Original Article
Non-apnea sleep disorders will increase subsequent liver cancer risk –
A nationwide population-based cohort study
Ji-An Liang
a,b,1, Li-Min Sun
c,1, Chih-Hsin Muo
d,e, Fung-Chang Sung
d,e, Shih-Ni Chang
e,f,g,⇑,
Chia-Hung Kao
b,h,⇑a
Department of Radiation Oncology, China Medical University Hospital, Taichung, Taiwan
b
Institute of Clinical Medicine Science and School of Medicine, College of Medicine, China Medical University, Taichung, Taiwan
c
Department of Radiation Oncology, Zuoying Armed Forces General Hospital, Kaohsiung, Taiwan
dManagement Office for Health Data, China Medical University Hospital, Taichung, Taiwan e
Institute of Environmental Health, College of Public Health, China Medical University, Taichung, Taiwan
f
The Ph.D. Program for Cancer Biology and Drug Discovery, China Medical University, Taichung, Taiwan
g
Institute of Biomedical Sciences, Academia Sinica, Taipei, Taiwan
h
Department of Nuclear Medicine and PET Center, China Medical University Hospital, Taichung, Taiwan
a r t i c l e
i n f o
Article history:
Received 2 November 2011
Received in revised form 13 February 2012 Accepted 14 February 2012
Available online 13 April 2012 Keywords: Sleep disorders Liver cancer Cohort study Apnea Breast cancer
National Health Insurance system
a b s t r a c t
Introduction: It is well known that patients with sleep disorders (SD) have an increased risk of cardiovas-cular disease, diabetes mellitus, obesity, and total mortality. However, little information exists regarding the relationship between non-apnea SD and the risk of cancer. The goal of this study was to determine if any association between SD and malignancy exists in Taiwan.
Methods: We used data from the National Health Insurance system of Taiwan to assess this issue. The SD cohort contained 42,351 patients, and each patient was randomly frequency-matched by age and sex with two people from the general population without SD. The Cox’s proportional hazard regression anal-ysis was conducted to estimate the effects of SD on cancer risk.
Results: In patients with SD, the overall risk of developing cancer was significantly higher than in normal healthy subjects (adjusted Hazard ratio [HR] = 1.12, 95% confidence interval = 1.06–1.18). This held true even when we analyzed males and females separately. In regards to individual types of cancer, the risk for developing liver cancer among patients with SD was significantly higher than in subjects without SD. For breast cancer the risk showed a marginally significant increase.
Conclusion: The nationwide population-based cohort study found Taiwanese patients with SD have a higher risk of developing cancer, particularly liver cancer and, possibly, breast cancer.
Ó 2012 Elsevier B.V. All rights reserved.
1. Introduction
Sleep problems are not uncommon in the general population
[33]. A previous study conducted in the Los Angeles metropolitan
area found the prevalence of sleep disorders to be 32.2%[2]. The
similar result was also reported for Asians[5]. A lack of quality sleep can cause accidents, affect relationships and mental prowess, and have detrimental effects on health outcomes such as increas-ing the risk of cardiovascular disease, diabetes mellitus, obesity, and total mortality[17]. When looking at the relationship between sleep problems and cancer, most researchers have focused on the issue of cancer induced sleep disturbance[1,7,10,21,27,28]. There is little information discussing the possible reverse causality – non-apnea sleep disorders (SD) may increase cancer risk.
Thomp-son et al. [31]found that a short duration of sleep can increase
the risk of colorectal adenoma, and a couple of studies have sug-gested that long sleepers or the combination of physical activity and a good-night’s sleep might reduce the risk of breast cancer
[19,32]. Poor sleep enhances proinflammatory cytokine production
and may induce immunosuppression[11,12]. Immunosuppression
plays an important role in the initiation process of cancer for pa-tients after organ transplant, as well as for papa-tients who are HIV 1389-9457/$ - see front matter Ó 2012 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.sleep.2012.02.005
Abbreviations: SD, sleep disorders; HR, Hazard ratio; NHI, National Health Insurance; NHRI, National Health Research Institute; CI, confidence interval; RR, rate ratio.
⇑ Corresponding authors. Address: Management Office for Health Data, China Medical University Hospital, No. 2, Yuh-Der Road, Taichung 404, Taiwan. Tel.: +886 4 22052121x7805; fax: +886 4 22339216 (S.-N. Chang), Department of Nuclear Medicine and PET Center, China Medical University Hospital, No. 2, Yuh-Der Road, Taichung 404, Taiwan. Tel.: +886 4 22052121x7412; fax: +886 4 22336174 (C.-H. Kao).
