Methods

Randomisation

• A randomised controlled trial with two arms: 4 escalated plus 4 standard courses of BEACOPP versus 8 courses of ABVD

Recruitment period

• 2002 to 2010

Median follow‐up time

• 3.8 years

Participants

Eligibility criteria:

• patients with untreated histologically documented HL

• no prior therapy for HL

• clinical stage: III or IV disease

• at least 1 two‐dimensionally measurable target lesion or extranodal lesion

• international prognostic score of at least 3

• performance status: WHO 0 to 3

• hematopoietic: WBC greater than 2000/mm³, platelet count greater than 100,000/mm³

• hepatic: no prior uncontrolled hepatitis B viral infection, bilirubin no greater than 2.5 times normal (unless due to HL)

• renal: creatinine no greater than 2.0 mg/dL (unless due to HL)

• cardiovascular: no severe cardiac disease that would limit normal life expectancy or preclude study, LVEF at least 50%

• pulmonary: no severe pulmonary disease that would limit normal life expectancy or preclude study, respiratory function at least 30%

• HIV negative

• HTLV1 negative

• no severe active infection

• no severe neurological or metabolic disease that would limit normal life expectancy or preclude study

• no other prior or concurrent malignancy except basal cell skin cancer or carcinoma in situ of the cervix

• no psychological, familial, sociological, or geographical condition that would preclude study

• not pregnant or nursing

• fertile patients must use effective contraception

• no concurrent radiotherapy

• patients between 16 years and 60 years

Patients recruited (not known)

• escalated BEACOPP (N = 274): 10 patients crossed over to ABVD

• ABVD (N = 275): 5 patients crossed over to BEACOPP

Mean age:

• 35.2 years

Gender (male, female):

• males 75%, females 25%

Country

• European Organization for Research and Treatment of Cancer

Interventions

escalated BEACOPP (every 22 days), (no information about dosage)

• Patients receive doxorubicin IV over 5 minutes and cyclophosphamide IV on day 1; etoposide IV over 30 minutes on days 1 to 3; oral procarbazine on days 1 to 7; oral prednisone on days 1 to 14; and vincristine IV and bleomycin IV or intramuscularly (IM) on day 8. Patients may receive dexamethasone in place of prednisone. Patients also receive filgrastim (G‐CSF) subcutaneously (SC) beginning on day 9 and continuing until blood counts recover or pegfilgrastim SC on day 9 only.

ABVD (every 28 days) (no information about doses)

• Patients receive doxorubicin IV over 5 minutes, bleomycin IV or IM, vinblastine IV, and dacarbazine IV over 5 to 10 minutes on days 1 and 15.

Additional therapy

• no irradiation

Outcomes

Outcomes and time points from the registered protocol of the study that are considered in the review:

reported

• EFS

• OS

• PFS

• CR rate

• Time to progression

• Quality of life

• Adverse events (secondary malignancies)

not reported

• Adverse events (treatment‐related mortality, Haematological grade III or IV toxicity, infertility, cardiopulmonary toxicity)

• Quality of life

Notes

None

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

Low risk

Quote: "Between October 2002 and January 2010, 550 patients were randomly assigned. "

Allocation concealment (selection bias)

Low risk

Quote: "Random assignments were stratified by institution and JPS ... "

Blinding (performance bias and detection bias)
Overall survival

Low risk

No information about blinding provided. This is judged not to be a source of bias for OS.

Blinding (performance bias and detection bias)
Other outcomes

High risk

No information about blinding provided, blinding in this setting uncommon.

Incomplete outcome data (attrition bias)
Overall Survival

Low risk

CONSORT flow diagram provided.

Selective reporting (reporting bias)

High risk

Comment: The protocol was registered at http://clinicaltrials.gov (Identifier: NCT00049595). The authors recognised all predefined outcomes except quality of life.

Other bias

Unclear risk

No information provided.

Methods

Randomisation

• A randomised controlled trial with two arms: 2 courses of escalated BEACOPP followed by 2 courses of ABVD plus involved field radiotherapy (IF‐RT) arm versus 4 courses of ABVD plus IF‐RT arm.

