Increased depression risk among patients with chronic
osteomyelitis
Chun-Hung Tseng
a,b,1, Wei-Shih Huang
a,b,⁎
, Chih-Hsin Muo
c, Yen-Jung Chang
d, Chia-Hung Kao
b,e,⁎
Introduction
Depression is a relatively commonmood disorder that has become a
primary cause of functional impairment, disability, or loss of work productivity
[1–3]. Therefore, early recognition of risk factors for depression prevention is a prime factor for decreasing depression-related
social and economic burdens. Through themechanisms involving afferent nerves and humoral pathways, infammatory substances formed by peripheral infammations could induce certain pathophysiological
changes, such as a change in neurotransmittermetabolism, varied neuroendocrine function and alternating neuroplasticity in the brain, and
depression formation [4–7]. Because of similar pathophysiological effects on the brain, depression, including vascular depression, can also
be triggered by certain chronic medical disorders associated with chronic infammatory processes, such as hypertension [8–10], diabetes
[4,8–10], coronary heart disease (CHD) [4,8–10], stroke [8–10], and neurodegenerative courses, including Parkinson's disease (PD) [11,12]
and dementia [13,14]. However, surveys have evidenced no traditional medical risk factors for depression in certain depression patients [6,15, 16]. Therefore, exploring depression risk factors beyond those that are well documented, including hypertension [8,9], diabetes [4,8–10], CHD
[4,8–10], stroke [8–10], PD [11,12], and dementia [13,14], is necessary.
The potential role of chronic infammatory processes in altering neurotransmitter metabolism, varied neuroendocrine function, and alternating
neuroplasticity in the brain is well evidenced [4–7]. Chronic
osteomyelitis (COM), a condition involving powerful chronic infammation caused by bone infection, can last for weeks, months, or years and involve pathological processes that sustain intense infammation in the foci because of the formation of abscesses, bone debris, and sinus tracts [17]. COM can occur at the long bone and spine after a period of bacteremia or fungemia and later can spread to adjacent vertebra or into the meninges [17]. The COM patients are characterized by male predominance and the majority of them are older in age [18,19]. COM
could increase the risk of some medical conditions, such as ischemic stroke [20], coronary artery disease [21], epilepsy [22], and dementia
[23]. Certain autoimmune disorders have also exhibited increased depression risk, including rheumatoid arthritis (RA) [4,24,25], systemic
lupus erythematosus (SLE) [24,25], and infectious diseases, such as human immunodefciency virus (HIV) [26,27] and hepatitis C virus (HCV) [26,27] infections. However, the degree to which diseases exhibiting chronic infammation contribute to depression pathogenesis beyond conventional risk factors for depression, particularly for vascular depression, such as hypertension [8–10], diabetes [4,8–10], CHD [4, 8–10], and stroke [8–10], is unknown.
There have been no studies ever connecting COM, a well-known chronic infammatory disease, to the development of depression. Our nationwide study was based on a large National Health Insurance (NHI) claims database, available in Taiwan, to investigate the linkage of COM and depression in a cohort exceeding 22 million enrollees for an 8-year period from January 1, 1999 to December 31, 2006. Methods
Data source
This nationwide retrospective cohort study used the NHI benefciary fles of the inpatient claims database obtained for the 1996–2010 period from in Taiwan, which began in 1995. The inpatient claims database used in this study was extracted from the NHI Research Database
(NHIRD) [28]. NHI covers over 98% of the Taiwanese population. The International Classifcation of Diseases, Ninth Revision, Clinical Modifcation
(ICD-9-CM) was used for coding the diseases of interest in the present study.
Study subjects
Based on the NHI database, enrollees who were newly diagnosed with COM (ICD-9-CM code 730.1) from January 1, 1999, to December 31, 2006, without a history of previous depression (ICD-9-CM codes 296.36, 296.82, 298.0, 300.4, 301.12, and 311) before COM diagnoses were collected and their dates of COM diagnoses were defned as the entry dates. The study group comprised 15,529 patients with COM. The control groupwas compiled fromthe NHIRD, using a randomselection of age- and gender-matched participantswithout COMand depression with corresponding entry dates, at a ratio of 4:1 in the study group (n = 62,116).
