B-type natriuretic peptides and mortality after stroke: A systematic review and meta-analysis
Teresa García-Berrocoso, Dolors Giralt, Alejandro Bustamante, Thorleif Etgen, Jesper Jensen, Jagdish Sharma, Kensaku Shibazaki, Ayhan Saritas, Xingyong Chen, William Whiteley, Joan Montaner.
Supplemental files
Appendices 2
Appendix e-1. Electronic search strategy. 2
Appendix e-2. Data collected for systematic review and meta-analysis. 3
Appendix e-3. Quality score (adapted from QUADASe1 and Qualitye2 questionnaires). 3
Appendix e-4. Individual participants’ information compiled for IPD analysis. 4
Tables 5
Table e-1. Quality score report for each included study. 5
Table e-2. Additive predictive value of BNP/NT-proBNP over clinical variables for ischemic stroke mortality. 6
Figures 7
Figure e-1. Mortality association of BNP/NT-proBNP depending on ischemic stroke etiologies. 7
Figure e-2. Mortality association of BNP/NT-proBNP depending on time of sample collection. 8
e-References 9
Appendices
Appendix e-1. Electronic search strategy.
1. cerebrovascular disorders/ or basal ganglia cerebrovascular disease/ or exp brain ischemia/ or carotid artery diseases/ or carotid artery thrombosis/ or carotid stenosis/ or cerebrovascular accident/ or exp brain infarction/ or exp hypoxia-ischemia, brain/ or exp intracranial arterial diseases/ or exp "intracranial embolism and thrombosis"/
2. ((brain or cerebr$ or cerebell$ or vertebrobasil$ or hemispher$ or intracran$ or intracerebral or infratentorial or supratentorial or middle cerebr$ or mca$ or anterior circulation) adj5 (isch?emi$ or infarct$ or thrombo$ or emboli$ or occlus$ or hypoxi$)).tw.
3. (isch?emi$ adj6 (stroke$ or apoplex$ or cerebral vasc$ or cerebrovasc$ or cva or attack$)).tw.
4. 1 or 2 or 3
5. brain natriuretic peptide.mp. or exp Natriuretic Peptide, Brain/
6. b-type natriuretic peptide.mp.
7. (brain natriuretic peptide or nesiritide or b-type natriuretic peptide or bnp gene product or bnp-32 or brain natriuretic peptide-32 or natrecor or natriuretic factor-32 or natriuretic peptide type-b or type-b natriuretic peptide or ventricular natriuretic peptide, b-type).mp. [mp=title, abstract, original title, name of substance word, subject heading word, protocol supplementary concept, rare disease supplementary concept, unique identifier]
8. NT-pro BNP.mp.
9. 5 or 6 or 7 or 8
10. Incidence/ or exp mortality/ or follow up studies/ or mortality/ or prognos$.tw. or predict$.tw. or course.tw. or rankin.tw. or Glasgow outcome scale.tw. or NIHSS.tw.
11. (prognos$ or outcome$ or follow-up or predict$).tw,sh.
12. exp Prognosis/
13. Disease Progression/
14. ((clinical or natural$ or disease$) adj (progress$ or course$ or histor$)).tw,sh.
15. Time Factors/
16. 10 or 11 or 12 or 13 or 14 or 15
17. 4 and 9 and 16
Appendix e-2. Data collected for systematic review and meta-analysis.
- PubMed unique identifier (PMID).
- First author, year and journal of publication.
- Biological material (e.g. plasma, serum) and time from stroke symptoms onset to sample collection.
- Biomarker being analyzed (i.e. BNP or NT-proBNP) and method of analysis.
- When used, cut-off point for biomarker, with sensitivity and specificity percentages.
- Time of death assessment and mortality rate.
- Mean values and standard deviation (SD) for neurological scale NIHSS scores at admission and age. Male sex rate.
- Total sample size and sample size for both death and survival groups.
- Mean values and SD for biomarker in both death and survival groups.
Items included in our standardized template form to be completed from articles selected for meta-analysis.
NIHSS: National Institutes of Health Stroke Scale.
Appendix e-3. Quality score (adapted from QUADASe1 and Qualitye2 questionnaires).
