Inflammation in multimorbidity and disability: An integrative review

Abstract

Multimorbidity is a robust predictor of disability in aging adults, but the mechanisms by which multimorbidity is disabling are not clear. Most existing research focuses on disease-specific phenomena, such as diminished lung capacity in COPD, which can result in functional limitations. This review takes a different approach by highlighting the potential role of a biological process - inflammation - that is common to many chronic medical conditions and thus, from a medical perspective, relatively disease non-specific. Beginning with a description of inflammation and its measurement, this paper will provide an overview of research on inflammation as a predictor of disease risk in healthy adults and of adverse outcomes (e.g., disability) in those with multimorbidity. The discussion of inflammation is then situated in the context of biopsychosocial influences on health, as inflammation has been shown to be sensitive to a wide range of social and psychological processes that are thought to contribute to healthy aging, including successful adaptation to multimorbidity and reduced risk of disability. Finally, implications of this broader perspective for interventions to improve outcomes in aging adults with multimorbidity are briefly considered.

INTRODUCTION

The focus of this special issue of Health Psychology - multimorbidity - presents the opportunity to view age-related chronic disease through a different lens than the one typically applied. Specifically, medical care is organized around the diagnosis and treatment of individual chronic diseases, an endeavor that rests on the ever more precise identification of pathological processes that are unique to each disease. The current focus on person-centered medicine based on genomic information (Joyner & Paneth, 2015) is an example of the search for more precise and individualized treatments. The disease-focused biomedical approach has been remarkably successful in reducing disease-related morbidity and mortality. Multimorbidity, in contrast, highlights the extent to which diverse diseases have characteristics in common, characteristics that by definition are disease non-specific. Most obviously, chronic diseases tend to cluster in aging adults; they also cluster in populations defined by various forms of adversity, most notably low socioeconomic status. Such clustering demands a focus not only on what makes chronic diseases and their treatments unique, but also on what makes them similar. More to the point, it invites the identification of processes, biological or otherwise, that are common to most chronic diseases that explain the fact that they tend to cluster in specific populations. Geroscience, for example, focuses on fundamental processes related to aging itself that ultimately give rise to an array of age-related diseases (Franceschi & Campisi, 2014; Sierra & Kohanski, 2017). One such biological process that has received a good deal of attention is inflammation.

The aim of this paper is to review the evidence that inflammation contributes to multimorbidity and to adverse outcomes related to multimorbidity, particularly functional limitations and disability. We begin with a consideration of the definitions of multimorbidity and disability as they apply in our work. We then present an overview of inflammation: its components, its role in host defense, and its links to age-related disease. This is followed by a review of the relatively small literature examining the role of inflammation in mediating the relationship between multimorbidity and functional impairments. We next embed the examination of inflammation in the broader context of psychological and social factors. Finally, we consider potential avenues for future research efforts.

Measurement considerations

Multimorbidity has been construed and assessed in diverse ways (see chapter by Suls, Green, & Boyd, 2019 in this issue). For the purposes of this paper, we define multimorbidity as the coexistence of two or more chronic medical conditions within the same individual. In our work we use self-reported data on chronic conditions; in some instances participants are asked specifically about physician diagnoses, in others whether they have experienced or been treated for a specific set of conditions. Epidemiological studies have shown individuals tend to be accurate in their reports of their own chronic conditions (Kriegsman, Penninx, van Eijk, Boeke, & Deeg, 1996), although there tend to be lower rates of multimorbidity in data from community samples compared with clinical and administrative sources (Fortin, Hudon, Haggerty, Akker, & Almirall, 2010). As reviewed elsewhere in this issue, there remains no gold standard definition; the number and types of conditions that are included in measures of multimorbidity can vary dramatically from one study to the next. In our work we have selected conditions that are most likely to result in adverse health outcomes and reduced quality of life (Institute of Medicine, 2012). Finally, our research to date has employed a measure based on numbers of chronic conditions. Future work will examine multimorbidity formulations that account for differences in severity among conditions and specific clusters of conditions.

