OUP user menu

Chronic fatigue syndrome is associated with the risk of fracture: a nationwide cohort study

C.-S. Chen, W.-M. Lin, T.-Y. Yang, H.-J. Chen, C.-N. Kuo, C.-H. Kao
DOI: http://dx.doi.org/10.1093/qjmed/hcu037 635-641 First published online: 11 March 2014

Abstract

Purpose: Chronic fatigue syndrome (CFS) is a complex disorder that is associated with unreasonable persistent fatigue. CFS has also been reported to be a possible risk factor for osteopathy. We propose that CFS might be associated with an increased risk of fracture.

Methods: We used the National Health Insurance Research Database to conduct a prospective cohort study, identifying 3744 patients with a CFS diagnosis (International Classification of Diseases, Ninth Revision, Clinical Modification code 780.71) and 14 976 patients without CFS until 2006, with follow-up observed until the end of 2010.

Results: The incidence rate of fracture was higher in the CFS cohort than in the non-CFS cohort (17.44 vs. 14.53 per 1000 person-year, respectively), with an adjusted hazard ratio of 1.14 (95% confidence interval = 1.00–1.30). The risks of fracture between CFS and non-CFS were shown without comorbidity for each would be elevated than with other comorbidities, particularly in osteoporosis. The patients without osteoporosis in the CFS cohort exhibited a 1.16-fold higher risk of fracture than did those in the non-CFS cohort.

Conclusions: We propose that CFS-related fracture might not be associated with osteoporosis. The mechanism for developing CFS-related fracture remains unclear; however, we recommend noticing the prevention of fracture for CFS patients before clarifying the aetiology of CFS-related fracture.

What’s already known about this topic?

  1. Chronic fatigue syndrome (CFS) is a complex disorder that is associated with unreasonable persistent fatigue.

  2. CFS has also been reported to be a possible risk factor for osteopathy.

What does this article add?

  1. The incidence rate of fracture was higher in the CFS cohort than in the non-CFS cohort.

  2. The patients without osteoporosis in the CFS cohort exhibited a 1.16-fold higher risk of fracture.

Introduction

Chronic fatigue syndrome (CFS) affects at least 3% of the population in Western countries, and women have a higher prevalence for CFS than men do.1 CFS is characterized by persistent fatigue signs for at least 6 months and its majority definition is based on unreasonable fatigue syndrome for substantial lower individual activity.2–4 CFS is a debilitating disorder that differs from common fatigue causes involved in hepatitis, hypothyroidism and depressive disorder.5,6 Previous studies have shown that the predilection population is within the age range of 18–50 years.7,8

According to the diagnostic criteria for CFS, which are based on the 1994 US Center for Disease Control criteria, severe persistent fatigue for at least 6 months indicates CFS, which enables well-known fatigue causes to be ruled out. Furthermore, four or more symptoms other than fatigue must be observed, including clinically evaluated, unexplained, persistent, or relapsing chronic fatigue, unusual post-exertion fatigue, impaired memory or concentration, low-quality sleep, headaches, muscle pain, joint pain, sore throat and tender cervical nodes.9

The efficiency of treatments and the specificity of diagnostic tools for CFS remain insufficient. According to the definition of Fukuda et al., published in 1994, the first step to confirming CFS is to exclude other possible causes of fatigue, such as cancer and common chronic disorders. Diagnostic workflows are obtained to exclude fatigue related to common disorders, such as abnormal liver function, diabetes, thyroid disease, autoimmune disease, renal disease and malignancy disease. Recently, studies have suggested that fatigue is associated with osteopathy,10 and that CFS is also associated with another surrogate disorder, fracture, which prompts the hypothesis of an association between osteopathy and CFS.10–12

We used a population-based prospective study of the National Health Insurance Research Database (NHIRD) to survey the hazard ratio (HR) and cumulative HR to assess whether the risk for fracture differs between patients with and without CFS.

Materials and methods

Study design

We conducted a population-based prospective cohort study based on the original claims data of 1 million beneficiaries randomly sampled from the Taiwanese NHIRD. All personal identifications were encrypted by the National Health Research Institutes (NHRI) before the data were released. We also obtained the approval of the Institutional Review Board of China Medical University Hospital (CMU-REC-101-012). Previous studies have presented related information about the NHIRD13,14 and demonstrated diagnosis accuracy and validity.15,16

Study population

We used International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) diagnostic codes to identify CFS cases. We identified 3744 patients with newly diagnosed CFS (ICD-9-CM code 780.71) and 14 976 patients without CFS from the registry of ambulatory and inpatient claims data between 2000 and 2006, and the CFS diagnosis date was defined as the index date. We excluded patients who had a history of facture (ICD-9-CM codes 800829) and artificial menopause (ICD-9-CM code 627.4), or missing information on sex or age. For each CFS patient, four controls were randomly selected and frequency matched for sex, age (every 5 years) and index year. Both cohorts were followed until they were diagnosed with fracture, or until the patients were censored because of loss to follow-up, withdrawal from the NHI system, or the end of 2010.