Allocation ratios

• 1:1

Recruitment period

• from January 2003 to July 2008

Median follow‐up time

• 43 months

Participants

Eligibility criteria:

• patients with previously untreated HL

• clinical stage: IA, IB, IIA (with one of the following risk factors: large mediastinal mass (a), extranodal disease (b), elevated erythrocyte sedimentation rate (c), or ≥ 3 nodal areas (d)), IIB (with risk factors c and d)

• age: 16 to 60 years

Patients recruited (N = 1655),

• BEACOPP analysed (N = 763), excluded: 15 not sufficiently documented, 42 violation of inclusion criteria

• ABVD analysed (N = 765) excluded: 17 not sufficiently documented, 53 violation of inclusion criteria

Mean age:

• for both arms 32 years (range, 18 to 60 years)

Gender (male, female):

• male 47.7%, female 53.3%

Stage of disease:

• for both arms: stage I 4.8%, stage II 95,2%

Country

• Germany, Switzerland, Netherlands, Czech Republic, Austria

Interventions

BEACOPP (repeated on day 22)

• cyclophosphamide 1,250 mg/m2 (day 1), doxorubicin 35 mg/m2 (day 1), etoposide 200 mg/m2 (days 1 to 3), procarbazine 100 mg/m2 (days 1 to 7), prednisone 40 mg/m2 (days 1 to 14), vincristine 1.4 mg/m2 (day 8; maximum, 2 mg), and bleomycin 10 mg/m2 (day 8). G‐CSF had to be administered from day 8 of each BEACOPPesc cycle until recovery of WBC to at least 1,000/L on 3 consecutive days.

ABVD (repeated on day 29)

• doxorubicin 25 mg/m2 (days 1 and 15), bleomycin 10 mg/m2 (days 1 and 15), vinblastine 6 mg/m2 (days 1 and 15), and dacarbazine 375 mg/m2 (days 1 and 15). G‐CSF was administered if clinically indicated, starting from day 7 or 21

After chemotherapy all patients received IF‐RT

• 30 Gray

Others: Treatment was postponed until recovery of WBC to at least 2 500/L and platelet count to at least 80 000/L on the day scheduled for re‐treatment. In cases of treatment postponement of more than 2 weeks or pronounced toxicity during treatment, dose reductions were foreseen in the trial protocol.

Outcomes

Outcomes and time points from the study that are considered in the review:

reported

• OS (4‐year outcome)

• CR rate

• PFS (4‐year outcome)

• Adverse events (haematological grade III or IV toxicity, secondary malignancies, treatment‐related mortality, infertility)

not reported

• Adverse events (cardiopulmonary toxicity)

• Time to progression

Notes

Supported by Grant No. 106164 from German Cancer Aid (Deutsche Krebshilfe) and the Swiss Federal Government.

Employment or Leadership Position: None Consultant or Advisory
Role: None, Stock Ownership: None, Honoraria: None, Research Funding: None, Expert Testimony: None, Other, Remuneration: JuliaMeissner, Amgen, Celgene

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

Unclear risk

Quote: "patients (...) were randomized"

Comment: The HD 14 trial is organized by the GHSG, that also led the HD 9 trial. Sequence generation and allocation concealment should be judged to probably be adequate, because this is a large multi‐centre trial and other trials of this study group have previously been adequately randomised and allocation of these trials have previously been adequately concealed.

Allocation concealment (selection bias)

Unclear risk

No information provided.

Blinding (performance bias and detection bias)
Overall survival

Low risk

No information about blinding provided. This is judged not to be a source of bias for OS.

Blinding (performance bias and detection bias)
Other outcomes

High risk

No information about blinding provided, blinding in this setting uncommon.

Incomplete outcome data (attrition bias)
Overall Survival

Low risk

Quote: "[..] 1.655 patients [...] were randomized. [...] The full analysis set comprised 1623 patients [...]. Patient characteristics were well balanced between both arms [...]."

Comment: Missing outcome data are balanced in numbers across intervention groups, with similar reasons for missing data across groups.

Selective reporting (reporting bias)

Unclear risk

A protocol of this trial was not registered.

Other bias

Unclear risk

When the third planned interim analysis yielded a significant group sequential test (adjusted overall P.0452), the trial was stopped according to protocol.

Methods

Randomisation

• A randomised controlled trial with three arms:

1. 8 escalated courses of BEACOPP

2. 8 standard courses of BEACOPP

3. 8 courses of cyclophosphamide, vincristine, procarbazine, and prednisone (COPP) alternating with doxorubicin, bleomycin, vinblastine, and dacarbazine (ABVD).

Allocation ratios

• Randomisation to the COPP/ABVD‐arm was stopped after first interim analysis due to the fact that both BEACOPP groups together were significantly superior to COPP/ABVD in terms of the rate of freedom from treatment failure (Diehl 2003). Subsequently, unequal randomization probabilities were used to equalize the numbers in each group.

