Follicular Lymphoma in Sweden: Nationwide Improved Survival in the Rituximab Era, Particularly in Elderly Women: a Swedish Lymphoma Registry Study

Home , Era

ORIGINAL ARTICLE Follicular lymphoma in Sweden: nationwide improved survival in the rituximab era, particularly in elderly women: a Swedish Lymphoma Registry Study

HR Junlén1, S Peterson2, E Kimby1, S Lockmer1, O Lindén3, H Nilsson-Ehle4, M Erlanson5, H Hagberg6, A Rådlund7, O Hagberg2 and BE Wahlin1

Treatment for follicular lymphoma (FL) improved with rituximab. In Sweden, first-line rituximab was gradually introduced between 2003 and 2007, with regional differences. The first national guidelines for FL were published in November 2007, recommending rituximab in first-line therapy. Using the population-based Swedish Lymphoma Registry, 2641 patients diagnosed with FL from 2000 to 2010 were identified and characterized by year and region of diagnosis, age (median, 65 years), gender (50% men), first-line therapy and clinical risk factors. Overall and relative survivals were estimated by calendar periods (2000–2002, 2003–2007 and 2008–2010) and region of diagnosis. With each period, first-line rituximab use and survival increased. Survival was superior in regions where rituximab was quickly adopted and inferior where slowly adopted. These differences were independent in multivariable analyses. Ten-year relative survival for patients diagnosed 2003–2010 was 92%, 83%, 78% and 64% in the age groups 18–49, 50–59, 60–69 and ⩾ 70, respectively. With increasing rituximab use, male sex emerged as an adverse factor. Survival improved in all patient categories, particularly in elderly women. The introduction and the establishment of rituximab have led to a nationwide improvement in FL survival. However, rituximab might be inadequately dosed in younger women and men of all ages.

Leukemia (2015) 29, 668–676; doi:10.1038/leu.2014.251

INTRODUCTION MATERIALS AND METHODS Follicular lymphoma (FL), an indolent CD20+ malignancy of Swedish registries germinal B-cell origin, is the most common indolent lymphoma.1 Every clinician and pathologist/cytologist in Sweden is legally obliged to The disease is usually widespread at diagnosis and considered report each occurrence of cancer to the nationwide Swedish Cancer incurable with conventional therapy (although most kinds Registry (SCR). The registry includes diagnosis, gender, date of birth, date of therapy induce remission), because it will relentlessly of diagnosis and the hospital where the diagnosis was made, but no 2 clinical information. The SCR and its diagnostic classification were relapse. fi Rituximab, a chimeric mouse/human anti-CD20 monoclonal introduced in 1958. By the 1990s, they were insuf cient for the proper antibody, was introduced in the 1990s. Its effector mechanisms registration of most lymphoma entities. The Swedish Lymphoma Group therefore commissioned the Swedish Lymphoma Registry (SLR), which include complement-mediated cytotoxicity, antibody-dependent 17 ’ 3 was launched nationwide on 1 January 2000. The SLR s purpose was to cellular cytotoxicity and direct induction of apoptosis. In the amend the lymphoma registration with information on exact subtype, single-agent pivotal study of 1998, rituximab induced complete first-line treatment, localization, stage and other prognostic lymphoma- 4 or partial remission in half of relapsed patients. Several and patient-specific data. The SLR excludes patients o18 years old and first-line treatment trials conducted around the turn of the cases diagnosed at autopsy.17 When a patient with lymphoma is millennium showed that rituximab improved remission rates reported to the SCR, the SLR is notified and sends a form to the and remission durations when combined with conventional attending clinician, requesting additional information. The coverage in fi 5–8 SLR compared with the mandatory SCR has been 95–97% since its chemotherapeutic rst-line regimens and that is was effective 18 as monotherapy.9–12 initiation, according to yearly reports. The SLR has been validated, showing a 95% agreement between the diagnosis reported to the SLR The median overall survival (OS) in FL patients was about 9 fi 18 years in the 1990s.13 Several centers have reported improved and the actual diagnosis in the local patient les. In June 2007, 14–16 additional prognostic factors were added to the report form, allowing survival since rituximab was introduced. The aim of this the calculation of the FL International Prognostic Index (FLIPI)19 and the national population-based study was to assess patterns of FL separation of FL grades 3A and 3B. For every resident in Sweden, date of survival after the introduction of first-line rituximab and to identify death is centrally registered in the Causes of Death Registry, from which important prognostic factors in the rituximab era. survival data was obtained.

