Childhood diabetes in Arab countries* Diabetes Epidemiology Research International Study Group' Insulin-dependent diabetes mellitus (IDDM) is a chronic disease of childhood that is associated with high costs, mortality and morbidity, but which is of unknown etiology. Globally, the incidence and prevalence of the disease are highly variable. Study of IDDM among Arab children, who have similar genetic characteristics, but markedly different environmental backgrounds, could provide important insight into its cause. Few studies of IDDM in Arab populations have been carried out, but the limited data available indicate that there are marked variations in the risk of the disease and in its distribution between the sexes. It is therefore very important that IDDM registries be established in Arab countries since this could lead to a greater understanding of the disease and perhaps its prevention.

Introduction Diabetes mellitus is a disorder characterized by a decreased secretion of insulin by the pancreas that results in a reduction in the body's ability to metabolize carbohydrates, and is frequently associated with disturbances in the metabolism of proteins and fats (1). Insulin-dependent diabetes mellitus (IDDM) is characterized by a low level or absence of circulating endogenous insulin and a proneness towards ketosis. It is usually associated with the presence of islet cell antibodies and is correlated with certain human histocompatibility leukocyte-associated (HLA) antigens (2). In the USA, diabetes is one of the leading, if not the leading, chronic disease of children (3). The mortality associated with IDDM is high, and about 12% of persons with the disease die within 20 years of its onset (4); this represents a risk of dying that is 7-10 times that of nondiabetics of the same age. The complications of IDDM are numerous and include the following: ketoacidosis, nephropathy, neuropathy, retinopathy, heart disease, and atherosclerosis (5,6). The treatment of IDDM and its associated complications is extremely expensive, with *This article was prepared by Dr T. Dokheel, Department of Epidemiology, Graduate School of Public Health, University of Pittsburgh, Pittsburgh, PA 15261, USA. Requests for reprints should be sent to this address. ' Study Group member countries and representatives: Algeria (K. Bessaoud); Egypt (N. Hashem, A. Massoud, and M. Radwan); Finland (J. Tuomilehto); Iraq (A.A.S. Alwan); Japan (N. Tajima); Jordan (K. Ajlouni, A.F. Al-Hader, R. Al-Kotob, 0. Al-Sartawi, H. Al-Qassab, and D.A. Lateef); Kuwait (M. Khogali, A. Gumaa, and N. Abdella); Lebanon (C. Macaron); Libyan Arab Jamahiriya (O.A. Kadiki); Poland (M. Rewers); Saudi Arabia (A. Abanamy, A.S. Al-Kassab, and M. Khalil); Sudan (A.A. Elamin); Tunisia (R. Bouguerra, R. Boukhris, S. Chamakhi, N. Jamel, F. Ben Khalifa, and K. Nagati); and USA (T. Dokheel, J. Dorman, R. LaPorte, and C. Moy).

Reprint No. 5067 Bulletin of the World Health Organization, 6 (2): 231-236 (190)

an estimated average cost per patient from onset to age 40 years of over US$ 50 000 (7). The current global effort to evaluate the patterns of diabetes continues to provide insight into its etiology and prevention. In the present article a case is presented that the evaluation of IDDM among Arab children could provide important information about its etiology as well as the cause of the marked variation in its incidence with climate and socioeconomic and living conditions.

Overview of the epidemiology of IDDM Globally, the incidence and prevalence of IDDM are highly variable. Depending on the country, the risk of developing it varies 60-fold (8), with the highest incidence in Finland (28.8 per 100 000 persons per year) and the lowest in the Republic of Korea (0.5 per 100 000 persons per year). There is also a large variation in its incidence across different racial groups; for example, in the USA, the incidence among Whites is higher than that among Blacks or Hispanics (9). Studies of migrant populations, such as children of French or Jewish backgrounds in Montreal (10) have demonstrated that migrants who move to a high-risk area for IDDM tend to assume a risk similar to that in the new location. Marked increases in the incidence of IDDM have been reported in many countries (11). These data indicate the importance of environmental factors in the development of diabetes (12). Identification of the external risk factors is therefore critical for potential prevention of the disease. Studies of the global differences in the incidence of IDDM could lead to the discovery of the causes of the disease. Similarly, international studies have led to a better understanding of the etiological factors of many conditions, including heart disease and certain cancers (13,14). Identification of the large differences in © World Health Organization 1990

231

Diabetes Epidemiology Research International Study Group

the incidences and prevalences of these diseases in various geographical areas and races has stimulated more specific studies designed to determine their possible environmental or genetic determinants.

