To report our experience with the use of local antibiotic co-delivery with a synthetic bone graft substitute during a second stage re-implantation of an infected proximal humeral replacement.

A 72 year old man was admitted to our department with a pathological fracture through an osteolytic lesion in the left proximal humerus, due to IgG Myelomatosis. He was initially treated with a cemented proximal humerus replacement hemiarthroplasty. Peri-prosthetic joint infection (PJI) with significant joint distention was evident three weeks post operatively. Revision surgery confirmed presence of a large collection of pus and revealed disruption of the soft tissue reattachment tube, as well as complete retraction of rotator cuff and residual capsule. All modular components were removed and an antibiotic-laden cement spacer (1.8g of Clindamycin and Gentamycin, respectively) was implanted onto the well-fixed cemented humeral stem. Initial treatment with i.v. Amoxicillin/Clavulanic acid was changed to Rifampicin and Fusidic Acid during a further 8 weeks after cultures revealed growth of S. epidermidis. During second stage revision, a hybrid inverse prosthesis with silver coating was implanted, with a total of 20 ml Cerament ™G (injected into the glenoid cavity prior to insertion of the base plate and around the humeral implant-bone interface) and again stabilized with a Trevira tube. Unfortunately, this prosthesis remained unstable, ultimately requiring re-revision to a completely new constrained reverse prosthesis with a custom glenoid shell and silver-coated proximal humeral component.

18 months postoperatively, the patient's shoulder remains pain free and stable, without signs of persistent or reinfection since the initial second stage revision. The function however, unfortunately remains poor.

This case report illustrates the application of an antibiotic-eluting bone graft substitute in a specific clinical situation, where co-delivery of an antibiotic together with a bone remodeling agent may be beneficial to simultaneously address PJI as well as poor residual bone quality.

To document early in-vivo concentrations of gentamicin in plasma and drain fluid after bone defect reconstruction using a gentamicin-eluting bone graft substitute.

Introduction

Reconstruction of bone defects after surgical bone tumor resection is associated with an increased risk of infection and some surgeons therefore prefer extended antibiotic prophylaxis in these patients. A gentamicin-eluting bone graft substitute consisting of sulphate and apatite has been shown to be effective for treatment of osteomyelitis(1) and may be a valuable addition to the therapeutic and/or prophylactic antibiotic regime for this and many other indications.

We performed a prospective pilot study from December 2014 to February 2015 in 7 patients (M/F: 4/3, mean age 51 (37–79) years) who underwent bone defect reconstruction with a gentamicin-eluting bone graft substitute (CERAMENT™|G – BONESUPPORT AB) containing 175 mg gentamicin per 10 mL. Indications for surgery were metastatic bone disease (n=3, proximal humerus), giant cell tumor (n=2, distal femur), aseptic prosthetic loosening (n=1, knee) and chondroid tumor (n=1, distal femur). Additional endoprosthetic reconstruction with a tumor prosthesis was performed in 3 patients (2 proximal humerus and 1 distal femur). Drain fluid and plasma was collected immediately postoperatively and each postoperative day until the drain was removed. In 2 cases we were unable to collect drain fluid directly postoperatively due to minimal fluid production. Gentamicin concentrations were analyzed using an antibody technique (Indiko™ – Thermo Scientific).

A mean of 14 (10–20) mL gentamicin-eluting bone graft substitute was used, either alone or in combination with cancellous allograft and/or a bone graft substitute not containing gentamicin (CERAMENT™|BVF – BONESUPPORT AB). Mean drain fluid concentrations of gentamicin were 1200 (723–2100) mg/L immediately postoperative (0–2 hours), 1054 (300–1999) mg/L on day 1 (17–23 hours) and 509 (38–1000) mg/L on day 2 (39–45 hours). Mean plasma concentrations of gentamicin were 1.26 (1.08–1.42) mg/L immediately postoperative, 0.95 (0.25–2.06) mg/L on day 1 and 0.56 (0.20–0.88) mg/L on day 2.

Discussion. As gentamicin induces a concentration-dependent bacterial killing effect, the obviously high local peak concentrations of gentamicin found in this study would be expected to deliver a substantial prophylactic effect after long operations with an increased risk of intraoperative bacterial contamination.

Local implantation of a gentamicin-eluting bone graft substitute for bone defect reconstruction results in high concentrations of gentamicin in the drain fluid in the first postoperative days and low plasma concentrations.