1Division of Hematology, Department of Medicine at Huddinge, Karolinska Institutet, and Hematology Center, Karolinska University Hospital, Stockholm, Sweden; 2Department of Statistics, Regional Cancer Center in South Sweden, Lund, Sweden; 3Department of Oncology, Lund University Hospital, Lund, Sweden; 4Section of Hematology and Coagulation, Department of Medicine, The Sahlgrenska Academy at the University of Gothenburg, Gothenburg, Sweden; 5Department of Radiation Sciences, Oncology, Umeå University, Umeå, Sweden; 6Department of Oncology, Uppsala University Hospital, Uppsala, Sweden and 7Department of Oncology, Ryhov County Hospital, Jönköping, Sweden. Correspondence: Dr BE Wahlin, Division of Hematology, Department of Medicine at Huddinge, Karolinska Institutet, and Hematology Center, Karolinska University Hospital, M54 Hematologiskt Centrum, Karolinska Huddinge, Stockholm 141 86, Sweden. E-mail: [email protected] Received 17 July 2014; revised 11 August 2014; accepted 20 August 2014; accepted article preview online 25 August 2014; advance online publication, 19 September 2014 Follicular lymphoma survival in Sweden 2000–2010 HR Junlén et al 669 Patients population life tables stratified by sex, age and calendar period, expected 32 The studied population includes all patients in the SLR newly diagnosed survival was assessed using the Ederer II method. RSRs were calculated with FL of any grade from 1 January 2000 to 31 December 2010, excluding for patients in the three calendar periods and in four age orders (18–49, patients with a coterminous (within 90 days) diagnosis of aggressive 50–59, 60–69 and ⩾ 70 years). Multivariable models for excess mortality lymphoma. Patients were not excluded because of other diseases. All were constructed using the expectation-maximization algorithm, using a patients were observed from date of diagnosis until death, emigration or smoothed baseline excess hazard, yielding excess mortality rate ratios as 33 end of follow-up (10 February 2014). From the SLR, we obtained estimates. The excess mortality rate ratio denotes excess mortality (the information on first-line therapy, age, gender, World Health Organization difference between observed and expected mortality in the general (WHO) performance status, Ann Arbor stage, lactate dehydrogenase, population); for example, if men show an excess mortality rate ratio of 1.55, B symptoms, bulky disease and the FLIPI. We also investigated the health- it would indicate that they have 55% higher excess mortality than women. care regions: Sweden is divided into six health-care regions, each of which The RS calculations were performed using R software (Maja Pohar Perme has a population between 0.9 and 2.2 million people and a tertiary (2013): relsurv, R package version 2.0–4 (http://CRAN.R-project.org/ lymphoma center. Each region maintained regional guidelines for FL until package = relsurv)); all other calculations with Stata version 9.2 (StataCorp, November 2007, when nationwide guidelines for FL were implemented. College Station, TX, USA). This study was approved by the Ethics Committee, Stockholm, Sweden. RESULTS An overview of treatment for FL in Sweden during the study period A total of 2641 FL patients (1334 women and 1307 men) without concurrent transformation were diagnosed in Sweden from 2000 Since the early 1970 s, local radiation20 for Ann Arbor stage I and limited to 2010 (Table 1). The average number of new patients per year stage II disease was the standard of care, and this did not change during the – study period. Asymptomatic patients with widespread lymphoma have been was 240 (range, 208 266). The age-adjusted incidence rates did managed with wait-and-watch policies since decades,21 and this practice was not increase over time (Supplementary Information). The median also continued through the study period. However, with the advent of age at diagnosis was 65 years (range, 18–100) and age at rituximab, treatment algorithms for symptomatic generalized disease diagnosis increased over time (P = 0.004), whereas the number of changed radically between 2000 and 2010. In the early 2000s, first-line patients who presented at diagnosis with poor WHO performance treatment was similar to the situation in the 1990s, with the most status (P = 0.003) and bulky disease (P = 0.006) decreased. There common regimen being the polychemotherapeutic cyclophosphamide- were no changes in lactate dehydrogenase levels or B symptoms hydroxydaunorubicin-vincristine-prednisone (CHOP) regimen, followed by over time. The median follow-up time was 7.8 years (2000–2002, single oral alkylators, mostly chlorambucil. In Sweden, CHOP was particularly 12.5 years; 2003–2010, 6.6 years (2003–2007, 8.6 years; 2008–2010, used for patients with high-risk disease.22 Relapse therapy included several 4.5 years)). The FLIPI, available from June 2007, predicted OS chemotherapy combinations, and autologous stem cell transplantation had o been established in the 1990s for relapsed23 or transformed24 lymphoma. (P 0.00005; Supplementary Figure). From 1998, the Swedish Lymphoma Group participated in a Nordic Lymphoma Group trial where 123 patients received rituximab without 12 First-line rituximab use chemotherapy, the last patient entering the trial in November 1999. The – encouraging experience from that trial prompted some doctors to use first- Nationwide, among patients diagnosed 2000 2002, 13% of line rituximab off-label at the start of the study period. Another Nordic reported patients had rituximab as part of systemic primary Lymphoma Group trial was launched in November 2002: in total, 145 treatment, among those diagnosed 2003–2010, 71% did (2003– Swedish FL patients received rituximab without chemotherapy in this trial 2007, 61%; 2008–2010, 86%; Table 2). In patients diagnosed from November 2002 until August 2008; all Swedish health-care regions between 2003 and 2007, first-line rituximab use increased every participated in the trial and those patients are included in the present year, 31% in 2003, 68% in 2005 and 85% in 2007. The speed of 25 population. During that period, there were different regional algorithms adoption of rituximab was not equal in all health-care regions. whether to give first-line rituximab or not to patients outside trials. In One region adopted rituximab quickly and, during 2003–2007, this fi November 2007, the Swedish Lymphoma Group published the rst region’s frequency of rituximab use in systemic first-line therapy nationwide guidelines for FL, in which rituximab was recommended for fi (92%) was comparable to modern usage. Another region rst-line therapy, either as monotherapy or in combination with chemother- fi apy, retaining chlorambucil as an alternative for the very old. The 2007 incorporated rituximab slowly and its rst-line use of rituximab national guidelines also incorporated rituximab maintenance after relapse remained low (31%) throughout 2003–2007. The other four therapy.26 Between 2008 and 2010, subsequent editions of the guidelines regions reported a more moderate introduction of first-line added rituximab maintenance after first-line therapy (with yttrium-90- rituximab (in total, 62% during 2003–2007). In the transitional ibritumomab tiuxetan as an alternative) and bendamustine as part of first-line period, that difference in rituximab use between the three immunochemotherapy.27,28 We divided the patients by their date of categories of region was significant (Po0.00005). The regions diagnosis into two eras, the prerituximab era (the calendar period 2000– had almost identical rituximab first-line use during 2008–2010, – 2002) and the rituximab era (2003 2010); the latter was further subdivided because, following the national guidelines introduced in Novem- into the transitional period (2003–2007), when rituximab became gradually fi – ber 2007, the slow-adopting and moderate regions caught up more common as rst-line therapy, and the established period (2008 2010), fi when rituximab was part of standard first-line therapy nationwide. We their rst-line rituximab use (91% and 84%, respectively, assessed the fraction of patients reported to SLR having received rituximab as compared with 92% in the fast-adopting region; P = 0.91). part of systemic first-line therapy by year of diagnosis and health-care region. Throughout the study period, rituximab, when given, was always intravenous OS analysis and the universal standard dose was 375 mg/m2. Improvement in OS. OS was superior in the rituximab era (P = 0.0001) and it improved incrementally with each calendar Survival analysis period (Po0.00005; Figure 1a). The OS differences between the FL survival in the three calendar periods (2000–2002, 2003–2007 and calendar periods were independent from other clinical predictors 2008–2010) and by health-care regions was estimated using univariable in multivariable Cox analysis (Supplementary Information). In all and multivariable Cox OS analysis29 and univariable and multivariable 30,31 2641 patients, the 5- and 10-year OS was 71% and 54%, relative survival (RS) analysis. Because RS analysis measures the total respectively. The 5-year OS in the three respective calendar excess mortality associated with a specific diagnosis, regardless of whether periods was 68% (95% confidence interval (CI), 65–71%), 74% the excess mortality is directly or indirectly caused by the diagnosis, the – – precise cause of death is irrelevant. The RS ratio (RSR) is defined as the OS (95% CI, 71 76%) and 77% (95% CI, 73 80%), respectively. in the diagnosed group divided by the expected OS of a comparable group from the general population. The 5- and 10-year RSRs show the proportion Regional differences in survival. In patients diagnosed during the of patients who survive FL at 5 and 10 years, respectively. From Swedish transitional period (2003–2007), OS was superior in the region

