The study team observed that higher lactic acid on the palmar surface was associated with lower visual hand scaling scores. Participants (n = 49; aged 20–49 years) washed hands with a standardized soap, acclimated under controlled conditions (20°C, 40% RH), and then underwent sampling. Hand scaling was scored on a five-level visual scale derived from flatbed-scanner images (1 = none, 5 = severe). The authors reported negative Spearman correlations between lactic acid content and the scaling score, and lactic acid remained associated with lower scaling in partial correlation models that adjusted for other measured parameters (significant in two models, trend-level in one model). These findings are based on water-extractable lactic acid captured by filter-paper sampling and quantified by LC–MS; the source did not report direct surface concentration or secretion rates beyond the described sampling method.
Anti-pathogenic activity of palmar extracts was quantified as log10 reduction against influenza A/H3N2, Escherichia coli NBRC3301, and Staphylococcus aureus NBRC13276. Higher anti-pathogenic activity correlated with lower scaling in univariate analyses. When the investigators introduced lactic acid into partial correlation models, the simple correlation between anti-pathogenic activity and scaling was attenuated, indicating that measured lactic acid may account for part of the association between functional pathogen inactivation and skin roughness. The assay used ethanol-extracted surface components, dried and reconstituted, and therefore reflects in vitro extract-based activity rather than in situ, on-skin killing dynamics.
Total stratum corneum ceramides and total amino acids were not significantly associated with the hand scaling score. The authors did observe negative univariate correlations for several ceramide classes (NS, ADS, and AS) with scaling, but they did not pursue class-level independence analyses within multivariable models, stating that the multivariable focus was on surface lactic acid. Amino acid totals likewise failed to show significant associations with scaling. Quantification of ceramides used reverse-phase LC–MS on tape-stripped stratum corneum normalized to μg-protein; amino acids were measured via ion-exchange chromatography. Detailed class distributions are provided in the supporting tables, and the manuscript notes that class-level architecture of epidermal lipids was beyond the primary scope.
Sampling followed a defined protocol. Surface lactic acid was collected by placing three round filter papers on the palm, adding 60 μL ultrapure water to each, leaving them for 5 minutes, desiccating, reconstituting, and analyzing by LC–MS. Anti-pathogenic extracts were obtained by swabbing a defined palmar region with 50% ethanol; split aliquots were used for antibacterial and antiviral assays after drying and reconstitution. Bacterial assays monitored growth in SCD or LB media for E. coli and S. aureus, respectively, whereas antiviral activity against influenza A/H3N2 was assessed in MDCK cells by neuraminidase-based readouts and focus-forming assays. Stratum corneum ceramides and amino acids were extracted by tape stripping, solvent partitioning, hydrolysis for protein-derived amino acids, and quantified by LC–MS and an amino acid analyzer; results were normalized to protein content.
The investigators used Spearman’s rank correlations and partial correlations to evaluate associations, applying Benjamini–Hochberg correction to control the false discovery rate. Partial correlations conditioned on all other measured factors when estimating the association between two variables; the authors explicitly noted that partial correlations were reported as descriptive associations and not evidence of mediation. With N = 49, the study team reported approximately 80% power to detect bivariate correlations of |ρ| ≈ 0.38 and a minimum detectable |ρp| ≈ 0.40 for five-variable partial models. Statistical significance was defined as P < 0.05, and P < 0.1 was interpreted as a trend. All analyses were performed in R, and the supporting information includes a de-identified participant-level dataset and the R code used for analyses.
The authors emphasized that the cross-sectional design precludes causal inference; associations do not establish whether increased lactic acid reduces scaling or whether lower scaling permits accumulation of lactic acid on the palm. The measured lactic acid reflects water-extractable surface content under the specified sampling conditions and does not quantify secretion rates or absolute surface concentrations beyond those methods. The team noted that although several ceramide classes correlated with scaling in univariate tests, class-level independence was not established in multivariable models. Sample size and the detectable effect-size thresholds were reported; the study was powered to detect moderate correlations but may have been underpowered for smaller effects. These points outline concrete unanswered questions suitable for mechanistic or interventional follow-up studies rather than high-level prognostic statements.