SEARCH

SEARCH BY CITATION

Keywords:

  • GENETIC RESEARCH;
  • EPIDEMIOLOGY;
  • HIP FRACTURE;
  • MORTALITY;
  • TWIN

ABSTRACT

  1. Top of page
  2. ABSTRACT
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. Acknowledgments
  9. References
  10. Supporting Information

Several studies have shown a long-lasting higher mortality after hip fracture, but the reasons for the excess risk are not well understood. We aimed to determine whether a higher mortality after hip fracture exists when controlling for genetic constitution, shared environment, comorbidity, and lifestyle by use of a nationwide cohort study in hip fracture discordant monozygotic twins. All 286 identical Swedish twin pairs discordant for hip fracture (1972 to 2010) were identified. Comorbidity and lifestyle information was retrieved by registers and questionnaire information. We used intrapair Cox regression to compute multivariable-adjusted hazard ratios (HRs) for death. During follow-up, 143 twins with a hip fracture died (50%) compared with 101 twins (35%) without a hip fracture. Through the first year after hip fracture, the rate of death increased fourfold in women (HR = 3.71; 95% confidence interval [CI] 1.32–10.40) and sevenfold in men (HR = 6.67; 95% CI 1.47–30.13). The increased rate in women only persisted during the first year after hip fracture (HR after 1 year = 0.99; 95% CI 0.66–1.50), whereas the corresponding HR in men was 2.58 (95% CI 1.02–6.62). The higher risk in men after the hip fracture event was successively attenuated during follow-up. After 5 years, the hazard ratio in men with a hip fracture was 1.19 (95% CI 0.29–4.90). On average, the hip fracture contributed to 0.9 years of life lost in women (95% CI 0.06–1.7) and 2.7 years in men (95% CI 1.7–3.7). The potential years of life lost associated with the hip fracture was especially pronounced in older men (>75 years), with an average loss of 47% (95% CI 31–61) of the expected remaining lifetime. We conclude that both women and men display a higher mortality after hip fracture independent of genes, comorbidity, and lifestyle. © 2014 American Society for Bone and Mineral Research.


Introduction

  1. Top of page
  2. ABSTRACT
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. Acknowledgments
  9. References
  10. Supporting Information

Hip fractures constitute a major and growing public health problem among the elderly worldwide. Approximately 1.5 million hip fractures occur per year, of which 340,000 take place in the United States.[1] The number of hip fractures is projected to more than double by 2050.[1, 2] These fractures have a profound impact on societal costs and quality of life, with a low proportion of individuals regaining their prefracture level of function.[3, 4] Premature mortality after hip fracture is recognized, and despite improved hospital care, there is no trend of decreasing mortality after hip fracture.[5] The mortality rate after hip fracture has been reported to be long lasting in many[6-8] but not all[9, 10] studies. Nonetheless, in a recent meta-analysis, an excess annual mortality lasting up to 10 years after hip fracture was observed, along with higher mortality rates in men compared with women.[7]

Whether the higher mortality rate after the hip fracture event is a direct consequence of the fracture event,[11] the underlying comorbidity,[11, 12] an unhealthy lifestyle,[11] or genetic constitution is not well understood.[6, 7, 11] The co-twin control design[13] in hip fracture discordant identical twins offers a natural experiment as an attempt to evaluate the impact on mortality that is the result of the hip fracture event. If identical twins have a comparable mortality rate regardless of the hip fracture event in one of them, genetic constitution is a powerful explanation for the excess risk of mortality. Conversely, a higher mortality rate found in the hip fracture–affected monozygotic co-twin, after controlling for lifestyle and comorbidity in addition to genes and shared familial environment by pairwise comparisons, would suggest that the hip fracture event contributes to the higher risk of death. We, therefore, aimed to compare sex-specific mortality rates in pairs of identical twins discordant for hip fracture, taking comorbidity and lifestyle factors into account. As a comparison, we additionally performed analysis in like-sex unrelated twins of the same age.