Arab heritage and the etiology of IDDM The unique history of Arab populations has yielded a relatively homogeneous gene pool spread across markedly varied living conditions. Investigation of IDDM among Arab heritage populations in various countries could therefore provide important information about its etiology. After the rise of Islam in 622 AD and the emergence of the Arabian Muslims as the founders of a great empire that extended over most of the Middle East, the name "Arab" came to be used by the Muslims themselves, and by the countries with whom they came into contact, to indicate all people of Arab origin from the Arabian peninsula (15,16) The League of Arab States comprises countries that extend across North Africa and the Middle East, from the Atlantic Ocean to the Indian Ocean, and from the borders of Turkey to equatorial Africa (17). The total area of the Arab countries is approximately 13 million km2. Only one country, Somalia, is crossed by the equator. Tunisia is furthest from the equator at latitude 370 north. This is important, since there appears to be an association between latitude and the incidence of IDDM; however, little is known about its incidence in equatorial areas (9). Among the Arab states the climate varies markedly (desert, tropical, subtropical, or Mediterranean), both from one country to another and in different areas of the same country. In summer, the highest temperature (54.4 C) occurs in Saudi Arabia and the lowest ( < 0 °C) in the mountains of Lebanon. There is an association between the incidence of diabetes and climate; for a given population the incidence is higher in cooler months, and between populations it increases as the climate coals (9). Study of the incidence of IDDM among Arab children should therefore permit investigation of the relationship between this and climate. The total population in Arab countries is approximately 160 million (18). Although there are no accurate statistics about the racial composition, it is likely that, because of religious and cultural constraints, the degree of mixing is lower than in other populations. There are several other racial and ethnic groups living in Arab countries that are genetically quite distinct from the Arabs. One of these is the Kurds, who represent about 12% (1.5 million) of the population of Iraq, as well as a small percentage of the population of the Syrian Arab Republic. In 232

Sudan, Somalia, Djibouti and Mauritania there are large Negroid populations, and a small number of Negroids live in every Arab country. A large number of Iranians, Pakistanis, Indians, Koreans, and Filipinos live in the countries of the Gulf. The populations of Lebanon and the Syrian Arab Republic include a small percentage of Turks, Europeans, and Armenians. The Berbers, who constitute over half the population of North Africa, represent a distinct ethnic group. There are therefore many ethnic and racial groups in Arab countries; however, of importance for understanding the interaction between environmental and genetic factors that produces IDDM is that these groups are probably genetically distinct. There is a sufficient variation in the socioeconomic status of Arab countries (range in annual per capita income, US$ 100 to US$ 15 000) to permit evaluation of the degree to which IDDM is related to this parameter (19). Arab countries, with their dramatic differences and rapid changes in socioeconomic status, are therefore ideal for studying the effect of socioeconomic factors on IDDM (20). Lifestyle also varies considerably in Arab countries, from nomadic, semi-nomadic, to urban, and to date there has been no investigation of IDDM among nomadic cultures. The most common language in Arab countries is Arabic and the most common religion is Islam. To date, the incidence of IDDM has been studied largely only in Christian countries. With the exception of Kuwait, there are no well-established adequately standardized national diabetes registries in Arab countries (21). Table 1 provides an overview of the limited epidemiological data that are available for Arab populations in various countries. A population-based national study of the incidence of IDDM in Kuwait was carried out in 198081. The annual incidence for children aged 0-14 years was 3.96 per 100 000 per year (21). In a retrospective hospital-based study, the incidence of IDDM in 1980-82 in the Eastern Province of Saudi Arabia was evaluated among Aramco employees and their dependents at the company hospital. The incidence was high, 7 per 100 000 per year, and the prevalence, 20 per 100 000.' The study also determined that in 1977-79 there were only four new cases of IDDM, while seven new cases were reported in 1980-82. This apparent increase could be due to a greater overall number of patients being treated at a

Mathew, P.M. & Hamdam, J.A. Presenting features and prevalence of juvenile diabetes in Saudi Arabian children. Unpublished document, Pediatric Services Division, Dhahran Health Center, Dhahran, Saudi Arabia, 1982. WHO Bulletin OMS. Vol 68 1990.

