Abstract. Reports dealing with childhood cancer population-based survival have not yet been pub- lished in Cuba. A survival study including cases re- ported to ...
European Journal of Epidemiology 16: 763±767, 2000. Ó 2000 Kluwer Academic Publishers. Printed in the Netherlands.
Childhood cancer survival in Cuba M.G. Boschmonar1, Y.G. Alvarez1, A.M. GarcõÂ a1, T.N. Soto1, M.C. Roger2 & L.F. Garrote3 1
National Cancer Registry, National Institute of Oncology and Radiobiology; 2National Cancer Control Program, National Institute of Oncology and Radiobiology; 3Vicedirector of Research, National Institute of Oncology and Radiobiology, Havana City, Cuba Accepted in revised form 9 September 2000
Abstract. Reports dealing with childhood cancer population-based survival have not yet been published in Cuba. A survival study including cases reported to the National Cancer Registry in the period 1988±1989 have been accomplished in the National Cancer Registry of Cuba. A comprehensive view of its results in childhood cancer is intended to be shown in this paper. All sites childhood cancer cases aged below 15 years (400) and reported in this period were included derived from a total of 578 incident cases. Vital status of cases was checked-up to December of 1994 by a mixed follow-up. Survival analysis was achieved by the life-table method using the SPSS for Windows software. Relative survival rates were not included because no dierences were appreciated with the observed ones. Cases between 0±4 years and females account for the highest frequencies. Leukaemia accounts for 27% of the cases, followed by lymphomas and tumours of the central nervous system.
Lymphoid and non-lymphoid acute leukaemia had 41 and 9% 5-year survival rate respectively being the latter comparable with Bangalore, India (10%). For lymphomas, Hodgkin's disease present better prognosis compared with non-Hodgkin lymphoma with 73 and 35% 5-year survival rates, respectively. Figures are comparable to the corresponding to Bangalore, India (72 and 33%, respectively). For time trends survival distribution were compared by period of diagnosis (1982/1988±1989). In all cases dierences resulted statistically non-signi®cant with lower rates for 1988±1989 except for lymphomas. International reports show increasing time trends for childhood cancer. Nevertheless, these results may serve as a milestone for future comparisons, corresponding with the creation of the National Program for Childhood Cancer Control and future studies will con®rm the success of its measures.
Key words: Cancer Registry, Childhood cancer, Survival Introduction In spite of the low proportion that childhood cancer represents in the global cancer burden of a country, these cancers are very important due to the emotional involvement and the high number of lost years-person they cause. In Cuba, the occurrence of childhood cancer had a stable trend in the last decades being leukaemia the more frequent diagnostic group, followed by lymphomas and tumours of the central nervous system [1, 2]. Since the 1960s, important advances in childhood cancer treatment and its centralization were achieved mainly in developed countries. These resulted in a dramatic improving of survival and a decrease of mortality for several sites, specially for leukaemia [3±5]. In Cuba, since the 70th decade great eorts have been done in the centralization of childhood cancer management along the country. These eorts culminated in the creation in 1989 of the National Program for Childhood Cancer Control, as a new step in the actions related to cancer control. Organ-
ised eorts in this sense started in Cuba in 1964 with the creation of the National Cancer Registry (NCR) a population-based cancer registry. The NCR have published several reports of childhood cancer incidence [6±8]. Few studies of population-based survival rates of childhood cancer have been published, mainly from developed countries [9±21]. The NCR of Cuba has not published any report of this concern. In 1994, a survival study was started including all ages cancer cases incident to NCR in 1988±1989 and only the results for adult cases were published [22±23]. A comprehensive view of its results in childhood cancer are intended to be shown in this paper. Materials and methods A descriptive survival study was achieved with the inclusion of cancer cases in ages lower than 15 years, incident to NCR in the period 1988±1989. From 578