Materials and Methods

  1. Top of page
  2. ABSTRACT
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. Acknowledgments
  9. References
  10. Supporting Information

Participants

Subjects were selected from the Swedish Twin Registry.[14] Twin pairs born in 1944 or earlier in which both were alive in 1972 were considered eligible for this study. Hip fractures were ascertained by international classification of diseases (ICD) codes (codes in the latest version ICD-10 S720 to S722) through the National Patient Register. The overall validity of the hip fracture diagnosis is high in this register.[15, 16] This register started in 1964, covered 83% of the Swedish population in 1972, and has covered all inpatient care in Sweden since 1987. Dates of death were obtained by linkage to the National Cause of Death Register. The individual personal identification number provided to all Swedish residents enabled linkage between health-care registers and questionnaire information. We identified all hip fractures and deaths that occurred after the age of 50 years during follow-up through December 31, 2010. A total of 286 hip fracture discordant identical twin pairs (210 female pairs and 76 male pairs) were identified: These twins form our primary study sample. A secondary aim was to investigate mortality rates in 1137 unrelated twin pairs (833 female pairs and 304 male pairs) to enable interpretation and comparison of our identical twin results with results from previous studies with an ordinary cohort design. In the unrelated analyses, twins with a hip fracture were individually matched by age and sex to an unrelated twin without a hip fracture history who was alive at the time of the hip fracture event. The study was approved by the Regional Ethical Review Board in Stockholm, Sweden.

Covariate information and statistical analysis

Comorbidity ICD diagnosis codes before the date of the hip fracture were collated from the National Patient Registry (with data from 1964 and onward[17]) to calculate the number of comorbidities and Charlson's comorbidity index.[18, 19] In addition, we used this register to retrieve ICD code information about alcohol or drug abuse (yes/no) and any psychiatric disease (yes/no).

Supplementary information was assessed by a comprehensive computer-assisted telephone interview conducted between 1998 and 2002. Previously, the twin pairs participated in mailed questionnaire surveys in 1970 (twins born in 1896 to 1925) and 1973 (twins born in 1926 to 1958). The overall response rate for the old questionnaires was about 90%,[20] and the response rate for the telephone interview was 84%.[21] The interviews included a number of items on lifestyle behavior, diseases, and symptoms related to fracture risk and survival,[14, 22] including smoking status (never, former, current), physical activity level (low, medium, high; similar categories have been found to be predictive both of hip fracture[23] and of mortality[24]) marital status (married or cohabitant, widowed, single), visual impairment (yes/no), hearing aid (yes/no), use of estrogen-replacement therapy (never/ever use), any prescribed medication (yes/no), nonprescribed or prescribed medication use (yes/no), present use of corticosteroids (yes/no), body mass index (BMI, kg/m2, continuous), alcohol abstainer (yes/no), and an index for activity of daily living (AD).[25] The ADL index was constructed from self-reported responses to the capacity of cooking/shopping, dressing/bathing, help with medication, help with household work, or other daily support. The questionnaire data information was time-updated when performing the analysis, ie, if the hip fracture event occurred before the telephone interview, we used the earlier questionnaire information, but if the hip fracture happened after the telephone interview, we used this latter information.

Statistical analyses were performed using SAS software version 9.3 (SAS Institute Inc., Cary, NC, USA), Stata 11.2 (Stata Corporation Inc., College Station, TX, USA), and R (R Foundation for Statistical Computing, Vienna, Austria, 2008). Missing questionnaire values were replaced by multiple imputation and the Markov chain Monte Carlo (MCMC) method and five imputations by PROC MI and MIANALYZE in SAS. In the imputation procedure, we used number of comorbidities, Charlson index, any psychiatric disease, alcohol or drug abuse, the interview/questionnaire information described above, age, sex, hip fracture status, and finally mortality status at end of follow-up. To compensate for the nonrandomized design of our observational study, we used propensity-score methods.[26] The individual propensity scores, defined as the conditional probability of obtaining a hip fracture based on covariates (ie, those described above but excluding mortality status), were estimated with a multiple logistic-regression model.

Follow-up time was accrued from date of the hip fracture event, ie, the same date for each hip fracture discordant pair, until date of death or the end of the study period (December 31, 2010). We estimated sex-stratified age and propensity score-adjusted hazard ratios (HRs) of death by intrapair Cox proportional hazards regression (PROC PHREG, SAS) and their 95% confidence intervals (CIs) for the twin with a hip fracture compared with her co-twin without a hip fracture. Thus, in the analyses, the pair status variable was used as a stratum variable, fixing the baseline hazard within a matched pair. Because the rate of death varied by time of follow-up, we display results from piece-wise models with cut points by 1-year of observation. Adjustment for other types of osteoporotic fractures (eg, distal forearm, spine, and proximal humerus) that occurred during follow-up affected our estimates only marginally (data not shown).

