Ageing and the epidemiology of multimorbidity

Mortality and disability-adjusted life-years

Two important metrics in determining the burden of disease are disease-related mortality and disease-related disability-adjusted life-years (DALYs). While death is a firm end-point, attribution of cause of death is sometimes difficult, particularly in the elderly or among those with multiple diseases. Most countries have systems in place where the underlying cause of death (of which there can only be one) is determined based on algorithms that look at all data contained on the death certificate and further review by nosologists. DALYs are an additional measure of disease burden that incorporates both premature death from a disease in addition to years of disability spent living with that disease (table 1).

The number of deaths, globally, increased from 46.5 million in 1990 to 52.8 million deaths in 2010, although the overall mortality rates were trending downward during this period. Ischaemic heart disease, stroke, COPD, lower respiratory infections and lung cancer were the five leading causes of death in 2010 (fig. 4a) [6]. Combined, pulmonary diseases accounted for 7.3 million deaths or 14% of all deaths worldwide. COPD climbed to the third place among the leading causes of death worldwide by 2010. In absolute numbers, COPD was responsible for a total of 3.1 million deaths (95% uncertainty interval (UI) 2.9–3.4 million) in 1990 and 2.9 million (95% UI 2.7–3.3 million) in 2010. Lower respiratory tract infections accounted for 2.8 million deaths in 2010, and lung cancer increased from 1.0 million deaths in 1990 to 1.5 million deaths in 2010 and became the fifth cause of death worldwide (fig. 4b).

DALYs of leading diseases are noted in figure 5 for three different age groups. In both of the older age groups (50–69 years old and ≥70 years old), COPD is the third leading cause of DALYs.

Another important pulmonary disease affecting older adults is ILD. Although the absolute number of deaths for this condition does not parallel that described above for COPD, there has been a 77% increase in the number of deaths from 65 000 in 1990 to 115 000 in 2010 [6]. Deaths related to ILD are probably greatly underestimated, particularly in parts of the world where modern imaging technology, such as computed tomography scans, are not routinely available. As this technology becomes more available (and with the recent data demonstrating the efficacy of computed tomography scans for early detection of lung cancer [12, 13]), the burden of these diseases will probably increase.

Prevalence

The estimated prevalence of COPD in 28 countries among adults aged >40 years is 9–10% [14]. There are important regional differences in the estimated prevalence, due in part to methodological differences in case definition and classification using pre-bronchodilator spirometric values, the latter being important as the pre-bronchodilator spirometry could overestimate the presence of airway obstruction by up to 50% in some subgroups [15]. The best estimates of prevalence are provided by the Latin American Project for the Investigation of Obstructive Lung Disease (PLATINO) (five countries in Latin America) and Burden of Obstructive Lung disease (BOLD) study (12 countries) also demonstrated a significant variation in prevalence ranging from 5–20% [16] (fig. 6). Both studies are comparable as they used the same case definition and standardised post-bronchodilator spirometric values to estimate COPD prevalence. The large variation by region and country suggest differences in susceptibility and exposures.

• Download figure
• Open in new tab
• Download powerpoint

Figure 6–

Chronic obstructive pulmonary disease prevalence by region. Data from [17, 18].

In figure 7, we arranged the prevalence in the 17 regions that participated in the PLATINO and BOLD studies, and compared prevalence of smoking, indoor and outdoor air pollution, and tuberculosis in these same countries. This comparison suggests that although cigarette smoking is the main risk factor for developing the disease, other environmental factors, exposures, and the cumulative effect of socioeconomic status and geographical latitude may influence disease progression, and therefore prevalence. This variability in the specific risk factors for COPD in different parts of the world and in different age groups may also be related to the multimorbidities that are found in different populations.

• Download figure
• Open in new tab
• Download powerpoint

Figure 7–

Global prevalence of chronic obstructive pulmonary disease (COPD) and risk factors. PM₁₀: particulate matter <10 μm in diameter; TB: tuberculosis. Data from [17, 18].

