Comparison Of Treatment Outcomes Between Epoch And Chop Regimens In Patients With Hiv Associated NHL In A Low Resource Setting

Purpose: The optimal chemotherapy regimen for treating HIV associated NHL in low resource settings is unknown. We conducted a retrospective study to compare survival rates, treatment response rates and adverse events in patients with HIV associated NHL treated with EPOCH and CHOP regimens at the Uganda Cancer Institute. Methods: A retrospective study of patients diagnosed with HIV and lymphoma and treated at the Uganda Cancer Institute from 2016 – 2018 was done. Results: One hundred eight patients treated with CHOP and 12 patients treated with EPOCH were analysed. Patients completing 6 or more cycles of chemotherapy were 51 (47%) in the CHOP group and 8 (67%) in the EPOCH group (p=0.20). One year overall survival (OS) rate in patients treated with CHOP was 54.5% (95% CI, 42.8 – 64.8) and 80.2% (95% CI, 40.3 – 94.8) in those treated with EPOCH. The survival rates were not statistically different in the two groups; hazard ratio, 0.43 (95% CI, 0.10 -1.78; p=0.24). Factor associated with favourable survival were BMI 18.5-24.9 kg/m 2 , (p=0.03) and completion of 6 or more cycles of chemotherapy, (p<0.001). The overall response rate was 40% in the CHOP group and 59% in the EPOCH group (P = 0.66). Severe adverse events occurred in 19 (18%) patients in the CHOP group and 3 (25%) in the EPOCH group (p=0.53). Severe adverse events were neutropenia (CHOP=13, 12%; EPOCH=2, 17%; p=0.65), anaemia (CHOP=12, 12%; EPOCH=1, 8%; P=0.71), thrombocytopenia (CHOP=7, 6%; EPOCH=0; p=0.36), sepsis (CHOP=1), treatment related death (EPOCH=1) and hepatic encephalopathy (CHOP=1). Conclusion: Overall survival rate, treatment response rates and adverse events were not different in patients treated with CHOP and EPOCH.


Introduction
Globally, the estimated incidence of non-Hodgkin's lymphoma (NHL) was 5/100,000 in 2012 (1). In Uganda, the incidence of NHL was 1,426/100,000 from the year 1991-2010 (2). The vast majority of people living with HIV are in low-and middle-income countries(3). In 2016, an estimated 1.4 million people were living with HIV in Uganda (4). The incidence of NHL remains significantly higher in HIVpositive patients compared with the HIV negative patients, even in the era of combination antiretroviral therapy (ART) (5)(6)(7)(8). The outcomes for patients with HIV associated NHL and non-HIV associated NHL treated with chemotherapy in resource limited settings is still disappointingly low (9).
Notwithstanding, the advent of combination ART in the 1990s resulted in reduced morbidity and mortality from HIV infection (10)(11)(12), and improved NHL specific outcomes (13) compared to the early days of the AIDS epidemic (14).
The optimal chemotherapy regimen for the treatment of HIV associated NHL in low resource settings is still unclear. Moreover, first-line chemotherapy regimens that are effective for lymphomas of similar histology in the HIV negative population have been associated with lower rates and durability of complete response in the HIV positive population (15). First line chemotherapy regimens used to treat HIV associated NHL in the post ART era include but not limited to the infusional cyclophosphamidedoxorubicin-etoposide (CDE) with complete remission rate (CR) of 42%, median survival time of 17.8month, and 1-year survival rate of 55% (16); and Cyclophosphamide, Doxorubicin, Vincristine and Prednisolone (CHOP) regimen with complete remission in 57.6% and 47% in patients treated with R-CHOP and CHOP respectively, with an overall survival of about 35 months for R-CHOP and 28 months for CHOP (17).
Etoposide, Prednisone, Vincristine, Cyclophosphamide, and Doxorubicin (EPOCH) infusional chemotherapy is a relatively recent combination in this category. Hitherto, it has been reported to achieve 79% CR rate and 72% overall 2-year survival rate in patients with HIV associated NHL (18).
Use of the EPOCH regimen has been suggested as unreasonable in low resource settings (19)  Continuous infusion of EPOCH regimen was done using a Y-junction infusion sets. Normal saline was infused concurrently through the Y-junction pot until the chemotherapy infusion was completed. No patient received rituximab or G-CSF. Data concerning additional medications, especially ART and pneumocystis jiroveci pneumonia (PCP) were included. All study data was abstracted from patients' charts as recorded by the treating oncologists. Diagnosis was made by examination of the lymphnode/tissue, and bone marrow histopathology and immunohistochemistry. The NHL was stage using the Ann Arbor staging system (20). Adverse events were recorded at each chemotherapy cycle and graded according to the NCI Common Terminology Criteria for Adverse Events v5. In participants who had CT scans at baseline and end of therapy, treatment response was assessed using the Lugano Criteria but without the use of positron Emission Tomography (PET). The last date of hospital review or death was recorded for survival analysis. Data was manually abstracted from charts using a standard data collection tool by the study assistant. Completed data collection tool was checked for completeness and accuracy by the principle investigator. Data was then coded, and double entered into a computer using Epidata version 3.1 (Epidata association, Denmark) before exporting into STATA Version 14 (StataCorp, USA) for analysis.