Disability is defined in terms of both the individual and the environment. Individuals may have functional limitations, for example, that put them at risk for disability. Functional limitations are often assessed using objective assessments of functional capacity, such as grip strength, time taken to rise from a chair after sitting for an extended period, or time taken to walk a standardized distance. These measures may not constitute disability (i.e., if the individual is able to function independently), but performance below normative levels may represent risk factors for disability. In contrast, measures of disability focus on the extent to which an individual is unable to perform such tasks and/or requires assistance in performing them. In our work we have generally employed self-reported assessments of functional limitations due to poor health, a measure based on the widely used SF-36 (Syddall, Martin, Harwood, Cooper, & Sayer, 2009). A related concept is frailty, which is reduced ability to respond to environmental challenge (Fried et al., 2001). While frailty may increase risk of disability, it is not in itself a measure of disability (Fried, Ferrucci, Darer, Williamson, & Anderson, 2004).

The literature reviewed in the rest of this paper uses diverse measures of multimorbidity and disability, often a reflection of the attempt to balance expense and participant burden on the one hand with measurement quality on the other. Each provides important perspectives on the links among multimorbidity, disability, and inflammation, but each also has important limitations, and measures of the same construct are not always interchangeable.

Inflammation and links to multimorbidity

From an immunological perspective, inflammation describes a complex process involving the carefully coordinated activities of multiple cell types and chemical messengers. As reviewed recently by Netea et al (2017), the inflammatory response is typically triggered by tissue damage from injury or infection - stimulatory signals can be from invading pathogens or endogenous stressors - and is designed to repair the damage and restore homeostasis. In response to encountering these signals, immune cells, typically monocytes, macrophages, or dendritic cells, release pro-inflammatory cytokines, such as interleukin-1 (IL-1), tumor necrosis factor-α (TNF-α), and interleukin-6 (IL-6) locally and into the blood resulting in the recruitment of other immune cells to the site of tissue damage, changes to the local environment (e.g., increased permeability of blood vessels to improve access for immune cells), and diverse effects on other physiological systems and behavior (e.g., fever, fatigue, anorexia). The inflammatory response is typically short-lived, lasting only as long as required to clear the pathogen and/or repair the damaged tissue. Resolution of the inflammatory response involves both the withdrawal of pro-inflammatory stimulation (e.g., through the removal of the pathogen) and the actions of anti-inflammatory mechanisms, including anti-inflammatory cytokines, such as IL-10, and the anti-inflammatory influence of neuroendocrine systems like the hypothalamic-pituitary-adrenal axis that become activated during the inflammatory response.

Inflammation is a critical component of host defense. Experimental animals in which the inflammation response is suppressed or genetically inactivated are unable to survive (Goldsby, Kindt, & Osborne, 2006). Nonetheless, inflammation can result in pathology when it becomes dysregulated in the form of exaggerated responses to stimuli and/or loss of sensitivity to the actions of anti-inflammatory mechanisms. A large variety of conditions arise from dysregulation of the inflammatory response, including sepsis, acute intestinal inflammation, rheumatoid arthritis, diabetes, and a number of cancers (Netea et al., 2017). Inflammation has also been implicated in the development of Alzheimer’s Disease and other dementias (Fotuhi, Hachinski, & Whitehouse, 2009). Although the specific inflammatory components vary among disease conditions, a common feature is exaggerated or otherwise inappropriate production of inflammatory mediators (Netea et al., 2017).

A growing epidemiological literature involving large-scale studies of non-clinical populations shows that inflammation is a risk factor for an array of chronic illnesses. These studies typically define inflammation in terms of concentrations of specific inflammatory factors in the blood, detection of which has been significantly advanced by the development of high-sensitivity assays and less invasive sample collection methods, such as blood spots (McDade, Williams, & Snodgrass, 2007). The bulk of the research centers on cardiovascular disease. A number of recent population-based studies from diverse national cohorts have shown that circulating levels of IL-6 and CRP, among other inflammatory markers, are robust predictors of future cardiac events, including incident heart disease and cardiac mortality, in otherwise healthy adults (Danesh et al., 2008; Kaptoge et al., 2014). Most of these studies involved long-term follow-up periods of 10–20 years, and they found that the predictive associations for inflammation were as robust as those for established cardiovascular disease risk factors. These results are bolstered by accompanying systematic reviews and meta-analyses that show similar robust associations between blood levels of inflammatory proteins and risk of incident cardiovascular disease or mortality (Danesh et al., 2008; Kaptoge et al., 2014). Even among long-lived adults from the Cardiovascular Health Study All Stars Study (mean age 85 years), circulating levels of IL-6 and CRP were significant predictors of subsequent risk of cardiac events and mortality over and above associations with diverse traditional cardiac risk factors (Jenny et al., 2012). The impact of inflammation on cardiovascular disease is not limited to healthy adults. Higher levels of inflammatory markers like IL-6 and CRP predict greater risk of adverse cardiac events in those with existing cardiovascular disease (Blake & Ridker, 2003). Data from the Women’s Health and Aging Study showed that in patients with existing heart disease, high levels of IL-6 were associated with a 4-fold increase in risk of subsequent mortality compared to patients with lower IL-6 levels (Volpato et al., 2001). These latter results underscore the role of inflammation in both incidence and progression of cardiovascular disease.

