Nigerian Journal of Paediatrics

Treatment outcome and Abandonment rates in a retrospective cohort of children with Burkitt lymphoma in Kano Nigeria

Niger J Paediatr 2018; 45 (3):159 -162

ORIGINAL

Abdullahi SU

CC – BY Treatment outcome and abandon-

Ibrahim M

Ahmed AK

ment rates in a retrospective

Abdulazeez I

cohort of children with Burkitt

Ali BU

lymphoma in Kano, Nigeria

DOI:http://dx.doi.org/10.4314/njp.v45i3.4

Accepted: 19th September 2018

Abstract : Background: Burkitt

sis of BL were admitted during the

lymphoma (BL) is highly chemo-

study time period, of which 56

Abdullahi SU (

)

therapy-sensitive tumor and one

(88.9%) medical records were re-

Ibrahim M, Ahmed AK

of most common childhood can-

trieved and included in the final

Abdulazeez I

cers in Nigeria. Unlike in high

analysis. The median age was 8

Department of Paediatrics, Bayero

income settings where cure rate is

years (range 3 – 18 years), with a

University/ Aminu Kano Teaching

very high, the outcome of children

male: female ratio of 1.6:1. Only

Hospital, Kano

with BL is still poor in low in-

eight patients completed therapy

Email: dr_suak@yahoo.com

come countries like Nigeria. The

without relapse (14.3%). Of those

main contributing factor to dispar-

who did not complete planned

Ali BU

ity in survival rate is treatment

therapy, 7(14.6) died prior to the

Department of Histo-pathology,

abandonment (TA). This study

completion of therapy, leaving 41

Bayero University/ Aminu Kano

aimed to determine treatment out-

(85.4%) who met criteria for TA.

Teaching Hospital, Kano

comes and TA rates in a retro-

Conclusion: Treatment completion

spective cohort of Nigerian chil-

and overall survival rates are poor

dren admitted with BL.

among children with BL, due to in

Methodology: We conducted ret-

part high treatment abandonment

rospective charts review of all

rate. Leaving outside the treat-

children (<13 years) with BL ad-

ment is a significant predictor of

mitted at paediatric oncology unit

TA. Local oncology services might

of AKTH between January1 ,

need to be decentralized in our

2014 and December 31 , 2016.

environment to substantially re-

Data extracted from patients’

duced abandonment to improve

medical chart records included

overall survival.

age, gender, duration of symp-

toms, clinical stage of tumor, and

Key words: Burkitt lymphoma,

treatment outcome.

Treatment outcome, Treatment

Results: 63 children with diagno-

abandonment, Low income setting

Introduction

lasted for a period of 4 weeks ⁶ is a possible explanation

for this disparity and is described as a significant con-

Burkitt lymphoma (BL) is an aggressive childhood B-

tributor to treatment failure and overall cancer mortality

in LICs.

4,7

cell tumor occurring mainly in Sub-Saharan Africa

Reasons for cancer-related TA would differ

where its

incidence ranges between 0.09 – 7.5 per

geographically depending on local prevailing factors in

100,000.

In Nigeria, it is one of the most common

these LICs. As such, knowledge of these factors could

cause of admissions for childhood malignancies with

help further understanding of cancer-related TA in these

hospital prevalence ranging between 18.3 and 65 percent

countries and potentially inform policy at reducing its

of all paediatric cancer-related admissions. Previous

burden. In Nigeria, very few studies have determined the

burden of TA and its associated factors. This study thus

study from AKTH reported 29.8% prevalence of BL in a

retrospective cohort of children admitted with malig-

investigated the treatment outcome and treatment aban-

nancy . Burkitt lymphoma is potentially curable disease,

donment rate in a retrospective cohort of children admit-

with cure rate exceeding 90% in high-income settings

ted with BL at a tertiary referral health facility in Kano,

however, in low income countries where majority of

north-western Nigeria.

affected children live, the cure rate is less than 35%.

4,5

Treatment abandonment (TA), defined as failure to initi-

ate chemotherapy or any interruption of treatment which

160

Methodology

Table 1: Socio-demographic and Clinical characteristics

Study site

of the study population (n=56)

Variable

Frequency (%)

We conducted retrospective charts review of children

Socio-demographic

(<13 years) with BL admitted at paediatric oncology

Median age in years (IQR)

8.0 (4.0)

unit of AKTH between January1 , 2014 and December

Gender

31 , 2016. Aminu Kano Teaching Hospital is a fee for

Male

37 (66.1)

service tertiary referral health facility located in Kano,

Female

19 (33.9)

the second largest metropolis in Nigeria. The hospital

Clinical Characteristics

receives patients from Kano, neighboring states and

Tumour site

from across the border (Niger republic).

