Study population

We selected the study population from a cohort of all consecutive adult patients who initiated DAA regimens for CHC between July 1, 2015 and December 31, 2024, across 22 specialized hepatology centers in Poland. We performed the analysis based on data from the EpiTer‑2 database, a multicenter, retrospective study evaluating real‑world HCV antiviral therapy, conducted with the support of the Polish Society of Epidemiologists and Infectiologists (www.epiter‑2.pl).

Therapeutic decisions regarding the DAA regimen for each patient were made individually by the treating physician. The choice of treatment depended on the available therapeutic options, and the therapy was administered in accordance with the product leaflet, the protocols of the National Health Fund treatment program, and the recommendations of the Polish Expert Group on Hepatitis C.^19-21

Data collection

Data were collected retrospectively via an online platform administered by Tiba Sp. z o. o. (Wrocław, Polska), utilizing patients’ medical records as the primary source of information. Baseline data encompassed both demographic and clinical characteristics, including sex, age, body mass index (BMI), coinfections with HIV and hepatitis B virus, comorbid conditions, liver disease severity, concurrent medications, and prior exposure to antiviral treatment.

Laboratory parameters assessed at treatment initiation included serum concentrations of bilirubin, albumin, and creatinine, alanine aminotransferase activity, hemoglobin concentration, international normalized ratio, platelet count, HCV genotype, and viral load.

The patients included in the study were divided into 2 groups: 1) patients with RA, and 2) patients without RA or any other AID (non‑AID). Patients with autoimmune diseases other than RA were excluded from the analysis to ensure that these would not influence the final results of this study. We subsequently compared these 2 patient subpopulations in terms of baseline characteristics, therapeutic regimens, treatment effectiveness, and safety profiles.

Diagnosis of rheumatoid arthritis

RA is a systemic AID characterized by inflammatory arthritis and extra‑articular involvement. We established the diagnosis of RA following a comprehensive medical interview and an in‑depth analysis of patient clinical records. The disease was diagnosed by rheumatologists in accordance with the criteria of the American College of Rheumatology / European League Against Rheumatism (ACR/EULAR), and the patients with RA remained under the ongoing care of rheumatologists.

Analysis of liver disease severity

We determined the severity of liver disease based on liver stiffness assessment utilizing noninvasive methods for fibrosis quantification. These included real‑time shear‑wave elastography using the Aixplorer system (SuperSonic Imagine, Aix‑en‑Provence, France) and transient elastography with the FibroScan device (Echosens, Aix‑en‑Provence, France). We assessed the fibrosis staging in accordance with the recommendations issued by the European Association for the Study of the Liver (EASL) and the METAVIR classification system.²² We adopted the liver stiffness threshold of 13 kPa as the cutoff value indicative of a high probability of cirrhosis (F4).²²

For all patients with cirrhosis, we evaluated the liver function status using the Child–Pugh (CP) score. Additionally, we collected clinical data regarding hepatic decompensation in the past or at the time of treatment initiation, the presence of esophageal varices, a history of HCC, and orthotopic liver transplant.

We conducted the study in accordance with the Declaration of Helsinki, and the use of retrospective data was approved by the Bioethics Committee of the Jan Kochanowski University in Kielce (57/2024). As the investigation was retrospective in nature, obtaining additional patient consent was not obligatory; nonetheless, the participants were asked to submit informed consent forms for inclusion in the therapeutic program.

Assessment of treatment effectiveness

The primary end point of the study and the principal indicator of DAA therapeutic effectiveness was the achievement of sustained virologic response (SVR). It was defined as undetectable HCV RNA at least 12 weeks after completion of therapy. The patients in whom viral genetic material remained detectable at this time point were virologic nonresponders. The individuals who did not undergo HCV RNA testing 12 weeks after treatment were categorized as nonvirologic treatment failures. In accordance with EASL and national recommendations, an HCV RNA level not exceeding 15 IU/ml was adopted as the diagnostic threshold for viral detectability.^19,22 We included all patients who started treatment in the intention‑to‑treat analysis. The per‑protocol analysis encompassed only the individuals with available HCV RNA testing data at 12 weeks after treatment completion. Furthermore, we assessed treatment effectiveness in the subgroup of patients whose complete data on both liver fibrosis stage and HCV genotype were available.

Assessment of treatment safety

We evaluated treatment safety throughout the entire therapeutic period and during 12‑week post‑treatment follow‑up. Safety data included all cases of treatment modifications in the form of changes in ribavirin (RBV) dosage or discontinuation, as well as the occurrence of adverse events, serious adverse events, and deaths. We placed particular emphasis on hepatic safety outcomes, including gastrointestinal bleeding, ascites, and hepatic encephalopathy.