Inflammation may be linked to cardiovascular events and mortality due to its role in the development and progression of atherosclerotic plaques on blood vessel walls. The earliest stages of atherogenesis involve the development of lesions containing cells that produce inflammatory factors, such as macrophages, and the growth of atherosclerotic plaques is considered a chronic inflammatory process (Pant et al., 2014). Cascading events that ultimately result in plaque rupture and ischemia are also linked to inflammatory factors acting locally within plaques and on endothelial cells (Pant et al., 2014). Atherosclerosis is linked not only to activation of cells that produce inflammatory factors, but also to the failure of mechanisms designed to resolve the inflammatory response, including anti-inflammatory cytokines (e.g., IL-10) and endocrine systems, such as the hypothalamic-pituitary-adrenal axis (Viola & Soehnlein, 2015). Atherosclerosis and consequent risk of cardiac events is thus a system failure, resulting in a shift in multiple biological elements toward a pro-inflammatory state (Viola & Soehnlein, 2015).

Inflammation is also associated with disorders of metabolic regulation, most notably diabetes. Large, prospective, population-based studies, such as the Multi-Ethnic Study of Atherosclerosis (MESA; (Bertoni et al., 2010)) and the Nurse’s Health Study (Hu, Meigs, Li, Rifai, & Manson, 2004) along with studies of community samples from other countries, including Greece (Koloverou et al., 2018) have shown that circulating levels of inflammatory proteins (e.g., TNF-α, IL-6, and CRP) are associated with markedly increased risk of incident Type 2 Diabetes Mellitus (T2DM) over follow-up periods as long as 10 years. A recent meta-analysis based on a set of 20 case-control studies concluded that IL-6 levels are also elevated in patients with Type I diabetes compared to case-matched controls (Chen et al., 2017), suggesting broad effects of inflammatory factors on the regulation of blood glucose. Inflammation is argued to be a component of pathophysiological processes that result in loss of insulin sensitivity (Calle & Fernandez, 2012; Fernandez-Real & Ricart, 2003), a potential mechanistic explanation of the links between inflammation and T2DM. Obesity is a robust risk factor for T2DM in particular (Nathan, 2015), and obesity is strongly linked to circulating levels of inflammatory factors, most likely produced by activated mononuclear cells resident in adipose tissue (Calle & Fernandez, 2012; Fernandez-Real & Ricart, 2003). Like atherosclerosis, therefore, inflammation may play a role in the development and progression of the loss of sensitivity to insulin in those at risk for diabetes; studies aimed at determining the most effective strategies for restoring optimal levels of inflammatory proteins (e.g., exercise, diet) are ongoing (Calle & Fernandez, 2012).

Cancer is another set of conditions where inflammation is thought to play a role (Aggarwal & Gehlot, 2009; Caruso, Lio, Cavallone, & Franceschi, 2004; Li, Withoff, & Verma, 2005; Naugler & Karin, 2008), and IL-6 has been a particular focus of interest. Infectious agents, such as Hepatitis viruses, Epstein Barr Virus, Human Papilloma Virus, and Helicobacter pylori are estimated to contribute to 15% of cancers worldwide (Kuper, Adami, & Trichopoulos, 2000). A common feature of infections and resulting inflammation is activation of the nuclear transcription factor NF-□B (Naugler & Karin, 2008), and NF-□B has itself been linked to cancer pathology (Karin, 2006). As NF-□B stimulates IL-6 production (Brasier, 2010), IL-6 is one hypothesized mechanism linking exposure to infectious disease and resulting NF-□B activation with cancer incidence and progression (Naugler & Karin, 2008).