Jaw

21 (37.5)

Abdomen

18 (32.1)

Data collection

Jaw + Abdomen

14 (25.0)

Ocular

3 (5.4)

Ethical approval for the study was obtained from the

Histologic diagnostic method

ethics committee of AKTH. Data extracted data from

FNAC

52 (92.9)

patients’ medical record included patients’ age, gender,

Tissue biopsy

4 (7.1)

date of admission, Address (whether resident within

Tumour clinical staging*

Kano metropolitan area or its environ or from neighbor-

9 (16.1)

ing states), time to onset of 1 symptom, number of cy-

1 (1.8)

cles of chemotherapy completed, last visit date, , and

6 (10.7)

clinical stage of tumour at presentation using modified

27 (48.2)

Ziegler staging,

and treatment outcome (completed

13 (23.2)

treatment, refused treatment/abandonment, did not re-

* Zeigler and Magrat staging

turn for treatment; or death during therapy from any

cause). For the treatment, we adapted the practical rec-

ommendations for the management of children with

Table 2: Comparison of potential risk factors between treat-

endemic BL in a resource limited setting, using combi-

ment abandonment group and those who did not abandon treat-

nation of cyclophosphamide, vincristine and intra-the cal

ment

methotrexate

Variable

Abandoned

Did not

P-

Total

treatment

Abandon

value

(%)

group (%)

treatment

n=56

Statistical analysis

group

n=41

(%) n=15

Basic descriptive statistics were performed on retrieved

History of use of traditional medication

data using frequency and percentages. We used chi-

Yes

11 (26.8)

3 (20.0)

0.74

14(22.9)

square and t-tests for univariate analysis comparing be-

= No

30 (73.2)

12 (80.0)

42(77.1)

tween the two groups: Those who did not abandon ther-

*Cancer Stage

0.28

apy and those who abandoned therapy. Those who died

Advanced

8(19.5)

5 (33.3)

13(23.2)

from any cause during treatment were included. Statisti-

Not Advanced

33(80.5)

10 (66.7)

43(76.8)

Place of residence

cal significance was set at 5% for this comparison.

0.055

Within Kano

13(31.7)

9 (60.0)

22(39.3)

Outside Kano

28 (68.3)

6 (40.0)

34(60.7)

Median time from

0.79

first onset of symp-

8(16.0)

8(12.0)

Results

toms to presentation

at treatment centre in

weeks (IQR)

Two hundred and twenty eight (228) children with ma-

13.0 (3.0)

11.0 (3.0)

0.63

Median in-patient

13.0

lignancies were admitted over the study period, of which

time to final diagno-

(8.5)

63 (27.6%) were BL. Of the 63 BL cases, clinical re-

sis in days (IQR)

2 (1)

3 (1)

0.35

cords were recovered for 56 (88.9%) children, and these

Number of previous

2 (1)

health encounters

were included in the final analysis.

before presenting at

treatment centre

n=54

Patients’ Characteristics

*-Advanced: stage C and D; not advanced: stage A, B, and

Median age (IQR) of the study population was 8.0 years

AR, -Two missing variables, IQR – interquartile range, a-

(4.0) and about two-thirds of the study population were

Fischer’s exact test b - Mann-Whitney U test c- Chi-square test.

male with M: F ratio of 1.9:1. Majority of the patients

&- combines those who completed therapy and those who died

presented with jaw swellings (39.6%) and about a fourth

from any cause during therapy

of the cohort presented with advanced disease (Burkitt

stage D). All patients were HIV sero-negative, and all

Treatment outcome

had diagnoses confirmed by histo-pathology.(Table 1)

Of the 56 patients included in our analysis, only eight

patients completed therapy without relapse (14.3%). Of

161

those who did not complete planned therapy, 7(12.5%)

lymphoma is a highly chemo-sensitive cancer that gives

died prior to the completion of therapy, while 41

a picture of clinical cure even after a single chemother-

(73.2%) met criteria for treatment abandonment. Major-

apy cycle and misperception that further treatment is not

ity (51.8%) of patients had discontinued chemotherapy

needed once the tumor has receded, while leukaemia on

by the second cycle. (Figure 1) Median time for stop-

the other hand tends to be aggressive requiring frequent

ping therapy was 16.5 weeks. Lower parental income

hospital admissions and as such, patients might be less

and longer distance from the treatment center were sig-

likely to abandon treatment. The El-Salvadorian study

nificantly associated with higher rate of treatment aban-

on the other hand was conducted in a health facility that

donment.

provided free cancer treatments as opposed to our study

where virtually all patients paid out of pocket for their

Fig 1: Bar chart showing number of chemotherapy cycles

treatments. Mostert et al have shown patients who had

completed

financial help with cancer treatment were less likely to

abandon their treatment. In high-income countries, for

many patients, a portion of the medical expenses is paid

by their health insurance plan. For individuals without

health insurance or who need financial assistance to

cover care costs, resources are available, including gov-

ernment sponsored programs and services supported by

voluntary organizations.

