Introduction

Hip fractures, defined as any fracture of the femur between the articular cartilage of the hip joint to 5 cm below the distal point of the lesser trochanter, can occur at any age but are most common in older persons [1, 2]. Most patients presenting with hip fracture are women aged over 50 years, and the mean age at first presentation is approximately 80 years [1, 2]. Johnell and Kanis estimated a worldwide incidence of 1.6 million osteoporotic fractures of the hip in people aged 50 years and older in 2000, of which about 70% (1.14 million) were in women [3]. The absolute global annual incidence of hip fracture is expected to increase to 2.6 million by 2025 and to 4.5 million by 2050 alongside an expanding and increasingly elderly population [4, 5].

Most cases of hip fractures arise because of low-impact trauma in an individual with underlying bone fragility. In individuals aged 50 years or older, 53% of all fractures are associated with low-impact trauma (generally arising from a fall), rising to 80% of hip fractures in those aged 75 years or older [6]. The bone fragility that places older persons at increased risk of fracture following a low-impact trauma is most often due to underlying osteoporosis, suggesting that hip fracture is almost always due to underlying osteoporosis.

It has been suggested that as many as one in three women and one in eight men over 50 years of age have osteoporosis [7]. Moreover, recent surveys suggest that even after a diagnosis of osteoporosis, which is usually precipitated by a fragility-related fracture, patients often do not receive the recommended or adequate treatment [8–10].

Particularly worrying in light of the low rate of diagnosis and lack of sustained intervention following diagnosis is the apparent considerable risk of mortality following hip fracture. The U.S. Congress Office of Technology Assessment (1994) estimated that an average of 24% of patients 50 years and older with hip fracture die in the year following their index fracture. Johnell and Kanis estimated that there were almost 750,000 deaths worldwide associated with hip fracture in people aged 50 years and older in 1990 [5]. Other studies have indicated that the community mortality rates associated with hip fracture may be higher than for other better known life-threatening conditions such as pancreatic or stomach cancer [11] and myocardial infarction [12].

Although mortality following hip fracture is apparently high, to our knowledge, there has been no systematic evaluation of the current evidence base with regard to excess mortality rates. We undertook a systematic review of the literature in order to better define the mortality risk faced by individuals experiencing hip fracture. We have examined both crude mortality rates and the excess mortality these patients face in relation to control populations with the aim of highlighting at-risk groups where active intervention could modify risk factors for excess mortality.

Methods

PubMed was searched in October 2008 using the following search terms: [hip fracture OR femoral neck fracture OR neck of femur] AND [death OR mortality OR survival]. The following restriction was applied: language English. The search covered titles and abstracts.

Noninterventional/nonrandomized observational, prospective, and retrospective studies were eligible for inclusion in this analysis. Excluded study types were interventional studies, case studies, and meta-analyses. Studies with populations <100, those examining mortality rates primarily focusing on patients with a pre-existing serious medical illness (such as myocardial infarction, Parkinson's disease, or renal dysfunction), and those not specific for hip fracture mortality were excluded. The outcome measures that were of particular interest were unadjusted mortality (the absolute, observed mortality rate within a defined study population) and excess mortality (mortality beyond that expected/observed for matched control/population groups) and the relative risk (RR) for death compared with control groups without fracture.

Statistical analyses

No formal meta-analysis was possible due to the lack of consistency in study design across the included studies. Descriptive statistics are presented throughout.

Results

The initial PubMed search returned a potential 1,114 studies for consideration. In all, 1,052 studies were discounted after the exclusion criteria were applied, including 15 studies that were nonspecific for mortality associated with hip fracture, 14 studies in which mortality was not the primary outcome, four studies in which participants had an a priori medical diagnosis, two studies that reported extrapolated rather than actual mortality estimates, and one population surveillance study that did not specifically report data for hip fracture. In all, 63 studies were considered suitable for inclusion in the present systematic analysis (Table 1). The majority of studies included samples from populations older than 50 years, with a mean age of approximately 80 years. Subjects were mainly female and had largely been treated surgically. The studies had mostly been conducted in the USA or Europe, with additional studies reported for Japanese, Australian, and New Zealand populations.

