Chronic osteomyelitis increases long-term mortality risk in the elderly: a nationwide population-based cohort study

The proportion of elderly individuals (≥65 years old) comprised 6.2 % of the world population in 1992, which is expected to rise to 20 % by 2050 [1]. By 2030, 20 % of the U.S. population is estimated to be more than 65 years old [2]. In Taiwan, owing to a decrease in the fertility rate, medical advances, and a comprehensive national healthcare system, the proportion of elderly individuals increased from 7 % in 1993 to 11.33 % in 2013 [3]. This steady increase in the elderly population also necessitates the increase of medical and healthcare resources [4]. In 2013, the cost of medical expenditure for the elderly population was estimated to be 33 % of the total Taiwan National Health Insurance program, which is almost three to four times than that of the non-elderly in terms of the average cost per person [4].

Osteomyelitis is a common musculoskeletal infectious disease in the elderly, second only to skin and soft tissue infection [5]. The elderly are predisposed to osteomyelitis because of the immunocompromised nature of aging [6] and comorbidities such as diabetes mellitus (DM), peripheral vascular disease, pressure ulcers, and surgical interventions [5, 7]. Osteomyelitis can be divided into three subgroups according to the acuteness of the infectious process: acute, subacute, and chronic osteomyelitis (CO) [5]. Osteomyelitis in the elderly is most often caused by pyogenic organisms, followed by Mycobacterium tuberculosis [5]. Acute osteomyelitis is usually caused by hematogenous spread, and Staphylococcus aureus is the most common pathogen in this context [5, 8]. Antimicrobial therapy alone for 4–6 weeks is usually effective for treating acute and subacute osteomyelitis [5, 9]. CO may be caused by S. aureus or Gram-negative organisms such as Bacteroides fragilis [5, 9]. Because of the poor blood supply to the infected bone, CO usually requires not only antibiotic therapy but also adequate surgical debridement [5, 6, 9]. Other adjuvant therapies such as antibiotics, hyperbaric oxygen, nutritional supplementation, advice on smoking cessation, tight blood glucose control, arterial bypass surgery, and discontinuation or alteration of medications are recommended according to individual conditions [6]. Despite the treatments, the persistence of CO is not uncommon [5], which results in chronic disability, impairment of the quality of life, and even an increased risk of long-term mortality. Several epidemiological studies have reported that CO increases the risk of subsequent coronary heart disease [10], stroke [11], DM [12], renal disease [13], and depression [14]. However, the long-term mortality risk of CO in the elderly has not yet been clarified. Therefore, we designed a retrospective nationwide population-based cohort study to delineate this issue. We hypothesized that long-term mortality increased in the elderly with CO.

There was no significant difference in age and gender between the two matched cohorts (Table 1). The mean ages of patients with CO and those without CO were 74.89 ± 6.45 and 74.87 ± 6.43 years, respectively. Young elderly (65–74 years) comprised the majority (~55 %) followed by the moderately elderly (75–84 years) (~37 %) in each cohort. There were more male patients with CO than female patients (52 % vs. 48 %). Patients with CO had significantly more number of comorbidities such as DM, HTN, renal disease, stroke, CHF, cancer, and COPD than patients without CO (all p < 0.05). There was no significant difference in the monthly income between the two cohorts.

After the 13-year follow-up period until 2011, patients with CO showed a significantly higher mortality risk than patients without CO [IRR: 2.29; 95 % confidence interval (CI): 2.01–2.59] (Table 2). Stratification analysis of age showed the highest mortality risk of CO in the old elderly (≥85 years) (IRR: 3.27; 95 % CI: 2.22–4.82). Male patients with CO had a higher mortality risk than their female counterparts (IRR: 2.7 vs. 1.77). Subgroup analyses of comorbidities such as DM, HTN, renal disease, stroke, cancer, and COPD also showed a higher risk of mortality in the patients with CO than those without CO. Stratification analysis of the follow-up period data revealed the highest mortality risk in the first month (IRR: 5.01; 95 % CI: 2.02–12.42), and the risk was persistently higher throughout the follow-up period, even after 6 years (IRR: 1.53; 95 % CI: 1.13–2.06), among the patients with CO compared with those without CO. The Kaplan–Meier survival analysis also showed that patients with CO had a significantly lower survival rate than patients without CO (p < 0.0001) (Fig. 2).

Cox proportional hazard regression analysis showed that CO was an independent predictor of mortality after adjusting for age, gender, comorbidities, and monthly income [adjusted hazard ratio (AHR): 1.89; 95 % CI: 1.66–2.15] (Table 3). In addition, advanced age, male gender, and all of the comorbidities included in this study were independent predictors of mortality.