Recruitment period

• from February 1993 to March 1998.

Median follow‐up time

• 111 months (range 3 to 167 months).

Participants

Eligibility criteria:

• patients with histologically documented, previously untreated HL

• clinical stage: IIB or IIIA (with a large mediastinal mass, i.e. > one third of the maximal thoracic diameter, or extranodal lesions or massive spleen involvement, i.e. with diffuse infiltration or five local lesions), IIIA (with at least three involved nodal areas or elevated erythrocyte sedimentation rate, i.e. more than 50 mm/h in asymptomatic patients and more than 30 mm/h in patients with B symptoms), IIIB, IV

• Karnofsky performance status higher than 70%

• age between 16 and 65 years

• not positive for HIV

• free of concurrent disease

• patients must have provided written informed consent

Patients recruited (N = 1282)

• escalated BEACOPP (N = 496): excluded (N = 28, 27 because of revised histology or met clinical exclusion criteria, 1 declined to give consent), eligible (N = 468), analysed (N = 466, two eligible were excluded because they were lost to follow‐up)

• standard BEACOPP (N = 498): excluded (N = 28, all because of revised histology or met clinical exclusion criteria), eligible (N = 470), analysed (N = 469, one eligible was excluded because they were lost to follow‐up)

• COPP/ABVD (N = 288): excluded (N = 25, all because of revised histology or met clinical exclusion criteria), eligible (N = 263), analysed (N = 261, three eligible were excluded because they were lost to follow‐up). Randomisation to the COPP/ABVD‐arm was stopped after first interim analysis due to the fact that both BEACOPP groups together were significantly superior to COPP/ABVD in terms of the rate of freedom from treatment failure (Diehl 2003).

Mean age:

• escalated: BEACOPP 32 years (15% of 466 patients were 50 to 65 years old); standard: BEACOPP 33 years (18% of 469 patients were 50 to 65 years old); COPP/ABVD: 32 years (16% of 260 patients were 50 to 65 years old)

Gender (male, female):

• escalated BEACOPP: 289, 177; standard BEACOPP: 295, 174; COPP/ABVD: 149, 112

Stage of disease:

• Stage IIB: escalated BEACOPP: 16%; standard BEACOPP: 14%; COPP/ABVD: 9%

• Stage III: escalated BEACOPP: 52%; standard BEACOPP: 52%; COPP/ABVD: 60%

• Stage IV: escalated BEACOPP: 33%; standard BEACOPP: 35%; COPP/ABVD: 32%

Bulky disease: escalated BEACOPP: 67%; standard BEACOPP: 68%; COPP/ABVD: 58%

Country

• Germany, Switzerland, Austria, Czech Republic

Interventions

escalated BEACOPP (every 22 days)

• bleomycin (10 mg/m², day 8), etoposide (200 mg/m², day 1 ‐ 3), doxorubicin (35 mg/m², day 1), cyclophosphamide (1200 mg/m², day 1), vincristine (1.4 mg/m² (2 mg/m² maximum), day 8), procarbazine (100 mg/m², day 1 ‐ 7), prednisone (40 mg/m², day 1 ‐ 14)

standard BEACOPP (every 22 days)

• bleomycin (10 mg/m², day 8), etoposide (100 mg/m², day 1 ‐ 3), doxorubicin ([25 mg/m², day 1), cyclophosphamide (650 mg/m², day 1]), vincristine (1.4 mg/m² (2 mg/m² maximum), day 8], procarbazine (100 mg/m², day 1 ‐ 7), prednisone (40 mg/m², day 1 ‐ 14)

COPP/ABVD (every 57 days)

• bleomycin (10 mg/m², day 29 and 43), doxorubicin (25 mg/m², day 29 and 43), cyclophosphamide (650 mg/m², day 1 and 8), vincristine (1.4 mg/m² (2 mg/m² maximum), day 1 and 8), procarbazine (100 mg/m², day 1 ‐ 14), prednisone (40 mg/m², day 1 ‐ 14), vinblastine (6 mg/m², day 29 and 43), dacarbazine (375 mg/m², day 29 and 43)

Additional therapy

At the end of chemotherapy, radiotherapy was scheduled for sites of initial bulky disease (those at least 5 cm in diameter) and any residual tumour.