E-mail addresses: u9665856@cmu.edu.tw(S.-N. Chang), d10040@mail.cmuh. org.tw(C.-H. Kao).
1 These authors equally contributed to this work.
Contents lists available atSciVerse ScienceDirect
Sleep Medicine
positive[23,25]. Based on the links between the conditions and diseases mentioned above, we were interested in exploring the possible relationship between SD and cancer risk. In this study we tried to test the hypothesis that SD can increase the risk of developing cancer.
To the best of our knowledge, no published population-based study has investigated the relationship between SD and cancer risks. The aim of this study is to determine if there is any associa-tion between cancer risk and SD in Taiwan. The results presented in this paper were from a retrospective cohort study to assess the possibility of a higher risk of developing malignancy in patients with SD. The original database was derived from the National Health Insurance (NHI) system in Taiwan.
2. Methods
2.1. Study design and data collection
This study used data retrieved from the medical claims data-base of Taiwan’s Universal NHI program. The NHI program covers more than 96% of the country’s population and has contracted with 97% of all hospitals and clinics in Taiwan[16]. The National Health Research Institute (NHRI), which established and maintains the claims database, created a research dataset containing all reim-bursement claims records from 1996 to 2008 for one million ran-domly selected insured patients. Details of this population-based
program have been described in a previous study[14].
In the period from 2000 January to 2001 December newly diag-nosed subjects with SD were identified using the ICD-9-CM code 780.5 and 307.4 (except for sleep apnea syndrome [ICD-9-CM code 780.51, 780.53, 780.57]) as the exposure cohort. We defined the first diagnosis of SD in the database as index date, and we also ex-cluded subjects with SD before 2000 to ensure our study subjects had never had SD before the beginning of the insurance program. Subjects with a history of malignant cancer (ICD-9-CM code 140–208) diagnosed before the index date were excluded. We also used the database to randomly select two insured people without SD or a history of cancer as a control group in the same period, fre-quency matched with the SD cohort using age and sex. This study analyzed a total of 127,053 subjects.
2.2. Study end-point
We used the unique personal-identification number of the sub-ject as a link to the registry of Catastrophic Illness Patient Database in order to search for new diagnoses of cancer as the outcome of this study. The diagnosis of cancer in the program needs histolog-ical confirmation. Both the SD patient cohort and the comparison cohort were started using the same index date; observation was completed when cancer was diagnosed or censored prior to or on December 31, 2008. Length of follow-up time was calculated for each patient diagnosed with cancer. The completion date was de-fined as the date of death in the follow-up period or the date of last withdrawal from the NHI program.
2.3. Statistical analysis
We compared the distribution differences of socio-demographic variables and baseline co-morbidities between the SD cohort and non-SD cohort using the Chi-square test. We also estimated sex, age, urbanization level, and co-morbidities specific cancer inci-dence density and rate ratios by the population person-year for two cohorts. The townships within which subjects registered for insurance were grouped into four levels of urbanization, level 1 being the highest and level 4 the lowest level of urbanization,
based on a score calculated by incorporating variables indicating
population density (people/km2) and the population ratio of the
elderly, agriculture workers, different educational levels, and the
number of physicians per 100,000 people[15].
The Cox proportional-hazards regression models were con-ducted to assess the influence of the SDs on cancer risk. When the association was significant, a Hazard ratio (HR) and 95% confi-dence interval (95% CI) were reported. Models were adjusted for age, sex, levels of urbanization, and co-morbidities. The co-morbid-ities included hypertension (ICD-9-CM code: 401–405), diabetes mellitus (ICD-9-CM code: 250), hyperlipidemia (ICD-9-CM code: 272), and heart disease (ICD-9-CM code: 410–429). The same mod-el was used for the sex- and cancer type-specific analysis. We also analyze the Hazard ratio of liver cancer after controlling for liver cancer-related co-morbidities: hypertension, diabetes mellitus, hyperlipidemia, heart disease, alcoholic liver damage (ICD-9-CM code: 571.0, 571.1 and 571.3), non-alcoholic liver damage (ICD-9-CM code: 571.8), hepatitis B (ICD-(ICD-9-CM code: V02.61, 070.20, 070.22, 070.30, 070.32), hepatitis C (ICD-9-CM code: V02.62, 070.41, 070.44, 070.51, 070.54), cirrhosis (ICD-9-CM code: 571.2, 571.5, 571.6), and alcoholism (ICD-9-CM code: 303 and 305.0).
