An epidemiological study of multiple myeloma in Southern Africa

Abstract

Globally, there are both striking geographic and ethnic differences in myeloma incidence. Myeloma is twice as common in American blacks compared to Caucasians. The ageadjusted incidence rates for Caucasians is 4.1/100 000 and for blacks is 9.1/100 000 (Ries et al. 1994). The higher rate of myeloma in blacks is not confined to the USA, but may also include other regions of the world. Blattner et al (1979), found that a similarly high rate of myeloma was present in South African blacks. Despite having made this important observation two decades ago, there have been no epidemiological studies in South Africa, and indeed in Africa, that have explored the role of the potential and suspected risk factors that are implicated in the aetiopathogenesis of the disease. This thesis was designed to address this deficiency and fill the gap that exists.This thesis examines the role of some of the suspected risk factors in the aetiopathogenesis of myeloma, in a population of patients in whom myeloma is the commonest iymphohaematopoietic malignancy. The influence of occupational,sociodemographic and other characteristics were evaluated using a case-control study.Environmental factors such as Kaposi’s sarcoma associated herpesvirus/human herpesvirus-8 (KSHV/HHV-8) were indirectly evaluated by detecting the presence of KSHV DNA sequences in bone marrow aspirates, bone marrow biopsy material and/or cultured bone marrow adherent cells, using a nested PGR (polymerase chain reaction) assay. In addition, class ......HLA (Human Leucocyte Antigen) types were identified in the myeloma patients and compared to an ethnically matched control population. Finally, the flow cytometric characteristics of myelomatous plasma cells including DNA pioidy analysis and relevant plasma cell antigen expression were determined. The clinical profile of the background myeloma population in whom these studies were conducted was documented. This, together with the relevant prognostic factors is presented and compared to the findings in Africa and the Western world.A total of 170 patients with myeloma were seen from January 1992 to December 1997.This constitutes the background myeloma population in this study. Initially, two aspects of the study were undertaken, viz., the casa-control study and the flow cytometry study.Interviews were possible in 130 of the 170 patients. The other 40 patients were not interviewed for primarily two reasons. Firstly, a number of patients died during the initial admission and secondly, patients who were not interviewed during the first admission did not return for follow up (were lost to follow up). Similarly, flow cytometry was possible in 103 of the 170 patients. The reduced numbers were due to the following reasons: i) some natients had bone marrow aspirates performed at a peripheral hospital or by a private physician prior to being referred to our hospital and the bone marrow was not repeated by us and ii) for technical reasons a flow cytometric analysis was not possible - insufficient or grossly haemodilute specimen, and in some instances there was a dry tap. The HLA study was commenced in 1995, by which time a number of the original patients had died and some were lost to follow up. Sixty two patients had HLA studies performed (57 of whom were part of the 170 patients). Towards the latter part of this ‘epidemiological study’, important information regarding the role of KSHV and myeloma emerged. A brief analysis regarding this association was deemed pertinent and was included prior to concluding the thesis. Thus, the numbers represented in this aspect of the study are much smaller (25 of the 170 patients).The me dian age at presentation of 61.4 years in our patients is intermediate between the higher figures quoted for patients in the Western world, and the lower median age noted for patients from Africa. The clinical profile is essentially similar to that described for myeloma elsewhere in the world, except that the majority of our patients present with advanced stage disease (stage IE). In general, the symptoms and signs are overt and exaggerated when compared to the Western world and occur in a similarly high proportion of patients as documented in other African series. The laboratory characteristics are also typical of that described. Other differences regarding the clinical and laboratory profile which have been noted in our series are highlighted.All the patients were treated with conventional chemotherapy. The standard chemotherapy regimen used initially in the vast majority of patients was melphalan and prednisone. For refractory disease, combination intravenous chemotherapy such as vincristine and an anthracycline together with dexamethasone was used in most patients.None of the patients were subjected to peripheral stem cell or bone marrow transplantation. Poor prognostic factors of statistical significance that were found in the study following a univariate analysis include: a beta 2 microglobulin of >6mg/l (both uncorrected and corrected for renal impairment), hypercalcaemia >2.65mmol/l, urea >8mmol/l, creatinine >180pmol/l, CRP >12mg/l, platelets <100 x 109/1, serum paraprotein >10g/l, intermediate and advanced stage disease (based on the Durie and Salmon classification), urine total protein >lg/l and a white cell count of <4 x 109/1 (especially <3.4 x 109/1).The case-controi study showed a significantly increased risk of myeloma in agriculture (farming), but we were unable to determine the exact nature of the factor responsible for the increased risk. A significantly higher number of cases (71.5%) compared to controls (51.5%) spent time at a farm (p=0.0009; odds ratio 2.36). More of the cases (patients) spent time as the owners or workers on a farm, compared to the controls (p=0.008). The number of cases (66 - 50.7%), who worked on a farm was significantly higher than the control group (28 - 21.5%) (p-0.00005; odds ratio 3.68). There was no significant difference between the cases and controls with regard to animal exposure (p=0.635), pesticide exposure (p=0.995) or exposure to exhaust fumes (p=0.945), although in each of these categories the exposures were more in the cases than controls.Radiation did not feature as a risk factor in our patients. With respect to benzene, lead, asbestos, mining, Sbreglass/mineral fibres, working at a building construction site, work as a caterer/cook and in the forestry industry, exposure was higher in the cases compared to controls, but the difference was not statistically significant. Regarding non-occupational exposures, there was no association between smoking or alcohol intake and myeloma. Intake of herbal toxins, ibuprofen and laxatives was higher in the cases than the controls, but the difference was fiot significant (NS). Allergies were higher in the cases compared to controls (NS), as were certain childhood infections such as measles and mumps. Interestingly, a significantly higher number of controls (123 - 94.6%) were immunized (against the common childhood illnesses based on the government immunisation schedule) compared to cases (104 - 80%) (p=0.0019). Thus, immunization may play a protective role against myeloma.The role of viruses was not specifically examined in the case-control study. However,there were only two patients of the 130 cases and no additional cases in the background 170 patients with myeloma who were HLV seropositive. The seropositive rate appears to be lower than in an aged matched hospital population without myeloma. In the KSHV/HHV-8 study, KSHV DNA sequences were detected in 4/10 (40%) of the adherent cell cultures and 1/20 (5%) of the bone marrow aspirate samples. None of the bone marrow biopsy samples (0/9), or control bone marrow aspirate samples (0/19) were positive. Based on the small sample, a similar background seroprevalence rate compared to the detection rate, as well as the conflicting data in the literature, the exact «ole of KSHV/HHV-8 in the aetiopathogenesis of myeloma stills remains to be elucidated and clarified.With respect to socioeconomic factors, there were no significant differences noted in relation to home ownership, income, type of home, educational status and occupational rank (social class) between the cases and controls. Thus, socioeconomic status does not appear to have an influence on myeloma risk in our patients.Analysis of class I and class II human leucocyte antigens in 62 myeloma patients revealed a corresponding association of statistical significance (to that documented in the literature) only with regard to HLA BIS. An additional association not documented previously was HLA B35. The relative risk for myeloma in individuals with HLAB35 is 11.82. Haplotype frequencies were also studied. Haplotypes that confer a relative higher risk o f myeloma include A30,DR8; A43,B22; A43,B35; B70,Cw3 and DR11,DQ3.Those associated with a lower risk (i.e. may be protective) are A2,B42; A30,B70.Thn results of the DNA ploidy analysis revealed that 63.1% o f the patients had a diploid pattern, while 36.9% o f the patients demonstrated aneuploidy. Hyperdiploidy accounted for 95% of the aneuploid population, while hypodiploidy was present in only 5% of the aneuploid group. The percentage of patients with aneuploidy is very variable and appears to be generally lower in our patients. When comparing the aneuploid and diploid groups,the mean plasma cell numbers were found to be significantly higher in the aneuploid group (p=0.0005). Additional factors of statistical significance between the two groups were a higher paraprotein level, IgG level and abnormal cytogenetics in the aneuploid group.Plasma cell antigen co-expression of CD38-56, CDS8-33, CD38-10 and CD38-45 were specifically studied. Of the four antigens, only CD38-56 expression was higher in myelomatous plasma cells. This is consistent with the reported bone marrow myeloma cell phenotype of CD56-H-. When the CD38-56 expression was correlated with the prognostic factors noted in this study, only stage of disease was of statistical significance (p=0.0013), i.e. more patients with CD38-56 expression had advanced stage (stage m disease). Furthermore, a shorter survival was noted in the CD38-56 positive group (24 months compared to the 37 months in the negative group), although the difference was not statistically significant (p=0.1599).