Protein restriction has been shown to reduce plasma levels of insulin-like growth factor-I (IGF-I) by inducing resistance to the action of growth hormone in the liver and increasing the metabolic clearance rate of the growth factor [19-23]. Furthermore, evidence shows that protein depletion may blunt the effect of IGF-I on target organs [20]. Thus, low protein intake in elderly persons may be detrimental to skeletal integrity, muscle strength, and immune response [11, 24-27] because of decreased production and action of IGF-I, which favorably influences these systems [28-33].

Previous studies [16, 17] have shown that a 5-week course of protein supplements can reduce the medical complication rate and duration of hospital stay in patients with recent hip fracture. These findings were independent of energy, calcium, and vitamin D intake [17]. Whether these observations were related to the restoration of decreased IGF-I levels and whether bone may benefit from long-term protein supplementation are not known.

We investigated whether protein supplements in vitamin D-replete patients with a recent hip fracture who were receiving calcium supplements could increase circulating IGF-I levels and favorably influence bone mineral density.

Methods

Top Methods Results Discussion Author & Article Info References

Patients

Patients were recruited in the orthopedic ward of Geneva University Hospital, the referral center that receives 94.6% of all patients with hip fracture from an area with a population of approximately 400 000 persons [5]. The protocol was approved by the ethical committee of the Geneva University Hospital Department of Surgery. Inclusion criteria were age greater than 60 years; recent hip fracture (within 2 weeks) attributable to osteoporosis (that is, a fracture after a minor trauma, such as a fall from standing height); and the ability to give written, informed consent. Exclusion criteria were pathologic fracture; fracture caused by severe trauma; history of contralateral hip fracture; severe mental impairment; active metabolic bone disease; renal failure (plasma creatinine concentration 200 µmol/L); acute illness that could interfere with the study protocol; severe malnutrition (serum albumin level <15 g/L); consumption of drugs known to alter bone metabolism, such as calcitonin, fluoride, sex hormones, corticosteroids, or bisphosphonates; and life expectancy of less than 1 year.

Using a random number table, we assigned patients to receive an oral protein supplement composed of 90% milk proteins or a placebo made isocaloric by the addition of maltodextrins. Patients took the assigned intervention 5 days a week for 6 months in addition to their regular diet. All patients received one oral dose of vitamin D₃, 200 000 IU (vitamine D₃B.O.N., Doms-Adrian, Courbevoie, France), to correct any possible vitamin D deficiency [34]. The daily protein supplement (Meritene, Sandoz Nutrition Ltd., Berne, Switzerland) provided 1050 kJ (250 kcal) of energy in the form of 20 g of proteins, 3.1 g of lipids, and 35.7 g of carbohydrates (54.4 g in placebo). The other constituents of the 65 g powder supplement were vitamin A (1000 IU), vitamin K₁ (30 µg), vitamin C (20 mg), calcium (550 mg), magnesium (91 mg), phosphorus (429 mg), and sodium (228 mg). The supplement is designed to normalize but not overcompensate for the insufficient dietary intake of protein of elderly persons with a recent fracture of the proximal femur [16]. Compliance was verified by weekly phone calls and by counting the remaining nutritional supplement bags, which were forwarded monthly by mail.

Clinical Data

Medical history, clinical characteristics, anthropometric data, and performance status according to an Activities of Daily Living score [35] were evaluated. Dietary intakes were recorded by using a food-frequency questionnaire [36] about the week before the fracture. Food intakes were analyzed by using a nutrient software system (Fruitdor, Astra-Calve, Paris, France). Dietary intake of calcium, phosphorus, and protein was calculated from intake of dairy products, meat, fish, and vegetables. Because many patients were unable to stand a few days after hip fracture, their height was determined with a fathom measure while they were recumbent. Body weight was determined with a scale that correlated highly with that provided by measurement of whole body bone, fat, and lean mass as assessed by dual x-ray absorptiometry [37]. Mid-arm circumference was measured with a tape measure (coefficient of variation, 2%). Isometric muscle strength of the biceps of the dominant arm was evaluated with a dynamometer (Lido, Lidoactive Isokinetic System, Loredan Biomedical Inc., West Sacramento, California), which measures the peak torque (coefficient of variation <5%). The grip strength of the same arm was measured with a dynamometer (Vigorimeter, Martin Medizin-Technik, Tuttlingen, Germany) [38]. The mean of three measurements was used for calculation.

