ORIGINAL ARTICLE
Malignancies in South African Children With HIV Alan Davidson, MD, MPhil,* Rosalinda D. Wainwright, MD,w David K. Stones, MD,z Mariana Kruger, MD, PhD,y Marc Hendricks, MD,* Jennifer Geel, MD,8 Janet Poole, MD,8 David Reynders, MD,z Fareed Omar, MD,z Rema Mathew, MD,# and D. Cristina Stefan, MD, PhDy
Objectives: In 2008 the South African Children’s Cancer Study Group decided to review the epidemiology, management, and chemotherapy response of HIV-positive children with malignancy. Methods: This is a retrospective analysis of data collected from the records of HIV-positive children diagnosed with malignancy at 7 university-based pediatric oncology units serving 8 of the 9 provinces in South Africa. Results: Two hundred eighty-eight HIV-positive children were diagnosed with 289 malignancies between 1995 and 2009. Age at diagnosis ranged from 17 days to 18.64 years; median 5.79 years. Of the 220 HIV-associated malignancies, there were 97 Kaposi sarcomas, 61 Burkitt lymphomas, 47 other B-cell lymphomas including 2 primary central nervous system lymphomas, 12 Hodgkin lymphomas, and 3 leiomyosarcomas. Sixty-nine patients presented with non–AIDS-defining malignancies. More than 80% presented with advanced disease. Most patients (76.7%) were naive to antiretroviral therapy; 22.2% did not have access because it only became available in public hospitals in 2004. One hundred ninety-seven children were treated with curative intent; 91 received palliative care due to advanced malignancy and/or advanced HIV disease. Nearly one third had coexisting pathology, mostly tuberculosis. Overall survival for the whole group was 33.7%, but was 57.8% for those treated with antiretroviral therapy and chemotherapy. Conclusions: The study shows more Kaposi sarcoma and fewer primary central nervous system lymphomas among HIV-positive children than that is reported in the developed world, but confirms a higher incidence of non-Burkitt B-cell lymphoma than in HIVnegative children. The high number of non–AIDS-defining malignancies underscores the prevalence of HIV-AIDS in South Africa. The overall survival should improve with universal access to antiretrovirals and earlier diagnosis. Key Words: HIV, infectious disease, pediatric oncology, Africa
(J Pediatr Hematol Oncol 2014;36:111–117)
Received for publication December 2, 2012; accepted April 22, 2013. From the *Department of Paediatrics and Child Health, Red Cross Children’s Hospital and the University of Cape Town; yDepartment of Paediatrics and Child Health, Tygerberg Hospital and Stellenbosch University, Tygerberg, Cape Town, Western Cape; wDepartment of Paediatrics and Child Health, Chris Hani Baragwanath Academic Hospital; 8Department of Paediatrics and Child Health, Charlotte Maxeke Johannesburg Academic Hospital, University of the Witwatersrand, Johannesburg; zDepartment of Paediatrics, Steve Biko Academic Hospital, University of Pretoria, Pretoria, Gauteng; zDepartment of Paediatrics and Child Health, Universitas Academic Hospital Complex and the University of the Free State, Bloemfontein, Free State; and #Department of Paediatrics, Frere Hospital and Walter Sisulu University, East London, Eastern Cape. The authors declare no conflict of interest. Reprints: Alan Davidson, MD, MPhil, Department of Paediatrics and Child Health, Ward G1, Red Cross Children’s Hospital, Private Bag, Rondebosch, Cape Town 7700, South Africa (e-mail: [email protected]). Copyright r 2013 by Lippincott Williams & Wilkins
J Pediatr Hematol Oncol
T
he advent of the HIV epidemic has brought about new challenges for pediatric oncologists. Although the incidence of cancer in HIV-negative children is estimated to be 0.13 to 0.15 per 1000,1,2 early reports from the United States found an incidence of 20/1000 in HIV-infected children.3,4 Similarly elevated figures were subsequently reported by European investigators: in Italy, before the widespread introduction of highly active antiretroviral therapy (HAART), the cancer rate in HIV-infected children was 4.49/10005; in Spain, an audit of diagnoses on discharge of all children admitted to hospitals found a malignant disease rate of 9.4/1000 for HIV-positive children, whereas HIV-negative children had a rate of 1.2/1000.6 This increase was due to an increased incidence of non-Hodgkin lymphomas, leiomyomas and leiomyosarcomas as well as Hodgkin lymphomas, acute leukemias, and other less commonly encountered tumors.7 The incidence of Kaposi sarcoma has also increased dramatically in children with AIDS.8–10 However, this increase is almost solely limited to Africa. Elsewhere it is seen mostly in adults living with HIV infection.11 Certain malignancies encountered more often in children with AIDS are associated with viral infections: the Epstein-Barr virus is very frequently identified in Burkitt lymphoma, mixed cellularity Hodgkin lymphoma, and leiomyosarcoma,12 whereas the human herpes virus 8 (HHV-8) is considered to be the etiologic agent of Kaposi sarcoma.13 The higher incidence of viral-related cancers in children with AIDS suggests an association between HIV and other viruses.14 The rising prevalence of HIV has also resulted in HIV-positive children presenting with non–AIDS-defining malignancies such as Wilms tumor and acute lymphoblastic leukemia.9,15–17 We believe that this is a function of both an increase in the number of children with HIV and increased survival of this cohort as the number of children on HAART expands.18 HIV seems to have affected not only the incidence but also the presentation of these malignancies. These patients typically present with unusual histological subtypes in unusual sites. Diffuse large B-cell lymphomas and primary central nervous system lymphomas (PCNSLs) make up a substantial portion of the HIV-related malignancies reported in studies from the United States3,4,19 and Italy,20 which is in strong contrast with the spectrum of disease seen in immunocompetent children. Even non–AIDS-defining malignancies may present atypically in immunocompromised children. For Africa, the higher incidence rates of malignant disease in children with HIV are augmented by the huge numbers of those infected. Out of a total of 3,400,000 HIV-positive children younger than 15 years of age in 2010, 3,100,000 (91.2%) were resident in sub-Saharan Africa.18 Although comprehensive prevention and treatment
Volume 36, Number 2, March 2014
www.jpho-online.com |
111
Davidson et al
strategies have reduced and stabilized the prevalence of HIV in Europe and North America, only 21% of children in subSaharan Africa who need to undergo HAART are receiving treatment.18 In South Africa, according to the mid-2009 population statistics, there were 15,500,700 children aged 0 to 15 years.21 At the end of 2010, 108,682 children were reported to be on HAART in South Africa, out of an estimated 300,000 (270,000 to 340,000) children requiring immediate HAART, an estimated coverage of 36%.18 Using the published European estimates, South Africa should expect at least 1000 new cases of HIV-related malignancies in children per year,1,5,22 but this figure is probably overstated in our context because far fewer children in the Italian cohort would have succumbed to gastroenteritis and pneumonia; leaving proportionally more survivors vulnerable to AIDS-related lymphomas. The widespread introduction of HAART has reduced the number of HIV-related malignancies but even with HAART, the overall rate of cancer in these children remains higher (0.76/1000 in the Italian study) than in those not infected by the virus.5 Moreover, the decrease in cancers related to HAART is significant only after 2 years of administration.23 It seems that although HAART has increased the life expectancy of children with HIV, it has not reduced the burden of malignant disease associated with the virus. Although a number of cases of Kaposi sarcoma might improve on HAART alone, most children eventually require chemotherapy.10 For other malignancies occurring in HIV-positive children many oncologists now recommend full-dose chemotherapy and HAART. It should be borne in mind that the interactions between HAART and standard chemotherapy regimens are insufficiently studied and potentially dangerous.24,25 There are still insufficient data to support the growing body of opinion that the outcomes that can be achieved by combining the 2 groups of drugs are similar to those in HIV-negative children. Treatment is made even more challenging by multiple comorbidities such as chronic lung disease, tuberculosis, and other infections. These have multiple effects, impacting on treatment decision-making and clinical outcomes by way of drug-drug interactions and non–treatment-related morbidity. For more than 15 years, South African pediatric oncologists have treated HIV-positive children with malignancy. Many of them have achieved remission and have been long-term survivors on HAART.
MATERIALS AND METHODS In 2008 members of the South African Children’s Cancer Study Group, representing the 7 university-based pediatric oncology units (POUs) and several satellite centers, decided to audit this cohort of HIV-positive children with malignancy in the hope that it would provide meaningful insights, and support future collaborative prospective clinical trials. Six of the 7 POUs in South Africa participated in addition to 1 satellite center (which is also now a university-based POU). The participating institutions include referral centers that serve 8 of the 9 provinces. All authors participate in the activities of the South African Children’s Cancer Study Group that holds annual meetings, collects registry data, and sponsors joint protocols. Since all but one university-based POU in South Africa participated, our data are representative but the cases reported here do not represent all pediatric cases of HIV and malignancy diagnosed during this period.