Childhood dlabetes In Arab countries Table 1: Incidence and prevalence of Insulin-dependent diabetes mellitus among Arabs living in various countries Period of

Age group

study

(years)

1980-81

0-14

1980-82 1987 1960-70 1975-80

0-19 0-14 7-13 0-15 0-20

Study area Kuwait

National

Saudi Arabia Sudan

Eastern Province Khartoum

France Israel

Rh6ne D6part6ment National

Population at risk

Incidence (per 100 000 per year)

568 000 714 000 -

3.96 5.61 7

63 000

-

-

10.2 1.2

2 million

Prevalence (per 100 000)

' Mathew, P.M. & Hamdam, J.A. Presenting features and prevalence of juvenile diabetes in document. Pediatric Services Division, Dhahran Health Center, Dhahran, Saudi Arabia.

-

-

20 95.35

Type of study

Reference

Population Hospital

21

22 23 24

School

-

Population

-

Population

Saudi Arabian children. Unpublished

Table 2: Prevalence, Incidence and sex ratio for Insulin-dependent diabetes meilitus In selected Arab and other countries Age group (years)

Saudi Arabia Libyan Arab Jamahiriya Sudan

Israel" Kuwait Finland Scotland France Japan

USA(AlleghenyCo., PA)

0-14 0-14 7-13 0-20 0-14 0-14 0-14 0-14 0-14 0-14

Incidence (per 100 000 per year) 7

Prevalence

(per 100 000) 20

-

95

1.2 3.96 29.5 19.9 4.4 0.6 11.4

-

-

Sex ratio (male-to-female) 1:0.375 1:1.083 1:1.409 1:1.409 1:1.449 1:0.9 1:0.84 1:0.93 1:1.429 1:1.22

M Mathew, P.M. & Hamdam, J.A. Presenting features and prevalence of juvenile diabetes in Saudi Arabian document. Pediatric Services Division, Dhahran Health Center, Dhahran, Saudi Arabia. ° Among Arab residents.

the hospital or to a true increase in the incidence of IDDM. It is essential to determine whether there has been a similar increase in the incidence of IDDM in Arab countries as that currently being observed in Europe (8). In Sudan, the prevalence of IDDM among children attending elementary school has been reported. A relatively high prevalence (95 per 100000) was found for the age group 7-13 years (22), which is similar to that in many European countries. This pattern may be due to the particular situation in Sudan, where only 5% of the population consists of individuals of pure Arab origin. Some information on the risk of IDDM among Arab children living in non-Arab countries is also available. In France, for example, the incidence of the disease in the Maghrebi population in the Rhone Ddpartement is relatively high (10.2 per 100 000 per year) (23). These data need, however, to be reexamined because of the large numbers of undocumented individuals living in the study area. This could lead to an underestimation of the at-risk WHO Bulletin OMS. Vol 68 1990.

Reference -' 25 22 24 21 9 9 9 9 9

children. Unpublished

population and therefore to an artificially high estimate of the incidence. However, it might suggest that migration affects the incidence of IDDM among Arabs. In Israel the incidence of IDDM among Arabs is 1.2 per 100 000 per year (24). This is very low, but must be viewed cautiously because of the possibility of missing the high proportion of Arab patients who seek treatment at Arab hospitals in Israel and' in neighbouring countries. In Europe, the difference in the incidence of IDDM between the sexes varies little from country to country, with a slight excess of males. Oriental populations also exhibit a consistent pattern in this respect, with an excess of females (8). In contrast in Arab countries the incidence of IDDM varies widely (Table 2). For example, in Saudi Arabia the male-tofemale ratio was 1:0.38;b in the Libyan Arab Jama-

