Morbidity and Mortality

A lack of social relationships has been associated with an increased risk of mortality, comparable to traditional risk factors such as being overweight or smoking (Holt-Lunstad, Smith, & Layton, 2010). A meta-analysis comparing the effects of objective isolation and loneliness suggests that both have an effect on mortality, with similar effect sizes (Holt-Lunstad, Smith, Baker, Harris, & Stephenson, 2015). Most studies examining mortality or other health outcomes, however, fail to examine both isolation and loneliness in conjunction. Of the two studies that do examine isolation and loneliness together in relation to mortality, one found that objective isolation but not loneliness was associated with mortality following adjustment (Steptoe, Shankar, Demakakos, & Wardle, 2013), while the other found the reverse (Holwerda et al., 2012).

A meta-analysis of 3 studies found that loneliness was associated with a nearly 30% increase in incident coronary heart disease (Valtorta, Kanaan, Gilbody, Ronzi, & Hanratty, 2016). As loneliness is often identified as an issue that is particularly pertinent to older adults, many studies have examined it as a risk factor for outcomes that are important to aging populations. These include outcomes such as declines in physical function, disability and frailty. Studies have found that loneliness is related to declines in motor function among older adults (Buchman et al., 2010; Shankar, McMunn, Demakakos, Hamer, & Steptoe, 2017), although findings with regard to reported increases in difficulties with basic daily activities such as dressing oneself, bathing and getting around are mixed (Perissinotto, Cenzer, & Covinsky, 2012; Shankar et al., 2017; Stessman, Rottenberg, Shimshilashvili, Ein-Mor, & Jacobs, 2014). In a cross-sectional analyses, loneliness was also found to be associated with frailty (defined as meeting at least three of the following criteria weight loss, exhaustion, low physical activity, slowness, or weakness) among community-dwelling Mexicans aged 70 years and older (Herrera-Badilla, Navarrete-Reyes, Amieva, & Avila-Funes, 2015) and urinary incontinence (Ramage-Morin & Gilmour, 2013).

Cognitive Function and Dementia

As with the studies of functional decline and frailty, there has been a growing interest in the role of social relationships in cognitive decline and dementia. Interactions with others are believed to contribute to building cognitive reserve (Fratiglioni, Wang, Ericsson, Maytan, & Winblad, 2000), thus protecting against deterioration in cognitive function as one ages, even in the presence of pathological brain changes. In their review of studies of loneliness and cognitive function, Boss et al. found that half the published studies were cross sectional, making it difficult to disentangle whether loneliness is a cause or consequence of declining cognitive function. Of the five longitudinal studies, two found that loneliness was associated with incident Alzheimer’s disease/dementia, one found that loneliness was associated with declining cognitive function after 10 years, and two found loneliness to be associated with declines in memory. No significant associations were found with executive function (Boss, Kang, & Branson, 2015). Kuiper et al. (2015), in their meta-analysis of three studies, found that loneliness was associated with a nearly 60% increase in the risk of incident dementia. Overall, however, the evidence for social relationships as a risk factor for cognitive decline remains poor (Baumgart et al., 2015), as studies often are unable to determine whether changes seen in cognitive decline are clinically significant. Hence, more longitudinal studies using larger samples with better measures of loneliness, using standardized assessments of cognitive function and dementia, and adjusting for important covariates and are necessary.

Furthermore, bidirectional associations between loneliness and cognitive decline are possible. Marked declines in cognitive ability are distressing, and can cause the sufferer to withdraw from social contact. Others may also feel unsure about how to interact with someone who has known cognitive difficulties (Ayalon et al., 2016). Hence, changes in functioning may lead to greater isolation and loneliness. A few studies have examined the bidirectional associations between loneliness and cognitive function. Two studies found stronger associations between cognitive function measures and later loneliness than vice versa (Ayalon, Shiovitz-Ezra, & Roziner, 2016; Zhong, Chen, Tu, & Conwell, 2016), while others have found that the reverse is true (Donovan et al., 2016; Wilson, Krueger, Arnold, & et al., 2007).

