As seen in the references above, the majority of data come from patients with more frequent aggressive lymphomas—i.e., diffuse large B-cell lymphoma (DLBCL), commonly treated with the CHOP-R regimen
[27]. Currently, people are witnessing a positive trend in successful treatment of older patients with aggressive lymphoma, as shown by a meta-analysis conducted by Tavares et al.
[28]. In this particular analysis, however, only 3 out of 38 studies involving very elderly patients (age ≥ 80) with DLBCL focused on cGA. However, the trend is changing, and more studies have focused on this approach in DLBCL, since it is primarily lymphoma of the elderly. Eyre et al. retrospectively examined infection-related mortality in 690 elderly patients, with curative intent being 7.2% at 1 year. In a multivariate analysis, alongside high International Prognostic Index values and albumin levels, a CIRS-G score≥ 6 was the strongest variable to predict this outcome
[7]. The G8 scale was associated with poor outcomes in 388 elderly patients (HR for OS = 0.871)
[13]. Concerning the loss of IADL, in a retrospective study on 142 patients (age ≥ 80) treated with miniCHOP-R, this factor was the most predictive for death event (HR = 12.39)
[18]. These and other studies are “proof of concept” that cGA plays an important role in the treatment and decision making in elderly patients with malignant lymphoma, and researchers believe that, given the paucity of data in HL, these results from appropriately designed up-to-date cGA studies in DLBCL may be translated to HL. The earliest registry study of historical importance carried out in the Netherlands found that comorbidities were associated with the delivery of treatment to elderly patients with HL
[29]. In the modern era, a retrospective multicenter study on 88 elderly patients, the majority of whom were treated with the ABVD regimen, found two risk factors for adverse outcome in multivariate analysis: age ≥ 70 (HR for OS = 2.24) and loss of ADL (HR for OS = 2.71)
[30]. The largest study conducted thus far was based on SEER data on 1315 patients receiving the full ABVD regimen and analyzed the risk factors for 1-year mortality
[2][31]. Increases in Charlson Comorbidity Index values
[32] (OR = 1.41) and age (OR = 1.33) were predictive of a worse outcome. Interestingly, disability status was not found to be a significant predictor, possibly due to patient selection bias (i.e., only fit patients offered ABVD regimen). Orellana-Noia et al. carried out research across 10 medical centers in the US (
n = 244) and showed that superior outcomes were obtained after treatment with the conventional regimens compared to alternative therapies
[11]. In a multivariate analysis, only loss of ADL was found to be predictive of poorer OS (HR = 2.13). Treatment was feasible, with treatment-related mortality (TRM) being 3.3%. However, toxicities did impact the total number of delivered cycles and reduced CR rates accordingly, resulting in lower PFS for patients experiencing them. The study suffers from several limitations. Due to inherent selection bias and retrospective analysis, it does not reflect “real-world” patients. Furthermore, only one patient underwent cGA before treatment initiation, with other functionality and geriatric assessments conducted in a retrospective manner. The latter matter should be assessed prospectively to draw firmer conclusions in HL patients; results from an observational study examining the relationship between cGA, type of treatment used and the long-term outcomes currently underway are eagerly awaited
[33]. The cGA of an elderly HL patient is a valuable first step in treatment decision making, recommended by experts in this field
[34][35][36]. Such assessment is shown to be more effective than clinical judgement in stratifying lymphoma patients according to fitness status and should be used whenever possible. However, given its laborious nature, it is often not feasible in every day clinical practice. On the other hand, indexes such as ADL
[14][15][16], IADL
[17] or G8
[12] (for other examples see
Table 1) are widely available and manageable within couple of minutes.
3. Does ABVD Regimen Treatment Have a Role in Elderly Patients with HL?
3.1. Early and Intermediate Disease
GHSG HD 10 and HD 11 phase III randomized clinical trials (RCTs) have established the historical standard of care for younger patients
[37][38]. A sub analysis of HD10 and HD11 included 177 elderly patients, with the majority of patients receiving four cycles of ABVD (93%)
[39]. The complete remission (CR) rate was high (89%), and 5-year OS and PFS rates were 81.2% and 74.8%, respectively. An additional analysis of HL-specific outcomes was undertaken, showing higher rates of OS and PFS and suggesting the existence of various competing factors influencing survival. One of the possible factors was the high rate of grade III and IV adverse events (AEs), which was 68%. Infections and respiratory disorders were more prevalent in this population in comparison to younger patients, and treatment-related mortality was 5%. Moreover, toxicities probably influenced relative dose intensity (RDI) of chemotherapy (with 41% of patients having an RDI < 80%), with frequent delays in therapy seen. Concerning other treatment modalities, IFRT was found to be feasible, with mild toxicity. Patients who received two cycles of ABVD experienced fewer AEs with a greater RDI; however, the final outcomes were similar (for details, see the appendix of the original study)
[39]. On the other hand, feasibility of the BEACOPP baseline regimen was low in terms of the high rate of AEs, more protocol deviations and TRM favoring ABVD (see the appendix of the original study)
[39]. The HD13 trial examined the role of omission of dacarbazine and/or bleomycin for early-stage HL and found that it compromises efficacy and should not be generally recommended
[40]. Subsequent analysis from the same group included 287 older patients from HD10 and HD13 trials and stratified them by regimen received: ABVD (two or four cycles) and AVD (two cycles)
[41]. The CR rate was high and similarly independent of the regimen used. The PFS and OS rates in elderly patients, however, seem to be inferior to those reported in younger patients. Considering only the results extrapolated from the HD13 trial (two cycles of ABVD vs. two cycles of AVD), there seems to be no difference in final outcomes (HR for PFS = 1.28; HR for OS = 1.3); however, researchers must note that the 95% CI are wide, leading to the low level of evidence, and as the authors stated, the analysis was underpowered to draw a noninferiority conclusion. Concerning feasibility, the RDI was like that of younger patients, with the early termination rate being low. The rates of grade III or IV AEs were similar whether receiving ABVD or AVD (40% vs. 39%, respectively), with negligible TRM. The authors concluded that the application of a maximum of two cycles of ABVD is feasible in older patients; however, they stated that the omission of bleomycin is mandatory when more than two cycles are used.
