Exposure Bias Research Articles

Hepatitis B Vaccine Is Not Associated with Liver Problems in U.S. Children Less Than 6 Years Old

To the Editor: Fisher and Eklund question the problem of external validity in previous studies in high-risk populations by evaluating the benefit of hepatitis B vaccination for the general population of American children. 1 The authors claim that their data do not support a protective effect of HB vaccine against liver problems for U.S. children; to the contrary, they report on liver problems among U.S. children that they consider associated with hepatitis B vaccination. The data presented, however, do not support the suggested associations and do not support the authors’ conclusions. First, the ICD-9 codes that Fisher and Ecklund used do not allow any conclusion on the protective effect of HB vaccine in U.S. children (used ICD-9 codes: acute and sub-acute necrosis of liver, chronic passive congestion of liver, unspecified fetal and neonatal jaundice, unspecified jaundice (not of new-born) and biliuria, each excluding any instances of viral hepatitis). The hepatitis B vaccine aims to protect against HB infection (including its chronic sequelae), not to prevent against these non-infectious acute or sub-acute pathologies. Second, some of these ICD-9 pathologies could have been already present before or at the moment of administration of the HB vaccine (eg, neonatal jaundice). Third, the authors did not consider the possibility of an exposure bias: in many industrialized countries it is good clinical practice to vaccinate children with pre-existing liver problems against hepatitis B to prevent further deterioration of the liver. The associations calculated in this paper show serious methodological limitations, and the results give no information on causality between hepatitis B vaccine and liver problems in American children. At several occasions in this paper, the authors consider the hepatitis B infection risk negligible for 99.5% of U.S. children. Indeed, in the U.S., each year 20,000 neonates are born to hepatitis B surface antigen positive mothers. In the United States, however, the lifetime risk to contract hepatitis B has been estimated at approximately 5%. 2 Infant immunization programs have not been set up to protect only infants against perinatal hepatitis B infections, but to offer a cohort of infants life-long protection against an infection they could contract from birth to death. Indeed, in the United States, HBV infection occurs mainly in adolescents and young adults. 3 The age distribution of HBV infections is indicative of a life-style disease linked with at risk behavior, such as sexual activity and injecting drug use. Moreover, part of the aim of the infant immunization programs is to vaccinate a higher proportion of the population than could be reached by targeting a teenage population, and to control the infection in the general population through an increasing herd immunity. In contrast with the authors’ statement, several trials have been conducted in low risk populations with a well-documented immunogenicity and safety profile of these vaccines, 4,5 and currently implemented universal hepatitis B immunization programs in low endemic countries have been shown effective in controlling HBV infection. 6,7 Fisher and Eklund are unclear on the eligibility criteria of the study (21.9% of the children eliminated) and on the completeness of the vaccination course. Does the classification “HB vaccinated” mean having been administered the full course? In addition, the use of parent’s memory to recall vaccination status of their children introduces the likelihood of recall bias. This possibility is reinforced by the lack of clarity over the period of time elapsed between immunization and questionnaire survey. In addition, the help-seeking behavior in the exposed group could be different from that of the un-exposed group, a difference that could introduce a selection bias. Parents who did not have their children participate in a universal HB vaccination program could be less willing to seek any kind of medical advice for their children. A number of liver problems could therefore remain unnoticed or undiagnosed, causing an under-registration of these disorders in the non-vaccinated group. Thus, the internal validity of this study leaves much to be desired. The methodology used does not permit a sound conclusion on the causality between hepatitis B vaccines and liver problems in American children. A prospective cohort study would have been more appropriate to the task. Pierre van Damme Koen van Herck André Meheus

Epidemiologic studies of fatal and nonfatal cardiovascular disease and ETS exposure from spousal smoking.

This article reviews the epidemiologic studies of the association of ischemic heart disease risk and environmental tobacco smoke (ETS) exposure from a spouse who smokes. Seventeen studies (nine cohort, eight case-control) comprising more than 485,000 lifelong nonsmokers and 7,345 coronary heart disease (CHD) events were included in a meta-analysis. Together, these studies include 36% more CHD events and 58% more study subjects than were available for review by the U. S. Occupational Safety and Health Administration (OSHA) in 1994. The relative risk (RR) for fatal or nonfatal coronary events among never smokers married to smokers, compared to those whose spouses did not smoke, was RR = 1.25 (95% confidence interval [95% CI], 1.17-1.33) across the combined studies. This association was statistically similar in men (RR = 1.24; 95% CI, 1.15-1.32) and women (RR = 1.23; 95% CI, 1.15-1.32); in studies of cohort (RR = 1.23; 95% CI, 1.15-1. 31) and case-control (RR = 1.47; 95% CI, 1.19-1.81) design; in the United States (RR =1.22; 95% CI, 1.13-1.30) and other countries (RR = 1.41; 95% CI, 1.21-1.65); and in studies of fatal (RR = 1.22; 95% CI, 1.14-1.30) and nonfatal (RR = 1.32; 95% CI, 1.04-1.67) heart disease. In three studies that presented data separately for nonsmokers married to current or former smokers, the association was stronger when the spouses continued to smoke (RR = 1.16, 1.06-1.28) than with former smokers (RR = 0.98; 95% CI, 0.89-1.08). The aggregate data are unlikely to be attributable to chance, publication bias, confounding, or misclassification of exposure. The evidence linking heart disease and ETS exposure from a spouse has become substantially stronger since OSHA first proposed including heart disease in its risk assessment of ETS in 1994.