Moreover, we estimated mean difference in survival time between the twin with a hip fracture and her co-twin up to 1 year and up to 10 years after hip fracture (also including estimates of sex difference in survival) and finally to the point of expected survival according to Official Statistics Sweden at the median age of the hip fracture event. Average expected survival was defined based on average Swedish population data during the study period. The number of years lost with 95% CIs was calculated using a bias corrected method and accelerated bootstrap CI derived from the draw of 10,000 bootstrap twin pairs.[27] We expressed the difference in survival between the fractured and the nonfractured twins in years as well as a proportion relative to the expected survival time. In addition, the estimated total years lost because of the fracture were multiplied by the average annual number of hip fracture cases in Sweden during the study period based on data from the National Patient Register.

In the analysis of unrelated twins, we used flexible parametric models[28, 29] for continuous estimation of the sex-specific propensity score–adjusted HRs together with 95% CIs by time of follow-up. The 95% CIs were derived from 5000 bootstrap samples of twin pairs. The propensity score was based on the same model as the identical twin analysis.

Results

  1. Top of page
  2. ABSTRACT
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. Acknowledgments
  9. References
  10. Supporting Information

Descriptive characteristics of identical twins by sex and hip fracture status are described in Table 1. The men and women had a common median age of 75 years. Current smoking (p = 0.049 in men) and comorbidities (p = 0.002) were more frequently present in the hip fracture–affected twin. The within-pair differences of other covariates were not statistically significant.

Table 1. Descriptive Characteristics by Hip Fracture Status and Sex
 Women (n = 420)Men (n = 152)
Hip fractureNo hip fractureHip fractureNo hip fracture
n = 210n = 210n = 76n = 76
  1. Proportion (%) of twins with missing information: age (0%); weight, height, and body mass index (8%); smoking status (5%); leisure physical activity (25%); Charlson comorbidity index and number of comorbidities included in the Charlson score (0%); any psychiatric disease based on register data (0%); alcohol or drug abuse from register data (0%); abstainer (3%); visual impairment (28%); hearing loss (28%); estrogen-replacement therapy (28%); use of medication (28%); marital status (28%); and activities of daily living (3%).

 Mean (SD)
Age at hip fracture/entry (years)74.2 (10.3)74.2 (10.3)73.4 (10.7)73.4 (10.7)
Weight (kg)63.6 (7.5)64.8 (8.9)76.8 (9.4)77.2 (9.6)
Height (cm)164.1 (4.9)164.1 (4.2)176.0 (5.6)176.0 (4.7)
Body mass index (kg/m2)23.8 (2.9)24.3 (3.4)24.8 (2.7)25.1 (2.7)
 Number (%)
Smoking status
Former23 (11.5)17 (8.5)13 (18.1)20 (27.8)
Current32 (16.0)28 (14.0)22 (30.5)11 (15.3)
Leisure physical activity
Low69 (41.8)54 (34.4)25 (41.7)21 (36.2)
Moderate90 (54.5)97 (61.8)32 (53.3)34 (58.6)
High6 (3.6)6 (3.8)3 (5.0)3 (5.0)
Charlson index
No comorbidity139 (66.2)147 (70.0)39 (51.3)54 (71.0)
One comorbidity45 (21.4)43 (20.5)19 (25.0)18 (23.7)
Two or more26 (12.4)20 (9.5)18 (23.7)4 (5.3)
Any psychiatric disease18 (8.6)15 (7.1)6 (7.9)3 (4.0)
Alcohol or drug abuse4 (1.9)2 (1.0)5 (6.6)4 (5.3)
Abstainer50 (24.4)52 (25.4)8 (11.1)11 (15.3)
Visual impairment64 (42.1)67 (44.1)15 (28.3)16 (30.2)
Hearing loss47 (30.9)47 (30.9)17 (32.1)17 (32.1)
Estrogen-replacement therapy31 (20.4)34 (22.4)NANA
Use of cortisone7 (4.6)4 (2.6)00
Use of any type of prescribed medication69 (45.4)75 (49.3)24 (45.3)23 (43.4)
Use of prescribed or unprescribed medication80 (52.6)82 (54.0)26 (49.1)27 (50.9)
Married or cohabitant39 (25.7)49 (32.2)27 (50.9)28 (52.8)
No need for help with activities of daily living180 (87.8)185 (90.2)62 (86.1)68 (94.4)