Statistical analysis

We executed statistical procedures using Statistica version 13 (StatSoft, Tulsa, Oklahoma, United States) and GraphPad Prism 5.1 (GraphPad Software, Inc., La Jolla, California, United States) packages. We summarized the continuous variables as medians with interquartile ranges (IQRs). We performed intergroup comparisons for continuous nonparametric data using the Mann–Whitney test. The assumption of normal distribution was examined with the Shapiro–Wilk test. We described the categorical variables using absolute frequencies and corresponding proportions. Between‑group comparisons of categorical data were performed with the Pearson χ² test; when expected cell counts were below 5, the Fisher exact test was applied. Since this analysis was exploratory and descriptive, no adjustment for multiple testing was applied.

Additionally, we utilized multivariable logistic regression models to estimate the probability of treatment failure across the 2 study cohorts. We presented results as odds ratios (ORs) with their respective 95% CIs. A P value below 0.05 was considered significant for all applied tests.

Baseline characteristics of the study population

Our analysis initially included 21 500 patients, of which we excluded 855 due to having an AID other than RA: systemic lupus erythematosus, autoimmune hepatitis, psoriasis, multiple sclerosis, systemic sclerosis, Sjögren syndrome, vasculitis, Graves disease, Hashimoto disease, autoimmune encephalitis, atopic dermatitis, type 1 diabetes mellitus, celiac disease, myasthenia gravis, inflammatory bowel disease, primary biliary cholangitis, and primary sclerosing cholangitis.

Among the 20 645 patients treated for HCV infection included in the analysis, 206 (1%) were diagnosed with RA (Figure 1). The individuals with RA were significantly older than those in the non‑AID group, with a median (IQR) age of 63 (58–68) vs 58 (48–66) years (P <⁠0.001). Women predominated in the RA cohort, comprising 79.6% of the patients, as compared with 51.4% of the non‑AID group (P <⁠0.001). The patients with RA had lower median BMI than those without AID (24.7 [22.8–27] vs 26.8 [24.1–30] kg/m²; P = 0.01; Table 1).

The patients with RA were more likely to present with comorbid conditions other than RA, including arterial hypertension (47.1% vs 31.2%; P <⁠0.001) and non‑HCC malignant tumors (4.4% vs 2.3%; P = 0.047), as compared with the non‑AID group. They were also more often on concomitant medications (other than RA treatment; 78.2% vs 57%, respectively; P <⁠0.001; Table 1).

The medications most commonly used for RA treatment were glucocorticoids (27.2%), methotrexate (MTX; 6.8%), or a combination of both drugs (6.8%; Figure 2). Notably, 36.9% of the patients were not on any disease‑modifying therapy for RA due to remission of the disease at the time of DAA therapy initiation.

Among laboratory parameters, the patients with RA had significantly lower serum albumin and hemoglobin concentrations, as well as significantly higher platelet counts. Baseline HCV viral load, bilirubin, and creatinine concentrations were comparable in both groups (Table 1).

The predominant HCV genotype (GT) in the entire population was GT1b (74.2%), with a significantly higher prevalence among the patients with RA than the non‑AID individuals. Only 2 patients in the RA group (1%) were infected with GT1a, as compared with 4.9% in the non‑AID group (P = 0.001). The percentage of patients infected with GT3 was also lower in the RA group, but the difference did not reach significance (8.9% vs 13.5%; P = 0.07). Advanced liver fibrosis and cirrhosis (F3–F4) were less common in the patients with RA than in the non‑AID population (30.3% vs 38.4%; P = 0.02). HIV coinfection was significantly less frequent in the RA group (Table 2).

Treatment characteristics

In both groups, most patients were treatment‑naive. The most common therapeutic approach in the RA group consisted of genotype‑specific regimens with the most frequent option of ombitasvir / paritaprevir / ritonavir with or without dasabuvir and with or without RBV (20.6%). However, the sofosbuvir / ledipasvir (SOF/LDV) ± RBV regimen was more frequent in the RA group (P = 0.01). In the non‑AID group, pangenotypic regimens were the most frequently used treatment approach, with glecaprevir / pibrentasvir being the predominant regimen (29.3%); it was also used more often in this group than in the RA patients (P = 0.003; Table 3).

Treatment effectiveness

The effectiveness of DAA therapy in the patients with RA was very high, with SVR reaching 96.2% in the intention‑to‑treat analysis. Six patients (4 women) were lost to follow‑up. The majority of this group was infected with GT1b (n = 4) and was treatment‑naive (n = 5); 1 patient had liver cirrhosis. Most of them completed therapy according to schedule (n = 5). One patient died. After exclusion of these patients, SVR increased to 98.4% in the per‑protocol analysis. These SVR rates were comparable in both groups regardless of sex, HCV genotype, stage of liver fibrosis, and the presence of cirrhosis (Figure 3A and 3B).

In univariable analysis of patients with RA, those who achieved a treatment response did not differ significantly from those who did not respond to therapy with respect to the analyzed clinical and virologic variables. Only a history of HCC occurred significantly more often in the nonresponders. In the group of patients with RA, only 3 individuals failed to respond to treatment, including 2 women; all had infection with GT1b and were treatment‑naive. Liver cirrhosis was present in 2 of these patients (Table 4).