1. Diagnosis of stroke is based on expert clinical opinion supported by neuroimaging.
2. All patients (including controls, in case) have expert opinion + neuroimaging.
3. The biomarker is not used to determine the end-point.
4. Specify that biomarker measurement is blinded to clinical data.
5. Specify that clinical data collection is blinded or collected before biomarker measurement.
6. All patients who entered into study complete it or withdrawals are explained.
7. Biomarker cut-off is previously established, based on literature or pilot study.
8. Specify any or none disclosure.
9. Report if the study is prospective.
10. Definition of time period, follow-up end and median follow-up time of the study (at least two items).
11. Clinical end-points to be measured are defined prior to analysis.
12. Rationale for used sample size.
13. Reporting a list of candidate variables and estimated effect (Odd Ratio/Hazard Ratio) with 95% CI for all variables in multivariate analysis.
14. Specify the assay method and provide or reference a detailed protocol.
15. Cases are not representing a selected subgroup of patients (e.g. only atrial fibrillation patients).
Appendix e-4. Individual participants’ information compiled for IPD analysis.
- BNP/NT-proBNP blood levels and blood collection time.
- Death/survival and time from onset to death assessment.
- Age, sex, NIHSS and glucose blood levels at admission.
- Type of stroke, TOAST etiological classification.
- Presence/absence of several risk factors for stroke: hypertension, diabetes mellitus, dyslipidemia, atrial fibrillation, smoking, alcohol intake, previous stroke.
- Presence/absence of thrombolytic treatment (rt-PA) administration.
Items requested to corresponding authors from articles included in meta-analysis.
NIHSS: National Institutes of Health Stroke Scale; TOAST: trial of ORG 10172 in acutestroketreatment; rt-PA: recombinant tissue Plasminogen Activator.
All information was gathered in an Excel database for each anonymized individual patient included in each original published article. Finally, data from all articles was compiled in a single whole database, including a study identifier (first author and year of publication). When articles from the same research group shared some patients from the same cohort e3, e4 or the whole cohort was followed up any longer, e5, e6 after authors’ agreement, both studies were considered together as a unique cohort in the whole database to avoid double counting of subjects. The provided data remain entirely under the principal investigator property and was only used for this analysis.
In order to have more consistency, some variables were transformed when needed:
- All BNP/NT-proBNP values were given as pg/mL. To convert NT-proBNP levels expressed in pmol/L to pg/mL, values were multiplied by 8.457.
- Time from symptoms onset to blood sample collection was given as hours.
- Time from onset to death assessment was given as days.
- Neurological severity was assessed by NIHSS. When SSS (Scandinavian Stroke Scale) was used, scores were transformed to NIHSS by the formula:
NIHSS = 25.68-0.43*SSS.e7
Tables
Table e-1. Quality score report for each included study.
Ref. / 1 / 2 / 3 / 4 / 5 / 6 / 7 / 8 / 9 / 10 / 11 / 12 / 13 / 14 / 15 / Quality scoree8 / + / + / + / - / - / + / + / + / + / - / + / - / + / + / + / 11
e9 / + / + / + / - / - / + / - / + / + / + / + / + / + / + / + / 12
e10 / + / + / + / - / + / - / - / + / + / + / + / - / - / + / + / 10
e6 / + / + / + / - / - / + / - / - / + / + / + / - / + / + / + / 10
e11 / ? / ? / + / - / - / + / + / - / - / - / + / - / + / - / - / 5
e12 / + / + / + / - / - / + / ? / + / - / ? / + / - / - / + / + / 8
e4 / + / + / + / - / - / - / - / - / + / + / + / - / + / + / + / 9
e5 / + / + / + / + / - / - / - / + / - / - / + / - / - / + / + / 8
e13 / ? / ? / + / - / - / + / ? / + / + / - / + / - / - / + / + / 7
e14 / + / + / + / + / - / + / ? / - / - / + / + / - / + / + / + / 10
e3 / + / + / + / - / - / - / - / - / + / + / + / - / + / + / - / 8
e15 / + / + / + / + / + / - / ? / + / + / + / + / - / + / + / + / 12
e16 / + / + / + / - / + / - / ? / + / + / + / + / - / + / + / + / 11
e17 / + / + / + / + / + / + / - / + / + / + / + / - / + / + / + / 13
e18 / ? / ? / + / - / - / + / - / + / + / ? / ? / - / + / + / + / 7
e19 / + / - / + / - / - / - / - / + / - / - / + / - / + / + / + / 7
# 1-15 referred to items evaluated for quality assessment (see Appendix 2). +: yes; -: no; ?: insufficient information.
Only positive answers accounted for quality score (i.e. negative answers to items did not penalize).
Table e-2. Additive predictive value of BNP/NT-proBNP over clinical variables for ischemic stroke mortality.