• 30 Gray for initial bulky disease

• 40 Gray for any residual tumour

Outcomes

Outcomes and time points from the study that are considered in the review:

reported

• OS (10‐year outcome)

• FFTF (analysed as PFS in this review)

• CR

• Adverse events (treatment‐related mortality, Haematological grade III or IV toxicity, secondary malignancies, infertility)

not reported

• Time to progression

• Adverse events (cardiopulmonary toxicity)

Notes

At the first interim analysis in September 1996, the early stopping boundary was crossed, with the demonstration that both BEACOPP groups together were significantly superior to COPP/ABVD in terms of the rate of FFTF. Therefore, assignment to the COPP/ABVD group was stopped (Diehl 2003a).

This study was supported by a grand from: Deutsche Krebshilfe, Swiss Group for Clinical Cancer Research (Diehl 2003a)

The author(s) indicated no potential conflicts of interest (Engert 2009).

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

Low risk

Quote: "random assignment was performed by computer"

Allocation concealment (selection bias)

Low risk

Quote: "after clinical staging patients were randomly assigned by telephone call to the trial coordination center to one of three chemotherapy regimens"

Blinding (performance bias and detection bias)
Overall survival

Low risk

No information about blinding provided. This is judged not to be a source of bias for OS.

Blinding (performance bias and detection bias)
Other outcomes

High risk

No information about blinding provided, blinding in this setting uncommon.

Incomplete outcome data (attrition bias)
Overall Survival

Low risk

Quote: "Of 1282 patients, 86 were excluded due to incorrect
diagnosis, concurrent disease, incorrect staging, other clinical exclusion criteria and insufficient data to determine the treatment outcome. Thus, a total of 1196 patients were fully assessable"

Comment: Missing outcome data are balanced in numbers across intervention groups, with similar reasons for missing data across groups. Therefore, see Diehl 2003a who has given a detailed overview of the reasons for the exclusion of patients: escalated BEACOPP (15 did not have HD, 12 met clinical exclusion criteria, 1 lost to follow‐up), standard BEACOPP (19 did not have HD, 9 met clinical exclusion criteria, 1 lost to follow‐up), COPP/ABVD (12 did not have HD, 13 met clinical exclusion criteria, 3 lost to follow‐up).

Selective reporting (reporting bias)

Unclear risk

A protocol of this trial was not registered.

Other bias

Unclear risk

Quote: "Assignment to the COPP/ABVD group was stopped by the safety board after the first interim analysis indicating that both BEACOPP groups were significantly superior in terms of FFTF to COPP/ABVD (P .03)."

Comment: In the GHSG HD9 trial randomisation for the escalated BEACOPP‐arm started later than randomisation to both other arms. Furthermore, randomisation to the COPP/ABVD‐arm was stopped after first interim analysis due to the fact that both BEACOPP groups together were significantly superior to COPP/ABVD in terms of the rate of freedom from treatment failure (Diehl 2003). According to this potential risk of bias the analyses of OS and PFS were based on data of patients recruited in parallel, i.e. during the same time period (COPP/ABVD N = 177, BEACOPP N = 265).

Methods

Randomisation

• A randomised controlled trial with two arms: 4 escalated plus 4 standard courses of BEACOPP versus 6‐8 courses of ABVD

Allocation ratios

• 1:1

Recruitment period

• from March 2000 to March 2007.

Median follow‐up time

• 61 months (range not reported).

Participants

Eligibility criteria

• patients with histologically documented HL

• clinical stage: IIB, III, IV or ≥ 3 international‐prognostic‐score HL

• no previous anticancer treatment

• adequate bone marrow, renal and hepatic function

• absence of severe cardiac disease and LVEF ≥ 50%

• absence of lung diseases

• absence of previous invasive cancer

• absence of HIV

• adult patients up to 60 years of age

Patients recruited (331 patients)

• BEACOPP (N = 163 were analysed)

• ABVD (N = 168 were analysed)

Mean age:

• not reported. (escalated BEACOPP 20% of 155 patients ≥ 45 years; ABVD 23% of 166 patients ≥ 45 years)

Gender (male, female):

• escalated BEACOPP: 88, 66; ABVD: 100, 66

Stage of disease:

• not reported.