Table 1. Clinical characteristics at diagnosis

Prerituximab era Rituximab era

(2000–2002) Transitional period (2003–2007) Established period (2008–2010) Total

No. % No. % No. % No. %

Patients with follicular lymphoma 757 100.0 1155 100.0 729 100.0 2641 100.0

Age (years) Median (range) 64 (21–95) 64 (19–100) 66.5 (20–98) 65 (19–100) 18–49 100 13.2 160 14.0 84 11.6 344 13.1 50–59 189 25.0 230 20.1 117 16.1 536 20.4 60–69 187 24.7 326 28.5 240 33.1 753 28.7 ⩾ 70 281 37.1 427 37.4 285 39.3 993 37.8

Sex Female 398 52.6 567 49.1 369 50.6 1334 50.5 Male 359 47.4 588 50.9 360 49.4 1307 49.5

WHO performance level 0 505 68.3 808 71.4 536 75.2 1849 71.6 1 176 23.8 242 21.4 141 19.8 559 21.6 ⩾ 2 58 7.9 81 7.2 36 5.1 175 6.8

Ann Arbor disease stage I 196 26.7 230 20.7 138 20.1 564 22.3 II 95 12.9 151 13.6 112 16.3 358 14.1 III 149 20.3 290 26.1 204 29.7 643 25.4 IV 294 40.1 442 39.7 234 34.0 970 38.3

Serum lactate dehydrogenase Normal 487 69.3 778 71.4 497 69.8 1762 70.3 Elevated 216 30.7 312 28.6 215 30.2 743 29.7

B symptoms Absent 543 75.4 856 76.6 557 78.3 1956 76.8 Present 177 24.6 261 23.4 154 21.7 592 23.2

Bulky disease Absent 591 81.4 973 87.1 608 86.5 2172 85.3 Present 136 18.6 144 12.9 95 13.5 374 14.7

FLIPI Low risk —— 10 32.3 202 33.8 212 33.8 Intermediate risk —— 10 32.3 196 32.8 206 32.8 High risk —— 11 35.5 199 33.3 210 33.4 Abbreviations: FLIPI, follicular lymphoma International Prognostic Index; WHO, World Health Organization.

Table 2. Reported ﬁrst-line therapy

2000–2002 2003–2007 2008–2010

No. % Of % Of systemic No. % Of % Of systemic No. % Of % Of systemic all therapy all therapy all therapy

Reports on therapy 650 100.0 981 100.0 583 100.0 No therapy given (at the time of report) 226 34.8 302 30.8 144 24.7 Unspecified antitumoral therapy 279 42.9 80 8.2 0 0.0 Specified local radiation only 33 5.1 105 10.7 74 12.7 Specified systemic therapy 112 17.2 100.0 494 50.4 100.0 365 62.6 100.0 Chlorambucil 21 18.8 74 15.0 39 10.7 CHOP 70 62.6 78 15.8 5 1.4 Other chemotherapy 6 5.4 43 8.7 8 2.2 Rituximab without chemotherapy 4 3.6 99 20.0 89 24.4 Rituximab-CHOP 9 8.0 160 32.4 160 43.8 Rituximab-other chemotherapy 2 1.8 40 8.1 64 17.5 Rituximab containing 15 13.4 299 60.5 313 85.8 Abbreviation: CHOP, cyclophosphamide-hydroxydaunorubicin-vincristine-prednisone.

