One safe [1] and promising substance group for this aim are the statins, where were primarily developed to inhibit cholesterol biosynthesis [2] and reduce its plasma levels [3]. Besides other pleiotropic effects [4,5], Simvastatin furthermore stimulates bone formation in a human osteoblast cell line (MG-63) via increased expression of bone morphogenetic protein (BMP-2) [6]. BMP-2 is one of the most potent growth factors targeting bone formation in vivo [7], which stimulates osteoblast differentiation and -proliferation [8-10].

With systemic [11-14] or local delivery of statins to the bone from biodegradable carriers or percutaneous application forms, bone formation and fracture healing were improved in several experimental investigations [15-22].

Statins were also shown to improve implant integration following arthroplastic surgery. Although there have been multiple studies using systemic administration [11,23-25], only a few studies have investigated local application of statin-containing gels or microspheres and their effect on implant-bone integration [26-28]. In summary, combining the bone anabolic effect of Simvastatin with other local drug application techniques is a promising approach. Local application seems to be a reasonable strategy, as statins target liver metabolism and are poorly distributed to bone due to first pass elimination [29]. Furthermore, local application helps prevent systemic side effects of drugs.

Hence, it was the aim of the present study to investigate the effect of Simvastatin, locally applied in a biodegradable Poly(D,L-lactide) coating from intramedullary implants [30], on bone integration using biomechanical and histomorphometrical analyses.

After 56 days at the x-ray control, no radiographic implant dislocation was observed (Figure ​(Figure1).1). However, n=2/20 animals of the SIM high-group showed osteolysis around the implants (Figure ​(Figure2)2) and n=1/20 animal from both the PDLLA and SIM high group had broadened medullary canals. However, microbiological analysis of these affected femurs showed no evidence of bacterial infections.

Calculated strength of fixation was highest for uncoated implants when compared to all other control and experimental groups. Fixation of SIM high coated implants was significantly weaker than with uncoated- (p = 0.002) or PDLLA-coated implants (p = 0.005, Figure ​Figure33).

Histomorphometrical analysis with regard to total bone/implant contact showed only marginal differences between the control groups (uncoated, PDLLA-coated, Figure ​Figure4).4). However, lower results were detected for both experimental groups, reaching level of significance for SIM high when compared to both control groups (minus 46.0% and 32.8%, respectively, p < 0.001) and for SIM low vs. uncoated controls (p = 0.006).

Neither direct nor indirect bone/implant contact differed significantly between uncoated and PDLLA-coated implants. In terms of direct contact, SIM high results were significantly decreased when compared to both control groups (vs. uncoated: p = 0.001; vs. PDLLA: p = 0.006; Figure ​Figure44).

Indirect contact was highest in the uncoated and PDLLA-coated groups. Both SIM coated groups had lower values than the controls, differing significantly between the PDLLA and SIM high group (minus 51.0%; p = 0.004, Figure ​Figure44).

Though the two cases of radiographic osteolysis around the implant occured in the histomorphometric group, biomechanical stability was significantly weaker for SIM high when compared to uncoated and carrier-only coated implants. Histomorphometry confirmed a significant reduction in total bone/implant contact, (in-)direct contact and new bone formation for the experimental groups when compared to the control groups.

In summary, the hypothesis of the present paper with regard to improved implant integration must be rejected for both experimental groups.

The question arises why bone integrational processes deteriorated under Simvastatin exposure when compared to controls. One potential obvious explanation, intramedullary infections of the femurs, was ruled out since no bacteria were identified after microbiological analysis over 14 days.

Another possible reason for the adverse effects might be the PDLLA-coating itself. However, this polymer was previously shown to be biocompatible, mechanically stable [30] and a reproducibly degradable carrier to locally deliver agents to the bone, without evidence of osteolysis [20,31-34]. Furthermore, the PDLLA-group showed (significantly) superior biomechanical, histomorphometric and radiographic properties when compared to the experimental groups. Thus, PDLLA is likely not the reason for the deteriorated bone integration, even though its biomechanical stability was slightly inferior when compared to uncoated controls.