Biochemical Data

Within 120 hours after surgery, we used standard methods to analyze venous blood samples for protein-corrected plasma levels of calcium, phosphate, creatinine, total proteins, albumin, and prealbumin. We also measured levels of IGF-I (Nichols Institute, San Juan Capistrano, California) after acid-ethanol extraction and cryoprecipitation [39], intact parathyroid hormone (Immulite, Diagnostic Products, Los Angeles, California), calcitriol, calcidiol (Incstar, Stillwater, Minnesota), and osteocalcin (CIS-BIO International, Gif-sur-Yvette, France). Biochemical measurements were repeated 6 and 12 months after the fracture. Serum IGF-I binding proteins were measured by Western ligand blot analysis [40]. Briefly, 3 µL of serum was boiled and separated onto 15% sodium dodecylsulfate polyacrilamide gel electrophoresis under nonreducing conditions. Proteins were electrically blotted onto a nitrocellulose membrane, and IGF-I binding proteins were detected after incubating the membranes with radioiodinated IGF-I at 4°C for 24 hours. Insulin-like growth factor binding proteins were quantified by phosphorimaging (Molecular Dynamics, Sunnyvale, California). The intensity of the bands corresponding to the appropriate molecular weights [40] was expressed as the percentage of total bound radioactive IGF-I. Measurements of calcium, phosphate, and creatinine were obtained from the second fasting morning urine. The ratio of hydroxyproline to creatinine in the same sample was taken as a reflection of bone resorption. Markers of bone resorption (pyridinoline and deoxypyridinoline) in the first morning spot urine were measured by detecting fluorescence emission after acid hydrolysis and separation with isocratic reverse-phase high-performance liquid chromatography (BioRad System, Munich, Germany), and the values were adjusted to the creatinine concentration.

Immunologic Data

Concentrations of IgA, IgG, IgM, and isohemagglutinins were measured by using standard methods. Cell-mediated immunity was assessed by using a skin-test antigens for cellular hypersensitivity system (STACH, Institut Merieux, Lyon, France), which evaluates the size of the skin reaction to various antigens injected intradermally [41].

Bone Mass Assessment

We measured areal bone mineral density at the lumbar spine (anteroposterior and lateral views), the contralateral proximal femur (femoral neck and trochanter), and the contralateral mid-femoral shaft. We also measured whole-body bone mineral content, fat, and lean mass by using dual x-ray absorptiometry (Hologic QDR-2000, Waltham, Massachusetts). The coefficients of variation of these measurements are reported elsewhere [37, 42].

Vertebral Deformity

Two series of lateral radiography of the thoracic and lumbar spine were taken at the start of the study and at 12 months. The films were examined by a single investigator (who was unaware of treatment group assignment) according to a 6-point analysis procedure for each vertebral body from T4 to L4 [43, 44] (intra-observer coefficient of variation, 1.8% to 3.1%). At baseline, fracture was defined by a decrease of 20% in the ratio of anterior or middle height to posterior height. Patients were considered to have a new vertebral deformity if the anterior, middle, or posterior height decreased by more than 20% between the two examinations.

Statistical Analysis

All values are given as the mean ±SD for baseline measurement and the mean ± SE for comparisons of outcome results. All analyses were performed by using the SAS procedure (Cary, North Carolina). One-way analysis of variance was used. A two-tailed P value less than 0.05 was considered significant.

Role of the Funding Sources

This study was supported by grants from Sandoz Nutrition Ltd., Berne, Switzerland, and the Swiss National Research Science Foundation. The skin-test antigens for cellular hypersensitivity system was supplied by Rhone-Poulenc, Thalwil, Switzerland. No organization influenced the design, conduct, or reporting of the study.

Results

Top Methods Results Discussion Author & Article Info References

Of 842 patients evaluated from April 1992 to February 1994, 82 were recruited into the randomized, double-blind, placebo-controlled trial. This low enrollment rate was due to patients' poor medical or cognitive conditions and reluctance to give informed consent.

Baseline Characteristics

Ninety percent of the patients enrolled were women; this reflects the higher incidence of hip fracture in this population (Table 1). At baseline, the anthropometric, functional, and nutritional characteristics of patients who received protein supplements and those who received isocaloric placebo (controls) did not differ.