b See footnote a, p. 232. 233

Diabetes Epidemiology Research International Study Group

hiriya it was 1: 1.1 (25,26); and for Arabs in Israel it was 1:1.41 (24). The lowest male-to-female ratio (1:1.45) was in Kuwait (21). The reasons why the sex ratio of the incidence of IDDM varies so markedly are not yet clear. Several studies have evaluated the relationship between HLA antigens and the incidence of IDDM in Arab populations. In Algeria, compared with nondiabetics, diabetics exhibit a decrease in HLA-A1 and HLA-DR2 and an increase in HLA-Aw 19.2, HLAB8, and HLA-B18, and there is a marked association with HLA-DR3 (27). In Kuwait for non-Gulf Arabs a strong positive association has been detected between the incidence of IDDM and both HLA-DR3 and HLA-DR4 (28); however, the incidence of IDDM among Gulf Arabs seems to be significantly associated only with HLA-DR3 (28). This is different from the situation in Caucasoid populations for which there is a strong association between the incidence of IDDM and HLA-DR3, HLA-DR4, and both HLA-DR3 and HLA-DR4 (2). Similar associations with HLA-DR3 and HLA-DR4 have been observed for Tunisian diabetics (29,30). It is interesting that the prevalence of HLA antigens that is associated with IDDM in Arab countries appears to be heterogeneous and different from that in European populations. Fig. 1 shows the prevalence of HLA-B8 and HLA-B15 in selected Arab countries compared with that for European Caucasoids. There is considerable variation in the prevalence of HLAB8 in Arab countries, with an antigen frequency that ranges from a value much higher to one much lower

Table 3: Percentage prevalence of HLA-DR phenotypes among Caucasolds, Tunisians, and Saudi Arabians Caucasoids

Tunisians

Saudi Arabians

23.6 23.8

28.4 20.2

27.0 30.0

HLA-DR3 HLA-DR4

than that among Europeans. In contrast, the prevalence of HLA-B15 is uniformly lower than that among Europeans. Limited data are available for the prevalence of HLA-DR3 and HLA-DR4 phenotypes for Tunisians, Saudi Arabians, and Caucasoids (Table 3) (27). The prevalence of the HLA-DR3 phenotype appears to be higher among Tunisians and Saudi Arabians, while the latter also have a higher prevalence of the HLA-DR4 phenotype. The results therefore indicate that the HLA antigen pattern that is associated with IDDM in Arab countries varies considerably with the particular country and appears to be different from that in Europe. It is thus very important to determine the degree to which the incidence of childhood diabetes reflects the heterogeneity of the prevalence of "diabetogenic" HLA antigens.

Conclusions Although Arabs have similar genetic characteristics, Arab countries vary dramatically in their geography and climate, as well as in the socioeconomic status and way of life of their inhabitants. They are

Fig. 1. Prevalence of HLA-B8 and HLA-B15 among the populatlons of selected Arab countries and among European Caucasolds. Caucaso

Tunisians

AlerIns (Kabylie) -.

Saudi Arabians

Iraqis Egyptians

.I

o

5

10 % prealnoe of HLA-BB

15

Caucasolds Tunisians Algerians (Kabyble) Saudi Arabians L nese Iraqis

Egyptans

II o

5

10

a5 % prevalence of HLA-815 234

21n

.C1

1:15

2Co

WHO Bulletin OMS. Vol 68 1990.

Childhood diabetes In Arab countries

therefore an important population group for studying the effect of the interaction of environmental factors and host susceptibility on the incidence of IDDM. Comparison of the incidence of IDDM among Arabs and non-Arabs who live in the same environments should also shed light on the importance of genetic and environmental factors and how they interact to cause diabetes. Available data on the incidence and prevalence of diabetes among Arabs suggest that there is a marked difference in its frequency and distribution between the sexes. These results should, nevertheless, be interpreted with caution because of the very different methods used to collect the data; however, the findings do demonstrate the importance of carrying out epidemiological evaluations of IDDM in Arab countries. A first step towards this would be to establish standardized IDDM registries. The incidence data obtained from these registries could help to determine the effect of factors such as climate, altitude, latitude, socioeconomic status, and life-style on the risk of IDDM. Cases identified by these registries could then be used for more in-depth genetic and virological testing. The results from studies of this type could also explain the apparent variation in the incidence of IDDM in Arab countries and provide considerable insight into the development of childhood diabetes in other countries.