During 3877 person-years of follow-up, 143 twins with a hip fracture died (50%) compared with 101 twins (35%) without a hip fracture (Table 2). Despite a similar median age at entry (75 years), more men (66%; 95% CI 54–76) than women (44%; 95% CI 37–51) died after the hip fracture event. Compared with the co-twin, the average 10-year loss of survival was 2.7 years in men (95% CI 1.7–3.7) and 0.6 years (95% CI 0.1–1.2) in women. The sex difference in survival was 2.0 years (95% CI 0.9–3.2; p < 0.001). At the age of 75 years, the average expected remaining survival time in Swedish men during the study period was 10 years compared with 13 years in women. These estimates were used to calculate the predicted average relative reduction in life expectancy associated with the hip fracture. It was found to be 26.6% (95% CI 17.1–36.9) in men and 6.9% (95% CI 0.4–13.4) in women.

Table 2. Deaths After Hip Fracture Analyzed in Identical Twin Pairs Discordant for Hip Fracture
 WomenMen
Hip fractureNo hip fractureHip fractureNo hip fracture
  1. a

    Adjusted by a propensity score that included age, number of comorbidities, Charlson index, smoking status, physical activity level, visual impairment, hearing aid, marital status, use of estrogen-replacement therapy, any prescribed medication, nonprescribed medication or supplement use, present use of corticosteroids, body mass index (BMI), weight, height, abstainer, alcohol or drug abuse, any psychiatric disease, and an index for activity of daily living (ADL).

  2. b

    Years lost resulting from the hip fracture up to the expected time of death from the median age at the hip fracture event (75 years, the median age at the hip fracture event in both men and women). The mean expected survival time during the study period at the age of 75 years was 10 years in men and 13 years in women. The relative reduction in life expectancy is the ratio between years lost to the expected survival.

Number of twins2102107676
Number of deaths93755026
Proportion dead (%; 95% CI)44 (37–51)36 (30–43)66 (54–76)34 (24–46)
Median follow-up (range) in years5.3 (26.1)5.9 (30.5)3.5 (24.5)5.1 (27.8)
Estimated years lost within 10 years0.6 (0.1–1.2)Reference2.7 (1.7–3.7)Reference
Estimated years lost in relation to life expectancya,b (95% CI)0.9 (0.1–1.7)Reference2.7 (1.7–3.7)Reference
Estimated relative reduction (%) in life expectancya,b (%; 95% CI)6.9 (0.4–13.4)Reference26.6 (17.1–36.9)Reference

The hazard ratio of death strongly depended on time after the hip fracture event (Fig. 1). Through the first year after the hip fracture, the rate of death was sevenfold higher in men (multivariable-adjusted HR = 6.67; 95% CI 1.47–30.13) and increased fourfold in women (adjusted HR 3.71; 95% CI 1.32–10.40) compared with the co-twin without a hip fracture. The average sex difference in survival time after the hip fracture during this first year of observation was 26 days (95% CI 3–57; p = 0.023). The increased rate in women only persisted during the first year after hip fracture, HR after 1 year = 0.99 (95% CI 0.66–1.50), whereas the corresponding HR in men was 2.58 (95% CI 1.02–6.62). The higher risk in men after the hip fracture event was successively attenuated during follow-up. After 5 years, the hazard ratio in men with a hip fracture was 1.19 (95% CI 0.29–4.90). The proportional hazards assumption was met for all subcategories of follow-up (proportionality tests p > 0.3). An age-adjusted curve (Supplemental Fig. S1) provided a similar pattern as that of the propensity-adjusted one in Fig. 1.

image

Figure 1. Hazard ratios (HRs) of death after hip fracture analyzed by pairwise Cox regression analysis in identical twin pairs discordant for hip fracture by sex and time of follow-up. The HRs were adjusted by a propensity score that included age, number of comorbidities, Charlson index, smoking status, physical activity level, visual impairment, hearing aid, marital status, use of estrogen-replacement therapy, any prescribed medication, nonprescribed medication or supplement use, present use of corticosteroids, body mass index (BMI), weight, height, abstainer, alcohol or drug abuse, any psychiatric disease, and an index for activity of daily living (ADL).