1.00 1.00

Fast rituximab 2008-2010 adopter 0.75 0.75

2003-2007 Moderate Slow rituximab rituximab adopter 0.50 0.50 adopter

2000-2002 Overall survival Overall survival 0.25 0.25

0.00 0.00 0 5 10 15 0 5 10 15 Years Years No. of patients at risk No. of patients at risk 2000-2002 757 517 383 Fast 81 65 13 2003-2007 1155 845 153 Moderate 958 702 125 2008-2010 729 195 0 Slow 116 80 17

1.00 1.00

2008-2010 0.75 0.75 2003-2007

0.50 0.50

’00-’02 ’03-’10 2000-2002 Overall survival 0.25

18-49 Overall survival 50-59 0.25 60-69 ≥ 0.00 70 0 5 10 15 0.00 Years 0 5 10 15 No. of patients at risk 18-49, ’00-’02 100 90 83 Years 18-49, ’03-’10 244 175 28 No. of patients at risk 50-59, ’00-’02 189 162 134 2000-2002 398 267 203 50-59, ’03-’10 347 243 41 2003-2007 567 415 83 60-69, ’00-’02 187 143 111 2008-2010 369 101 0 60-69, ’03-’10 566 346 57 ≥70, ’00-’02 281 125 58 ≥70, ’03-’10 712 268 30

1.00 1.00

2008-2010 0.75 0.75 Women 2003-2007

0.50 0.50 2000-2002

Overall survival Men

0.25 Overall survival 0.25

0.00 0.00 0 5 10 15 0 5 10 15 Years No. of patients at risk Years 2000-2002 359 251 181 No. of patients at risk 2003-2007 588 431 72 Women 109 45 3 2008-2010 360 95 0 Men 86 32 3 Figure 1. OS of FL in Sweden. Kaplan–Meier graphs by (a) calendar period of diagnosis, (b) regional adoption of rituximab during 2003–2007, (c) age and the 2003 introduction of ﬁrst-line rituximab, (d) calendar period of diagnosis in women, (e) calendar period of diagnosis in men and (f) gender in patients ⩾ 70 years who received rituximab in ﬁrst-line therapy. with fast adoption of rituximab and inferior in the region with slow differences in clinical characteristics between the three categories adoption (P = 0.039; Figure 1b). This was also demonstrated in of region. multivariable analysis (Supplementary Information). No OS differences were detected between the categories of region in patients Improvement in grades 1 and 2. To ensure an analysis of patients diagnosed 2000–2002 (P = 0.72) or 2008–2010 (P = 0.58). Patients with only low-grade lymphoma, we conducted another analysis of diagnosed 2003–2007 in the slow-adopting region showed higher OS in patients with the exclusion of FL grade 3 (n = 430). Also, in lactate dehydrogenase levels (P = 0.001), but there were no other this grade 1 and 2 subset (n = 2211), OS improved with each

© 2015 Macmillan Publishers Limited Leukemia (2015) 668 – 676 Follicular lymphoma survival in Sweden 2000–2010 HR Junlén et al 672 calendar period (multivariable P = 0.0005) and, during the transi- analysis (HR for men, 1.8; 95% CI, 1.1–2.7; P = 0.013), also when tional period, with increased regional use of rituximab (multi- extending the analysis to include patients ⩾ 60 years (HR, 1.5, 95% variable P = 0.011; Supplementary Figure). CI, 1.1–2.2; P = 0.026; n = 398). There were no OS differences between genders in patients o60 years who received first-line Improvement in different age groups. Improvement in OS after rituximab (P = 0.36). the introduction of first-line rituximab was seen in all age groups (Figure 1c). OS rates at 10 years are shown in Table 3. In patients Relative survival o60 years (median 52 years), 5- and 10-year OS rates were 87% With each calendar period of diagnosis, the RSRs improved in all and 74% in the pre-rituximab era (2000–2002) and 91% and 83% – ⩾ age groups in general, and in the elderly patients in particular, in the rituximab era (2003 2010), and in those 70 years (median although patients ⩾ 70 years maintained a higher excess mortality 77 years), 5- and 10-year OS rates were 44% and 21% (2000–2002) – than others (Figure 2). Also, cumulative RS increased with the and 53% and 32% (2003 2010), respectively. Survival rates at 5 calendar periods (Figures 3 and 4). Excess mortality rate ratios and 10 years are given in detail in Supplementary Information. were significantly lower in later calendar periods and in younger patients (both Po0.00005; multivariable analysis in Improvement in women and men. Although OS improved by each Supplementary Information). RSRs at 10 years are given in Table 3. calendar period in both women (univariable and multivariable o In patients diagnosed during the transitional period between 2003 P 0.00005) and men (univariable P = 0.046; multivariable and 2007, RS differed between the three categories of region, with P = 0.047), the improvement was greater in women (Figures 1d the best outcome in the region with fast adoption of rituximab and e). Furthermore, male sex was noted as an independent and the worst in the region with slow adoption of rituximab adverse factor in OS multivariable analysis (Supplementary (Figure 4b). There were no RS differences between these regions fi Information). In multivariable analysis strati ed by calendar period, in patients diagnosed 2000–2002 or 2008–2010 (data not shown). male sex was not adverse in the prerituximab era (hazard ratio With each calendar period, women showed a larger improvement – (HR), 1.1; 95% CI, 0.9 1.4; P = 0.26), showed a trend in the in RS than men (Figure 4c), because of increasingly divergent – transitional period (HR, 1.2; 95% CI, 1.0 1.5; P = 0.069), but was gender outcomes in higher ages (Supplementary Information). fi highly signi cant in the established period (HR, 1.7; 95% CI, Finally, we performed a multivariable RS analysis of patients – 1.2 2.4; P = 0.005). Gender differences were seen in two clinical diagnosed in the rituximab era (2003–2010), stratified by age: fi characteristics: men were signi cantly younger both in the among patients ⩾ 60 years, male sex was an independent adverse prerituximab (P = 0.003) and rituximab (P = 0.005) era (overall, factor for higher excess mortality at 5 (P = 0.007) and 10 (P = 0.020) median 64 vs 66 years). Men also presented more cases of bulky years (multivariable analysis in Supplementary Information), but disease, both in the prerituximab (P = 0.008) and rituximab not among patients o60 years (P = 0.50 and P = 0.89 at 5 and 10 (P = 0.004) era (overall, 17% vs 12%). In age-stratified multivariable years, respectively). analysis of the 1884 patients diagnosed in the rituximab era (2003–2010), male sex was a significant risk factor among those 60 years or older (P = 0.0004), but in those younger than 60 years men DISCUSSION did not have inferior OS (P = 0.69). In the rituximab era, survival In this large, population-based, nationwide study of the 2641 rates at 10 years were similar between the genders in younger Swedish patients with FL, there was a significant survival patients, but diverged in the elderly (Supplementary Table). improvement in patients diagnosed during 2000–2010. This Among patients 60–69 years, primary therapy with rituximab was improvement correlated with increasing use of first-line rituximab reported as equally common in women (69%) and men (70%). over time and with regional differences in first-line rituximab use. In those ⩾ 70 years, women appeared to have received more first-line rituximab than men (60% vs 49%; P = 0.054). Still, an analysis restricted to the 195 patients ⩾ 70 years who actually 5-year relative survival 10-year relative survival received first-line rituximab showed better OS in women (Figure 1f). This difference was independent in multivariable 18-49