Do Power Frequency Magnetic Fields Cause Leukemia in Children?

The possibility of a causal relationship between exposure to electromagnetic fields and childhood leukemia has been studied for more than 10 years. This review evaluates the epidemiologic findings relevant to a possible causal association between exposure to power frequency magnetic fields during childhood and leukemia. The nine published studies on this topic were analyzed, with a focus on the selection of subjects, assessment of exposure, and control of confounding. Consistency of a positive association and dose-response relationship are evident for assessment of past exposure: among the five recent studies, relative risk estimates vary from 1.5 to 2.7 for past exposure assessment, and a significant dose-response relation was found in three studies. These observations cannot be explained by misclassification of exposure or consistent selection bias, and a confounder effect could hardly explain such results. Even if there is no currently understood biological plausibility for such an association, its possible causal nature cannot be dismissed. The impact on public health of such a possible causal association is difficult to assess precisely but could be significant. Therefore, it might be prudent to consider interim preventive measures while we await results of further research.

Epidemiology of silicone-related disease

The epidemiology of silicone-related disease (SRD) is complicated by the variety of disease endpoints that have been associated with silicone exposure and the atypical nature of these diseases in silicone-exposed women. Current research reviewed here suggests that SRD may constitute a new disease entity, thus complicating disease definition and rendering studies of classic disease unlikely to detect risks of silicone exposure. This report addresses the most important study design issues (disease and exposure definitions, bias, confounding, and power) in the context of studies of SRD. The variety of silicones used complicates the definition of exposure for all studies, and for some populations simply determining who was implanted will be difficult. For any of these studies, inadequate patient follow-up is likely to underestimate disease risk. Studies of SRD are also complicated by confounding. That is, whether or not a woman chooses to receive an implant is related to her age, race, and other variables also related to rheumatoid and autoimmune disease. The absence of an appropriate control group also plagues published studies of silicone-related disease. Finally, inadequate sample size, resulting in studies of low statistical power, is a critical problem for rare diseases such as SRD. These points are illustrated using two published studies and five studies in progress.

Formaldehyde and the risk of squamous cell carcinoma of the sinonasal cavities.

A study of 759 histologically verified cancers of the nasal cavity (287 cases), paranasal sinuses (179 cases), and nasopharynx (293 cases) and 2465 cancer controls diagnosed in Denmark between 1970 and 1982 was conducted to investigate the importance of occupational exposure to formaldehyde. Information on job history for cases and controls was derived from a national data linkage system and exposure to formaldehyde and wood dust was assessed by industrial hygienists unaware of the case-control status of the patients. The exposure rates for formaldehyde among male and female controls were 4.2% and 0.1%, respectively. After proper adjustment for contemporary wood dust exposure, relative risks of 2.3 (95% CI = 0.9-5.8) for squamous cell carcinoma and 2.2 (95% CI = 0.7-7.2) for adenocarcinoma of the nasal cavity and paranasal sinuses were detected among men who have ever been exposed to formaldehyde in their job compared with those never exposed. The introduction of a 10 year latency period did not change the risk estimates substantially. It was considered unlikely that the results were due to bias or misclassification of exposure although the effect of chance could not be excluded.

Proximity effect reduction in high voltage electron-beam lithography by bias exposure method

The proximity effect reduction was achieved by electron beam (EB) uniform exposure (bias exposure) on the target covered with PMMA resist of 1 μm in addition to conventional high voltage EB writing. At an acceleration voltage of 50 kV and a dosage of 50 μC/cm2, a pattern size error of 0.5 μm pattern due to the proximity effect amounted to 0.25 μm, which was improved to 0.12 μm by bias exposure. When dosage was large, the pattern size error was further improved by bias exposure. For 50 kV and 140 μC/cm2 writing, 0.25 μm pattern including isolated window, line and window, and isolated line was formed with an accuracy of ±0.025 μm. The pattern size error for conventional EB writing was almost independent of dosage. For 20 kV and 20 μC/cm2 writing, the pattern size error was hardly improved by bias exposure.

Effect of the threshold detection property of photographic films on hologram linearity and efficiency

Optimum linearity, and maximum efficiency, for thin-amplitude and thin-phase holograms of small modulation depth are realized at different values of average exposure (bias). We utilize a threshold model for the photographic detection process to predict the bias values. These values are shown to be in agreement with available experimental results. For large modulation, the biasing point depends upon both modulation depth and average exposure.