Comorbidity variables

The following diseases were considered potential confounders in the association between CFS and fracture. The baseline comorbidity history for each patient included all types of cancer (ICD-9-CM codes 140–208), rheumatoid arthritis (ICD-9-CM code 714), hyperthyroidism (thyrotoxic crisis) (ICD-9-CM code 242), diabetes (ICD-9-CM codes 250 and A181), renal disease (ICD-9-CM codes 582–583.7, 585, 586 and 588), chronic hepatitis (ICD-9-CM codes 571, 572.2, 572.3, 572.8, 573.1–573.3, 573.8, 573.9 and A347), depression (ICD-9-CM codes 296.2–296.9) and osteoporosis (ICD-9-CM code 733.0).

Statistical analysis

We compared the distributions of age, sex and comorbidities between the case and control groups by using a chi-square test. The incidence rates of fracture were calculated in follow-up time until the end of 2010 or the date of fracture diagnosis, death, or loss to follow-up.17 We used the Kaplan–Meier (K–M) estimation method to depict cumulative incidence curves of fracture in the CFS and non-CFS cohorts, and used a log-rank test to examine whether the K–M curves differed statistically. Cox’s proportional hazard regression analysis was conducted to measure the effects of CFS on the risk of fracture. HRs and 95% confidence intervals (CIs) were calculated in the model. We performed all statistical analyses by using the SAS statistical package (Version 9.2 for Windows; SAS Institute, Inc., Cary, NC, USA). We set the statistical significance at α = 0.05 and depicted the survival curves using R statistical software (Version 2.14.1 for Windows; R Development CT, Vienna, Austria).

Results

Among the claims data from 2000 to 2006, 3744 patients with CFS and 14 976 people without CFS met the eligibility criteria (Table 1). The two groups were similar in sex and age distributions, with a mean age of 48 years. Cancer, hyperthyroidism, diabetes, renal disease, chronic hepatitis, depression and osteoporosis were more prevalent in the CFS group than in the non-CFS group at baseline.

View this table:
Table 1

Demographic factors and comorbidity of study participants according to CFS status

Non-CFSCFSP-value
N = 14 976N = 3744
Variablen%n%
Gender0.99
    Female774851.74193751.74
    Male722848.26180748.26
Age, years0.99
    <50817654.59204454.59
    ≥50680045.40170045.41
Comorbidity
    Cancer3392.261102.940.02
    Rheumatoid arthritis170.1180.210.21
    Hyperthyroidism2051.37872.32<0.001
    Diabetes12838.5745012.02<0.001
    Renal disease4442.961624.33<0.001
    Chronic hepatitis271918.16139137.15<0.001
    Depression2481.661233.29<0.001
    Osteoporosis8185.462556.810.002

In Table 2, the incidence rate for fracture in the CFS cohort (17.44 per 1000 person-year) was higher than that in non-CFS cohort (14.53 per 1000 person-year). The adjusted HR (aHR) in the CFS cohort indicated a 1.14-fold increased risk of fracture compared with the non-CFS cohort when adjusted for age, sex and comorbidities. When stratified according to sex, the crude HR for fracture indicated that men in the CFS cohort exhibited a significantly higher risk of fracture (1.26, 95% CI = 1.05–1.53) than did those in the non-CFS cohort. However, the aHR indicated a non-significant 1.21-fold increased risk after adjusting for age, sex and comorbidities. The stratified age group, ≥50 or <50 years, exhibited an increased crude HR, indicating an approximately 1.17- to 1.29-fold increased risk of fracture, and the aHR was approximately 1.10- to 1.20-fold higher, despite being non-significant. The CFS patients without cancer, rheumatoid arthritis, diabetes, renal disease, chronic hepatitis, or osteoporosis were more likely to suffer a fracture than non-CFS patients were (HR = 1.16, 95% CI = 1.02–1.32 for those without cancer; HR = 1.15, 95% CI = 1.01–1.31 for those without rheumatoid arthritis; HR = 1.21, 95% CI = 1.05–1.40 for those without diabetes; HR = 1.15, 95% CI = 1.01–1.32 for those without renal disease; HR = 1.24, 95% CI = 1.06–1.46 for those without chronic hepatitis; and HR = 1.16, 95% CI = 1.01–1.34 for those without osteoporosis).