Table 1 Overview of studies included in this analysis
Full size table

A total of 54 studies presented unadjusted mortality data (deaths as a proportion of the study population; Table 2). These studies were conducted in Argentina, Australia, Canada, Denmark, Finland, France, Greece, New Zealand, Norway, Singapore, Spain, Sweden, The Netherlands, UK, and USA. Cumulative unadjusted mortality rates increased over time following the index fracture, from 2.3–13.9% during the index hospitalization to 5.9–50% up to 1 year after the fracture.

Table 2 Unadjusted post-fracture mortality rates
Full size table

In all, 22 studies reported excess mortality rates compared with local population norms (Table 3). The studies were conducted in northern Europe (Sweden, Denmark), France, Germany, New Zealand, UK, and USA. Excess mortality rates compared with population-controlled cohorts (including general population rates) during the first year post-fracture were reported in 12 studies [13–24] and ranged from 8.4% in a Swedish population [13] to 36% in the USA [18].

Table 3 Relative mortality risk in hip fracture cases compared with nonhip fracture control populations
Full size table

Five studies reported an RR analysis for death following hip fracture and four other studies reported mortality hazard ratios (HRs) [13, 19, 21, 22, 24, 25–28]. The study by Wolinsky and co-workers reported a mortality risk analysis following hip fracture relative to institutionalized elders without hip fracture rather than relative to a general age-matched population [24]. Overall, the risk of mortality following hip fracture was at least double that for age-matched population values (Fig. 1).

Fig. 1
figure 1

Forest plot of risk (relative risk [RR], odds ratio [OR], or hazard ratio [HR], with 95% confidence intervals where available) of death following hip fracture compared with general population values

Full size image

Temporal profile of post-fracture mortality

Unadjusted mortality rates over time

The cumulative mortality rate over the first 12 months after hip fracture ranged from 5.9% among US patients aged 50–74 years (deaths identified via the mortality listings from the Vital Records Section of the Washington State Department of Social and Health Services) [23] to 50% among all patients admitted to a single US hospital for hip fracture between 1956 and 1961 [29] (Table 2). The inpatient mortality rate following hip fracture ranged from 2.3% among women attending a US urban orthopedic referral hospital [30] to 13.9% in patients treated at a single hospital in Norway [31].

Twelve studies examined the accumulation of mortality up to 1 year after fracture [13, 15, 32–41]. Of these, four studies reported mortality rates at 1 and 12 months post-fracture, all of which found that between one-quarter and one-third of the observed mortality occurred in the first month after fracture [34, 35, 37, 41]. Seven studies indicated that around half of the observed mortality occurs within the first 3 months post-fracture [15, 32, 33, 36, 38–40], with four studies indicating that around 70% of all observed deaths had occurred by 6 months post-fracture [13, 32, 38, 40].

Mortality risk over time

In the five studies that evaluated mortality risk over time, the highest risk of death was in the first 6 months after fracture [19, 21, 24, 25, 42] (Table 3). A standardized mortality ratio (SMR; observed/expected deaths) of 6.0 was calculated for the first 6 months after hip fracture among women in France compared with nonfracture controls matched for age and baseline health status and a history of falls; the SMR fell to 2.0 after 6 months [25]. Using a Cox proportional hazards model, Tosteson and co-workers found that hip fracture patients were 11.6 times more likely to die than controls within the first 6 months post-fracture after adjustment for age, sex, and race, with the risk reducing to 1.37 times that of controls thereafter [21]. In this study, the excess mortality risk was no longer significant beyond 6 months after adjustment for age, sex, race, prefracture functional status, socioeconomic status, facility residence, body mass index, comorbid conditions, and overall health status. Two studies utilized data from the Longitudinal Study of Aging to determine excess risk over time. Magaziner and co-workers compared 529 white, female community-dwelling hip fracture patients aged >70 years with 3,773 gender- and aged-matched nonhip fracture participants in the Longitudinal Study of Aging [42]. The authors determined an expected mortality rate for the nonhip fracture population (using a Cox regression analysis) and compared this with observed mortality rate among participants with hip fracture. In the first 2 months after fracture, the observed/expected ratio was 6.08, indicating an excess mortality, decreasing to 1.29 between months 6 and 12 post-fracture, and returning to equality (1.0) thereafter.

