This nationwide, noninterventional, retrospective study is the first to assess real-world treatment patterns in patients with mUC in Hungary. Although nationwide studies in mUC have been carried out in other countries [20,21,22,23,24], to our knowledge, no similar study has been performed in Central-Eastern Europe. The population of patients with mUC identified in Hungary was mostly men and median age was 67 years, similar to other real-world studies [21,22,23, 25].

Overall, study data showed that treatment of mUC in Hungary follows established European guidelines [5,6,7]. Only half of patients received any systemic treatment for mUC. This rate is similar or higher than rates observed in other countries across a similar period [25,26,27,28,29,30,31,32]. However, rates of 1L systemic treatment increased during the study period when full-year data were available (2016–2020); a similar increase in 1L systemic treatment rate was also seen in a real-world study of mUC in Germany between 2015 and 2019 [21,22,23, 25]. No conclusions could be drawn for the year 2021 because data were incomplete. Of patients who received treatment, most received 1L PBC, consistent with treatment guidelines [5,6,7]. However, 1L ICI use in Hungary may have been limited by availability in the NPBR program, which occurred only later in the study period (since 2018), with use of 1L ICIs increasing thereafter. In our study, only 24% of patients who were treated with 1L therapy received subsequent 2L therapy. This rate is slightly lower than rates seen in some recent real-world studies [21,22,23, 30, 33]. The limited proportion of patients with mUC who receive systemic treatment, and the high rate of attrition between 1L and 2L therapy, are well recognized and highlight the need to use the most effective available frontline agents to improve outcomes in this population [34].

Patients who received 1L ICI treatment were older and had more comorbidity than patients who received 1L PBC, similar to patients in a previous real-world study [29]. In addition, patients who received 1L carboplatin + gemcitabine had more comorbidities than patients who received 1L cisplatin + gemcitabine. These observations are consistent with 1L ICIs being approved by the EC only in patients who were cisplatin-ineligible with PD-L1+ tumors during the study period [12, 13] and the recommendation of carboplatin + gemcitabine as an alternative to cisplatin-based chemotherapy in patients who were cisplatin-ineligible in guidelines [2, 5,6,7]. Specifically, established criteria for cisplatin ineligibility include specific comorbidities (reduced performance status, renal impairment, hearing loss, neuropathy, or heart failure) [2, 5,6,7]; thus, patients who were cisplatin-ineligible would be expected to have more comorbidities and to be generally frailer than patients receiving cisplatin-based chemotherapy.

Unadjusted OS was longer in patients with mUC who received any 1L treatment than in those who did not receive a relevant systemic treatment. Of those who received 1L treatment, unadjusted OS was longer in patients who received any 1L PBC than in those who received 1L ICIs or 1L non-PBC (median, 12.8 vs 6.3–7.5 months). These outcomes are consistent with those in previously published real-world data from unadjusted analyses [23, 24, 28, 29]. OS was also longer in patients who received 1L cisplatin + gemcitabine than in patients who received 1L gemcitabine + carboplatin, which was reported in previous studies [20,21,22, 29]. However, because detailed patient demographic and disease specific data were not available at baseline, we were unable to perform adjusted OS analyses. Potential predictors for OS include patient demographics (e.g., performance status), comorbidity, and site of metastasis at the start of 1L treatment [35,36,37]. Thus, comparisons of OS between treatment subgroups in our analyses should be interpreted with caution.

This study had some limitations. First, the NHIFA is a payer-specific database, and information on mUC disease-specific parameters, such as stage at diagnosis, treatment outcome, or known mUC risk factors, including smoking history, is lacking. Because of the lack of disease-staging information in the database, mUC diagnosis was identified either by ICD-10 codes or receipt of 1L systemic anticancer treatment. No information was available on the potential reasons for nonreceipt of 1L systemic treatment. Furthermore, claims-based studies are also inherently associated with factors such as coding errors and missing information. Because detailed baseline characteristics could not be collected, descriptive statistical analyses were conducted for efficacy results and OS was unadjusted, as discussed earlier. Median follow-up in the overall population was short because of the short median OS in some subgroups (untreated patients and patients receiving 1L ICI or 1L non-PBC treatment), which reflects the poor prognosis of patients with mUC. Moreover, subgroups of fewer than 10 patients were not included in the analysis, in line with the data protection regulations of the NHIFA database. Sizes of some subgroups were relatively small, specifically those receiving 1L ICI and 1L non-PBC vs any 1L PBC, meaning that data may not be fully representative. Lastly, 1L ICIs were only accepted in the NPBR program later in the study period (2018) [38].