LAUSR

In Response: We read with great interest a meticulous letter written by Li et al in response to our recent article entitled, “NfL (Neurofilament Light Chain) Levels as a Predictive Marker for Long-Term Outcome After Ischemic Stroke” The authors encouraged us to comment on potential confounders, measurement of additional biomarkers, and temporal evolution of serum NfL (sNfL) upon specific treatments. We fully agree with the authors that potential confounders might explain, at least in part, the observed predictive value of sNfL for short-term functional outcome and long-term cardiovascular survival. Indeed, we reported the following confounders with regard to functional short-term outcome at the univariate level in Table I in the online-only Data Supplement: age, arterial hypertension, diabetes mellitus, coronary artery disease, National Institutes of Health Stroke Scale at admission, and age-related white matter changes, which were all included in multivariate logistic regression analysis. Because of the small sample size and limited occurrence of outcome events, we decided not to include details of medications such as antiplatelet drugs and antidiabetic drugs, or elevated hemoglobin A1C levels and increased fasting glucose, as we explained them by the more frequent occurrence of coronary artery disease and diabetes mellitus within patients with an unfavorable functional outcome, both representing cardiovascular risk factors for which we already corrected in multivariate logistic regression analysis as previously mentioned. With regard to further inflammatory biomarkers, we are not able to report neutrophil-to-lymphocyte ratio, as a differential blood count was not routinely performed within the Find-AF study. Nonetheless, we are able to provide data on C-reactive protein, which was found not to differ between patients with unfavorable and good functional outcome (please see Table I in the online-only Data Supplement). Moreover, we concede that by investigating the temporal evolution of sNfL levels in various clinical situations, such as mechanical thrombectomy and intravenous thrombolysis, we would get further insight into the role of sNfL as a marker for neuroaxonal injury within these distinct patient cohorts. Unfortunately, we are not able to provide this information within the Find-AF study. We agree with Li et al that the addition of sNfL to the combination of clinical variables (diabetes mellitus, National Institutes of Health Stroke Scale at admission, and age-related white matter changes-rating) showed a trend towards an improvement of the area under the curve in receiver operating characteristic analysis (0.792–0.820, P=0.052) with regard to the rating of longterm cardiovascular survival. This effect might also be explained by the low sample size of our study cohort. Thus, the potentially limited added value of sNfL compared with much cheaper clinical variables cannot be answered finally within our study cohort and needs clarification in future studies. In summary, the concerns described in the letter by Li et al call for further validation studies of sNfL, especially with regard to temporal evolution within specific clinical situations and patient cohorts. Nonetheless, we once more want to highlight the value of nervous-system specific NfL as a marker for neuroaxonal tissue damage within the central nervous system, which is superior to less nervous system–specific markers such as brain natriuretic peptide, and represents a promising biomarker within various neurological diseases including stroke.