Low protein intake is highly prevalent among orthopaedic elderly patients. We studied the effects of an isocaloric low protein diet on the resistance to pull-out of titanium rods implanted into rats proximal tibia. Isocaloric low protein intake impairs titanium implant osseointegration, with a decreased strength needed to completely loose the implant and altered bone microarchitecture in its vicinity.
Introduction: Low protein intake is highly prevalent among elderly patients in orthopaedic wards and could retard fracture healing. It was previously shown that reduced protein intake decreases bone strength. Whether dietary protein intake could influence titanium implant osseointegration is unknown. We studied the effects of an isocaloric low protein diet on the resistance to pull-out of titanium rods implanted into rats proximal tibia.
Materials and Methods: Forty-eight 11-month-old female rats were fed isocaloric diets containing 2.5% (low protein) or 15% (normal protein) casein from 2 weeks before the implantation of a 1-mm-diameter cylindrical titanium rod in the proximal metaphysis of each tibia. Four, 6, and 8 weeks after implantation, the tibias were removed for μCT histomorphometry to quantify bone-to-implant contact and bone trabecular microarchitecture around the implant. Resistance to implant pull-out was tested by recording the maximal force necessary to completely loosen the implant.
Results: Pull-out strength was significantly lower in rats fed an isocaloric low protein diet by 6 and 8 weeks after implantation (−43%, p < 0.001 and −42%, p < 0.001, respectively) compared with rats fed a normal protein diet. Bone-to-implant contact was significantly lower in the low protein group 8 weeks after implantation (p < 0.05). Bone-to-implant contact and pull-out strength were correlated (r2 = 0.57, p < 0.0001). BV/TV around the implant was 19.9 ± 2.2% (SE) versus 31.8 ± 3.3% (p < 0.05) at 6 weeks and 20.1 ± 1.9% versus 29.8 ± 3.2% (p < 0.05) at 8 weeks after implantation in the low protein and normal protein intake groups, respectively. Trabecular thickness was 96.2 ± 3.7 versus 113.0 ± 3.6 μm (p < 0.01) at 6 weeks and 101.4 ± 2.7 versus 116.2 ± 3.3 μm (p < 0.01) at 8 weeks in the corresponding groups. In a structure model index analysis, there was a significant shift to a more rod-like pattern in the low protein diet groups. All these changes were associated with lower plasma IGF-I levels.
Conclusions: Isocaloric low protein intake impairs titanium implant osseointegration, with decreased strength needed to completely loosen the implant and altered bone microarchitecture in the vicinity of the implant.
MALNUTRITION IS HIGHLY prevalent among elderly patients in orthopedic wards(1–4) and especially among those with hip fracture. Malnutrition could contribute to increased risk of falling and to a low bone mass. Furthermore, undernourished patients display higher rates of medical complications, of postoperative wound healing retardation, and of death, together with a longer duration of hospital stay in the postfracture period. (5–10)
Among the various nutrients, protein seems to be of major importance. Indeed, low protein intake has been associated with decreased bone strength in rats, (11, 12) decreased bone mass in humans, (13–15) and increased osteoporotic fracture risk. (14, 16, 17) Reduced protein intake has been shown to decrease intrinsic bone tissue quality in rats. (18) The clinical outcome of elderly with a recent fracture of the proximal femur was more favorable when they received an oral nutritional protein supplement. (9, 19, 20) Besides its influence on clinical outcome, a low protein intake could also affect fracture healing. Indeed, acute protein deprivation in the postfracture period has been shown to impair fracture healing in rats, both morphologically and mechanically. (21–23) Chronic isocaloric protein deprivation was also associated with detrimental effects on long bone fracture healing in rats. (24) As a potential mechanism, there is reduced IGF-I levels, whose hepatic production and plasma concentrations are under the influence of protein intakes. (25–29)
A substantial proportion of elderly admitted into orthopedic wards for osteosynthesis or arthroplasty are undernourished. (3, 9) Whether integration of orthopedic material could be altered under these conditions, particularly, the specific role of reduced protein intake, is not known.