Bone mineral density of the femoral neck, trochanter, or mid-femoral shaft contralateral to fracture side was measured 9.6 ± 3.8 days after hip fracture (Table 1). For femoral sites, the low bone mineral density values fulfilled the World Health Organization criteria for the diagnosis of osteoporosis [45]. As previously reported, the difference in bone mineral density of the lumbar spine between patients with hip fracture and young, healthy adults was of lesser magnitude than the difference in bone mineral density of the proximal femur [46, 47]. Bone mineral density values and whole body bone mass were similar in patients who received protein supplements and controls. Biochemical markers of bone resorption measured 6.5 ± 2.8 days after hip fracture were increased in patients and controls (Table 2). Levels of prealbumin and IGF-I were –3.0± 1.4 and –1.1± 0.8, respectively, below mean values for healthy elderly persons. These values are compatible with some degree of protein malnutrition. All other values were within the normal range and were similar in patients who received protein supplements and controls. Cell-mediated immunity as assessed by a skin-test antigens for cellular hypersensitivity test was similar in both groups (sum of the diameters of the skin reactions, 4.5 ± 7.9 cm in patients who received protein supplements and 4.9 ± 8.2 cm in controls).

Outcomes

During the 6-month intervention period, 9 patients who received protein supplements and 7 controls dropped out because of nausea (4 patients and 4 controls), diarrhea (2 patients and 1 control), and refusal to pursue the trial (3 patients and 2 controls). Among patients who received protein supplements, 1 died of pulmonary embolism and 1 died of sepsis; 1 control died of myocardial infarction. Thus, of the 41 persons enrolled in each group, 30 patients who received protein supplements (73%) and 33 controls (80%) completed the intervention. During post-treatment follow-up, 4 patients dropped out, 2 patients died, and 2 controls died. During the 1-year study, dietary intake (excluding supplements) as assessed by the mean of three frequency questionnaires at 6-month intervals was 629 ± 58 mg/d in patients who received protein supplements and 679 ± 41 mg/d in controls for calcium, 681 ± 46 mg/d in patients and 709 ± 35 mg/d in controls for phosphorus, and 44.2 ± 3.8 g/d in patients and 46.4 ± 2.5 g/d in controls for proteins (P > 0.2).

Function and Length of Stay

At the end of the intervention period, changes in the Activities of Daily Living score were 0.1 ± 0.2 in patients who received protein supplements and 0.7 ± 0.3 in controls (difference, 0.6 [95% CI, –0.2 to 1.4]; P = 0.124). After 6 months of protein supplementation, biceps muscle strength increased 28.2% ± 8.2% in patients who received supplements and 12.5% ± 5.4% in controls (difference, 15.7 percentage points [CI, –6.7 to 32.5 percentage points]; P = 0.177). Body weight, whole body lean and fat mass, and hand grip did not change significantly, nor did the groups differ significantly for these measures (data not shown). During recovery from surgery, patients who received protein supplements stayed in the orthopedic ward for 18.0 ± 1.4 days, whereas controls stayed 16.9 ± 0.9 days to start the rehabilitation program. Patients were then transferred to various rehabilitation hospitals. In these hospitals, where the staff was unaware of treatment group assignment, the rehabilitation stay of patients who received protein supplements was significantly shorter than that of controls (42.2 ± 6.6 days and 53.0 ± 4.6 days; median stay, 33 days [CI, 29 to 56 days] and 54 days [CI, 44 to 62 days]; difference, 21 days [CI, 4 to 25 days] P = 0.018).

Nutritional and Immunologic Status

Six months after hip fracture, levels of albumin, prealbumin, and IGF-I increased compared with baseline in patients who received protein supplements. Protein supplementation was associated with a significantly higher increase in prealbumin and IGF-I levels (Table 3, Figure 1). This favorable effect seemed to last beyond the end of the intervention period. During this period, the relative proportion of the 24-kD and 34-kD IGF-I binding protein decreased significantly (which probably represented IGF binding proteins 4 and 2) and the 41.5 ± 38.5 kD complex (IGF binding protein 3) increased (data not shown). Variability of reading on the various blots reduced the power to compare the groups.

View larger version (14K): [in this window] [in a new window]   Figure 1. Effect of protein supplements on serum levels of insulin-like growth factor-I (IGF-I) in patients with recent hip fracture. Results are the mean ± SE and are expressed as a percentage of baseline values. The solid line represents patients who received protein supplements; the dashed line represents controls. *P = 0.055; **P = 0.003 for comparison with controls (analysis of variance).

