Long-term use of valproic acid and the prevalence of cancers in bipolar disorder patients in a Taiwanese population: An association analysis using the National Health Insurance Research Database (NHIRD)

Author’s Accepted Manuscript Long-term use of valproic acid and the prevalence of cancers in bipolar disorder patients in a Taiwanese population: an association analysis using the national health insurance research database (NHIRD) Cheng Chia Lin, Tsung-Cheng Hsieh, Lawrence Shih-Hsin Wu

PII: DOI: Reference:

www.elsevier.com/locate/jad

S0165-0327(17)32206-1 https://doi.org/10.1016/j.jad.2018.02.047 JAD9593

To appear in: Journal of Affective Disorders Received date: 24 October 2017 Revised date: 24 January 2018 Accepted date: 16 February 2018 Cite this article as: Cheng Chia Lin, Tsung-Cheng Hsieh and Lawrence ShihHsin Wu, Long-term use of valproic acid and the prevalence of cancers in bipolar disorder patients in a Taiwanese population: an association analysis using the national health insurance research database (NHIRD), Journal of Affective Disorders, https://doi.org/10.1016/j.jad.2018.02.047 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting galley proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Long-term use of valproic acid and the prevalence of cancers in bipolar disorder patients in a Taiwanese population: an association analysis using the national health insurance research database (NHIRD) Cheng Chia Lina, Tsung-Cheng Hsiehb1*, Lawrence Shih-Hsin Wub1 a Department of Urology, Chang Gung Memorial Hospital, Keelung Division, Taiwan b Institute of Medical Sciences, Tzu Chi University, Hualien, Taiwan * Correspondence to: Institute of Medical Sciences, Tzu Chi University, #701, Zhongyang Road, Section 3, Hualien, 97004, Taiwan. Tel.: +886 3 856 5301x2014; Fax: +886 3 857 3053 * Correspondence to: Tel.: +886 3 856-5301x2015; Fax: +886 3 857 3053

Abstract Background: Epigenetic events play a major role in the carcinogenesis of many cancers. A retrospective cohort study had been performed to evaluate the effects of exposure to the anticonvulsant agent valproic acid (VPA), a histone deacetylase inhibitor, on the risk of developing cancers. Methods: The study was based on the 1998 through 2009 National Health Insurance Research Database (NHIRD), provided by the Taiwan National Health Research Institute. Patients with a diagnosis of bipolar disorder (ICD-9-CM codes 296.0, 296.1, 296.4– 8) from 1998 to 2009 were identified. VPA and lithium were the primary index drugs. Patients treated with anticonvulsants who did not use VPA or lithium were selected as the control group. Competing risk regression analysis were used to 1

These authors contributed equally to this work 1

estimate hazards ratios (HR) and 95% confidence intervals (95% CI) reflecting the association between use of VPA and cancer incidence. Results: The cancer incidence of bipolar disorder patients treated with VPA was no significant difference than treated with lithium and other anticonvulsants. In subgroup analysis, VPA associated to higher risk of genitourinary cancer in the duration <1 year group (HR: 3.49; 95%CI: 1.04, 11.67). No significant differences in other cancers incidence in any duration of VPA treatment. Limitations: The cancer prevalence in selected bipolar disorder patients was still low. The sample size was not enough for some types of cancer. Conclusions: A role of VPA in cancer prevention was not found in this study. An increased subgroup risk of genitourinary cancer was observed.

1. Introduction Epigenetic changes, involving either DNA methylation or changes in chromatin structure (Herman and Baylin, 2003; McCabe et al., 2009), are early carcinogenic events in many cancer sites, including the lung (Belinsky et al., 2002, 2006; 2