Wolinsky and co-workers found that the mortality risk was greatest during the first 6 months after fracture in their population of 368 participants aged >70 years in the Longitudinal Study of Aging compared with 7,159 age-matched participants in the same study [24]. They calculated an adjusted HR of 38.93 (95% confidence interval [CI], 29.58–51.23) for the first 6 months post-fracture compared with an adjusted HR of 1.17 (95% CI, 0.95–1.44) for the subsequent 7.5-year observation period (Table 3).

Rapp and co-workers examined the mortality risk following hip fracture among a population of >69,000 elderly people newly admitted to German nursing homes [19]. Using sex, age, and level of care-matched nursing home residents as the control group, they found that excess mortality was limited to the first 3 months post-fracture for women and the first 6 months post-fracture for men (HR women, 0–3 months 1.72 [95% CI, 1.59–1.86], >3–6 months 1.00 [95% CI, 0.89–1.13]; men, 0–3 months 2.14 [95% CI, 1.80–2.53], >3–6 months 1.40 [95% CI, 1.08–1.82]).

Farahmand and co-workers reported an RR of just under three times (2.7) that of the Swedish population in the 12 months post-fracture, with an RR of just over three times (3.3) that of the Swedish population in the first 6 months post-fracture [13]. Dahl and co-workers reported a mortality rate 15 times greater than for the general Norwegian population in the first month post-fracture and seven times greater in the second month [31]. Thereafter, for up to 4 years post-fracture, they found that the mortality rate was comparable with that for the general population.

Duration of increased mortality risk

Twelve studies examined the duration of the mortality risk in the years following hip fracture [11, 14, 16, 19, 21, 24, 25, 27, 28, 36, 43, 44]. Nine of the 12 studies reported that patients face an increased risk of death for several years following a hip fracture.

Two studies examined the mortality risk for up to 2 years following the index hip fracture [14, 43]. The risk of death at 2 years after an index hip fracture was 1.34 (95% CI, 0.83–2.16) in women and 7.18 (95% CI, 2.04–21.99) in men in one study [14], with an SMR at 2 years of 1.4 (p < 0.001 versus expected) reported in the second study [43]. Fisher and co-workers found that among a cohort of Medicare users in the USA, the excess risk for death among patients with hip fracture persisted for up to 3 years post-fracture [36]. At 5 years post-hip fracture, the mortality rate among men and women was significantly higher than for age-matched general population cohorts for all 5-year age groups from 50 to 90 years [11]. In a Japanese population, the mortality rate remained higher for individuals following hip fracture compared with the general population for up to 10 years after the index fracture [44]. Forsén and co-workers followed patients for up to 9 years after their index fracture [26]. They found that both men and women <75 years old experienced a 2- to 3-fold excess risk of death for at least 6.5 years for women and 5 years for men. Similarly, Paksima and co-workers reported that the excess SMR for patients aged 65–84 years persisted for up to 10 years post-fracture among a cohort of 1,109 patients with hip fracture admitted to a single US hospital [16]. Consistent with this, Johnell and co-workers found that the RR for death following hip fracture remained higher than among the general population for both men and women up to 5 years post-fracture [27]. Finally, the study reported by Vestergaard and co-workers suggested that the risk of death may persist for at least 20 years after the index fracture [28]. However, as noted above, in the study reported by Tosteson and co-workers, in which patients were followed for a median of 1.5 years after fracture, the excess mortality risk was no longer significant beyond 6 months after adjustment for age, sex, race, prefracture functional status, socioeconomic status, facility residence, body mass index, comorbid conditions, and overall health status [21]. Wolinsky and co-workers also found that the excess mortality was limited to the first 6 months post-fracture among patients aged >70 years who took part in the US-based Longitudinal Study of Aging [24]. Finally, Rapp and co-workers found that the excess mortality faced by institutionalized elders with hip fracture compared with institutionalized elders without hip fracture was also limited to the first 6 months post-fracture [19].

