Limits...
Predictive value of early brain atrophy on response in patients treated with interferon β.

Pérez-Miralles FC, Sastre-Garriga J, Vidal-Jordana A, Río J, Auger C, Pareto D, Tintoré M, Rovira A, Montalban X - Neurol Neuroimmunol Neuroinflamm (2015)

Bottom Line: Larger PBVC and WMVc% decreases were observed in patients with disability worsening at 4 years of follow-up, whereas no differences were found in BPVc% or GMVc%.Patients with decreases of PBVC and WMVc% below cutoff values were more prone to develop disability worsening (unadjusted hazard ratio [HR] 3.875, p = 0.005; HR 4.246, p = 0.004, respectively).PBVC (HR 4.751, p = 0.008) and the interaction of new T2 lesions with WMVc% (HR 1.086, p = 0.005) were found to be independent predictors of disability worsening in the multivariate analysis.

View Article: PubMed Central - PubMed

Affiliation: Servei de Neurologia/Neuroimmunologia (F.C.P.-M., J.S.-G., A.V.-J., J.R., M.T., X.M.) Multiple Sclerosis Centre of Catalonia (Cemcat) and Unitat de Ressonància Magnètica (Servei de Radiologia) (C.A., D.P., A.R.), Hospital Universitari Vall d'Hebron, Barcelona, Spain; and Departament de Medicina (F.C.P.-M.), Universitat Autònoma de Barcelona, Barcelona, Spain.

ABSTRACT

Objective: To investigate the association between brain volume loss during the first year of interferon treatment and clinical outcome at 4 years.

Methods: Patients with multiple sclerosis initiating interferon β were clinically evaluated every 6 months for the presence of relapses and assessment of global disability using the Expanded Disability Status Scale (EDSS). MRI scans were performed at baseline and after 12 months, and the percentage of brain volume change (PBVC), brain parenchymal volume change (BPVc%), gray matter volume change (GMVc%), and white matter volume change (WMVc%) were estimated. Patients were divided based on the cutoff values for predicting confirmed EDSS worsening obtained by receiver operating characteristic analysis for all atrophy measurements. Survival curves and Cox proportional hazards regression to predict disability worsening at last observation were applied, adjusting for demographic, clinical, and radiologic variables.

Results: Larger PBVC and WMVc% decreases were observed in patients with disability worsening at 4 years of follow-up, whereas no differences were found in BPVc% or GMVc%. Cutoff points were obtained for PBVC (-0.86%; sensitivity 65.5%, specificity 71.4%) and WMVc% (-2.49%; sensitivity 85.3%, specificity 43.8%). Patients with decreases of PBVC and WMVc% below cutoff values were more prone to develop disability worsening (unadjusted hazard ratio [HR] 3.875, p = 0.005; HR 4.246, p = 0.004, respectively). PBVC (HR 4.751, p = 0.008) and the interaction of new T2 lesions with WMVc% (HR 1.086, p = 0.005) were found to be independent predictors of disability worsening in the multivariate analysis.

Conclusions: At the patient level, whole-brain and white matter volume changes in the first year of interferon β therapy are predictive of subsequent clinical evolution under treatment.

No MeSH data available.


Related in: MedlinePlus

Cox proportional hazards regression models for predicting confirmed Expanded Disability Status Scale worsening at 4 years of follow-upHazard ratios are plotted with 95% confidence interval. (A) Model including percentage of brain volume change (PBVC) categorized into ≤−0.86% or higher. (B) Model including percentage of white matter volume change (WMVc%) categorized into ≤−2.49% or higher as an interaction with number of new or enlarged T2 lesions after 12 months of therapy (NEL). ARR = annualized relapse rate as the mean of relapses of the 2 years before starting treatment; Bas. BPV = baseline brain parenchymal volume; Bas. WMV = baseline white matter volume; CEL = contrast-enhancing lesions; TtT = time in years since the first attack until the start of therapy with interferon β; WMVc% = (100 × [first year WMV − baseline WMV]/[baseline WMV]).
© Copyright Policy - open-access
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC4496631&req=5