Palmisano et al., 2000), prostate (Alumkal et al., 2008; Nelson et al., 2007), colon (Brandes et al., 2005; Schuebel et al., 2007), bladder (Hoffman and Cairns, 2011; Kim and Kim, 2009), and head and neck (Wong et al., 2012; Ha and Califano, 2006; Hoque et al., 2008). DNA methyltransferases (DNMTs) and histone deacetylases (HDACs) are major epigenetic mediators for which pharmacologic inhibitors are available. In animal models, inhibition of DNMTs and HDACs has been shown to prevent the development of both lung (Belinsky et al., 2003) and prostate cancers (McCobe et al., 2006). In addition, HDACs 1, 2, and 3 are not only associated with increased DNMT1 protein levels in patients with lung cancer as compared with normal controls, but are also directly responsible for stabilizing DNMT1 expression (Brodie et al., 2013). Valproic acid (VPA), which is widely used among patients with psychiatric or neurologic disorders as a mood stabilizer or antiepileptic drug, has recently been reported to act as a class 1 HDAC inhibitor (Gottlicher et al., 2001). HDAC inhibition is observed at VPA concentrations as low as 40 μg/mL (Brodie et al., 2013). Epigenetic therapies such as the DNMT inhibitor azacytidine and the HDAC inhibitor Vorinostat have proven effective against several hematologic malignancies such as myelodysplastic syndrome (Lubbert et al., 2011; Fenaux et al., 2009; Kantarjian et al., 2006) and cutaneous T-cell lymphomas (Olsen et al., 2007). A recent phase 2 study demonstrated that the combination of 3

azacytidine with HDAC inhibitor Entinostat for the treatment of patients with lung cancer showed promising results (Juergens et al., 2011). However, to the best of our knowledge, there is no clinical evidence to date of an association between the use of HDAC inhibitors and cancer risk. Given the importance of epigenetic mechanisms in early carcinogenesis (Sharma et al., 2010) and preclinical evidence supporting the anticarcinogenic effects of VPA (Duenas-Gonzalez, 2008), we conducted a retrospective cohort study to evaluate the risks of various malignancies in relation to use of VPA in bipolar disorder patients. 2. Methods 2.1. Data source We used a longitudinal health insurance database, the National Health Insurance Research Database (NHIRD), provided by the Taiwan National Health Research Institute. Taiwan launched its compulsory social insurance program, National Health Insurance (NHI), to provide health care for all the island’s residents in 1995. The annual coverage rate of the NHI program has ranged from 96.1% to 99.6%, with more than 20 million Taiwanese residents enrolled since 1997. In the Taiwan NHI system, the government defined a set of diseases, such as mood disorders and cancer, as ‘‘catastrophic illnesses’’, and instituted regulations that enabled insured affected individuals to apply for a catastrophic illness certificate. 4

Patients with catastrophic illness certification receive free care for their illness or related conditions within the certificate’s period of validity. The study was approved by the Institutional Review Board of Chang Kung Memorial Hospital, Keelung branch, Keelung, Taiwan.

2.2. Study cohort Patients with a diagnosis of bipolar disorder (ICD-9-CM codes 296.0, 296.1, 296.4–8) from 1998 to 2009 were identified (WHO, 1978). VPA and lithium were the primary index drugs. To identify a comparison group with the same indications for VPA and lithium, patients treated with anticonvulsants who did not use VPA or lithium were selected as the control group. Patients younger than 18 years of age and those who died due to cancer within 90 days after the index date were excluded. The earliest date of lithium or VPA exposure was defined as the index date. The study cohort was divided into three groups: VPA only, lithium only, and control (patients who had not used VPA or lithium). The end of the observation period was the date of cancer diagnosis or the end of 2010. Cancer diagnoses were identified from the Registry for Catastrophic Illness (ICD-9-CM codes for malignant neoplasm 140–208). Diagnoses were further divided into different types of cancer, including oral cavity (141, 144–147), digestive (150– 5

159), respiratory (160–165), bone, skin, and breast (170–175), genitourinary (179– 189), other and unspecified (190–199), and hemopoietic (200–208).

