Dear Editor,

Classical Hodgkin lymphoma (cHL) is a B-cell derived malignancy that is characterized by a bimodal incidence peak in the 20–30 and 60–75 years age groups [1]. Curative treatment for cHL has been feasible since the 1960s, after the introduction of mantle-field irradiation [2, 3] and combination chemotherapy protocols, including mechlorethamine, vincristine, procarbazine, and prednisone (MOPP) and MOPP–doxorubicin, bleomycin, vinblastine, and dacarbazine (ABVD) [4]. Data collected by IARC NORDCAN indicated that the 5-year age-standardized relative survival rates improved from 51.7% and 55.8% in 1971–1975 to 88.2 and 90% in 2016–2020 for males and females, respectively [5]. However, age of the patient at diagnosis consistently affects survival significantly. Furthermore, improvement in therapy efficacy has come at the expense of severe long-term toxicities. The less toxic ABVD regimen [6] with a shorter duration [7] has substituted the MOPP–ABVD backbone to mitigate side effects while maintaining disease control. Advances in radiotherapy (RT) planning and delivery have enabled local control maintenance while minimizing exposure to surrounding tissues [8]. Considering these continuous adjustments to therapeutic regimens, the up-to-date reporting of survival and late effects within patients with cHL of different age cohorts remains highly relevant.

Our retrospective cHL registry study utilized existing nationwide data (Fig. S1) available from six Finnish electronic healthcare registries, including the Finnish Cancer Registry (FCR), the National Institute of Health and Welfare, the Social Insurance Institution of Finland, the Finnish Centre for Pensions, Statistics Finland, and Digital and Population Data Services Agency (DPA). The analysis cohort included patients with record of HL based on FCR database (ICD-10: C81*), excluding those with nodular lymphocyte-predominant HL (ICD-O-3 morphology code 9659/3) from January 1, 2000, to December 31, 2019 (n = 2510). An age, sex, and region-matched control cohort in a 1:1 ratio was determined using DPA records, and eligible control participants included individuals without lymphoma diagnosis (C81–C85) from January 1, 2000, and December 31, 2019, and who were alive at the index of the corresponding case. Tables S1 and S2 present the baseline characteristics of the participants. Table S3 presents the methods. The end of follow-up (FU) was defined as death, end of study, or loss to FU. The mean FU time were 9.7 and 8.3 years in the younger and older patients (9.9 and 8.3 years for respective control cohorts), respectively.

Consistent with other reports [9], we observed excellent 5-, 10-, and 20-year survival rates of 96.6% (95% confidence interval [CI]: 95.6–97.4%), 94.6% (95% CI: 93.1–95.7%), and 90.9% (95% CI: 88.4–92.8%) in younger patients with cHL (Fig. 1A) and 99.7% (95% CI: 99.3–99.9%), 99.2% (95% CI: 98.5–99.6%), and 97.3% (95% CI: 95.2–98.4%) in the younger control cohort, respectively. Importantly, in concordance with data from the United States [10] and other Nordic countries [11], we observed a temporal improvement in overall survival (OS) and a reduction in the overall risk of mortality in younger patients from the 2000–2009 to 2010–2019 treatment era (p = 0.0033) (Fig. 1C, Table S4). The 2010–2019 era was also associated with a lower mortality risk than the 2000–2010 era in the Cox multivariable analysis (hazard ratio [HR]: 0.41, 95% CI: 0.22–0.74, p < 0.01) (Table S5). The reasons for the improved survival could not be thoroughly described because the hospital medical records of the administered therapy protocols were not retrieved for this study. However, this finding may be attributed to more effective salvage treatments, such as brentuximab vedotin or immune checkpoint inhibitors, because no major changes have occurred in the first-line treatment of cHL since 2000 [12]. General healthcare improvement is an unlikely explanation because the survival of the control cohort remained similar between the two eras of interest (Fig. S2).

Fig. 1: Overall survival of cHL patients.
figure 1

A Younger patients with cHL and their controls, B Younger patients by the diagnosis year period of 2000–2009 or 2010–2019, C Older patients with cHL and their controls, D Older patients by the diagnosis year period.

