Kendra Schwartz, MD, MSPH

BACKGROUND: Various interventions to prevent hip fractures have been tested with mediocre success. Most have treated underlying risk factors, such as osteoporosis and fall propensity. This study evaluated the effectiveness of an external hip protector to prevent hip fractures.

POPULATION STUDIED: The trial involved 1801 ambulatory but frail elderly adults (1409 women and 392 men, mean age=82 years) from 22 community-based health care centers in Finland. All patients were aged at least 70 years, were ambulatory (assisted or unassisted), and had at least one identifiable risk factor for hip fracture.

STUDY DESIGN AND VALIDITY: Patients were randomized using adequate concealment of allocation, in an unblinded manner, to receive a hip protector or to not receive one. The hip protector (KPH Hip Protector, Respecta, Helsinki, Finland) covered the greater trochanter, was 19 cm × 9 cm with a convex shape. It was designed to shunt the energy of an impact away from the greater trochanter, the most common site of hip fracture. The 2 padded protectors were worn inside pockets of a stretchy undergarment and did not limit walking or sitting. The subjects in the hip protector group were asked to wear the protector whenever they were on their feet and especially when they were at risk for falling. Many patients randomized to receive the hip protector (204 of 650) refused to participate. The dropout rate during the 18-month study period was high (657 out of 1427), mostly because of death, inability to walk, or refusal to continue in the study. The subjects from a waiting list replaced the dropouts. The sample size was sufficient to identify a 50% reduction in hip fractures over 1 year. As a group, hip protector subjects had significantly more risk factors for falls. However, statistical adjustment for baseline differences did not alter the results. The authors compensated for the high dropout rate by using an intention-to-treat analysis and including the subjects in the analysis for the time period of participation. Information on other factors associated with hip fractures (race, presence of osteoporosis, and use of osteoporosis medications) would have been helpful for generalizing the results to US patients.

OUTCOMES MEASURED: The primary outcome was hip fracture. Secondary outcome variables were the number and rate of falls in the hip protector group and the number of days the subjects wore the protector.

RESULTS: During follow-up, 13 subjects in the hip protector group had a hip fracture compared with 67 controls. Hip fracture risk was significantly lower in the treatment group (21.3 vs 46.0 per 1000 person-years; relative hazard 0.4; 95% confidence interval [CI], 0.2-0.8; P=.008). The risk of other fractures was similar in the 2 groups, which supports the effectiveness of the hip protector (ie, these patients were not at risk for fractures in general). Subjects in the hip protector group wore them during 48% of all days and during 74% of all falls, suggesting that they were being worn during higher risk times. In the hip protector group, 4 subjects had a hip fracture while wearing the device; 9 subjects had a hip fracture while not wearing it (P=.002). A total of 41 people would have to wear a hip protector for 1 year to prevent one hip fracture (95% CI, 25-115).

BACKGROUND: Various interventions to prevent hip fractures have been tested with mediocre success. Most have treated underlying risk factors, such as osteoporosis and fall propensity. This study evaluated the effectiveness of an external hip protector to prevent hip fractures.

POPULATION STUDIED: The trial involved 1801 ambulatory but frail elderly adults (1409 women and 392 men, mean age=82 years) from 22 community-based health care centers in Finland. All patients were aged at least 70 years, were ambulatory (assisted or unassisted), and had at least one identifiable risk factor for hip fracture.

STUDY DESIGN AND VALIDITY: Patients were randomized using adequate concealment of allocation, in an unblinded manner, to receive a hip protector or to not receive one. The hip protector (KPH Hip Protector, Respecta, Helsinki, Finland) covered the greater trochanter, was 19 cm × 9 cm with a convex shape. It was designed to shunt the energy of an impact away from the greater trochanter, the most common site of hip fracture. The 2 padded protectors were worn inside pockets of a stretchy undergarment and did not limit walking or sitting. The subjects in the hip protector group were asked to wear the protector whenever they were on their feet and especially when they were at risk for falling. Many patients randomized to receive the hip protector (204 of 650) refused to participate. The dropout rate during the 18-month study period was high (657 out of 1427), mostly because of death, inability to walk, or refusal to continue in the study. The subjects from a waiting list replaced the dropouts. The sample size was sufficient to identify a 50% reduction in hip fractures over 1 year. As a group, hip protector subjects had significantly more risk factors for falls. However, statistical adjustment for baseline differences did not alter the results. The authors compensated for the high dropout rate by using an intention-to-treat analysis and including the subjects in the analysis for the time period of participation. Information on other factors associated with hip fractures (race, presence of osteoporosis, and use of osteoporosis medications) would have been helpful for generalizing the results to US patients.