2.3 Statistical analysis The data elements obtained for each study subject included scrambled social security number, sex, age, first date of encounter, last date of encounter, first VPA prescription filled, and last VPA prescription filled. Propensity score (PS) matching is a statistical matching technique that attempts to estimate the effect of a treatment, policy, or other intervention by accounting for the covariates that predict receipt of treatment (Rosenbaum et al., 1983). We performed PSM using SAS 9.2 in an attempt to reduce bias due to confounding variables present in estimating the treatment effect by simply comparing outcomes among patients that received the treatment versus those that did not. We calculated expected numbers of cancers by multiplying the number of person-years accumulated in each stratum of age, sex, and follow-up duration by the corresponding background specific rate. The incidence rate ratio, calculated as the incidence of VPA use/incidence of non-VPA use of cancer patients, was used as a measure of relative risk, with 95% confidence intervals (CIs) assuming a Poisson distribution of the observed number of cancers. The Competition regression analysis 6

adjusted by the effects of gender, age, and CCI was used to derive age-adjusted cancer hazard ratios (HRs) for patients with bipolar disorder. The statistical analyses were performed using SAS 9.2.

3. Results Data of a total of 22,865 participants who had been diagnosed with bipolar disorder during 1998–2009 were extracted from the longitudinal health insurance database; 1362 patients were excluded due to age <18 or death caused by any cancer within 90 days after the index date. For 3528 patients, there were no records of VPA or lithium treatment; 3416 and 925 patients used VPA only and lithium only after the initial diagnosis of bipolar disorder, respectively. The sampling flow was as shown in Figure 1. The demographic characteristics of the groups are presented in Table 1. The standardized differences were low for each characteristic [gender, age, and Charlson Comorbidity Index (CCI)] between the groups. The propensity score was similar in each comparison group. The incidence and risk of cancers in BP patients in the VPA only/control, VPA/lithium and lithium only/control groups were compared. No statistically significant differences were found in all cancer cases and in cases of the various 7

types of cancer (Table 2). Hazard ratios were generated by competing risk analysis and adjusted for age, gender and CCI. No statistically significant differences were found (Table 3). Furthermore, competing risks analysis of the duration of drug exposure and CCI subgroups was performed. As shown in Table 4, the VPA only group was at higher risk of genitourinary cancer in the duration <1 year group (HR: 3.49; 95%CI: 1.04, 11.67). No statistically significant differences were found in the CCI subgroups.

4. Discussion It is known that HDAC inhibition, such as by VPA, can lead to reduced levels of DNMT1 expression (Du et al., 2010). Class I HDAC-mediated stabilization of DNMT1 protein expression is an early event in smoke carcinogen-induced transformation of bronchial epithelial cells (Brodie et al., 2013). This is associated with uncoupling of DNMT1 expression from the usually tight limitation to the S-phase of the cell cycle, leading to de novo methylation and epigenetic silencing of tumor suppressor genes. Importantly, treatment with VPA partially reverses aberrant DNA methylation, leading to re-expression of previously-silenced genes and suppression of anchorage-independent colony formation. In Taiwan, in the bipolar disorder cohort identified from the NHIRD, the 8

standardized incidence ratio (SIR) for all cancers was 1.29, but excess risk was found in males (SIR: 1.42; 95%CI: 1.14–1.77) and not females (Lin et al., 2013). We hypothesized, based on these data, that VPA may play a role in increasing or decreasing the risks of malignancies in bipolar patients. Another mood stabilizer, lithium, is used to treat bipolar disorder and is associated with a reduced overall cancer risk in bipolar disorder patients (Huang et al., 2016). In this study, bipolar disorder patients under lithium treatment were included as a reference group. In the group with long-term VPA use, there was no significant increase or decrease in the risk of overall cancer or of any type of cancer as compared with the lithium and control groups. In the subgroup analysis, it was found only that the risk of genitourinary cancer was higher in the group with a VPA treatment duration of less than 1 year than in the control group (HR: 3.49; 95%CI: 1.04–11.67). In the CCI>1 subgroup, there was a trend towards an increase in cases of genitourinary cancer (HR: 3.18; 95%CI: 0.84–12.08). In a retrospective cohort study of 439,628 US veterans with indications for routine clinical use of VPA (bipolar disorder, seizure d/o, PTSD, migraines), only the risk of squamous cell carcinomas of the head and neck was significantly reduced in the 26,911 patients with long-term VPA use (HR: 0.66; 95%CI: 0.48–0.92) (Kang et al., 2014). The risks of lung, bladder, prostate and colon cancers were not statistically 9

