Intended for healthcare professionals


Genetic factors and osteoporotic fractures in elderly people: prospective 25 year follow up of a nationwide cohort of elderly Finnish twins

BMJ 1999; 319 doi: (Published 20 November 1999) Cite this as: BMJ 1999;319:1334
  1. Pekka Kannus, chief physician (klpeka{at},
  2. Mika Palvanen, research fellowa,
  3. Jaakko Kaprio, professorb,
  4. Jari Parkkari, research fellowa,
  5. Markku Koskenvuo, professorc
  1. a Accident and Trauma Research Center and the Tampere Research Center of Sports Medicine, UKK Institute for Health Promotion Research, PO Box 30, FIN-33501 Tampere, Finland
  2. b Department of Public Health, University of Helsinki, PO Box 41, FIN-00014, Helsinki, Finland
  3. c Department of Public Health, University of Turku, FIN 20520, Turku, Finland
  1. Correspondence to: P Kannus UKK Institute, Kaupinpuistonkatu 1, FIN-33500 Tampere, Finland
  • Accepted 24 August 1999


Objective: To determine whether genetic factors partly explain variation in risk of osteoporotic fracture, the true end point of the osteoporosis problem.

Design: Prospective 25 year follow up of a nationwide cohort of elderly Finnish twins.

Setting: The Finnish twin cohort and the national hospital discharge register, covering the entire 5 million population of Finland.

Subjects: All same sex twin pairs born before 1946. The cohort contained 2308 monozygotic and 5241 dizygotic twin pairs (15 098 people) at the beginning of follow up.

Main: outcome measure The number and concordance of osteoporotic fractures in the twin pairs, 1972-96.

Results: 786 cohort members sustained an osteoporotic fracture. In women, the pairwise concordance rate for fracture (that is, the relative number of twin pairs in whom the fracture affected both twins in a pair) was 9.5% (95% confidence interval 5.3% to 15.5%) in monozygotic pairs and 7.9% (5.2% to 11.4%) in dizygotic pairs. In men, the figures were 9.9% (4.4% to 18.5%) and 2.3% (0.6% to 5.7%).

Conclusions: Susceptibility to osteoporotic fractures in elderly Finns is not strongly influenced by genetic factors, especially in elderly women. The traditional strategy for prevention of osteoporotic fractures—that is, increasing peak bone mass and preventing age related bone loss—should be changed to include new elements such as prevention of falls and protection of the critical anatomical sites of the body when a fall occurs.

Key messages

  • Genetic factors have a substantial role in explaining age specific variation in bone mass and density, but no previous study has directly evaluated whether they have a role in the variation of risk of osteoporotic fracture, the true end point of the entire osteoporosis problem

  • Genetic factors are not strongly related to likelihood of osteoporotic fracture, particularly in elderly women

  • For this reason, the traditional prevention strategy of osteoporotic fractures—increasing peak bone mass and preventing age related bone loss—could include new additional elements, such as prevention of falls in elderly people and protection of the critical anatomical sites of the body when a fall occurs


The large number of osteoporotic fractures among elderly people represents a worldwide epidemic, and the predicted ageing of populations will further increase the burden of these minimal trauma fractures on our healthcare systems.14 In addition to high costs, osteoporotic fractures are associated with high morbidity and disability, high risk for long term institutionalisation, and increased risk of death.15

Bone mineral density and bone mineral content, as measured by absorptiometry, are predictors of osteoporotic fractures of the spine and proximal femur, the sites of clinically important fractures. 1 6 7 Twin and other types of family studies have, in turn, consistently shown that genetic determinants have a substantial role in explaining age specific variation between individual people in bone mineral density and bone mineral content at various anatomical sites of a skeleton,813 heritability thus being an important determinant of risk for osteoporosis in elderly women.13 Nevertheless, despite the fact that reduction of the number of fractures can be the only ultimate goal in the prevention and treatment of osteoporosis, previous twin studies have not directly examined whether genetic factors can explain some of the variation in risk of osteoporotic fracture in elderly people.

We examined whether genetic factors in elderly individual people are related to their susceptibility to osteoporotic fracture. We thought that this information would be valuable and of help in planning the strategies for fracture prevention. Our hypothesis or suspicion was that the role of genetic factors is not so clear cut when the end point of the study is changed from osteoporosis to actual fractures.