112 | www.jpho-online.com
J Pediatr Hematol Oncol
Volume 36, Number 2, March 2014
The aim was to audit malignancies occurring in HIVpositive South African children in order to document patient demographics, relative frequency, stage at diagnosis, degree of immunosuppression (WHO criteria), comorbidities and their impact on survival, and outcome correlated with the use of chemotherapy, combined with radiotherapy or surgery as indicated, and HAART, alone or in combination. Data were collected from the records of 288 HIV-positive children diagnosed with malignancy between 1995 and 2009, and entered into a Microsoft Access database. Institutions assigned a numerical identifier to each case in order to blind the principal investigator who managed the data. Uniformity of histological evaluation was achieved by having most samples processed by pathologists of the National Health Laboratory Service, where similar laboratory protocols and assessment criteria are applied nationwide. Treatment outcomes were analyzed using Statistica software; survival analysis was performed using KaplanMeier curves. A P value of r0.05 was regarded as significant. Confidentiality for individual patients was assured as the data provided to the principal investigator were anonymized. As this is a retrospective study, we did not seek informed consent from individual patients, but obtained the consent of the custodian of the data to access patient files. Approval was obtained from the relevant Human Research Ethics Committees of the participating centers.
RESULTS Between January 1995 and December 2009, 288 children with HIV infection and cancer were admitted to the pediatric hematology-oncology units participating in the study. The median age was 5.79 years, with a range of 17 days to 18.64 years, and there was a male to female ratio of 1.64:1. Only 4 of the patients in this series were over the age of 15 at diagnosis (2 children with diffuse large B-cell lymphomas, 1 with Kaposi sarcoma, and 1 with adenocarcinoma of the rectum). All these children were being followed up by pediatric infectious disease services. Four children were of ethnically mixed race and the remaining 284 were black South Africans. The annual number of cases increased slowly from 1995, following the gradual spread of the epidemic (Fig. 1). Kaposi sarcoma registered the highest increase in annual numbers, followed by Burkitt lymphoma and non-Burkitt B-cell lymphoma. In addition, an increased frequency of non–AIDSdefining malignancies was noted. The incidence of Hodgkin lymphoma did not increase significantly (Fig. 1). One child had a second malignant diagnosis so that there were 289 malignancies among the 288 children. The relative frequency of the numerically predominant cancers is presented in Table 1. Kaposi sarcoma, rarely seen in South African children before 1995, is the main contributor, with 97 cases, just over one third of the total cancers in this group. The B-cell lymphomas constitute much of the balance of the HIV-associated malignancies, with 61 Burkitt lymphomas, 45 non-Burkitt B-cell lymphomas, and 2 PCNSLs. Sixty-nine patients presented with non–AIDSdefining malignancies. Excluding the 16 cases of leukemia, and 33 cases where staging information was unavailable, the majority of children were diagnosed with advanced (stage III or IV) cancer (210 of 239 children or 87.9%). The HIV-associated malignancies, especially the B-cell lymphomas, demonstrated a propensity to present in unusual sites. Thirty-eight (35.8%) patients with B-cell lymphoma r
2013 Lippincott Williams & Wilkins
J Pediatr Hematol Oncol
Volume 36, Number 2, March 2014
Malignancy in HIV-positive South African Children
FIGURE 1. Annual number of cancers in children with HIV.
presented with primary tumors in the abdomen and pelvis but there were unusual sites such as the ampulla of Vater, the adrenal, and the periurethral area. Other primary sites included the jaw (7), paraspinal masses (8), the bone marrow (10), and lymph nodes (11), but there were 8 patients with mediastinal primaries (including 1 intracardiac tumor), 20 patients with disease in the face or sinuses (including 6 with orbital disease) and 4 patients with primary bone lesions. Among the patients with Kaposi sarcoma 71 (73.2%) had nodal involvement, 61 (62.9%) had skin lesions, 40 (41.2%) had oropharyngeal disease, 32 (33%) had lesions involving lung or pleura, and 21 (21.6%) had lesions involving abdominal viscera. Eleven of the 12 patients with Hodgkin lymphoma had nodal primaries; 1 patient had an appendiceal primary resected at presentation. The patients with non–AIDS-defining malignancies had tumors representing the full spectrum of pediatric malignancy (Table 1). Twenty-four (35%) had leukemia or lymphoma, there were 7 brain tumors (10%) and 25 embryonal tumors (36%). The median age of this group (2.93 y) was younger than that for the whole group. Of interest were several carcinomas occurring among older children. Apart from the nasopharyngeal carcinomas (8.56 and 12.78 y) there was 1 undifferentiated carcinoma (12.9 y), 1 adenocarcinoma of the rectum (16.85 y), and 1 squamous cell carcinoma of the anus (9.04 y). The latter may be related to human papilloma virus infection.26 The child who developed 2 malignancies was a 3-yearold boy who was diagnosed with orbital embryonal rhabdomyosarcoma and had nodular opacities on his chest radiograph and computed tomography scan consistent with lung metastases. He was commenced on an IRS-based regime without radiotherapy because he developed orbital infection, but chemotherapy was stopped at week 20 when the metastases failed to respond. Four months later he represented with a Kaposi sarcoma of his feet, legs, and scrotum. He was treated with 8 cycles of vincristine and bleomycin to which he had a good clinical response. He died at home 3 months later and the cause of death was never established. The absolute CD4 counts were not severely low (median: 415/mm3; range: 1 to 4594/mm3), nor were the HIV viral loads excessively raised in this patient population (median: 15,000 RNA copies/mL; range: 0 to 3,000,000 RNA copies/mL). CD4 percentage was lower than 15% in 111 of r
2013 Lippincott Williams & Wilkins
the 222 patients (50%) where this parameter was measured. The patients whose CD4 percentage was >15% were divided almost equally between a CD4 percentage of 15% to 20% (33%), 20% to 25% (34%), and >25% (32%). When we consider the degree of immunosuppression by diagnosis in terms of median absolute CD4 counts, those TABLE 1. Demographics and Diagnoses
N (%) Characteristics Age (y) Median Range (4 children over 15 y) Sex Males Females Male:female ratio HIV-associated malignancy (n = 220) Kaposi sarcoma Burkitt lymphoma Non-Burkitt B-cell lymphoma Diffuse large B-cell lymphoma Plasmablastic/body cavity B-cell lymphoma Other Hodgkin lymphoma Primary CNS lymphoma Leiomyosarcoma Non–AIDS-defining malignancy (n = 69) Acute lymphoblastic leukemia Acute myeloid leukemia T-cell non-Hodgkin lymphoma Anaplastic large cell lymphoma Brain and spinal tumors Nephroblastoma Other renal tumors Neuroblastoma Hepatocellular carcinoma Hepatoblastoma Rhabdomyosarcoma Retinoblastoma Melanotic progonoma Osteogenic sarcoma Nasopharyngeal carcinoma Other cancers Total cases
5.79 0.05-18.64 179 109 1.64:1 97 (33.6) 61 (21.1) 45 (15.6) 24 6 15 12 (4.2) 2 (0.6) 3 (1) 10 6 3 5 7 9 3 6 2 1 2 4 2 1 2 6
(3.5) (2.1) (1) (1.7) (2.4) (3.1) (1) (2.1) (0.6) (0.3) (0.6) (1.4) (0.6) (0.3) (0.6) (2.1) 289
CNS indicates central nervous system.
www.jpho-online.com |
113
J Pediatr Hematol Oncol
Davidson et al
with B-cell lymphoma (425/mm3) and Hodgkin lymphoma (377/mm3) were better off than those with Kaposi sarcoma (325/mm3). Patients with non–AIDS-defining malignancy were the least immunosuppressed with a median absolute CD4 count of 496/mm3, and a median HIV viral load of 6150 RNA copies/mL. Kaposi sarcoma patients whose CD4 percentage exceeded 15% (n = 35) had a mean age of 6.43 years and a median age of 5.34 years, and 17.1% had primary nodal disease. Those with a CD4 percentage below 15% (n = 49) had a mean age of 7.03 years and a median age of 7.23 years, and 18.4% had primary nodal disease. Only 23.3% (67 patients) were already on HAART at the time of cancer diagnosis. Another 41.3% (119 patients) commenced antiretrovirals after diagnosis. It is important to state that HAART only became available in public hospitals in 2004, and this is the reason at least 22.2% of the patients did not receive HAART.27 Of the 186 who received antiretrovirals, 14 (7.5%) discontinued HAART during the period of follow-up. A third of patients had coexisting pathology (100 children, 34.7%), mostly active tuberculosis that was seen in 82 children (28.5%). Treatment for tuberculosis was instituted in these children before starting chemotherapy. Due to the number of institutions and the number of different treatment protocols involved, the outcome of management was analyzed only in broad terms (Table 2). The overall disease-free survival was relatively low at 33.7%; survival according to diagnosis is shown in Figure 2. This was an estimated 5-year survival based on follow-up among the disease-free survivors varying from 6 to 118 months (with a mean of 36 mo and a median of 31 mo). The group of 197 children treated with curative intent (chemotherapy, surgery, and/or radiotherapy) had an overall survival of 47.9%, and 154 of these children, who also had access to HAART, had an overall survival of 57.8% (survival according to diagnosis is shown in Fig. 3). Notably the group of 26 children treated with HAART and definitive therapy for non–AIDS-defining malignancies had an overall survival of 73% (Fig. 3). Excluding those 39 patients who were lost to follow-up, 31 of whom are presumed dead, the analysis of causes of death presented in Table 2 shows that the majority (122 of 147 or 83%) of deaths were due to disease progression. The Kaplan-Meier curves indicate that approximately 30% of children died soon after diagnosis, which is in agreement with the high frequency of advanced disease at presentation. The mortality rate due to treatment toxicity was 10.2% (20 of 197) among those receiving curative therapy. Apart from 2 deaths due to tumor lysis syndrome, and one each from cardiomyopathy and druginduced hepatitis, all the other patients died from infection. There were only 2 PCNSLs. One patient was initially believed to have a high-grade glioma and was referred back to a shared care center to continue HAART and receive palliation as necessary. When he was found to be alive 1 year later, with complete radiologic resolution of the lesion, the pathology was reviewed and found to be a B-cell lymphoma. The second patient was commenced on HAART and treated with corticosteroids and radiotherapy. He was discharged free of disease after 57 months of follow-up.