different between the VPA users and non-users. The lack of effects on the risks of lung, colon and prostate cancer was confirmed in a study of VPA users and non-users from Denmark, in which no significant correlation between VPA use and cancer risk was found (Hallas et al., 2009). A third study based on the UK General Practice Research database found no impact of VPA on the total cancer incidence, but did detect an increase in colorectal cancers (HR: 3.95; 95%CI: 1.97–7.92; P = 0.001) and trends towards an increased risk of prostate cancer (HR: 2.15; 95%CI: 0.92–5.02; P = 0.08) and a decreased risk of breast cancer (HR: 0.40; 95%CI: 0.14– 1.30; P = 0.08) (Singh et al., 2012). There were some limitations in this study. After our stringency selection and performing PS matching, the finally sample size was small to investigate the prevalence of cancers. We did not have detailed information regarding patients’ life style related to cancer developing, such as smoking and drinking. Also, the detailed information of co-treated non-psychotic drugs was incomplete to analyze. In conclusion, a role of VPA in cancer prevention was not found in this study. An increased subgroup risk of genitourinary cancer (including prostate cancer) was observed, making an uneven distribution of risk factors the most likely explanation for this phenomenon.

10

Ethical considerations This research was conducted in accordance with the Helsinki Declaration. Ethical approval for this study was obtained from the Research Ethic Committee of Chang Gung Memorial Hospital, Keelung division.

Data sharing statement All register information was provided by the Taiwan National Health Research Institute. Information about how to apply for data from Taiwan health care registers is available at http://www.nhri.org.tw. Role of funding source None Acknowledgements None Contributors Cheng-Chia Lin, MD, Tsung-Cheng Hsieh, Ph.D., Lawrence Shih-Hsin Wu, Ph.D. Conflicts of interest No conflicts of interest.

References Alumkal, J.J., Zhang, Z., Humphreys, E.B., Bennett, C., Mangold, L.A., Carducci, 11

M.A., Partin, A.W., Garrett-Mayer, E., DeMarzo, A.M., Herman, J.G., 2008. Effect of DNA methylation on identification of aggressive prostate cancer. Urology 72, 1234-1239. Belinsky, S.A., Klinge, D.M., Stidley, C.A., Issa, J.P., Herman, J.G., March, T.H., Baylin, S.B., 2003. Inhibition of DNA methylation and histone deacetylation prevents murine lung cancer. Cancer Res. 63, 7089-7093. Belinsky, S.A., Liechty, K.C., Gentry, F.D., Wolf, H.J., Rogers, J., Vu, K., Haney, J., Kennedy, T.C., Hirsch, F.R., Miller, Y., Franklin, W.A., Herman, J.G., Baylin, S.B., Bunn, P.A., Byers, T., 2006. Promoter hypermethylation of multiple genes in sputum precedes lung cancer incidence in a high-risk cohort. Cancer Res. 66, 3338-3344. Belinsky, S.A., Palmisano, W.A., Gilliland, F.D., Crooks, L.A., Divine, K.K., Winters, S.A., Grimes, M.J., Harms, H.J., Tellez, C.S., Smith, T.M., Moots, P.P., Lechner, J.F., Stidley, C.A., Crowell, R.E., 2002. Aberrant promoter methylation in bronchial epithelium and sputum from current and former smokers. Cancer Res. 62, 2370-2377. Brandes, J.C., van Engeland, M., Wouters, K.A., Weijenberg, M.P., Herman, J.G., 2005. CHFR promoter hypermethylation in colon cancer correlates with the microsatellite instability phenotype. Carcinogenesis 26, 1152-1156. 12