Another possible reason for the lack of osseointegration may have been the drug dose used, since dose-dependent effects of statins on bone metabolism were suggested due to differing sensitivity of osteoblasts and osteoclasts. Bone resorption and formation were elevated with high-dose Simvastatin while low-dose SIM decreased formation and increased bone resorption [35]. The present results seem contrary since SIM high rather than SIM low had bone catabolic effects and neither exerted bone anabolic effects. The dose-dependent, bone-anabolic effect of a comparable SIM high-dose was previously shown to have a similar effect to that of BMP-2 in a rat fracture model [20]. Additionally, one other failure option is the type of the incorporated drug. However, the identical substance and coating technique were successfully investigated previously [20]. Several other experimental studies confirmed beneficial effects of statins on fracture healing using different local application approaches [18,19,21,22]. In addition, statins improved defect regeneration when locally applied in cranial/mandibular bone defect models without metal implants [15-17,36]. These studies used absorbable collagen and gelatin sponges or injections for drug delivery which are prone to dissolve at the site of application. A femoral defect model with local small molecule drug delivery (but no metal implant) revealed bisphosphonates significantly improved bone formation while lovastatin did not [37]. Piskin et al. demonstrated Simvastatin-loaded electrospun nanofibers enhanced bone mineralization (histological and micro-CT analysis) [38]. Even though the same drug was used, the different dosage and the use of the polymer caprolactone represent different approaches than in the present setting, hence impede comparability.

In contrast to fracture- and bone defect healing, implant integration was investigated presently. In this regard, different studies investigated the effect of local statin application. A similar rodent model was utilized by Moriyama et al. who observed improved tibial implant integration dose-dependently after 7–14 days of local fluvastatin-release from a PGA-coating [26]. In contrast to the present findings, their higher-dosed group (2.5 mg/ml) showed the best results in terms of bone formation and push-out strength. The short observational period of 7 or 14 days may be one reason for different results when compared to the present 56 day period. Further distinctions to the present study are the chosen type of carrier and the incorporated drug Fluvastatin (vs. Simvastatin), although both are lipophilic, penetrate cell membranes and enhance osteogenesis. These authors later investigated injectable PGA-gel around tibia implants in rodents and found similar results [28]. They observed a significant decrease in implant integration at one week comparable to the present results after 56 days. However, stability recovered and significantly increased after 14 and 28 days, respectively. Nevertheless, injected gel or mobile nanoparticles may dissolve from the intramedullary destination, while solid bioactive implant coatings may reduce this effect.

Other experimental studies reported on improved orthopedic implant integration in animals even under systemical exposure to statins, administered orally [23], percutaneously [24,27] or intraperitoneally [11,25]. Effects were observed with up to 10-50 mg/kg bodyweight, far exceeding the statin dose rates applied in humans, while the equivalent dose used in humans was ineffective [25]. Hence, systemical application does not seem to be useful for improved implant integration with normal human drug doses. Even though one observational clinical study reported on reduced risk of hip implant revision, deep infection and aseptic loosening among statin users under normal dose rates [39], no direct conclusion on the drug was feasible. It was suggested that statin-users in general might show a more health oriented behaviour (i.e. medication or rehabilitation compliance). One prospectively randomized clinical study found no effect on bone healing between low Simvastatin-intake (20 mg/d orally) and placebo [40].

Since less than 5% of an oral statin dose reaches the circulation due to hepatic first pass elimination [29], systemical application requires rather high drug doses. Targeted, local application of drugs from bioactive carrier polymers seems more efficient and may help to improve drug availability within the bone while lowering necessary drug doses, hence preventing systemic side effects.

As another potential limitation of this study, the observation period of eight weeks may be inappropriate to observe differences in implant integration since other authors observed effects of statins with this regard after 7–30 days [11,25-28]. During fracture healing, beneficial effects were reported after 5–14 days of local statin exposure to the fracture site [18,21,22], suggesting that statins cause a delayed onset of endogenous BMP-2 production [20]. Mundy et al. reported on a quick BMP-2 response to statins within 3–5 days in vivo and in vitro [6] while other authors found improved implant integration after 42–84 days following high systemical doses of Simvastatin [23,24].

After 8 weeks, remodeling “back to normal” may occur and initial improvements in implant integration may vanish over time. However, mid- and long term data are important with regard to prosthetic implant integration in humans.

Nevertheless, the timepoint does not explain the osteolysis in two SIM high coated animals.