A two-tail P value of less than 0.05 was considered to indicate statistical significance. SAS 9.1 statistical software (SAS Institute, Inc., Cary, NC, USA) was used to perform the analysis.
3. Results
3.1. Subject characteristics
For the period 2000–2001 we identified 42,351 patients with SD; this group was comprised of 26,079 (61.6%) women and 16,272 (38.4%) men, with a mean age of 49.0 at baseline (Table 1). Age and sex distributions were similar in both groups, with the most prevalent age group being 50–64 years of age (23.8%). Com-pared with the non-SD group, subjects with SD were more likely to have hypertension (33.6% vs. 20.9%, p < 0.0001), diabetes melli-tus (10.9% vs. 7.3%, p < 0.0001), hyperlipidemia (17.2% vs. 9.4%, p < 0.0001), and heart disease (27.8% vs. 13.5%, p < 0.0001). 3.2. Rate ratio and risk of cancer
During the follow-up period of 310,866 person-years in the SD cohort and 618,434 person-years in the non-SD cohort, we identi-fied a total of 5898 cancer cases (Table 2). Overall, the incidence of cancer was higher in the SD cohort than in the non-SD group (6.77 vs. 6.13 per 1000 person-years). In both cohorts, men (8.02 and 8.99 per 1000 person-years, RR = 1.12, 95% CI = 1.04–1.21) had greater incidence of cancer than women (5.00 and 5.46 per 1000 person-years, RR = 1.09, 95% CI = 1.01–1.18). Compared with the non-SD cohort, the RR was highest in those less than 30 years old (RR = 1.92, 95% CI = 1.25–2.95).
In the Cox proportional-hazards regression model adjusting for covariates inTable 1, subjects with SD were more associated with increased risk of cancer (HR = 1.12, 95% CI = 1.06–1.18). The risk of cancer associated with SD was statistically significantly in both women (HR = 1.08, 95% CI = 1.00–1.17) and men (HR = 1.14, 95% CI = 1.06–1.24) (Table 3).
In addition, the cancer specific analysis is displayed inTable 4. We only observed that patients with SD had significantly increased risk of developing liver cancer (HR = 1.64, 95% CI = 1.42–1.89) and marginally significantly increased risk of developing breast cancer (HR = 1.17, 95% CI = 0.98–1.39). The SD patients are more likely to have liver cancer-related co-morbidities, including alcoholic liver damage, non-alcoholic liver damage, hepatitis B, hepatitis C, cirrhosis, and alcoholism, than non-SD patients (all p-values were
less than 0.0001,Table 5).Table 6also shows that the SD patients had an appreciatively increased risk of 40% for developing liver cancer (HR = 1.40, 95% CI = 1.21–1.62) after controlling liver cancer-related variables. We also analyzed the relations between different risk factors and liver cancer, and the results reveal all but alcoholism had higher risks for liver cancer (Table 7).
4. Discussion
Our studies found that patients with SD had a significantly higher risk of developing all cancer, especially liver cancer. For
female patients, it also showed a marginally significant higher risk of developing breast cancer.
The epidemiology data from Surveillance Epidemiology and End Results showed that the overall cancer rates for all racial and eth-nic groups combined decreased by 0.7% per year between 1999 and 2006[9]. However, it is an inverse trend for the age-adjusted can-cer rate in Taiwan, which revealed a steady increase, and reached 270 new cases per 100,000 people in 2007. Since 1982 cancer has become the leading cause of death in the Taiwanese general population[30]. As this issue continues to be a challenge for the public health system in Taiwan, it has come to the attention of the government, thus resulting in population-based investigations Table 1
Baseline characteristics between the sleep disorder group and the non-sleep disorder group in 2000–2001.