ILDs, including idiopathic pulmonary fibrosis (IPF), are less common than COPD but appear to be increasing in the population. The overall prevalence of IPF ranges from 0.5 to 27.9 per 100 000 population [19]. IPF represents about 20% of the burden of ILDs, so the total burden of ILD may be much higher [20].

Functional effects relevant to ageing

While financial burden, mortality and acute events requiring emergency treatment or hospital admissions (e.g. COPD exacerbations) are the usual metrics used to describe disease impact, the complex effects of chronic diseases in the elderly population, both from the individual and the societal perspectives, are insufficiently captured by these descriptors. Additional relevant measures, such as the need for informal caregivers (subtracting their participation in the regular workforce), home healthcare, nursing homecare and years lived with disability (or loss of disability-free years), are becoming more important to identify areas of intervention and prevention. Research efforts in this area are still limited, but growing. For example, COPD is associated with federal disability benefits in the USA, with an association strong or stronger than other chronic diseases such as cardiovascular disease [21], and subjects with COPD have significant loss of disability-free years, after adjusting for the effects of other chronic diseases [22]. Earlier steps in the development of disability (functional limitation to and dependency for activities of daily living) are also adversely impacted by COPD [23–25]. A growing understanding of these effects of COPD [26] will help to identify trajectories of decline important for future interventions.

Impact of ageing

The physiological changes of ageing and the impact of multimorbidity are discussed elsewhere in this article. Ageing is a heterogeneous process, with individuals ageing differently depending on their genetic background, environmental exposures and other factors. A component of ageing is the imbalance between inflammatory and anti-inflammatory networks in individuals resulting in a low-grade chronic inflammatory state referred to as “inflammaging” [27].

This chronic inflammatory state is also part of the development and progression of COPD, which is increasingly recognised as a systemic illness [28] Multiple diseases can coexist in an individual for a number of reasons, including: random chance (where two diseases are common in a population); two diseases that are part of the same continuum; two diseases that share a common risk factor; two diseases where one disease causes the second disease; and scenarios where the presence of one disease increases the risk of a second disease [29]. Ageing and multimorbidity contribute to frailty, which, in turn, confers a higher risk of a number of complications and poor outcomes such as falls, disability, hospitalisation and mortality [30, 31].

Multimorbidity increases in the elderly as was demonstrated in the study by Barnett et al. [31] that included 1.7 million patients in Scotland, UK, where 30.4% of adults aged 45 to 64 years reported at least two chronic conditions, increasing to 64.9% of adults aged 65 to 84 years and more than 80% for those above 85 years old (fig. 8). Similar findings were described in the USA Medicare enrolee population [32, 33]. These results are not unexpected, in that the longer one lives, the more likely one is to be diagnosed with any chronic disease. While many chronic diseases are ultimately lethal, there may be a long time before death occurs. Thus, increasing longevity with diseases that are not immediately lethal, such as hypertension, can be expected to result in multimorbidity in ageing populations.

• Download figure
• Open in new tab
• Download powerpoint

Figure 8–

Number of chronic comorbidities by age stratum. The number of comorbidities increases with age and is larger in individuals 65 years and older. Reproduced from [28] with permission from the publisher.

Ageing can lead to the appearance of chronic diseases in different ways. In some instances, diseases become more common in the older population, with malignancies such as lung cancer or prostate cancer being examples. In other instances, the disease may have been present for some time but becomes more apparent with ageing because of the general senescence in multiple organ systems. Examples of this include arthritis and its complications. Often, there is an overlap of these processes, which is true for chronic respiratory disease. COPD is also a disease of ageing, with prevalence doubling for each decade of life above the age of 40 years [17]. Some authorities suggest that pulmonary fibrosis may also be related to injury to the ageing lung [34].