3.2. Advanced Disease
Concerning ABVD in advanced disease, randomized data are available from the ECOG E2496 phase III RCT comparing ABVD vs. Stanford V regimen
[42]. An additional analysis was undertaken for elderly patients
[43]. Since this trial established ABVD as the regimen of choice at the time of publication, researchers focus on the ABVD arm (
n = 23). The ORR was 74%. Concerning outcomes, the 5-year failure-free survival (FFS) rate was 53%, and the 5-year OS rate was 64%, showing a statistically significant difference (
p = 0.002 and
p < 0.001, respectively) when compared to the rates in younger patients. The grade III or IV AE rate was 92%. AEs of clinical interest were febrile neutropenia and respiratory disorders (as discussed in the section “Should We Be Afraid of Bleomycin Use in Elderly Patients with HL?”), with TRM being 8%. RDI was 73% in this population, showing a poor feasibility. The authors concluded that novel approaches are needed to improve tolerability and efficacy in advanced disease setting. In advanced disease, the role of ABVD is controversial. However, if practicing physicians decide to use ABVD in advanced stage disease, researchers' recommendation is to implement an interim positron emission tomography (iPET) adapted strategy, such as per RATHL trial design
[44]. Patients with positive iPET results should not continue with the same treatment protocol, given the anticipated lack of efficacy and unacceptable toxicity, but could rather be offered enrollment in clinical trials, alternative treatment or receive best supportive care. Although there has been no formal analysis in this particular study concerning elderly patients, the retrospective study from Bentur et al. represents a “proof-of-concept” study on iPET
[45]. Concerning treatment, most subjects were treated using the ABVD protocol (67%), with most patients achieving complete metabolic remission (CMR) as per Lugano criteria
[46]. The iPET positivity was associated with poor outcomes in terms of PFS and OS. In a regression analysis, the only predictor for poor outcomes was positive iPET (HR for progression = 8.5, HR for death = 6.9). The study suggests that iPET is useful in guidance of the treatment goal, i.e., to continue with curative intent or palliative care. This notion is further confirmed by the sub analysis of the GATLA LH-05 trial
[47][48]. Treatment with three cycles of ABVD and iPET CMR had excellent outcomes, with median PFS and OS not reached. More importantly, no grade III and IV AEs were noted, representing a possible treatment option for a subset of elderly patients.
The data obtained on ABVD regimen are presented in Table 2: “ABVD Regimen in Elderly Hodgkin Lymphoma.”
Table 2. ABVD Regimen in Elderly Hodgkin Lymphoma.
Study |
n |
Disease Stage |
Regimen |
CR |
Outcome Rate * |
p ** |
AE Rate N (Rate) *** |
CVD-Related Death N (Rate) |
Respiratory AE N (Rate) *** |
Infection AE N (Rate) *** |
TRM Due to AEs |
Boll et al. [39] |
117 |
early |
ABVD 4x |
89% |
5-year PFS = 74.8% 5-year OS = 81.2% |
p < 0.001 |
79 (68%) |
8 (7%) |
6 (5%) |
11(10%) |
5% |
Boll et al. [41] **** |
287 |
early |
AVD 2x (HD13, N = 82) |
98% |
5-year PFS = 79% 5-year OS = 91% |
p = NR |
31(40%) |
NR |
NR |
2(3%) |
1% |
ABVD 2x (HD13, N = 67) |
99% |
5-year PFS = 78% 5-year OS = 86% |
26 (42%) |
3 (4%) |
1 (2%) |
4 (6%) |
NR |
ABVD 2x (HD10, N = 70) |
96% |
5-year PFS = 79% 5-year OS = 84% |
24 (37%) |
NR |
1(2%) |
5 (8%) |
3% |
ABVD 4x (HD10, N = 68) |
88% |
5-year PFS = 79% 5-year OS = 87% |
45(65%) |
5 (7%) |
7 (10%) |
(8) 12% |
6% |
Evens et al. [43] ***** |
24 |
advanced |
ABVD x6 |
65% |
5-year FFS = 53% 5-year OS = 64% |
p = 0.002 p < 0.0001 |
22 (92%) |
NR |
Grade 3 = 5 (20%) Grade 4 = 7 (27%) |
Grade 3 = 10 (4%) Grade 4 = 10 (4%) |
9% |
Published trials suffer from multiple limitations. The first and the most important one is selection bias, meaning that only a low number of patients were included, not representing the “real world” population of elderly HL. The main reason for this is rigorous inclusion criteria, mainly fitness and adequate organ function. Yet, as shown in Table 1, toxicities are common even in fit subjects, sending an important safety signal. Furthermore, it is evident that extrapolation of the treatment of younger patients is not possible to the elderly, primarily due to AEs occurring in this setting.
