Clarifying the Longitudinal Interpretation of Wind-Up Ratio After Lower-Limb Amputation [Response to Letter] Authors: Jana Wagner, Angela Knaf-Serian, Gregor Reiter, Jamila Andoh, Herta Flor Published: 2026 Dear editor We thank the authors for their careful reading of our manuscript, “Sensory Sensitivity Over Time After Lower Limb Amputation and Its Association with Phantom Limb Pain,” and welcome the opportunity to clarify these points.1,2 Regarding participant ID 6, we would like to clarify that this participant did not withdraw consent from the study. She was unable to continue with in-person study procedures, including the T2 QST assessment, for health reasons, but remained willing to complete the clinical questionnaires at T2. A T2 Phantom Limb Pain Severity score and T2 medication data are therefore available and reported in Table 1, whereas no T2 QST data, including wind-up ratio (WUR), exist for this participant. ID 6 was accordingly excluded from Figure 3, Table 4, and all longitudinal QST analyses, but retained in Table 1 to provide a complete clinical picture of the cohort. We take this opportunity to clarify that the corresponding statement in the Limitations section of the published manuscript refers to this participant’s inability to continue with in-person study procedures, rather than a formal withdrawal of consent. We confirm the reconstruction of the grouping underlying Table 4. Unlike Figure 3 and Table 2, which classify participants by concurrent phantom limb pain (PLP) status at each time point separately, so that group membership and size can differ between T1 and T2, the paired comparisons in Table 4 used PLP status anchored at baseline (T1). The PLP+ trajectory group in Table 4 therefore comprises the five participants classified as PLP+ at T1 (IDs 2, 3, 5, 7, 9), and the PLP- trajectory group comprises the five T1 PLP- completers with T2 QST data (IDs 1, 4, 8, 10, 11); ID 6, who was PLP- at T1, is excluded due to missing T2 QST. This distinction was not stated in the Statistical Analysis section or the Table 4 note. We clarify here that Table 4 reflects trajectories defined by baseline PLP status, with group membership as given by the participant identifiers above. We agree that demonstrating a significant within-group change in WUR in the PLP+ group (P = 0.003) alongside a non-significant change in the PLP- group (P = 0.199) does not, by itself, establish that the change differs between the two groups. To address this directly, we computed the individual change in WUR (T2 minus T1) for each participant in the two baseline-anchored groups described above and compared them using Welch’s t-test for unequal variances. The mean change was 0.114 in the baseline PLP+ group and 0.098 in the baseline PLP- group, and this difference was not statistically significant (t(4.64) = 0.241, P = 0,820, 95% CI [0.016, - 0.17508]). We agree that this direct between-group comparison, rather than the separate within-group tests alone, is the appropriate way to evaluate whether WUR trajectories differ according to baseline PLP status. On this basis, the longitudinal WUR finding is best regarded as suggestive of divergent trajectories within the small PLP+ subgroup, but not confirmed by a direct test of the group-by-time interaction, particularly given the limited sample size. The T2 WUR comparison between PLP+ and PLP- (t(8) = 2.057) and the T1 warm detection threshold comparison (t(8) = 1.851) were based on one-tailed calculations. We therefore re-examined all remaining pairwise comparisons in Tables 2 and 3 to ensure consistency with the Methods. This review confirmed that the PLP+ versus PLP- comparisons in Table 2 were calculated one-tailed, whereas the other comparisons in Table 2 (healthy controls versus PLP+ and healthy controls versus PLP-) and all comparisons in Table 3 were two-tailed. For the five measures assessed using Student’s t-test (WDT, HPT, MDT, MPT, and WUR), the corresponding two-tailed P-values were: WDT, 0.101 (T1) and 0.348 (T2); HPT, 0.122 and 0.740; MDT, 0.596 and 0.107; MPT, 0.700 and 0.566; and WUR, 0.263 and 0.074. None reached significance at α = 0.05 under two-tailed testing, including the T2 WUR comparison (originally P = 0.037). For the three Mann–Whitney U-test measures (MPS, DMA, and PPT), the original one-tailed P-values ranged from 0.227 to 0.600, and none can reach significance under two-tailed testing. Importantly, this correction does not alter the conclusions for these measures. Overall, the re-analysis provides a consistent and conservative statistical assessment of the reported comparisons. We thank the authors of the Letter again for their careful and constructive reading, which we believe has meaningfully improved the transparency and rigor of our reporting. Disclosure The authors report no conflicts of interest in this communication. References