Inflammation and multimorbidity

Fewer studies to date have examined links between inflammation and multimorbidity, although such links are predicted by the concurrent and prospective associations between inflammation and single conditions, as detailed above. Moreover, age-related increases in systemic levels of inflammation - referred to as “inflammaging” (Franceschi & Campisi, 2014) - are thought to contribute to diverse age-related disease conditions, including multimorbidity (Barnes, 2015). Cross-sectional data from the Mid-life in the United States (MIDUS) study showed that circulating levels of IL-6 and CRP increase linearly with increasing numbers of chronic medical conditions (Friedman, Christ, & Mroczek, 2015). Data from a study of older adults living in the Chianti region of Italy (InCHIANTI) also showed that levels of IL-6 were higher in older adults with existing multimorbidity (Fabbri et al., 2015). Importantly, the same study included several follow-up assessments over a 9-year period, and the results of longitudinal analyses showed that the rate of increase in chronic conditions was steeper in those with higher baseline levels of IL-6 and in those with faster increases in IL-6 over time (Fabbri et al., 2015).

Multimorbidity, disability, and quality of life

Disability rates among older adults in the United States and other Western countries have declined substantially in the past few decades (Manton, 2008; Schoeni, Freedman, & Martin, 2008). Whether these declines will be sustained is less clear; current high rates of obesity among middle-aged adults, for example, may result in higher disability rates once they approach old age (Manton, 2008). Nevertheless, loss of functional capacity and independence remain chief concerns for aging adults. Not surprisingly, having more than one chronic condition increases the risk of functional impairment compared to having a single condition, but these associations are often non-linear. Risk of functional impairment can increase exponentially with increasing numbers of conditions (Wolff, Boult, Boyd, & Anderson, 2005). Moreover, as noted below, some combinations of conditions are more likely than others to result in impairment. Finally, different types and combinations of conditions can result in different kinds of impairments.

The broadest perspectives on the impact of multimorbidity involve assessments of health-related quality of life. One widely-used instrument is the EuroQol-5 Dimensions (EQ-5D) scale (EuroQol Group, 1990), items for which cover 5 health domains: mobility, self-care, usual activities, pain or discomfort, and anxiety or depression. Multiple, large-scale studies from diverse countries and cultures (e.g. United States, United Kingdom, Canada, Korea, Germany), show that single and multiple chronic conditions are cross-sectionally associated with lower scores on the EQ-5D (Agborsangaya, Lau, Lahtinen, Cooke, & Johnson, 2013; Brettschneider et al., 2013; Chin, Lee, & Lee, 2014; Mujica-Mota et al., 2015) or similar quality of life assessments, such as the NIH Patient-Reported Outcomes Measurement Information System (PROMIS) measures (Rothrock et al., 2010). Importantly, most of these studies showed that the presence of multiple conditions was associated with greater decrements in quality of life than what would have been expected from adding up the decrements associated with each individual condition, suggesting adverse synergistic interactions among conditions in adults with multimorbidity. In general neurological disorders (e.g., Parkinson’s Disease, stroke), mental illness, and pain tended to be most robustly linked to lower quality of life scores when examined alone or in combination with other conditions. And quality of life domains related to mobility, self-care, and engaging in “usual activities” - akin to basic and instrumental activities of daily living (ADLs and IADLs, respectively) measures that are widely used for disability - were generally more consistently impaired than other quality of life domains (e.g., participation in usual activities; pain; fatigue) in those with greater numbers of conditions regardless of the specific conditions involved (Agborsangaya et al., 2013; Brettschneider et al., 2013; Chin et al., 2014; Hunger et al., 2011; Rothrock et al., 2010).

A large literature focuses more specifically on assessments of functional capacity and ability to function independently; these are measures of functional limitations and disability. As discussed earlier, some measures are objective (e.g., grip strength, timed walk, chair stands) and may or may not indicate impaired ability to live independently, while others (e.g., ADLs, IADLs) directly assess the individual’s capacity to meet the demands of daily life.

Cross-sectional associations of multimorbidity and diverse measures of disability have been examined in multiple studies employing population samples from diverse regions, including North America (Griffith et al., 2017; Quinones, Markwardt, & Botoseneanu, 2016), The European Union (Laan et al., 2013; Yokota et al., 2016), and Asia (Wang et al., 2017). All find that greater numbers of conditions are associated with higher levels of functional limitations and disability. Moreover, combinations of chronic conditions that include neurological (e.g., dementia) or psychiatric (e.g., depression) problems are most strongly linked to greater disability, despite relatively low prevalence compared to, for example, hypertension or arthritis.