The Finnish twin cohort

The Finnish twin cohort comprises all same sex twin pairs born before 1958 with both cotwins alive in 1975. An extensive questionnaire was posted to the twins in 1975 to confirm twinship, determine zygosity, and obtain data on health related variables. The overall response rate was 89%.

Twin zygosity was determined by examining the responses of both members of each twin pair to two questions on the similarity of appearance at school age, items similar to those used in other large twin samples.1416 A set of decision rules was then used to classify the twin pairs as monozygotic, dizygotic, or undetermined zygosity. The validity of the zygosity was studied in a subsample of 104 pairs, and the agreement in classification from the questionnaires and 11 blood markers was 100%.17 The estimated probability of misclassification was 1.7%

The total number of twin pairs born before 1946 was 7549 at the beginning of the prospective follow up of the cohort. Of these, 2308 were monozygotic pairs and 5241 dizygotic pairs. Among the monozygotic pairs, 1221 pairs were female and 1087 male, while among the dizygotic pairs, 2618 pairs were female and 2623 male. The approximately 2:1 dizygotic to monzygotic ratio reflects the high rate of twinning in Finland during the first half of the 20th century, a phenomenon discussed in detail elsewhere.18

Identification of osteoporotic fractures among twins

Using the guidelines from previous epidemiological studies of osteoporotic fractures 2 3 19 20 we defined an osteoporotic fracture as a fracture that occurred in a person aged 50 years or more as a consequence of a only minimal trauma—that is, a fall from standing height or less. Fractures caused by a vehicular accident or other high energy trauma could be excluded as the Finnish national hospital discharge register also contains data on cause of injury. Previous investigations indicated that most osteoporotic fractures occur at hip, pelvis, knee (distal femur, patella, or proximal tibia), ankle, thoracic and lumbar spine, ribs, proximal and distal humerus, and wrist, and therefore these anatomical sites were used in this study too.

According to the above described age criterion, all 50 year old or older twin cohort members who were admitted to Finnish hospitals (1972-96) for primary treatment of an osteoporotic fracture were selected from the national hospital discharge register. Our register is the oldest nationwide discharge register in the world and its accuracy and coverage have been shown to be good (in injuries, 95% and over) and these percentages are particularly good in severe injuries such as bone fractures.2123

Determination of concordance and risk of fracture among twins

Twin similarity for osteoporotic fractures was summarised by estimates of concordance. This could be assessed from two types of concordance (termed pairwise and probandwise), each calculated separately for monozygotic and dizygotic pairs.24 The 95% confidence intervals for concordance of fractures were also computed.

Pairwise distribution of twins with osteoporotic fracture in Finnish twin cohort*

View this table:

The pairwise concordance is the relative number of twin pairs in whom disease (osteoporotic fracture) has affected both twins in a pair, and it is calculated with a formula of C/C+D, in which C is the number of concordant pairs and D is the number of discordant pairs.25 The probandwise concordance is defined as an individual's risk of disease (the conditional probability that one twin is affected, given that his or her cotwin is affected) and as such can be compared with the probability of disease for an individual in the general population. The probandwise concordance is calculated with a formula of 2C/(2C+D).25

The overall cumulative risk for fracture (with 95% confidence intervals) was calculated by dividing the number of fracture cases by the total number of individuals. The relative fracture risk (with 95% confidence intervals) was, in turn, calculated by dividing the above noted probandwise concordance by this overall cumulative fracture risk.


Overall cumulative risk for fracture—Between 1972 and 1996, 786 cohort members sustained an osteoporotic fracture that required hospital treatment, the overall cumulative risk for fracture being similar in monozygotic twins (all 5.4%; men 41%; women 6.6%) and dizygotic twins (all 5.1%; men 3.5%; women 6.8%) (table).

Pairwise concordance for fracture—The pairwise concordance for fracture was 9.6% (95% confidence interval 6.2% to 14.2%) in monozygotic pairs and 5.9% (4.0% to 8.4%) in dizygotic pairs, the observed difference of 3.7% having a confidence interval of −0.6% to 8.1%. By sex, this concordance was 9.9% in monozygotic male pairs (9.5% in women) and 2.3% in dizygotic male pairs (7.9% in women) (table). In hip fractures, the pairwise concordance was 7.8% (3.4% to 15.0%) in monozygotic pairs and 6.7% (3.5% to 12.0%) in dizygotic pairs, the difference of 1.0% having a confidence interval of −5.3% to 7.4%.

