Direct-acting antiviral treatment in real world patients with hepatitis C not associated with psychiatric side effects: a prospective observational study

This study was initiated in 2013, shortly before new DAA treatments were introduced, with the aim to study the longitudinal relationship between psychiatric symptoms and HCV treatment. Patients were recruited from a group of patients treated by a specialist in infectious disease (AL) at the Department of Infectious Diseases, Uppsala University Hospital. The eligible patients for DAA treatment were those without malignancy. Patients with advanced liver fibrosis and cirrhosis, extrahepatic diseases, as well as those with strong psychosocial reasons for treatment were prioritised for inclusion.

Exclusion criteria were inability to read or write Swedish and low cognitive or intellectual ability. These criteria were considered incompatible with completing an extensive research protocol and undergoing an in-depth inquiry with recurrent interviews. Sixty-three consecutive patients were considered for participation in this study between June 2014 and April 2015. Eleven were excluded because of a lack of Swedish language skills or low cognitive or intellectual functioning. This selection was made by the treating specialist (AL) at the Department of Infectious Diseases, Uppsala University Hospital, based on clinical judgement of the patient’s ability to understand and complete the study protocol with repeated assessments in the Swedish language. Of the 52 eligible patients, 19 (37%) accepted and were included in the study. Two of those patients declined to start the study before the first visit. Thus, the final sample was composed of 17/52 patients (33%). The patients represented the spectrum of individuals considered suitable for treatment of HCV. None of the patients had HBV-HCV or HIV-HCV co-infection. A flow chart of the patient inclusion process is provided in Fig. 1.

The study was a prospective observational study. The choice of treatment regimen was based on actual Swedish national recommendations by The Swedish Medical Products Agency at the time of treatment [33]. The prescription was combined with a treatment schedule that was carefully supervised by research nurses at the outpatient ward. Before treatment start, the physician reviewed laboratory data, liver elasticity, biopsy data and concomitant medication. A visit was scheduled every four weeks during treatment. After the treatment was completed, follow-up visits were performed at 12 and 24 weeks. At each time point, research nurses asked the participants about adherence to medication (taking DAA medication as prescribed or not), which was then noted in the medical records. Psychiatric assessment and patient self-report measures were employed at baseline, after 4 weeks of treatment, after 8 weeks of treatment, at the end of treatment and 12 weeks post-treatment.

Information was gathered on sociodemographic data, medical history and treatment from the patients and medical records. Patients and providers were blinded to the HCV RNA results at the time of completion of the questionnaires, but not blinded to the HCV RNA results from earlier time points in the study.

At baseline, patients were assessed for past and current psychiatric morbidity by a trained psychiatrist (IS) using the Structured Clinical Interview for DSM-IV Axis I Disorders (SCID-I) clinical version. Previous HCV treatment and psychiatric side effects were also addressed in the interview. In following visits (including the 12-week post-treatment visit), patients were interviewed by the principal author, using module A of the SCID-I to assess the presence of depressive episodes according to DSM-IV.

Symptoms of depression were measured using the self-rating version of the Montgomery Åsberg Depression Rating Scale (MADRS-S). It has been shown to be a reliable and sensitive self-report tool for depressive symptoms [34] and thus suitable to follow patients with depressive symptoms over time. Scores on the MADRS-S range from 0 to 54, with higher scores indicating a greater severity of depression [35]. Sleep quality was measured with the Pittsburgh Sleep Quality Index (PSQI), a validated self-rated questionnaire assessing sleep quality and disturbances over a one month period [36]. The PSQI measures several aspects of sleep (e.g., sleep latency, sleep duration, habitual sleep efficiency and sleep disturbances). This tool includes seven component scores (range 0–3), as well as a composite global score (range 0–21). At each visit, patients filled in the MADRS-S and PSQI questionnaires. To screen for alcohol and drug use, the Alcohol Use Disorders Identification Test (AUDIT) [37] and Drug Use Disorders Identification Test (DUDIT) [38] were completed at baseline, after 12 weeks of treatment and 12 weeks post-treatment. The AUDIT is a 10-item screening instrument to assess alcohol consumption, drinking behavior and problems related to alcohol consumption. The DUDIT serves as a parallel tool to the AUDIT to identify persons with drug-related problems.

HCV RNA was obtained from blood samples and analysed at each time point. All analyses were conducted at the Department of Clinical Chemistry at Uppsala University Hospital. A negative test for HCV RNA (detection limit of ˂15 IU/mL) at both 12 and 24 weeks post-treatment was considered as a SVR. The rate of the SVR was calculated on an “intention to treat” basis – i.e. all the patients who started treatment.

Non-parametric tests were used to calculate differences in MADRS-S, PSQI, AUDIT and DUDIT over time (Friedman’s test) and between depressive symptoms and HCV viral load at baseline (Spearman’s rank correlation test). In a post-hoc analysis Wilcoxon’s test was applied to compare baseline depressive symptoms with post-treatment depressive symptoms. All statistical analyses were conducted with the Statistical Package for the Social Sciences version 2 (SPSS Inc., Chicago, IL, USA). Reported p-values are two-sided with statistical significance set at p < 0.05.

At 95% of the visits, patients reported taking their DAA medication as prescribed. Fifteen of 17 (88%) patients reached SVR at both 12 and 24 weeks. One male patient had a virological relapse. A female patient was diagnosed with lung cancer during the treatment; this individual had a negative HCV test (< 15 IU/mL) at 4 and 8 weeks of treatment and discontinued treatment at week 11.