Volume 36, Number 2, March 2014
TABLE 2. Cancer Stage and Comorbidity, Therapy, and Outcome
N (%) Malignancies by stage (239 cases) Stage I Stage II Stage III Stage IV Comorbidity On TB treatment or diagnosed with TB at the onset of malignant disease Interventions Antiretroviral therapy On HAART at presentation HAART started after diagnosis No HAART Unknown Chemotherapy + / radiotherapy (and receiving HAART) Burkitt lymphoma Other B-cell lymphoma Primary CNS lymphoma Kaposi sarcoma Hodgkin lymphoma Other Outcome Alive cancer free Alive with cancer Died due to disease progression (cancer) Died due to disease progression (AIDS) Died due to treatment toxicity, or treatment-related infection Died of unrelated causes Lost to follow-up: presumed dead Lost to follow-up: presumed alive Survival Overall survival with chemotherapy/surgery/ radiotherapy (CSR) Overall survival with CSR and HAART
11 18 120 90
(4.6) (7.5) (50.2) (37.7)
82 (28.5)
67 119 95 7 43 34 1 65 10 44 97 5 108 14 20
(23.3) (41.3) (33.0) (2.4) (36) (30) (1) (54) (7) (26) (33.7) (1.7) (37.5) (4.9) (6.9)
5 (1.7) 31 (10.8) 8 (2.8) 197 (47.9) 154 (57.8)
CNS indicates central nervous system; HAART, highly active antiretroviral therapy; TB, tuberculosis.
B-cell lymphoma) than in HIV-negative children and we noted a propensity for B-cell tumors to develop in unusual sites. Leiomyosarcoma and Hodgkin lymphoma were rare, even at the zenith of the epidemic in South Africa.
DISCUSSION The study shows more Kaposi sarcoma and fewer primary CNS lymphomas among HIV-positive children than is reported in the developed world. There was a higher incidence of non-Burkitt B-cell lymphoma (42.5% of all
114 | www.jpho-online.com
FIGURE 2. Malignancy in HIV-positive children: overall survival by diagnosis. CNS indicates central nervous system. r
2013 Lippincott Williams & Wilkins
J Pediatr Hematol Oncol
Volume 36, Number 2, March 2014
FIGURE 3. Malignancy in HIV-positive children: overall survival by diagnosis among those treated with HAART and chemotherapy. HAART indicates highly active antiretroviral therapy.
Surprisingly, the patients were not severely immunosuppressed on the basis of absolute CD4 counts and HIV viral load, with lymphoma patients being less immunosuppressed than those with Kaposi sarcoma. Nonetheless we expect the results for these HIV-related malignancies to improve with universal access to HAART and earlier diagnosis. The fact that Kaposi sarcoma was the single most common malignancy in this study (accounting for 34% of all tumors) is consistent with other reports from Africa28–30 and is undoubtedly related to the high rate of HHV-8 infection. We would hope for a decline in new cases as HAART continues to be rolled out. In regions of Africa where Burkitt lymphoma is endemic in children, such as Malawi, the tumor does not seem to occur more frequently in HIV-infected children.31 In South Africa, however, where this malignancy was less common before the AIDS epidemic,32 there have been a large number of cases among these children in the HIV-AIDS era. We report a large number of non–AIDS-defining malignancies. The spectrum of disease is very similar to that seen in the HIV-negative population and these patients had relatively well preserved immune systems and low HIV viral loads. We believe these tumors are primarily due to the high prevalence of HIV-AIDS in South Africa and expect to see more over time as the number of survivors established on HAART grows. Our results for these non–AIDS-defining tumors in the HAART era are already very encouraging. To explain the excess of Kaposi sarcoma, Burkitt lymphoma, Hodgkin lymphoma, and leiomyosarcoma observed in children with HIV, it is postulated that the immunologic and cellular changes due to HIV activity pave the way for infection with other viruses and enhance their oncogenic potential.33 Although Kaposi sarcoma is associated with HHV-8, the other 3 malignancies often demonstrate positivity for Epstein-Barr virus antigen in their histological samples.34,35 Such synergism between viral effects may well exist, but other factors may modulate it, as no significant increase in PCNSL, Hodgkin lymphoma, or leiomyosarcoma was observed in our series. In a 2007 meta-analysis, Grulich et al26 compared cancer incidence among adult HIV-AIDS patients with the incidence of cancers among transplant recipients. There was a significantly increased incidence in both populations in 20 out of 28 types of cancer, and most of these were cancers r
2013 Lippincott Williams & Wilkins
Malignancy in HIV-positive South African Children
with a known infectious cause such as Kaposi sarcoma, non-Hodgkin and Hodgkin lymphoma, cervical cancer, liver cancer, and stomach cancer. Rates of most of the common epithelial cancers were not increased. This provides strong evidence that it is immune deficiency that is responsible for the increased risk of cancer rather than other risk factors. It seems that the range of infectionrelated cancers is much wider than previously appreciated, and that infection-related cancer will become an increasingly important complication of long-term HIV infection. Published trends in cancer risk among people with AIDS provide further support for this assertion. There was a drastic decline in the incidence of Kaposi sarcoma and non-Hodgkin lymphoma in the United States between 1990 to 1995 and 1996 to 200236 with the introduction of HAART, but adults with AIDS remained at increased risk compared with the general population. There was no change in the risk of cervical cancer and there was an increase in the incidence of Hodgkin lymphoma. Similarly the Italian pediatric registry data5 show a decline in cancer rates since the introduction of HAART, but the incidence of cancer still remains 5 times higher than that for HIVuninfected children. Thus, we conclude that the degree of immune suppression as measured by the CD4 count is not the sole risk factor for cancer. Most African children infected with HIV acquire the virus perinatally, with a slight female predominance.37 In this series of malignancies, similar to other studies,8,9,38 a male predominance of 1.64:1 was observed. This could be ascribed either to the slightly higher tendency for the male sex to develop malignancy or to the characteristic male predominance in Kaposi sarcoma and Burkitt lymphoma. Over 80% of the 239 cancers where staging information was available had advanced disease (stages III or IV) that had a negative impact on eventual survival. The factors associated with delay in diagnosis of cancers in children have been shown to be complex, including the site of the cancer, the age of the child, family factors, and health system factors.39,40 Current recommendations should include regular clinical evaluation of HIV-positive children paying particular attention to the warning signs of cancer.41 Such regular examinations should take place even before initiation of HAART; in our study half of the subjects developed a malignancy while their CD4 counts indicated only mild to moderate immune deficiency. What adds to the difficulty is that these children frequently present with disease in unusual sites42 and at the same time antiretroviral therapy–naive children may have gross lymphoproliferation in the absence of malignancy. For this reason it is important to develop a framework and clinical guidelines to foster early detection.43 Active tuberculosis was found in 28.5% of the children at diagnosis of their cancer; and the treatment for tuberculosis was instituted before starting chemotherapy. Although the possibility of adverse effects resulting from the interaction of tuberculosis drugs with chemotherapy and antiretrovirals was a concern, there were no documented cases in this series. More than 10% of the 197 children treated with chemotherapy and/or radiotherapy died from toxicity. This percentage is high when compared with published data from well-resourced countries, where the risk of death attributed to infections in HIV-negative children on chemotherapy is usually in the order of 1%,44 but low when compared with poorly resourced countries where figures of www.jpho-online.com |
115
J Pediatr Hematol Oncol
Davidson et al
