Risk of incident fractures in individuals hospitalised due to unexplained syncope and orthostatic hypotension

The Malmö Diet and Cancer (MDC) study is a large prospective population-based study launched in the 1990s. A total of 30,446 participants were included (41% participation rate; age, 57.5 ± 7.6; women, 60.2%) and attended the baseline examination between 1991 and 1996. Details of the MDC study have been described elsewhere [9].

In the current study, we excluded individuals with fractures that had occurred prior to hospitalisation for syncope or orthostatic hypotension (n = 39), and individuals with missing follow-up time data (n = 8), leaving 30,399 individuals that were included in the final analysis. Of these, 8201 subjects had incident fractures occurring after baseline examination (Fig. 1). One of the primary outcomes were incident fractures occurring after hospital admissions for unexplained syncope or orthostatic hypotension compared with the individuals without a history of unexplained syncope or orthostatic hypotension hospitalisation.

Data on hospitalisation for unexplained syncope, orthostatic hypotension, and incident fractures were retrieved from the Swedish National Patient Register [10], which includes complete coverage of public in-patient care since 1987 and out-patient care since 2001. Thus, fractures occurring prior to 2001 were only included if they resulted in hospital admission.

Follow-up data on hospitalisations was available through December 31, 2016. Follow-up time was defined as the time from baseline examination until the first incident fracture, death, emigration, or end of follow-up on December 31, 2016, whichever occurred first.

Unexplained syncope and orthostatic hypotension were defined according to the International Classification of Diseases 9th and 10th Revisions (ICD-9 and ICD-10), based on the primary or main secondary discharge diagnoses: 780.2, R550.9; 458 and I951.

Fractures were defined according to the following ICD-10 codes: spine and thorax (S12.x, S22.x, S32.x); arms, shoulders, and hands (S42.x, S52.x, S62.x); and pelvis (S32.x), hips, femur, and lower leg (S72.x, S82.x). In addition, ICD codes signalling spontaneous fractures and peri-implant fractures were also sought (M48.4 non-traumatic vertebral fracture, M84.3 stress fracture, and M96.6 peri-implant fractures; however, M97 is not used in Sweden). A total of 503 fractures occurring prior to 1997 and coded by corresponding ICD-9 codes were included.

The following variables were obtained at baseline examination: age (n = 30,393), sex (n = 30,399), body mass index (BMI) (n = 30,341), systolic blood pressure (SBP) (n = 30,344), and diastolic blood pressure (DBP) (n = 30,341). Blood pressure was measured using a sphygmomanometer and properly sized right arm cuff after a 10-min rest in the supine position. The following variables were self-reported in the questionnaire: current smoking, defined as regular or occasional smoking (n = 28,523), and family history of fracture, defined as any fracture in mother/father/sibling after 50 years of age (n = 30,399). Since bone density data is unavailable in the MDC cohort, we used data on prevalent fracture and family history of fracture as a proxy of poor bone health. Data on fractures was collected from the Swedish National Patient Register.

Two Cox regression models were constructed to assess the association of unexplained syncope and orthostatic hypotension hospitalisations with subsequent fractures. The basic model included age and sex, whereas the multivariable model included age, sex, prevalent fractures, family history of fractures, BMI, and current smoking.

The long-term cumulative incidence of fractures was stratified according to the presence or absence of hospital admission for unexplained syncope and orthostatic hypotension and calculated using the Kaplan-Meier method with quantification using the log-rank test.

All statistical analyses were performed in IBM SPSS Statistics 27 (IBM Corporation, Armonk, NY, USA).

To our knowledge, this is the first prospective observational study investigating the association between incident fractures following hospitalisations due to unexplained syncope and orthostatic hypotension. The strength of this study is the use of a large population-based cohort with over 30,000 middle-aged individuals enrolled and minimal loss to follow-up. Moreover, data on hospitalisations due to syncope and orthostatic hypotension were retrieved from a national register with complete coverage of public healthcare in Sweden, making our findings reliable and robust. Previous studies have focused on examining the association between unexplained falls in individuals with syncope and orthostatic hypotension [4, 7, 12]. Our findings provide additional knowledge related to the importance of fracture prevention in the population. The mortality in the MDC cohort has been explored in several previous studies. For example, a recent study showed that the cumulative survival after 10 years is around 90% [13]. A previous study from our research group concluded that syncope and OH increased the risk of death in the MDC cohort [14]. Thus, based on the assumption of competing risk, we anticipate that this would underestimate rather than overestimate our current associations.