On the other hand, in low

resource countries, parents of children with malignancy

have to pay out-of pocket for treatment and no health

insurance and resources to help families through finan-

cial difficulties and the health infrastructure are ill-

equipped to provide appropriate supportive care for the

affected children. In most LMICs, cancer is often de-

tected late very late, that even the most effective treat-

ment will not result in long-term cure, and out-of-pocket

spending is often wasted because it contributes nothing

to improve health. Additionally, care may be coupled

with prohibitive transportation costs and investments of

Number of chemotherapy cycles

time that include long turn-around time to access care.

Our study also documented that roughly half of patients

had abandoned treatment early in chemotherapy, as it

Discussion

has been reported in other studies.

10,17

This finding might

be explained by the highly chemo-sensitive nature of BL

Our study investigated treatment outcome and abandon-

which could give a false impression of cure to parents.

ment rates in a retrospective cohort of children who

In our study, place of residence was the most significant

were admitted at a referral tertiary health facility in

predictor for TA in our cohort. Patients who lived out-

North-Western Nigeria. We found treatment abandon-

side the state where our facility is located where more

ment (TA) was high (73.2% of our cohort) with greater

likely to abandon treatment compared to those who lived

than half of children discontinuing therapy by the 2

near the treatment facility. We used this categorization

cycle of chemotherapy. Among the risk factors for TA

as a proxy to investigate the influence of distance on TA

investigated, living outside treatment area was the only

among our cohort. Such findings of distance from treat-

significant predictor of TA.

ment centers affecting cancer treatment completion have

been shown in other previous studies in the developing

Similar TA rates have been previously described in

world.

12,18

Other factors such as parental inability to pay

other regions in Nigeria and parts of Africa. In a study

for treatment and low parental education have been

by Meremikwu etal in South-Eastern Nigeria, two-

demonstrated to be associated with TA in other stud-

thirds of children with Burkitt lymphoma discontinued

ies.

11,13

Due to the retrospective nature of our data and

cancer treatment. In Western Kenya, Njuguna et al

missing information regarding these variables we were

reported a 54% prevalence of treatment abandonment,

unable to determine what influence this might have had

while Slone et al and Boer et al, both in Zambia re-

in our cohort.

ported prevalence of 45.7% and 51% respectively. Our

findings however contrast those of Kulkarni and Marwa-

Our study is one of few studies that investigated treat-

who documented a TA prevalence of 18.1% and

ment outcome and TA rates in Nigeria. We however

Bonilla et al who documented 13% in Indian and El-

recognize some limitations of our study. Due to the ret-

Salvadorian children respectively. Differences in our

rospective nature of the study we could not determine

study populations might explain why our findings differ

the cause of death of patients with BL or investigate

from these two studies. While our study was conducted

outcomes for those who abandoned treatment. Future

in children with Burkitt lymphoma, the Indian study was

studies in this region might benefit from a prospective

conducted among those with acute leukaemia. Both can-

mix method design, where phone interviews are con-

cers are known to have varying natural histories. Burkitt

162

ducted with parents of children who abandoned cancer

systems might need to be developed to actively monitor

treatment.

and trace children who have defaulted from chemother-

apy.

Similarly, there is need to improve treatment outcome

through advocacy and education of local health workers,

Conclusion

through collaborations between LMIC and HIC centers,

capacity building, provision of cancer centers of excel-

Our study found high rate of treatment abandonment as

lence, and standards of care. Approaches need to be inno-

the major cause of treatment failure and overall poor

vative, yet realistic, and the LMIC team must set priori-

outcome, and distance from our treatment site seems to

ties.

be an explanation for this high rates. There is a clear

need for decentralization of oncology services such that

Conflict of interest: None

they are more assessable. In the interim however,

Funding: None

References

Fisher SG, Fisher RI. The epidemi-

9. Ziegler JL. Burkitt’s lymphoma. N

16. Felicia Knaul, Susan Horton,

ology of non- Hodgkin’s lym-

Engl J Med. 1981;305(13):735 –

Pooja Yerramilli, Hellen Gel-

phoma. Oncogene. 2004; 23

745.

band, RifatAtun. Financing Can-

(38):6524 – 34.

