For creatinine, there is peer-reviewed data demonstrating that in people without renal disease, increasing the protein in your diet tends to improve your eGFR, not decrease it (which is what happened here). So for him, it either fluctuated or was a marker of true kidney disease, but in any case this is not just a marker of adequate protein intake.
Also, while you may very well be right that GFR would improve, the little bits of reading I've had time to do indicates that unless they did eGFR by measuring his urine over 24 hours, then they likely estimated it from his creatinine levels, and so anything increasing creatinine levels would presumably cause the eGFR value to drop whether or not the increased creatinine levels are due to renal function or anything else.
Also, creatinine is measured in serum to estimate gfr. Urine creatinine is more used as part of a test to look for nephrotic syndrome.
Creatinine is produced constantly, filtered and not resorbed, without active secretion, which is why it is an excellent proxy for glomerular filtration rate.
It could be just fluctuating in his case, or it could very well be, given his past diet (how could he survive on that), that he was finally getting decent amounts of protein.