Download figure to PowerPoint

Subgroup analysis

Age at the hip fracture event affected our results, ie, a statistical interaction between the hip fracture event and age in men was detected (p for interaction = 0.017). The same effect modification was not evident in women (p = 0.35). We performed stratified analysis by our median hip fracture age 75 years (Table 3). At older ages (>75 years), the proportion of deaths and loss in life expectancy after the hip fracture was much greater in men than in women. In these 38 older male twin pairs, 12 (32%) men with a hip fracture died during the first year compared with none of their co-twins. During the first 5 years of follow-up, 26 (68%) men with a hip fracture had died compared with 5 (18%) men without a hip fracture. In total, 29 older men (>75 years) or 76% (95% CI 59–88) with a hip fracture were dead at end of follow-up (Table 3). The potential years of life lost resulting from the hip fracture was therefore especially pronounced in older men, with an average loss of 47% (95% CI 31–61) of the expected remaining lifetime. The corresponding estimate was 3.8% (95% CI –4.7 to 12.5) in women older than 75 years.

Table 3. Deaths After Hip Fracture Analyzed in Identical Twin Pairs Discordant for Hip Fracture by Age at the Hip Fracture Event
 WomenMen
Hip fractureNo hip fractureHip fractureNo hip fracture
  1. a

    Adjusted by a propensity score that included age, number of comorbidities, Charlson index, smoking status, physical activity level, visual impairment, hearing aid, marital status, use of estrogen-replacement therapy, any prescribed medication, nonprescribed medication or supplement use, present use of corticosteroids, body mass index (BMI), weight, height, abstainer, alcohol or drug abuse, any psychiatric disease, and an index for activity of daily living (ADL).

  2. b

    The average expected survival time among those twins aged <75 years (median age 66 years) was 17 years in men and 20 years in women, whereas those >75 years had an expected survival of 7 years (men, median age 81 years) and 7.5 years (women, median age 83 years). The relative reduction in life expectancy is the ratio between years lost to the expected survival.

Age <75 years    
Number of twins1011013838
Number of deaths37242115
Proportion dead (%; 95% CI)37 (28–47)24 (16–34)55 (38–71)39 (24–56)
Median follow-up (years)7.3 (26.1)8.2 (30.5)8.9 (24.5)9.5 (27.4)
Estimated years lost in relation to life expectancya,b (95% CI)2.5 (0.3–4.8)Reference1.5 (–0.6 to 3.8)Reference
Estimated relative reduction (%) in life expectancya,b (%; 95% CI)12.7 (1.4–23.8)Reference8.9 (–3.8 to 22.2)Reference
Age >75 years    
Number of twins1091093838
Number of deaths56512911
Proportion dead (%; 95% CI)51 (41–61)47 (37–57)76 (59–88)29 (15–46)
Median follow-up (years)3.3 (22.4)3.8 (16.5)1.2 (9.0)3.7 (13.8)
Estimated years lost in relation to life expectancya,b (95% CI)0.3 (–0.4 to 0.9)Reference3.3 (2.1–4.3)Reference
Estimated relative reduction (%) in life expectancya,b (%; 95% CI)3.8 (–4.7 to 12.5)Reference46.9 (30.6–61.0)Reference

We did not detect significant effect modification by Charlson's comorbidity. Thus, the p value for the interaction term between the hip fracture event and Charlson's comorbidity index was 0.32 among men and 0.23 among women.

Theoretical total number of annual years of life lost resulting from a hip fracture

Based on the average annual number of hip fractures in Sweden during the study period (4802 in men and 12,307 in women), the total number of annual years of life lost resulting from a hip fracture event in Sweden is 12,762 years (95% CI 8216–17,733) in men compared with 10,989 years (95% CI 713–21,483) in women. At older age (>75 years), the loss in life expectancy after the hip fracture in our study was much greater in men than in women. Therefore, the annual total years of lives lost in men at these ages were fourfold higher in men (11,430 years) than in women (2893 years), irrespective of the lower life expectancy in men.

Comparison to unrelated twins

To ease the comparison to previous ordinary cohort studies, we compared death rate after a hip fracture with that of an unrelated twin not suffering a hip fracture (Supplemental Fig. S2). The hazard ratios when performing comparisons with unrelated female twins tended to be increased for a longer follow-up time compared with the pairwise comparison in monozygotic female twins (Fig. 1), whereas the excess mortality pattern among men after a hip fracture was less affected by the analytical approach.

