Patients
Thirty patients who had been managed with a PROSTALAC implant for the treatment of infection at the site of a shoulder arthroplasty between 2000 and 2006 were identified from the case logs of the three senior authors (J.M., L.H., and J.J.P.W.). All three surgeons were highly experienced fellowship-trained shoulder surgeons who followed the same protocol for the use of the PROSTALAC implant. With use of an institutional review board-approved protocol, the medical records were reviewed and patients who had less than one year of follow-up after the insertion of the PROSTALAC implant were excluded. Two of thirty consecutive patients were excluded, leaving twenty-eight patients available for inclusion in the present study.
The initial arthroplasty procedures included five hemiarthroplasties, fifteen total shoulder arthroplasties, two reverse total shoulder arthroplasties, one revision total shoulder arthroplasty, and five revision reverse total shoulder arthroplasties. Eight patients were women, and twenty were men. The average age was sixty-three years (range, thirty-one to eighty-two years). Nine patients presented from an outside hospital, and nineteen had the original procedure performed at one of our institutions. Ten patients underwent irrigation and debridement with retention of the prosthesis prior to placement of the PROSTALAC implant. The average duration of follow-up was 27.6 months (range, twelve to sixty-nine months).
Diagnosis
The clinical diagnosis of deep infection at the site of the shoulder arthroplasty was established on the basis of a combination of symptoms, physical findings, and laboratory tests. The laboratory values that were considered to be elevated were a white blood-cell count of >11.0 × 109/L; an erythrocyte sedimentation rate of >22 mm/hr, and a C-reactive protein level of >2.0 mg/L. The white blood-cell count, erythrocyte sedimentation rate, and C-reactive protein values were available for twenty-eight, twenty-three, and nineteen patients, respectively. Intraoperative gram-stain and culture results as well as histopathological findings were available for twenty-eight and twenty-one patients, respectively (Table I).
The average time from the index arthroplasty to the diagnosis of infection was 12.8 months (range, 0.5 to thirty months). According to the definitions of Strickland et al.8, six infections were acute (occurring less than three months postoperatively), fourteen were subacute (occurring three to twelve months postoperatively), and eight were late (occurring more than twelve months postoperatively). It is not known how many of the subacute and late infections were due to acute hematogenous infections. In addition, these definitions differ from those of infections following total hip and knee arthroplasty, which are defined as acute postoperative (occurring less than four weeks postoperatively), late chronic (occurring more than four weeks postoperatively), acute hematogenous, and intraoperative5.
Treatment
All patients underwent thorough debridement with removal of the original prosthesis and cement. A PROSTALAC spacer was made for each patient with a combination of antibiotic powder (typically 2 g of gentamicin and 3 g of tobramycin per 40-g bag of cement) and Simplex cement (Stryker, Kalamazoo, Michigan) (typically three 40-g bags with four polymethylmethacrylate monomers). The spacers were made without any additional structural support (three cases), around a one-third tubular plate (fourteen cases), or around a 3.5-mm limited contact-dynamic compression plate (eleven cases). The plate was bent to recreate the neck shaft angle of the prosthesis, and the length was made similar to a broach trial. The humeral head was created with a custom 44-mm mold (Figs. 1-A through 1-D).
With the exclusion of two patients who had a revision PROSTALAC implant, sixteen patients (57%) had a second-stage procedure. These procedures included ten reverse shoulder arthroplasties, three hemiarthroplasties, two total shoulder arthroplasties, and one resection arthroplasty. The remaining twelve patients declined additional procedures and retained the PROSTALAC implant. All patients were managed with culture-specific intravenous antibiotics in consultation with an infectious-disease specialist, as detailed in Table I. The average interval between the first and second-stage procedures was 8.9 months (range, two to twenty months).
Source of Funding
There was no external funding source for this investigation.
Preoperative and Intraoperative Findings
Preoperatively, the white blood-cell count was elevated in only one (4%) of the twenty-eight patients, the erythrocyte sedimentation rate was elevated in sixteen (70%) of twenty-three patients, and the C-reactive protein level was elevated in fifteen (79%) of nineteen patients. The average white blood-cell count, erythrocyte sedimentation rate, and C-reactive protein level were 7.4 ×109/L (range, 2.3 to 12.7 ×109/L), 52 mm/hr (range, 6 to 112 mm/hr), and 22 mg/L (range, 1 to 93 mg/L), respectively.
Positive culture results were found for twenty-four (86%) of the twenty-eight shoulders. Propionibacterium acnes was identified on culture of specimens from fifteen (54%) of the twenty-eight shoulders, and coagulase-negative Staphylococcus was identified on culture of specimens from thirteen shoulders (46%); a combination of both of these organism grew on culture of specimens from nine shoulders. Peptostreptococcus was identified on culture of specimens from three shoulders (11%); in all three cases, Propionibacterium acnes and coagulase-negative Staphylococcus also were identified on culture. Staphylococcus aureus was identified on culture of specimens from two shoulders (7%); in one of these two cases methicillin-resistant Staphylococcus aureus (MRSA) was identified. Corynebacterium was identified on culture of specimens from two shoulders (7%), and Serratia was identified on culture of specimens from one shoulder (4%). The cultures were routinely held for two weeks.