Our study has some limitations that need to be addressed. The number of subjects that was excluded due to incident fracture before hospitalisation for syncope or OH was small. Two of the baseline variables (current smoking and family history of fracture) were self-reported in the questionnaire, suggesting that there may be some response bias. Also, data on bone density was unavailable in the cohort; consequently, we used prevalent fracture and family history of fracture as a proxy of poor bone health. Data on serum vitamin D levels, vitamin D supplementation, dementia, and baseline resting heart rate was unavailable in the cohort, which otherwise would have been of interest to analyse. One explanation for the observed lower hazard ratio in the syncope groups compared with OH hospitalisation may be that individuals with OH were older and may have more co-morbidities. Although therapy for syncope is available once the arrhythmic or hypotensive aetiology is established, patients with syncope and OH may demonstrate treatment failure due to comorbidities that are commonly observed at this age, e.g. cognitive and physical impairment. Unfortunately, we have no data available on intervention against syncope hospitalisation.

To the best of our knowledge, our study is the first to provide effect sizes of unexplained syncope and orthostatic hypotension in terms of incident fracture risk in the population. A correlation between cardiovascular instability, hypotension, and fractures has been previously demonstrated, suggesting that fractures due to low-energy trauma (falling from a standing height or less) may serve as a valid surrogate marker for syncope and blood pressure instability [15, 16]. In a previous study of individuals aged 50 and above, the prospective risk of fracture associated with delayed BP recovery was 7% [17]. As only 5–10% of falls lead to a fracture [18], we believe our study, examining incident fractures as an outcome, provides additional clinically relevant information, since previous studies on syncope and orthostatic hypotension have only investigated falls as an outcome. Thus, studying the risk of fracture due to falls is highly relevant, as falls are the single strongest risk factor for fracture, suggesting that the focus in fracture prevention may benefit from a shift from osteoporosis to falls [19].

In 90% of cases in the elderly, femoral fractures are preceded by a fall [20]. Unexplained falls commonly lead to hospitalisations and increased healthcare costs [21], and over 33% of falls at orthopaedic wards are unexplained [22]. A previous study found that 25% of adults aged above 50 years presenting to the emergency department due to a fall had symptoms suggestive of syncope or an unexplained fall [4]. In addition, nearly 20% of individuals above 70 years with orthostatic hypotension presented as falls, signalling that orthostatic hypotension is also an important independent risk factor for future falls [23].

A syncopal mechanism may underlie an unexplained fall [1]. Our results emphasise the importance of fracture risk evaluation in individuals that have been hospitalised for syncope or orthostatic hypotension, since substantial clinical consequences result from overlooked and missed diagnoses of syncope, that could otherwise have been addressed by adequate management of the underlying syncopal cause. Systolic blood pressure tends to increase over time but decreases in later life [24]. One of the underlying explanations may be arterial stiffness. Arterial stiffness is independently associated with orthostatic hypotension [25]. An increase in peripheral vascular resistance at rest is not routinely observed in healthy older persons, but is often associated with increased stiffness of central elastic arteries, as hallmarks of ageing effects on the vasculature [26]. Impaired stretching of the carotid arterial wall close to the baroreceptor due to arterial stiffness and superimposed atherosclerosis may lead to impaired baroreceptor function in response to change of body position, thus being a potential marker of orthostatic reactions [26].

Our study results indicate that syncope and orthostatic hypotension are two factors that may be overlooked during fracture prevention. During the last decade, falls clinics have been implemented on a small scale [27]. Falls clinics and syncope units constitute an important tool in the management of unexplained falls, syncope, and orthostatic hypotension [1]. Although large randomised studies are currently lacking, our findings support the importance of implementing multidisciplinary teams with falls and syncope specialists. In addition, general preventive efforts against fracture and fall should be undertaken, for example, restrictive use of medications known to increase fall risk (e.g. anti-hypertensive and psychotropic drugs), implementation of home hazard reduction (e.g. rails, nonslip mats, the use of shoes with nonslip soles), and tailored exercise programmes improving mobility and quality of life [19, 20, 28‐31]. Lifestyle changes with proper nutrition are also essential, since nutritional status is an independent predictor of falls in older people living in the community [32].

The observed increased risk of incident fractures in individuals with orthostatic hypotension compared with syncope in the present study may result from possible concomitant diseases such as Parkinson’s disease or dementia, which are well-established factors associated with postural instability and gait dysfunction that may lead to an increased risk for falls and injuries [33‐35]. Although such investigations are beyond the scope of this paper, future research warrants assessment of the level and potential role of dementia in patients with concomitant orthostatic hypotension or unexplained syncope who are at high risk of incident fractures.