Ethics statement
Because identifcation numbers of patients had been encrypted, patient consent was not required for this study. This study was approved by the Research Ethic Committee at China Medical University (CMUREC-101-012). The committee waived the requirement for consent.
Outcome and relevant variables
In this study, the end point was depression (ICD-9-CM codes 300.4, 296.36, 296.20-296.36, 298.0, 301.12, and 311) during the study period. The variables of relevancewere age, gender, and comorbidities, including hypertension (ICD-9-CM codes 401–405), diabetes (ICD-9-CM codes 250), CHD (ICD-9-CM codes 410–414), stroke (ICD-9-CM codes 430–438 at discharge), PD (ICD-9-CM codes 332), and dementia (ICD-9-CM codes 290.0-290.4, 291.2, 292.82, 294.10, 294.11, 331.0). Statistical analysis
The chi-square test and the t-testwere used to assess the differences of discrete and continuous variables between the COM cohort and the control group. Person-years were calculated from the entry dates to the frst dates of the occurrence of depression, withdrawal from the insurance program, death, or the end of 2010. The gender- and agespecifc incidence rates (per 1000 person-years) of depression were
compared between the 2 groups. Using the Cox proportional hazards regression model, hazard ratios (HRs)were derived to compare the risk of
developing depression between the groups. The demographic factors and comorbidities relevant to depression were retrieved for group comparison. HRs for depression were stratifed by age and gender (Table 2), and the risk of depression with the stratifcation of each of the 6 depression-relevant comorbidities was also compared between groups (Table 3). The association between COMseverity and depression was analyzed according to COMseverity,whichwas defned as the total length of hospital stay because of COM during the follow-up duration, divided by the length of follow-up duration (Table 5). Using the tertile, COM severity was further classifed as mild (the frst tertile in COM severity), moderate (the second tertile in COM severity), and severe (the
third tertile in COM severity) [20]. The cumulative incidences for depressionwere plotted using the Kaplan–Meiermodel and the difference
between groups was calculated by applying the log-rank test (Fig. 1). A 2-tailed P value b 0.05 was considered signifcant. SAS Version 9.1 (SAS Institute Inc., Carey, NC, USA) was used in the present study.
Results
Men were more likely than women were to be infected with COM (66.9% vs. 33.1%) (Table 1). Compared with the control group, the prevalence of comorbidities considered
to be depression risk factors, including hypertension, diabetes mellitus, CHD, stroke, PD,
and dementia, was signifcantly higher in the osteomyelitis group (p b 0.0001) (Table 1).
The overall incidence rates for depression in the study and in the control groups were 3.94
and 1.21 per 1000 person-years, respectively (Table 2). In both groups, women had a higher
incidence of depression thanmen did (Table 2). We observed a U-shaped distribution of depression
incidence with higher rates in both the youngest (age b 45) and oldest (age N 65)
age-groups in the COM group, which increased with age in the control group (Table 2).
The depression risk was 3.22-fold [95% confdence interval (CI): 2.82–3.67] higher in the
COM group than in the control cohort (Table 2). When adjusting for age, gender, hypertension,
diabetes, CHD, stroke, PD, and dementia, using the Cox proportional-hazards regression, the riskwas 2.84-fold (95% CI: 2.47–3.26) signifcantly higher in the COMgroup than the control
cohort (Table 2). The age-specifc adjusted hazard ratios (aHRs) were highest in the
youngest age-group (b45: 6.08, 95% CI: 1.71–7.85) with a steady decline with increasing
age (N65: 1.75, 95% CI: 1.39–2.19) (Table 2). The Kaplan–Meier analysis in the 2 groups
showed that depression risk rose during follow-up in both groups,with a signifcantly higher cumulative incidences for depression in the study group than in the control group (log-rank p b .0001) (Fig. 1).
