Bollettino epidemiologico nazionale

Radio Vatican and Leukemia: The Difficulty of Decision-Making in the Presence of Weak Epidemiologic Evidence

Radio Vatican Study Group: Dr. Donato Greco, Director of the Laboratory of Epidemiology and Biostatistics, Istituto Superiore di Sanità, Rome, Professor Peter Boyle, Chairman, Prevention and Control, Imperial Cancer Research Fund, London, Giuseppe Masera, Director of the Pediatric Clinic of Monza Hospital, University of Milan, and Professor Roland Mertelsmann, Chief, Department of Hematology of the University of Freiburg

In the 1950s, a large radio installation (Radio Vatican) was built in the north-east periphery of Rome that contained several transmitters varying in power from 50 to 500kw that transmit radio broadcasts throughout the world. The installation covers an area of approximately 450 hectares in the area of Cesano. In the years following construction of the transmission installation, housing was constructed in the immediate vicinity. In 1999, in the area included within a 10 km radius of the installation, there are approximately 60,000 persons, including almost 10,000 children 0-14 years of age.

In recent years, the populations living near the transmission site have complained that transmissions of Radio Vatican interfere with domestic electrical devices (intercoms, radios, etc). In addition, a local physician reported a suspect excess of cases of childhood leukemia, which was attributed to the radio installation. In response to the growing preoccupation of the local population, the Local Health Agency Rome E involved the Regional Epidemiologic Observatory (OER) of the Lazio Region. Between 1996 and 2001, descriptive ecologic studies of mortality and incidence of leukemia were conducted for the implicated area [1,2]. Reports were issued that suggested an excess incidence of childhood leukemia in the area compared with the Commune of Rome. Furthermore, it was suggested that there was a decreasing gradient with distance from the installation.

In April 2001, the Italian Minister of Health convened a group to examine the relationship between radiofrequency waves and childhood leukemia, with specific reference to this highly publicized cluster. The group, which was comprised of experts within and outside Italy, reviewed available scientific literature on radiofrequency electromagnetic fields and their effects on health and on the etiology of childhood leukemia; they also examined the extensive epidemiologic studies conducted by the Regional Observatory of Lazio and performed a secondary analysis of the existing data. The following report summarizes the main findings regarding a possible association. A full report will be published in the November 2001 issue of ISTISTAN.

Epidemiologic findings: The OER analysis based on the Regional Childhood Leukemia Register the Commune of Rome for 1987-1999 showed no changes in leukemia incidence over time. The standardized incidence rate (SIR) of childhood leukemia for the 13-year period in the area within a 10 km radius of the Radio Vatican installation was calculated using the entire Commune of Rome to obtain expected values. No statistically significant increase in incidence was observed (SIR = 1.22; 95% confidence intervals = 0.56-2.27).

The OER identified 8 children with childhood leukemia in the study area between 1987 and 1999 in children 0-14 years of age and living in the area for at least 6 months prior to their leukemia diagnosis. Seven had acute lymphocytic leukemia (ALL); one had acute myelogenous leukemia (AML). Figure 1 shows the temporal trends by sex and age. Of the eight cases, six had resided in the area since birth and two were born elsewhere but moved to the area later (a 12 year old with AML and a five-year old with ALL).

In the OER analyses, leukemia rates were calculated for the five concentric rings, each of which was 2 km wide (respectively at 2, 4, 6, 8, and 10 km from the installation). The pediatric population living in each ring was estimated, and the cases were considered to belong to that ring if they were living there at the time of diagnosis. Ninety-five percent confidence intervals around the SIR calculated for each ring consistently overlapped 1.0. However, the Stone test for homogeneity showed a statistically significant gradient of incidence with distance from the center of the transmission installation. Furthermore, one case of ALL was diagnosed in 1993 in a 9 year old whose census tract of residence fell within the first ring; the expected incidence for the 13-year period was 1.6. This resulted in national media stories stating that those living near the Radio Vatican installation were at a six-fold increased risk of childhood leukemia and a public outcry that the installation should be closed or its transmissions drastically reduced.