Brodie, S.A., Li, G., El-Kommos, A., Kang, H., Ramalingam, S.S., Behera, M., Gandhi, K., Kowalski, J., Sica ,G.L., Khuri, F.R., Vertino, P.M., Brandes, J.C., 2014. Class I HDACs are mediators of smoke carcinogen-induced stabilization of DNMT1 and serve as promising targets for chemoprevention of lung cancer. Cancer Prev. Res. (Philadelphia, Pa.) 7: 351-361. Brodie, S.A., Li, G., El-Kommos, A., Kang, H., Ramalingam, S.S., Behera, M., Gandhi, K., Kowalski, J., Sica, G.L., Khuri, F.R., Vertino, P.M., Brandes, J.C., 2014. Class I HDACs are mediators of smoke carcinogen induced stabilization of DNMT1 and serve as promising targets for chemoprevention of lung cancer. Cancer Prev. Res. (Philadelphia, Pa.) 7: 351–361. Du, Z., Song, J., Wang, Y., Zhao, Y., Guda, K., Yang, S., Kao, H.Y., Xu, Y., Willis, J., Markowitz, S.D., Sedwick, D., Ewing, R.M., Wang, Z., 2010. DNMT1 stability is regulated by proteins coordinating deubiquitination and acetylation-driven ubiquitination. Sci. Signal. 3:ra80. Duenas-Gonzalez, A., Candelaria, M., Perez-Plascencia, C., Perez- Cardenas, E., de la Cruz-Hernandez, E., Herrera, L.A., 2008. Valproic acid as epigenetic cancer drug: preclinical, clinical and transcriptional effects on solid tumors. Cancer Treat. Rev. 34: 206-222. Fenaux, P., Mufti, G.J., Hellstrom-Lindberg, E., Santini, V., Finelli, C., Giagounidis, 13

A., Schoch, R., Gattermann, N., Sanz, G., List, A., Gore, S.D., Seymour, J.F., Bennett, J.M., Byrd, J., Backstrom, J., Zimmerman, L., McKenzie, D., Beach, C., Silverman, L.R., International Vidaza High-Risk MDS Survival Study Group, 2009. Efficacy of azacitidine compared with that of conventional care regimens in the treatment of higher-risk myelodysplastic syndromes: a randomised, open-label, phase III study. Lancet Oncol. 10:223-232. Gottlicher, M., Minucci, S., Zhu, P., Krämer, O.H., Schimpf, A., Giavara, S., Sleeman, J.P., Lo, Coco F., Nervi, C., Pelicci, P.G., Heinzel. T., 2001. Valproic acid defines a novel class of HDAC inhibitors inducing differentiation of transformed cells. EMBO J. 20: 6969-6978. Ha, P.K., Califano, J.A., 2006. Promoter methylation and inactivation of tumour-suppressor genes in oral squamous-cell carcinoma. Lancet Oncol. 7, 77-82. Hallas, J., Friis, S., Bjerrum, L., Stovring, H., Narverud, S.F., Heyerdahl, T., Gronbaek, K., Andersen, M., 2009. Cancer risk in long-term users of valproate: a population-based case-control study. Cancer Epidemiol. Biomarkers Prev. 18: 1714–1719. Herman, J.G., Baylin, S.B., 2003. Gene silencing in cancer in association with promoter hypermethylation. N. Engl. J. Med. 349, 2042-2054. 14

Hoffman, A.M., Cairns, P., 2011. Epigenetics of kidney cancer and bladder cancer. Epigenomics 3, 19-34. Hoque, M.O., Kim, M.S., Ostrow, K.L., Liu, J., Wisman, G.B., Park, H.L., Poeta, M.L., Jeronimo, C., Henrique, R., Lendvai, A., Schuuring, E., Begum, S., Rosenbaum, E., Ongenaert, M., Yamashita, K., Califano, J., Westra, W., van der Zee, A.G., Van Criekinge, W., Sidransky, D., 2008. Genome-wide promoter analysis uncovers portions of the cancer methylome. Cancer Res. 68, 2661-2670. Huang, R.Y., Hsieh, K.P., Huang, W.W., Yang, Y.H., 2016. Use of lithium and cancer risk in patients with bipolar disorder: population-based cohort study. British J. Psych. 209, 393-399. Juergens, R.A., Wrangle, J., Vendetti, F.P., Murphy, S.C., Zhao, M., Coleman, B., Sebree, R., Rodgers, K., Hooker, C.M., Franco, N., Lee, B., Tsai, S., Delgado, I.E., Rudek, M.A., Belinsky, S.A., Herman, J.G., Baylin, S.B., Brock, M.V., Rudin, C.M., 2011. Combination epigenetic therapy has efficacy in patients with refractory advanced non-small cell lung cancer. Cancer Discov. 1, 598-607. Kang, H., Gillespie, T.W., Goodman, M., Brodie, S.A., Brandes, M., Ribeiro, M., Ramalingam, S.S., Shin, D.M., Khuri, F.R., Brandes, J.C., 2014. Long-term use of valproic acid in US veterans is associated with a reduced risk of smoking-related cases of head and neck cancer. Cancer. 120, 1394–1400. 15