View this table:
Table 2

Incidence density rates and HR for facture according to CFS status stratified by demographic factors and comorbidity

CFSCFS to non-CFS
NoYesCrude HR (95% CI)Adjusted HR (95% CI)
VariablesCasePerson-yearIRCasePerson-yearIR
Overall104671976.3114.5330617548.7617.441.20 (1.06–1.36)*1.14 (1.00–1.30)*
Gender
    Female59137367.8815.821679168.4818.211.15 (0.97–1.37)1.09 (0.92–1.30)
    Male45534608.4313.151398380.2816.591.26 (1.05–1.53)*1.21 (1.00–1.47)
Age, years
    <5034740977.458.4711010083.9310.911.29 (1.04–1.60)*1.20 (0.96–1.50)
    ≥5069930998.8622.551967464.8226.261.17 (0.99–1.37)1.10 (0.94–1.29)
Comorbidity
    Cancer
        No101270640.3114.3329917155.7617.431.22 (1.07–1.38)**1.16 (1.02–1.32)*
        Yes341336.0025.457393.0017.810.68 (0.30–1.53)0.75 (0.33–1.73)
    Rheumatoid arthritis
        No104371913.1414.5030617513.9117.471.21 (1.06–1.37)**1.15 (1.01–1.31)*
        Yes363.1647.50034.840.00
    Hyperthyroidism
        No103071011.0214.5029717132.9817.331.20 (1.05–1.36)*1.14 (1.00–1.30)
        Yes16965.2916.589415.7821.651.32 (0.58–2.99)1.35 (0.58–3.15)
    Diabetes
        No89466334.2813.4825815614.9716.521.23 (1.07–1.41)**1.21 (1.05–1.40)*
        Yes1525642.0326.94481933.7924.820.92 (0.67–1.27)0.87 (0.63–1.21)
    Renal disease
        No99470242.0914.1528616918.1616.901.20 (1.05–1.36)**1.15 (1.01–1.32)*
        Yes521734.2229.9820630.6031.721.06 (0.63–1.77)1.06 (0.63–1.80)
    Chronic hepatitis
        No78559519.0113.1918011035.9116.311.24 (1.05–1.45)*1.24 (1.06–1.46)*
        Yes26112457.3020.951266512.8519.350.93 (0.75–1.14)1.02 (0.82–1.26)
    Depression
        No102070846.7014.4028817000.5716.941.18 (1.03–1.34)*1.13 (0.99–1.29)
        Yes261129.6123.0218548.1932.841.42 (0.78–2.59)1.65 (0.89–3.05)
    Osteoporosis
        No91668395.6713.3926216398.9015.981.19 (1.04–1.37)*1.16 (1.01–1.34)*
        Yes1303580.6436.31441149.8538.271.05 (0.75–1.48)1.00 (0.70–1.42)
  • IR, incidence density rates, per 1000 person-years; Adjusted HR, mutually adjusted for age, gender and comorbidity in Cox proportional hazards regression. *P < 0.05, **P < 0.01.

The cumulative incidence rate of the CFS cohort was significantly correlated with a higher risk of fracture compared with the non-CFS cohort were ascending for the follow-up years, as shown in Figure 1A (log-rank test, P = 0.005). We also demonstrated that the cumulative incidence rate for fracture in the CFS cohort was significantly higher than in the non-CFS cohort for patients aged between 15 and 65 years, as shown in Figure 1B (log-rank test, P = 0.03). The female patients aged 15–65 years exhibited a non-significant difference between the CFS and non-CFS cohorts, as shown in Figure 2A (log-rank test, P = 0.22), and the male patients aged 15–65 years exhibited a borderline significant difference between the CFS and non-CFS cohorts, as shown in Figure 2B (log-rank test, P = 0.06). The cumulative incidence rate of fracture in male patients in the CFS cohort indicated a significant 1.2-fold increased risk of fracture.

Figure 1.

Cumulative incidence rates of fracture for CFS and non-CFS groups: (A) follow-up years and (B) age.

Figure 2.

Cumulative incidence rates of fracture in female (A) and male (B) according to CFS status.