Figure 3: Cox proportional hazards regression models for predicting confirmed Expanded Disability Status Scale worsening at 4 years of follow-upHazard ratios are plotted with 95% confidence interval. (A) Model including percentage of brain volume change (PBVC) categorized into ≤−0.86% or higher. (B) Model including percentage of white matter volume change (WMVc%) categorized into ≤−2.49% or higher as an interaction with number of new or enlarged T2 lesions after 12 months of therapy (NEL). ARR = annualized relapse rate as the mean of relapses of the 2 years before starting treatment; Bas. BPV = baseline brain parenchymal volume; Bas. WMV = baseline white matter volume; CEL = contrast-enhancing lesions; TtT = time in years since the first attack until the start of therapy with interferon β; WMVc% = (100 × [first year WMV − baseline WMV]/[baseline WMV]).

Mentions: Signs of a moderate interaction between NEL and the presence or absence of a WMVc% of −2.49% or lower were detected, but not for PBVC. Although the likelihood of having CW after 4 years of follow-up increases with NEL for both groups of WMVc%, it increases faster for the group with WMVc% ≤−2.49% (figure e-3), so these variables were expressed as an interaction in the regression model. Two models for the Cox proportional hazards regression were used (figure 3). Common variables used are those already explained in the methods section. Model 1 (n = 105) showed that an annual PBVC ≤−0.86% after treatment (adjusted HR 4.647, 95% CI 1.479–14.603, p = 0.009), the number of NEL, and the presence of relapses during the first year of therapy were independent predictors for developing CW. Model 2 (n = 84) indicated that the interaction between NEL and annual WMVc% ≤−2.49% after treatment (adjusted HR 1.073, 95% CI 1.012–1.137, p = 0.018), together with the presence of relapses in the first year of therapy, were independent predictors for CW. BPVc% and GMVc%, either alone or in combination, were not found to predict sustained progression as per the models above (data not shown).


Predictive value of early brain atrophy on response in patients treated with interferon β.

Pérez-Miralles FC, Sastre-Garriga J, Vidal-Jordana A, Río J, Auger C, Pareto D, Tintoré M, Rovira A, Montalban X - Neurol Neuroimmunol Neuroinflamm (2015)

Cox proportional hazards regression models for predicting confirmed Expanded Disability Status Scale worsening at 4 years of follow-upHazard ratios are plotted with 95% confidence interval. (A) Model including percentage of brain volume change (PBVC) categorized into ≤−0.86% or higher. (B) Model including percentage of white matter volume change (WMVc%) categorized into ≤−2.49% or higher as an interaction with number of new or enlarged T2 lesions after 12 months of therapy (NEL). ARR = annualized relapse rate as the mean of relapses of the 2 years before starting treatment; Bas. BPV = baseline brain parenchymal volume; Bas. WMV = baseline white matter volume; CEL = contrast-enhancing lesions; TtT = time in years since the first attack until the start of therapy with interferon β; WMVc% = (100 × [first year WMV − baseline WMV]/[baseline WMV]).
© Copyright Policy - open-access
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC4496631&req=5