Taking into account all these elements, we hypothesized that reduced protein intake could negatively influence implant osseointegration by altering bone turnover. (11) To test this hypothesis, we studied the osseointegration and the pull-out strength of titanium cylindrical implants in the proximal metaphysis of adult female rats fed either isocaloric low or normal protein diets, providing identical amounts of minerals.
MATERIALS AND METHODS
Animals and diet
All experimental designs and procedures received approval of the Animal Ethics Committee of the Geneva University Faculty of Medicine. Forty-eight 11-month-old Sprague-Dawley female rats (Charles River Laboratories, L'Arbresle, France) were housed individually at 25°C with a 12:12-h light-dark cycle. They were strictly pair-fed a laboratory diet provided by Provimi Kliba AG (Kaiseraugst, Switzerland) containing 15% or 2.5% casein, 0.8% phosphorus, 1% calcium, 70–80% carbohydrates, and 5% fat throughout the whole experimental period. Isocaloric intakes were ensured by the addition of corn carbohydrate to the low protein diet, thus providing the same energy intake. Demineralized water was available ad libitum. Because the minimal protein intake in adult rats necessary to maintain normal bone homeostasis is 5%, (11, 12) a restriction to 2.5% corresponds to a reduction of protein intake of 50% of the animal necessary amount, a reduction that is often observed in patients with hip fracture. (9, 14)
After 1 week of adaptation to a 15% casein-containing diet, rats were divided into six groups: three groups received a diet containing 15% casein and three groups received a diet containing 2.5% casein until the end of the study. Two weeks after the beginning of protein restriction, a titanium implant was inserted in each tibia of the animals under general anesthesia. The implants were made of commercially pure titanium (Institut Straumann AG, Waldenburg, Switzerland) and were sandblasted and acid-etched on the nonthreaded part. They were 1.0 mm in diameter and 4.1 mm long. A 1.25-mm-long threaded part (type M1) remained outside of the bone, thus permitting prehension for the pull-out test. We selected this type of implant because titanium is well tolerated by bone tissue. The rough surface produced by sandblasting and acid-etching ensures a good osseointegration. The implants were cleaned in phosphate-free Deconex (Borer Chemie AG, Zuchwil, Switzerland), rinsed in pure water in an ultrasonic bath, and sterilized with ethylene oxide. At 4, 6, and 8 weeks after implantation, blood from rats in one 2.5% and one 15% casein group was obtained by aortic puncture under general anesthesia, and animals were killed by an overdose of ketamine hydrochloride. The tibias were removed for μCT histomorphometry and mechanical testing.
Animals were anesthetized by abdominal injection of ketamine (100 mg/kg) and xylazine (10 mg/kg). The skin of the bilateral tibial region was shaved and cleaned with 70% ethanol. Under aseptic conditions, an anterior approach with a 10-mm incision was made to gain access to the proximal medial aspect of the tibia metaphysis. The periosteum was reflected medially. A hole with a diameter of 1.0 mm was drilled through one of the cortices with a hand-held drill, and the implant was inserted (Fig. 1). Rotatory speed did not exceed 2000 rpm, and drilling was accompanied by profuse saline irrigation to avoid bone thermal necrosis. After insertion of the implant, the skin was sutured using a 3–0 resorbable polyglactic suture (Vycril; Ethicon, Spreintenbach, Switzerland). Postoperative analgesia was ensured by subcutaneous injection of buprenorphine (0.06 mg/kg) twice a day for 3 days.
The tibias were carefully excised immediately after death and frozen at −20°C in plastic bags. The night before μCT analysis, the bones were thawed slowly at 7° and maintained at room temperature. During the whole analysis, they were immersed in saline solution. Parameters of mass, architecture, and percentage of implant surface in direct contact with trabecular bone (percentage of attachment, bone-to-implant contact) were studied using μCT histomorphometry with a high-resolution μCT system (μCT 40; Scanco Medical AG, Bassersdorf, Switzerland). The voxel size was 20 μm in all spatial directions. For penetrating radio-opaque titanium and improving signal-to-noise ratio, the system was set to 70 kEV and 350-ms integration time. The titanium and bone were segmented individually using two distinct threshold values (Fig. 2) because of their different densities. Trabecular bone was analyzed on a circular band of 0.5 mm around the implant. Relative bone volume (BV/TV), trabecular number (Tb.N), trabecular thickness (Tb.Th), and trabecular separation (Tb.Sp) were calculated by measuring directly the 3-D distances in the trabecular network. Connectivity density based on Euler number (Conn.D) and the structure model index (SMI) were calculated. Bone-to-implant contact was calculated from the μCT voxel data. When a bone voxel was in the vicinity (60-μm boundary) of a titanium voxel, this titanium voxel was classified as “attached surface.” The number of attached surface voxels divided by the total number of titanium surface voxels yielded the percentage bone-to-implant contact.