Only clinical / Clinical + BNP / Only clinical / Clinical + NT-proBNP
Logistic regression / NIHSS admission / 1.17 (1.13-1.22), p<0.001 / 1.17 (1.12-1.21), p<0.001 / 1.11 (1.08-1.13), p<0.001 / 1.10 (1.07-1.13), p<0.001
Age / 1.03 (1.01-1.06), p=0.015 / 1.03 (1.00-1.05), p=0.076 / 1.05 (1.03-1.07), p<0.001 / 1.04 (1.02-1.06), p<0.001
BNP / - / 2.30 (1.32-4.01), p=0.003 / - / -
NT-proBNP / - / - / - / 2.63 (1.75-3.94), p<0.001
Categorical NRI / NRI events / - / -1.4% / - / -1.9%
NRI
non-events / - / 0.6% / - / 10.0%
NRI / - / -0.9% (-7.4-5.7) / - / 8.1% (2.8-13.3)
p-value / Ref. / 0.794 / Ref. / 0.003
IDI statistics / IDI events / - / 0.017 / - / 0.024
IDI
non-events / - / 0.001 / - / 0.003
IDI / - / 0.018
(0.003-0.034) / - / 0.028
(0.013-0.043)
p-value / Ref. / 0.020 / Ref. / <0.001
ROC curves / AUC / 0.840
(0.815-0.863) / 0.848
(0.823-0.870) / 0.752
(0.722-0.781) / 0.774
(0.745-0.802)
p-value / Ref. / 0.285 / Ref. / 0.029
Logistic regression models gave odd-ratios (OR) adjusted by gender with 95% CI and p-value for each included variable. Biomarkers were added to clinical logistic regression model using highest quartile cut-off point. Bootstrapping gives 95% CI for OR of BNP (1.35-4.29) and NT-proBNP (1.79-4.04), which increases consistency to our regression results.
NRI: net reclassification improvement index (risk categories used: ≤10%, 10-90% and >90%); percentage of reclassification given for both events (i.e. patients who died) and non-events and for the sum of both (with 95% CI) when biomarker was added to clinical predictive model. IDI: integrated discrimination improvement index; index given for both events and non-events and for the sum of both (with 95% CI). AUC: area under ROC Curve; area with 95%CI given for each model. Clinical model always used as reference model to compare. NIHSS: National Institutes Health Stroke Scale.
6
Figures
Figure e-1. Mortality association of BNP/NT-proBNP depending on ischemic stroke etiologies.
Median and interquartile range for normalized BNP (A) and NT-proBNP (B) in alive (white boxes) and dead (grey boxes) patients, depending on TOAST etiology.
* p<0.05; ** p<0.01; *** p<0.001 (among etiologies; alive patients); $ p<0.05; $$ p<0.01; $$$ p<0.001 (between alive and dead patients). All p-values given after Bonferroni correction. There were not differences among etiologies for dead patients.
6
Figure e-2. Mortality association of BNP/NT-proBNP depending on time of sample collection.
Median and interquartile range for normalized BNP (A) and NT-proBNP (B) in alive (white boxes) and dead (grey boxes) patients, depending on time of sample collection.
* p<0.05; ** p<0.01; *** p<0.001 (among times; alive patients); $ p<0.05; $$ p<0.01; $$$ p<0.001 (between alive and dead patients). All p-values given after Bonferroni correction. There were not differences among times for dead patients.
e-References
e1. Whiteley W, Tseng MC, Sandercock P. Blood biomarkers in the diagnosis of ischemic stroke. A systematic review. Stroke 2008;39:2902-2909.
e2. Whiteley W, Chong WL, Sengupta A, Sandercock P. Blood markers for the prognosis of ischemic stroke: a systematic review. Stroke 2009;40:e380-e389.
e3. Shibazaki K, Kimura K, Iguchi Y, Aoki J, Sakai K, Kobayashi K. Plasma brain natriuretic peptide predicts death during hospitalization in acute ischaemic stroke and transient ischaemic attack patients with atrial fibrillation. Eur J Neurol 2011;18:165-169.
e4. Shibazaki K, Kimura K, Okada Y, et al. Plasma brain natriuretic peptide as an independent predictor of in-hospital mortality after acute ischemic stroke. Inter Med 2009;48:1601-1606.
e5. Idris I, Hill R, Ross I, Sharma JC. N-terminal probrain natriuretic peptide predicts 1-year mortality following acute stroke: possible evidence of occult cardiac dysfunction among patients with acute stroke. Age Ageing 2010;39:752-755.
e6. Sharma JC, Ananda K, Ross I, Hill R, Vassallo M. N-terminal probrain natriuretic peptide levels predict short-term poststroke survival. J Stroke Cerebrovasc Dis 2006;15:121-127.