Discussion

  1. Top of page
  2. ABSTRACT
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. Acknowledgments
  9. References
  10. Supporting Information

We attempted to estimate the excess mortality associated with a hip fracture event, not only by consideration of the impact from comorbidity and lifestyle factors as in previous studies but also by controlling for genetic constitution and shared familial environmental factors through following identical twin pairs discordant for hip fracture. Regardless of these covariates, we found a higher mortality in both men and women that might therefore be attributable to the hip fracture itself. The excess mortality was particularly pronounced during the first year after the hip fracture.

Several studies have shown an enduring higher mortality after a hip fracture in both men and women, lasting at least 5 years.[6-8] The results, however, have been inconsistent, where the time frame of the excess risk of mortality has varied from 6 months[9] to more than 15 years.[7] Differences in the pattern of mortality risk in our study and in most earlier ordinary cohort analyses[7] are also an indication of genetic and common environmental confounding, as well as effects caused by gene-environmental interactions.[30] Intriguingly, even though the mortality rates in hip fracture–affected twins decreased substantially over time in our unrelated twin analysis, they did not return to rates seen in age- and sex-matched controls—a finding compatible with results from most earlier observational studies[7] and in contrast to our findings in identical twins, especially in women. Accordingly, part of the long-term excess risk of mortality observed in earlier studies can be explained by a mutual genetic liability to both hip fracture and death, eg, genetic variation in common between hip fracture and cardiovascular disease pathogenesis.[31] Unlike LeBlanc and colleagues,[32] we found no interaction between comorbidity and hip fracture status. This discrepancy between the studies can well be explained by our study design and analysis, in which we controlled for shared genetic and environmental factors that might explain associations between comorbidity, hip fracture risk, and mortality. Moreover, not only comorbidities but also variation in lifestyle factors, such as smoking,[33] diet,[34] and physical activity,[35] have a heritable component.

In our analysis of identical twins, men tended to lose more of their remaining lifetime both during the first year and within 10 years after the hip fracture. Notwithstanding, the number of years lost in association with the hip fracture event was on average increased in both women and men, but the magnitude is dependent on the person's age at the time of fracture. Our analysis indicated that especially vulnerable are men >75 years who seemed to experience a halved survival time compared with expected survival. In Sweden, about 72% of all hip fractures in men and 83% of all hip fractures in women occur in those aged >75 years of age. Accordingly, more emphasis should be targeted to the prevention of hip fractures in older men, by the use of bone antiresorptive treatment[36] and prevention of falls,[37] because both low BMD[38] and fall-related factors[25, 37] are of importance for the occurrence of fractures in men, as well as in women. There is no indication that fracture preventive efforts are likely to be less successful in men compared with women[39, 40] and also in those with comorbidities.[39, 41] Interestingly, the estimated excess risk of death in older men associated with the hip fracture event in our study was similar to the risk of prostate cancer death (ie, deaths specifically resulting from the prostate cancer) in men with noncuratively treated prostate cancer with distant metastases.[42] This implies a 30% difference after 1 year and a 50% difference after 5 years in cumulative probability of death related to the hip fracture and prostate cancer. The survival probability of a man after a hip fracture in our study is also similar to that after a stroke.[43]

Potential mechanisms

Our study established a more long-lasting influence on mortality by the hip fracture in men. Accordingly, in modern societies, women tend to have a lower mortality rate at every age level. Excess mortality in men after trauma, infectious diseases, cancer, and cardiovascular diseases is a well-known but unexplained phenomenon.[44-46] Further, for most animal species, males have a shorter life span than females.[47] The severity of the trauma is an unlikely explanation for the higher mortality in our study, in that the hip fracture is often caused by a low-energy fall from standing height. Other alternative conceivable explanations include oxidative stress and telomere length. High oxidative stress is a key element in the multifactorial process of aging.[48] Oxidative stress in humans is higher in males than in females,[49] which might be one explanation of higher male mortality after trauma and disease.[50, 51]

A related additional explanation for the sex difference in survival after a hip fracture is telomere length, which may serve as a biological clock.[48] Telomeres shorten faster in men than in women,[52, 53] and oxidative stress escalates telomere erosion.[48, 54] In fact, several of the genes regulating oxidative stress and inflammatory response are found on the X chromosome.[55] During a woman's lifetime, cells with a nonfunctioning X-chromosome gene expression are downregulated, and cells with more potent gene expression are retained.[55, 56] In addition, estrogen diminishes oxidative stress[48] and stimulates the transcription of the gene encoding the telomerase reverse transcriptase enzyme that adds telomere repeats.[48] Therefore, men who suffer a hip fracture might not have the same preconditions to interfere with disabilities and complications that are triggered by the hip fracture event.[57-60]