Study variables
The study variables included participant's age and sex; type and stage of lymphoma, comorbidities, baseline ECOG performance score, body mass index (BMI), and number of chemotherapy cycles received, ART regimen and other additional concomitant medications received, and B-symptoms; haematological adverse events as demonstrated by the complete blood count (CBC) and other adverse events during chemotherapy, and disease response at completion of treatment.

DATA ANALYSIS
Continuous variables were expressed as means and standard deviation (SD) if normally distributed or medians and inter quartile ranges (IQR) if skewed; categorical variables were described using frequencies and percentages; the overall treatment response rates (complete response, partial response) were estimated for both CHOP and EPOCH chemotherapy arms using the binomial proportion and its 95% confidence interval as separate categories using the total number of participants enrolled at baseline as the denominator in each study arm. The two treatment regimens were compared with respect to overall response rate using Fisher's exact test; the proportion of patients who completed each treatment regimen were described as separate categories using the total number of participants enrolled at baseline as the denominator in each study arm and compared using Fisher's exact test. Reasons for non-completion of chemotherapy were also described and compared using Fisher's exact test. The one year overall survival was described for patients in the two treatment regimens using the Kaplan-Meier curves and compared using log-rank test. A two sided alpha of 0.05 was used for comparison. Cox proportional hazards model was constructed to evaluate the association between patient characteristics and OS. Hazard ratios and 95% confidence intervals were generated, with hazard ratio < 1.0 indicating survival benefit.
There were no statistical differences in BMI, age at diagnosis, types of lymphoma, ECOG performance score, presence of B-symptoms, comorbidities and stage of NHL between the two treatment groups (Table 1). Majority of patients,105(97.2%) in the CHOP group were already on ART prior to starting chemotherapy, and only 3 patients (2.8%) in the CHOP group started ART after completion of chemotherapy cycles. All patients in the EPOCH group were on ART prior to starting chemotherapy.

Treatment response and survival
Only patients who completed at least 6 cycles of chemotherapy were evaluated for response. Overall treatment response rate was 40% in the CHOP group and 59% in the EPOCH group (P=0.66).
Complete response (CR) was achieved in 29(27%) patients in the CHOP group and 5(42%) patients in the EPOCH group. Partial response was observed in 16(13%) patients in the CHOP group and 2(17%) patients in the EPOCH group. There was no statistical difference between all the response categories (Table 4).