Variables Sleep disorder p-value
Total N = 127,053 No N = 84,702 Yes N = 42,351 N (%) n (%) n (%) Sex 1.00 Women 78,237 (61.6) 52,158 (61.6) 26,079 (61.6) Men 48,816 (38.4) 32,544 (38.4) 16,272 (38.4) Age (years) 1.00 <30 19,500 (15.4) 13,000 (15.4) 6500 (15.4) 30–39 22,194 (17.5) 14,796 (17.5) 7398 (17.5) 40–49 26,904 (21.2) 17,936 (21.2) 8968 (21.2) 50–64 30,198 (23.8) 20,132 (23.8) 10,066 (23.8) P65 28,257 (22.2) 18,838 (22.2) 9419 (22.2) Mean (SD)a 49.0 (18.0) 48.9 (18.2) 49.2 (17.7) 0.005 Urbanization level <0.0001 1 37,169 (29.3) 25,514 (30.1) 11,655 (27.5) 2 36,010 (28.3) 24,158 (28.5) 11,852 (28.0) 3 22,370 (17.6) 14,771 (17.4) 7599 (17.9) 4 31,500 (24.8) 20,255 (23.9) 11,245 (26.5) Co-morbidity Hypertension 31,919 (25.1) 17,701 (20.9) 14,218 (33.6) <0.0001 Diabetes mellitus 10,793 (8.5) 6186 (7.3) 4607 (10.9) <0.0001 Hyperlipidemia 15,234 (12.0) 7968 (9.4) 7266 (17.2) <0.0001 Heart disease 23,203 (18.3) 11,448 (13.5) 11,755 (27.8) <0.0001 Chi-square test. a Student t-test. Table 2
Comparisons of incidence density of cancer between the sleep disorder group and the non-sleep disorder group by characteristics.
Variables Sleep disorder RR (95% CI)
No Yes
Cases Person-years Ratea
Cases Person-years Ratea
All 3793 618,434 6.13 2105 310,866 6.77 1.10 (1.05–1.16)** Sex Women 1927 385,710 5.00 1067 195,481 5.46 1.09 (1.01–1.18)* Men 1866 232,724 8.02 1038 115,405 8.99 1.12 (1.04–1.21)* Age (years) <30 42 98,044 0.43 41 49,886 0.82 1.92 (1.25–2.95)* 30–39 207 111,856 1.85 135 56,847 2.37 1.28 (1.03–1.59)* 40–49 506 138,306 3.66 333 69,176 4.81 1.32 (1.15–1.51)** 50–64 1135 152,105 7.46 648 75,427 8.59 1.15 (1.05–1.27)* P65 1903 118,083 16.1 948 59,550 15.9 0.99 (0.91–1.07) Urbanization level 1 1038 188,671 5.50 518 86,725 5.97 1.09 (0.98–1.21) 2 1012 177,557 5.70 547 87,998 6.22 1.09 (0.98–1.21) 3 639 107,305 5.95 353 55,467 6.36 1.07 (0.94–1.22) 4 1104 144,870 7.62 687 80,696 8.51 1.12 (1.02–1.23)*
RR = rate ratio, compared to non-sleep disorder group.
a
Per 1000 person-year.
*p-Value <0.05. ** p-Value <0.001.
regarding cancer preventive epidemiology. The Taiwanese NHI sys-tem health insurance program covers more than 96% of the
popu-lation and maintains contracts with 97% of hospitals and clinics, which has been the case since the end of 1996, so the generalizabil-ity is undoubted. It provides us with a good resource to approach population-based studies. We recently used it as the material to evaluate the risk of malignancy for patients with end-stage renal
disease and published some positive findings[14]. This current
study used a similar design and tried to determine if SD increases the risk of cancer.
In order to create a control group we randomly frequency matched each SD patient with two people from the general
popu-lation without SD but with similar age and same sex.Table 1
re-veals that more SD people live in areas with lower urbanization levels. The possible explanation is that the areas of lower urbaniza-tion in Taiwan normally represent lower socioeconomic status as well as lower incomes and educational levels. Prior papers indi-cated that sleep disorders are independently associated with lower
income and lower educational levels [5,26]. Major medical
illnesses had significantly higher prevalence in the SD group (Table 1). This is unsurprising due to the fact that a significant association exists between sleep disorders and chronic physical diseases (such as diabetes, hypertension, hypercholesterolemia, or other chronic diseases)[24].
Unadjusted analysis found patients with SD had a higher risk of developing cancer overall, and this was also true when we ana-lyzed females and males separately. All age groups except for the oldest age group showed significantly higher cancer risks for the SD cohort. All the urbanization levels showed the same trend of Table 3
Hazard ratios and 95% confidence interval of cancer associated with sleep disorder in Cox’s regression analysis.
Variables Model 1 Model 2 Model 3
HR (95% CI) HR (95% CI) HR (95% CI)
leep disorder
No 1.00 (reference) 1.00 (reference) 1.00 (reference)
Yes 1.11 (1.05– 1.17)** 1.12 (1.05– 1.17)** 1.12 (1.06– 1.18)*** Women Sleep disorder
No 1.00 (reference) 1.00 (reference) 1.00 (reference)
Yes 1.10 (1.02–1.19)* 1.10 (1.02–1.18)* 1.08 (1.00–1.17)* Men
Sleep disorder
No 1.00 (reference) 1.00 (reference) 1.00 (reference)
Yes 1.13 (1.05–1.22)* 1.12 (1.04–1.21)* 1.14 (1.06–1.24)* Model 1: unadjusted.