The increase in serum IgM level after hip fracture was significantly larger in patients who received protein supplements than in controls (P = 0.016). Changes in the size of the skin reaction as determined by the skin-test antigens for cellular hypersensitivity test of cell-mediated immunity were –0.3± 1.3 cm and –2.2± 1.1 cm, respectively, at 6 months (P > 0.2) and –0.1± 1.8 cm and –1.7± 1.8 cm, respectively, at 1 year (P > 0.2) in patients who received protein supplements and controls. The difference in outcome of isohemagglutinins anti-A and anti-B was not statistically significant.

Bone Mass and Bone Remodeling

During the year after hip fracture, a significant decrease in bone mineral density was detected, particularly at the lumbar spine in lateral view, the proximal femur (Figure 2), and the midfemoral shaft. Whole-body bone mineral content also decreased significantly (Table 4). Protein supplementation was associated with an almost 50% reduction of proximal femur bone loss at 1 year. When women were analyzed separately (23 patients who received protein supplements and 27 controls), the same pattern was seen, and the attenuation of bone mass loss with protein supplements was statistically significant for the femoral neck and the proximal femur (data not shown).

View larger version (13K): [in this window] [in a new window]   Figure 2. Effect of protein supplements on bone mineral density (BMD) of the proximal femur in patients with hip fracture 12 months earlier. Results are given as the mean ± SE and are expressed as a percentage of baseline values. The solid line represents patients who received protein supplements; the dashed line represents controls. *P = 0.029 for comparison with controls (analysis of variance).

Vertebral deformity was evaluated by quantitative morphometry on lateral radiography at 1 year. Seven and 13 vertebral deformities were detected among patients who received protein supplements and controls, respectively. This difference was not statistically significant.

All patients received a single oral dose of vitamin D, 200 000 IU. This was associated with an increase of 25-hydroxyvitamin D₃ levels at 6 months (Table 3). Plasma levels of calcium, phosphate, creatinine, serum parathyroid hormone, and 1,25-dihydroxyvitamin D₃ did not change during the study and did not differ between the patients who received protein supplements and controls. Osteocalcin levels increased in both groups but did not differ significantly between patients who received protein supplements and controls.

Discussion

Top Methods Results Discussion Author & Article Info References

Among elderly patients with recent hip fracture, oral protein supplementation was associated with a more pronounced increase in IGF-I levels and a more favorable outcome, including attenuation of proximal femur bone loss and shorter stay in rehabilitation hospitals.

Protein restriction in rats has been shown to reduce plasma IGF-I levels by inducing resistance of the liver to the effects of growth hormone through receptor and post-receptor defects [20]. Decreased translation, higher metabolic clearance rate, and lower sensitivity to the anabolic effects of IGF-I in peripheral tissues also occur during protein malnutrition [20, 48, 49]. Skeletal growth was not restored in rats that received a low-protein diet when they were given IGF-I in doses that normalized their plasma levels [50]. In elderly persons, low protein intake may be particularly detrimental not only for skeletal integrity but also for muscle function, probably because IGF-I production and circulating levels are lowered [11, 19-23]. Impairment of both systems may contribute to the occurrence of hip fracture [12, 13, 46, 51]. Furthermore, because IGF-I can stimulate the proliferation of immunocompetent cells and modulate cellular and humoral immune response [28], protein deficiency may also predispose patients to higher rates of infectious complication after fracture [7, 9, 16, 17], resulting in prolonged hospital stay.

The type of protein supplement that we used was well tolerated and had no major side effects, confirming the observations of previous studies [16, 17]. The protein content (20 g) corrected but did not overcompensate for the low protein intake recorded in patients with hip fracture [16]. Three frequency questionnaires administered at 6-month intervals produced values of daily protein intake of approximately 40 g. Thus, with the 20-g supplement, the total daily protein intake was close to the recommended amount of 1 g/kg of body weight. Compliance with the oral protein supplement was verified by making regular phone calls and counting the remaining bags. A more pronounced increase in prealbumin levels, which was associated with a higher IGF-I level, confirmed patient compliance.

