Osteoporosis, diabetes, and hypertension are major risk factors for mortality in older adults: an intermediate report on a prospective survey of 1467 community-dwelling elderly healthy pensioners in Switzerland

In developed countries, the number of seniors over 60 years of age has grown over the last century. In Switzerland, the proportion of adolescents in relation to the total population fell from 40.7% in 1900 to 20.2% in 2014 and was accompanied by a rise in the proportion of pensioners (aged over 64 years) from 5.8% (1900) to 17.8% (2014). This rate is expected to increase to 26% by 2045 [1]. In addition, the generation over 64 years of age has become wealthier in recent decades, and an estimated 80% of homeowners are above 65 years of age [2].

The aging generation is in better health than in the past. The Swiss Federal Statistical Office maintains mortality statistics; these data include the prevalence of multimorbid terminal phases, among which one nosological entity is assigned as the cause of death [1]. Currently, little is known about the risk factors for mortality in the older generation. The best evidence is from the Cardiovascular Health Study, which was published in 1998 [3]. That study demonstrated that arteriosclerosis (i.e., aortic stenosis and stenosis of the internal carotid artery) was a principal risk factor for mortality, accounting for a two- to fivefold excess mortality risk. The report found that systolic hypertension (over 169 mmHg, a 2.4-fold excess mortality risk), use of diuretics, renal failure, and diabetes with fasting glucose > 7.2 mmol/L (approximately twofold) were less important factors. That study also identified both age and smoking as minor risk factors. Cancer was not mentioned as a mortality factor in that cohort [3]. In addition, that investigation did not include osteoporosis, which is generally considered an important factor in morbidity and mortality in older adults [4, 5]. One report found a slightly higher mortality risk for osteoporosis than the expected general mortality risk. However, that study did not compare the risk of mortality from osteoporosis with other risk factors with respect to general health status [6]. Other studies have reported slightly higher (between 1.2 and 1.5) relative risks of mortality associated with an osteoporosis diagnosis [7‐9], with the exception of one study from Japan that actually showed a higher mortality risk [10].

Studies investigating osteoporosis typically focus on fractures and the subsequent diagnosis of osteoporosis. Data on clinically asymptomatic individuals diagnosed with osteoporosis are scarce, although some studies are cited above.

Today, osteoporosis is known to be a disease of the elderly. However, the epidemiological frequency is underestimated, and its prevalence in some places, such as in Switzerland, is unknown. Another example of this underestimation is the treatment gap. In Switzerland, estimates suggest that between 36 and 58% of diagnosed cases of osteoporosis are not treated as recommended by the guidelines [11]. Epidemiological data from Sweden estimate that osteoporosis affects approximately 7.8% of men and 27.9% of women among citizens aged 70–74 years [12]. In Europe, the prevalence of osteoporosis is estimated to be between 5.9 and 7.2% in men and 19.1 and 23.5% in women [12].

One of the aims of the present prospective cohort study was to evaluate general health-impairing conditions in subjectively healthy retired people in Switzerland. This study also aimed to explore under-evaluated potential risk factors for mortality in this age group and to compare the magnitudes of their effects.

This report describes mortality and associated risk factors in an older Swiss population. Beyond the effects of these risk factors on mortality, a central concern is the time lag between their identification and the occurrence of death. As of March 2017, 1401 study participants survived, and 66 had died. Fortunately, that number is relatively low; however, these data allow us to employ health risk management techniques when assessing the potential causes of death for target prevention.

In addition to T2DM, osteoporosis was a strong predictor of mortality in our senior cohort. This finding was evident in the bivariate analysis (Table 3) and our survival analysis. In our multivariate model using Cox regression models with covariates (Table 4), osteoporosis had a higher relative risk of mortality than type 2 diabetes. The relative risk of osteoporosis seems to be high. Based on the absolute numbers, we assumed that osteoporosis in our cohort was responsible for approximately 6% of deaths; hypertension was responsible for 26% of deaths and diabetes was responsible for 9% of all fatalities. Although comparisons with other mortality risk factors in a cohort of healthy senior citizens are lacking, osteoporosis has been identified in several studies as an important risk factor for all-cause mortality. Our data confirmed the results of a Danish cohort study, which found a shorter life expectancy among osteoporosis patients. One conclusion of the present study was significant comorbidity in such patients with a higher mortality risk [14]. Additional evidence for osteoporosis as an important all-cause mortality risk factor comes from studies in the Netherlands [15], United Kingdom [17], and Austria [18]; however, those studies were conducted retrospectively. In addition, the relatively low prevalence of osteoporosis in our cohort compared to that in the general Swedish population in the same age group (osteoporosis prevalence rates of 7.8% in men and 27.9% in women) would predict a much higher death rate and most likely a higher public health relevance than was found in our study. Based on our results, we propose that osteoporosis is relevantly underreported in the study population, which may lead to an underestimation of its public health relevance compared to diseases such as hypertension.