Model 2: adjusted for age and urbanization.
Model 3: adjusted for age, urbanization, and co-morbidity (included hypertension, diabetes mellitus, hyperlipidemia and heart disease).
*p-Value <0.05. ** p-Value <0.001. ***p-Value <0.0001.
Table 4
Hazard ratios and 95% confidence interval of cancer associated with sleep disorder in Cox’s regression analysis in different cancer.
Variables Sleep disorder HR (95% CI)
No Yes Case % Case % Colorectal cancer 568 0.67 275 0.65 0.97 (0.84–1.13) Liver cancer 432 0.51 354 0.84 1.64 (1.42–1.89)*** Lung cancer 509 0.60 264 0.62 1.07 (0.91–1.24) Melanoma 16 0.02 4 0.01 0.37 (0.12–1.14) Skin cancer 76 0.09 31 0.07 0.83 (0.54–1.27)
Breast cancer (women only) 369 0.71 212 0.81 1.17 (0.98–1.39)
Cervical cancer (women only) 151 0.29 68 0.26 0.96 (0.72–1.29)
Prostate cancer (men only) 167 0.51 104 0.64 1.15 (0.89–1.49)
Brain tumor 46 0.05 17 0.04 0.75 (0.42–1.33)
Other cancers 1459 1.72 776 1.83 1.05 (0.96–1.15)
ICD-9-CM: colorectal cancer, 153.xx and 154.xx; liver cancer, 155.xx; lung cancer, 162.xx; melanoma: 172.xx; skin cancer: 173.xx breast cancer, 174.xx and 175.xx; cervical cancer, 180.xx; prostate cancer, 185.xx; brain cancer, 191.xx.
Adjusted for age, urbanization, and co-morbidity (included hypertension, diabetes mellitus, hyperlipidemia and heart disease).
***p-Value <0.0001.
Table 5
Liver cancer-related co-morbidity between the sleep disorder group and the non-sleep disorder group.
Variables Sleep disorder p-Valuea
No N = 84,702
Yes N = 42,351
n % n %
Alcoholic liver damage 273 0.32 456 1.08 <0.0001
Non-alcoholic fatty liver disease 559 0.66 606 1.43 <0.0001
Hepatitis B 626 0.74 773 1.83 <0.0001
Hepatitis C 332 0.39 468 1.11 <0.0001
Cirrhosis 676 0.80 755 1.78 <0.0001
Alcoholism 163 0.19 346 0.82 <0.0001
Alcoholic liver damage ICD-9-CM: 571.0, 571.1 and 571.3. Non-alcoholic fatty liver disease ICD-9-CM: 571.8.
Hepatitis B ICD-9-CM: V02.61, 070.20, 070.22, 070.30, 070.32. Hepatitis C ICD-9-CM: V02.62, 070.41, 070.44, 070.51, 070.54. Cirrhosis ICD-9-CM: 571.2, 571.5, 571.6.
Alcoholism ICD-9-CM: 303 and 305.0.
a
higher cancer risk within the SD group with a marginally signifi-cant difference. The relatively small sample size caused by separat-ing them into four groups could weaken the statistical power of this finding.
In our Cox-regression model with adjustments for sex, age, urbanization, and co-morbidity (model 3) we found the HR for the overall cancer is 1.12 with a 95% CI of 1.06–1.18. This value is more reliable because we were able to control for many of the possible confounders from the database. The same model (except without the adjustment for sex) was used for women and men and also showed similar findings. Women with SD had a margin-ally significantly higher cancer risk, and men had a significantly higher cancer risk. We validated our hypothesis that SD did increase the risk of cancer, and this novel finding can be partially attributed to the large sample size in this population-based study. One of the possible reasons for a higher cancer risk in the SD cohort could be the pathway of sleep disorders-immunosuppression-cancer. Sleep deprivation leads to the suppression of immune surveillance, which may permit the establishment and growth of malignant clones (e.g., immunosuppression from increased corti-sol, etc.)[7].
