POLG A467T Substrate Crossover Is Not Identified by Marginal Kinetic Intervals

**Objective:** Determine which comparisons of PZL-A-treated human POLG A467T with untreated wild type survive a declared sensitivity set at the eight deposited dNTP concentrations. **Design:** Exploratory incremental reanalysis of published active-complex-normalized incorporation summaries and existing Michaelis–Menten fits, using exact equality and the Cartesian product of four marginal kinetic intervals. **Results:** The deposited point fits give equality at 0.389815 µM dNTP. Nine of sixteen endpoint scenarios cross at positive substrate, but other allowed scenarios never cross. Monotonicity gives continuous-box ratio bounds without treating corner frequencies as probabilities. Every box curve is below untreated wild type for substrate greater than 8.331306 µM; only 10 µM among the eight deposited conditions meets that criterion. None guarantees above-WT incorporation. At 10 µM the point ratio is 0.6405 and the box range is 0.4180–0.9680. **Conclusions:** The marginal summaries identify a conditional below-WT direction at 10 µM, while they do not identify a persistent low-substrate crossover. Joint-fit covariance and active-complex normalization uncertainty remain unavailable; this sensitivity envelope is not a confidence region, a correction of the source analysis, or evidence of cellular restoration.