OUTCOMES MEASURED: The primary outcome was hip fracture. Secondary outcome variables were the number and rate of falls in the hip protector group and the number of days the subjects wore the protector.

RESULTS: During follow-up, 13 subjects in the hip protector group had a hip fracture compared with 67 controls. Hip fracture risk was significantly lower in the treatment group (21.3 vs 46.0 per 1000 person-years; relative hazard 0.4; 95% confidence interval [CI], 0.2-0.8; P=.008). The risk of other fractures was similar in the 2 groups, which supports the effectiveness of the hip protector (ie, these patients were not at risk for fractures in general). Subjects in the hip protector group wore them during 48% of all days and during 74% of all falls, suggesting that they were being worn during higher risk times. In the hip protector group, 4 subjects had a hip fracture while wearing the device; 9 subjects had a hip fracture while not wearing it (P=.002). A total of 41 people would have to wear a hip protector for 1 year to prevent one hip fracture (95% CI, 25-115).

BACKGROUND: Various interventions to prevent hip fractures have been tested with mediocre success. Most have treated underlying risk factors, such as osteoporosis and fall propensity. This study evaluated the effectiveness of an external hip protector to prevent hip fractures.

POPULATION STUDIED: The trial involved 1801 ambulatory but frail elderly adults (1409 women and 392 men, mean age=82 years) from 22 community-based health care centers in Finland. All patients were aged at least 70 years, were ambulatory (assisted or unassisted), and had at least one identifiable risk factor for hip fracture.

STUDY DESIGN AND VALIDITY: Patients were randomized using adequate concealment of allocation, in an unblinded manner, to receive a hip protector or to not receive one. The hip protector (KPH Hip Protector, Respecta, Helsinki, Finland) covered the greater trochanter, was 19 cm × 9 cm with a convex shape. It was designed to shunt the energy of an impact away from the greater trochanter, the most common site of hip fracture. The 2 padded protectors were worn inside pockets of a stretchy undergarment and did not limit walking or sitting. The subjects in the hip protector group were asked to wear the protector whenever they were on their feet and especially when they were at risk for falling. Many patients randomized to receive the hip protector (204 of 650) refused to participate. The dropout rate during the 18-month study period was high (657 out of 1427), mostly because of death, inability to walk, or refusal to continue in the study. The subjects from a waiting list replaced the dropouts. The sample size was sufficient to identify a 50% reduction in hip fractures over 1 year. As a group, hip protector subjects had significantly more risk factors for falls. However, statistical adjustment for baseline differences did not alter the results. The authors compensated for the high dropout rate by using an intention-to-treat analysis and including the subjects in the analysis for the time period of participation. Information on other factors associated with hip fractures (race, presence of osteoporosis, and use of osteoporosis medications) would have been helpful for generalizing the results to US patients.

OUTCOMES MEASURED: The primary outcome was hip fracture. Secondary outcome variables were the number and rate of falls in the hip protector group and the number of days the subjects wore the protector.

RESULTS: During follow-up, 13 subjects in the hip protector group had a hip fracture compared with 67 controls. Hip fracture risk was significantly lower in the treatment group (21.3 vs 46.0 per 1000 person-years; relative hazard 0.4; 95% confidence interval [CI], 0.2-0.8; P=.008). The risk of other fractures was similar in the 2 groups, which supports the effectiveness of the hip protector (ie, these patients were not at risk for fractures in general). Subjects in the hip protector group wore them during 48% of all days and during 74% of all falls, suggesting that they were being worn during higher risk times. In the hip protector group, 4 subjects had a hip fracture while wearing the device; 9 subjects had a hip fracture while not wearing it (P=.002). A total of 41 people would have to wear a hip protector for 1 year to prevent one hip fracture (95% CI, 25-115).