50-59 18-49 0.8 Table 3. Overall and relative survival in Sweden at 10 years in age 60-69 groups, stratiﬁed by era 50-59 60-69 Analysis Age group Prerituximab era Rituximab era (years) (2000–2002) (2003–2010) 0.6 ≥70 Rate (95% CI) Rate (95% CI)

Overall survival 0.4 Relative survival ≥70 All 51% (47–54) 59% (56–62) 18–49 82% (73–88) 90% (85–94) 50–59 70% (63–76) 78% (71–83) 60–69 59% (51–65) 68% (61–73) ⩾ 70 21% (16–25) 32% (27–38) 0.2

Relative survival All 63% (59–68) 76% (72–80) 18–49 84% (76–92) 92% (87–96) 50–59 75% (68–82) 83% (77–90) 2000−2002 2003−2007 2008−2010 2000−2002 2003−2007 60–69 69% (61–78) 78% (72–86) ⩾ 70 40% (31–50) 64% (55–76) Period of diagnosis Abbreviation: CI, conﬁdence interval. Figure 2. Relative survival of FL in Sweden. At 5 and 10 years in calendar groups by age category (years).

2000-2002 2003-2007 2008-2010

1.0 18-49 18-49 50-59

60-69 50-59 0.8 18-49

60-69 50-59 ≥70

≥70 60-69 0.6 Relative survival Relative

≥70 0.4

0 5 10 0 5 10 0 5 Time since diagnosis (years)

18-49 years 50-59 years 1.0 2008-2010 2003-2007 2008-2010 2003-2007 0.8 2000-2002

2000-2002 0.6

0.4

60-69 years ≥70 years 1.0

2008-2010 Relative survival ratio Relative

0.8 2008-2010 2003-2007

2003-2007 0.6 2000-2002

0.4 2000-2002

0 5 10 0 5 10 Time since diagnosis (years) Figure 3. Cumulative RS of FL in Sweden. (a) Cumulative RS in calendar periods by age category (years). (b) Cumulative RS in age categories by calendar period.

Elderly patients, particularly female ones, appeared to benefit the that 5-year OS was 95% in their cohort treated with rituximab (the most from the introduction of rituximab. previous cohort had about 80% 5-year OS).15 More recently, the Our study period encompasses a tectonic shift in FL treatment: Stanford group presented data with longer follow-up (6.1 years) in the first years of the study, about 1 out of 10 patients received for their 257 patients diagnosed between 1997 and 2003; in total, rituximab in systemic primary therapy, whereas in the past years, 9 18% of those patients received first-line rituximab either as out of 10 did. This accompanied a greatly improved outcome. monotherapy or combined with chemotherapy and they showed Several single-center studies have shown a benefit from rituximab. a 73% 10-year OS.16 The median ages in these three single-center Using different trial cohorts, the Southwest Oncology Group cohorts were 52, 50, and 50 years. Thus, a direct comparison already in 2005 demonstrated that 4-year OS was better in between them and the present nationwide, population-based patients given CHOP+monoclonal antibody (91%) than in those cohort with a median age of 65 is impossible. In the 591 Swedish given CHOP only (69%).14 Similarly, MD Anderson’s report showed patients o60 years (median age, 52 years) diagnosed in the

2008-2010

0.8 2003-2007

0.6 2000-2002

0.4 Relative survival Relative

0.2

0 2 4 6 8 10 12 14 Years

Fast rituximab adopter

0.8 Moderate rituximab adopter

0.6 Slow rituximab adopter

0.4 Relative survival Relative

0.2

2 4 6 8 10 Years

0.8

0.6

Women, 2008-2010 0.4 Men, 2008-2010 Relative survival Relative Women, 2003-2007 Men, 2003-2007 0.2 Women, 2000-2002 Men, 2000-2002

0 2 4 6 8 10 12 14 Years Figure 4. Cumulative RS of FL in Sweden. (a) By calendar period of diagnosis. (b) By regional adoption of rituximab during 2003–2007. (c)By calendar period of diagnosis and gender.

