Long-term survival and survivorship in non-Hodgkin lymphoma patients in Sweden

Abstract

Non-Hodgkin lymphoma (NHL) is one of the top ten most common cancer types in Sweden. Although sometimes referred to as one disease, NHL is truly an umbrella term representing a heterogeneous group of diseases with varying clinical course and prognosis. The main data source for all four studies included in this thesis is the Swedish lymphoma register (SLR). This national quality register provides population-based data, detailed clinical information and the possibility to distinguish between different morphological subtypes of NHL.

In Study I we provide a systematic presentation of temporal trends in absolute numbers of prevalent patients by NHL subtypes, linking them to trends in incidence, survival and mortality. Poisson regression was used to test for temporal trends. We found that an increasing incidence and improved survival have led to an increase in the prevalence of NHL overall and for almost all investigated subtypes between 2000 and 2016. This increase was most notable for diffuse large B-cell lymphomas (DLBCL) among aggressive subtypes and marginal zone lymphomas among indolent subtypes. The prevalence provides a measure of burden of disease, useful for health care planning and to optimize resource allocation. The prevalence also represents the number of survivors in the population, at risk for relapses and psychological and physiological side effects of their lymphoma or treatment. The increase in number of prevalent NHL patients underscores the need to develop and evaluate alternative follow-up schemes of lymphoma survivors since especially patients diagnosed with indolent lymphoma subtypes are followed in the clinic for many years.

The most common subtype of NHL, DLBCL is the focus in study II-IV. In recent years, the addition of rituximab to the standard combination chemotherapy has improved outcomes in patients with DLBCL. Nevertheless, every fourth patient treated curatively is expected to experience progressive disease or relapse. Study II aimed to quantify trends and remaining loss in life expectancy due to DLBCL in a population-based cohort. Loss in life expectancy was predicted using flexible parametric models from diagnosis and among two-year survivors, by age, sex and age-adjusted international prognostic index (aaIPI). The number of life-years lost decreased over the study period 2000-2013 in all patient groups. However, especially younger patients (≤60 years) with aaIPI≥2 were still estimated to lose many life years in 2013. Among two-year survivors, the loss in life-expectancy was reduced to two years or less by the end of the study period, regardless of age, sex and aaIPI. By using novel measures, we illustrated the improvement of DLBCL survival in a population-based context and over the entire life-span. The standard chemotherapy for curative treatment of DLBCL contains the cardiotoxic anthracycline doxorubicin. An increased rate of heart failure is well documented following this treatment, whereas incidence and outcome of other cardiac complications, e.g. myocardial infarction, are less well studied. In Study III we assessed the incidence, characteristics and outcome of acute myocardial infarctions (AMI) in curatively treated patients with DLBCL. Patients were matched to lymphoma-free comparators and the rate of AMI was estimated using flexible parametric survival models incorporating repeated events. Overall, DLBCL patients had a 33% excess rate of AMI compared to the general population. However, the excess rate was most pronounced during the first year after diagnosis and diminished after 2 years. The strongest risk factors for AMI were advanced age, male sex and pre-existing comorbidity. There was no difference in AMI characteristics, extent of treatment or 30-day survival following hospitalization for AMI between DLBCL patients and comparators. The increased risk of AMI especially during the first 2 years and among elderly patients calls for improved cardiac monitoring. In Study IV we estimated real-world probabilities for lasting remission by clinical disease characteristics using a multi-state model approach. DLBCL patients who achieved remission after primary treatment were followed for repeated relapses and death. Flexible parametric models were used to model transition rates between disease stages accounting for competing events at each transition. At 2 years after end of primary treatment, 81% of the patients remained in remission, 13% had relapsed and 6 % of patients had died in first remission. The probability of remaining in remission for at least 2 years was reduced by 24 percentage units for patients with international prognostic index, IPI 4-5 compared to patients with IPI 0-1. On average, these patients lost 4.4 months of being in remission the first 2 years. Only 43% of relapsing patients achieved a second remission and half of them (51%) relapsed again - reflecting the difficulties in treating relapsing patients.