Hip fracture mortality and gender

Unadjusted mortality rates

The unadjusted mortality rates following hip fracture support a gender bias in favor of women both during the index hospitalization and in the months and years following the index fracture.

In a retrospective US analysis, the mortality rate in men was almost twice that in women while in hospital following hip fracture (unadjusted rate 4.3% versus 2.3% for women) [30]. The study also found that men had a higher preoperative risk (according to the American Society of Anesthesiologists' classification system) and were more likely to experience at least one postoperative complication. Similarly, Jiang and co-workers observed a significantly higher risk of inpatient mortality among men than in women (10.2% versus 4.7%; p < 0.001), a dichotomy that became more pronounced with advancing age so that for patients older than 90 years, inpatient mortality for men was 17.5% compared with 8.7% for women (p = 0.01) [45]. A significantly higher case fatality rate for males compared with females (males 11.9% versus females 5.3%; p < 0.001) was also reported by Benet-Travé and co-workers in their analysis of in-hospital mortality following hip fracture in a Spanish population [46]. Beals found that the in-hospital mortality rate was higher among males over 70 years compared with younger patients [29].

Studies of the mortality rates in the first year following hip fracture confirm a disparity between male and female patients. Within 1 month of the index fracture, the mortality rate among men was 17.1% compared with 9.8% for women (p < 0.01) in a Norwegian patient population [31] and men were less likely to survive to 30 days post-fracture in a large UK population (30-day mortality rate: men 12%, women 7%; odds ratio [OR] 1.93; 95% CI, 1.73–2.14) [47]. At 3 months post-fracture, the mortality rate for men was higher than in women (13% versus 6%) among Canadian hip fracture patients presenting at one of two acute care centers [48]. Several studies have reported that the increased mortality rate for men compared with women was still evident up to 1 year post-fracture [30, 32, 36, 45, 49–51]. Endo and co-workers reported that the increased risk of death for men versus women was still evident up to 1 year post-fracture (16.5% versus 9.4%; p < 0.01), while Jiang and colleagues observed a mortality rate of 37.5% for men and 28.2% for women (p < 0.001) after 1 year [30, 45]. Tosteson and co-workers noted that although mortality rates were higher among men than women in their study, this difference became less pronounced beyond 6 months after fracture [21]. In contrast, Parker and Anand found that among 703 consecutive patients admitted for hip fracture to a single UK hospital, the actuarial mortality rate to 1 year post-fracture was comparable between men and women (37.1% versus 36.6%, respectively) but was considerably in excess of the expected 1-year mortality for age-matched population norms (31.1% and 30.2% excess for men and women, respectively) [17].

Several studies have also revealed a markedly higher mortality rate among men compared with women for up to 20 years after fracture regardless of age [11, 16, 26, 28, 38, 40].

Excess mortality rates compared with age-matched controls

Consistent with the gender specificity of the unadjusted mortality rates, excess mortality compared with age-matched controls was higher among men than women regardless of the measurement employed or age group studied in five studies [11, 15, 19, 22, 27] but similar in one study [17] (Table 3).