Stronger evidence that multimorbidity causes disability comes from studies that use longitudinal data with prospective determination of disability. In a study of adults age 65 and over who participated in the Medicare Beneficiary Survey, Wolff and colleagues (2005) found the odds of becoming functionally dependent over a 3-year follow-up period doubled with the onset (at 1 year) of a single chronic condition, quadrupled with two, and was 13 times greater with three or more. Data from the MIDUS study using a count measure of multimorbidity showed that greater numbers of chronic conditions predicted greater loss of mobility function over an 18–20 year follow-up period (Friedman, Mroczek, & Christ, 2018). Sensitivity analyses showed that changes in mobility were not attributable to any single condition but rather to overall multimorbidity burden. Stenholm and colleagues (2015) examined change over time in physical functioning in relation to changes in chronic conditions in 24,000 men and women aged 60–107 from the Health and Retirement Study. Results showed that at all ages having more chronic conditions predicted poorer function, but that the impact of chronic conditions on functional capacity became more pronounced with age (Stenholm et al., 2015). Similar results have been observed in studies using data from the Australian Longitudinal Study on Women’s Health (Jackson et al., 2015), and from the Kungsholmen Project in Sweden (Calderon-Larranaga et al., 2018; Rizzuto, Melis, Angleman, Qiu, & Marengoni, 2017; Vetrano et al., 2018). Results of some of these studies show that increases in physical limitations and disability over time are more likely to be associated with the presence of neurological and psychiatric problems than with other conditions, paralleling the results of cross-sectional work.

Inflammation and disability

Given the robust links between multimorbidity and disability, there is growing interest in underlying mechanisms. While myriad biological processes are likely to contribute to disability, a number of studies have documented associations between circulating levels of inflammatory factors and measures of disability and functional capacity. Cross-sectional analyses of data from 3,500 adults age 65 and older from the PolSenior study showed IL-6 and CRP were significantly associated with poorer physical function, determined by scores on the Katz ADL scale (Puzianowska-Kuznicka et al., 2016). Using cross-sectional data from the National Health and Nutrition Examination Survey (NHANES), Kuo and colleagues (2006) found that higher levels of CRP were associated with significantly greater odds of disability, assessed using multiple measures, including ADLs, IADLs, leisure activities, and general mobility. Higher CRP levels were also associated with less muscle strength, assessed as walking speed and leg power, and reduced muscle strength partially explained the association of CRP and disability (Kuo, Bean, Yen, & Leveille, 2006). Results from the Health Aging and Body Composition study of adults age 70–79 showed baseline associations between high levels of inflammation (IL-6 and TNF-α) and reduced grip and knee extensor strength and appendicular muscle mass (Visser et al., 2002).

In one prospective study of disability-free adults 71 years of age and older from the Established Populations for Epidemiologic Studies of the Elderly, inflammation predicted mobility and ADL disability 4 years later. Those with baseline IL-6 levels in the highest tertile were 62% and 76% more likely to report new ADL and mobility disability, respectively, than those with levels in the lowest tertile (Ferrucci et al., 1999). Nüesch and colleagues (2012) created a composite measure of inflammatory and coagulation factors using data from the British Women’s Heart and Health Study. Results showed that higher scores on the inflammation composite significantly predicted incident mobility disability 3 and 7 years later; odds ratios for incident disability ranged from 1.8 to 2.4 (Nuesch et al., 2012). Cross-sectional (Friedman et al., 2015) and longitudinal (Friedman et al., 2018) analyses of data from the MIDUS study showed that higher levels of IL-6, CRP, and fibrinogen, modeled as a latent inflammatory factor, significantly predicted greater functional limitations, assessed using items from the SF-36 (Syddall et al., 2009). The ongoing InCHIANTI study assessed physical performance – an aggregate measure of results from chair stand, balance, and walking speed tasks - and grip strength. Results from cross-sectional analyses showed significantly poorer physical performance and weaker grip among those with IL-6 and CRP levels in the top quartiles compared to those in the lowest quartile (Cesari et al., 2004). Similarly, data on physical function and inflammation in adults age 55 or older from a set of clinical studies (all participants were diagnosed with a chronic condition or reported a disability) showed that higher levels of IL-6 and CRP were associated with longer times to complete 4-meter walk, with chair stand tests, and with weaker grip (Brinkley et al., 2009). Schaap and colleagues (2006) found that higher levels of IL-6 and CRP predicted greater loss of strength and decline in appendicular muscle mass over time in 1,000 older men and women from the Longitudinal Aging Study Amsterdam. Indeed, analyses adjusted for a range of sociodemographic, health, and lifestyle factors showed that having high circulating levels of IL-6 (>5 pg/mL) and CRP (>6.1 μg/mL) increased by 2–3 times the risk of experiencing a 40% drop in grip strength (Schaap, Pluijm, Deeg, & Visser, 2006).

