Retrospective Studies in Very Elderly DLBCL Patients
The difficulty of designing prospective studies in the population of very elderly patients with DLBCL has led researchers to analyze outcomes in the very elderly by conducting retrospective analyses. However, few of these reports included exclusively patients over the age of 80, since every analysis defined “elderly” differently. Hasselblom et al performed a retrospective study of 535 DLBCL patients (median age, 73 years) diagnosed in western Sweden between 1995 and 2000.[47] High IPI score and male sex had a negative impact on OS and PFS. Treatment with a curative intent was administered to only 376 patients (70%), without details as to why others did not receive curative therapy. However, this analysis was not restricted to patients > 80 years. These same investigators subsequently evaluated the outcome of a small cohort of patients > 80 who were treated in two sequential eras: before rituximab(Drug information on rituximab) (1997 to 2000) and after rituximab (2006 to 2009).[48] Forty patients were identified with an improved 3-year PFS—and 30 patients with an improved OS—in the post-rituximab period compared to before (P = .01 for both). However, only 53% of patients received treatment with a curative intent. Thieblemont et al retrospectively studied 205 NHL patients over the age of 80 (median age, 83 years) to evaluate prognostic factors and outcomes.[49] Of these, 82 patients (40%) had DLBCL, 87% had at least one comorbidity, and 43% had a low Charlson Index score. Of the evaluable patients, 15% did not receive any treatment or received corticosteroids alone; 35% had surgery, radiotherapy, or mono-chemotherapy; 18% received poly-chemotherapy without anthracycline; and 32% received anthracycline-based therapy. Median OS was 1.3 years, and the main reason for death was disease progression (57%). Factors that were independently prognostic of survival were poor ECOG-PS (P < .0001) and high LDH level (P < .00001); comorbidities were not found to influence survival.
A larger multicenter retrospective analysis of 303 NHL patients ≥ 80 years treated between 1999 and 2009 was recently reported.[18] Of these patients, 127 (42%) had DLBCL. In this study, comorbidities were present in 80% of patients; 26% presented with at least one geriatric syndrome at initial diagnosis. Dementia was the most common geriatric syndrome, identified in 21 patients (26%), followed by osteoporosis in 18 (23%), depression in 16 (20%), falls in 15 (19%), and incontinence in 10 (12%). Additionally, 41 patients (14%) had a loss of at least one ADL; loss of ADLs was more common with aggressive compared with indolent NHLs (18% vs 6%, respectively; P = .002). At a 49-month median follow-up, 4-year PFS and OS for aggressive NHLs were 31% and 44%, respectively (stage I/II: PFS 53% and OS 66%; stage III/IV: PFS 20% and OS 32%; P < .0001 and P < .0002, respectively). Multivariate regression analysis identified two key factors that predicted inferior PFS and OS: lack of complete remission and loss of ADLs.[18] Exact details regarding the dose intensity and density used in treating these patients were not clearly identified.
Dose density and intensity proved important in providing patients with better outcomes. Of 109 patients with NHL > 80 years who were studied by Bairey et al, 66 (61%) had DLBCL; these patients were treated at three Israeli institutions between 1984 and 2004. While 18 patients out of the entire cohort of 109 patients (17%) received no therapy, 37 (34%) received radiation at some point (with or without chemotherapy) and 79 (72%) received chemotherapy. Only half of the chemotherapy-treated patients were able to complete the full prescribed program.[50] Complications were frequent, mostly hematologic toxicity, despite the administration of growth factors to one-third of patients. Patients who received more than 80% of the calculated combination chemotherapy dose (38% of patients) had a longer median survival than those who did not (72 vs 20 months). Patients who received any form of therapy had better OS than those who received no therapy (20 months vs 13 months), despite the fact that 26% of patients died from progressive disease.[50]
Van de Schans et al retrospectively investigated data on 419 patients with advanced-stage DLBCL, aged 75 years or older, who were diagnosed between 1997 and 2004 and were included in five regional population-based cancer registries in the Netherlands.[51] Only 46% of the patients received the standard therapy, and only 56% of these completed the planned full 6 cycles of treatment. Reasons for withholding chemotherapy were refusal by patient/family, poor PS, or estimated short life expectancy. Toxicities were observed in 67% of patients. Among patients older than 85 years, only 4% received standard optimal therapy. Among patients who received CHOP-like therapy, 73% of patients aged 80 to 84 years and 85% of patients aged > 85 years experienced hematologic toxicity. The 5-year survival rate for patients who received at least 6 cycles of CHOP-like therapy was 38%, compared with 22% in patients who received fewer than 6 cycles, and 4% in patients who received no chemotherapy.[51] While details on patients who were older than 80 years were not available (although they were available in those who were 75 to 80 years of age), these data suggest that, even in the elderly, full prescribed courses of therapy likely provide better outcomes than abbreviated courses when treating advanced disease.
