Association Between Viral Hepatitis Infection and Parkinson's Disease

A Population-based Prospective Study

Hwa-Young Choi; Thi Ha Mai; Kyung-Ah Kim; Hyunsoon Cho; Moran Ki


J Viral Hepat. 2020;27(11):1171-1178. 

In This Article


Data Source

We used health insurance claims data provided by the Korea National Health Insurance Service (NHIS). The data included details of medical treatment (diagnosis, length of stay, treatment costs, services received, drug code, days prescribed and daily dosage) prescriptions and socio-economic (income-based insurance contributions, demographic variables and date of death characteristics) of inpatient and outpatient who joined the National Health Insurance (NHI) system.[9] In Korea, NHI is mandatory; it covered approximately 98% of the Korean population in 2015.[10] We used both the NHI data for all Korean populations to collect information about patients with viral hepatitis, and data from the NHIS-National Sample Cohort (NSC) to select participants without viral hepatitis. The NHIS-NSC cohort database contains data of 1 020 340 individuals (a 2.2% sample) stratified by sex, age, and income level and selected by random sampling from the 2002 target population of 4 6605 433 individuals. The selected individuals were followed for 11 years until 2013.[11] Based on the current National Health Insurance Act, these data were designed for research purposes only; therefore, participants' consent was not required. However, identification of participants was difficult due to the large and population-dwelling sample size.

Information about the diagnosis of hepatitis was based on health insurance claim data. However, as information about individuals without hepatitis was not available from the claim data, we used data from the sample cohort database for the group with individuals without hepatitis (Figure 1).

Figure 1.

Flow chart of study subject selection. aHBV (ICD-10 : B18.0, B18.1), HCV (ICD-10 : B18.2), bSecondary parkinsonism (ICD-10 : G21), Parkinsonism in diseases classified elsewhere (ICD-10 : G22), Other degenerative diseases of basal ganglia (ICD-10 : G23) and Other extrapyramidal and movement disorders (ICD-10 : G25)

Selection of Participants

Patients who were newly diagnosed with chronic HBV infection (International Classification of Diseases, tenth edition [ICD-10]: B18.0 or B18.1), chronic HCV infection (ICD-10: B18.2) and co-infection (HBV/HCV) from 2005 to 2014 were selected as participants for this study. We excluded the following patients: (a) those who were <40 years of age and (b) those who had been diagnosed with PD (ICD-10: G20) or any Parkinson's-related diseases (secondary parkinsonism [ICD-10: G2], parkinsonism in diseases classified elsewhere [ICD-10: G22], other degenerative diseases of the basal ganglia [ICD-10: G23], and other extrapyramidal and movement disorders [ICD-10: G25]) before being diagnosed with HBV and HCV infections.

For the group without viral hepatitis, we used NHIS-NSC claims data from 2002 to 2013. The following patients were excluded: (a) those who were <40 years of age in 2003, (b) those with unknown information on age and sex in 2003, (c) those who had been diagnosed with acute or chronic HBV or HCV infection in 2002–2013, and (d) those who had been diagnosed with any PD or Parkinson's-related diseases in 2002.

End Point and Comorbidity

The study end point at follow-up was PD diagnosis, death or censored (31 December 2015 for the group with viral hepatitis and 31 December 2013 for the group without viral hepatitis), whichever was first. The definition of PD was defined as the patients who were prescribed antiparkinsonism drugs among those diagnosed with ICD-10 code (G20). Comorbidities were defined with more than one admission or two outpatient visits between 2002 and the end point based on ICD-10 codes and comprised hyperlipidemia (E78), hypertension (I10-I15), ischaemic heart disease (I20-I25), epilepsy (G40), type 2 diabetes mellitus (E11), liver cancer (C22, C22.0 and C22.9), stroke (I60-I69), head injury (S00-S09), alcoholic liver disease (K70) and depressive disorder (F32). We included these comorbidities in the analysis because they are potential risk factors for PD.

Statistical Analyses

The incidence density of PD per 10 000 person-years was calculated as follows: The numerator was the number of newly diagnosed cases of PD during the observation period, and the denominator was the sum of the observation periods for all the individuals in the group. To assess the risk of PD, we estimated the cumulative incidence and hazard ratios for adjusted age, sex, comorbidity and death. The Fine and Gray competing risks regression models were used to formally analyse the association between risk factors and PD, considering that all-cause mortality was a competing risk.[12] The cumulative incidence rate in the HBV/HCV group was not calculated due to the small number of patients. Statistical significance was defined as P values < .05, and all statistical analyses were performed using SAS version 9.4 (SAS Institute Inc.).