Can preoperative CRP levels predict infections of bipolar hemiarthroplasty performed for femoral neck fracture? A retrospective, multicenter study

Current recommendations urge us to operate quickly on femoral neck fractures to reduce the risk of comorbidity decompensation. In some cases, this leads us to operate when an underlying infection is present. In this study, we evaluated the infection rate of bipolar hemiarthroplasty after femoral neck fracture and attempted to relate it to preoperative C-reactive protein (CRP) levels.

The infection rate of bipolar hemiarthroplasty was evaluated in a dual-center, retrospective study of 260 patients over a 2-year period. During the first year, the preoperative CRP levels were not taken into account when scheduling the procedure. During the second year, if preoperative CRP levels were above 50 mg/L, the procedure was delayed to look for and treat any ongoing infections.

The overall periprosthetic infection rate in this study was 4.85 % (range 4.8–4.9), with 33 % of patients passing within 1 year due to the infection. In the group where CRP was not taken into consideration, 59 of the 143 operated away patients (41 %) had their preoperative CRP levels measured. Twenty-nine of these patients had CRP > 50 mg/L when they were operated. Of the seven infections in the group, one patient had CRP > 50 mg/L, two had CRP < 50 mg/L, and four patients did not have preoperative CRP levels measured. In the group where CRP was taken into consideration, 104 of the 117 patients (89 %) had their preoperative CRP assessment. Thirty of these patients had CRP > 50 mg/L upon admission; their procedure was delayed to determine the etiology of this CRP elevation. No cause was found in 16 of these 30 patients, and they were operated despite having CRP > 50 mg/L. There were five infections in this group: four patients had CRP > 50 mg/L and were treated accordingly; one patient had preoperative CRP < 50 mg/L. In patients where the preoperative CRP levels were taken into account, the delay before surgery was twice as long as those where CRP levels were not considered.

C-reactive protein (CRP) level alone is not a good preoperative predictive factor for periprosthetic joint infection, although 80 % of the patient with an infected bipolar hemiarthroplasty had CRP > 50 mg/L upon admission. The increased delay due to the CRP analysis is not without consequence for this cohort. Two previous studies have looked into the predictive ability of CRP levels, but these involved scheduled surgical procedures.

This study could not validate the use of CRP levels, nor a 50 mg/L threshold, as predictive factors for a preexisting infection during bipolar hemiarthroplasty for femoral neck fracture. As a consequence, new infection screening tools must be developed and validated.

Retrospective, no control group.