Table 3 shows the comorbidity stratifcation analyses of the Cox proportional-hazards
regression model adjusted for age and gender. In the absence of any of the 6 relevant comorbidities that are considered to be depression risk factors (hypertension, diabetes,
CHD, stroke, PD, and dementia), the risk of depression was signifcantly higher in the
COM group (aHR= 3.04, 95% CI: 2.55–3.62) (Table 3). Comorbidities that are recognized
to increase depression riskmight affect COMas risk factors for depression. In the presence of common comorbidities associated with depression, patients with COMcarried a signifcantly higher risk of developing depression (aHRs: hypertension, 1.74 [95% CI: 1.31–2.30]; diabetes, 1.92 [95% CI: 1.36–2.71]; CHD, 2.21 [95% CI: 1.52–3.23], and stroke, 1.88 [95% CI:
1.24–2.86]) (Table 3). In the COMgroup, hypertension, diabetes, CHD, stroke, and dementia
were confrmed to increase depression risk (Table 4), consistent with prior reports [4,
8–10,13,14]. Comparedwith the control group,we observed a severity-dependent risk for
depressionwith the stratifcations of COMseverity after controlling age, gender, hypertension, diabetes, coronary heart disease, stroke, Parkinson's disease, and dementia (mild:
aHR = 1.42, 95% CI: 1.17–1.71; moderate: aHR = 4.90, 95% CI: 3.94–6.09; severe:
aHR = 26.3, 95% CI: 21.3–32.6) (Table 5).
Discussion
by infectious processes, such as HIV [26,27] and HCV [26,27] infections, and infammatory courses, such as RA [4,24,25] and SLE [24,25], could increase depression risk in addition to conventional risk factors such as hypertension, diabetes, CHD, stroke, PD, and dementia. Besides these well-known sources of infammation, other less noticed infammatory disorders, such as COM, should also be taken into consideration for the possible risk to cause depression. By accessing the large patient population in the Taiwan NHI claims data set, we tried to investigate whether
COM, a condition involving persistent long-term and high-intensity infammation, would increases depression risk in patients with or without traditional relevant depression comorbidities.
Similar to published reports [18,19], COM patients were predominantly men, 40 years of age or older (Table 1). These fndings indicate the validity of the COM patient population identifed in the Taiwanese NHI database. We associated COM diagnosis with a higher risk of depression (HR= 3.22, 95% CI: 2.87–3.67) (Table 2). Although men dominated
the COM group, COM increased the depression risk in both genders (Table 2).
Age was the most crucial risk factor for depression, particularly for the older patients [1–3,8–14], because depression-related comorbidities were more prevalent among the older patients [8–14]. We also observed this age-related effect on depression in the control group
(Table 2). However, a U-shaped distribution of depression incidence rates with relatively high rates in both young and old patients (4.43 per 1000 person-years in the age b 45 group and 4.53 per 1000 personyears in the age N 65 group) was noted in the COMgroup (Table 2). Despite a U-shaped incidence distribution in age, the relative weight of
COM in increasing depression risk was higher in the younger agegroup, with the highest aHR of 6.08 noted in the age b 45 year group,
which declined steadily with the lowest aHR (1.75) noted in the
age N 65 year group (Table 2). Since the comorbidities known to increase depression risk were “less prevalent”, i.e., “relatively clean,” in
the younger patient population, COM stood out as a more prominent risk factor than in the older patient population, in whom comorbidities known to increase depression risk played increasingly stronger roles with advancing age (Table 2). Depression risk increased gradually with the progression of COM severity, from the lowest risk in the mildest subgroup (aHR = 1.42, 95% CI: 1.17–1.71) to the highest risk
in the most severe subgroup (aHR = 26.3, 95% CI: 21.3–32.6) (Table 5). These fndings strengthen the relationship between COM and the development of depression.