14% have been reported in HIV-negative children.45 Such findings underscore the risks of administering cytotoxic medication to patients who are already immune compromised. Notwithstanding concerns about toxicity most children died as a result of disease progression. The overall disease-free survival rate was low at 33.7%, due mainly to late presentation, but results were better for those treated with chemotherapy and HAART who achieved a 57.8% overall survival rate. The best results were seen in patients with Hodgkin lymphoma who had an estimated 5-year survival rate of 71.4%. REFERENCES 1. Stack M, Walsh PM, Comber H, et al. Childhood cancer in Ireland: a population-based study. Arch Dis Childhood. 2007; 92:890–897. 2. Ries LAG, Melbert D, Krapcho M, et al. SEER Cancer Statistics Review, 1975-2004. Bethesda, MD: National Cancer Institute . Available at: http://seer.cancer.gov/csr/1975_2004/. Accessed December 9, 2008. 3. Centers for Disease Control and Prevention. US HIV and AIDS cases reported through December 1996. HIV/AIDS Surveill Rep. 1996;8:1–39. 4. Biggar RJ, Frisch M, Goedert JJ. Risk of cancer in children with AIDS. JAMA. 2000;284:205–209. 5. Chiappini E, Galli L, Tovo P-A, et al. Cancer rates after year 2000 significantly decrease in children with perinatal HIV infection: A study by the Italian register for HIV infection in children. J Clin Oncol. 2007;25:97–101. 6. Alvaro-Meca A, Micheloud D, Jensen J, et al. Epidemiologic trends of cancer diagnoses among HIV-infected children in Spain from 1997 to 2008. Pediatr Infect Dis J. 2011;30:764–768. 7. Lanzkowsky P. Manual of Pediatric Hematology-Oncology. London: Academic Press; 2011:101. 8. Ziegler J, Katongole-Mbidde E. Kaposi’s sarcoma in childhood: an analysis of 100 cases from Uganda and relationship to HIV infection. Int J Cancer. 1996;65:200–203. 9. Newton R, Ziegler J, Beral V, et al. Uganda Kaposi’s Sarcoma Study Group. A case-control study of human immunodeficiency virus infection and cancer in adults and children residing in Kampala, Uganda. Int J Cancer. 2001;92:622–627. 10. Cairncross L, Davidson A, Millar AJW, et al. Kaposi sarcoma in children with HIV: a clinical series from Red Cross Children’s Hospital. J Ped Surg. 2009;44:373–376. 11. Mueller BU. Cancers in children infected with the human immunodeficiency virus. Oncologist. 1999;4:309–317. 12. McClain KL, Leach CT, Jenson HB, et al. Association of Epstein-Barr virus with leiomyosarcomas in children with AIDS. N Engl J Med. 1995;332:12–18. 13. Chang Y, Cesarman E, Pessin MS, et al. Identification of herpesvirus-like DNA sequences in AIDS-associated Kaposi’s sarcoma. Science. 1994;266:1865–1869. 14. Flint SJ, Enquist LW, Racaniello VR, et al. Infection of a Susceptible Host. In: Flint SJ, Enquist LW, Racaniello VR, Skalka AM, eds. Principles of Virology, Volume II, Pathogenesis and Control. 3rd ed. Washington: ASM Press; 2009:3–27. 15. Chitsike I, Siziya S. Seroprevalence of human immunodeficiency virus type 1 infection in childhood malignancy in Zimbabwe. Cent Afr J Med. 1998;44:242–245. 16. Sinfield RL, Molyneux EM, Banda K, et al. Spectrum and presentation of paediatric malignancies in the HIV era: experience from Blantyre, Malawi, 1998-2003. Pediatr Blood Cancer. 2007;48:515–520. 17. Davidson A, Hendricks M, Geel J, et al. Malignancy in HIVpositive South African children. Pediatr Blood Cancer. 2009; 53:719. 18. WHO, UNAIDS, UNICEFGlobal HIV/AIDS response: update 2011. Epidemic update and health sector progress towards Universal Access. Available at: http://
116 | www.jpho-online.com
19.
20.
21.
22.
23. 24. 25. 26.
27.
28. 29. 30. 31.
32. 33. 34. 35. 36. 37. 38. 39. 40.
Volume 36, Number 2, March 2014
whqlibdoc.who.int/publications/2011/9789241502986_eng.pdf. Accessed May 27, 2012. Granovsky MO, Mueller BU, Nicholson HS, et al. Cancer in human immunodeficiency virus-infected children: a case series from the Children’s Cancer Group and the National Cancer Institute. J Clin Oncol. 1998;16:1729–1735. Caselli D, Klersy C, de Martino M, et al. Human immunodeficiency virus-related cancer in children: incidence and treatment outcome—report of the Italian register. J Clin Oncol. 2000;18:3854–3861. Statistics South Africa. Statistical release P0302: mid-year population estimates 2010. Available at: http://www.statssa. gov.za/publications/P0302/P03022010.pdf. Accessed May 27, 2012. Horner MJ, Ries LAG, Krapcho M, et al. SEER Cancer Statistics Review, 1975-2006. Bethesda, MD: National Cancer Institute. Available at: http://seer.cancer.gov/csr/1975_2006/. Accessed January 17, 2010. Kest H, Brogly S, McSherry G, et al. Malignancy in perinatally human immunodeficiency virus-infected children in the United States. Pediatr Infect Dis J. 2005;24:237–242. Chanock SJ, Pizzo PA. Infection prevention strategies for children with cancer and AIDS: contrasting dilemmas. J Hosp Infec. 1995;30(suppl):197–208. Bower M, McCall-Peat N, Ryan N, et al. Protease inhibitors potentiate chemotherapy-induced neutropenia. Blood. 2004; 104:2943–2946. Grulich AE, van Leeuwen MT, Falster MO, et al. Incidence of cancers in people with HIV/AIDS compared with immunosuppressed transplant recipients: a meta-analysis. Lancet. 2007; 370:59–67. Department of Health, South AfricaOperational plan for comprehensive HIV and AIDS care, management and treatment for South Africa. Available at: http://www.hst.org.za/ uploads/files/aidsplan.pdf. Accessed January 27, 2012. Stefan DC, Stones DK, Wainwright L, et al. Kaposi sarcoma in South African children. Pediatr Blood Cancer. 2011;56: 392–396. Davidson A. Kaposi sarcoma: the African HIV epidemic’s partner in crime. Pediatr Blood Cancer. 2010;54:657–658. Gantt S, Kakuru A, Wald A, et al. Clinical presentation and outcome of epidemic Kaposi sarcoma in Ugandan children. Pediatr Blood Cancer. 2010;54:670–674. Mutalima N, Molyneux EM, Johnston WT, et al. Impact of infection with human immunodeficiency virus-1 (HIV) on the risk of cancer among children in Malawi—preliminary findings. Infect Agent Cancer. 2010;5:5. Hesseling P, Wood RE, Nortje´ CJ, et al. African Burkitt lymphoma in the Cape province of South Africa and in Namibia. Oral Surg Oral Med Oral Pathol. 1989;68:162–166. Angeletti PC, Zhang L, Wood C. The viral etiology of AIDSassociated malignancies. Adv Pharmacol. 2008;56:509–557. Kutok JL, Wang F. Spectrum of Epstein-Barr virus-associated diseases. Annu Rev Pathol. 2006;1:375–404. Bajaj BG, Murakami M, Robertson ES. Molecular biology of EBV in relationship to AIDS-associated oncogenesis. Cancer Treat Res. 2007;133:141–162. Engels EA, Pfeiffer RM, Goedert JJ, et al. Trends in cancer risk among people with AIDS in the United States 1980-2002. AIDS. 2006;20:1645–1654. Biggar RJ, Taha TE, Hoover DR, et al. Higher in utero and perinatal HIV infection risk in girls than boys. J Acquir Immune Defic Syndr. 2006;41:509–513. Amir H, Kaaya EE, Manji KP, et al. Kaposi’s sarcoma before and during a human immunodeficiency virus epidemic in Tanzanian children. Pediatr Infect Dis J. 2001;20:518–521. Stefan DC, Siemonsma F. Delay and causes of delay in the diagnosis of childhood cancer in Africa. Pediatr Blood Cancer. 2011;56:80–85. Haimi M, Peretz Nahum M, Ben Arush MW. Delay in diagnosis of children with cancer: a retrospective study of 315 children. Pediatr Hematol Oncol. 2004;21:37–48. r
2013 Lippincott Williams & Wilkins
J Pediatr Hematol Oncol
Volume 36, Number 2, March 2014
41. Stones DK. Sub-Saharan African children and cancer: how do we improve the outcome? Afr J Cancer. 2011;3: 91–93. 42. Mueller BU. Cancers in human immunodeficiency virusinfected children. Monogr Natl Cancer Inst. 1998;23:31–35. 43. Davidson A, Eley B. HIV and childhood cancer. Cont Med Ed J. 2010;28:337–342.
r
2013 Lippincott Williams & Wilkins
Malignancy in HIV-positive South African Children
44. Lund B, A˚sberg A, Heyman M, et al. for the Nordic Society of Paediatric Haematology and Oncology. Risk factors for treatment related mortality in childhood acute lymphoblastic leukemia. Pediatr Blood Cancer. 2011;56:551–559. 45. Gupta S, Bonilla M, Gamero M, et al. Microbiology and mortality of pediatric febrile neutropenia in El Salvador. J Pediatr Hematol Oncol. 2011;33:276–280.