On histological examination, eighteen of twenty-one specimens showed signs of acute or acute and chronic inflammation. However, two of the patients with a negative finding on histological examination had a positive result on culture. The remaining patient, who had a negative result on both culture and histopathological examination, had exceptionally high inflammatory markers and clinical signs of infection.
The four patients who had a negative culture had other findings that were consistent with infection. The first patient had an erythrocyte sedimentation rate of 74 mm/hr and a C-reactive protein level of 62 mg/L. The second patient had an erythrocyte sedimentation rate of 25 mm/hr and a C-reactive protein level of 9 mg/L, with acute and chronic inflammation on histopathological analysis. The third patient had a C-reactive protein level of 4 mg/L and had acute inflammation on histopathological analysis. The last patient had an erythrocyte sedimentation rate of 20 mm/hr and had acute inflammation on histopathological analysis.
Complications and Subsequent Procedures
There were nine complications in eight shoulders, with the need for six additional procedures. One patient had a dislocation of the PROSTALAC implant; three patients had a fracture of the PROSTALAC implant; and five patients, including one with a PROSTALAC fracture, had recurrent infections. One of the five patients with an infection who had had successful conversion to a hemiarthroplasty had a reinfection and required a second PROSTALAC implant. All three fractures were associated with the use of the semitubular plate in the spacer (Fig. 2). In addition, the patient with the PROSTALAC dislocation had no remaining subscapularis muscle and had substantial glenoid bone loss after removal of the cemented component. Six additional procedures were performed; all were revisions to a PROSTALAC implant. One of these six procedures was performed because of a PROSTALAC fracture; four, because of recurrent infection with a current PROSTALAC implant in place; and one, because of a recurrent infection at the site of a second-stage hemiarthroplasty. Importantly, all four patients who had an infection with a PROSTALAC implant in place originally had had a subacute or chronic infection, with an average time to infection of 17.5 months. Three of these reinfections were with Propionibacterium acnes, and the fourth was with MRSA. The reinfection after a second-stage procedure was also with Propionibacterium acnes, with the original onset of infection occurring thirty months after the original procedure.
Functional Outcomes
Overall
The average forward flexion at the shoulder after the latest procedure for all patients was 77° (range, 30° to 130°). Fifteen patients indicated that they had mild or no pain, eight patients had moderate pain, and five patients had severe pain.
Retained PROSTALAC Cohort
With the inclusion of two patients who had revision with a second PROSTALAC implant, the average forward flexion for the twelve patients with a retained PROSTALAC implant was 83° (range, 40° to 130°). Four patients had no pain, six patients had mild pain, and two patients had moderate pain.
Reverse Total Shoulder Arthroplasty Cohort
The ten patients who had a reverse total shoulder arthroplasty as a second-stage procedure had an average forward flexion of 74° (range, 30° to 120°). Five had moderate pain, and five had severe pain.
Total Shoulder Arthroplasty and Hemiarthroplasty Cohort
The five patients who had total shoulder arthroplasty or hemiarthroplasty as a second-stage procedure had an average forward flexion of 61° (range, 35° to 110°). Four patients had mild pain, and one had moderate pain.
There have been few published reports evaluating the treatment of infection after shoulder arthroplasty. Consequently, there is no clearly defined treatment strategy. In addition, we are not aware of any studies that have focused on the potential complications and benefits of PROSTALAC use.
Two sentinel reports on multiple treatment cohorts resulted in similar conclusions. Sperling et al.2 and Coste et al.6 both found antibiotics or debridement alone to be an ineffective treatment. Resection arthroplasty was beneficial for eradicating infection in the majority of patients, but the functional outcomes were poor.
The two largest studies evaluating the use of the PROSTALAC implant in the setting of two-stage reimplantation demonstrated the procedure to be beneficial, but with suboptimal functional results8,9. In addition, those two studies demonstrated very different rates of reinfection. Strickland et al.8, in what we believe to be the largest report on two-stage reimplantation (seventeen patients), found eradication of infection in 63% of the patients, with an average forward flexion of 89°. It is not clear if the reinfection rates were higher in patients with acute, subacute, or late infections. Jerosch and Schneppenheim9 evaluated ten patients who received a PROSTALAC, with eight proceeding to second-stage reimplantation. None of the patients had recurrent infection, but all had some degree of functional limitation.
Two major studies evaluated the treatment of glenohumeral infection with one-stage arthroplasty without the use of a PROSTALAC implant7,10. Both studies showed good results. Ince et al.10 evaluated sixteen patients who were managed with one-stage reimplantation for the treatment of acute infection (two patients) or late infection (fourteen patients) and found no recurrence of infection and reasonable function, with an average Constant score of 33.6. However, that study was limited by a small cohort size and the loss of seven of the sixteen patients to long-term follow-up. Cuff et al.7 compared one-stage with two-stage arthroplasty and found no recurrent infections in either group, with similar limited functional results, even with the use of reverse total shoulder arthroplasty for second-stage procedures. Importantly, however, five of the twenty-one patients had native shoulder infections before the initial arthroplasty was done7.