Discussion

The 1994 Fukuda definition provides a clear definition of CFS, and the diagnostic flowchart also stipulated that other causes of CFS must first be ruled out when diagnosing CFS.2 CFS is similar to other disorders, such as fibromyalgia, but no efficient treatments for CFS exist;18,19 treatment and medication for similar disorders are provided through national health insurance. We did not exclude disorders similar to CFS in the non-CFS cohort because it would have caused underestimation of the risk of fracture between the CFS and non-CFS cohorts. Fortunately, the CFS and non-CFS cohorts did not include any fibromyalgia patients. A previous study indicated that fibromyalgia might be associated with osteoporosis.20 However, we observed that the CFS cohort without fibromyalgia exhibited an increased risk of fracture compared with the non-CFS cohort. Furthermore the CFS was without any diagnostic tool and specific treatment for application of insurance benefits, we considered the hidden bias of insurance benefits would be less.

This study had several limitations. First, the CFS-related fracture was heterogeneous in the non-CFS cohort composition; we restricted the people which were not diagnosed as CFS into the comparison group of national wide population database during 2000–06, and followed up to end of 2010. One previous study of CFS incidence and prevalence rate,21 the incidence rate of CFS was estimated as annual incidence of CFS as 3.7 per 1000 and the prevalence as 7.4 per 1000. However, the sample was non-random and relatively small. And the other study was shown the CFS prevalence rate as 0.42% of random community-based samples. The incidence of CFS patients was under-estimated by lower accuracy of diagnosis of CFS and competition of other disease with common insurance benefits. Therefore, we conducted random matched-pair analyses to describe the risk for developing fracture between CFS and non-CFS. Thus, we had underestimated the risk for fracture despite adjusting for CFS- and fracture-related comorbidities, such as rheumatoid arthritis, hyperthyroidism, diabetes, renal disease, chronic hepatitis, depression and osteoporosis. Second, a recent study showed that CFS might overlap with cancer-related fatigue in clinical diagnosis,21 even with similar mechanisms such as immunologic responses.22 However, because cancer-related fatigue does not have an ICD-9-CM code, we could not clearly identify it in the NHIRD. We attempted to adjust for all cancer types to stratify cancer-related fatigue. Among the cancer-free patients, the risk for developing fracture between CFS and non-CFS cohorts was also statistically significant. Third, we determined CFS to be a possible risk factor for developing fracture when we adjusted for cancer, osteoporosis and other comorbidities. Our study supports the finding of a previous study that proposed that CFS might be indirectly associated with osteopathies, such as osteoporosis and osteopenia. However, the results indicated that the aetiology of CFS-related fracture was unique, and that CFS does not develop directly through osteoporosis. However, it is difficult to clarify whether misclassification of and competition among disorder diagnoses in clinics is a reason for the association between CFS and osteopathy. We considered that CFS-related fracture might exist when comparing CFS and non-CFS cohorts without accounting for osteoporosis.

We observed that patients in the CFS cohort without comorbidities for each exhibited a significantly higher risk of fracture compared with the non-CFS cohort, indicating the association between CFS and fracture, which had reduced the disturbance of cancer and osteoporosis. However, immobility or history of syncope also might be increasing the risk for developing fracture; we must further clarify the aetiology of CFS and fracture by improving more accurate diagnoses of CFS.

The cumulative incidence rate for CFS-related fracture stratified according to age was first determined in this study. The overall cumulative incidence rate for fracture revealed the significant difference between the CFS and non-CFS cohorts, and we propose that menopause might account for the highest risk among 50- to 60-year-olds. However, the cumulative incidence rates stratified according to sex showed that the male CFS patients aged 15–65 years were at significantly high risk of fracture, whereas female patients were not.

The results showed that CFS might be associated with the risk of fracture, and even CFS without comorbidity for fracture might also be associated with an increased risk of fracture. The male CFS patients were noticed with risks for fracture needed further investigation. However, we recommended the male CFS patients to avoid dangerous behaviours among labour force and suggested that postmenopausal female patients with CFS require improved fracture-prevention.

Authors’ contribution

Conception/design: T.-Y.Y. and C.-H.K.; provision of study materials and patients: C.-S.C., C.-N.K. and H.-J.C.; collection and/or assembly of data: C.-S.C., C.-H.K. and H.-J.C.; data analysis and interpretation: C.-S.C., C.-H.K., W.-M.L.; manuscript writing and final approval of manuscript: all authors.

Acknowledgements

This study was supported by grants from China Medical University Hospital (DMR-103-018 and DMR-103-020); the Clinical Trials and Research Center for Excellence, Ministry of Health and Welfare, Taiwan (DOH102-TD-B-111-004); the Cancer Research Center for Excellence, Ministry of Health and Welfare, Taiwan (MOHW103-TD-B-111-03); and the International Research-Intensive Center of Excellence, Taiwan (NSC101-2911-I-002-303).

Conflict of interest: None declared.

Footnotes

  • *These authors contributed equally to this work.

References

View Abstract