There are likely to be a number of routes by which inflammation may be linked to loss of functional capacity, but there is at least some evidence of the potential for direct effects on muscle tissue. Specifically, inflammation is thought to contribute to age-related loss of muscle mass, known as sarcopenia (Peake, Della Gatta, & Cameron-Smith, 2010). The number of macrophages is lower while levels of inflammatory proteins are higher in the muscle tissue of older adults compared to younger individuals, differences that may reflect the loss of ability to repair damaged muscle efficiently as well as increases in the expression of genes related to inflammation (Peake et al., 2010). Laboratory studies have shown that inflammatory proteins reduce the production of muscle cells and increase muscle catabolism (Michaud et al., 2013), suggesting that even low-grade inflammation may act directly on muscle tissue and muscle-related genes to impair function (Michaud et al., 2013; Peake et al., 2010).

Biopsychosocial context and inflammation

The interactions among multimorbidity, disability, and inflammation take place within individuals who are embedded in psychological and social contexts that affect propensity for disease and disability. Notably, there is a well-established social gradient of health whereby incremental increases in socioeconomic status (SES: e.g., educational attainment, income, wealth, occupational status) are matched by incremental improvements in health across the full socioeconomic continuum (Marmot & Wilkinson, 1999). These socioeconomic gradients are observed for prevalence of individual conditions (Marmot & Wilkinson, 1999) and multimorbidity (Marengoni et al., 2011) alike, and also for prevalence of disability (Hosseinpoor et al., 2013). Similar social patterning of inflammation would lend additional support to the perspective that inflammation has a role in the onset, progression, and adverse outcomes related to multimorbidity. Data from multiple large-scale epidemiological studies, including NHANES (Alley et al., 2006), MIDUS (Friedman & Herd, 2010; Morozink, Friedman, Coe, & Ryff, 2010), the Coronary Artery Risk Development in Young Adults (CARDIA) study (Gruenewald, Cohen, Matthews, Tracy, & Seeman, 2009), the Multi-Ethnic Study of Atherosclerosis (MESA) (Ranjit et al., 2007), and the Atherosclerosis Risk in Communities (ARIC) study (Pollitt et al., 2008) have all shown that socioeconomic indicators like education and income are inversely associated with inflammatory markers, typically IL-6 and CRP. In many cases, the associations between inflammation and SES are linear across all levels of education and income; they are also preserved when multiple health and lifestyle influences are taken into account. Similar graded associations between SES and inflammation have been reported in studies from other countries, including Finland (Jousilahti, Salomaa, Rasi, Vahtera, & Palosuo, 2003), Russia (Glei et al., 2013), and the United Kingdom(Davillas, Benzeval, & Kumari, 2017), and in results based on other markers of socioeconomic status, such as occupation (Steptoe et al., 2003).

Adverse psychological experiences, particularly depression, are also associated with disease and disability. Major depression is both a risk factor for and a consequence of a number of chronic conditions, including cardiovascular disease (Bradley & Rumsfeld, 2015; Elderon & Whooley, 2013; Irwin, 2002) and cancer (Bortolato et al., 2017). Depression is also a significant risk factor for disability (Drageset, Eide, & Ranhoff, 2011; Nakamura, Michikawa, Imamura, Takebayashi, & Nishiwaki, 2017), and as noted above, mental illness (usually depression) in combination with other conditions is consistently linked to reduced quality of life. Circulating levels of inflammatory proteins have long been known to be elevated in patients with major depression (Howren, Lamkin, & Suls, 2009; Irwin, 2002; Raison, Capuron, & Miller, 2006), but inflammation is also directly related to variations in depressive symptoms at sub-clinical levels (Deverts et al., 2010; Howren et al., 2009; Penninx et al., 2003). The possibility that inflammation may mediate the link between depression and age-related disease and disability has been suggested (Wolkowitz, Epel, Reus, & Mellon, 2010). That said, the complexity of the associations among depression, chronic illness, and inflammation underscores the need for carefully designed studies to illuminate the precise roles depression may play.