www.jpho-online.com |
117
Malignancies in South African Children With HIV Alan Davidson, MD, MPhil,* Rosalinda D. Wainwright, MD,w David K. Stones, MD,z Mariana Kruger, MD, PhD,y Marc Hendricks, MD,* Jennifer Geel, MD,8 Janet Poole, MD,8 David Reynders, MD,z Fareed Omar, MD,z Rema Mathew, MD,# and D. Cristina Stefan, MD, PhDy
Objectives: In 2008 the South African Children’s Cancer Study Group decided to review the epidemiology, management, and chemotherapy response of HIV-positive children with malignancy. Methods: This is a retrospective analysis of data collected from the records of HIV-positive children diagnosed with malignancy at 7 university-based pediatric oncology units serving 8 of the 9 provinces in South Africa. Results: Two hundred eighty-eight HIV-positive children were diagnosed with 289 malignancies between 1995 and 2009. Age at diagnosis ranged from 17 days to 18.64 years; median 5.79 years. Of the 220 HIV-associated malignancies, there were 97 Kaposi sarcomas, 61 Burkitt lymphomas, 47 other B-cell lymphomas including 2 primary central nervous system lymphomas, 12 Hodgkin lymphomas, and 3 leiomyosarcomas. Sixty-nine patients presented with non–AIDS-defining malignancies. More than 80% presented with advanced disease. Most patients (76.7%) were naive to antiretroviral therapy; 22.2% did not have access because it only became available in public hospitals in 2004. One hundred ninety-seven children were treated with curative intent; 91 received palliative care due to advanced malignancy and/or advanced HIV disease. Nearly one third had coexisting pathology, mostly tuberculosis. Overall survival for the whole group was 33.7%, but was 57.8% for those treated with antiretroviral therapy and chemotherapy. Conclusions: The study shows more Kaposi sarcoma and fewer primary central nervous system lymphomas among HIV-positive children than that is reported in the developed world, but confirms a higher incidence of non-Burkitt B-cell lymphoma than in HIVnegative children. The high number of non–AIDS-defining malignancies underscores the prevalence of HIV-AIDS in South Africa. The overall survival should improve with universal access to antiretrovirals and earlier diagnosis. Key Words: HIV, infectious disease, pediatric oncology, Africa
(J Pediatr Hematol Oncol 2014;36:111–117)
Received for publication December 2, 2012; accepted April 22, 2013. From the *Department of Paediatrics and Child Health, Red Cross Children’s Hospital and the University of Cape Town; yDepartment of Paediatrics and Child Health, Tygerberg Hospital and Stellenbosch University, Tygerberg, Cape Town, Western Cape; wDepartment of Paediatrics and Child Health, Chris Hani Baragwanath Academic Hospital; 8Department of Paediatrics and Child Health, Charlotte Maxeke Johannesburg Academic Hospital, University of the Witwatersrand, Johannesburg; zDepartment of Paediatrics, Steve Biko Academic Hospital, University of Pretoria, Pretoria, Gauteng; zDepartment of Paediatrics and Child Health, Universitas Academic Hospital Complex and the University of the Free State, Bloemfontein, Free State; and #Department of Paediatrics, Frere Hospital and Walter Sisulu University, East London, Eastern Cape. The authors declare no conflict of interest. Reprints: Alan Davidson, MD, MPhil, Department of Paediatrics and Child Health, Ward G1, Red Cross Children’s Hospital, Private Bag, Rondebosch, Cape Town 7700, South Africa (e-mail: [email protected]). Copyright r 2013 by Lippincott Williams & Wilkins
J Pediatr Hematol Oncol
T
he advent of the HIV epidemic has brought about new challenges for pediatric oncologists. Although the incidence of cancer in HIV-negative children is estimated to be 0.13 to 0.15 per 1000,1,2 early reports from the United States found an incidence of 20/1000 in HIV-infected children.3,4 Similarly elevated figures were subsequently reported by European investigators: in Italy, before the widespread introduction of highly active antiretroviral therapy (HAART), the cancer rate in HIV-infected children was 4.49/10005; in Spain, an audit of diagnoses on discharge of all children admitted to hospitals found a malignant disease rate of 9.4/1000 for HIV-positive children, whereas HIV-negative children had a rate of 1.2/1000.6 This increase was due to an increased incidence of non-Hodgkin lymphomas, leiomyomas and leiomyosarcomas as well as Hodgkin lymphomas, acute leukemias, and other less commonly encountered tumors.7 The incidence of Kaposi sarcoma has also increased dramatically in children with AIDS.8–10 However, this increase is almost solely limited to Africa. Elsewhere it is seen mostly in adults living with HIV infection.11 Certain malignancies encountered more often in children with AIDS are associated with viral infections: the Epstein-Barr virus is very frequently identified in Burkitt lymphoma, mixed cellularity Hodgkin lymphoma, and leiomyosarcoma,12 whereas the human herpes virus 8 (HHV-8) is considered to be the etiologic agent of Kaposi sarcoma.13 The higher incidence of viral-related cancers in children with AIDS suggests an association between HIV and other viruses.14 The rising prevalence of HIV has also resulted in HIV-positive children presenting with non–AIDS-defining malignancies such as Wilms tumor and acute lymphoblastic leukemia.9,15–17 We believe that this is a function of both an increase in the number of children with HIV and increased survival of this cohort as the number of children on HAART expands.18 HIV seems to have affected not only the incidence but also the presentation of these malignancies. These patients typically present with unusual histological subtypes in unusual sites. Diffuse large B-cell lymphomas and primary central nervous system lymphomas (PCNSLs) make up a substantial portion of the HIV-related malignancies reported in studies from the United States3,4,19 and Italy,20 which is in strong contrast with the spectrum of disease seen in immunocompetent children. Even non–AIDS-defining malignancies may present atypically in immunocompromised children. For Africa, the higher incidence rates of malignant disease in children with HIV are augmented by the huge numbers of those infected. Out of a total of 3,400,000 HIV-positive children younger than 15 years of age in 2010, 3,100,000 (91.2%) were resident in sub-Saharan Africa.18 Although comprehensive prevention and treatment
Volume 36, Number 2, March 2014
www.jpho-online.com |
111
Davidson et al
strategies have reduced and stabilized the prevalence of HIV in Europe and North America, only 21% of children in subSaharan Africa who need to undergo HAART are receiving treatment.18 In South Africa, according to the mid-2009 population statistics, there were 15,500,700 children aged 0 to 15 years.21 At the end of 2010, 108,682 children were reported to be on HAART in South Africa, out of an estimated 300,000 (270,000 to 340,000) children requiring immediate HAART, an estimated coverage of 36%.18 Using the published European estimates, South Africa should expect at least 1000 new cases of HIV-related malignancies in children per year,1,5,22 but this figure is probably overstated in our context because far fewer children in the Italian cohort would have succumbed to gastroenteritis and pneumonia; leaving proportionally more survivors vulnerable to AIDS-related lymphomas. The widespread introduction of HAART has reduced the number of HIV-related malignancies but even with HAART, the overall rate of cancer in these children remains higher (0.76/1000 in the Italian study) than in those not infected by the virus.5 Moreover, the decrease in cancers related to HAART is significant only after 2 years of administration.23 It seems that although HAART has increased the life expectancy of children with HIV, it has not reduced the burden of malignant disease associated with the virus. Although a number of cases of Kaposi sarcoma might improve on HAART alone, most children eventually require chemotherapy.10 For other malignancies occurring in HIV-positive children many oncologists now recommend full-dose chemotherapy and HAART. It should be borne in mind that the interactions between HAART and standard chemotherapy regimens are insufficiently studied and potentially dangerous.24,25 There are still insufficient data to support the growing body of opinion that the outcomes that can be achieved by combining the 2 groups of drugs are similar to those in HIV-negative children. Treatment is made even more challenging by multiple comorbidities such as chronic lung disease, tuberculosis, and other infections. These have multiple effects, impacting on treatment decision-making and clinical outcomes by way of drug-drug interactions and non–treatment-related morbidity. For more than 15 years, South African pediatric oncologists have treated HIV-positive children with malignancy. Many of them have achieved remission and have been long-term survivors on HAART.
MATERIALS AND METHODS In 2008 members of the South African Children’s Cancer Study Group, representing the 7 university-based pediatric oncology units (POUs) and several satellite centers, decided to audit this cohort of HIV-positive children with malignancy in the hope that it would provide meaningful insights, and support future collaborative prospective clinical trials. Six of the 7 POUs in South Africa participated in addition to 1 satellite center (which is also now a university-based POU). The participating institutions include referral centers that serve 8 of the 9 provinces. All authors participate in the activities of the South African Children’s Cancer Study Group that holds annual meetings, collects registry data, and sponsors joint protocols. Since all but one university-based POU in South Africa participated, our data are representative but the cases reported here do not represent all pediatric cases of HIV and malignancy diagnosed during this period.