Discussion
This retrospective study observed no statistical difference in one year OS, treatment response rate or adverse events between patients with HIV associated NHL treated with CHOP and EPOCH regimens in low resource setting.
The one year OS of patients treated with CHOP and EPOCH in our study is comparable to other results in Africa. A study in Malawi reported a one year OS of 59.4% in patients with HIV associated lymphomas treated with CHOP(21); in Botswana, the 1 year survival rate in patients with DLBCL was 52.8% following treatment with CHOP(± R)(22); a retrospective study in south Africa on patients with HIV associated DLBCL treated with CHOP and concomitant ART reported a 2 year OS of 40.5% (23). It has been noted that the outcomes of treating aggressive B cell NHL with chemotherapy appear to be similar in HIV-positive and HIV-negative populations especially in the era of combination ART (8,24,25). Some studies report a CD4 count < 100/uL as a negative prognostic finding (17,26). However, our study did not have data on CD4 counts.
Normal BMI and completion of 6 or more cycles of chemotherapy were associated with favourable survival in our study. A retrospective study on HIV associated lymphomas in Nigeria reported stage of lymphoma as the only factor predictive of survival (27). Other factors that have been noted to predict survival include type of lymphoma (28), age, ECOG performance scores, stage of lymphoma and LDH level (29). The numerous losses to follow up of patients in our study might have influenced our results.
Our result show a trend towards better survival with the EPOCH regimen. The initial study on EPOCH in patients with DLBCL reported better OS rate at 62 months of 73% than with CHOP (30). Subsequent addition of Rituximab to EPOCH produced even better results of a 12-month PFS rate of 85% (31,32).
Our study showed a trend in favour of EPOCH regimen in all the treatment response categories.
Nonetheless, the CR rate in our study is less impressive than reported in other studies. A study by the AIDS-Malignancies Consortium Trial 010, a phase 3 trial of CHOP vs R-CHOP in patients with HIVassociated NHL showed a better CR of 47% for CHOP (17) than was observed in our study (27%).
Other studies on the treatment of HIV associated NHL with CHOP or EPOCH in the sub-Saharan Africa show similar results with our study. De Witt (2013) in their retrospective study on patients with HIV associated DLBCL treated with CHOP (n = 34) and CHOP-like (n = 2) regimens in south Africa reported CR of 38.9% (23). A smaller study in Malawi (n = 12) on patients with plasmablastic lymphoma in HIV positive (n = 6) and HIV negative patients (n = 6) in which 8 patients were treated with CHOP and 4 patients were treated with modified EPOCH reported an overall CR in 42% of the patients (CHOP = 25%; EPOCH = 75%) (33). In another retrospective study in south Africa where only 4 cases (< 1%) were HIV(+) and no specific chemotherapy regimens were defined, the overall CR range was 46-75% for all subtypes of NHL(34); and in a large retrospective study of paediatric Burkitt's Lymphoma in Uganda where 70 of the 228 patients were HIV positive with a mean age of 6.7 years and no specific chemotherapy was mentioned, CR was 36%(35).
The low completion rate of chemotherapy in our study may have partly contributed to the low treatment response rates. Adverse events contributed to non-completion of chemotherapy in only 10% of the patients whereas a majority of patients did not have clearly documented reasons for noncompletion of chemotherapy in our study. This might be due to the large number of losses to follow up in our study.
Our study show that haematological adverse events (AEs) were the most prevalent in the two treatment groups but with no statistical difference between the groups. However, this may be due to the limitation associated with data abstraction from patient charts that may not easily capture non lab based AEs. Takondwa et al. (33) in their study in Malawi reported treatment delays in patients receiving EPOCH (n = 4/4) and patients treated with CHOP (n = 4/8) due to grade 3/4 neutropenia and grade 3 anaemia (CHOP = 1). It is possible that the small number of patients in our study and that by To the best of our knowledge, our study is one of the few studies comparing CHOP and EPOCH regimens in the treatment of HIV associated NHL in the sub-Saharan Africa. It included a relatively larger sample size than other previously related studies (33). However, the inherent limitations of retrospective studies such as inadequate documentations may limit its generalizability. It is also not clear if the differences in the baseline parameters of the patients in our study such as sex and intake of Tenofovir/Lamivudine/Efavirenz between the CHOP and EPOCH arms contributed to the reported observations. Another limitation was the inadequacy of immunohistochemistry to refine the diagnosis.
It is possible that some patients who were treated with the CHOP regimen might have had more aggressive histological subtypes of NHL such as Burkitt lymphomas or plasmablastic lymphomas that were treated inadequately. CHOP regimen is considered less intensive and therefore inadequate for the treatment of Burkitt lymphoma(36, 37). Also, standard CHOP is considered an inadequate therapy for Plasmablastic lymphoma (38).
In conclusion, our study showed comparable treatment outcomes between infusional EPOCH and CHOP regimens in HIV associated NHL in a resource limited setting. However, these results should be taken with caution due to the inherent limitations of the retrospective nature and the small sample size; a prospective study with a larger sample size may be more informative.

Ethics approval and consent to participate
Waiver of consent and study approvals were obtained from the Uganda Cancer Institute Research Ethics Committee (Reference number: 15-2018) and the study was registered at the Uganda National Council for Science and Technology (Reference number: HS 2568).

Availability of data and materials
All data generated or analysed during this study are included in this published article

Competing interests
All the authors have declared no conflicts of interest.

34.
Sissolak G, Juritz J, Sissolak D, Wood L, Jacobs P.     NB: There was no statistical significant difference between arms CHOP and EPOCH in terms of CR, PR, and PD (P > 0.05).   Overall survival among study population Overall survival among patients treated with CHOP vs EPOCH chemotherapy regimens

Supplementary Files
This is a list of supplementary files associated with this preprint. Click to download.