Meta-analysis of cardiovascular adverse events in patients receiving frontline treatment for non-Hodgkin lymphoma

The risk of cardiovascular (CV) adverse events (AEs; CVAEs) and heart failure is elevated in patients with non-Hodgkin lymphoma (NHL), and survivors of NHL, when compared with the general population and other patient groups. This marked increase is believed to be associated with the frontline treatments currently employed in the NHL setting. Components of the cyclophosphamide, doxorubicin, vincristine, and prednisone (CHOP) and rituximab plus CHOP (R-CHOP) regimens are thought to be particularly cardiotoxic. Specifically, cyclophosphamide and doxorubicin have been linked to left ventricle dysfunction.

Marijke Linschoten, Janine A. M. Kamphuis, and colleagues at the University of Utrecht, NE, carried out a meta-analysis to determine the frequency of CVAEs in patients with NHL that had received CHOP or R-CHOP as a frontline therapy and to identify potential risk factors. Below is a summary of the findings, recently published in Lancet Hematology.¹

Study design

• The study systematically gathered patient data from PubMed, EMBASE, and the Cochrane Library
• The analysis included adult patients involved in clinical trials and observational studies receiving frontline CHOP and/or R-CHOP for
  • diffuse large B-cell lymphoma (DLBCL)
  • follicular lymphoma (FL)
  • mantle cell lymphoma (MCL)
  • NHL not otherwise specified
• Primary outcomes: frequency of Grade 3–4 CVAEs and heart failure
• Sub-group analyses were carried out to identify possible risk factors such as
  • the number of CHOP or R-CHOP cycles
  • cycle interval
  • age
  • sex

Results

• Between April 1984 and June 2019, 137 studies involving patients receiving treatment for NHL (n = 21,211) met the inclusion criteria, and the study characteristics are given in Table 1

 Table 1. Characteristics of the 137 studies eligible for analysis¹

CHOP, cyclophosphamide + doxorubicin + vincristine + and prednisone; CV, cardiovascular; CVAE, cardiovascular adverse event; IQR, interquartile range; R-CHOP, rituximab + CHOP; SD, standard deviation
Study characteristics (N = 137 studies)
Number of patients receiving frontline treatment CHOP R-CHOP CHOP + R-CHOP	21,211  9,541 11,293 377
Median follow-up, months (IQR)	39.0 months (25.5–52.8)
Mean cycle interval, days (SD)	19.9 (3.4)
Mean number of cycles (SD)	6.3 (1.3)
Number of studies reporting Grade 1–2 CVAEs Grade 3–4 CVAEs Overall heart failure Treatment-related CV death Treatment discontinuation due to heart failure	23 77 38 118 51

  Sub-group analysis

• The subgroup analysis identified several factors that significantly influenced the proportion of Grade 3–4 CVAEs and overall heart failure in patients receiving frontline CHOP and/or R-CHOP for the treatment of NHL (Table 2)
• Ninety-six subgroups were derived from the 77 studies (n = 14,351 patients) reporting on Grade 3–4 CVAEs
• The proportion of Grade 3–4 CVAEs ranged from 0.0–15.1%
  • Pooled proportion: 2.35% (95% CI, 1.81–2.93)
• Thirty-eight studies, covering 47 subgroups of treatment and 5,936 patients, reported on the occurrence of overall heart failure
  • Pooled proportion of overall heart failure: 4.62% (95% CI, 2.25–7.65)
  • When cardiac function was evaluated following chemotherapy, there was a significant increase in the proportion of reported heart failure cases: 1.64% (95% CI, 0.82–2.65) to 11.72% (3.00–24.53); p = 0·017
• Discontinuation of CHOP or R-CHOP following treatment-related heart failure in patients was rare (< 0.001%)

 Table 2. Factors influencing the development of Grade 3–4 CVAEs and overall heart failure¹

CHOP, cyclophosphamide + doxorubicin + vincristine + prednisolone; R-CHOP, rituximab + CHOP *Meta-regression †Tested with a fixed-effect model across subgroups ‡Analysis was done with data of % of females per study group, rather than comparing the proportion of adverse events or occurrence of heart failure between men and women
	Number of subgroups of treatment	Regression coefficient treatment (95% CI)	Relative risk (95% CI)	p
Grade 3–4 CVAEs
Age, years < 65 (young) ≥ 65 (old) Old vs young	91 58 33 —	0.39 (0.21–0.57) — — —	—  —  — 3.18 (2.54–3.98)	< 0.0001* < 0.0001† — —
Sex, % female‡	92	0.24 (0.09–0.40)	—	0.0023*
Overall heart failure
Screening No Yes	44 30 14			— 0.017† —
Number of cycles ≤ 6 (low) > 6 (high) High vs low	46 31 15 —	4.99 (0.66–9.32) — — —	— — — 1.78 (1.43–2.23)	0.024* 0.31† — —

 Conclusion

Although the proportion of patients that developed Grade 3–4 CVAEs was relatively low, this multi-study analysis underpins frontline treatment with CHOP and/or R-CHOP as a contributing factor towards heart failure. Older (≥ 65 years) age and female sex significantly influenced the establishment of Grade 3–4 CVAEs, while the number of treatment cycles had a significant impact on the development of overall heart failure. Additionally, when active cardiac function screening took place following treatment with CHOP/R-CHOP, an increased number of patients with overall heart failure were identified. These data may encourage clinicians to monitor cardiac events in patients receiving frontline chemotherapy for NHL, as treating pre-symptomatic individuals could halt the progression to serious cardiac AEs. Additionally, as the study uncovered the negative impact older age has on the risk of developing CVAEs, and given the increased prevalence of NHL with age, the use of alternative less cardiotoxic frontline treatments could be beneficial. The authors noted a number of limitations associated with the study, including the lack of access to individual patient data, inter-study heterogeneity, exclusion of potentially applicable studies, and a lack of uniform CVAE toxicity grading across the analysis.