Kantarjian, H., Issa, J.P., Rosenfeld, C.S., Bennett, J.M., Albitar, M., DiPersio, J., Klimek, V., Slack, J., de Castro, C., Ravandi, F., Helmer, R. 3rd, Shen, L., Nimer, S.D., Leavitt, R., Raza, A., Saba, H., 2006. Decitabine improves patient outcomes in myelodysplastic syndromes: results of a phase III randomized study. Cancer 106, 1794-1803. Kim, Y.K., Kim, W.J., 2009. Epigenetic markers as promising prognosticators for bladder cancer. Inter. J. Urol. 16,17-22. Lin, G.M., Chen, Y.J., Kuo, D.J., Jaiteh, L.E., Wu, Y.C., Lo, T.S., Li, Y.H., 2013. Cancer incidence in patients with schizophrenia or bipolar disorder: a nationwide population-based study in Taiwan, 1997–2009. Schizophr. Bull. 39, 407-416. Lubbert, M., Suciu, S., Baila, L., Rüter, B.H., Platzbecker, U., Giagounidis, A., Selleslag, D., Labar, B., Germing, U., Salih, H.R., Beeldens, F., Muus, P., Pflüger, K.H., Coens, C., Hagemeijer, A., Eckart Schaefer, H, Ganser, A., Aul, C., de Witte, T., Wijermans, P.W., 2011. Low-dose decitabine versus best supportive care in elderly patients with intermediate- or high-risk myelodysplastic syndrome (MDS) ineligible for intensive chemotherapy: final results of the randomized phase III study of the European Organisation for Research and Treatment of Cancer Leukemia Group and the German MDS Study Group. J. Clin. Oncol. 29, 1987-1996. 16

McCabe, M.T., Brandes, J.C., Vertino, P.M., 2009. Cancer DNA methylation: molecular mechanisms and clinical implications. Clin. Cancer Res. 15, 3927-3937. McCabe, M.T., Low, J.A., Daignault, S., Imperiale, M.J., Wojno, K.J., Day, M.L., 2006. Inhibition of DNA methyltransferase activity prevents tumorigenesis in a mouse model of prostate cancer. Cancer Res. 66: 385-392. Nelson, W.G., Yegnasubramanian, S., Agoston, A.T., Bastian, P.J., Lee, B.H., Nakayama, M., De Marzo, A.M., 2007. Abnormal DNA methylation, epigenetics, and prostate cancer. Front. Biosci. 12, 4254-4266. Olsen, E.A., Kim, Y.H., Kuzel, T.M., Pacheco, T.R., Foss, F.M., Parker, S., Frankel, S.R., Chen, C., Ricker, J.L., Arduino, J.M., Duvic, M., 2007. Phase IIb multicenter trial of vorinostat in patients with persistent, progressive, or treatment refractory cutaneous T-cell lymphoma. J. Clin. Oncol. 25, 3109-3115. Palmisano, W.A., Divine, K.K., Saccomanno, G., Gilliland, F.D., Baylin, S.B., Herman, J.G., Belinsky, S.A., 2000 Predicting lung cancer by detecting aberrant promoter methylation in sputum. Cancer Res. 60, 5954-5958. Rosenbaum, P.R., Rubin, D.B., 1983. The central role of the propensity score in observational studies for causal effects. Biometrika 70, 41–55. Schuebel, K.E., Chen, W., Cope, L., Glöckner, S.C., Suzuki, H., Yi, J.M., Chan, T.A., 17