Leukemia (2015) 668 – 676 © 2015 Macmillan Publishers Limited Follicular lymphoma survival in Sweden 2000–2010 HR Junlén et al 675 rituximab era OS was 92% and 83% at 5 and 10 years, respectively. The main strengths of our population-based material are its lack Also, a Swiss single-center study showed improved disease- of bias and its size, describing a nationwide pattern of improved specific survival in the rituximab era.34 Using data from all outcome after the introduction of first-line rituximab (either as randomized rituximab trials, a Cochrane meta-analysis demon- monotherapy or combined with chemotherapy; Table 2). Because strated better OS with the addition of rituximab (HR for OS, 0.60).35 we based our analyses on calendar periods and geographical Our population-based data agree with all studies above. regions, we minimized the selection bias that is always present Compared with the prerituximab period, the HR was 0.67 (OS) when retrospectively comparing different therapeutic algorithms. for the period when rituximab had been established as first-line Our analysis made use of a natural experiment induced by the therapy. During the transitional period, patients diagnosed in the regionally different introduction rates of first-line rituximab. The region with fast adoption of rituximab had an HR of 0.46 (OS) aggressive45 grade 3B was specified from June 2007 and onwards compared with those diagnosed in the region with slow adoption (in total, 31 grade 3B cases; about 4% of all patients reported with (multivariable analyses in Supplementary Information). the new form). Therefore, an exclusion of grade 3B would have An interesting finding in this study was that male sex has introduced bias. Nonetheless, also an analysis restricted to the 2212 become an independent adverse factor for survival. When closely non-grade 3 cases showed improved OS with each time period and examined, the increased risk emerged in the rituximab era and it with increasing regional rituximab use. A caveat is that some was only seen in patients ⩾ 60 years. The development of a patients are diagnosed within one calendar period and, being gender difference in B-cell lymphoma outcome in the era of asymptomatic for some years, they might subsequently be treated immunochemotherapy has been reported earlier.36–40 The reason according to an algorithm of a later calendar period. However, that for the gender difference in diffuse large B-cell lymphoma has effect would only dilute the large differences between the calendar been explained by a significantly lower rituximab clearance in periods, and it would not lead to false conclusions. Also, in this women than in men and thus a prolonged elimination half-life.41 indolent disease, survival is affected not only by first-line therapy A more recent study by the same group showed that the female but also by second- and third-line regimens, and we only had superiority in pharmacokinetics and outcome of immunochemo- access to first-line therapy data, which were not given in complete therapy-treated diffuse large B-cell lymphoma was only seen detail in the first calendar period. Still, the reported data agree among the elderly patients (no difference in outcome or exactly with what was expected from the history of Swedish first- pharmacokinetics was seen between young women and men).42 line treatment algorithms. Our results clearly show that survival The influence of gender and age on the pharmacokinetics of rates were highest in patients who were diagnosed when rituximab rituximab should be the same in FL. Although this study is the first was established as part of first-line treatment nationwide. to show male sex as an independent adverse factor in a large This study of all Swedish FL patients diagnosed 2000–2010 is cohort of FL patients, a greater OS benefit in women than in men the first nationwide analysis of recent prognostic improvements. after rituximab has been shown previously.16 Taken together, the We show that after the introduction of rituximab, in unselected findings in diffuse large B-cell lymphoma, mantle cell lymphoma patients the 10-year RS is 92% for those 18–49 years, 83% for and FL imply that the current intravenous rituximab 375 mg/m2 those 50–59, 78% for those aged 60–69 years and 64% in patients dosing is suboptimal for younger female and all male patients ⩾ 70 years. Rituximab should also be given to older patients (the with CD20+ lymphoma. This must be taken into account when median age was 77 years in the ⩾ 70 years category) who showed rituximab is compared with newer CD20 antibodies that are given much better outcome after the drug was introduced, although in higher doses. It remains to be seen whether men will continue they still have inferior RS. Among the elderly, men now show the to show inferior outcome with the 1400 mg dosing of the recently highest excess mortality, because outcome in elderly women have registered subcutaneous rituximab. We do not know the optimal improved more. Rituximab might be inadequately dosed in rituximab dose(s) for FL patients of different ages and genders. younger women and in men of all ages. Increased toxicity was not reported when chronic lymphocytic leukemia patients were treated with 1500 mg/m2 intravenous rituximab.43 CONFLICT OF INTEREST Another noteworthy finding is that the median age at diagnosis The authors declare no conflict of interest. increased over time and that the overall median of 65 years is significantly higher than the peak incidence between 50 and 60 ACKNOWLEDGEMENTS years reported in reference literature.1 This is likely an effect of the source population becoming older. Swedes are one of the oldest We thank the Swedish Lymphoma Registry and the Swedish Lymphoma Group for nations in the world, with a median age of 40.8 years in 2010 providing data for this study. This work was funded by grants from Cancerfonden, (furthermore, the national median ages in women (41.8 years) and Svenska Sällskapet för Medicinsk Forskning (SSMF) and KI/SLL (a joint grant from men (39.7 years) explain why the median Swedish female FL Karolinska Institutet and Stockholm County Council). patient is 2 years older than the median male).44 The high median age of recruited patients in recent first-line trials from other REFERENCES countries (e.g., 70 years in the bendamustine trial28) also suggests ’ 1 Harris NL, Swerdlow SH, Jaffe ES, Ott G, Nathwani D. Follicular lymphoma. that, in today s western world, the median age at diagnosis must In: Swerdlow SH, Campo E, Harris NL, Jaffe ES, Pileri SA, Stein H (eds). WHO be over 60 years. Although patients became older at diagnosis, Classification of Tumours of Haematopoietic and Lymphoid Tissues. IARC: Lyon, they were not more ill. Rather, fewer showed poor performance France, 2008, pp 220–226. status at diagnosis, probably reflecting earlier diagnoses; this 2 Hiddemann W, Cheson BD. How we manage follicular lymphoma. Leukemia 2014; could be because of earlier self-admission (higher patient 28: 1388–1395. awareness) and increasing use of computerized tomography by 3 Bezombes C, Fournie JJ, Laurent G. Direct effect of rituximab in B-cell-derived general practitioners. Still, the improving OS during the 2000s was lymphoid neoplasias: mechanism, regulation, and perspectives. Mol Cancer Res 9 – independent of all clinical prognostic factors, including age and 2011; :1435 1442. WHO performance status (Supplementary Information). An 4 McLaughlin P, Grillo-Lopez AJ, Link BK, Levy R, Czuczman MS, Williams ME et al. Rituximab chimeric anti-CD20 monoclonal antibody therapy for relapsed indolent improving health status in the general population must also be lymphoma: half of patients respond to a four-dose treatment program. J Clin taken into account when one considers OS by calendar periods. Oncol 1998; 16: 2825–2833. However, also RS increased with each calendar period, verifying 5 Marcus R, Imrie K, Belch A, Cunningham D, Flores E, Catalano J et al. CVP that the longer survival was independent of any changes in the chemotherapy plus rituximab compared with CVP as first-line treatment for general population. advanced follicular lymphoma. Blood 2005; 105:1417–1423.