Why might osteoporosis lead to a higher mortality rate? The mechanism is complex and multifactorial, and the precise pathways are not known. As shown by Sakem and coworkers, a low vitamin D level is involved in an altered immune response [13]. However, we found no direct association between vitamin D and all-cause mortality and osteoporosis; specifically, 3.8% of osteoporosis patients had high vitamin D levels, whereas 15.8% of patients without osteoporosis had a vitamin D level of less than 13 ng/mL (Fig. 4). In contrast to the results for 25-hydroxyvitamin D3 levels, 20.7% of our participants with low BMIs (defined as less than 21 kg/m²) were in the osteoporotic group compared with only 9.8% in the non-osteoporotic population. This result indicates that an impaired nutritional state could be reflected by osteoporosis. This finding supports the results of a US study by Thomas-John et al., which reported that osteoporosis was associated with weight loss and low consumption of dairy products, suggesting that malnutrition might be an important cofactor [19]. In addition, an Australian longitudinal study observed a clinically relevant reduction in mortality among older seniors diagnosed with osteoporosis without fractures undergoing bisphosphonate treatment [20].

The most prevalent contemporary causes of death are cardiovascular and cerebrovascular disorders and cancer. Recent decades have seen rapid advances in diagnoses and surgical and pharmaceutical therapies for these conditions. Comparative studies of mortality in Europe have shown that the all-cause mortality per 1000 people aged 70–74 years is 25.1 for women and 47.1 for men [21]. The all-cause mortality rates in Alaska and Florida were estimated to be 43.5 and 44.2, respectively. In our cohort (mean age, 72.1 years; 95% CI, 71.7–72.5), the overall mortality was 12.2 per 1000. In the present study of an apparently healthy Swiss population of pensioners, we found that T2DM, osteoporosis, and hypertension were important risk factors for mortality.

Unsurprisingly, hypertension and T2DM are leading causes of mortality, as shown by the Cardiovascular Health Study [3]. The prevalence of T2DM is greatest in people over 65 years of age, and the highest incidence also occurs in that age group [22‐24]. Some authors believe that the burden of diabetes in the older generation will increase further due to demographic changes, the rising incidence of T2DM in all groups [22], and the longer survival times of people diagnosed with diabetes at a younger age [25]. However, older people are under-represented in clinical trials; thus, they are often treated according to guidelines based on expert opinions and extrapolation of results obtained from trials on younger people [26]. People diagnosed with T2DM at an older age account for approximately half of all new cases. Consequently, diabetes and mortality data in older populations have produced heterogenic results and have led to a mortality risk estimation with an OR of 3.0 [27] but only a slight, non-significant risk elevation [28]. One systematic review demonstrated a pooled OR of 1.4 for mortality and T2DM in an elderly population over 65 years of age [29]. Our data from this population-based prospective cohort indicate that T2DM is a major risk factor for death, even for the older generation. Diabetes appears to be a leading cause of arteriosclerosis and death, even with the relatively short follow-up period of 3.7 years used by this study.

In addition to diabetes, hypertension is an independent predictor of mortality in elderly populations. This result is not surprising and has been convincingly demonstrated by multiple studies [30‐32]. The magnitude of its effect is associated with the degree of hypertension [26]. Weidung et al. showed that mortality due to hypertension increased in older adults who were active and robust [27]. We assumed that the individuals in our cohort were robust, since being subjectively healthy was a requirement for inclusion in the study. Huynh et al. found that hypertension was less of a risk factor for all-cause mortality than T2DM. In their models obtained from a cohort of older adults, T2DM was responsible for 50% of cardiovascular mortality, although it may have overlapped with the effect of hypertension [33]. A similar magnitude of mortality risk was reported in the multivariate models of Fried et al. in the Cardiovascular Health Study [3].

We did not find that a history of cancer was a predictor of death. That result was expected, since the observational time was too short to demonstrate an effect of cancer on all-cause mortality. In this context, comorbidities other than cancer play more important roles in all-cause mortality in older patients diagnosed with cancer, at least in cases with a relatively short surveillance period [34], such as the present interim report.

One strength of the present study is its prospective design and follow-up of 3.7 years in subjectively healthy seniors. A weakness is the lack of osteodensitometry measurements to define the disease. Additionally, our data are based on self-reporting and may have underestimated conditions. On the other hand, in the absence of other representative samples, this study provides the best available estimates, as stated elsewhere [35]. Thus, our results may have overestimated the association between osteoporosis and mortality in this older population as a result of recall bias. Osteoporosis and malnutrition may be easily treated, but they have large impacts on public health. We believe that osteoporosis in the elderly may be under-diagnosed as a result of reduced physician awareness of the condition. Moreover, our population showed a smaller prevalence of osteoporosis than the reported prevalence in Sweden (women, CH: 6.4%; women SWE: 27.9%). This finding suggests underreporting in our sample. Diagnosing osteoporosis should be undertaken more proactively in older people with subjectively good health.

We demonstrated that hypertension and T2DM followed by osteoporosis were major factors of all-cause mortality in subjectively healthy seniors. Osteoporosis is probably underestimated due to underreporting and underestimation by clinicians. Osteoporosis should be more actively diagnosed in healthy older people before they experience osteoporosis-associated incidents.