112 | www.jpho-online.com
J Pediatr Hematol Oncol
Volume 36, Number 2, March 2014
The aim was to audit malignancies occurring in HIVpositive South African children in order to document patient demographics, relative frequency, stage at diagnosis, degree of immunosuppression (WHO criteria), comorbidities and their impact on survival, and outcome correlated with the use of chemotherapy, combined with radiotherapy or surgery as indicated, and HAART, alone or in combination. Data were collected from the records of 288 HIV-positive children diagnosed with malignancy between 1995 and 2009, and entered into a Microsoft Access database. Institutions assigned a numerical identifier to each case in order to blind the principal investigator who managed the data. Uniformity of histological evaluation was achieved by having most samples processed by pathologists of the National Health Laboratory Service, where similar laboratory protocols and assessment criteria are applied nationwide. Treatment outcomes were analyzed using Statistica software; survival analysis was performed using KaplanMeier curves. A P value of r0.05 was regarded as significant. Confidentiality for individual patients was assured as the data provided to the principal investigator were anonymized. As this is a retrospective study, we did not seek informed consent from individual patients, but obtained the consent of the custodian of the data to access patient files. Approval was obtained from the relevant Human Research Ethics Committees of the participating centers.
RESULTS Between January 1995 and December 2009, 288 children with HIV infection and cancer were admitted to the pediatric hematology-oncology units participating in the study. The median age was 5.79 years, with a range of 17 days to 18.64 years, and there was a male to female ratio of 1.64:1. Only 4 of the patients in this series were over the age of 15 at diagnosis (2 children with diffuse large B-cell lymphomas, 1 with Kaposi sarcoma, and 1 with adenocarcinoma of the rectum). All these children were being followed up by pediatric infectious disease services. Four children were of ethnically mixed race and the remaining 284 were black South Africans. The annual number of cases increased slowly from 1995, following the gradual spread of the epidemic (Fig. 1). Kaposi sarcoma registered the highest increase in annual numbers, followed by Burkitt lymphoma and non-Burkitt B-cell lymphoma. In addition, an increased frequency of non–AIDSdefining malignancies was noted. The incidence of Hodgkin lymphoma did not increase significantly (Fig. 1). One child had a second malignant diagnosis so that there were 289 malignancies among the 288 children. The relative frequency of the numerically predominant cancers is presented in Table 1. Kaposi sarcoma, rarely seen in South African children before 1995, is the main contributor, with 97 cases, just over one third of the total cancers in this group. The B-cell lymphomas constitute much of the balance of the HIV-associated malignancies, with 61 Burkitt lymphomas, 45 non-Burkitt B-cell lymphomas, and 2 PCNSLs. Sixty-nine patients presented with non–AIDSdefining malignancies. Excluding the 16 cases of leukemia, and 33 cases where staging information was unavailable, the majority of children were diagnosed with advanced (stage III or IV) cancer (210 of 239 children or 87.9%). The HIV-associated malignancies, especially the B-cell lymphomas, demonstrated a propensity to present in unusual sites. Thirty-eight (35.8%) patients with B-cell lymphoma r
2013 Lippincott Williams & Wilkins
J Pediatr Hematol Oncol
Volume 36, Number 2, March 2014
Malignancy in HIV-positive South African Children
FIGURE 1. Annual number of cancers in children with HIV.
presented with primary tumors in the abdomen and pelvis but there were unusual sites such as the ampulla of Vater, the adrenal, and the periurethral area. Other primary sites included the jaw (7), paraspinal masses (8), the bone marrow (10), and lymph nodes (11), but there were 8 patients with mediastinal primaries (including 1 intracardiac tumor), 20 patients with disease in the face or sinuses (including 6 with orbital disease) and 4 patients with primary bone lesions. Among the patients with Kaposi sarcoma 71 (73.2%) had nodal involvement, 61 (62.9%) had skin lesions, 40 (41.2%) had oropharyngeal disease, 32 (33%) had lesions involving lung or pleura, and 21 (21.6%) had lesions involving abdominal viscera. Eleven of the 12 patients with Hodgkin lymphoma had nodal primaries; 1 patient had an appendiceal primary resected at presentation. The patients with non–AIDS-defining malignancies had tumors representing the full spectrum of pediatric malignancy (Table 1). Twenty-four (35%) had leukemia or lymphoma, there were 7 brain tumors (10%) and 25 embryonal tumors (36%). The median age of this group (2.93 y) was younger than that for the whole group. Of interest were several carcinomas occurring among older children. Apart from the nasopharyngeal carcinomas (8.56 and 12.78 y) there was 1 undifferentiated carcinoma (12.9 y), 1 adenocarcinoma of the rectum (16.85 y), and 1 squamous cell carcinoma of the anus (9.04 y). The latter may be related to human papilloma virus infection.26 The child who developed 2 malignancies was a 3-yearold boy who was diagnosed with orbital embryonal rhabdomyosarcoma and had nodular opacities on his chest radiograph and computed tomography scan consistent with lung metastases. He was commenced on an IRS-based regime without radiotherapy because he developed orbital infection, but chemotherapy was stopped at week 20 when the metastases failed to respond. Four months later he represented with a Kaposi sarcoma of his feet, legs, and scrotum. He was treated with 8 cycles of vincristine and bleomycin to which he had a good clinical response. He died at home 3 months later and the cause of death was never established. The absolute CD4 counts were not severely low (median: 415/mm3; range: 1 to 4594/mm3), nor were the HIV viral loads excessively raised in this patient population (median: 15,000 RNA copies/mL; range: 0 to 3,000,000 RNA copies/mL). CD4 percentage was lower than 15% in 111 of r
2013 Lippincott Williams & Wilkins
the 222 patients (50%) where this parameter was measured. The patients whose CD4 percentage was >15% were divided almost equally between a CD4 percentage of 15% to 20% (33%), 20% to 25% (34%), and >25% (32%). When we consider the degree of immunosuppression by diagnosis in terms of median absolute CD4 counts, those TABLE 1. Demographics and Diagnoses
N (%) Characteristics Age (y) Median Range (4 children over 15 y) Sex Males Females Male:female ratio HIV-associated malignancy (n = 220) Kaposi sarcoma Burkitt lymphoma Non-Burkitt B-cell lymphoma Diffuse large B-cell lymphoma Plasmablastic/body cavity B-cell lymphoma Other Hodgkin lymphoma Primary CNS lymphoma Leiomyosarcoma Non–AIDS-defining malignancy (n = 69) Acute lymphoblastic leukemia Acute myeloid leukemia T-cell non-Hodgkin lymphoma Anaplastic large cell lymphoma Brain and spinal tumors Nephroblastoma Other renal tumors Neuroblastoma Hepatocellular carcinoma Hepatoblastoma Rhabdomyosarcoma Retinoblastoma Melanotic progonoma Osteogenic sarcoma Nasopharyngeal carcinoma Other cancers Total cases
5.79 0.05-18.64 179 109 1.64:1 97 (33.6) 61 (21.1) 45 (15.6) 24 6 15 12 (4.2) 2 (0.6) 3 (1) 10 6 3 5 7 9 3 6 2 1 2 4 2 1 2 6
(3.5) (2.1) (1) (1.7) (2.4) (3.1) (1) (2.1) (0.6) (0.3) (0.6) (1.4) (0.6) (0.3) (0.6) (2.1) 289
CNS indicates central nervous system.
www.jpho-online.com |
113
J Pediatr Hematol Oncol
Davidson et al
with B-cell lymphoma (425/mm3) and Hodgkin lymphoma (377/mm3) were better off than those with Kaposi sarcoma (325/mm3). Patients with non–AIDS-defining malignancy were the least immunosuppressed with a median absolute CD4 count of 496/mm3, and a median HIV viral load of 6150 RNA copies/mL. Kaposi sarcoma patients whose CD4 percentage exceeded 15% (n = 35) had a mean age of 6.43 years and a median age of 5.34 years, and 17.1% had primary nodal disease. Those with a CD4 percentage below 15% (n = 49) had a mean age of 7.03 years and a median age of 7.23 years, and 18.4% had primary nodal disease. Only 23.3% (67 patients) were already on HAART at the time of cancer diagnosis. Another 41.3% (119 patients) commenced antiretrovirals after diagnosis. It is important to state that HAART only became available in public hospitals in 2004, and this is the reason at least 22.2% of the patients did not receive HAART.27 Of the 186 who received antiretrovirals, 14 (7.5%) discontinued HAART during the period of follow-up. A third of patients had coexisting pathology (100 children, 34.7%), mostly active tuberculosis that was seen in 82 children (28.5%). Treatment for tuberculosis was instituted in these children before starting chemotherapy. Due to the number of institutions and the number of different treatment protocols involved, the outcome of management was analyzed only in broad terms (Table 2). The overall disease-free survival was relatively low at 33.7%; survival according to diagnosis is shown in Figure 2. This was an estimated 5-year survival based on follow-up among the disease-free survivors varying from 6 to 118 months (with a mean of 36 mo and a median of 31 mo). The group of 197 children treated with curative intent (chemotherapy, surgery, and/or radiotherapy) had an overall survival of 47.9%, and 154 of these children, who also had access to HAART, had an overall survival of 57.8% (survival according to diagnosis is shown in Fig. 3). Notably the group of 26 children treated with HAART and definitive therapy for non–AIDS-defining malignancies had an overall survival of 73% (Fig. 3). Excluding those 39 patients who were lost to follow-up, 31 of whom are presumed dead, the analysis of causes of death presented in Table 2 shows that the majority (122 of 147 or 83%) of deaths were due to disease progression. The Kaplan-Meier curves indicate that approximately 30% of children died soon after diagnosis, which is in agreement with the high frequency of advanced disease at presentation. The mortality rate due to treatment toxicity was 10.2% (20 of 197) among those receiving curative therapy. Apart from 2 deaths due to tumor lysis syndrome, and one each from cardiomyopathy and druginduced hepatitis, all the other patients died from infection. There were only 2 PCNSLs. One patient was initially believed to have a high-grade glioma and was referred back to a shared care center to continue HAART and receive palliation as necessary. When he was found to be alive 1 year later, with complete radiologic resolution of the lesion, the pathology was reviewed and found to be a B-cell lymphoma. The second patient was commenced on HAART and treated with corticosteroids and radiotherapy. He was discharged free of disease after 57 months of follow-up.