Together, these limited data point to the need for explantation in the setting of infection, regardless of timing, but they do not show a clear advantage in association with one-stage or two-stage reconstruction or PROSTALAC use. All studies demonstrated functional outcomes to be worse than for primary arthroplasty.
To our knowledge, the present report represents the largest study on the use of the PROSTALAC for the treatment of infection following shoulder arthroplasty. The findings suggest that the use of a PROSTALAC implant for possible two-stage reimplantation can successfully allow for the eradication of infection and the recovery of reasonable, but decreased, function. In addition, a shoulder PROSTALAC implant alone can provide enough stability and pain relief that reimplantation often is not necessary. This was also seen in the report by Jerosch and Schneppenheim9, in which two of ten patients retained the PROSTALAC implant and did not require a second-stage procedure. However, similar to the findings reported by Strickland et al.8, Coste et al.6, and Jerosch and Schneppenheim9, the ultimate function is limited in comparison with that of primary total shoulder replacements and the recovery of function is not comparable with that after two-stage reimplantation of total hip and total knee replacements5. In our series, five of the twenty-eight patients had severe pain and the average forward flexion was only 77°. The difference likely lies in the importance of the limited function of the rotator cuff and adjacent soft tissue after multiple surgical procedures.
Importantly, all cohorts (patients with retained PROSTALAC implants, those managed with reverse total shoulder arthroplasty, and the combined group managed with hemiarthroplasty or total shoulder arthroplasty) had similar forward flexion of the shoulder at the time of the latest follow-up. Although reverse total shoulder arthroplasty does not require an intact rotator cuff for motion, patients who had a second-stage reverse shoulder arthroplasty did not have increased forward flexion compared with other groups. In fact, patients in this group had more pain than those with a retained PROSTALAC implant or a hemiarthroplasty or total shoulder arthroplasty. Although the importance of these findings is not entirely clear, it appears that reverse total shoulder replacement does not have a distinct advantage over other treatment options. The detrimental effects of infection and multiple surgical procedures are likely more important to outcomes than the final implant type is.
Use of the PROSTALAC implant does not always lead to eradication of infection, as we noted in four patients who required revision debridement and PROSTALAC placement. However, it is important to note that, after revision with a PROSTALAC implant, these four infections were eradicated. The remaining infection was a recurrence only after completion of the second stage, which points to the tenacity of these often indolent infections. Specifically, four of the five recurrent infections were due to Propionibacterium acnes and were diagnosed at an average of 17.5 months after the index arthroplasty.
In addition, PROSTALAC implants can fracture or dislocate, as we observed in the cases of four patients (only one of whom needed a revision). All three fractures occurred around implants with a one-third tubular plate, and therefore we now use a limited contact-dynamic compression plate scaffold. We cannot explain why three implants without a scaffold did not fracture.
Importantly, 43% of patients declined a second-stage procedure because they were satisfied with the PROSTALAC in terms of pain relief and function. Although we do not advocate leaving the PROSTALAC implant in place, its use provides another important treatment option in these complex cases.
Last, consistent with the findings of previous studies, the diagnosis of infection is difficult to make because markers such as C-reactive protein and the erythrocyte sedimentation rate are not always elevated8. In addition, the organisms that most commonly grew on culture were Propionibacterium acnes and coagulase-negative Staphylococcus, which are indolent infections that do not always present with overt signs and symptoms and do not always show growth on culture. It is important to note that all four patients without growth on culture had elevated inflammatory markers, with three showing inflammation on histopathological analysis. For this reason, it is important always to consider infection as a possible diagnosis in patients with an unsuccessful arthroplasty2,7,8. It may be necessary to keep cultures for up to two weeks for indolent organisms like Propionibacterium acnes11, although this will inherently increase the risk of contaminants.
The present study is inherently limited by its retrospective nature. In addition, there was no comparison group (specifically, one-stage reimplantation without the use of a PROSTALAC implant). However, the current series points to the effectiveness of PROSTALAC use in terms of eradicating infection following shoulder arthroplasty and in providing reasonable, yet limited, function with either retention or second-stage reconstruction. A large, multicenter comparison of single and two-stage reimplantation with a clear delineation of early postoperative, late chronic, and acute hematogenous infection is needed.
Nonetheless, the data show that either one-stage reconstruction or the use of a PROSTALAC implant for possible two-stage reconstruction is effective. One-stage procedures may have similar reinfection rates to two-stage reconstruction7, although the reason for this is unclear and inconsistent with the hip and knee arthroplasty literature and general surgical principles. Use of the PROSTALAC implant for possible two-stage reimplantation is a more conservative approach that has been well studied in patients managed with hip and knee arthroplasty and has been found to provide acceptable outcomes in terms of the eradication of infection and the restoration of function. In addition, PROSTALAC use may serve as an acceptable permanent treatment. However, any unnecessary complex secondary surgery should be avoided if the outcomes are not clearly different. At this time, the optimal treatment strategy is not clear.