Mortality risk and gender

As observed for general inpatient mortality, male patients appear to remain at higher risk of mortality in the months following the index hip fracture, with women having a 38% lower risk of death than men [33]. Three studies have reported that men face at least twice the risk of death following hip fracture compared with women [47, 49, 51]. Holt and co-workers reported an OR for death of 1.93 (95% CI, 1.73–2.14) for men compared with women at 30 days post-fracture [47]. Hindmarsh and co-workers reported an RR of death for men versus women of 2.2 (95% CI, 2.03–2.38) up to 1 year post-fracture [51]. Finally, Boereboom and co-workers found that during a 4-year follow-up, the RR for death was 1.88 (95% CI, 1.40–2.53) for men compared with women [49].

The gender specificity for excess mortality described above was reflected in the higher RR of death compared with the general population for male hip fracture patients than for female hip fracture patients. Forsén and co-workers found that, among patients <75 years old, men were at a 9-fold increased risk and women at a 5-fold increased risk of dying compared with controls during the first 3 months post-fracture [26]. Rapp and co-workers noted that institutionalized male and female elders with hip fracture faced at least twice the risk of death as institutionalized elders without hip fracture, with an increased risk persisting among male residents to 6 months post-fracture [19] (Table 3).

Hip fracture mortality and age

Unadjusted mortality rates

Unadjusted mortality rates following hip fracture increase with age both during the index hospitalization [29, 45, 46, 52] and in the subsequent months and years [13, 17, 20, 25, 28, 31, 35, 36, 40, 41, 49, 51, 53]. One study found that in a cohort of consecutive patients with hip fracture admitted to a single US hospital, separation by general health status negated the effect of age on mortality rates for all except those over 90 years of age [54]. However, data from this study conflict with those reported in the other 13 studies (in which an increase in mortality with increasing age was noted) in that no difference in crude mortality rates was noted for patients <85 years of age compared with those ≥85 years.

Several studies have identified increasing age as a predictor of mortality following hip fracture [16, 18, 22, 32, 33, 41, 48, 50, 55, 56–58].

Excess mortality rates

A number of studies have shown that the excess mortality following hip fracture is highest among younger patients, suggesting that the excess mortality due to hip fracture decreases with increasing age [11, 13, 20, 26–28, 36, 43, 44, 59]. Three studies found that the SMR was higher among younger versus older patients [20, 51, 60].

Mortality risk and age

Bass and co-workers used a Cox proportional hazards model to show that increasing age was positively associated with mortality and that the risk of mortality following hip fracture increased by approximately 5% for each additional year [33].

Two studies conducted risk analyses using younger patients with hip fracture as controls [60, 61]. Both studies found that the RR for death was increased in older compared with younger age groups. In their analysis of data from the Scottish Hip Fracture Audit, Holt and co-workers found that, compared with hip fracture patients aged 50 to 59 years, the OR for death was 1.78 (95% CI, 0.95–0.33) for those aged 60 to 69 years, 3.46 (95% CI, 1.94–6.15) for those aged 70 to 79 years, 5.68 (95% CI, 3.21–10.1) for those aged 80 to 89 years, and 7.11 (95% CI, 3.98–12.7) for those aged 90 years and over [60]. Similarly, Mortimore and co-workers found that among community residents of Baltimore (USA), the RR for death was 1.13 (95% CI, 0.76–1.67) for hip fracture patients aged 75 to 84 years and 1.59 (95% CI, 1.06–2.38) for those aged 85 years and over when compared with those aged 65 to 74 years [61].

Three studies conducted risk analyses compared with general population controls (Fig. 1) [13, 26, 27]. In Norway, the RR was highest among those aged 50–74 years (RR 4.2 in men versus 3.3 in women) in the 12 months after fracture [26]. Similarly, Johnell and co-workers reported that the RR for death following hip fracture in the Swedish population was higher among those aged 60 years than among those aged 80 years [27]. Farahmand and co-workers found that even though the absolute mortality rate increased with increasing age, the RR of mortality following hip fracture compared with general population values decreased from 8.4 among those younger than 70 years to 2.1 among those 76 years or older [13].