4. Should We Be Afraid of Bleomycin Use in Elderly Patients with HL?
Bleomycin-induced lung injury (BILI) is a syndrome that presents with a variety of clinical features (ranging from acute respiratory distress syndrome to pneumonitis or pulmonary edema), with diagnosis being based on the exclusion of other possible respiratory disorders
[49]. Based on registry data obtained from 835 patients with germ cell tumors, the contributing risk factors for bleomycin toxicity are age greater than 40 years (HR = 2.3) and impaired glomerular filtration rate (HR = 3.3)
[50]. In researchers' opinion, these findings readily translate to HL patients.
The actual incidence of BILI in elderly patients with HL is not known, with variable rates reported due to the use of retrospective study designs and reporting biases
[11][30][43][51][52]. Concerning prospective trials, only Evens et al. used exact methods for testing pulmonary lung function
[43]. In total, 10 patients treated with ABVD developed BILI, with AE commonly being grade I or II, with two death events. The experience from the Mayo Clinic shows the rates of BILI and BILI-related mortality being 18% and 24%, respectively
[51]. Patients with BILI had a median 5-year OS rate of 63% vs. 90% in unaffected patients (
p = 0.001). However, there was no significant HL-related OS difference among patients surviving bleomycin toxicity compared to ones not developing it, suggesting that the survivors of BILI have the same long-term outcome. Due to the small patient sample, multivariate analysis was not performed, but univariate analysis suggests age greater than 40, the use of G-CSF factors and the application of the ABVD regimen as possible risk factors. French data on elderly patients (
n = 117, with the majority having advanced disease and the mean N of ABVD cycles = 6) reported the rate of occurrence of pulmonary toxicity of grade III or IV
[52]. Early pulmonary AEs occurred in 31 patients, with 7 death events, while late pulmonary AEs occurred in 22 patients. In total, the mortality rate related to pulmonary events was 22%. Despite some of these reports suggesting no influence on HL-related outcomes for survivors of acute toxic events, one must keep in mind that long-term consequences can severely impair the survivor’s quality of life.
In conclusion, the data on BILI in older patients are strong enough to question the use of bleomycin in the treatment algorithm. If used, it should be given cautiously and by an experienced clinician, able to recognize and treat BILI promptly if it occurs.
5. ABVD and Leading Guidelines
Recently published British guidelines recommend that GA should be undertaken prior to deciding on treatment, using GRADE criteria
[36][53]. Concerning fit patients, they suggest the use of the AVD regimen as a treatment of choice, especially in early stage (1B). Furthermore, a maximum of three cycles of ABVD may be used as part of the iPET adapted strategy for advanced disease (2B). It is important to note that cardiac comorbidity must be assessed before initiating a doxorubicin-containing regimen due to the high rate of cardiovascular mortality
[34][35][36]. The option recommended by the National Comprehensive Cancer Network (NCCN) guidelines for early, favorable disease is two cycles of A(B)VD followed by irradiation
[54]. For advanced disease, the advised course is two cycles of ABVD followed by four cycles of AVD in the case of iPET negativity, as shown by the RATHL study
[44]. Furthermore, when using the ABVD regimen, the use of G-CSF is not recommended due to higher incidence of BILI. Concerning the European Society for Medical Oncology guidelines on HL, the authors state that ABVD-based chemotherapy is the therapy of choice, with omission of bleomycin after two cycles (grade III, B–C)
[55]. The main emphasis of these guidelines’ recommendation is limiting toxicity in first-line treatment.
The first part of the approach in treating these patients with ABVD is multidisciplinary cGA to identify fit patients who may benefit from ABVD. Concerning early or intermediate disease, researchers recommend two to four cycles of A(B)VD followed by IFRT. Omission of bleomycin remains at the treating physician’s discretion; however, its use should not exceed two cycles of treatment
[41]. In advanced disease, six cycles of ABVD are unfeasible and should not be used in this setting. If a treating physician chooses ABVD, it should follow PET-adapted strategy such as the RATHL approach with modification that iPET positive patients should be referred to clinical trial, if possible, or palliative care
[44].