Introduction: For chronic prosthetic joint infections (PJI), complete removal of the infected prosthesis is necessary in order to cure the infection. Unfortunately, a subgroup of patients is not able to undergo a revision surgery due to high surgical risk. Alternatively, these patients can be treated with antibiotic suppressive therapy (AST) to suppress the infection. Aim: To evaluate the efficacy and tolerability of AST. Methods: We retrospectively collected data (period 2009-2015) from patients with a PJI (of hip, knee or shoulder) who were treated with AST at the University Medical Center Groningen, the Netherlands. AST was defined as antibiotic treatment for PJI that was started after the usual 3 months of antibiotic treatment. The time of follow-up was defined from the time point AST was started. Treatment was considered as failed, when the patient still experienced joint pain, when surgical intervention (debridement, removal, arthrodesis or amputation) was needed to control the infection and/or when death occurred due to the infection. Results: We included 21 patients with a median age of 67 years (range 21 - 88) and with a median follow-up of 21 months (range 3 - 81). Coagulase negative staphylococci (CNS) (n=6), S. aureus (n=6) and polymicrobial flora (n=4) were the most frequently found causative pathogens. Most patients with CNS and S. aureus were treated with minocycline (67%) and clindamycin (83%) as AST, respectively. Overall, treatment was successful in 67% of patients. Failure was due to persistent joint pain (n=1), surgical intervention because of an uncontrolled infection (n=3), and death due the infection (n=3). We observed a treatment success of 90% in patients with a 'standard' prosthesis (n=11), compared to only 50% in patients with a tumor-prosthesis (n=10). Also, treatment was successful in 83% of patients with a CNS as causative microorganism for the infection, compared to 50% in patients with a S. aureus. Patients who failed on AST had a higher ESR in comparison to patients with a successful treatment (mean 73 ± 25SD versus 32 ± 19SD mm/hour (p = 0.007), respectively. 43% of patients experienced side effects and led to a switch of antibiotic treatment or a dose adjustment in almost all of these patients. Conclusions: Removal of the implant remains first choice in patients with chronic PJI. However, AST is a reasonable alternative treatment option in a subgroup of patients with a PJI who are no candidate for revision surgery, in particular in patients with a 'standard' prosthesis and/or CNS as the causative micro-organism.