Comorbidities for depression
Previous studies have established numerous risk factors for depression
[4,8–14]. Among these, age [1–3,8–14], female gender [2,3,10,13, 14], hypertension [8–10], diabetes [4,8–10], CHD [4,8–10], stroke
[8–10], PD[11,12], and dementia [13,14] have been shown to carry various weights. Depression risk increased to various extents in both COM
and control groups with superimposed comorbidities (Table 4). These fndings confrm the validity of using the NHIRD for searching supplementary risk factors for depression, such as COM, beyond the conventional
risk factors in the present study. Patients with COM without
those comorbidities that are known to increase stroke risk had higher risk for depression (aHR= 3.04, 95% CI = 2.55–3.62) than the control
group did without the same comorbidities (Table 3). These fndings suggest that COM is an independent predictor for depression, with
greater weight in younger patients (Table 2). Strengths and limitations
This population-based cohort study has numerous strengths. First, we extracted patientswith COMand age- and gender-matched controls from a data set exceeding 22 million enrollees in a national insurance
programencompassing over 98% of the entire population of Taiwan. Insurance claims for reimbursement for in-hospital management have
been under strict NHI monitoring and auditing to prevent health care fraud. This NHI surveillance program strengthens diagnosis reliability based on insurance claims. The demographic stratifcations showing higher female incidence and age distribution profles are comparable to those reported in the literature [1–3,8–14]. The demonstration of well-established depression risk factors including hypertension [8–10], diabetes [4,8–10], CHD [4,8–10], stroke [8–10], PD [11,12], and dementia
[13,14], associated with increased incidence rates of depression in both COM and control groups also supports the validity of the data presented in this report. Second, the large sample size obtained adequate power for stratifcation into subgroups for statistical analyses and enabled us to ascertain the impact of COM on depression risk, particularly in the younger age-group inwhom the cause of depression might be less certain. The
risk strengthens the role of COM as a risk factor for depression.
Finally, the increased incidence of comorbidities, including hypertension,
diabetes, CHD, PD, and dementia, which are recognized risk factors for depression among COM patients, raises the possibility that the underlying
chronic infammatory condition might also predispose COM patients to a higher risk of developing relevant depression comorbidities.
However, this study also has certain limitations. First, we could not exclude the compounding possibility that other vascular risk factors, such as altered immunity, reduced physical activity, and medications
prescribed for COM, might predispose patients to increased risk of depression, particularly vascular depression. Thus, the risks derived from
the present study might carry factors superimposed on COM. Second, personal health habits, including smoking and alcohol consumption,
were unavailable in the Taiwan NHI data set for assessing their effect on the apparent increase in depression risk in the COMgroup. However, COM increased depression risk in both genders (Table 2), and the low smoking rate (b4.5%) among adult women in Taiwan [29] suggests
that smoking is unlikely to be a confounding variable for the noted increase in depression risk in COM patients. A higher depression risk in
younger COM patients (Table 2), who have potentially less cumulative smoking exposure than older adults do, also suggests that smoking
does not play a crucial role in developing depression in the COM population. Third, in this study, the evaluation of incidence rate and risk of
depression in both COM and controls cohorts only represents the incidence and risk of depression among patients who have used health services. Hence, with this study bias, the risk of depression in this study
might be over- or under-estimated. Fourth, we demonstrated an increased risk of depression in COM patients and noted the higher prevalence rates of depression-relevant comorbidities among COM patients, compared with those without COM (Table 1). Fifth, the results of this studymay not necessarily be generalized to COM patients in community settings since the database used in this study was from inpatient claims and those COMpatients in community settings may be less likely to be diagnosed with depression and treated directly in hospital. The current study results show that even among those without risk factors for depression, COM imposed a higher risk for developing depression (Table 3). However, whether a causal relationship exists between COM and depression is unclear. Further future researches are required
to try and identify any causalmechanisms that would explain the association found in the current study or to look at bidirectional relationships
between depression and chronic infammatory diseases such as COM.
Conclusion
Our study results demonstrated that COM is a risk factor for depression. Patients with COM exhibit signifcantly higher prevalence rates of comorbidities known to increase depression. Similar as the fndings in RA, a conditionwith persistent infammation like COM, that the younger patients may be at higher risk of experiencing depression in a recent study about the prevalence of depression in rheumatoid [30], the relative weight of COM as a risk factor for depression was also greater in
the younger patient population in our study. The fndings suggest greater assertiveness in applying preventionmeasures for depression in COM patients, such as treatment of COM patients with adequate antibiotics, effective abscess drainage and wound debridement of the infected bones, hyperbaric oxygen therapy, and, if needed, limb amputation as early as possible [17], particularly those who contract the disease at a younger age.