We compared the patients who achieved and those who did not achieve a virologic response to DAA therapy based on available data on both the degree of liver fibrosis and HCV genotype (n = 17 623; Supplementary material, Table S1). In the analyzed population, the patients who failed to achieve an SVR were more often male (P <⁠0.001) and had higher BMI (P <⁠0.001). The patients who responded to DAA treatment received genotype‑specific regimens significantly more commonly. In addition, the nonresponder group significantly more often used medications other than those prescribed for RA.

The prevalence of HCC was higher among the nonresponders than in those in whom therapy was successful (3.8% vs 1.4%; P <⁠0.001). The occurrence of RA, arterial hypertension, diabetes mellitus, chronic kidney disease, and malignancies other than HCC was comparable between the 2 groups and did not reach significance. The patients who did not respond to DAA therapy were more often treatment‑experienced (25.5% vs 18.3%; P <⁠0.001). Moreover, among the nonresponders, a history of hepatic decompensation and the presence of decompensated cirrhosis at treatment initiation were more frequently observed (P <⁠0.001).

In the nonresponder group, as compared with the responders, GT3 infection (32.7% vs 12.5%; P <⁠0.001) and liver cirrhosis (F4; 50.5% vs 23.8%; P <⁠0.001) were more prevalent. In contrast, among the patients who achieved a virologic response, GT1b infection (75% vs 58.4%; P <⁠0.001) and low‑grade fibrosis (F1; 40.2% vs 23.6%; P <⁠0.001) were more common. In the subgroup of patients with liver cirrhosis (F4), the nonresponders were more often classified as CP class B or C (22.9% vs 9.2%; P <⁠0.001) and more frequently had esophageal varices (P <⁠0.001).

Logistic regression analysis assessing virologic failure in the entire cohort of patients infected with HCV identified the following significant predictive factors: male sex (OR, 2.3), GT3 infection (OR, 2.9), a history of prior antiviral therapy (OR, 1.4), a history of esophageal varices (OR, 2), cirrhosis (OR, 2), use of medications other than those for RA (OR, 1.3), and CP class B or C (OR, 2.3; Figure 4).

Treatment safety

In both groups, the proportion of patients who completed antiviral therapy as planned was very high and comparable. We found no significant differences between the analyzed groups in terms of serious adverse events leading to treatment discontinuation or in treatment cessation due to hepatic decompensation. In contrast, the proportion of patients with at least 1 adverse event was significantly higher in the RA group. Mortality was similar across the groups, and none of the reported deaths in either group were considered related to antiviral therapy (Supplementary material, Table S2).

Limitations

When interpreting the results of this study, several important limitations should be taken into account. The key limitation that should be acknowledged is the paucity of data on RA. We cannot rule out that RA patients treated with currently recommended agents, in particular those with extensive joint involvement, differ from the group defined using medical records in our study. We based the diagnosis of RA on medical records and clinical history rather than on the uniform application of validated classification criteria, such as the ACR/EULAR criteria. However, to our best knowledge, the diagnosis made by rheumatologists was based on validated criteria. Secondly, the retrospective nature of the study and its reliance on actual data may introduce systematic errors leading to an underestimation of the frequency of adverse events. Thirdly, detailed data on RA disease activity, duration, serological status, and extra‑articular manifestations were not available. Fourthly, information on past immunosuppressive therapy was limited, and we only had information on the medications taken during DAA therapy. It should be added that although we had data on steroid and MTX use, data on total treatment duration, cumulative doses, and changes in immunosuppressive therapy before DAA treatment were incomplete. Therefore, it was not possible to conduct a more detailed analysis of the potential impact of different RA treatment strategies on antiviral efficacy and safety. Finally, the RA subgroup represented a relatively small proportion of the overall cohort, and the number of virologic failures within this group was very low. As a result, the statistical power to detect subtle differences in treatment effectiveness or to perform robust subgroup analyses within the RA population was limited.

Despite these restrictions, the study has some significant advantages, supported by its large sample size and data collection from 22 specialized hepatology centers, which enhances the representativeness of the results. Furthermore, detailed patient characterization enabled thorough subgroup analyses, and the low proportion of patients who were lost to follow‑up reduced the risk of selection bias. All these factors, taken together, increase the credibility of the findings and support their broad generalizability to clinical populations treated with DAA‑based regimens.

Conclusions

Patients with RA constitute a small but distinct subgroup among individuals treated for CHC. Despite being older and burdened with more comorbidities, the RA patients had similar liver disease severity and responded to DAA therapy as effectively and safely as the individuals without RA. Multivariable analysis showed that advanced liver disease, presence of esophageal varices, male sex, GT3 infection, prior antiviral therapy, and use of non‑RA medications—but not RA itself—were independent predictors of virologic failure. These findings indicate that RA does not compromise DAA treatment outcomes.