© 2015 Macmillan Publishers Limited Leukemia (2015) 668 – 676 Follicular lymphoma survival in Sweden 2000–2010 HR Junlén et al 676 6 Hiddemann W, Kneba M, Dreyling M, Schmitz N, Lengfelder E, Schmits R et al. 25 Wahlin BE, Sundstrom C, Sander B, Christensson B, Jeppsson-Ahlberg A, Frontline therapy with rituximab added to the combination of cyclophosphamide, Hjalmarsson E et al. Higher World Health Organization grades of follicular doxorubicin, vincristine, and prednisone (CHOP) significantly improves the lymphoma correlate with better outcome in two Nordic Lymphoma Group trials outcome for patients with advanced-stage follicular lymphoma compared with of rituximab without chemotherapy. Leuk Lymphoma 2014; 55:288–295. therapy with CHOP alone: results of a prospective randomized study of the 26 van Oers MH, Klasa R, Marcus RE, Wolf M, Kimby E, Gascoyne RD et al. Rituximab German Low-Grade Lymphoma Study Group. Blood 2005; 106: 3725–3732. maintenance improves clinical outcome of relapsed/resistant follicular 7 Herold M, Haas A, Srock S, Neser S, Al-Ali KH, Neubauer A et al. Rituximab added non-Hodgkin lymphoma in patients both with and without rituximab during to first-line mitoxantrone, chlorambucil, and prednisolone chemotherapy fol- induction: results of a prospective randomized phase 3 intergroup trial. Blood lowed by interferon maintenance prolongs survival in patients with advanced 2006; 108:3295–3301. follicular lymphoma: an East German Study Group Hematology and 27 Salles G, Seymour JF, Offner F, Lopez-Guillermo A, Belada D, Xerri L et al. Oncology Study. J Clin Oncol 2007; 25: 1986–1992. Rituximab maintenance for 2 years in patients with high tumour burden follicular 8 Salles G, Mounier N, de Guibert S, Morschhauser F, Doyen C, Rossi JF et al. lymphoma responding to rituximab plus chemotherapy (PRIMA): a phase 3, Rituximab combined with chemotherapy and interferon in follicular lymphoma randomised controlled trial. Lancet 2011; 377:42–51. patients: results of the GELA-GOELAMS FL2000 study. Blood 2008; 112: 28 Rummel MJ, Niederle N, Maschmeyer G, Banat GA, von Grunhagen U, Losem C 4824–4831. et al. Bendamustine plus rituximab versus CHOP plus rituximab as first-line 9 Hainsworth JD, Burris III HA , Morrissey LH, Litchy S, Scullin Jr DC, Bearden JD III treatment for patients with indolent and mantle-cell lymphomas: an open-label, et al. Rituximab monoclonal antibody as initial systemic therapy for patients with multicentre, randomised, phase 3 non-inferiority trial. Lancet 2013; 381: low-grade non-Hodgkin lymphoma. Blood 2000; 95: 3052–3056. 1203–1210. 10 Colombat P, Salles G, Brousse N, Eftekhari P, Soubeyran P, Delwail V et al. 29 Cox DR. Regression models and life-tables. J R Stat Soc 1972; 34: 187–220. Rituximab (anti-CD20 monoclonal antibody) as single first-line therapy for 30 Henson DE, Ries LA. The relative survival rate. Cancer 1995; 76: 1687–1688. patients with follicular lymphoma with a low tumor burden: clinical and molecular 31 Dickman PW, Adami HO. Interpreting trends in cancer patient survival. J Intern evaluation. Blood 2001; 97:101–106. Med 2006; 260:103–117. 11 Ghielmini M, Schmitz SF, Cogliatti SB, Pichert G, Hummerjohann J, Waltzer U et al. 32 Ederer F, Heise H. Instructions to IBM 650 Programmers in Processing Survival Prolonged treatment with rituximab in patients with follicular lymphoma Computations. Methodological Note No. 10, End Results Evaluation Section. National significantly increases event-free survival and response duration compared with Cancer Institute: Bethesda, MD, USA, 1959. the standard weekly x4 schedule. Blood 2004; 103: 4416–4423. 33 Perme MP, Henderson R, Stare J. An approach to estimation in relative survival 12 Kimby E, Jurlander J, Geisler C, Hagberg H, Holte H, Lehtinen T et al. Long-term regression. Biostatistics 2009; 10:136–146. molecular remissions in patients with indolent lymphoma treated with rituximab 34 Conconi A, Motta M, Bertoni F, Piona C, Stathis A, Wannesson L et al. Patterns of as a single agent or in combination with interferon alpha-2a: a randomized phase survival of follicular lymphomas at a single institution through three decades. II study from the Nordic Lymphoma Group. Leuk Lymphoma 2008; 49:102–112. Leuk Lymphoma 2010; 51: 1028–1034. 