Volume 36, Number 2, March 2014
TABLE 2. Cancer Stage and Comorbidity, Therapy, and Outcome
N (%) Malignancies by stage (239 cases) Stage I Stage II Stage III Stage IV Comorbidity On TB treatment or diagnosed with TB at the onset of malignant disease Interventions Antiretroviral therapy On HAART at presentation HAART started after diagnosis No HAART Unknown Chemotherapy + / radiotherapy (and receiving HAART) Burkitt lymphoma Other B-cell lymphoma Primary CNS lymphoma Kaposi sarcoma Hodgkin lymphoma Other Outcome Alive cancer free Alive with cancer Died due to disease progression (cancer) Died due to disease progression (AIDS) Died due to treatment toxicity, or treatment-related infection Died of unrelated causes Lost to follow-up: presumed dead Lost to follow-up: presumed alive Survival Overall survival with chemotherapy/surgery/ radiotherapy (CSR) Overall survival with CSR and HAART
11 18 120 90
(4.6) (7.5) (50.2) (37.7)
82 (28.5)
67 119 95 7 43 34 1 65 10 44 97 5 108 14 20
(23.3) (41.3) (33.0) (2.4) (36) (30) (1) (54) (7) (26) (33.7) (1.7) (37.5) (4.9) (6.9)
5 (1.7) 31 (10.8) 8 (2.8) 197 (47.9) 154 (57.8)
CNS indicates central nervous system; HAART, highly active antiretroviral therapy; TB, tuberculosis.
B-cell lymphoma) than in HIV-negative children and we noted a propensity for B-cell tumors to develop in unusual sites. Leiomyosarcoma and Hodgkin lymphoma were rare, even at the zenith of the epidemic in South Africa.
DISCUSSION The study shows more Kaposi sarcoma and fewer primary CNS lymphomas among HIV-positive children than is reported in the developed world. There was a higher incidence of non-Burkitt B-cell lymphoma (42.5% of all
114 | www.jpho-online.com
FIGURE 2. Malignancy in HIV-positive children: overall survival by diagnosis. CNS indicates central nervous system. r
2013 Lippincott Williams & Wilkins
J Pediatr Hematol Oncol
Volume 36, Number 2, March 2014
FIGURE 3. Malignancy in HIV-positive children: overall survival by diagnosis among those treated with HAART and chemotherapy. HAART indicates highly active antiretroviral therapy.
Surprisingly, the patients were not severely immunosuppressed on the basis of absolute CD4 counts and HIV viral load, with lymphoma patients being less immunosuppressed than those with Kaposi sarcoma. Nonetheless we expect the results for these HIV-related malignancies to improve with universal access to HAART and earlier diagnosis. The fact that Kaposi sarcoma was the single most common malignancy in this study (accounting for 34% of all tumors) is consistent with other reports from Africa28–30 and is undoubtedly related to the high rate of HHV-8 infection. We would hope for a decline in new cases as HAART continues to be rolled out. In regions of Africa where Burkitt lymphoma is endemic in children, such as Malawi, the tumor does not seem to occur more frequently in HIV-infected children.31 In South Africa, however, where this malignancy was less common before the AIDS epidemic,32 there have been a large number of cases among these children in the HIV-AIDS era. We report a large number of non–AIDS-defining malignancies. The spectrum of disease is very similar to that seen in the HIV-negative population and these patients had relatively well preserved immune systems and low HIV viral loads. We believe these tumors are primarily due to the high prevalence of HIV-AIDS in South Africa and expect to see more over time as the number of survivors established on HAART grows. Our results for these non–AIDS-defining tumors in the HAART era are already very encouraging. To explain the excess of Kaposi sarcoma, Burkitt lymphoma, Hodgkin lymphoma, and leiomyosarcoma observed in children with HIV, it is postulated that the immunologic and cellular changes due to HIV activity pave the way for infection with other viruses and enhance their oncogenic potential.33 Although Kaposi sarcoma is associated with HHV-8, the other 3 malignancies often demonstrate positivity for Epstein-Barr virus antigen in their histological samples.34,35 Such synergism between viral effects may well exist, but other factors may modulate it, as no significant increase in PCNSL, Hodgkin lymphoma, or leiomyosarcoma was observed in our series. In a 2007 meta-analysis, Grulich et al26 compared cancer incidence among adult HIV-AIDS patients with the incidence of cancers among transplant recipients. There was a significantly increased incidence in both populations in 20 out of 28 types of cancer, and most of these were cancers r
2013 Lippincott Williams & Wilkins
Malignancy in HIV-positive South African Children
with a known infectious cause such as Kaposi sarcoma, non-Hodgkin and Hodgkin lymphoma, cervical cancer, liver cancer, and stomach cancer. Rates of most of the common epithelial cancers were not increased. This provides strong evidence that it is immune deficiency that is responsible for the increased risk of cancer rather than other risk factors. It seems that the range of infectionrelated cancers is much wider than previously appreciated, and that infection-related cancer will become an increasingly important complication of long-term HIV infection. Published trends in cancer risk among people with AIDS provide further support for this assertion. There was a drastic decline in the incidence of Kaposi sarcoma and non-Hodgkin lymphoma in the United States between 1990 to 1995 and 1996 to 200236 with the introduction of HAART, but adults with AIDS remained at increased risk compared with the general population. There was no change in the risk of cervical cancer and there was an increase in the incidence of Hodgkin lymphoma. Similarly the Italian pediatric registry data5 show a decline in cancer rates since the introduction of HAART, but the incidence of cancer still remains 5 times higher than that for HIVuninfected children. Thus, we conclude that the degree of immune suppression as measured by the CD4 count is not the sole risk factor for cancer. Most African children infected with HIV acquire the virus perinatally, with a slight female predominance.37 In this series of malignancies, similar to other studies,8,9,38 a male predominance of 1.64:1 was observed. This could be ascribed either to the slightly higher tendency for the male sex to develop malignancy or to the characteristic male predominance in Kaposi sarcoma and Burkitt lymphoma. Over 80% of the 239 cancers where staging information was available had advanced disease (stages III or IV) that had a negative impact on eventual survival. The factors associated with delay in diagnosis of cancers in children have been shown to be complex, including the site of the cancer, the age of the child, family factors, and health system factors.39,40 Current recommendations should include regular clinical evaluation of HIV-positive children paying particular attention to the warning signs of cancer.41 Such regular examinations should take place even before initiation of HAART; in our study half of the subjects developed a malignancy while their CD4 counts indicated only mild to moderate immune deficiency. What adds to the difficulty is that these children frequently present with disease in unusual sites42 and at the same time antiretroviral therapy–naive children may have gross lymphoproliferation in the absence of malignancy. For this reason it is important to develop a framework and clinical guidelines to foster early detection.43 Active tuberculosis was found in 28.5% of the children at diagnosis of their cancer; and the treatment for tuberculosis was instituted before starting chemotherapy. Although the possibility of adverse effects resulting from the interaction of tuberculosis drugs with chemotherapy and antiretrovirals was a concern, there were no documented cases in this series. More than 10% of the 197 children treated with chemotherapy and/or radiotherapy died from toxicity. This percentage is high when compared with published data from well-resourced countries, where the risk of death attributed to infections in HIV-negative children on chemotherapy is usually in the order of 1%,44 but low when compared with poorly resourced countries where figures of www.jpho-online.com |
115
J Pediatr Hematol Oncol
Davidson et al