Our data showed different directions for various cancers. Higher risks of liver, lung, breast, prostate, and other cancers, and lower risks of brain tumor, melanoma, skin, cervical, and colorectal can-cers were observed. However, only liver cancer showed a signifi-cantly higher risk. One of the possible reasons for higher liver cancer risk is that liver cancer is thought to be a virus-related
malignancy[3,13]; therefore, it may be linked to our hypothesis
of a sleep disorders-immunosuppression-viral infection-cancer pathway. Patients with SD in Taiwan tend to try Chinese herbal medicines and that could have been liver toxic[4]. Primary liver
cancer is the second most common malignancy in Taiwan[30],
and the large number of cases in the study could enforce the statis-tics power. Breast cancer showed a marginally significant increase in risk in the SD group. Epidemiologic studies are now beginning to emerge suggesting that women who work at night, and who expe-rience sleep deprivation, circadian disruption, and exposure to
light-at-night are at an increased risk for breast cancer[6,8]. Wu et al. found a significant decline in postmenopausal breast cancer risk with increasing self-reported hours of sleep among Chinese women in Singapore, and possibly via its effect on melatonin levels
[35]. In addition, the suggestion of lower breast cancer risk in long sleepers in the Finnish Twin Cohort also adds to the body of evi-dence for a possible anticarcinogenic effect of melatonin [32]. In our target group, we collected all patients who fit the ICD9 code for SD and sleep disturbances (except for obstructive sleep apnea in our analysis because it has different causes and biological effects compared to other sleep problems), which included the phase-shift disruption of the 24-h sleep–wake cycle (code number 30745) and disruptions of the 24-h sleep–wake cycle (code number 78055). Although the majority of our patients suffered from sleep disor-ders, we still had eight patients in these two codes (data not shown). Patients with similar symptoms might have been coded into others, which may have diluted our power to detect any statis-tical significance in breast cancer.
One of the strengths of this study is its population-based design due to its generalizability. However, one major limitation that needs to be addressed is that there is no information regarding life style or behavior of the patient on the NHI database, so it is impos-sible to adjust for behavior related factors such as smoking and alcohol consumption. Smokers might have a higher risk of sleep
disorders[18], and smoking might be a confounder between SD
and smoking-related cancers. According to the Surgeon General’s Report, lung, oral, pharynx, larynx, esophagus, kidney, bladder, pancreas, stomach, and cervical cancers are all smoking-related
[22]. Our data did not show any significant difference in these can-cers. However, a study from Taiwan suggested habitual alcohol drinking, betel quid chewing, and cigarette smoking are associated with an increased risk of liver cancer[34]. For alcohol consump-tion, clinical investigations supported a relationship between sleep disturbance and alcohol use, but variability in the definition and measurement of these domains and a preponderance of cross-sectional studies made uncertain the strength and direction of the association[29]. Chronic alcohol use of greater than 80 g/day for more than 10 years increases the risk of liver cancer approxi-mately 5-fold, but alcohol use of less than 80 g/day is not associ-ated with a significant increase in risk[20].Table 5validated the theory that alcoholic liver damage and alcoholism are more fre-quently seen in the SD group, and it may partially explain the sig-nificantly higher risk of liver cancer for patients with SD. However, after controlling the possible confounders,Table 6reveals the ad-justed HR of liver cancer for patients with SD is still significantly
higher, and Table 7does not show a higher risk of liver cancer
for patients with alcoholism.
In conclusion, this population-based retrospective cohort study suggested that SD may associate with an increased risk for the subsequent liver cancer, and possibly breast cancer. The apparent increased alcohol use in the SD group remains a possible con-founder. Further studies with controls for alcohol use and other possible confounders are mandatory to truly uncover the possible link between SD and cancer risk. Nevertheless, the findings from Table 6
Hazard ratios and 95% confidence interval of liver cancer associated with the sleep disorder group and the non-sleep disorder group.
Variables Non-sleep disorder Sleep disorder Crude Adjusted
Cases Cases HR (95% CI) HR (95% CI)
All 432 354 1.63 (1.42–1.88)*** 1.40 (1.21–1.62)***
Female 174 132 1.51 (1.20–1.89)***
1.23 (0.97–1.56)
Male 258 222 1.73 (1.45–2.08)*** 1.54 (1.28–1.87)***
Adjusted for age, urbanization, and co-morbidity (included hypertension, diabetes mellitus, hyperlipidemia, heart disease, alcoholic liver damage, non-alcoholic liver damage, Hepatitis B, Hepatitis C and cirrhosis) and alcoholism.
***p < 0.0001.
Table 7
Hazard and 95% confidence interval of liver cancer-related comorbidity.