764 cases included in the NCR database, 84 were cases reported by death certi®cate only (DCO) for which date of diagnosis could not be traced back and were excluded from the study. A mixed follow-up was achieved by which vital status of cases was established up to 31 December 1994 guaranteeing a 5 years minimum period of follow-up for each case, details were described elsewhere [22]. Cases lost to follow-up also lacking date of last contact (94; 16%) were excluded. Finally 400 cases were included in the survival analysis. The exclusion of 178 cases from the study may account for certain dierences in relative frequencies by age group. Univariate survival analysis was made by the lifetable method [24], relative survival proportions were not included in this paper since no dierences were appreciated with the observed ones apparently due to low mortality from other causes in Cuban children. Global estimations of observed survival ®gures were made as well as by sex, age group, province and diagnostic group. The latter were de®ned by Birch and Mardsen classi®cation (1987) [25]. Results were compared with the non-published results of a former survival study dealing with cases reported to the NCR in 1982. Results Table 1 gives the number, proportion and 5 years observed survival rates of cases registered by age group and gender for all childhood cancers. The number of cases included in the study was higher for the 0±4 year age group and for females. There was little dierence in survival overall between the age groups and the dierence in observed survival rates between genders was not signi®cant (p = 0.6465) although females present higher ®gures. Regarding diagnostic category (Table 2), leukaemia is the most frequent, accounting for 25% of all the cases included in the study, followed by lymphomas and tumours of the central nervous system. Retinoblastoma showed the highest 5-year survival proportion, the lower ®gures corresponding to leukaemia and bone tumours. Table 1. Number, relative frequencies and survival of cases (1988±1989) for all childhood cancers by age group and gender Age group (years)
N
Frequency (%)
5-year survival (%)
0±4 5±9 10±14
187 87 126
47 22 32
46 50 48
Gender Male Female
190 210
48 53
46 49
Table 2. Number, relative frequencies and survival ®gures of cases (1988±1989) for all childhood cancers by diagnostic group Diagnostic group
N
Frequency (%)
5-year survival (%)
Leukaemias Lymphomas CNS Malignant bone tumours Retinoblastoma Renal tumours Soft tissue neoplasms Other malignant tumours
101 75 60 28
25 19 15 7
39 50 54 41
21 21 16 78
5 5 4 20
70 62 54 36
All sites
400
100
48
The geographic pattern of survival is shown in Table 3. Interpretation is dicult because of the small number of cases in some provinces. If we take only provinces with more than 10 cases, there is a gradient of decreasing survival from west to east provinces. Among provinces of treatment, Havana City presents signi®cant dierences with Holguin and Santiago de Cuba (p = 0.0181 and p = 0.0103, respectively). Survival curves by diagnostic groups for leukaemia and lymphomas are shown in Figures 1 and 2. Lymphoid and non-lymphoid acute leukaemia with 41 and 9% 5-year survival proportions, respectively and Hodgkin's disease and non-Hodgkin lymphoma with 73 and 35% 5-year survival proportions, respectively. Survival ®gures by the most frequent diagnostic groups and period of diagnosis for both sexes at 1, 3 and 5 years are shown in Table 4. More over, 1-year Table 3. Province of residence and province of attendance in 5-year survival Province of attendance
N
5-year survival (%)
Pinar del RõÂ o Havana Havana city Matanzas Villa Clara Cienfuegos Sancti Spiritus Ciego de Avila Camaguey Las Tunas Holguin Gramma Santiago de Cuba Guantanamo Cuba
18 1 195 5 23 6 6 5 18 6 31 12 66 8 400
56 ± 54 ± 48 ± ± ± 44 ± 29 50 39 ± 48
765
Figure 1. Survival distribution for leukaemias by diagnostic group. Both sexes.
Discussion
Figure 2. Survival distribution for lymphomas by diagnostic group. Both sexes.
survival was lower in 1988±1989 for leukaemia and CNS tumours, while for lymphomas and all cancer combined there were only small decreases compared to 1982. These survival distributions are shown in Figures 3±5 for leukaemia, lymphomas and overall childhood cancer. In all cases dierences resulted statistically non-signi®cant with p = 0.8830; p = 0.5612 and p = 0.6287, respectively.