14% have been reported in HIV-negative children.45 Such findings underscore the risks of administering cytotoxic medication to patients who are already immune compromised. Notwithstanding concerns about toxicity most children died as a result of disease progression. The overall disease-free survival rate was low at 33.7%, due mainly to late presentation, but results were better for those treated with chemotherapy and HAART who achieved a 57.8% overall survival rate. The best results were seen in patients with Hodgkin lymphoma who had an estimated 5-year survival rate of 71.4%. REFERENCES 1. Stack M, Walsh PM, Comber H, et al. Childhood cancer in Ireland: a population-based study. Arch Dis Childhood. 2007; 92:890–897. 2. Ries LAG, Melbert D, Krapcho M, et al. SEER Cancer Statistics Review, 1975-2004. Bethesda, MD: National Cancer Institute . Available at: http://seer.cancer.gov/csr/1975_2004/. Accessed December 9, 2008. 3. Centers for Disease Control and Prevention. US HIV and AIDS cases reported through December 1996. HIV/AIDS Surveill Rep. 1996;8:1–39. 4. Biggar RJ, Frisch M, Goedert JJ. Risk of cancer in children with AIDS. JAMA. 2000;284:205–209. 5. Chiappini E, Galli L, Tovo P-A, et al. Cancer rates after year 2000 significantly decrease in children with perinatal HIV infection: A study by the Italian register for HIV infection in children. J Clin Oncol. 2007;25:97–101. 6. Alvaro-Meca A, Micheloud D, Jensen J, et al. Epidemiologic trends of cancer diagnoses among HIV-infected children in Spain from 1997 to 2008. Pediatr Infect Dis J. 2011;30:764–768. 7. Lanzkowsky P. Manual of Pediatric Hematology-Oncology. London: Academic Press; 2011:101. 8. Ziegler J, Katongole-Mbidde E. Kaposi’s sarcoma in childhood: an analysis of 100 cases from Uganda and relationship to HIV infection. Int J Cancer. 1996;65:200–203. 9. Newton R, Ziegler J, Beral V, et al. Uganda Kaposi’s Sarcoma Study Group. A case-control study of human immunodeficiency virus infection and cancer in adults and children residing in Kampala, Uganda. Int J Cancer. 2001;92:622–627. 10. Cairncross L, Davidson A, Millar AJW, et al. Kaposi sarcoma in children with HIV: a clinical series from Red Cross Children’s Hospital. J Ped Surg. 2009;44:373–376. 11. Mueller BU. Cancers in children infected with the human immunodeficiency virus. Oncologist. 1999;4:309–317. 12. McClain KL, Leach CT, Jenson HB, et al. Association of Epstein-Barr virus with leiomyosarcomas in children with AIDS. N Engl J Med. 1995;332:12–18. 13. Chang Y, Cesarman E, Pessin MS, et al. Identification of herpesvirus-like DNA sequences in AIDS-associated Kaposi’s sarcoma. Science. 1994;266:1865–1869. 14. Flint SJ, Enquist LW, Racaniello VR, et al. Infection of a Susceptible Host. In: Flint SJ, Enquist LW, Racaniello VR, Skalka AM, eds. Principles of Virology, Volume II, Pathogenesis and Control. 3rd ed. Washington: ASM Press; 2009:3–27. 15. Chitsike I, Siziya S. Seroprevalence of human immunodeficiency virus type 1 infection in childhood malignancy in Zimbabwe. Cent Afr J Med. 1998;44:242–245. 16. Sinfield RL, Molyneux EM, Banda K, et al. Spectrum and presentation of paediatric malignancies in the HIV era: experience from Blantyre, Malawi, 1998-2003. Pediatr Blood Cancer. 2007;48:515–520. 17. Davidson A, Hendricks M, Geel J, et al. Malignancy in HIVpositive South African children. Pediatr Blood Cancer. 2009; 53:719. 18. WHO, UNAIDS, UNICEFGlobal HIV/AIDS response: update 2011. Epidemic update and health sector progress towards Universal Access. Available at: http://
116 | www.jpho-online.com
19.
20.
21.
22.
23. 24. 25. 26.
27.
28. 29. 30. 31.
32. 33. 34. 35. 36. 37. 38. 39. 40.
Volume 36, Number 2, March 2014
whqlibdoc.who.int/publications/2011/9789241502986_eng.pdf. Accessed May 27, 2012. Granovsky MO, Mueller BU, Nicholson HS, et al. Cancer in human immunodeficiency virus-infected children: a case series from the Children’s Cancer Group and the National Cancer Institute. J Clin Oncol. 1998;16:1729–1735. Caselli D, Klersy C, de Martino M, et al. Human immunodeficiency virus-related cancer in children: incidence and treatment outcome—report of the Italian register. J Clin Oncol. 2000;18:3854–3861. Statistics South Africa. Statistical release P0302: mid-year population estimates 2010. Available at: http://www.statssa. gov.za/publications/P0302/P03022010.pdf. Accessed May 27, 2012. Horner MJ, Ries LAG, Krapcho M, et al. SEER Cancer Statistics Review, 1975-2006. Bethesda, MD: National Cancer Institute. Available at: http://seer.cancer.gov/csr/1975_2006/. Accessed January 17, 2010. Kest H, Brogly S, McSherry G, et al. Malignancy in perinatally human immunodeficiency virus-infected children in the United States. Pediatr Infect Dis J. 2005;24:237–242. Chanock SJ, Pizzo PA. Infection prevention strategies for children with cancer and AIDS: contrasting dilemmas. J Hosp Infec. 1995;30(suppl):197–208. Bower M, McCall-Peat N, Ryan N, et al. Protease inhibitors potentiate chemotherapy-induced neutropenia. Blood. 2004; 104:2943–2946. Grulich AE, van Leeuwen MT, Falster MO, et al. Incidence of cancers in people with HIV/AIDS compared with immunosuppressed transplant recipients: a meta-analysis. Lancet. 2007; 370:59–67. Department of Health, South AfricaOperational plan for comprehensive HIV and AIDS care, management and treatment for South Africa. Available at: http://www.hst.org.za/ uploads/files/aidsplan.pdf. Accessed January 27, 2012. Stefan DC, Stones DK, Wainwright L, et al. Kaposi sarcoma in South African children. Pediatr Blood Cancer. 2011;56: 392–396. Davidson A. Kaposi sarcoma: the African HIV epidemic’s partner in crime. Pediatr Blood Cancer. 2010;54:657–658. Gantt S, Kakuru A, Wald A, et al. Clinical presentation and outcome of epidemic Kaposi sarcoma in Ugandan children. Pediatr Blood Cancer. 2010;54:670–674. Mutalima N, Molyneux EM, Johnston WT, et al. Impact of infection with human immunodeficiency virus-1 (HIV) on the risk of cancer among children in Malawi—preliminary findings. Infect Agent Cancer. 2010;5:5. Hesseling P, Wood RE, Nortje´ CJ, et al. African Burkitt lymphoma in the Cape province of South Africa and in Namibia. Oral Surg Oral Med Oral Pathol. 1989;68:162–166. Angeletti PC, Zhang L, Wood C. The viral etiology of AIDSassociated malignancies. Adv Pharmacol. 2008;56:509–557. Kutok JL, Wang F. Spectrum of Epstein-Barr virus-associated diseases. Annu Rev Pathol. 2006;1:375–404. Bajaj BG, Murakami M, Robertson ES. Molecular biology of EBV in relationship to AIDS-associated oncogenesis. Cancer Treat Res. 2007;133:141–162. Engels EA, Pfeiffer RM, Goedert JJ, et al. Trends in cancer risk among people with AIDS in the United States 1980-2002. AIDS. 2006;20:1645–1654. Biggar RJ, Taha TE, Hoover DR, et al. Higher in utero and perinatal HIV infection risk in girls than boys. J Acquir Immune Defic Syndr. 2006;41:509–513. Amir H, Kaaya EE, Manji KP, et al. Kaposi’s sarcoma before and during a human immunodeficiency virus epidemic in Tanzanian children. Pediatr Infect Dis J. 2001;20:518–521. Stefan DC, Siemonsma F. Delay and causes of delay in the diagnosis of childhood cancer in Africa. Pediatr Blood Cancer. 2011;56:80–85. Haimi M, Peretz Nahum M, Ben Arush MW. Delay in diagnosis of children with cancer: a retrospective study of 315 children. Pediatr Hematol Oncol. 2004;21:37–48. r
2013 Lippincott Williams & Wilkins
J Pediatr Hematol Oncol
Volume 36, Number 2, March 2014
41. Stones DK. Sub-Saharan African children and cancer: how do we improve the outcome? Afr J Cancer. 2011;3: 91–93. 42. Mueller BU. Cancers in human immunodeficiency virusinfected children. Monogr Natl Cancer Inst. 1998;23:31–35. 43. Davidson A, Eley B. HIV and childhood cancer. Cont Med Ed J. 2010;28:337–342.
r
2013 Lippincott Williams & Wilkins
Malignancy in HIV-positive South African Children
44. Lund B, A˚sberg A, Heyman M, et al. for the Nordic Society of Paediatric Haematology and Oncology. Risk factors for treatment related mortality in childhood acute lymphoblastic leukemia. Pediatr Blood Cancer. 2011;56:551–559. 45. Gupta S, Bonilla M, Gamero M, et al. Microbiology and mortality of pediatric febrile neutropenia in El Salvador. J Pediatr Hematol Oncol. 2011;33:276–280.
www.jpho-online.com |
117