Mental Health

Loneliness and depression are often thought of as being the same. Indeed scales such as the Centre for Epidemiologic Studies Depression scale include “feeling lonely” as one of the symptoms of depression (Radloff, 1977). A genome-wide association study also found some evidence in support of the co-heritability of loneliness and depressive symptoms (Gao et al., 2017). However, studies have identified depression and loneliness as distinct constructs (J. T. Cacioppo et al., 2006; VanderWeele, Hawkley, Thisted, & Cacioppo, 2011; Weeks, Michela, Peplau, & Bragg, 1980). Indeed, feelings of loneliness are an important predictor of future incidence and severity of depression (J. T. Cacioppo, Hawkley, & Thisted, 2010; Holvast et al., 2015). It has also been suggested that depression may be a pathway through which loneliness can affect health (S. Cacioppo, Capitanio, & Cacioppo, 2014). Data from the Longitudinal Aging Study Amsterdam found that depression was a strong predictor of mortality among lonely men, suggesting that loneliness and depression in combination is particularly harmful for health (Holwerda et al., 2016).

Research with respect to other mental health conditions is largely cross-sectional in nature. Feelings of loneliness are common among individuals diagnosed with psychotic and mood disorders (Badcock et al., 2015; Giacco, Palumbo, Strappelli, Catapano, & Priebe, 2016; Sündermann, Onwumere, Kane, Morgan, & Kuipers, 2014). Świtaj et al. found that loneliness mediates the relationship between stigma and depressive symptoms among people with major psychiatric conditions (Świtaj, Grygiel, Anczewska, & Wciórka, 2014; Świtaj, Grygiel, Anczewska, & Wciórka, 2015). Loneliness is associated with the use of psychotropic drugs (Boehlen et al., 2015; Canham, 2015), independent of depressive symptoms and symptom severity (Boehlen et al., 2015). Loneliness was also identified as a potential mediator of the association between emotional dysregulation and the psychopathology associated with binge eating and bulimia nervosa (Southward et al., 2014). However, the cross-sectional designs of these studies make it difficult to determine whether loneliness is a risk factor for these mental health conditions or is an outcome, given what is known about isolation and stigma following diagnosis of a mental health condition (Alonso et al., 2009). Further, individuals with certain mental health conditions may suffer from symptoms that make it more difficult to develop and sustain close social networks (Anderson, Laxhman, & Priebe, 2015), leading to feelings of loneliness.

Health Behaviors

There are several reasons why individuals who are lonely may be more likely to have poorer health behaviors. The social control hypothesis (House, 2001) suggests that individuals who are lonely may be more prone to unhealthy practices due to limited access to social support and social cues that promote healthy behaviors. Another reason why loneliness may lead to risky health behavior is because lonely individuals are less likely to use active coping strategies when faced with stressful situations. Hence, their response to feeling or loneliness or other stressors may be through sad passivity (Rubenstein, Shaver, & Peplau, 1979), resorting to risky behaviors such as smoking, drinking to excess, or taking drugs. Another explanation suggests that feelings of loneliness lead individuals to be hypervigilant to social threats, thereby limiting the resources available to them for self-regulation (Hawkley & Cacioppo, 2010).

While these explanations might have a certain intuitive appeal, the evidence in support of loneliness leading to poor health behaviors remains equivocal. A review of the research examining the association between loneliness and physical activity found that nearly two-thirds of the studies in this area were cross sectional (Pels & Kleinert, 2016). Evidence from the longitudinal studies showed that while loneliness was associated with decreased physical activity, there was also evidence suggesting that inactivity may lead to feelings of loneliness over time. Findings from intervention studies indicate that physical activity may play a role in alleviating loneliness (Pels & Kleinert, 2016). Similarly, a review of the association between loneliness and smoking found that 13 of 25 studies reported a significant association (Dyal & Valente, 2015), but here, too, most studies were cross-sectional. Thus, it is unclear whether feelings of loneliness lead to individuals taking up smoking or whether smokers are likely to become lonelier over time. There is some support for the latter hypothesis, with social network analysis of the Framingham Offspring Cohort showing that with a growing number of individuals quitting smoking over time, smokers are moved to the periphery of their social networks (Christakis & Fowler, 2008). This may, in turn, lead to feelings of loneliness. There is limited evidence suggesting greater solitary drinking (Arpin, Mohr, & Brannan, 2015) and poorer food choices (Henriksen, Torsheim, & Thuen, 2014) among lonely adults, and increased substance use among lonely when compared with non-lonely adolescents and young adults (J. T. Cacioppo et al., 2002; Stickley, Koyanagi, Koposov, Schwab-Stone, & Ruchkin, 2014). Further work in this area is necessary.