Discussion

A systematic review of the literature identified 22 studies that reported excess mortality for patients following hip fracture compared with the general population and a further 41 studies reporting survival data in fracture patients only. The majority of studies have shown that patients with hip fracture experience a significant excess risk for mortality that is at least double that of the age-matched population norms and which persists for several years after the index fracture. Both excess and unadjusted mortality rates among patients with hip fracture indicate that the greatest risk of death is within the first 6 months after the index fracture. In addition, most studies have confirmed that mortality following hip fracture increases with increasing age, although the excess mortality versus age-matched population norms decreases with increasing age. In other words, while older patients have higher mortality following hip fracture in absolute terms, the RR of death is greater in younger age groups where the expected risk of all-cause death is lower. Finally, in general, men face a greater excess risk of death after fracture than women regardless of the measurement employed or age group studied.

To our knowledge, the results reported here represent the first systematic analysis of the evidence base for excess mortality associated with hip fracture. However, there was a lack of consistency in the study designs and the statistical analyses used to determine excess mortality across the 22 studies that reported such data. Consequently, no meta-analysis or pooled analysis of the current dataset was possible.

The extent to which underlying conditions contribute to the excess mortality associated with hip fracture is unclear. Numerous studies have reported that the presence of concomitant medical illness or poor health status are negative predictors for survival following a hip fracture [16, 18, 20, 28, 31–33, 36, 37, 42, 43, 45, 49, 52, 54, 56, 57, 60, 62, 63], while other studies have found no association between concurrent life-threatening disease and mortality after hip fracture [50] nor an increased risk of death regardless of the presence of comorbid illness [13]. Tosteson and co-workers found that while adjustment for a variety of factors, including prefracture functional status and comorbid conditions, did not fully account for the excess mortality observed in the first 6 months after fracture, adjustment for these factors did eliminate the observed excess mortality beyond 6 months post-fracture [21]. Kanis and colleagues noted that hip fracture per se (rather than comorbidities) accounted for 17–32% of deaths in patients with hip fracture and was responsible for 1.5% of all deaths among persons aged 50 years or older [11]. In a separate large cohort study, Vestergaard and colleagues demonstrated that post-fracture conditions related to the trauma experienced had a greater influence on mortality than prefracture comorbidities [28]. Trauma-related complications accounted for 70.8% of the deaths occurring within 30 days of hip fracture, decreasing to 7.6% of deaths occurring more than 30 days after the fracture [28]. It is possible that the high proportion of in-hospital deaths classified as trauma-related on death certificates in this study may reflect the requirement to classify deaths in this way if there is any doubt that the death is due to natural causes. Accordingly, several studies have highlighted the contribution of selected comorbidities that increase or at least contribute to the higher risk of death following hip fracture including metastatic cancer, congestive heart failure, renal failure, liver disease, lymphoma, infection, and weight loss [13, 33, 45, 52, 64].

There are few data that can help determine whether hip fracture-related mortality has increased or decreased in recent years. Available studies have provided conflicting results: two studies suggest a trend toward increasing mortality following hip fracture in recent years [22, 65], while two other studies failed to identify either an increase or a decrease in hip fracture-related mortality in recent years [39, 66]. However, even with a stable rate of death following hip fracture, the actual number of fractures can be expected to increase in line with a growing and increasingly elderly global population.

