Hodgkin lymphoma (HL) is one of the most curable malignancies in adults. However, survival rates for elderly patients with HL (often defined as ≥ 60 years of age) are inferior to those achieved by younger populations. The 5-year event-free survival or freedom from treatment failure rates for elderly HL range from 30% to 40%,[1-4] with 5-year overall survival rates ranging from 40% to 55%.[2-5] This compares to 5-year event-free survival rates of > 70% to 80% and overall survival rates of > 80% to 90% for patients aged < 40 years.[1,3-5]
Suboptimal staging and inadequate treatment delivery for older patients may compromise the rate of cure.[6,7] Furthermore, comorbidities may preclude the delivery of standard chemotherapy.[8,9] Intensive treatments such as BEACOPP (bleomycin, etoposide(Drug information on etoposide), doxorubicin(Drug information on doxorubicin) [Adriamycin], cyclophosphamide(Drug information on cyclophosphamide), vincristine [Oncovin], procarbazine(Drug information on procarbazine), and prednisone) are too toxic for older HL patients, while bleomycin-containing regimens including ABVD (doxorubicin, bleomycin, vinblastine, dacarbazine(Drug information on dacarbazine)) are often poorly tolerated.[4,7,11-14] Treatment delivery and comorbidity do not appear to completely explain the observed differences in outcome, however, implicating the biology of HL in the elderly.[15-17]
Elderly HL has been aptly described as an orphan disease, defined as a disease that affects fewer than 200,000 persons at a given time. Furthermore, elderly HL is a disease for which (1) there is no agreed upon standard treatment approach; (2) it is difficult to investigate in classical trial format (due to the rarity of the disease, the advanced age of patients, and their underrepresentation in clinical trials); (3) there are large numbers of subjects who are frail or have excessive comorbidities; and (4) the approach to curative treatment is absent or inconsistent. Clinical data regarding elderly HL are derived mainly from registries and retrospective series, which have included heterogeneous patient populations. In addition, several sources of evidence suggest a different biology in patients with elderly HL.[1,15-17] An improved understanding of the epidemiology, biology, outcomes, and treatment options for elderly HL is needed.
Within population-based studies, the proportion of HL patients aged 60 years or more has ranged from 15% to 30%.[1,5,13,20-22] However, the proportion of patients ≥ 60 years in clinical trials has been considerably lower, typically constituting < 5% to 10% of participants.[4,5,7]
Among two Swedish population reports from 1979 to 1988 and 1973 to 1994, 31% and 26% of all HL patients, respectively, were ≥ 60 years. The British National Lymphoma Investigation (BNLI) compared the number of patients aged ≥ 65 years enrolled in clinical trials with the total number of older patients in their population-based cancer registry (1970–1987). Approximately 15% of HL patients in the population were ≥ 65 years, whereas < 5% of patients in the same age group were entered onto BNLI trials. Investigators from the United Kingdom prospectively studied the incidence and outcomes of HL patients diagnosed between 1991 and 1998.[1,21] They reported that 20% of newly diagnosed HL patients were ≥ 60 years old. Moreover, the median overall survival of elderly HL patients was significantly inferior compared with a “normal” age- and sex-matched population (26 vs 150 months, respectively; P < .0001).
Among recent data from the US National Cancer Institute’s Surveillance, Epidemiology, and End Results (SEER) program, evidence of an age-related bimodal incidence pattern in HL persists (Figure 1). The increased incidence beyond age 65 is more pronounced for men vs women. Furthermore, in the 2000–2005 SEER-17 database, the incidence rates for elderly HL according to race are highest among Hispanics (6.5/100,000), followed by whites (4.5/100,000) and blacks (3.4/100,000). In addition, in part due to the inferior outcomes for older HL patients, the US prevalence among patients aged > 60 years is 15,400 compared with 140,700 for age < 30 years.
Epstein-Barr virus (EBV) is associated with a variety of hematopoietic malignancies including HL. Approximately 30% to 50% of all HL cases are EBV-positive, as identified by latent membrane protein (LMP)-1 or EBV-encoded RNA (EBER) immunostaining of the malignant cells.[24,25] The highest percentages of EBV-positive cases are seen in HL patients aged < 10 years or > 55 years.[15,24,25] Two population-based studies showed that the survival of older patients with EBV-positive tumors was significantly inferior compared with EBV-negative tumors.[1,15] Among 922 classical HL patients, Keegan et al showed that the presence of EBV independently predicted inferior survival for older adults (age 45–96 years), with significantly increased HL-specific mortality (hazard ratio [HR] for death = 2.5).
In a prospective population-based elderly HL study, Stark et al reported that 34% of cases were EBV-positive and that EBV status correlated with stage at presentation (EBV-positive: 9% early stage vs 50% advanced stage, P = .0006). Moreover, EBV-positive HL was associated with an inferior overall survival (median overall survival = 20 months vs not reached, respectively; P = .007). Recent reports have confirmed the inferior survival for older patients with EBV-positive HL vs EBV-negative disease.[26,27] Although the basis for this prognostic association is unknown, reduced EBV-specific cellular immunity of older patients has been proposed. The weaker immunocompetence in older patients may allow increased proliferation of an EBV-related tumor, and/or EBV antigens might induce a less robust immune response.[16,28]
Other pathologic and genetic features have emerged as potentially important prognostic factors in HL. Kelley and colleagues showed that HL patients with more intratumor T-cell regulatory cells (Tregs) and fewer activated cytotoxic T/NK lymphocytes had superior survival. Interestingly, older age (> 45 years) in that analysis was associated with decreased forkhead box P3 (FOXP3) Tregs and increased granzyme-B–positive cells compared with younger patients. Diepstra et al reported that the lack of HLA class II cell-surface expression on HRS cells was associated with inferior failure-free and overall survival. The only other factor in that study associated with increased risk of death was older age (> 65: relative risk of death = 6.47).