A newer, growing literature suggests that positive psychological functioning, and not merely the absence of psychological adversity, makes contributions to health in aging adults. Positive affect (e.g., happiness) is linked to better health and reduced mortality (Boehm & Kubzansky, 2012; Chida & Steptoe, 2008; Pressman & Cohen, 2005). Similarly, engagement with meaningful life pursuits (e.g., a sense of life purpose) is associated with diverse positive health outcomes, including reduced risk of disability and mortality (Ryff, 2014). In studies including assessments of inflammatory factors, results suggest that positive psychological functioning may be protective. Low educational attainment, for example, typically predicts high levels of inflammation, but in MIDUS study participants, adults with a high school education who also had high levels of psychological well-being exhibited IL-6 levels that were comparable to participants who had completed college (Morozink et al., 2010). Germane to this review is another study, also using MIDUS data, showing that the positive linear relationship between numbers of chronic medical conditions and inflammation (IL-6 and CRP) was substantially weaker in adults with high levels of purpose in life (Friedman & Ryff, 2012), suggesting that positive psychological functioning may buffer against the increase in inflammation that typically accompanies the accumulation of chronic conditions. To illustrate, among adults with 5 chronic conditions, the difference in IL-6 levels between those with low scores on purpose in life and those with high scores was the equivalent of aging 10 years (Friedman & Ryff, 2012). Recent evidence suggests that differences in psychological well-being are reflected in differential expression of genes related to inflammation and immune function (Fredrickson et al., 2013), and engaging in behaviors that promote some forms of well-being produce changes in the expression of these genes in ways that suggest reduced propensity for inflammation (Nelson-Coffey, Fritz, Lyubomirsky, & Cole, 2017). Although the mechanisms underlying the link between well-being and gene expression have yet to be determined, β-adrenergic receptor activation has been shown to increase pro-inflammatory gene expression (Cole, 2013), suggesting that the sympathetic nervous system may be involved. In support of this possibility, circulating levels of catecholamines are lower in adults with higher levels of well-being (Ryff et al., 2006).

These diverse lines of research bolster the possibility that inflammation is a key player in age-related multimorbidity and related outcomes, particularly disability. Not only are there associations among inflammation, multimorbidity, and disability at the general population level, but among those most vulnerable – due to low SES, depression, or diminished psychological well-being – disease risk, disability rates, and inflammation are disproportionately higher.

Summary and considerations for future research

As suggested here and elsewhere in this special section, multimorbidity is a growing health issue that presents diverse complications for medical treatment. While medical advances in the ability to treat specific conditions will help to reduce the burden of multimorbidity, increased attention to and understanding of the ways in which diverse conditions are related are also needed in order to improve health outcomes and quality of life in older adults. Inflammation has been suggested as one important mechanism that may contribute to multiple chronic conditions and to their outcomes, such as disability and mortality. Circulating levels of inflammatory factors are higher in many single chronic conditions and in multimorbidity. Inflammation also represents an independent risk factor for future disease in healthy adults and for accelerated disease progression and disease-related mortality in those who are already sick. Inflammation is also an attractive candidate for mediation of well-established links between social and psychological factors and disease outcomes, as inflammation has been shown to track social status as well as psychological adversities and resources. Such associations extend to other disease-related factors not mentioned here, including the links between inflammation and experiences of racial discrimination (Lewis, Aiello, Leurgans, Kelly, & Barnes, 2010).

There are a number of important priorities for advancing this work. First, despite the volume of suggestive evidence gathered to this point, the extent to which inflammation lies in the causal pathways - driving incidence and progression of disease and disability – or is merely a robust correlate has yet to be fully established (Kaptoge et al., 2014; Pant et al., 2014). A recent clinical trial targeting interleukin-1β, however, showed marked declines in CRP along with reduced morbidity and mortality related to cardiovascular disease (Ridker, Everett, et al., 2017) and lung cancer (Ridker, MacFadyen, et al., 2017), supporting a causal role for inflammation in these diseases. Prospective epidemiological studies linking change in inflammation to changes in incident multimorbidity and related adverse outcomes, such as the longitudinal study by Fabbri and colleagues (2015), are also needed. Finally, interventions targeting social or psychological processes that are known to be linked to inflammation – social status, depression, well-being – could include proximate (e.g., risk factors) and/or distal (e.g., incidence and progression of multimorbidity and disability) measures with which to determine whether changes in inflammation are accompanied by comparable changes in health.

