Does Preoperative Glycemic Control Restore Immune Defense Against Implant-related Infection in Mice with Diabetes?

BACKGROUND The presence of Type II diabetes is a well-established risk factor for bone and joint infection, especially in patients with poor glycemic control. However, few studies have investigated the effect of the duration of preoperative glycemic intervention. For patients with poor glycemic control, the effect of the duration of preoperative glycemic intervention remains unknown. Many glycemic biomarkers including hemoglobin A1c (HbA1c), fructosamine, and 1,5-anhydroglucitol have different response rates to glycemic change. It is unclear which biomarker is more closely related to the decrease in infection proportion after preoperative glycemic intervention. QUESTIONS/PURPOSES (1) Is there an effect of the duration of preoperative insulin therapy in mice with diabetes receiving an experimental intra-articular implant? (2) Of the three commonly used biomolecules for monitoring blood glucose levels (HbA1c, fructosamine, and 1,5-anhydroglucitol), is one more closely related to decrease in infection proportion after presurgical insulin therapy? METHODS With a well-established protocol, Type II diabetes was modeled in female 10-week-old C57BL/6 mice by maintaining them on a high-fat diet (60% fat) for 8 months; control mice without diabetes received a normal low-fat diet (10% fat). Mice with Type II diabetes were randomized into groups to receive preoperative glycemic intervention with insulin for 0, 1, 3, 5, 7, 14, or 28 days, and investigators were blinded to the randomization. Mice with and without diabetes then received a surgically inserted wire into the femoral canal in a retrograde fashion and received a local or systemic challenge with Staphylococcus aureus or Escherichia coli (n = 20 for each bacteria challenge [systemic or local]/timepoint). The proportion of culture-positive joint samples was calculated. An additional 10 mice with Type II diabetes were treated with insulin for 28 days and the HbA1c, fructosamine, and 1,5-anhydroglucitol levels were consecutively monitored. Fisher exact tests and nonparametric Wilcoxon rank sum tests were used to analyze the different between different groups, with p < 0.05 taken as significant. RESULTS When insulin therapy was administered, the proportion of bone and joint infections decreased in mice with Type II diabetes, reaching asymptotic levels after 3 days of treatment for the systemic (S. aureus: 7 of 20 mice with diabetes on 3-day therapy, p < 0.001; 8 of 20 on 5-day, p = 0.002; 10 of 20 on 7-day, p = 0.01; 9 of 20 on 14-day, p = 0.006; and 8 of 20 on 28-day, p = 0.002 versus 18 of 20 in the no insulin therapy group; E. coli: 6 of 20 on 3-day therapy, p = 0.004; 7 of 20 on 5-day, p = 0.01; 7 of 20 on 7-day, p = 0.01; 6 of 20 on 14-day, p = 0.004; and 7 of 20 on 28-day, p = 0.01 versus 16 of 20 in the no insulin therapy group) or local bacterial challenge (S. aureus: 11 of 20 on 3-day therapy, p = 0.001; 12 of 20 on 5-day, p = 0.003; 10 of 20 on 7-day, p < 0.001; 12 of 20 on 14-day, p = 0.003; and 13 of 20 on 28-day, p = 0.008 versus 20 of 20 in the no insulin therapy group; E. coli: 10 of 20 on 3-day therapy, p = 0.003; 10 of 20 on 5-day, p = 0.003; 9 of 20 on 7-day, p = 0.001; 11 of 20 on 14-day, p = 0.008; and 10 of 20 on 28-day, p = 0.003 versus 19 of 20 in no insulin therapy group). Even after 28 days of insulin therapy, the proportion of bone and joint infections was still higher (statistically insignificant with large absolute difference, except for one instance) in mice with diabetes than in control mice without diabetes after systemic (S. aureus: 8 of 10 mice with diabetes on 28-day therapy versus 4 of 20 mice without diabetes, p = 0.30; E. coli: 7 of 20 on 28-day therapy versus 1 of 20 mice without diabetes, p = 0.04) or local challenge (S. aureus: 13 of 20 mice on 28-day therapy versus 8 of 20 mice without diabetes, p = 0.21; E. coli: 10 of 20 on 28-day therapy versus 5 of 20 mice without diabetes, p = 0.19). HbA1c and fructosamine levels were lagging indicators of the decrease in infection proportion after insulin treatment. In contrast, the 1,5-anhydroglucitol level increased quickly (reflecting lower blood glucose levels) in response to short-term glycemic control. Moreover, the time required for changes in 1,5-anhydroglucitol levels to be detected was no more than 3 days (3 days insulin therapy 1.86 ± 0.20 [95% CI -1.27 to -0.45]; p＜0.001 versus no insulin therapy 1.00 ± 0.11). CONCLUSION In a model of mice with Type II diabetes, prolonged preoperative glycemic intervention did not further reduce the proportion of bone and joint infections compared with that achieved with short-term intervention of 3 days. CLINICAL RELEVANCE Compared with HbA1c and fructosamine, 1,5-anhydroglucitol might be a better indicator for risk stratification and guiding the timing for elective surgery. Comparative study of these three biomarkers based on patient samples is warranted to further confirm this conclusion.