After μCT analysis, tibias were subjected to a pull-out test. The tibias were fixed with a metal device designed for this purpose, and another metal piece was screwed on the threaded part of the implant. The pull-out strength was determined as the peak force applied to fully loosen the implant from the bone as measured with a servo-controlled electromechanical system (Instron 1114; Instron Corp., High Wycombe, UK) with the actuator displaced at 2 mm/min. Reproducibility was 9.4%, as evaluated as the CV of pair sample measurements (left/right). A preliminary study showed that the freezing procedure did not alter pull-out strength values. Regression between values obtained before or after freezing procedure was characterized by an r2 coefficient of 0.96.
IGF-I was measured in plasma by immunoenzymometric assay (IEMA) with a kit from Nichols Institute (San Juan Capistrano, CA, USA) following manufacturer instructions.
All results are expressed as means ± SE. Significance of difference was evaluated with a two-sided unpaired t-test. Regression analysis was performed using the Statview software (Statview SE + Graphics 1; Abacus Concept, Berkeley, CA, USA).
Effect of isocaloric protein restriction on pull-out strength
Mean pull-out strength values were significantly lower in rats fed an isocaloric low protein diet at 6 and 8 weeks after implantation (−43% and −42%, respectively) compared with the normal protein intake groups (Fig. 3). This was because of a progressive increase in pull-out strength under normal protein intake, which did not occur in rats fed the isocaloric low protein diet. By 4 weeks, pull-out strength values were similar in both groups.
Effect of isocaloric protein restriction on μCT histomorphometry of the trabecular bone around the implant
To study the mechanisms of reduced implant pull-out strength under isocaloric protein restriction, we analyzed trabecular bone microarchitecture around the implant and bone-to-implant contact. By 4 weeks after implantation, BV/TV values were not statistically different between groups (Table 1). A significant change of the relative bone volume was detected at 6 and 8 weeks with a decrease in BV/TV. For bone-to-implant contact, a progressive difference developed between the groups on the two diets, becoming significant at 8 weeks (−22% in the low protein intake group; Fig. 4; Table 1). Bone-to-implant contact and pull-out strength were related with an r2 coefficient of 0.57 (p < 0.001; Fig. 5). In the normal and the low protein intakes groups, the regression analysis between pull-out strength and bone-to-implant contact was y = 73.3x + 5.2, r2 = 0.60 (p < 0.001) and y = 52.3x + 1.3, r2 = 0.49 (p < 0.001), respectively. For a given value of bone-to-implant contact, pull-out strength was lower in the presence of a low protein intake, suggesting the involvement of other factors, yet to be elucidated.
Table Table 1.. Effects of Isocaloric Dietary Protein Restriction on Bone-to-Implant Contact and on Bone Architecture Around the Titanium Implant
Other parameters of the trabecular bone architecture in the vicinity of the implant were affected by protein restriction (Table 1). A significant decrease of trabecular thickness was noted at 6 and 8 weeks after implantation. This was associated with an alteration of the trabecular structure, with a shift to a more rod-like bone pattern, as shown by changes in the SMI. The trabecular number was also significantly decreased by 6 weeks in the animals fed a low protein diet. The apparent decrease in connectivity density in the low protein intake groups at 6 and 8 weeks after implantation did not reach a level of statistical significance.
Effect of isocaloric protein restriction on IGF-I and on body weight
As previously reported, (11, 12, 29) isocaloric protein restriction significantly lowered plasma IGF-I levels (Table 2). Low protein intake was also associated with decreased body weight, despite a strict pair-feeding with isocaloric diets.