10. Meremikwu MM, Ehiri JE,

cer Care in Low-Resource Set-

Brown BJ. A review of the litera-

Nkanga DG, Udoh EE, Ikpatt OF,

tings. In Gelband H, Jha P,

ture on childhood Burkitt lym-

Alaje EO. Socioeconomic con-

Sankaranarayanan R, et al., edi-

phoma in Nigeria. Niger J Paedi-

straints to effective management of

tors. Cancer: Disease Control

atr. 2016;43(1):1 – 7.

Burkitt’s lymphoma in south -

Priorities. Washington (DC): The

Shehu UA, Adegoke SA, Abdul-

eastern Nigeria. Trop Med Int

International Bank for Recon-

salam U, Ibrahim M, Oyelami OA,

Health. 2005;10(1):92 – 98.

struction and Development /The

Adeodu OO. Pattern of childhood

11. Njuguna F, Mostert S, Slot A,

World Bank; 2015 Nov 1.

malignant tumours in two tertiary

Langat S, Skiles J, Sitaresmi MN,

17. Bonilla M, Rossell N, Salaverria

teaching hospitals in Nigeria: com-

et al. Abandonment of childhood

C, Gupta S, Barr R, Sala A, et al.

parative Study. Niger J Paediatr.

cancer treatment in Western

Prevalence and predictors of

2013;40(2):175 – 178.

Kenya. Arch Dis Child. 2014; 99

abandonment of therapy among

Molyneux EM, Rochford R, Grif-

(7):609 – 614.

children with cancer in El Salva-

fin B, Newton R, Jackson G,

12. Slone JS, Chunda-Liyoka C, Perez

dor. Int J Cancer. 2009; 125

Menon G, et al. Burkitt’s lym-

M, Mutalima N, Newton R, Chintu

(9):2144 – 2146.

phoma. The Lancet. 2012;379

C, et al. Pediatric Malignancies,

18. Metzger ML, Howard SC, Fu LC,

(9822):1234 – 1244.

Treatment Outcomes and Aban-

Peña A, Stefan R, Hancock ML,

Peter Hesseling, TrijnIsraels,

donment of Pediatric Cancer Treat-

et al. Outcome of childhood acute

MhamedHarif, Guillermo Chan-

ment in Zambia. Moormann AM,

lymphoblastic leukaemia in re-

tada, Elizabeth Molyneux. Practi-

editor. PLoSONE. 2014 Feb 21;9

source-poor countries. The Lan-

cal Recommendations for the Man-

(2):e89102.

cet. 2003; 362(9385):706 – 708.

agement of Children withEndemic

13. Boer JD, Boellaard TN, Parkinson

Burkitt Lymphoma (BL) in a Re-

S, Blanchard E, Heij HA. Patient

source Limited Setting. Pediatr

compliance in the treatment of

Blood Cancer 2013;60:357 – 362

Burkitt’s lymphoma in rural Zam-

Mostert S, Arora RS, Arreola M,

bia: A retrospective study on 80

Bagai P, Friedrich P, Gupta S, et

Burkitt’s lymphoma patients in

al. Abandonment of treatment for

Katete, Zambia. Afr J Paediatr

childhood cancer: position state-

Surg. 2009 Jan 1;6(1):3.

ment of a SIOP PODC Working

14. Kulkarni KP, Marwaha RK. Pat-

Group. Lancet Oncol. 2011;12

tern and implications of therapy

(8):719 – 20.

abandonment in childhood acute

Arora RS, Pizer B, Eden T. Under-

lymphoblastic leukemia. Asian Pac

standing refusal and abandonment

J Cancer Prev. 2010;11(5):1435 –

in the treatment of childhood can-

cer. Indian Pediatr. 2010;47

15. Mostert S, Njuguna F, Van de Ven

(12):1005 – 1010.

PM, Olbara G, Kemps L, Musimbi

Ishaya AI, Adah R. The Role of

J, et al. Influence of health-

Socio-economic Factors in Aban-

insurance access and hospital re-

donment of Cancer Treatment

tention policies on childhood can-

among Paediatric Patients in Jos

cer treatment in Kenya. Pediatr

Nigeria. [cited 2017 Feb 8]; Avail-

Blood Cancer. 2014; 61(5):913 –

able from: http://

918.

search.proquest.com/

open-

view/91d2470df5cf22ab062672cc2

33834b2/1?pq-

origsite=gscholar&cbl=626442