Full size image

Despite the excellent survival rates, we revealed a consistently higher overall mortality risk in younger patients with cHL in comparison to controls. The Cox multivariable analysis (Table S6) revealed that the HRs for death in patients with cHL in comparison to controls was 12.13 (95% CI: 4.37–33.67, p < 0.001) at 0–5 years after diagnosis and 2.24 (95% CI: 0.96–5.19, p = 0.06) even at 10 years after diagnosis. We categorized the causes of death into cHL, other lymphomas, solid malignancies, myeloid malignancies, cardiovascular diseases (CVDs), and others for the cause-specific analysis of mortality (Table 1). Of the 87 deaths in younger patients during the FU study, cHL was the cause of death in 53 (60.9%) (Table S7). Cumulative cHL-specific mortality increased to 3.7% by the 10-year FU and leveled off thereafter (Table 1). Mortality caused by other reasons accumulated during the later years of FU but remained surprisingly low. For example, the cumulative risk of mortality from CVD was only 0.9% (0.1% in controls) and that from solid cancers was 1.6% (1.1%) at 20 years. The rates are comparable to those from a recent Danish study from the same era [13]. Similar results between countries are not unexpected considering the joint Nordic treatment guidelines for cHL. Thus, mortality from secondary toxicities may have a decreasing tendency among younger patients with cHL treated from 2000 onward, given the caveat of short FU.

Table 1 Cumulative incidence (%) of other malignancies and CVDs and mortality (%) by cause of death in younger and older age cohorts.
Full size table

Alarmingly, however, our study revealed a rapidly and steadily increasing cumulative incidence of CVDs and solid malignancies from 5 years onward without a plateau in the younger cHL cohort, which contrasts with the low mortality rates (Table 1, Supplementary Fig. S3). The cumulative incidence of solid malignancies at 10- and 20-year FU was 7.7% and 13.8%, in patients with cHL and 1.0% and 3.6% in controls, respectively. Similarly, the cumulative incidence of CVDs at 10- and 20-year FU was 7.0% (3.9%) and 17.4% (9.6%), respectively. The numbers are in the same range or only slightly lower than those for patients treated before 2000 [14]. The high incidence, together with low death rates, could be attributed to improved patient monitoring, enabling early therapy interventions for treatment-induced CVDs, and a curative approach for early-stage secondary cancers [15]. Nevertheless, there is a risk that extended FU would turn the high incidence of CVDs in our study into increased late mortality. Notably, the incidence of myeloid malignancies of 1.0% (0.2%) at 20 years of FU remained moderate in our study, indicating successful conversion away from highly mutagenic protocols.

We observed significantly poorer survival rates of 55.1% (95% CI: 51.5–58.5%), 41.8% (95% CI: 37.8–45.7%), and 22.7% (95% CI: 15.1–31.3%) in older patients and 85.5% (95% CI: 82.7–87.9%), 68.5% (95% CI: 64.4–72.3%), and 40.1% (95% CI: 27.0–52.7%) in older controls at 5, 10, and 20 years of FU, respectively, in contrast to younger patients. Poor survival was mostly attributed to high cHL-specific mortality in older patients with cHL during early FU (21.7% during the first year, Table 1). No temporal improvement in OS or reduction in the overall risk of mortality was observed from 2000–2010 to 2010–2019, unlike in younger patients (Fig. 1D and Table S4). However, the median age of patients treated in the latter period (Table S2) was higher than in the former period (68.1 vs. 65.8 years), which may have countervailed the possible OS improvement. We did not observe any excess morbidity or mortality from other malignancies or CVD in the older patient cohort (Table 1). In particular, the 20-year cumulative mortality from solid malignancies was 7.0% and from CVDs was 9.5%, which were lower than in the control (11.4% and 17.0%, respectively, Table 1). This is most likely explained by the healthy survivor effect, i.e., the enrichment of fit and less morbidity among long-term survivors.

The strengths of our study include a population‐based approach using real‐life data, excellent coverage and reliability from national registries used for the study. This study has limitations, including a short FU time, which restricts the comprehensive characterization of late toxicities and mortality. Additionally, we have no access to detailed information on patients’ treatment or therapy intend, and missing or inaccurate data are possible because of differences in recording practices. Moreover, due to technical reasons, no lymphomas were allowed in the FU in control groups which may have a marginal biasing effect between the study groups.

In conclusion, the prognosis of patients aged <50 years is excellent and is still improving with the current therapeutic modalities. However, the high incidence of secondary solid cancers and CVDs remains a concern. Considering the long lifetime expectancy of these patients, our results emphasize the importance of close monitoring of secondary complications and evaluations for therapeutic de-escalation. The survival of patients aged >50 years is poor because of high mortality associated with either lymphoma or complications of conventional therapy. Evidently, more tolerable therapeutic modalities are required in older patients with cHL.