As other researchers have found [9, 16, 17], protein supplementation was associated with a better clinical outcome, leading to a stay in the rehabilitation hospital of about 10 days. This shorter stay may be related to the lower rate of medical complications, as shown by previous studies of the same type of patients and the same protein supplement [16, 17]. The frequency of the complications that could account for the lower need for hospital care was not reliably recorded in our trial. However, because our study was double-blind, interference with the medical team's decision about the duration of hospital stay was unlikely. Certainly, the higher response of immunoglobulin levels and, to a lesser extent, cell-mediated immunity and muscle function may have played a role in accelerating functional recovery. If these findings are verified in larger populations of patients with hip fracture, protein supplementation may have a great impact on the costs of health care (an approximately 20% decrease in rehabilitation time) and on quality of life [5].

Small but positive effects on bone mineral density have been described in trials in which growth hormone was administered at a dose that restored decreased IGF-I levels in elderly men over a 6-month period [52, 53]. Despite convincing results with IGF-I in experimental animals [31, 32], substantial side effects have prevented the use of this growth factor in trials in humans [54, 55]. In our study, oral protein supplements given over 6 months increased plasma IGF-I levels much as protein repletion normalizes IGF-I levels in malnourished patients [56, 57]. Protein repletion could also sensitize peripheral target tissues to the action of IGF-I [20]. Thus, by modulating endogenous IGF-I production and action, protein supplements offer another advantage over exogenous IGF-I administration.

The selectively lower bone mineral density that we detected at the proximal femur is in agreement with previous studies [17, 44, 46, 47, 58]. It is probably related to osteoporosis predating the fracture because bone mineral density was measured within an average of less than 10 days after the fracture, making post-fracture immobilization-induced osteoporosis unlikely. Bone mass and the rate of bone turnover were changed by the protein supplement. The marked decrease in bone mineral density at the proximal femur occurring after hip fracture, possibly related to chronic underloading, was attenuated by about 50%. Similarly, protein supplements seemed to improve bone mineral density at the lumbar spine, particularly when this was measured in the lateral view. However, the high variability of this measurement and the relatively small number of patients prohibited detection of significant differences from baseline. Vertebral deformity was also assessed by quantitative morphometry at 1 year. The 7 and 13 deformities found in patients who received protein supplements and controls, respectively, seem to agree with the positive findings on bone mineral density.

Insulin-like growth factor-I has been shown to increase bone mass [31, 32], and evidence shows that IGF-I has direct effects on bone-forming cells. Exogenously administered IGF-I is associated with an increase in bone formation and resorption in humans. We found a trend toward a decrease in the specific bone resorption marker pyridinium cross-links. Whether this decrease is related to some prevention of bone resorption when endogenous levels of IGF-I are modulated over a long time or whether it results from an indirect action through improved muscle strength (and thereby better functional recovery) is not yet known.

Our stringent inclusion criteria may have caused baseline biochemical variables recorded in patients to be slightly less abnormal than those commonly encountered in this type of patient [16, 58]. We cannot exclude the possibility that selection of less affected patients led to underestimation of the magnitude of the effects observed. Patients with lower albumin levels had greater response to protein supplements (data not shown). Alternatively, the fracture or the surgical procedure itself could have affected the biochemical pattern. However, because the trial was double-blind and placebo-controlled, the difference observed can reliably be attributed to the effects of protein supplements. Finally, the dropout rate may limit the strength of our conclusions.

In conclusion, oral protein supplements given to patients with recent hip fracture improved clinical outcome and muscle strength and lessened loss of bone mass. This effect may be related to changes in IGF-I levels. Because they significantly reduced length of stay in rehabilitation hospitals, protein supplements may prove cost-effective. Protein repletion in frail elderly persons at high risk for hip fracture may help prevent falls and increase bone mass [51]. Whether protein supplementation contributes to the prevention of osteoporotic fractures, particularly those of the hip, remains to be tested.

From the World Health Organization Collaborating Center for Osteoporosis and Bone Diseases and University Hospital, Geneva, Switzerland; and INSERM Unit 403, Hopital Edouard-Herriot, Lyon, France.

Dr. Slosman: Division of Nuclear Medicine, Department of Radiology, University Hospital, CH-1211 Geneva 14, Switzerland.

Dr. Vadas: Central Chemistry Laboratory, Department of Pathology, University Hospital, CH-1211 Geneva 14, Switzerland.

Dr. Vergnaud: INSERM Research Unit 403, Hopital Edouard-Herriot, F-69437 Lyon, France.