Collectively, these studies demonstrate that very elderly patients are treated less frequently, experience more toxicity, and have less favorable outcomes, with worse OS and PFS, compared with younger patients. Importantly, progressive disease was a major cause of death in the very elderly. However, these studies have significant limitations. The definition of “elderly” was not uniform. Some studies reported on patients over age 65 while others suggested an age of 75 as a cutoff. Only a few studies included very elderly patients (> 80 years) exclusively. Furthermore, not all studies evaluated these very elderly patients using CGA and other geriatric scales, since the data needed to comprehensively conduct these evaluations were missing due to the retrospective nature of these analyses. Nonetheless, these trials paved the way for a few prospective phase II trials that have targeted this patient population.
Prospective Studies in Elderly Patients With DLBCL
Selective Prospective Studies of DLBCL in the Elderly
Since pivotal prospective DLBCL studies have generally defined “elderly” as 60 to 80 years, no uniform consensus exists as to how best to treat the very elderly (> 80 years). Some argue that replacing the R-CHOP (rituximab [Rituxan] plus CHOP) regimen with a less toxic non-anthracycline program is preferred, while others simply implement dose reductions of various R-CHOP components. In the absence of prospective randomized studies, it is unclear whether one strategy is superior to another. Importantly, few prospective studies (Table 2) have included exclusively patients older than 80 years, and all of these trials were small, phase II, single-institution studies.
In patients who had an ECOG-PS of 2 or less, Peyrade et al investigated the efficacy and safety of a decreased dose of CHOP (cyclophosphamide, 400 mg/m2; doxorubicin(Drug information on doxorubicin), 25 mg/m2; vincristine, 1 mg; and prednisone(Drug information on prednisone), 40 mg/m2) with a conventional dose of rituximab (375 mg/m2) in 150 DLBCL patients over the age of 80 (median age, 83 years).[52] Advanced-stage disease was present in 75% of treated patients. After a median follow-up of 20 months (range, 0 to 45 months), median OS was 29 months (95% confidence interval [CI], 21 to upper limit not reached); 2-year OS was 59% (95% CI, 49%–67%). In multivariate analyses, OS was only affected by a serum albumin concentration of 35 g/L or less (hazard ratio, 3.2; 95% CI, 1.4–7.1; P = .0053). Median PFS was 21 months (95% CI, 13 to upper limit not reached), with a 2-year PFS of 47% (95% CI, 38%–56%). There were 58 deaths reported, 33 of which were secondary to lymphoma progression. Twelve deaths (8%) were attributed to treatment toxicity. The most frequent side effect was hematologic toxicity (grade ≥ 3 neutropenia in 59 patients; febrile neutropenia in 11 patients). Others have also looked at reducing the intensity of CHOP. Meguro et al treated 61 DLBCL patients over the age of 70 (15 patients > 80 years) with 70%-dose CHOP plus full-dose rituximab for 6–8 cycles.[53] When these patients were compared to younger counterparts (< 70 years) who were treated with full-dose CHOP, 3-year OS and PFS were similar between both groups, suggesting that the use of CHOP in elderly patients might demonstrate similar benefits despite lower dose intensity. Importantly, despite the dose reduction, elderly patients still experienced more adverse events and toxicities, which were mainly hematologic.
An alternative anthracycline-based program was investigated by Musolino et al when 23 patients (10 of whom were older than 80 years) received dose-adjusted infusional cyclophosphamide(Drug information on cyclophosphamide), doxorubicin, vincristine, and prednisone with rituximab (DA-POCH-R).[54] Treated patients had poor prognosis, defined as having at least two adverse risk factors based on the IPI. The overall response rate (ORR) was 90% (57% achieved a complete response[CR]), and the 3-year OS was 56%. Most adverse events were hematologic, specifically neutropenia in 48% of the patients. The number of patients over the age of 80 in this cohort is too small to specifically form a conclusion regarding their outcomes, however.
Non-anthracycline–based regimens were also studied. In the pre-rituximab era, Monfardini et al treated 30 frail patients with the combination of vinorelbine (25 mg/m2 on days 1 and 8) and prednisone (30 mg daily on days 1 through 8) every 3 weeks for 6 cycles.[55] The criteria for frailty were age older than 80 years, age older than 70 with three or more comorbidities of grade 3, or at least one comorbidity of grade 4 using the CIRS. The ORR was 40% (10% achieved a CR and 30% had a partial response [PR]), with 10 additional patients demonstrating stable disease. Three patients died of heart failure within 28 days of therapy. Median OS was 10 months. However, some patients in this study (24%) had different histologies, including mantle cell, peripheral T-cell, and follicular NHL. Weidmann et al studied the activity of bendamustine (Treanda) plus rituximab in 14 patients > 80 years of age (median age, 84 years) who were considered ineligible for R-CHOP or who refused aggressive treatment.[56] Ineligibility for R-CHOP was left to the investigator’s discretion without uniform criteria. Of 13 evaluable patients, 7 (54%) achieved CR, and 2 (15%) had a PR, with a median OS of 7.7 months. The major toxicity was also hematologic: 17% of the patients developed grade 3 neutropenia and 6% had grade 4.
Collectively, these studies demonstrate that anthracyclines can still be used in some patients and that other non-anthracycline–based regimens also have activity. Whether omitting anthracyclines in the very elderly has an adverse impact on their outcomes is yet to be seen and will depend on the results of future large prospective studies. Also, whether patients older than 80 years can be treated in a fashion similar to those between the ages of 70 and 80 remains unclear.