Prognosis andClinical Presentation
Prognosis: Age and IPS
Older age has been a consistent adverse prognostic factor for HL survival in population studies and clinical trial reports. The outcome for patients on the randomized Cancer and Leukemia Group B (CALGB) 8251 study (MOPP [mechlorethamine (Mustargen), vincristine, procarbazine, prednisone(Drug information on prednisone)] vs ABVD vs alternating MOPP/ABVD) was dependent on age. The 5-year overall survival was 31% for patients at least 60 years old vs 63% for patients aged 40 to 59 years, and 79% for age < 40 years (P < .0001). The median disease-free survival rates were 8.9 years for ages 16 to 45 years, 3.5 years for 46 to 55 years, 1.5 years for 56 to 65 years, and 0.7 years for > 65 years (P < .0001).
In a recent analysis of US SEER data (1979–2004), Brenner et al reported increasing survival associated with HL, including elderly patients. The 5-year survival rate increased 24% from 1980–1984 to 2000–2004 for HL patients over 60 in this analysis (35% to 59%, respectively). However, the 2000–2004 survival rates for older HL patients were still markedly inferior (> 30% lower) compared with younger populations. Current US SEER data show continued age-disparate outcomes in HL, as illustrated in Figure 2.
The BNLI compared survival of patients participating in randomized clinical trials from 1970 to 1987 with population-based data. Age was the strongest risk factor for survival. Of note, both 5- and 10-year overall survival among BNLI trial participants aged 65 to 74 years were higher than for population-based patients. Among older patients entered on German Hodgkin Study Group (GHSG) protocols, patients ≥ 60 years old also had significantly worse freedom from treatment failure than younger patients (60% vs 80%, P < .001) and inferior overall survival (65% vs 90%) at 5 years. This difference remained significant with the exclusion of events unrelated to HL.
One of the most commonly used prognostic tools in advanced-stage HL is the International Prognostic Score (IPS), which identified seven adverse prognostic factors, including age > 45 years. Of note, only 9% of patients in the study that established the IPS were over 55, and no patients over 65 were included. Enblad et al analyzed registry data according to IPS from 1985 through 1992, consisting of 139 HL patients aged 60 and older. When analyzing patients who were treated with curative intent, 29% had an IPS ≥ 4. These patients had a borderline worse prognosis compared to those with an IPS of 1 to 3.
Clinical characteristics of elderly HL patients are available through registry databases and some prospective clinical trials. Mir et al reported the clinical characteristics of HL patients at least 60 years old who were entered onto four consecutive CALGB randomized trials from 1969 to 1988 (Table 1). In the most recent trial (CALGB 8251), 8% of patients were at least 60 years old. Mixed cellularity histology was more frequent in older patients, whereas mediastinal disease was less common. Levis et al compared a registry cohort of HL patients < 65 years old to patients ≥ 65. These investigators also found frequent mixed cellularity among older patients, while younger patients more commonly presented with bulky mediastinal disease (22% vs 2%, respectively; P < .01).
Engert et al compared the characteristics and outcomes of all HL patients enrolled in GHSG clinical trials by age (> 60 years vs younger). In that series, B symptoms, mixed cellularity, and poorer performance status were more common in older patients. According to IPS, older patients had more unfavorable risk factors vs younger patients, although when age was removed from the IPS calculation, that difference was no longer significant (age-adapted IPS 4–6: 13% vs 11%, respectively). Kim et al reported characteristics of 52 elderly patients treated over a 26-year period, finding that nodular sclerosis histology predominated in early-stage disease, whereas mixed cellularity was more common in advanced disease.
The presence of comorbidity as an independent prognostic factor is particularly relevant for older patients. Levis et al reported results of a trial including 105 elderly HL patients in which all patients received a lower-intensity regimen, VEPEMB (vinblastine, cyclophosphamide, procarbazine, etoposide, mitoxantrone(Drug information on mitoxantrone), bleomycin(Drug information on bleomycin)). Approximately 37% of patients had a comorbidity, defined as the presence of a concomitant disease requiring specific treatment. On multivariate analysis, in addition to stage and presence of B symptoms, presence of comorbidity was independently associated with inferior survival. Of note, the effect on disease-specific survival was as significant as the impact on overall survival, which suggests ineffective treatment and/or biology related to comorbidity rather than death from other causes.
In a population-based study, van Spronsen et al reported on 194 HL patients diagnosed between 1993 and 1996. Among patients < 60 years old, 13% had a serious comorbid condition compared to 56% for older patients (P < .0001). The prevalence of comorbidity among HL patients was similar among early vs advanced stage. The most common comorbid conditions were cardiovascular disease (18%), hypertension (13%), chronic obstructive pulmonary disease (13%), and diabetes (10%). The proportion of elderly patients who received chemotherapy was 50% lower among those with a comorbidity vs no comorbidity (P = .01). Furthermore, elderly patients with early-stage HL and a comorbid condition received chemotherapy much less often (90% vs 33%, P = .05).
Guinee et al compared the outcomes for HL patients aged 60 to 79 years with those for patients aged 40 to 59 years from 1977 to 1983. The older group experienced twice the risk of dying from HL and four times the risk of dying from other causes compared with the younger group. Of note, older patients responded to therapy with similar complete response rates in their analysis (84% vs 88% in the younger group). As proposed by Levis et al and Proctor et al, documentation of comorbid conditions and objective assessment of functional status such as the Comprehensive Geriatric Assessment should be included in prospective studies. It may be possible to modify regimens and/or dosing based on objective criteria that predict for prohibitive morbidity and mortality.