Strengths and limitations

Although observational, the main strength of our study is the robust discordant-twin-pair design among identical twins,[13] approaching a natural experiment with an optimal control for genetic constitution and multiple shared environmental factors. Additionally, we could control our estimates for the influence of comorbid conditions identified from nationwide registers and lifestyle factors known to influence the occurrence of hip fracture as well as survival. We had complete register identification of hip fracture events, mortality, and comorbidities. Treatment of diseases is generally equally provided in the Swedish tax-funded public health-care system, and it is mandatory for all hospitals, public or private, to report ICD diagnoses codes, date of contact, and the personal registration number to the National Board of Health and Welfare.[17] Thus, misclassification of major diseases before the hip fracture event is likely to be small.

Nevertheless, there are also several limitations that might affect the certainty and interpretation of our findings. The questionnaire information was not assessed in every twin, but restricting the analysis to twins with nonimputed data only marginally influenced our results (data not shown). In addition, questionnaire information was assessed at several occasions, but the interval to the hip fracture event could have been long with risk of more than ordinary misclassification of the covariates and residual confounding, which could have led to both conservative and inflated estimates. Our comorbidity index was based on inpatient care and less severe diseases were not captured from our register data, which might lead to residual confounding. Regrettably, although we had access to comorbidity information, we had no information regarding whether the twins were community-dwelling or institutionalized before the hip fracture event. The average age at the hip fracture event in our study is somewhat lower than the mean hip fracture age of the general Swedish population. The difference can be explained by the age distribution of the Swedish twin registry in which there is underrepresentation of the oldest old. The subgroup analysis by age at the hip fracture event should be cautiously interpreted. Nevertheless, we have previously shown a reduced genetic liability of hip fractures with increasing age,[21] and our present data indicate that mortality in men after the hip fracture event is modified by age. BMD before the hip fracture event was not assessed in this study, but Swedish identical twins seem to have high correlation of BMD values.[61] Nevertheless, a previous study in UK twins has revealed self-reported birth weight as a predictor of BMD in adulthood,[62] but the association was not independent of adult body weight and height, which was included in our propensity score adjustment. Our cohort was comprised of mainly white twins, and our findings might not be relevant to populations of other ethnic origin. Moreover, the results might not be applicable to countries with other health-care systems. Finally, studies in twins may be questioned in terms of their representativeness to the general population, although cumulative risks of late-onset diseases and death do not differ in twins compared with singletons.[63]

In conclusion, our study indicates a strong impact by the hip fracture on subsequent mortality independent of genetic constitution, comorbidity, and lifestyle, particularly in older men (>75 years) and thus an extra effort in hip fracture prevention should be directed to these men.

Acknowledgments

  1. Top of page
  2. ABSTRACT
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. Acknowledgments
  9. References
  10. Supporting Information

This study was supported by funding from the Swedish Research Council. The funding source had no role in defining the research question, abstracting data, synthesizing results, preparing the manuscript, or in the decision to submit the manuscript for publication.

Authors' roles: Study design: KM. Study conduct: KM. Data collection: KM, PN, NLP, and HM. Data analysis: KM and HG. Data interpretation: KM, PN, AN, LB, HG, NLP, and HM. Drafting manuscript: KM. Revising manuscript content: PN, AN, LB, HG, NLP, and HM. Approving final version of manuscript: PN, AN, LB, HG, NLP, and HM. KM takes responsibility for the integrity of the data analysis.

References

  1. Top of page
  2. ABSTRACT
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. Acknowledgments
  9. References
  10. Supporting Information

Supporting Information

  1. Top of page
  2. ABSTRACT
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. Acknowledgments
  9. References
  10. Supporting Information

Additional Supporting Information may be found in the online version of this article.

FilenameFormatSizeDescription
jbmr2029-sm-0001-SupAppFig-A1.tif79KSupplementary Appendix Figure 1
jbmr2029-sm-0002-SupAppFig-A2.tif1202KSupplementary Appendix Figure 2
jbmr2029-sm-0003-SupAppLegend-S1.doc37KSupplementary Appendix Figures Legend.

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.