Country

• Italy

Interventions

escalated BEACOPP (every 21 days)

• bleomycin [10 mg/m², day 8], etoposide [200 mg/m², day 1‐3], doxorubicin [35 mg/m², day 1], cyclophosphamide [1250 mg/m², day 1], vincristine [1.4 mg/m² (2 mg/m² maximum), day 8], procarbazine [100 mg/m², day 1‐7], prednisone [40 mg/m², day 1‐14], G‐CSF [no information about dosage and period of application]

standard BEACOPP (every 21 days)

• bleomycin [10 mg/m², day 8], etoposide [100 mg/m², day 1‐3], doxorubicin [25 mg/m², day 1], cyclophosphamide [650 mg/m², day 1], vincristine [1.4 mg/m² (2 mg/m² maximum), day 8], procarbazine [100 mg/m², day 1‐7], prednisone [40 mg/m², day 1‐14]

ABVD (every 28 days)

• doxorubicin [25 mg/m², day 1 and 15], bleomycin [10 mg/m², day 1 and 15], vinblastine [6 mg/m², day 1 and 15], dacarbazine [375 mg/m², day 1 and 15]

Additional therapy

At the end of chemotherapy, radiotherapy was scheduled for sites of initial bulky disease and any residual tumour

• 30 Gray

For treatment failures

• re‐induction chemotherapy followed by stem‐cell supported BEAM plus up to 25 Gray radiotherapy

Outcomes

Outcomes and time points from the study that are considered in the review:

reported

• OS (3‐year outcome)

• EFS (analysed as PFS in this review)

• CR rate

• Adverse events (Haematological grade III or IV toxicity, secondary malignancies, cardiopulmonary toxicity)

not reported

• Time to progression

• Adverse events (infertility)

Notes

Funding resources: The study was designed by the investigators, and the coordination of the study, the collection and quality control of the data, and the statistical analyses were performed at the operations office of Fondazione Michelangelo (the sponsoring organization). No commercial support was provided.

The author(s) indicated no potential conflicts of interest.

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

Unclear risk

Quote: "patients (...) were randomized"

Comment: No further information provided

Allocation concealment (selection bias)

Unclear risk

No information provided.

Blinding (performance bias and detection bias)
Overall survival

Low risk

No information about blinding provided. This is judged not to be a source of bias for OS.

Blinding (performance bias and detection bias)
Other outcomes

High risk

No information about blinding provided, blinding in this setting uncommon.

Incomplete outcome data (attrition bias)
Overall Survival

Low risk

Quote: "a total of 331 patients were recruited (...); 168 patients were assigned to receive ABVD and 163 to receive BEACOPP. Two patients were subsequently found to be ineligible because of a wrong diagnosis, and 7 withdrew consent before starting chemotherapy. Thus, 166 patients were initiated on ABVD chemotherapy and 156 on BEACOPP chemotherapy (safety population), whereas all 331 patients who underwent randomization were assessed for efficacy (intention‐to‐treat population)."

Selective reporting (reporting bias)

Low risk

All outcomes reported as pre‐planned

Other bias

Unclear risk

No information provided.

Methods

Randomisation

• A randomised controlled trial with three arms: 4 escalated plus 2 standard courses of BEACOPP versus 6 courses of CEC (cyclophosphamide, lomustine, vindesine, melphalan, prednisone, epidoxorubicin, vincristine, procarbazine, vinblastine, and bleomycin) versus 6 courses of ABVD arm.

Allocation ratios

• 1:1:1

Recruitment period

• from April 2000 to June 2007.

Median follow‐up time

• 41 months (range 4 to 91 months).

Participants

Eligibility criteria.

• patients with histologically documented, previously untreated HL,

• clinical stage: IIB, III, IV

• Eastern Cooperative Oncology Group (ECOG) performance status of 0 to 3

• age >16 years

• patients must have provided written informed consent

Patients recruited (N = 307)

• BEACOPP (N = 102): excluded (N = 2, one because of revised histology, one because of not compliant), eligible (N = 100), analysed (N = 98, two eligible were excluded because of missing data)

• CEC (N = 102): excluded (N = 0), eligible (N = 102), analysed (N = 98, three eligible were excluded because of missing data, one because of lost of follow‐up)

• ABVD (N = 103): excluded (N = 0), eligible (N = 103), analysed (N = 99, three eligible were excluded because of missing data, one because of lost of follow‐up)

Mean age:

• escalated BEACOPP: 29 years (17% of 98 patients ≥ 45 years); CEC: 33 years (26% of 98 patients ≥ 45 years); ABVD: 32 years (18% of 99 patients ≥ 45 years)

Gender (male, female):

• escalated BEACOPP: 60, 38; CEC: 56, 42; ABVD: 43, 56

Stage of disease:

• Stage IIB: escalated BEACOPP: 31%; CEC: 29%; ABVD: 33%

• Stage III: escalated BEACOPP: 46%; CEC: 45%; ABVD: 44%

• Stage IV: escalated BEACOPP: 22%; CEC: 26%; ABVD: 22%

Bulky disease: escalated BEACOPP: 37%; CEC: 37%; ABVD: 31%

Country

• Italy

Interventions

BEACOPP (every 21 days)

• bleomycin [10 mg/m² intravenously, day 8)], etoposide [200 (escalated) ‐100 (standard) mg/m² intravenously, day 1‐3], doxorubicin [35 (escalated)‐25 (standard) mg/m² intravenously, day 1], cyclophosphamide [1250 (escalated)‐650 (standard) mg/m² intravenously, day 1], vincristine [1.4 mg/m² (2 mg/m² maximum) intravenously, day 8], procarbazine [100 mg/m² orally, day 1‐7], prednisone [40 mg/m² orally, day 1‐14], G‐CSF [300 µg total subcutaneously, day 8 until neutrophils 0.5/µL]

CEC (every 28 days)

• cyclophosphamide [650 mg/m² intravenously, day 1 (cycles 1, 3 and 5 only)], lomustine [100 mg/m² orally, day 1 (cycles 2, 4 and 6 only)], vindesine [3 mg/m² intravenously, day 1], melphalan [6 mg/m² orally, day 1‐3], prednisone [40 mg/m² orally, day 1‐14], epidoxorubicin [40 mg/m² intravenously, day 8], vincristine [1.4 mg/m² (2 mg/m² maximum) intravenously, day 8], procarbazine [100 mg/m² orally, day 8‐14], vinblastine [6 mg/m² intravenously, day 15], bleomycin [10 mg/m² intravenously, day 15]

ABVD (every 28 days)

• doxorubicin [25 mg/m² intravenously, day 1 and 15], bleomycin [10 mg/m² intravenously, day 1 and 15], vinblastine [6 mg/m² intravenously, day 1 and 15], dacarbazine [375 mg/m² intravenously, day 1 and 15]

Additional therapy

At the end of chemotherapy, radiotherapy was scheduled for sites of previous bulky disease or on slowly or partially responding sites.

• 30 to 36 Gray with a boost of 6 additional Gray to persisting disease sites

Outcomes

Outcomes and time points from the study that are considered in the review:

reported

• OS (5‐year outcome)

• PFS (5‐year outcome)

• CR rate

• Adverse events (Haematological grade III or IV toxicity)

not reported

• Time to progression

• Adverse events (treatment‐related mortality, cardiopulmonary toxicity, secondary malignancies, infertility)

Notes

This study was in parts funded by: Associazione Angela Serra per la Ricerca sul Cancro, Modena, and the Gruppo Amici Dell’Ematologia, Reggio Emilia, Italy.

The author(s) indicated no potential conflicts of interest.

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

Unclear risk

Quote: "patients were randomly assigned"

Comment: No further information provided

Allocation concealment (selection bias)

Unclear risk

No information provided

Blinding (performance bias and detection bias)
Overall survival

Low risk

No information about blinding provided. This is judged not to be a source of bias for OS.

Blinding (performance bias and detection bias)
Other outcomes

High risk

No information about blinding provided, blinding in this setting uncommon.

Incomplete outcome data (attrition bias)
Overall Survival

Low risk

Missing outcome data balanced in numbers across intervention groups, with similar reasons for missing data across groups (i.e. 4 patients of the BEACOPP arm were excluded from the analyses because of revised histology (N = 1), not compliant (N = 1), missing data (N = 2); 4 patients of the ABVD arm were excluded because of missing data (N = 3) and lost to follow‐up (N = 1))

Selective reporting (reporting bias)

Unclear risk

Although the protocol is published, no pre‐planned outcomes are reported, therefore we judge the risk of bias as unclear.

Other bias

Unclear risk

Quote: "The study was initially designed to compare myelotoxicity referred of leukopenia of CEC and BEACOPP regimens with respect of that of ABVD (...). Once, in the first 18 months, Gruppo Italiano per lo Studio dei Linfomi centers became familiar with BEACOPP, the protocol was amended and BEACOPP and CEC regimens were tested against ABVD primarily in terms of FFS."

Quote: "We used one‐sided test in consideration of the expected superiority of BEACOPP and CEC, in term of FFS in comparison with the historical knowledge about ABVD chemotherapy regimen."

Comment: This procedure might have led to a underpowered study.