In 1999, Mundy et al described a set of experiments, which indicated that a group of common cholesterol lowering drugs, the statins, have anabolic effects on bone[1]. Other experiments supporting this finding have followed [2–11]. However, other studies have not shown any such effect, most notably the study reported by Maritz et al, which in essence repeated the study by Mundy et al and found diametrically different results[12]. Also the experiments reported by von Stechow et al found no positive effect on undisturbed bone by simvastatin in mice[13]. Thus, there still remains some controversy concerning the effect of statins on bone formation. In 2002, the authors reported on a dramatic improvement of fracture repair in mice by simvastatin mixed in the feed[14].

Although effective, the dose used in that study (about 100 times the recommended maximum clinical dose, as set out in the official label text) seemed impractical if statins were to have any use in bone formation in a clinical situation.

Most of the orally administered simvastatin in our previous study would have been sequestered in the liver, as only a few per cent of orally administered simvastatin reaches the general circulation in an unbound form and are accessible to extra-hepatic cells (Official label text, [15–18]). Consequently, in order to achieve a dose which would be clinically useful, we would need to by-pass this first pass clearance of the liver. We therefore conducted a number of experiments on fracture repair in which we administered the simvastatin as one daily subcutaneous injection in doses ranging from 1 to 100 mg/kg body weight. We were unable to find any significant effect of the statins with this setup (data not shown).

With one daily injection, the concentration of simvastatin would reach a peak relatively quickly, and then leave the organism. The elimination half-life of simvastatin is about 2 hours in humans, and probably not longer in mice[19, 20].

Consequently, two questions arose. Firstly, is a continuous plasma concentration necessary for an effect on fracture repair? If so, subcutaneous injections would not work, whereas a continuous subcutaneous release of simvastatin would yield positive results similar to the ones achieved when mixing it in the feed. Secondly, since the effect of simvastatin on bone metabolism seems to be a local effect on bone cells; would local delivery to the fracture work?

In order to answer these questions, we performed three experiments. Firstly, we conducted an expanded experiment with subcutaneous injections. Secondly, subcutaneously implanted osmotic mini-pumps were used to deliver a continuous systemic dose of simvastatin. Thirdly, silicone tubes were led subcutaneously from implanted osmotic mini-pumps to the fracture area, delivering a local continuous dose.

70 mature male Balb-C mice were used. The study had been approved by the regional ethics board and institutional guidelines for the care and treatment of laboratory animals were adhered to. The mice were kept 1 per cage with free access to mouse-chow and water.

Simvastatin powder (kindly supplied by MSD) was dissolved in PEG 400 (Sigma-Aldritch Chemie Gmbh, Steinheim, Germany) and passed through a sterile filter (Millex™, pore-size 0.22 μm, Millipore Corporation) before injection or filling the mini-pumps.

Surgical procedure

The mice were anesthetized with isoflourane gas. Each mouse received a preoperative subcutaneous injection of 1.5 mg oxytretracycline and 0.003 mg of buprenorphine. Implants and surgical equipment were sterilized in an autoclave. Sterile gowns, gloves, surgical masks and theatre caps were used. The mouse leg on the operated side was shaved and the entire mouse was put into a sterile surgical glove. Subsequently, a hole was cut out of the glove and the leg pulled out through the hole using tweezers and the leg washed with chloro-hexidine alcohol.

A lateral incision was made along the distal femur. The patella was dislocated medially with the blunt side of the scalpel, so that the femoral condyles were exposed. Using an intercondylar approach, a hole was drilled through the medullary canal of the femur using a cannula (diameter 0.4 mm). A wider cannula (0.6 mm) was then inserted into the canal made. The sharp end of the cannula was blunted in order not to penetrate through to the hip. The cannula was cut off, so that the remaining bit was inside the bone and no end extended beyond the bone. A specially made pair of scissors with semi-lunar cutting edges were then slid along the bone to mid-diaphysis and the femur cut to produce a fracture (Figures 1, 2 &3). The patella was then repositioned over the knee-joint and the muscles and skin sutured separately.

The mice recovered well after surgery and did not display overt signs of discomfort. Weight-bearing on the operated leg began within the first post-operative week.

In the subcutaneous injections experiment, 2 mice were excluded because of faulty placement of the intramedullary pins. In the systemic continuous release experiment, 4 mice were excluded. One due to failed surgery, one due to death shortly after surgery and 2 due to technical errors while harvesting the femurs. In the local treatment experiment, one femur was excluded due to faulty intramedullary pin placement. Exclusions were done while blinded.

In contrast to our previous results, we could not demonstrate an effect on cross-sectional area. This, combined with a general increase in biomechanical parameters, seems to point to the possibility of improved material properties, indicating an increased callus maturity with more bone. Such an increase would be in line with several in vitro studies, which have reported an increase in osteoblast differentiation and calcification by statins[4–6, 21–25]. We could not demonstrate improved maturity by histology in our previous study with orally administered simvastatin. The difference between that study and the present one could be due to different local doses (as has been suggested as a possibility by Maritz et al[12]) or a failure to detect actual effects.

Enhanced repair by local simvastatin has been demonstrated in a critical size defect model in rabbits[9, 11], and there are plentiful in vitro data which support an anabolic effect on bone by statins [5, 23, 26–30]. Furthermore, it seems that, at least for this mice model, a sustained continuous release is necessary for an effect. From our experiments, it seems clear that the in vivo effects of simvastatin on bone is a local phenomenon not related to the established cholesterol-lowering effect, and that a local delivery system can be as effective as systemic treatment in promoting fracture healing.

From personal communication with Mundy's group, we had learned that above a certain dose, the statins did not have any effect on bone in their experiments. Further, indirect evidence for a possible biphasic effect on bone comes from research into possible angiogenesis effects of statins[31], and there seems to be quite a flat dose-response curve[32]. Therefore, to get the desired effect, one should perhaps maintain a certain local concentration of the drug over a longer period of time. This is what we most probably achieved in our previous study by mixing the drug in the feed, since mice eat more or less continuously throughout the day. In contrast, and similar to subcutaneous injections, gavage should lead to a high but relatively brief concentration peak of the drug. Such a regimen has been reported to have little effect in mice [13].

In conclusion, we found an effect of simvastatin treatment directly applied to the fracture area. Our current results are in parity with what we saw in the first study with orally administered simvastatin[14]. In that study, the required doses were excessive. By using local treatment we were able to reduce the total dose by three orders of magnitude. The resultant negligible systemic levels bring us one step closer to the feasibility of using statins to improve fracture healing in a clinical setting. A natural next step would be to deliver the simvastatin coated to the implants, as we have previously successfully done with bisphosphonates[33].