Sleep

Sleep has important restorative functions. Sleep duration (long or short sleep) have been associated with mortality and a range of chronic health conditions such as diabetes (Cappuccio, D’Elia, Strazzullo, & Miller, 2010a, 2010b). There is growing evidence suggesting that loneliness may affect sleep, with lonely individuals reporting poorer sleep quality, more disturbances, greater fragmentation, and less sleep efficiency when compared with non-lonely individuals (J. T. Cacioppo et al., 2002; Kurina et al., 2011). Daily feelings of loneliness have been found to be associated with feelings of tiredness and low energy on the following day, with evidence also in support of bidirectional associations (Hawkley, Preacher, & Cacioppo, 2010). Evidence in this area is, however, quite limited.

Stress

Lonely individuals tend to evaluate events in their lives as more stressful (J. T. Cacioppo et al., 2000). Indeed, the feeling of loneliness itself acts as a stressor (House, 2001). Further, lab-based and naturalistic studies show that lonely and non-lonely individuals differ in their reactions to stress, as evidenced by changes in a range of biological markers including cardiovascular profiles, cortisol profiles and immune function. In a lab-based study, lonely college students had higher total peripheral resistance and lower cardiac output in response to stressful tasks when compared with non-lonely students (J. T. Cacioppo et al., 2002). A similar pattern was observed in a study examining ambulatory blood pressure (Hawkley, Burleson, Berntson, & Cacioppo, 2003). The authors suggest that this pattern may contribute to higher blood pressure over time in lonely individuals (J. T. Cacioppo et al., 2002; Hawkley et al., 2003). Also age-related increases in blood pressure are more marked among lonely individuals (J. T. Cacioppo et al., 2002; Hawkley, Masi, Berry, & Cacioppo, 2006). On the other hand, loneliness associated with a naturalistic stressor (migration) was not found to be associated with heart rate or heart rate variability (Gouin, Zhou, & Fitzpatrick, 2015).

Lonely individuals also show dysregulation of the HPA axis, as marked by changed cortisol profiles. Lonely individuals show a higher cortisol awakening response when compared with non-lonely individuals (Adam, Hawkley, Kudielka, & Cacioppo, 2006; Steptoe, Owen, Kunz-Ebrecht, & Brydon, 2004), but flatter cortisol profiles over the day (Doane & Adam, 2010). However, findings to the contrary have also been reported, with Cole et al. (2007) finding no significant differences in circulating cortisol levels between lonely and non-lonely adults and loneliness not being associated with urinary cortisol in the Chicago Health, Aging, and Social Relations Study (Hawkley et al., 2006). Further, in a lab-based stress task, increasing loneliness was associated with decreased cortisol responsiveness (Hackett, Hamer, Endrighi, Brydon, & Steptoe, 2012).

Genome-wide analysis indicates that genes associated with the anti-inflammatory glucocorticoid receptor pathway were underexpressed, while the proinflammatory NF-kB/Rel transcription pathway was overexpressed among lonely participants when compared with non-lonely participants (Cole et al., 2007). Data from the Taiwanese Social Environment and Biomarkers Study also showed that leukocyte sensitivity to glucocorticoid regulation was lower among participants who reported being lonely than among non-lonely participants (Cole, 2008). These findings may represent pathways through which loneliness leads to inflammatory diseases. In laboratory studies, lonely participants respond to stress tasks with increases in fibrinogen, natural killer cell response, interleukin-6, interleukin-1 receptor antagonist, and chemokine monocyte chemotactic protein-1 (Hackett et al., 2012; Steptoe et al., 2004). However, some studies have reported no significant associations between loneliness and inflammatory markers (Mezuk et al., 2016; Shankar, McMunn, Banks, & Steptoe, 2011).

Cacioppo et al. propose a model of loneliness placing the brain at the center of the experience of social relationships (J. T. Cacioppo, Cacioppo, Capitanio, & Cole, 2015; S. Cacioppo et al., 2014). According to this model, loneliness as an aversive state leads the brain to take steps that help protect the individual in the short term. These changes include a greater vigilance for social threats, poorer sleep, biological changes such as activation of the HPA axis and changes to immune function, decreased impulse control, and increased depression. While some of these changes may be beneficial immediately, over prolonged periods of time these may be detrimental to health. Hence, chronic loneliness may be particularly damaging to health. Much of the work presented in support of this model is based on animal research. Further, the evidence in support of some mechanisms such as HPA axis activation, sleep disruption, and immune function in humans is mixed. However, this model offers a useful framework for future work into understanding the processes through which loneliness affects health.