Another interesting measure of ageing and chronic lung diseases is the changing referral patterns for lung transplantation in the USA. In the early era of lung transplantation, there was an age limit of ≤60 years, but in recent years the age limit has been increased. The International Society of Heart and Lung Transplantation reports an increase in the median age of lung transplant recipients from 45 years in 1985 to 55 years in 2012 [35]. During 2000–2005, 3% of recipients were ≥65 years and 0.3% were ≥70 years. By 2006–2012, this proportion had increased to 10% and 3%, respectively. The increased age of recipients occurred mainly in the diagnostic groups of ILDs and COPD excluding α₁-antitrypsin deficiency [35]. Interestingly, the presence of multiple chronic diseases is a contraindication for lung transplant.

Approaches to multimorbidity

Traditionally, researchers have focused on a single disease or disease pairs, but recent evidence supports departing from this reductionist approach towards more integrative studies assessing the relationship and impact of multimorbidities. This approach could reveal disease clusters and novel pathobiology mechanisms, and lead toward a better understanding and, possibly, integrated co-management [36–39].

Using the mathematics of network theory [40], it has been demonstrated that disease coexistence is not the result of simple chance. For chronic NCDs, co-occurrence of diseases that are the result of genetic predisposition in combination with the cumulative effect of generic responses to biological stress, evoked either by the principal mutation and/or environmental exposures [40]. These generic responses constitute the so-called intermediate pathophenotypes and include: inflammation; thrombosis and haemorrhage; fibrosis; the immune response; proliferation; and apoptosis/necrosis. Given that these intermediate pathophenotypes constitute the entire range of human response to injury or infection, most chronic diseases can be linked back to these pathways. Hidalgo et al. [41] generated a visual map representing disease association in the US Medicare claim data (age ≥65 years), as shown in figure 9.

• Download figure
• Open in new tab
• Download powerpoint

Figure 9–

US Medicare claims disease network map. Nodes are diseases classified according to International Classification of Diseases (ICD)9 codes at the five-digit level. Link weight corresponds to the association between diseases. The blue area represents the respiratory disease neighbourhood and the grey area represents the interlinked comorbidities neighbourhood. Reproduced and modified from [41].

This map is two dimensional but attempts to portray multidimensional relationships. For example, “chronic airway obstruction” appears in the cluster of respiratory diseases in the upper left hand corner of the map and is surrounded by other chronic respiratory diseases but is also linked to congestive heart failure, which has been well described [42]. Other data, however, have shown links between COPD and lung cancer [43] and diabetes [44], yet this map does not show a direct connection between these diseases.

Patients with COPD have, on average, four to six chronic comorbidities, compared with two to three comorbidities in age-matched controls [45–47]. Considering COPD as the index disease, prevalent comorbidities include cardiovascular, metabolic, musculoskeletal, cancer and psychiatric conditions, which are presented in detail in table 2 using data from multiple studies. Variability in comorbidity prevalence exists and probably represents the differences in disease ascertainment in each study, as they used administrative databases, patient self-reported disease, and data extracted from medical records or measured with specific cut-offs [42, 44, 46, 48, 49]. Nevertheless, compared with aged-matched subjects without COPD, the prevalence of such comorbidities are 1.5–2 times higher (table 2).

For IPF, there are no large-scale epidemiological studies looking at multimorbidity, although one small study showed higher mortality when comorbid cardiovascular disease was present [50] and another showed a high prevalence of comorbid depression [51]. As noted earlier, the prevalence estimates for ILDs vary greatly [20, 52–56] and are much lower than that of COPD. The wide range in these numbers is probably explained by the previous lack of a uniform definition used in identifying cases of IPF, as well as by differences in study designs and populations. 80% of patients with IPF die from respiratory failure, although heart failure, bronchogenic carcinoma, ischaemic heart disease, infection and pulmonary embolism are also common, suggesting that comorbidity may be associated with and play a significant role in the disease [57].