This is the ®rst report regarding population-based survival rates for childhood cancer in Cuba. Although there may be an important bias in survival and frequency estimations due to the exclusion of high proportions of DCO and lost to follow-up cases from the study, our report attempts to give an insight of the survival patterns of this cancer in a developing country. This is the ®rst report of this kind from Cuba and also from Latin America. Overall 5-year survival for childhood cancers in Cuba is low compared with ®gures available for developed countries as 70% for Italy (1986±1989) [19] and 64% for Denmark (1983±1987) [17], but higher than 37% for Bangalore (India) [9]. For total leukaemia, the 5-year survival ®gures from Cuba is equal to 39% for Slovakia (1983±1987) [11] and lower to 66% in USA (1983-1992) [21]. India [9] shows an comparable 10% rate for acute nonlymphoid leukaemia, other countries presenting values ranking from 13% for Slovakia [11], 27% for
Table 4. Survival rates by year of follow-up and major diagnostic groups of childhood cancer in Cuba Diagnostic group
Survival probability (SE) 1982 N
Leukaemia Lymphomas CNS All
1988±1989 1
3
5
N
1
3
5
44 50 17
72.22 (7.47) 72.84 (6.99) 78.57 (10.97)
36.11 (9.29) 54.63 (8.79) 49.11 (15.09)
28.09 (8.79) 54.63 (8.79) 49.11 (15.09)
101 75 60
62.19 (4.84) 58.11 (5.74) 61.02 (6.35)
45.14 (4.97) 49.81 (5.83) 57.53 (6.45)
39.12 (4.88) 49.81 (5.83) 53.96 (6.52)
172
73.76 (3.71)
53.55 (4.70)
51.53 (4.73)
400
63.08 (2.43)
52.27 (2.52)
47.91 (2.52)
766
Figure 3. Survival distribution for leukaemias by period of diagnosis. Both sexes. 1982/1988±1989.
and lymphoma, a similar feature is shown for the ®rst (Figure 3) and for lymphomas slightly higher values are noted (Figure 4). Leukaemia is the most frequent of childhood cancers and highly amenable to cytotoxic therapy, also, Cuba was under a very severe blockade in the follow-up period of the study (1990± 1994) with great restrictions regarding very important drugs. These facts could have in¯uenced survival trends in spite of the eorts on centralization of childhood cancer care. Also, the fact that the tracing back of cases informed by DCO to the NCR started with the cases reported in the period 1988±1989 as was stated elsewhere [23] reducing DCO proportions, could lowered survival probabilities for the period giving the eect shown when comparing with the previous period. It can be appreciated that there are many tasks to achieve in order to enhance the quality of childhood cancer care. The results of an on-going survival study including cases reported to the NCR in the period 1990±1995 will give further information and help to estimate survival trends in Cuba. References
Figure 4. Survival distribution for lymphomas by period of diagnosis. Both sexes. 1982/1988±1989.
Figure 5. Survival distribution for childhood cancer by period of diagnosis. Both sexes. 1982/1988±1989.
USA [21] and 40% for Italy [19]. For acute lymphoid leukaemia Cuban's value (41%) is lower than other countries ranking from 50% for Slovakia [11], 69% USA [21], 70% for Denmark [17] and 74% for Italy [17]. Regarding Hodgkin's disease Cuban ®gure is comparable with 72% for India [9], meanwhile, the other countries rank from 86% for Slovakia [11], 87% for Italy [19] and 88% for USA [21]. For nonHodgkin lymphomas only India and Slovakia present 5-year survival ®gures similar to Cuban's with 33 and 37%, respectively. Time trends analyses for overall childhood cancer in Cuba showed non-signi®cant decreases when comparing survival distribution for both periods, with lower ®gures for the latter. Regarding leukaemia
1. Alert J. Incidence of malignant tumours in childhood in Cuba in the years 1973±1975 and 1976±1978. Rev Cub Ped 1985; 57: 469 [in Spanish]. 2. Martin A, Grueiro S, ValdeÂs M, Longchong M. Incidence of childhood cancer in Cuba (1983±1987). Rev Cub Oncol 1994; 10(1±2): 50±58 [in Spanish]. 3. IARC±European Commission. Survival of cancer patients in Europe. In: Berrino F, Sant M, Verdecchia A, Capocaccia R, HakulinenT (eds), The Eurocare Study. IARC Scienti®c Publications No. 132, Lyon: IARC, 1995. 4. SaÂnchez V, SaÂnchez de Toledo J, Mingot M, BorraÂs JM. Trends in childhood mortality of tumours in CatalunÄa (1975±1992). Med Clin (Barc) 1995; 104: 241± 244 [in Spanish]. 5. Draper GJ. Childhood cancer: Trends in incidence, survival and mortality. Eur J Cancer 1995; 31A(5): 653±654. 6. Caraballoso M. National Cancer Registry of Cuba: 1971±1980. In: Parkin DM, Stiller CA, Draper GJ and others (eds), International Incidence of Childhood Cancer. IARC Scienti®c Publications No. 87, Lyon: IARC, 1988. 7. MartõÂ n A, et al. National Cancer Registry of Cuba: 1986±1990. In: Parkin DM, Kramarova E, Drapes GJ, Marsuyer E, Michaelis J, Neglia J, Qureshi S, Stiller SA (eds), International Incidence of Childhood Cancer. Vol. II, International Agency for Research on Cancer, Lyon, France. IARC Scienti®c Publications No. 144, 1998, pp. 79±82. 8. Martin A, Alert JA, Ren JS, Longchong M, Grueiro S. Incidence of childhood cancer in Cuba (1986±1990). Int J Cancer 1997; 72: 551±555. 9. Nandakumar A, Narayanappa A, et al. Descriptive epidemiology of childhood cancers in Bangalore, India. Cancer Causes and Control 1996; (7): 405±410.