Second, while this review has focused on inflammation and its potential role in age-related disease and disability, and while there are reasons inflammation makes for a particularly plausible focus of research on age-related disease and disability (Franceschi & Campisi, 2014), there are numerous other potential foci, including many related to the immune system. Cellular senescence, due to aging and/or lifetime exposures to pathogens, reduces the ability to respond to exogenous and endogenous challenges, increasing risk for disease (Effros, 2009), and also promotes inflammation in older adults (Lasry & Ben-Neriah, 2015; Tchkonia, Zhu, van Deursen, Campisi, & Kirkland, 2013). Beyond the immune system there are age-related changes in multiple physiological systems (e.g., autonomic, neuroendocrine, metabolic) that can increase risk of disease (Barnes, 2015; Hart & Charkoudian, 2014; Joyner, Barnes, Hart, Wallin, & Charkoudian, 2015; Kim & Choe, 2018). Importantly, though, these systems interact with one another in affecting health outcomes, and examining individual systems in isolation may be less informative than focusing on the ways they interact. Such a multi-system approach to disease risk is seen in indices of allostatic load (Gruenewald et al., 2012; Juster, McEwen, & Lupien, 2010), for example, although such measures examine the additive effects of multi-system dysregulation rather than the ways deficits in two or more systems may interact in predicting disease and disability. More integrative approaches have shown that combinations of markers of dysregulation across multiple systems predict adverse health outcomes (Gruenewald, Seeman, Ryff, Karlamangla, & Singer, 2006). In addition, network modeling approaches are revealing the ways in which diverse conditions that often co-occur within individuals have genetic (Goh et al., 2007) and metabolic (Lee et al., 2008) processes in common.

Finally, as noted earlier, positive psychological functioning is an independent predictor of positive health outcomes and is associated with favorable profiles of inflammation. Much of the research on multimorbidity appropriately centers on the adverse exposures and experiences that increase risk for multimorbidity and related outcomes, the underlying goal being prevention. A complementary goal, especially given the growing prevalence of multimorbidity, is to help older adults avoid or adapt to the negative sequelae of multimorbidity, notably disability. When asked, most older adults report having one or more chronic conditions, but they tend to report feeling that they are aging successfully, usually because of maintaining social contacts and a sense of life purpose (Reichstadt, Sengupta, Depp, Palinkas, & Jeste, 2010; Strawbridge, Wallhagen, & Cohen, 2002). The longest-lived adults are rarely free of chronic diseases, but they are nonetheless able to maintain a high quality of life, in part due to available psychosocial resources (Poon et al., 2010). These observations highlight the potential for positive psychological functioning to improve quality of life, even in the context of multimorbidity. The fact that psychological well-being, including a sense of life purpose, can be increased in older adults (Friedman et al., 2017), points to another line of potential interventions helping older adults maintain a high quality of life as they age and incur age-related disease by promoting well-being.

It has been 25 years since Ershler (1993) noted that IL-6 increases with age and may be linked to age-related disorders. In the intervening time, inflammation has emerged as a robust risk factor for multiple chronic conditions that are prevalent in older adults and for adverse outcomes associated with these conditions. Importantly, inflammation has also been situated within a theoretical framework – geroscience and inflammaging (Franceschi & Campisi, 2014) – that posits the existence underlying age-related processes that give rise to multiple diseases of aging and that may be amenable to interventions to improve health and quality of life by slowing the aging process itself (Sierra & Kohanski, 2017). Some of these interventions may involve the precise manipulation of systems at the cellular or molecular level. Given the robust and persistent links between contextual factors – especially socioeconomic status – and disease (Phelan, Link, & Tehranifar, 2010), though, efforts to improve health and quality of life in aging adults should also include attention to the environments in which people live. Moreover, the literature on well-being and health shows that positive emotions (Steptoe, Deaton, & Stone, 2015), social connections (Holt-Lunstad, Smith, & Layton, 2010), and a sense of life purpose (Ryff, 2014) make contributions to health that are often as large as established risk factors like smoking and obesity. It may be possible to foster improvements in the health of older adults by expanding the opportunities to engage in activities that engender well-being. There are multiple health benefits of volunteering in older adults (Anderson et al., 2014), for example, including increases in brain volume (Carlson et al., 2015). Expanding the scope of geroscience to include the social and psychological contexts in which older adults live will improve the odds of successfully improving health and well-being in this growing population.