Probandwise concordance for fracture—The probandwise concordance for fracture was 17.6% (11.2% to 24.0%) in monozygotic pairs and 11.2% (7.5% to 14.9%) in dizygotic pairs. By sex, this concordance was 18.0% in male monozygotic pairs (17.4% in women) and 4.4% in male dizygotic pairs (14.6% in women) (table). In hip fractures the probandwise concordance was 14.4% (5.5% to 23.0%) in monozygotic pairs and 12.6% (6.2% to 19.0%) in dizygotic pairs.

Relative risk for fracture—In men the relative risk for fracture was 4.39 (2.70 to 7.16) in monozygotic pairs and 1.28 (0.64 to 2.56) in dizygotic pairs (table). Among women, the risks were 2.64 (1.83 to 3.81) and 2.16 (1.65 to 2.83), respectively. In hip fractures, the relative fracture risk was 5.99 (3.68 to 9.78) in monozygotic pairs and 6.97 (4.67 to 10.4) in dizygotic pairs.

Site specific results—The numbers of specific fractures except hip fracture were not large enough for meaningful site specific analyses; a table of the full results can be found on the BMJ's website.


Our study indicates that in elderly women genetic factors are only weakly related to the likelihood of hospital admission for an osteoporotic fracture. As twin estimates of heritability are likely to represent the upper boundary of the genetic effects2628 the modest genetic effect seen in men should not be emphasised too much either. On the other hand because just a small number of male pairs were concordant for fractures (eight monozygotic pairs and four dizygotic pairs), only the coming years will show the true fracture development in our male twin cohort. In this respect our results for women are more convincing, but again, more incident cases will be needed to increase the statistical power of the study and thus for more definitive conclusions.29 This will be especially important for different fracture types as family history studies seem to suggest that a positive family history for a specific fracture (wrist, hip) increases risk only for that specific fracture. 30 31

Although the data on fractures were collected only from hospital admissions, which represent in some fracture types (such as wrist and vertebral fractures) only a proportion of the fractures in the population, it was unlikely that the people admitted to hospital were selected according to zygosity. So we think that the conclusions of this study are unbiased and valid.

The reason for our finding in women (that is, we could not show that the presumptive genetic effect on bone mass and density had an important role in explaining variation between individuals in risk for osteoporotic fracture) cannot be understood by looking at the results of previous studies that evaluated the risk factors for osteoporosis per se. A review of the recent studies of risk factors for osteoporotic fractures 30 3237 indicates that the determinants of an osteoporotic fracture are largely independent of bone mass and density. These studies have suggested that in the pathogenesis of osteoporotic fractures, the falling, the direction and mechanism of falling, the protective neuromuscular responses, the impact energy created by the fall, and the capacity of the soft tissues around the impact site to absorb energy rather than bone quality and quantity are the main determinants of the fracture, and it is easy to understand that these determinants, especially falling, are largely controlled by unshared environmental factors. 30 3237

Our results could help to enlarge our view on prevention of osteoporosis. As prevention of fractures in elderly people is the ultimate goal in prevention and treatment of osteoporosis, the population level strategies for fracture prevention could, in addition to the traditional means of increasing peak bone mass and preventing age related bone loss, include serious efforts for diminution of the number and severity of falls in older adults and protection of the critical anatomical sites of the body when a fall occurs. The first interventions in prevention of falls in elderly people and protection of their proximal hip by external protectors have been promising, 38 39 giving hope that the increasing number of age related fractures could be controlled.


We thank the Finnish Ministry of Health for its cooperation and permission to use the hospital discharge data in the study.

Contributors: PK carried out the study, was involved in the design and analysis of the data, and wrote the basic manuscript. MP and JP were involved in the study design, data analysis, and writing of the paper. JK was involved in the design, data management, analysis of the study, writing of the paper, and organisation of funding. MK contributed to the study design and analysis, finalisation of the manuscript, and organisation of funding. PK is the guarantor.


  • Funding Medical Research Fund of Tampere University Hospital, Tampere, Finland, and the Academy of Finland (grants no 38332 and 42044).

  • Competing interests None declared.

  • website extra Further results can be found on the BMJ's website


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
  25. 25.
  26. 26.
  27. 27.
  28. 28.
  29. 29.
  30. 30.
  31. 31.
  32. 32.
  33. 33.
  34. 34.
  35. 35.
  36. 36.
  37. 37.
  38. 38.
  39. 39.