767 10. Plesko I, Somogyi E, Dimitrova E, KramaÂrova E. Childhood cancer malignancies in Slovak. Neoplasm 1989; 36(2): 233±243. 11. KramaÂrova E, et al. Improving survival for childhood cancer in Slovak. Int J Cancer 1996; 65: 594±600. 12. Stiller CA, Bunch KJ. Trends in survival for childhood cancer in Britain diagnosed in 1971±1985. Brit J Cancer 1990; 62: 806±815. 13. Stiller CA. Population-based survival rates for childhood cancer in Britain. 1980±1991. Brit Med J 1994; 309: 1612±1616. 14. National Registry of Paediatric Tumours (RNTI). Basic Statistics 3 (1980±1990). Survival 1980±1989. Monogra®es SanitaÁries, SeÁrie A (Estudis) No. 25, 1993 (in Spanish). 15. Mc Whuirter WR, Ring I, et al. Childhood Cancer in Australia 1977 to 1982. Incidence and Survival. Australian Paediatric Cancer Registry, Brisbane, 1985. 16. Adami HO, et al. Trends in childhood and adolescent cancer survival in Sweden 1960±1984. Acta Oncol 1992; 31: 1±10. 17. de Nully Brown P, Olsen JH, Hertz H, Carstensen B, Bautz A. Trends in survival after childhood cancer in Denmark, 1943±1987: A population-based study. Acta Paediatric 1995; 84(3): 316±324. 18. Black R, Sharp L, Kendrick SW. Trends in Cancer Survival in Scotland. 1968±1990. ISD Pub, Edinburgh, Scotland, 1993.
19. Magnani C, Pastore G. Survival of childhood cancer patients in Italy, 1978±1989. ITACARE Working Group. Tumori 1997; 83(1): 426±489. 20. Giles G, Farrugia H, Thurs®eld V, Staples M. Childhood Cancer in Victoria. Cartoon South Victoria, Canada. Anticancer Council of Victoria, 1994; Canstat No. 18. 21. Novakovic B. US childhood cancer survival, 1973± 1987. Med Pediat Oncol 1994; 23: 480±486. 22. FernaÂndez L, Graupera M, MartõÂ n A, GalaÂn Y, Lezcano M. Survival of cancer in Cuba. In: Sankaranarayanan R, Black R, Parkin DM (eds), Cancer Survival in Developing. International Agency for Research on Cancer, Lyon, France. IARC Scienti®c Publications, No. 145, 1998, pp. 51±58. 23. Graupera M, et al. Trends in cancer survival in Cuba. Eur J Epid 1999; 15: 521±528. 24. EsteÁve J, Benhamou E, Raymond L. Statistical Methods in Cancer Research. Vol. 4. Descriptive Epidemiology. IARC Scienti®c Publications No. 132, Lyon: IARC, 1994. 25. Birch JM, Marsden HB. A classi®cation scheme for childhood cancer. Int J Cancer 1987; 40: 620±624. Address for correspondence: M. Graupera, National Cancer Registry, National Institute of Oncology and Radiobiology, 29 Y Vedado, Havana city, Cuba E-mail: rncinor@infomed.sld.cu