Further examination of the data by the Commission revealed that the case in the first ring actually resided in more than 2 km from the installation, even though the most of his census tract of residence fell within the first ring. In addition, the Commission observed that the Stone test method used to analyze spatial risk lacks precision when applied to rings without observed cases (3).  Therefore, a reanalysis was performed in which the cases were regrouped such that there was at least one in every ring. In addition, the Commission used distance from the center of the Vatican Radio site and the actual residence, rather than census tract of residence, to more accurately assess a possible association (Figure 2). The population denominators for each ring were based on census tract rather than actual distances. For the three concentric rings evaluated (0-4km, 2 cases; 4-6 km, 4 cases; and 6-10 km, 2 cases) the SIRs were 1.96, 1.50, and 1.69 , respectively, and was 1.20 overall (Stone test p= 0.16).

Comment: Leukemia is a relatively rare disease, affecting an estimated 40-45 children per million population (0-14 years of age) annually. In Italy, approximately 450 cases are diagnosed per year. Of these, approximately 75% are acute lymphoblastic leukemia (ALL), 20% are various forms of acute myoblastic leukemia (AML), and approximately 4% are chronic forms of leukemia. The disease most frequently occurs in the 2-5 year age group.

The word leukemia aggregates a number of diseases that differ according to their etiologic mechanism and population target such that it is not appropriate to aggregate these diagnoses into a unique category, especially when looking for etiologic factors. Different forms of leukemia appear to have different etiologies and are most likely multifactorial, with no single factor capable by itself of causing the disease. In childhood leukemia, especially ALL, there is convincing evidence of the existence of a leukemogenic process that begins during fetal development: this underlines the importance of genetic risk factors and environmental and behavioral exposures of the mother.

According to the World Health Organization in a report summarizing the effects of electromagnetic fields on health (4), approximately 25,000 studies have been performed on the health effects of radiofrequency (RF) radiation. RF (300 Hz - 300 GHz) are a form of non-ionizing radiation that is used in radio and television transmission, telecommunication, in many industrial processes involving heating and soldering, and in various diagnostic and therapeutic medical devices. RFs are capable of producing biologic effects that depend on the frequency and intensity of the field.  These biologic effects, however, do not necessarily translate into adverse health effects. The biologic effects are dependent upon frequency and are largely the results of heating produced by these waves. At high levels of exposure, unlikely to be experienced under normal living and working circumstances (eg, close proximity to powerful radar or at the top of a powerful radio transmitter), cataracts and skin burns may occur; in addition, high exposures may increase local or body temperatures. This induced heating can affect fetal development (although teratogenic effects are seen only if the temperature of the fetus increases 2-3 °C). These latter effects, however, occur following acute exposures to high levels of RF which are not encountered in daily life.

Studies on animals, cell lines, and other biologic models support an absence of a significant cancerogenic effect of electromagnetic fields and radio frequencies, such that it is highly unlikely that an effect exists in man (11-16). Indeed, six studies are available in the literature on the mortality or incidence of childhood tumors in residents living in proximity to radio and television repeaters [29-34].  These studies, which have been conducted in various sites in the United States, Great Britain, and Australia, have failed to demonstrate spatial clusters of childhood leukemia associated with distance from radio and television transmitters.

Ecologic studies, such as those conducted in the Cesano area, have limited capacity to produce useful insights into the association between electromagnetic fields and leukemia, especially when precise data on individual exposure are lacking. Particularly problematic in the study was the measure of exposure, which is complicated by the use of surrogate measures, which are likely to be inaccurate because of such factors as changing patterns of transmission over time, changes in RF used, the effect of geographic barriers and buildings on RF transmission.

Research is clearly needed into better methods of measuring exposure to RF. In the current situation, it is nonetheless important not to dismiss the public’s concern, even if the risk is only potential. Instead, it is important not to focus on unknown risks at the cost of losing track of risk factors known to affect the health of children. The possibility of avoiding even a few cases of leukaemia, especially in children is a moral responsibility in our society.

To better study such issues in the future, consideration should be given to the creation of a National Cancer Registry based on existing Cancer Registries and the development of regional registries where none currently exist and the establishment of a national small-area statistics unit which would work closely with the data from the National Cancer Registry and have a number of functions including monitoring the geographic pattern of cancer in Italy and investigating suspected clusters nationally and on a routine basis.  Furthermore, research is needed to improve methods for accurately measuring the exposure to RF.