One patient attained SVR but developed hepatocellular cancer shortly after completing DAA treatment. Ultrasound of the liver and computer tomography of the upper abdomen was performed before DAA treatment for this patient showing no indication of malignancy.

According to the SCID-I, 15/17 patients (88%) had any lifetime DSM IV psychiatric diagnosis. At baseline, 6/17 patients (35%) had an ongoing DSM IV psychiatric diagnosis. The number of lifetime psychiatric diagnoses ranged from 0 to 7 (median 2). Eleven of 17 patients (65%) had previous abuse or dependence of drugs or alcohol. Eleven of 17 patients (65%) were previously treated with IFN for HCV but the treatment had not been successful because of relapse (n = 3), intolerance (n = 3), partial response (n = 3) and viral breakthrough (n = 2). Retrospectively, at least two of the previously treated patients had developed a depressive episode during treatment (according to the psychiatric assessment with SCID-I) while another four had developed depressive symptoms. One patient had developed diabetes, which is a reported side effect of IFN/RBV treatment [39‐41].

There was no significant correlation between baseline symptoms of depression and HCV viral load at baseline (Spearman’s rho: r = 0.17, p = 0.55).

MADRS-S scores for each visit were available for 15/17 patients (one patient was lost to follow-up because of lung cancer and one missed filling in the MADRS-S at one of the assessment visits). For these 15 remaining patients, the mean MADRS-S score was 10.7 (range 1–30, SD 7.9) at baseline, 7.2 (range 0–26, SD 7.1) after 4 weeks, 7.3 (range 0–19, SD 5.7) after 8 weeks, 8.0 (range 0–25, SD 7.4) at the end of treatment and 8.3 (range 0–30, SD 8.3) at 12 weeks post-treatment (Fig. 2).

No statistical difference in MADRS-S levels was found when comparing all test points (Friedman test Χ ²  = 4.03, p = 0.40), indicating that depressive symptoms did not increase during treatment. We observed a significant difference when comparing baseline MADRS-S with 12-week post-treatment (Wilcoxon p = 0.01), i.e. there were less depressive symptoms post-treatment.

Reanalysis, omitting those three patients taking antidepressant medication again found no statistical difference in MADRS-S levels when comparing all test points (Friedman test Χ ²  = 4.6, p = 0.33) and a significant difference was still seen when comparing baseline MADRS-S with 12 weeks after treatment (Wilcoxon p = 0.02).

In one patient adjunctive RBV treatment was discontinued after four weeks because of affected blood count (from 146 to 115 g/L). For this patient, the MADRS-S score increased from 8 at baseline to 15 after four weeks and then decreased to 11, 9 and 8 during the rest of the study.

Three patients were depressed at baseline (based on the psychiatric assessment with the SCID-I). Two patients developed depression during DAA treatment. Of the two patients who developed depression, one described a subthreshold depressive episode at baseline (days with depressed mood but not during two weeks) and the other experienced family psychosocial problems during the treatment period. The latter patient also had a history of recurrent depressive episodes. One patient was depressed 12 weeks after treatment concluded.

Mean total PSQI scores were 7.7 (range 2–19, SD 4.8) at baseline, 7.6 (range 2–18, SD 4.7) after 4 weeks, 7.7 (range 2–19, SD 4.9) after 8 weeks, 7.2 (range 2–19, SD 4.7) at the end of treatment and 7.6 (range 3–18, SD 4.9) 12 weeks after treatment ended (Fig. 2). No significant variation in total PSQI was noted (Χ ²  = 3.4, p = 0.49). For PSQI component scores, post-hoc analysis of these measures revealed no significant variation during DAA treatment for sleep duration (Χ ²  = 2.4, p = 0.66), sleep disturbances (Χ ²  = 1.8, p = 0.77), sleep latency (Χ ²  = 2.7, p = 0.61), day dysfunction (Χ ²  = 5.5, p = 0.24), habitual sleep efficiency (Χ ²  = 2.4, p = 0.67), and overall sleep quality (Χ ²  = 3.4, p = 0.50) (Fig. 3).

Mean AUDIT scores were 1.9 (range 0–7, SD 2.2) at baseline, 1.5 (range 0–12, SD 3.0) 12 weeks of treatment and 1.9 (range 0–9, SD 2.5) 12 weeks post-treatment. There was no significant variation during treatment (Χ ²  = 1.15, p = 0.59).

Mean DUDIT scores were 2.8 (range 0–41, SD 10.0) at baseline, 2.3 (range 0–28, SD 7.0) after 12 weeks of treatment and 0.6 (range 0–7, SD 2.0) 12 weeks after treatment. One person with previous methamphetamine abuse was taken into a substance abuse treatment centre just before start of the study (and start of treatment). This patient’s DUDIT scores were 41 at baseline, 28 after 12 weeks of treatment and 7 12 weeks post-treatment. She did not report drug abuse during the study, but because the questions in DUDIT refer to “during the last year”, high scores can be obtained without ongoing abuse. One patient had slightly increasing DUDÌT scores (0 at baseline, 4 after 12 weeks of treatment and 7 at 12 weeks post-treatment). No significant variation was found during the treatment (Χ ²  = 1.2, p = 0.55).