Variables N Case HR (95% CI)
Alcoholic liver damage 729 23 4.66 (3.08–7.07)***
Non-alcoholic fatty liver disease 1165 25 3.03 (2.03–4.51)***
Hepatitis B 1399 58 8.68 (6.63–11.4)***
Hepatitis C 800 75 16.4 (12.9–20.8)***
Cirrhosis 1431 145 18.9 (157–22.6)***
Alcoholism 509 5 1.93 (0.80–4.67)
Alcoholic liver damage ICD-9-CM: 571.0, 571.1 and 571.3. Non-alcoholic fatty liver disease ICD-9-CM: 571.8.
Hepatitis B ICD-9-CM: V02.61, 070.20, 070.22, 070.30, 070.32. Hepatitis C ICD-9-CM: V02.62, 070.41, 070.44, 070.51, 070.54. Cirrhosis ICD-9-CM: 571.2, 571.5, 571.6.
HR, Hazard ratio, adjusted for age and sex.
this study may arouse the attention of the NHI of Taiwan to recon-sider the policy regarding follow-up and cancer screening in patients with SD.
Conflict of interest
The ICMJE Uniform Disclosure Form for Potential Conflicts of Interest associated with this article can be viewed by clicking on the following link: doi:10.1016/j.sleep.2012.02.005.
Acknowledgement
This work was supported by the study projects (DMR-101-061 and DMR-101-062) in our hospital and the Taiwan Department of Health Clinical Trial and Research Center for Excellence (DOH101-TD-B-111-004), and the Taiwan Department of Health Cancer Research Center for Excellence (DOH101-TD-C-111-005). Reference
[1] Berger AM. Update on the state of the science: sleep–wake disturbances in adult patients with cancer. Oncol Nurs Forum 2009;4:E165–77.
[2] Bixler EO, Kales A, Soldatos CR, Kales JD, Healey S. Prevalence of sleep disorders in the Los Angeles metropolitan area. Am J Psychiatry 1979;10:1257–62. [3] Chen JD, Yang HI, Iloeje UH, You SL, Lu SN, Wang LY, et al. Risk Evaluation of
Viral Load Elevation and Associated Liver Disease/Cancer in HBV (REVEAL-HBV) Study Group. Carriers of inactive hepatitis B virus are still at risk for hepatocellular carcinoma and liver-related death. Gastroenterology 2010;5:1747–54.
[4] Chen LC, Chen IC, Wang BR, Shao CH. Drug-use pattern of Chinese herbal medicines in insomnia: a 4-year survey in Taiwan. J Clin Pharm Ther 2009;34:555–60.
[5] Cho YW, Shin WC, Yun CH, Hong SB, Kim J, Earley CJ. Epidemiology of insomnia in Korean adults: prevalence and associated factors. J Clin Neurol 2009;1:20–3. [6] Costa G, Haus E, Stevens R. Shift work and cancer – considerations on rationale, mechanisms, and epidemiology. Scand J Work Environ Health 2010;2:163–79. [7] Davidson JR, MacLean AW, Brundage MD, Schulze K. Sleep disturbance in
cancer patients. Soc Sci Med 2002;9:1309–21.
[8] Davis S, Mirick DK. Circadian disruption, shift work and the risk of cancer: a summary of the evidence and studies in Seattle. Cancer Causes Control 2006;4:539–45.
[9] Edwards BK, Ward E, Kohler BA, Eheman C, Zauber AG, Anderson RN, et al. Annual report to the nation on the status of cancer, 1975–2006, featuring colorectal cancer trends and impact of interventions (risk factors, screening, and treatment) to reduce future rates. Cancer 2010;3:544–73.
[10] Fiorentino L, Ancoli-Israel S. Sleep dysfunction in patients with cancer. Curr Treat Options Neurol 2007;9:337–46.
[11] Hu J, Chen Z, Gorczynski CP, Gorczynski LY, Kai Y, Lee L, et al. Sleep-deprived mice show altered cytokine production manifest by perturbations in serum IL-1ra, TNFa, and IL-6 levels. Brain Behav Immun 2003;6:498–504.
[12] Kiecolt-Glaser J. Psychological stress, immune and endocrine function, and cancer risk. American Association for Cancer Research Seventh Annual
International Conference on Frontiers in Cancer Prevention, Research; 2008. p. 16–8.
[13] Lee MH, Yang HI, Lu SN, Jen CL, Yeh SH, Liu CJ, et al. Hepatitis C virus seromarkers and subsequent risk of hepatocellular carcinoma: long-term
predictors from a community-based cohort study. J Clin Oncol
2010;30:4587–93.
