This study provides new evidence suggesting that specific disease clusters may synergistically increase the risk of frailty, compared with the risk from single chronic inflammatory diseases alone. The presence of coexisting inflammatory diseases could therefore be a risk factor for frailty. Chronic inflammatory diseases, such as CKD, may accelerate the progression of frailty by virtue of their end-stage pathophysiology alone. Notably, chronic inflammatory diseases and inflammation are risk factors for other causally related diseases that may hasten the progression of frailty, either individually or jointly.
In their proposed cycle of frailty, Fried and colleagues (24
) hypothesize that some acute or chronic diseases can “potentially impact on every point in the cycle.” Beyond this, when vulnerable older adults develop multimorbid chronic diseases, the aggregate effect of a higher inflammatory disease burden could perpetuate a heightened proinflammatory state, which could elevate cortisol levels, as well as decrease their muscle mass, physiological reserves, and immunocompetence (24
). Each of these effects alone, or in combination, could render them less capable of tolerating physiological stressors, more susceptible to new diseases, whether subclinical or clinical, and ultimately at greater risk of becoming frail. The implications regarding treatment of comorbid inflammatory diseases before the onset of frailty are therefore immense. These could include benefit from comanaging diseases effectively to minimize their severity and decreasing the financial burden often associated with multimorbid diseases. Further evidence is also needed as to whether clinical outcome would benefit from treating or preventing inflammation.
A second notable finding is the evidence suggesting additive biologic interactions between the diseases of specific inflammatory pairs. One theory that may explain statistically detectable synergistic interactions between paired conditions, such as depressive symptoms and anemia, is that multimorbidity could impact risk factors and physiological mechanisms, beyond clinically apparent diseases, that contribute to the risk of frailty. Researchers have recently begun investigating potential interactions between biochemical mediators and co-occurring impairments that present in geriatric syndromes, such as disability and frailty. For example, Cappola and colleagues (26
) observed a synergistic effect between low insulin-like growth factor-1 and high IL-6 levels, which was associated with an increased risk of progressive functional decline and mortality in baseline physically disabled older women from WHAS I. Keller and colleagues (27
) demonstrated in older outpatients that the synergistic interaction between vision and hearing impairments in the same individual has a greater effect on declining physical function than each sensory impairment by itself, independent of mental status and comorbidity. The identification of intersecting, or overlapping, physiological pathways between shared risk factors for frailty and other geriatric syndromes may offer new targets for developing clinical treatments in order to manage multimorbid inflammatory diseases aggressively and early on.
This study had several limitations. First, because its design was cross-sectional, we were unable to draw any conclusions regarding causality. Our study was also limited to a population-based sample of older women in Baltimore City and Baltimore County, Maryland. Therefore, the disease clusters identified as most prevalent may not be generalizable to older men because of gender-specific differences in body size and composition that may influence mortality rates from chronic diseases and the development of frailty (28
). Furthermore, compared with an 8% black female older adult representation from the 1990 U.S. Census, approximately 25% of the screened WHAS population and 20.1% of our study sample consisted of black women aged 65 years or older (10
). Because having an African American descent has been found to be associated with frailty, the higher representation of blacks in our study may render our results less generalizable to the overall U.S. older adult population, even though we adjusted for race in our analyses (3
). However, these limitations in generalizability do not negate the internal validity of the study.
Second, there were inherent challenges in measuring the effects of multimorbidity as an exposure. We attempted to draw conclusions about the risk of frailty measured only by the effects of diagnosed chronic inflammatory diseases. The proinflammatory effects contributed by subclinical diseases on the risk of frailty could not be identified.
Next, CKD was determined based upon a calculated creatinine clearance using the MDRD formula. In a study by Froissart and colleagues (29
), the MDRD formula demonstrated higher precision and accuracy than the Cockcroft–Gault equation in estimating the measured glomerular filtration rate (GFR) of all subpopulations based upon age and gender, except for women aged 65 years or older with a measured GFR less than 60 mL/minute per 1.73 m2
. As a result, misclassification of CKD may have occurred. This phenomenon could have provided an inaccurate estimate of the number of women with impaired renal function and affected the assessment of additive interaction between disease pairs that include CKD. For example, of those with both CKD and anemia, 27% of the frail cases were attributable to the interaction between the two diseases, although not statistically significant (AP = 0.27, 95% CI -0.56–1.10).
Another factor that could influence the association between disease pair patterns and the likelihood of frailty is the socioeconomic profile of the community-based population from which the study participants were recruited. Educational attainment was lower in the screened Baltimore population than in the U.S. population of women aged 65 years or older. Of the screened Baltimore population, 42% had 12 or more years of education compared with 56% of U.S. women aged 65 years or older in 1991 (10
). Approximately 20% of those who were screened for the study did not know or disclose their annual income and 70.4% reported an annual income of less than $35,000. It is possible that study participants with a lower socioeconomic status are more likely to have limited access to health care resources that promote preventive health and enhanced physical activity, which may modify their risk of developing disease-related outcomes, leading to frailty.
Finally, several previous findings, which showed the presence of synergistic interactions between different physiological biomarkers and disease risk factors in frail older adults, were cited to support our results that potential biologic interactions between inflammatory disease pairs exist (7
). However, these studies were conducted on data from the same study cohort as ours. Similar findings from other study populations would further strengthen our conclusions about synergistic inflammatory disease pair interactions.