Van Neste, L., Van Criekinge, W., van den Bosch, S., van Engeland, M., Ting, A.H., Jair, K., Yu, W., Toyota, M., Imai, K., Ahuja, N., Herman, J.G., Baylin, S.B., 2007. Comparing the DNA hypermethylome with gene mutations in human colorectal cancer. PLoS Genet. 3, 1709-1723. Sharma, S., Kelly, T.K., Jones, P.A., 2010. Epigenetics in cancer. Carcinogenesis 31,27-36. Singh, G., Bell, G.S., Driever, P.H., Sander, J.W., 2012. Cancer risk in people with epilepsy using valproate sodium. Acta. Neurol. Scand. 125, 234–240. Wong, T.S., Gao, W., Li, Z.H., Chan, J.Y., Ho, W.K., 2012. Epigenetic dysregulation in laryngeal squamous cell carcinoma. J. Oncol. 2012, 739461. World Health Organization, 1978. International Statistical Classification of Diseases and Related Health Problem (ICD-9). WHO.

Figure 1 Flowchart of data analysis

Table 1 Demographic characteristics of the study groups. VPA

w/t

(N=26 VPA 63) (N=2 663) Mean (std)/

Mean (std)/

Standar

VPA

Lithi

dized (N=1 differen 820) ce

Standar

Lithi

w/t

um dized um Lithi (N=9 differen (N=9 um 25) ce 25) (N=1 850)

Mean (std)/

Mea n 18

Mea n

Mean (std)/

Standar dized differen ce

N (%)

N

N

(std)/

(std)/ n(%)

(%)

(%)

N (%)

n (%)

Male: Gende 1108 r (41.6 %) Age 43.0 (years (14.5) ) 200 Age≥ ( 7.5 65 %)

Male: 1097 (41.2 %)

865 (46.8 %)

439 0.01 (47.5 %)

439 857 (47.5 (46.3 %) %)

CCI

1.53

1.55

AVG

(1.76)

(1.71)

CCI> 1

1045 (39.2 %)

1096 (41.2 %)

-0.01

43.0 0 (13.9) 205 ( 7.7 %)

Durati on of n (%) drug

0.01

0.01

0.04

-

-0.02

41.8 41.5 -0.02 (15.0) (14.1 ) 132 49 -0.08

41.5 41.4 -0.01 (14.1 (13.9) ) 49 96 0

( 7.1 %) 1.34

( 5.3 %) 1.34

( 5.2 %) 1.32

(1.65) (1.58 ) 644 334 0.03 (34.8 (36.1 %) %) 374 202 0.05 (20.2 (21.8 %) %)

(1.58 ) 334 (36.1 %) n (%)

(1.56)

326 (17.6 %) 386 (20.9 %) 764 (41.3 %)

369 (39.9 %) 185 (20.0 %) 371 (40.1 %)

( 5.3 %) 1.34

0

660 (35.7 %)

-0.01

-0.01

-

use D <1yr 1≤ D< 3 yrs D≥3 yr Prope nsity score

929 (34.9 %) 583 (21.9 %) 1151 (43.2 %) 0.50 (0.12)

-

-

-

0.50 -0.01 (0.12)

167 (18.1 %) 185 (20.0 %) 371 (40.1 %)

0.22 0.22 0 (0.04) (0.04 )

19

0.24 0.24 0 (0.07 (0.07) )

Table 2 Incidences and risks of cancers in BP patients treated with VPA and lithium.

Overall cancer Oral cavity

VPA (N=26 63)

w/t VPA (N=26 63)

n

n

RR (95% CI)

n

n

RR (95% CI)

n

66

76

1.0 (0.7-

35

26

0.8 (0.5-

26

8

14

Digestive 26

32

Respirato 5 ry

7

Bone,

8

13

skin, and breast Genitouri 10 nary Other and unspecifi ed Hemopoi

6

5

7

0

5

VPA Lithi (N=18 um 20) (N=9 25)

1.3) 0.6 (0.31.5) 0.9 (0.51.5) 0.8 (0.32.5) 1.8 (0.74.3) 1.8 (0.75.1) 0.8 (0.32.5) 0.0

4

3

19

12

1

2

6

5

2

3

1

2

1

0

etic RR: relative risk; CI: confidence interval.

