVEDD <67.5 mm, or SLWMD >95 ms.
These findings can be summarized as follows: 1) an association exists between abnormally high baseline serum PICP and a positive long-term response to CRT in HF patients; 2) whereas the beneficial effects of CRT are associated with its ability to normalize serum PICP, the lack of benefit is associated with enhancement of serum PICP; and 3) serum PICP is more accurate than some echocardiographic parameters and NT-proBNP in the prediction of a positive long-term response to CRT.
Previous studies have shown that circulating PICP detected in HF patients is essentially of cardiac origin and that serum PICP is a reliable index of the amount of CTI present within the myocardium (2,3). Therefore, the present study suggests that ventricular dyssynchrony may induce excessive cardiac synthesis and deposition of CTI fibrils and that limitation of these alterations may be one of the mechanisms contributing to the positive effect of CRT. Our results would suggest also that stimulation of CTI synthesis and deposition may account for a negative response to CRT. Whatever are the factors determining these two patterns of response, it appears that long-term responses to this therapy are linked to its ability to interfere with myocardial fibrosis. Additional studies aimed to evaluate histologically measured myocardial collagen deposition in CRT patients are necessary to prove these hypotheses.
Some findings reported here suggest that PICP adds predictive value to other indices of response to CRT. First, serum PICP was the only independent predictor of a positive response to CRT. Second, serum PICP was a highly sensitive and specific parameter in the identification of a positive long-term response to CRT. Third, patients with serum levels of PICP >73 μg/l had an almost 28-fold higher probability of presenting a positive response to CRT, showing the highest performance for predicting a positive response to CRT.
In conclusion, the determination of serum PICP may be useful for predicting the response to CRT. However, we are aware that this was a study involving a relatively small number of patients with heterogeneous etiologies. Albeit preliminary, these findings set the stage for large-scale studies to definitively validate this approach.
- American College of Cardiology Foundation
- López B.,
- Querejeta R.,
- González A.,
- Sánchez E.,
- Larman M.,
- Díez J.
- Querejeta R.,
- López B.,
- González A.,
- et al.