State by slot not being fetched correctly

### Describe the bug

Starting attempts to fetch chain state by slot fails before chain has synced to the said slot (expected) but also after (unexpected).

### Expected behavior

The BN returns the state at slot after successful sync.

### Steps to reproduce

1. Request chain state for a slot (i.e. at `POST /eth/v1/beacon/states/{state_id}/validators`) couple of seconds before the chain is synced to the slot.
2. Continue retrying each 0.5-1 seconds.
3. The chain syncs to this slot after couple of seconds.
4. The responses to the request are still unknown state.

### Additional context

I have reproduced the issue on local Kurtosis devnet, running Charon DV. The setup I have is geth v1.15.5, lodestar CL v1.28.1, Charon v1.3.0, lodestar VC v1.28.1. I have added more logging to v1.28.1 tag and ran lodestar CL from this branch https://github.com/KaloyanTanev/lodestar/pull/1/files. I attach logs from the run.

[lodestar.log](https://github.com/user-attachments/files/19684297/lodestar.log)

### Operating system

macOS

### Lodestar version or commit hash

v1.28.1

Start by reproducing the request sequence against the local Kurtosis devnet using POST /eth/v1/beacon/states/{state_id}/validators, retrying until the requested slot is synced. Review the state-fetching path for why responses remain unknown after sync; done when the endpoint returns the state at that slot after successful synchronization.

Written by the indexing model from the issue text.