Hepatocellular carcinoma in Turkey: A nationwide health system database analysis of clinical patterns and survival

Hepatocellular carcinoma (HCC) is a leading cause of cancer-related mortality worldwide. However, comprehensive population-level data describing etiologic patterns, treatment exposure, healthcare utilisation, and survival remain limited in many countries. This study aimed to characterise the national burden, care patterns, and outcomes of HCC in Turkey.

We performed a nationwide, real-world analysis of adult patients diagnosed with HCC between 2016 and 2024 using administrative health data. Cases were identified by an administrative case definition requiring at least three separate records of ICD-10 codes C22, C22.0, or C22.9; at least one C22.0 code was not required, so HCC could not be confirmed at the individual level. Baseline comorbidities were restricted to conditions recorded on or before the date of diagnosis. Demographic characteristics, underlying liver disease, etiologic factors, treatment exposure, and the level and type of diagnosing institution were assessed. Overall survival (OS) was estimated using the Kaplan-Meier method. To reduce immortal-time bias, treatment exposure was analysed in a three-month landmark cohort; because treatment dates were unavailable, this reduces but does not eliminate the bias. Multivariable Cox regression was performed within the same cohort.

A total of 13,674 patients (64.6% male; median age 64 years, IQR 55-73) were included, of whom 61.4% had died at the data cutoff; median follow-up was 53.0 months. The median OS was 13.2 months (95% CI 12.1-14.5), with 1-, 3-, and 5-year OS of 50.9%, 41.1%, and 36.9%; because patients dying before three qualifying records accumulated could not enter the cohort, these estimates apply to patients meeting the case-identification algorithm and are likely to overstate population survival. OS was markedly longer in women than in men (median 50.9 vs 8.2 months), and longer in patients without cirrhosis recorded at diagnosis than in those with cirrhosis (18.5 vs 9.0 months; both P < 0.001). Hepatitis B was the most frequently recorded specific etiology (29.4%); among mutually exclusive categories, NAFLD-associated HCC had the most favourable survival (5-year OS 61.0%) and alcohol-related HCC the least (median 8.8 months), while HBV-HDV co-infection was associated with longer survival than HBV mono-infection (58.2 vs 12.0 months; P < 0.001), a difference that persisted after adjustment for age, sex, and cirrhosis; no specific etiology could be assigned in 56.1% of patients. Sensitivity analyses restricting etiologic ascertainment to codes recorded within 3 months before to 3 months after diagnosis, or incorporating viral serology, attenuated the survival of the specific viral etiologies, indicating survival-dependent ascertainment. In a multivariable Cox model fitted in the three-month landmark cohort (n = 8958), older age, male sex, cirrhosis, and systemic therapy were associated with higher mortality (all P < 0.001), with time-varying associations for age, cirrhosis, and systemic therapy. The association of local therapy with mortality was not constant over follow-up (HR 0.78, 95% CI 0.72-0.85 during 0-24 months after the landmark and 1.22, 1.04-1.43 thereafter); the overall association was no longer evident after liver transplantation was excluded from the local-therapy category. Care was highly centralised within secondary and tertiary institutions.

In this nationwide analysis, HCC in Turkey was characterised by limited OS and pronounced sex-based disparities, with survival differing by cirrhosis status and recorded etiology; etiologic ascertainment was itself survival-dependent, so these contrasts describe recorded categories. Hepatitis B was the most frequently recorded specific etiology, underscoring the continued importance of HBV prevention, screening, and surveillance, while HCV, HDV, alcohol-related and metabolic liver disease accounted for smaller recorded proportions; temporal trends in etiology were not analysed. Because tumour stage and liver function were unavailable, the treatment-survival associations are descriptive and subject to confounding by indication. Efforts focused on HBV control, earlier detection, risk-based surveillance of patients with cirrhosis, and improved referral pathways may improve population-level outcomes.