Figure 3: Cox proportional hazards regression models for predicting confirmed Expanded Disability Status Scale worsening at 4 years of follow-upHazard ratios are plotted with 95% confidence interval. (A) Model including percentage of brain volume change (PBVC) categorized into ≤−0.86% or higher. (B) Model including percentage of white matter volume change (WMVc%) categorized into ≤−2.49% or higher as an interaction with number of new or enlarged T2 lesions after 12 months of therapy (NEL). ARR = annualized relapse rate as the mean of relapses of the 2 years before starting treatment; Bas. BPV = baseline brain parenchymal volume; Bas. WMV = baseline white matter volume; CEL = contrast-enhancing lesions; TtT = time in years since the first attack until the start of therapy with interferon β; WMVc% = (100 × [first year WMV − baseline WMV]/[baseline WMV]).
Mentions: Signs of a moderate interaction between NEL and the presence or absence of a WMVc% of −2.49% or lower were detected, but not for PBVC. Although the likelihood of having CW after 4 years of follow-up increases with NEL for both groups of WMVc%, it increases faster for the group with WMVc% ≤−2.49% (figure e-3), so these variables were expressed as an interaction in the regression model. Two models for the Cox proportional hazards regression were used (figure 3). Common variables used are those already explained in the methods section. Model 1 (n = 105) showed that an annual PBVC ≤−0.86% after treatment (adjusted HR 4.647, 95% CI 1.479–14.603, p = 0.009), the number of NEL, and the presence of relapses during the first year of therapy were independent predictors for developing CW. Model 2 (n = 84) indicated that the interaction between NEL and annual WMVc% ≤−2.49% after treatment (adjusted HR 1.073, 95% CI 1.012–1.137, p = 0.018), together with the presence of relapses in the first year of therapy, were independent predictors for CW. BPVc% and GMVc%, either alone or in combination, were not found to predict sustained progression as per the models above (data not shown).

Bottom Line: Larger PBVC and WMVc% decreases were observed in patients with disability worsening at 4 years of follow-up, whereas no differences were found in BPVc% or GMVc%.Patients with decreases of PBVC and WMVc% below cutoff values were more prone to develop disability worsening (unadjusted hazard ratio [HR] 3.875, p = 0.005; HR 4.246, p = 0.004, respectively).PBVC (HR 4.751, p = 0.008) and the interaction of new T2 lesions with WMVc% (HR 1.086, p = 0.005) were found to be independent predictors of disability worsening in the multivariate analysis.

View Article: PubMed Central - PubMed

Affiliation: Servei de Neurologia/Neuroimmunologia (F.C.P.-M., J.S.-G., A.V.-J., J.R., M.T., X.M.) Multiple Sclerosis Centre of Catalonia (Cemcat) and Unitat de Ressonància Magnètica (Servei de Radiologia) (C.A., D.P., A.R.), Hospital Universitari Vall d'Hebron, Barcelona, Spain; and Departament de Medicina (F.C.P.-M.), Universitat Autònoma de Barcelona, Barcelona, Spain.

ABSTRACT

Objective: To investigate the association between brain volume loss during the first year of interferon treatment and clinical outcome at 4 years.

Methods: Patients with multiple sclerosis initiating interferon β were clinically evaluated every 6 months for the presence of relapses and assessment of global disability using the Expanded Disability Status Scale (EDSS). MRI scans were performed at baseline and after 12 months, and the percentage of brain volume change (PBVC), brain parenchymal volume change (BPVc%), gray matter volume change (GMVc%), and white matter volume change (WMVc%) were estimated. Patients were divided based on the cutoff values for predicting confirmed EDSS worsening obtained by receiver operating characteristic analysis for all atrophy measurements. Survival curves and Cox proportional hazards regression to predict disability worsening at last observation were applied, adjusting for demographic, clinical, and radiologic variables.

Results: Larger PBVC and WMVc% decreases were observed in patients with disability worsening at 4 years of follow-up, whereas no differences were found in BPVc% or GMVc%. Cutoff points were obtained for PBVC (-0.86%; sensitivity 65.5%, specificity 71.4%) and WMVc% (-2.49%; sensitivity 85.3%, specificity 43.8%). Patients with decreases of PBVC and WMVc% below cutoff values were more prone to develop disability worsening (unadjusted hazard ratio [HR] 3.875, p = 0.005; HR 4.246, p = 0.004, respectively). PBVC (HR 4.751, p = 0.008) and the interaction of new T2 lesions with WMVc% (HR 1.086, p = 0.005) were found to be independent predictors of disability worsening in the multivariate analysis.

Conclusions: At the patient level, whole-brain and white matter volume changes in the first year of interferon β therapy are predictive of subsequent clinical evolution under treatment.

No MeSH data available.


Related in: MedlinePlus