20

Lithi um (N=9 25)

1.4) 0.8 (0.23.7) 1.0 (0.52.0) 0.3 (0.03.4) 0.7 (0.22.4) 0.4 (0.12.5) 0.3 (0.03.4) -

3

w/t Lithiu m (N=18 50) n RR (95% CI) 48

9

12

19

2

6

5

4

3

4

2

6

0

4

1.0 (0.61.6) 0.6 (0.22.2) 1.1 (0.52.3) 0.6 (0.13.0) 2.2 (0.68.3) 1.3 (0.36.0) 0.6 (0.13.0) 0.0

Table 3 Hazard ratios of cancers generated by competing risk analysis in BP patients treated with VPA and lithium.

Overall cancer Oral cavity

Digestive

Respiratory

Bone, skin, and breast

VPA vs. w/t VPA

VPA vs. Lithium

Lithium vs. w/t Lithium

Adjusted HR (95% CI)

P value

Adjusted HR (95% CI)

P value

Adjusted HR (95% CI)

P value

1.04 ( 0.74, 1.45) 0.69

0.831

0.93 ( 0.56, 1.56) 0.99

0.792

0.98 ( 0.61, 1.59) 0.61

0.938

( 0.29, 1.67) 0.91 ( 0.54, 1.56) 0.79 ( 0.24, 2.63) 2.09 ( 0.87,

5.03) Genitourinary 2.10 ( 0.80, 5.48) Other and 0.79 unspecified ( 0.24, 2.63) Hemopoietic

0.414

0.740

0.705

0.102

0.131

0.705

-

( 0.23, 4.32) 1.04 ( 0.50, 2.15) 0.27 ( 0.02, 2.90) 1.02 ( 0.34, 3.04) 0.37 ( 0.06, 2.17) 0.27 ( 0.02, 2.90) -

0.992

0.920

0.277

0.973

0.271

0.277

-

( 0.17, 2.23) 1.19 ( 0.57, 2.47) 0.65 ( 0.13, 3.27) 1.98 ( 0.56, 7.03) 1.36 ( 0.29, 6.29) 0.65 ( 0.13, 3.27) -

0.452

0.643

0.598

0.292

0.697

0.598

-

HR: hazard ratio; CI: confidence interval. Table 4 Competing risk analysis of subgroups with differing durations of VPA treatment and Charlson Comorbidity Index (CCI). Adjusted hazard ratio (95% CI)

Overall

Duration < 1 year

1year ≤Duration<3 years

Duration≥3 years

CCI≤ 1

CCI >1

0.83 (0.48,

0.95 (0.47,

1.17 (0.80,

0.81 (0.46,

1.20 (0.80,

21

cancer

1.42)

1.90)

1.71)

1.45)

1.82)

Oral cavity

N/A

1.05 (0.23, 4.72) 0.71 (0.22, 2.36) N/A

1.01 (0.39, 2.65) 0.99 (0.54, 1.81) 1.19 (0.35, 4.09) 2.45 (0.94, 6.40) 1.62 (0.45, 5.82)

0.51 (0.13, 1.95) 1.08 (0.35, 3.37) 0.52 (0.09, 2.85) 1.70 (0.48, 6.07) 0.88 (0.14, 5.48)

0.88 (0.27, 2.88) 0.92 (0.51, 1.67) 1.24 (0.25, 6.17) 2.72 (0.78, 9.50) 3.18 (0.84, 12.08)

1.19 (0.35, 4.09)

0.52 (0.09, 2.85)

1.24 (0.25, 6.17)

Digestive

0.80 (0.35, 1.83) Respiratory 0.49 (0.06, 4.03) Bone, skin, 1.22 (0.25, and breast 5.84) Genitourinary 3.49 (1.04, 11.67)* Other and unspecified

0.49 (0.06, 4.03)

2.35 (0.48, 11.58) 2.21 (0.25, 19.85) N/A

N/A: data were not analyzed due to very limited numbers of cases; * P<0.05. CCI: Charlson Comorbidity Index; CI: confidence interval. Highlights  Epigenetic events play a major role in the carcinogenesis of many cancers.  VPA associated to higher risk of genitourinary cancer in the treatment duration <1 year in bipolar disorder.  A role of VPA in cancer prevention was not found in bipolar disorder patients.

22

23