This cohort study demonstrated that CO significantly increased the risk of long-term mortality in the elderly patients during the 13-year follow-up period. The effect was highest in the first month, which persisted even after 6 years, and the impact of CO was directly proportional to the age of the patients. Male patients had a higher mortality risk than female patients if they had CO. In addition to CO, comorbidities such as DM, HTN, renal disease, stroke, CHF, liver disease, cancer, and COPD were also independent predictors of mortality.

The increased long-term mortality in the elderly due to CO could probably be attributed to infection, chronic inflammatory reaction, and CO-related chronic complications such as impaired limb function and psychological status. CO can occur as a result of microorganisms originating from the hematogenous route, spread of infection to the bone from the adjacent soft tissues and joints, or post-trauma or post-surgery infection [5, 17]. S. aureus is the commonest causative organism of hematogenous CO, although multiple organisms are usually isolated in cases where CO results from direct inoculation or through contiguous spread [5, 18]. Underlying the pathological processes of infection are inflammation, suppuration, necrosis, exudation, vascular congestion and intraosseous HTN, intravascular thrombosis with occlusion of blood flow, and reactive new bone formation [19]. Because of the presence of sequestrum (dead bone) in CO, it is always very difficult to eradicate the infection completely [5, 7, 19]. The mainstay treatments for CO are surgical debridement, management of ensuing dead space, adjunctive antibiotic therapy, reconstruction of soft tissue defects, and hyperbaric oxygen therapy [5, 7, 19]. However, the elderly patients are always too frail to receive complete surgical debridement with adjunctive therapy in the real situation, which results in the non-curable nature of the disease and a vicious cycle of the general conditions.

Chronic inflammatory reaction is a risk factor for cardiovascular disease and even death [20, 21]. Inflammation triggers the production of proinflammatory cytokines in the arterial wall [22]. Primary cytokines such as tumor necrosis factor-α and interleukin-1 mediate the attraction and migration of inflammatory cells into the vascular tissue [22]. They also induce the “messenger” cytokines, which are released into the systemic circulation, causing the liver to increase the production of acute phase reactants such as C-reactive protein and serum amyloid A, which amplify the inflammatory and procoagulant responses [22‐24]. Other risk factors including concomitant DM, HTN, and smoking also add to the inflammation [21]. Finally, all these processes lead to vascular atherosclerosis and cardiovascular events.

The high number of comorbidities suggesting a poor underlying condition renders the elderly particularly vulnerable to CO, which in turns causes more chronic complications including limb deformities, limb length inequality, impaired limb function, pathological fractures, malignant transformation, compartment syndrome and Volkmann contracture, coronary heart disease [10], stroke [11], DM [12], chronic renal failure [13], loss of self-esteem and depression [14], and secondary amyloidosis leading to nephrotic syndrome [19]. Our results showed that elderly patients with chronic comorbidities had higher mortality risk when they had CO. Furthermore, chronic comorbidities also predicted the subsequent mortality in the elderly. The extremity disability caused by CO results in decreased daily activities and elevated risk of other chronic physical and psychological disorders [14], which also may increase the risk of mortality when combining with previous comorbidities. This is a vicious circle suggesting that both prevention of CO and control of comorbidities are very important management strategies for the elderly with CO.

There are some limitations to this study. First, there was no detailed information regarding the treatment strategies (e.g., surgery, antibiotics, hyperbaric oxygen, etc.), laboratory data, lifestyle, and personal health factors including smoking and obesity, which may be confounding factors. Second, we did not classify CO into subgroups based on the affected area (i.e., infection site) such as vertebrae, hip, sacrum, sternum, or mandible and the induced mechanisms such as traumatic, hematogenous, diabetic, post-surgery or pressure ulcer, which may have a different prognosis. Nevertheless, as this study was designed to investigate the long-term effects of CO in the elderly, subgroup analysis was beyond the primary goal. Third, potential information biases due to misclassification by the ICD-9-CM diagnosis codes in NHIRD may exist; however, this misclassification is likely nondifferential. Because nearly 100 % Taiwan’s 23.75 million individuals were enrolled into NHIRD, we thought the selection bias is very small. Fourth, we did not have the data of direct causes of death in this study. Further studies are warranted to address these issues. For example, a prospective well-designed study with detailed data including mechanism of infection, infection site, treatment strategies, and more important measures for physical as well as psychological assessment such as SF-36 (Short Form 36 Health Surveys), would better evaluate the effect of CO on mortality in the elder patients.

This is the first nationwide population-based cohort study delineating that CO increased the long-term mortality in the elderly, particularly in the males and those with an advanced age. The effect was most significant in the first month and persisted even after 6 years. In addition to CO, chronic underlying comorbidities also predict the mortality. Early prevention and treatment of CO and concomitant control of comorbidities are the suggested management strategies for the elderly with CO.

The dataset supporting the conclusions of this article is included within the article.