Surgical Site Infections at Donor and Recipient Sites in Patients with Iliac Crest Harvesting For Autologous Bone Grafting - A Pilot Evaluation

Surgeons harvest the iliac crest for bone grafting. The epidemiology of surgical site infections (SSI) associated with this procedure at the donor, or recipient site, is unknown. We perform a retrospective pilot evaluation of adult patients undergoing first-time orthopedic surgery at the Balgrist University Hospital between 2014-2019. We excluded patients with infection at the index surgery, diabetic foot surgeries, superficial SSIs, and revision surgeries. We included 20,088 episodes of primary orthopedic surgery, of which 467 with iliac crest bone sampling (467/20,088; 2%). Only two iliac sites (2/467; 0.4%) become infected. In contrast, surgeries with iliac crest sampling yielded more SSIs at the recipient site than those without (1.9% vs. 0.8%; χ 2 -test; p<0.01). These patients equally revealed more co-morbidities such as a longer duration of surgery (median 127 vs. 79 minutes), when compared to the general orthopedic population. In multivariate logistic regression analysis with the outcome “ SSI at the recipient site ”, the iliac harvesting was independently associated with deep SSIs requiring surgical revision (odds ratio 2.1; 95% confidence interval 1.1-4.2). In our pilot evaluation with 20,088 primary orthopedic surgeries, the SSI risk of the iliac harvest site was low. In contrast, surgeries with supplementary iliac crest harvesting revealed a higher SSI risk than the general orthopedic population, potentially due to a mix of local independent risks of grafting together with a prolonged surgery time.


Introduction
Surgeons often restore osseous defects with autologous bone grafting, which is believed to be osteo-inductive, osteo-genetic and osteo-conductive [1]. The iliac crest represents the perfect donor site due to the ample quantity of retrievable bony graft and the possibility for tricortical grafts when structural support is needed [2]. However, iliac bone harvesting may be associated with substantial complications in up to 10-30% of cases [3]: persistent wound drainage, hematoma requiring revision, prolonged pain, persistent loss of sensibility and unfavorable scar formation with cosmetic deformity [3]. The epidemiology of deep surgical site infections (SSI) at the iliac site equally remains largely unknown. This procedure represents another clean intervention, but usually without preceding perioperative antibiotic prophylaxis. Hence, the ideal timing regarding the antibiotic effects administered might be suboptimal for the donor site. In the sparse literature on this topic, the SSI incidence at the iliac site oscillates between 0% [3] and 7% [4], with less than 100 surgeries.

Patients and Methods
We edified a composite database including all first-time (index) orthopedic surgeries at the Balgrist University Hospital between April 2014 and June 2019. We closed our database in May 2020; one year after the inclusion of the last surgery and excluded patients with local orthopedic infections prior to index surgery, pediatric cases, revision surgeries, diabetic foot surgeries or superficial SSIs not requiring revision in the operating theatre. The primary outcome was the deep SSI incidence at the iliac (donor) site. The secondary outcome was the SSI risk at the recipient site among those with iliac harvesting. We defined deep SSI according to internationally accepted norms [5]. In a second step, we compared clinical variables of surgeries with bone grafts to a general cohort of 19,621 primary orthopedic surgeries without iliac crest harvesting, using the Pearson-χ 2 or the Wilcoxonranksum-test, as appropriate (Table 1). Finally, we adjusted for the large case-mix by incorporation the iliac crest harvesting into a multivariate logistic regression model with the secondary outcome "SSI at the recipient site" ( Table 2) and reviewed the available scientific literature (Table 3).

Results
We revealed 20,088 episodes of primary orthopedic surgery, of which 467 with iliac crest sampling (467/20,088; 2%). Only two iliac sites (2/467; 0.4%) become infected. With such few numbers, we were unable to perform group comparisons or multivariate analyses for iliac SSIs. The reason for our low proportion of harvested bone is the limitation to primary surgeries, as we had excluded revision surgeries that are more prone to bone grafting.
Regarding the secondary outcome, surgeries with iliac crest sampling yielded significantly more deep SSIs (9/467; 1.9%) at the primary (recipient) site than those without (149/19,621; 0.8%). Likewise, patients with iliac crest harvesting also revealed more classic risk factors for SSI, when compared to the general orthopedic population (Table 1). In multivariate logistic regression analysis, we confirmed iliac harvesting as an independent risk for deep SSI (odds ratio 2.1; 95% confidence interval 1.1-4.2) ( Table  2). Unfortunately, the wound dressing modalities on the wards and the timing of perioperative antibiotic prophylaxis were too aleatory to be included into this multivariate model. Indeed, the iliac harvesting occurred at different timepoints during operations (at begin, middle or end; with only one prophylaxis administered at the start). Some operations started with iliac crest harvestings, others finished with it. Likewise, the duration of the harvesting equally varied; with an estimated additional time between 25 and 50 minutes. Similarly, there were no exact incision techniques, leading to a considerable variation of the wound length for harvesting. Available literature is sparse regarding deep SSIs at the donor sites (Table 3). Using the MeSH terms "infection" or "surgical site infection", together with "iliac crest" or "graft" or "harvest", we retrieved 181 hits in PubMed at the end of July 2020, by excluding single case reports. The majority of the articles concerned the use of bone grafts for non-unions. Only eight publications reported the SSI risk at the graft donor site, while none indicated the SSI incidence at the recipient sites (Table 3) [6][7][8][9][10]. In this literature, the deep SSI incidence at the iliac crest oscillates between 0-7%, while no article investigates the associated variables due to the paucity of cases.   1, 1,1-1.1  1.1, 1.1-1.1 ASA-Score 2 compared to Score 0-1 1.9, 1.2-3

Discussion
Our single-center pilot evaluation with 20,088 primary orthopedic surgeries found a very low SSI risk (0.4%) in the iliac harvest (donor) site. In contrast, first-time surgeries with supplementary bone grafting yielded higher SSI risks (1.9%) at the recipient sites, when compared to the general orthopedic population (0.8%). The reasons might be multiple. First, we identified the bone grafting as a separate association for SSI in our multivariate analysis. Additionally, patients with bone grafting revealed more classic risk factors for SSI, such as a long operation time. Furthermore, we speculate that the order of the two interventions needs to be fixed. Probably, the perioperative antibiotic prophylaxis (with correct timing) should target the primary site, leaving the harvesting at the end. Alternatively, the repetition of the antibiotic prophylaxis during harvesting might be considered, in as much as the bone harvesting may last up to 50 minutes. These are future investigations.

Conclusion
In our pilot evaluation with 20,088 primary orthopedic surgeries, the SSI risk of the iliac harvest site was low. In contrast, surgeries with supplementary iliac crest harvesting revealed a higher SSI risk than the general orthopedic population, potentially due to a mix of local independent risks of grafting together with a prolonged surgery time.