Rfsum4 Dlab6te de l'enfant dans les pays arabes Le diabete sucre insulino-dependant (DSID) est une maladie chronique de 1'enfant associee a une mortalite et a une morbidite elevees, ainsi qu'a des couts importants, mais dont I'etiologie reste inconnue. L'incidence et la prevalence de cette maladie dans le monde sont tres variables, les chiffres obtenus pouvant etre 60 fois plus eleves la ou l'incidence est la plus forte, a savoir en Finlande (28,2 pour 100 000 par an) que la oui elle est la plus faible c'est-a-dire en Republique de Coree (0,5 pour 100 000 par an). Comme les enfants arabes appartenant aux 22 pays membres de la Ligue des Etats arabes ont des caracteristiques gen6tiques semblables, mais presentent des facteurs environnementaux tres differents, on a pense que l'etude de cette maladie chez eux pourrait permettre de mieux en connaitre les causes. Bien que ces populations arabes soient genetiquement semblables, la geographie et le climat des pays arabes presentent en effet des variations considerables, tout comme les conditions socio4conomiques qui y regnent et le mode de vie de leurs habitants. On sait peu de chose sur le diabete dans les pays WHO Bulletin OMS. Vol 68 1990.

arabes en raison de l'insuffisance des registres pour ceote maladie; toutefois, les quelques etudes disponibles indiquent qu'il y a des differences importantes d'un pays a l'autre. Par exemple, une etude menee dans la population du Koweit, entre 1980 et 1981, a montre que l'incidence du DSID chez les 0 a 14 ans etait de 3,96 pour 100 000 par an, alors qu'une etude retrospective menee dans les h6pitaux d'Arabie saoudite a fait etat d'une incidence de 7 pour 100 000 par an dans le meme groupe d'age. En France, dans le departement du Rhone, I'incidence de ceote maladie dans les populations d'origine arabe est de 10,2 pour 100 000 par an, et elle est de 1,2 pour 100 000 par an en Israel. Contrairement a l'uniformite de l'incidence et de la prevalence du DSID en Europe et en Extreme-Orient, ces mesures varient considerablement dans les pays arabes, avec un rapport de masculinite de 1:0,38 en Arabie saoudite et de 1:1,5 au KoweitPlusieurs etudes ont cherche a evaluer les relations existant entre les antigenes leucocytaires humains (HCA) d'histocompatibilite et le DSID dans les populations arabes. Les donnees recueillies indiquent qu'en plus de l'heterogeneite de la prevalence des antigenes HLA associes au DSID dans les pays arabes, les associations rencontrees sont tres differentes de celles observees chez les caucasiens d'Europe. La prevalence du HLA-B8 chez les diabetiques arabes varie considerablement et s'incrit dans un eventail de valeurs plus large que celui rencontre chez les europeens; cependant, la prevalence du HLA-B15 est inferieure a ce qu'elle est chez les caucasiens. Bien que l'HLA-DR3 et l'HLA-DR4 dominent chez les diabetiques d'Arabie saoudite, seule la prevalence de l'HLA-DR3 est elevee chez les diabetiques tunisiens. En raison des differences importantes que l'on observe dans les pays arabes concernant l'incidence, la prevalence, la distribution entre les sexes du diabete sucre insulino-dependant ainsi que son association avec les antigenes HLA, 1'etude de ce groupe de population unique de par ses caracteristiques genetiques semblables et ses facteurs environnementaux variables, pourrait permeftre de mieux cerner l'etiologie, la prevention et l'eventuelle guerison de cette maladie. La premiere etape consisterait a etablir dans les pays arabes des registres normalises pour le diab6te sucre insulino-dependant.

Acknowledgement This work supported by the National Institutes of Health (NIH grant No. DK-35905).