13 Johnson PW, Rohatiner AZ, Whelan JS, Price CG, Love S, Lim J et al. Patterns of 35 Schulz H, Bohlius JF, Trelle S, Skoetz N, Reiser M, Kober T et al. Immunochemo- survival in patients with recurrent follicular lymphoma: a 20-year study from a therapy with rituximab and overall survival in patients with indolent or mantle single center. J Clin Oncol 1995; 13:140–147. cell lymphoma: a systematic review and meta-analysis. J Natl Cancer Inst 2007; 99: 14 Fisher RI, LeBlanc M, Press OW, Maloney DG, Unger JM, Miller TP. New treatment 706–714. options have changed the survival of patients with follicular lymphoma. J Clin 36 Tomita N, Kodama F, Motomura S, Koharazawa H, Fujita H, Harano H et al. Oncol 2005; 23: 8447–8452. Prognostic factors in diffuse large B-cell lymphoma treated by risk-adopted 15 Liu Q, Fayad L, Cabanillas F, Hagemeister FB, Ayers GD, Hess M et al. Improvement therapy. Intern Med 2006; 45:247–252. of overall and failure-free survival in stage IV follicular lymphoma: 25 years of 37 Ngo L, Hee SW, Lim LC, Tao M, Quek R, Yap SP et al. Prognostic factors in patients treatment experience at The University of Texas MD Anderson Cancer Center. with diffuse large B cell lymphoma: before and after the introduction of rituximab. J Clin Oncol 2006; 24: 1582–1589. Leuk Lymphoma 2008; 49: 462–469. 16 Tan D, Horning SJ, Hoppe RT, Levy R, Rosenberg SA, Sigal BM et al. Improvements 38 Riihijarvi S, Taskinen M, Jerkeman M, Leppa S. Male gender is an adverse in observed and relative survival in follicular grade 1-2 lymphoma during 4 prognostic factor in B-cell lymphoma patients treated with immunochemotherapy. decades: the Stanford University experience. Blood 2013; 122:981–987. Eur J Haematol 2011; 86:124–128. 17 Nationellt kvalitetsregister för lymfom. Available at: http://www.cancercentrum.se/ 39 Abrahamsson A, Albertsson-Lindblad A, Brown PN, Baumgartner-Wennerholm S, sv/INCA/kvalitetsregister/Blodcancer/lymfom (last accessed 30 December 2013). Pedersen LM, D'Amore F et al. Real world data on primary treatment for mantle 18 Svenska lymfomgruppen. Available at: http://www.swedishlymphoma.se cell lymphoma: a Nordic Lymphoma Group observational study. Blood 2014; 124: (last accessed 30 December 2013). 1288–1295. 19 Solal-Celigny P, Roy P, Colombat P, White J, Armitage JO, Arranz-Saez R et al. 40 Carella AM, de Souza CA, Luminari S, Marcheselli L, Chiappella A, di Rocco A et al. Follicular lymphoma international prognostic index. Blood 2004; 104: 1258–1265. Prognostic role of gender in diffuse large B-cell lymphoma treated with rituximab 20 Vijungco J, Phillips R, Hendrickson FR, Millburn LF. Stage I and II non-Hodgkin's containing regimens: a Fondazione Italiana Linfomi/Grupo de Estudos em lymphoma. Results of regional radiation therapy. Am J Roentgenol Radium Ther Molestias Onco-Hematologicas retrospective study. Leuk Lymphoma 2013; 54: Nucl Med 1973; 117:45–49. 53–57. 21 Ardeshna KM, Smith P, Norton A, Hancock BW, Hoskin PJ, MacLennan KA et al. 41 Muller C, Murawski N, Wiesen MH, Held G, Poeschel V, Zeynalova S et al. The role Long-term effect of a watch and wait policy versus immediate systemic treatment of sex and weight on rituximab clearance and serum elimination half-life in for asymptomatic advanced-stage non-Hodgkin lymphoma: a randomised elderly patients with DLBCL. Blood 2012; 119:3276–3284. controlled trial. Lancet 2003; 362: 516–522. 42 Pfreundschuh M, Muller C, Zeynalova S, Kuhnt E, Wiesen MH, Held G et al. Sub- 22 Brandt L, Kimby E, Nygren P, Glimelius B. A systematic overview of chemotherapy optimal dosing of rituximab in male and female patients with DLBCL. Blood 2014; effects in indolent non-Hodgkin's lymphoma. Acta Oncol 2001; 40: 213–223. 123:640–646. 23 Schouten HC, Qian W, Kvaloy S, Porcellini A, Hagberg H, Johnson HE et al. High- 43 O'Brien SM, Kantarjian H, Thomas DA, Giles FJ, Freireich EJ, Cortes J et al. dose therapy improves progression-free survival and survival in relapsed follicular Rituximab dose-escalation trial in chronic lymphocytic leukemia. J Clin Oncol non-Hodgkin's lymphoma: results from the randomized European CUP trial. J Clin 2001; 19: 2165–2170. Oncol 2003; 21: 3918–3927. 44 Statistiska centra lbyrån. Available at: http://www.scb.se (last accessed 24 Williams CD, Harrison CN, Lister TA, Norton AJ, Blystad AK, Coiffier B et al. High- 1 August 2012). dose therapy and autologous stem-cell support for chemosensitive transformed 45 Wahlin BE, Yri OE, Kimby E, Holte H, Delabie J, Smeland EB et al. Clinical low-grade follicular non-Hodgkin's lymphoma: a case-matched study from the significance of the WHO grades of follicular lymphoma in a population-based European Bone Marrow Transplant Registry. J Clin Oncol 2001; 19: 727–735. cohort of 505 patients with long follow-up times. Br J Haematol 2012; 156:225–233.

Supplementary Information accompanies this paper on the Leukemia website (http://www.nature.com/leu)