Table Table 2.. Effects of Isocaloric Protein Restriction on IGF-I Plasma Levels
This study indicates that an isocaloric low protein diet impairs the osseointegration of titanium implants in adult female rats. The mechanical fixation of the implant is altered, as illustrated by the lower pull-out strength observed in the low protein groups by 6 and 8 weeks after implantation. The microarchitecture of the trabecular bone immediately adjacent to the implant as well as the bone-to-implant contact are decreased in protein restricted rats.
In this experimental model, isocaloric dietary protein restriction corresponds to ∼50% of the minimal protein intake necessary to maintain normal bone strength, microarchitecture, and turnover in adult rats. (11) Such a degree of protein undernutrition is very frequent in elderly in orthopedic wards.
Various determinants could account for the reduced values observed in the implant pull-out test. Thus, pull-out strength represents the overall result of various factors influencing the integration of the implant in bone. In our model, the mechanical fixation of titanium implant, together with bone mass, microarchitecture, and bone-to-implant contact show a progressive increase with time after surgery under normal protein intake. This gradual increase does not occur under an isocaloric low protein diet, explaining the lower strength required to remove the implant. The difference in pull-out strength between the two diets is already detectable by 6 weeks after surgery. To our knowledge, this is the first time that protein restriction has been shown to be associated with an impaired osseointegration of metal implants. This observation could be of significant clinical importance in elderly patients undergoing joint replacement. However, the relation between protein intake and implant loosening has not been extensively studied in humans yet.
The alterations of the bone-to-implant contact and of the trabecular microarchitecture around the implant could contribute to the negative effects of isocaloric dietary protein restriction on implant osseointegration. Bone-to-implant contact seems to be a strong predictor of pull-out strength, as indicated by an r2 of 0.57 in a linear regression analysis. Bone-to-implant contact is significantly lower in animals fed an isocaloric low protein diet. This is consistent with the decrease of the relative trabecular bone volume around the implant. Sandblasting and acid-etching of the implants produce an irregular surface. In this study, bone ingrowth into the rough titanium surface, which could also play a role in implant osseointegration, could not be assessed, because μCT resolution was not sufficient. Because dietary protein restriction has been shown to decrease BV/TV in trabecular bone of the rat tibia proximal metaphysis, (11, 12, 29) the relative effect of reduced protein intake on trabecular bone and on implant osseointegration cannot be distinguished.
IGF-I could represent a potential link between changes in dietary proteins and alterations of bone microarchitecture and bone strength. (30) In our study, as well as in previous experiments, (11, 12, 29) an isocaloric low protein diet is associated with an early decrease in plasma IGF-I levels, together with increased bone resorption. Thus, both mechanisms (i.e., increased resorption and decreased formation) are likely to impair implant osseointegration. In protein restriction-induced bone resorption, (11) bone resorption-stimulating cytokines, such as TNF-α, (11, 31–33) or sex hormone deficiency could be implicated. (11) With the age-related decline in renal function, blood pH slightly decreases, possibly in relation with excess dietary protein. (34, 35) In this study, we evaluated the influence of dietary protein deprivation, but did not address the role of increased protein intake, and the consecutive possible influence of acid production on implant osseointegration.
In conclusion, in an experimental model of isocaloric dietary protein restriction, we have shown for the first time that titanium implant osseointegration is impaired under protein undernutrition. Indeed, the strength needed to completely loose the implant and the bone-to-implant contact are reduced in protein-deficient animals, as is altered bone microarchitecture, with a diminished bone relative volume, thinner trabeculae, and a trabecular network shifted to a more rod-like bone. This negative influence of reduced protein intake on implant osseointegration could be of major clinical significance, because of the high prevalence of malnutrition among elderly admitted into orthopedic wards for osteosynthesis or joint replacement.
The authors thank S Clément for expert animal care, diet preparations, and blood collection; I Badoud for μCT measurements; Dr A Laib for technical help in using μCT; Dr M Cattani and V Châtelain for biomechanical testing; and M Perez and M Battistella for secretarial assistance. This research project was supported by Swiss National Research Foundation Grant 3200BO-100714.