This retrospective study was undertaken to (i) define the most appropriate thresholds for serum d-dimer and fibrinogen for differentiating aseptic failure from periprosthetic joint infection (PJI) and (ii) evaluate the predictive value of our d-dimer and fibrinogen threshold compared to previously proposed thresholds. This observational cohort study included consecutive patients who had undergone total knee arthroplasty (TKA) revision between January 2019 and December 2020. International Consensus Meeting diagnostic criteria were used to identify patients affected by the prosthetic infection. Receiver operating characteristic curve analyses assessed the predictive value of the parameters, and the areas under the curves were evaluated. We included 125 patients with a median age of 69 years (53–82) affected by painful TKA. Fifty-seven patients (47%) had PJI. Patients with PJI had higher median d-dimer, fibrinogen, ESR, and CRP when compared to patients believed to be free of PJI. The best threshold values for d-dimer and fibrinogen were 1063 ng/ml (sensitivity 0.72, specificity 0.74) and 420 mg/dl (sensitivity 0.67 and specificity 0.82), respectively. A d-dimer level >1063 ng/ml combined with a fibrinogen level >420 mg/dl had a sensitivity of 0.52, and a specificity of 0.90. We found that an increased d-dimer beyond 1063 ng/ml showed a better predictive value than the previously proposed threshold. The combined determination of d-dimer and fibrinogen displayed high specificity and should be considered an excellent tool to rule out an infection. The accuracy of the proposed cutoffs is more effective than previously reported.