Dlabetes Epidemiology Research International Study Group

References 1. Technical Report Series 727, 1985. (Diabetes mellitus: report of a WHO Study Group). 2. Tlwarl, J.L. & Terasakl, P.l. HLA and disease associations. New York, Springer 1985. 3. LaPorte, R.E. & Cruickshanks, K.J. Incidence and risk factors for insulin-dependent diabetes mellitus. In: National Diabetes Data Group, ed. Diabetes in America. Washington, DC, Department of Health and Human Services, 1985 (NIH Publication No. 85-1468). 4. Dorman, J.S. & LaPorte, R.E. Mortality in insulindependent diabetes. In: National Diabetes Data Group. ed. Diabetes in America. Washington, DC, Department of Health and Human Services, 1985 (NIH Publication No. 85-1468). 5. Klein, R. & Klein, B.E.K. Vision disorders in diabetes. In: National Diabetes Data Group, ed. Diabetes in America. Washington, DC, Department of Health and Human Services, 1985. (NIH Publication No. 851468). 6. Leland, O.S. & Maki, P.C. Heart disease and diabetes mellitus. In: Marble, A. et al., ed. Joslin's Diabetes Mellitus. 12th edition. Philadelphia, Lea & Febiger, 1985. 7. Songer, T.J. The economic and social influence of insulin-dependent diabetes mellitus (IDDM): a casecontrol study. Diabetes, 36 (Suppl. 1): 19A (1987). 8. Rower, M. et al. Trends in the prevalence and incidence of diabetes: insulin-dependent diabetes mellitus in childhood. World health statistics quarterly, 41: 179-189 (1988). 9. Diabetes Epidemiology Research InternatIonal Group. Geographic patterns of childhood diabetes mellitus. Diabetes, 37: 1113-1119 (1988). 10. Slemiatyckl, J. et al. Incidence of IDDM in Montreal by ethnic group and by social class and comparisons with ethnic groups living elsewhere. Diabetes, 37: 1096-1102 (1988). 11. Diabetes Epidemiology Rosearch International Group. Secular trends in ten countries in the incidence of childhood insulin-dependent diabetes (IDDM). Diabetes, (In press). 12. Diabetes Epidemiology Research International. Preventing insulin-dependent diabetes mellitus: the environmental challenge. British medical journal, 295: 479-481 (1982). 13. Keys, A. Coronary heart disease in seven countries. New York, American Heart Association, 1970 (American Heart Association Monograph 29). 14. Doll, R.S. & Peto, R. The causes of cancer: quan-

236

15.

16. 17.

18.

19. 20.

21. 22.

23.

24.

25. 26.

titative estimates of avoidable risks of cancer in the United States today. New York, Oxford University Press, 1981 pp. 1205-1207. Encyclopaedia Britannica. Chicago, Encyclopaedia Britannica, Inc., 1984. Encyclopedia Americana International Edition, Danbury, CT, Grolier Inc., 1987. Stephens, R. The Arabs' new frontier. Colorado, Westview Press, 1976. Raymond, S. Health and policy-making in the Arab Middle East. Washington, DC, Center for Contemporary Arab Studies, Georgetown University, Press, 1978. Abu-Labon, B. & Abu-Labon, S.M. The Arab world: dynamics of development. Kinderhook, NJ, E.J. Brill 1986. Editorial. Diabetes mellitus and social class. Lancet, 2: 530-531 (1982). Taha, T. et al. Diabetes mellitus in Kuwait. Incidence in the first 29 years of life. Diabetologia, 25: 306-308 (1983). Elamin, A. et al. Prevalence of diabetes mellitus in school children in Khartoum, Sudan. Diabetes care, 12: 430-431 (1989). Hours, M. et al. Diab6te insulin-dependent juvenile. Etude descriptive dans le d6partement du Rh6ne. Revue d'Epid6miologie et Sant6 publique 32: 107112 (1984). Laron, Z. et al. The incidence of insulin-dependent diabetes mellitus in Israeli children and adolescents 0-20 years of age. A retrospective study, 1975-1980 Diabetes care, 8: 24-28 (1985). KadikI, 0. et al. Childhood diabetes mellitus in Benghazi (Libya). Journal of tropical pediatrics, 33: 136-139 (1987). KadikI, 0. et al. Diabetes mellitus in Benghazi. Journal of tropical medicine and hygiene, 91: 19-22

(1988).

27. Mercler, P. et al. HLA-A, B, DR antigens and insulindependent diabetes in Algerians. Tissue antigens, 26: 20-24 (1985). 28. Ben Behanl, K. et al. HLA associations in an Arab type 1 diabetic population. Disease markers, 5: 165169 (1987). 29. Ayed, K. H. et al. HLA-A, -B and DR antigens and complotype in Tunisian patients with diabetes mellitus. Disease markers, 7: 43-47 (1989). 30. Ayed, K. et al. HLA-A,B,C and DR antigens in a sample of the Tunisian population. Tissue antigens, 29: 225-231 (1987).

WHO Bulletin OMS. Vol 68 1990.

Childhood diabetes in Arab countries. Diabetes Epidemiology Research International Study Group.

Insulin-dependent diabetes mellitus (IDDM) is a chronic disease of childhood that is associated with high costs, mortality and morbidity, but which is...
972KB Sizes 0 Downloads 0 Views