GBM data show faster mean spin loss in 4U 1538-52 in January-October 2026 than in 2009-2018

We compare the observed 2026 spin loss of the known X-ray pulsar 4U 1538-52 with its 2009-2018 decline. We use public Fermi GBM frequency estimates from a fixed October 2026 data set. We check two FITS formats and remove one exact repeated row. From 13 January to 9 October 2026, the frequency drops by 1.824 microhertz in 268.51 days. The mean loss rate between these endpoints is 7.86e-14 Hz/s. This is 4.37 times the historical endpoint rate. All 27 endpoint and date-range checks give faster recent loss, with ratios from 3.19 to 5.24. The period increases by 0.5057 s. The fixed nominal 2023 period model permits at most 0.0739 s over these endpoints. This bound applies to any sine period or phase. That fixed extension cannot explain the measured change. A permanent torque change is not established. A model with random walk noise gives a much wider drift error than a model that also permits independent scatter. The target and intervals were selected after inspection. The monitor results are preliminary. The error estimates omit binary-model uncertainty. The package contains fixed inputs, source records, code, figures, numerical checks, and instructions to repeat the work without network access.

4U 1538-52 · Fermi GBM · accreting pulsar · spin loss · torque noise · reproducible timing