Most patients presenting with hip fracture are treated surgically. Possible causes of death following surgical intervention for hip fracture include cardiac and pulmonary complications, infections (such as pneumonia, influenza, and septicemia), and an increased risk of thromboembolism [35, 37, 67]. A recent report indicated that 39% of inpatient deaths among patients with isolated limb and pelvic fractures were due to bronchopneumonia [35]. Lawrence and colleagues found that the risk of mortality increased with the number of postoperative complications and that serious cardiac and pulmonary complications were the most significant with respect to risk of death [37]. The relationship between the timing of surgery and the subsequent mortality risk has been the subject of some debate. There is evidence to suggest that patients who undergo hip fracture stabilization surgery within 48 h of the fracture event are at a reduced risk of death compared with those whose surgery is delayed [68–70]. However, there may be a number of barriers to achieving such early surgery including the patient's health status. Two studies found an increased risk of death, including death due to infection, among patients whose surgery was delayed beyond 72 h after admission [71, 72] but other studies have failed to find any significant benefit of early surgery (<24 h post-fracture) in terms of subsequent mortality [73, 74]. Furthermore, two studies found that for otherwise medically fit individuals, a delay of at least 4 days after admission did not appreciably affect survival [75, 76]. Surgery within 48 h of hospital admission may be difficult to achieve due to both organizational reasons and patient factors such as health status at the time of fracture [68]. While it has not yet been definitively demonstrated that early (<48 h after admission) treatment reduces the subsequent risk of mortality, it is widely regarded as prudent to surgically stabilize the fractured joint as soon as possible. The Royal College of Physicians guidelines recommend surgical repair within 24 h of admission [68].

Possible reasons for the increased mortality risk faced by men versus women following hip fracture are still poorly understood. One study suggested that a gender difference in terms of infection rates (notably pneumonia and septicemia) may contribute to the differential risk [77], although the etiological reasons for this remain unclear. In two studies, men were more likely to have a higher American Surgical Association (ASA) rating of operative risk (a system that assesses patients in terms of general disease burden [78]) than women, suggesting that men had more severe medical comorbidities prior to the index hip fracture [30, 47]. Endo and colleagues also reported that male gender was associated with an increased risk of postoperative complications, including pneumonia, arrhythmia, delirium, and pulmonary embolism, even after controlling for age and ASA rating [30]; other studies have failed to demonstrate such a link [79].

Patients experiencing one fragility-related hip fracture are at increased risk for subsequent fractures [80–83]. Despite this, it would appear that such patients are inadequately investigated [82] and often do not receive the recommended or adequate treatment [8–10, 84]. Notably, few patients who have experienced a hip fracture are prescribed osteoporotic treatments such as bisphosphonates, and in many cases only calcium and vitamin D are prescribed. Treatment rates of around 20–30% are generally cited [85–90] but estimated treatment rates vary considerably, possibly reflecting local practices in different countries [85–90]. Encouragingly, recent studies have indicated that pharmacologic treatment for osteoporosis may decrease the risk of subsequent hip fractures [91] and potentially also the increased risk of death [92]. Nonpharmacological approaches to maximizing peak bone mass, such as regular exercise and calcium and vitamin D supplementation, are established approaches to the management of osteoporosis and may also contribute to the prevention of fractures [93]. Indeed, there is some suggestion that interventions such as nutritional supplementation [94] and dietetic assessment [95], comprehensive multidisciplinary intervention programs [96, 97], and in-hospital rehabilitation programs [98] may also improve outcomes, including mortality.

By conducting a systematic review of the current evidence base with regard to hip fracture-related mortality, we have confirmed the assumption that patients with hip fracture experience a marked and significant excess risk of death.

Despite calls to improve the identification, assessment, and treatment of patients at risk of first or subsequent osteoporotic hip fractures [10, 87], many patients remain poorly treated on discharge from hospital [8, 9]. Our review has raised a number of questions, perhaps the most important of which is why there has been an apparent increase in mortality following hip fracture. Additional research is now needed to identify the reasons for the apparent increase in post-hip fracture mortality and to develop methods to distinguish between health outcomes that are a direct consequence of the fracture and those that result from pre-existing/comorbid medical conditions. Future research should also focus on establishing whether and which interventions, such as those for osteoporosis, can effectively reduce the risk of death following hip fracture. Properly viewed, the high and long-lasting excess mortality risk associated with hip fracture should be a strong incentive rather than a barrier for the establishment of tertiary prevention programs for osteoporotic fractures, including fracture liaison services. There is a need to ensure evaluation for osteoporosis in all patients following hip fracture and to implement and ensure long-term compliance with treatment regimens, including pharmacotherapy, with demonstrated improvement in treatment outcomes and adherence to therapy.