Table 2 Leukemia deaths in men in England and Wales in 1998 by age groupThe value, 0.612 is for all ages 20-75 combined and is not correct for soldiers who are younger. Leukaemia rates increase markedly in people above 50 as you can see from the table and this would suggest a higher expected number of deaths if this large age group were used as a basis for any comparison. It is unlikely that there would have been many soldiers older than 40. Assuming an age range of 20-40 (which is conservative) there should be 0.15 deaths per 10,000 exposed per year (i.e. the death rate in the men aged 20-40 is about 116/7832822 = 1.48 E-5 which is  0.148 per 10,000 per year. So in the year since the bombing we should expect approximately 0.15 per 10,000 or 1.5 deaths in 100,000).

In January 2001, Nippon TV were told of there were  7 leukemia deaths in Italian Kosovan peacekeepers (assume 50,000) and more recently Eddie Goncalves, a journalist in Portugal, reported 5 deaths from leukemia in the Portuguese Kosovan peacekeepers (5 deaths in 10,000 with two in the 20-30 age group).  Thus in those groups we observe 12 leukaemia deaths where 0.9 are expected, a relative risk of 13. Even if we use a two-year period since the war the Relative Risk is still 6.5 However, there has recently been an official investigation of cancer in the Italian Balkan peacekeepers and the incidence of cancer in this group has been tabled and assessed. The analysis of the findings made by the Italian epidemiologists concluded that there was no significant increase in cancer but there was a confusing reference to a 'non-significant' increase in lymphoma. I want to look more closely at these results and the analysis. But first I will briefly examine the biologically likely consequences of DU exposure and develop a prior hypothesis.

The radiological consequence of DU exposure

The dissonance between the conventional (ICRP/Royal Society) analysis of the health effects of DU exposure and those advanced by independent researchers and the new models of the European Committee on Radiation Risk (ECRR) pivots on the type of exposure involved (Busby 2000, 2001)  The ICRP model assumes uniform and averaged deposition of ionisation in the whole body or organ, leading to a very low overall dose. The independent model argues that because of the micron diameter particulate nature of the DU, there will be high local and fractionated doses to cells close to the particles and no doses to the majority of cells more distant than the 30 micron range of the particles. This has two consequences. First, the cell dose is in the dose-squared region of the dose-response curve, and thus unrepairable chromosome double strand breaks are very likely. Second, the fractionation of doses in time, makes it highly probable that Second Event enhancement of mutation hazard will occur. These arguments are reviewed in Busby 2001.

                 It is of some interest, however, to consider which of the organs/ tissues of the body will be most likely to suffer damage from particulate DU and thus predict the cancer sites most likely to suffer increased risk. From the particle inhalation route we might expect irradiation of the lung and upper respiratory tract with possible increases in lung cancer and respiratory tract tumours and then following translocation to the lymphatics, lymphoma and to a lesser extent leukemia. From the ingestion route, it is possible that colon cancer might result, and excretion of the Uranium through the kidneys might result in increased incidence of kidney cancer. Increases in incidence of cancer in any or all of these sites would be consistent with exposure to particulate DU.

The Italian Report

Seconda Relazione Della Commissione Institiuta Dal Ministro Della Difesa Sull'

Incidenza di Neoplasie Maligne tra I Militari impiegati in Bosnia 28 Maggio 2001

David Coggon, reviewed the results of this report for the members of the UK DU Oversight board and since he has outlined the case made by the authors of the report, I give his comments below in full.

Coggon's review

I have now had a chance to look at a preliminary report of this study in English, dated 19 March 2001. 

The report describes a cohort study of Italian military personnel who were followed up from the date of their first mission to Bosnia or Kosovo until 31 January 2001.  Cancer incidence, ascertained from various sources including voluntary unsolicited reports, was compared with that in the male populations covered by Italian tumour registries. 

With allowance for a minimum latent interval of 12 months from first deployment, there were 18 cases of cancer in the cohort as compared with 32 expected.  These cases of cancer included one non-Hodgkin’s lymphoma (3.3 expected), six Hodgkin’s lymphoma (2.2 expected) and one acute lymphoplastic leukaemia (0.42 expected).

The study has two major but unavoidable limitations.  First, cases of cancer in the cohort were ascertained from different sources than for the comparison population.  This could give rise to bias in either direction.

Second, the follow-up of the cohort was extremely short.  This meant that there was greater statistical uncertainty because relatively few cancers had occurred.  Furthermore, if the cohort was exposed to a cancer hazard in Bosnia or Kosovo, it is possible that the effects of that exposure would not yet be manifest (because of the latent interval that often occurs between exposure to a carcinogen and the subsequent manifestation of the disease) 

Overall, the findings seem unremarkable.  The statistically significant deficit of deaths from cancer overall is probably in part because of a “healthy worker effect”.  Military personnel known to have cancer would not have been sent on missions, and would have been selectively excluded from the cohort.

I would like now to examine these results and suggest, contrary to Prof. Coggon's view that they show a highly significant excess risk from lymphomas and that the temporal spectrum of the disease suggests that this was a result of their periods of duty in the areas where DU had been used.  The statistically significant excess of lymphoma was recorded in Table 7 of the original Italian language version of the report which recorded 5 cases of non-Hodgkin lymphoma and 11 cases of Hodgkin's lymphoma, contrary to Prof Coggon's assertion above.

                 The results given in the Italian version of this report are given below in Table 3.

Table 3 Seconda Relazione Della Commissione Institiuta Dal Ministro Della Difesa Sull'   

Incidenza di Neoplasie Maligne tra I Militari impiegati in Bosnia 28 Maggio 2001: Table 5

Descrizione dei case accertati al 30/04/2001