[14] Liang JA, Sun LM, Yeh JJ, Lin WY, Chang SN, Sung HC, et al. Malignancies associated with systemic lupus erythematosus in Taiwan: a nationwide population-based cohort study. Rheumatol Int 2012;32:773–8.
[15] Liu CY, Hung YT, Chuang YL, Chen YJ, Weng WS, Liu JS, et al. Incorporating development stratification of Taiwan townships into sampling design of large scale health interview survey. J Health Manag 2006;1:1–22.
[16] Lu JF, Hsiao WC. Does universal health insurance make health care unaffordable? Lessons from Taiwan. Health Aff (Millwood) 2003;3:77–88. [17] Mai E, Buysse DJ. Insomnia: prevalence, impact, pathogenesis, differential
diagnosis, and evaluation. Sleep Med Clin 2008;2:167–74.
[18] Mak KK, Ho SY, Thomas GN, Lo WS, Cheuk DK, Lai YK, et al. Smoking and sleep disorders in Chinese adolescents. Sleep Med 2010;3:268–73.
[19] McClain J. Sleep, physical activity and the prevention of breast cancer. American Association for Cancer Research Seventh Annual International Conference on Frontiers in Cancer Prevention, Research; 2008. p. 16–8. [20] Morgan TR, Mandayam S, Jamal MM. Alcohol and hepatocellular carcinoma.
Gastroenterology 2004;5(Suppl. 1):S87–96.
[21] O’Donnell JF. Insomnia in cancer patients. Clin Cornerstone 2004:S6–S14. [22] Office of the Surgeon General (US), Office on Smoking and Health (US). The
health consequences of smoking: a report of the surgeon general. Atlanta (GA): Centers for Disease Control and Prevention (US); 2004.
[23] Pakkala S, Ramalingam SS. Lung cancer in HIV-positive patients. J Thorac Oncol 2010;11:1864–71.
[24] Paparrigopoulos T, Tzavara C, Theleritis C, Psarros C, Soldatos C, Tountas Y. Insomnia and its correlates in a representative sample of the Greek population. BMC Public Health 2010:531.
[25] Rama I, Grinyo JM. Medscape. Malignancy after renal transplantation: the role of immunosuppression. Nat Rev Nephrol 2010;9:511–9.
[26] Rocha FL, Guerra HL, Lima-Costa MF. Prevalence of insomnia and associated socio-demographic factors in a Brazilian community: the Bambui study. Sleep Med 2002;2:121–6.
[27] Roscoe JA, Kaufman ME, Matteson-Rusby SE, Palesh OG, Ryan JL, Kohli S, et al. Cancer-related fatigue and sleep disorders. Oncologist 2007:35–42. [28] Sateia MJ, Lang BJ. Sleep and cancer: recent developments. Curr Oncol Rep
2008;4:309–18.
[29] Stein MD, Friedmann PD. Disturbed sleep and its relationship to alcohol use. Subst Abus 2005;1:1–13.
[30] Taiwan Cancer Registry. Cancer Statistics, <http://tcr.cph.ntu.edu.tw/ main.php?Page=N2>; [accessed 2010 November].
[31] Thompson CL, Larkin EK, Patel S, Berger NA, Redline S, Li L. Short duration of sleep increases risk of colorectal adenoma. Cancer 2011;4:841–7.
[32] Verkasalo PK, Lillberg K, Stevens RG, Hublin C, Partinen M, Koskenvuo M, et al. Sleep duration and breast cancer: a prospective cohort study. Cancer Res 2005;65:9595–600.
[33] Wallander MA, Johansson S, Ruigomez A, Garcia Rodriguez LA, Jones R. Morbidity associated with sleep disorders in primary care: a longitudinal cohort study. Prim Care Companion J Clin Psychiatry 2007;5:338–45. [34] Wang LY, You SL, Lu SN, Ho HC, Wu MH, Sun CA, et al. Risk of hepatocellular
carcinoma and habits of alcohol drinking, betel quid chewing and cigarette smoking: a cohort of 2416 HBsAg-seropositive and 9421 HBsAg-seronegative male residents in Taiwan. Cancer Causes Control 2003;14:241–50.
[35] Wu AH, Wang R, Koh WP, Stanczyk FZ, Lee HP, Yu MC. Sleep duration, melatonin and breast cancer among Chinese women in Singapore. Carcinogenesis 2008;29:1244–8.
