Epidemiology relative risk

Ibuprofen is the most thoroughly researched 2-ary lpropionic acid. It is a relatively weak, non-selective inhibitor of COX. In epidemiological studies, ibuprofen compared to all other conventional NSAIDs, has the lowest relative risk of causing severe gastrointestinal side effects. Because of this, ibuprofen is the most frequently used OTC ( over the counter , sale available without prescription) analgesic. Ibuprofen is highly bound to plasma proteins and has a relatively short elimination half-life ( 2 h). It is mainly glucuronidated to inactive metabolites that are eliminated via the kidney. 

Again, in epidemiological studies or clinical trials there is nearly always a degree of uncertainty due to bias, chance and confounders. In these studies uncertainty is measured in terms of / -values, odd ratios, and relative risks, and so on. 

In Chapter 6, in the section on epidemiology, risks were presented in relative terms, the ratio of the risk observed in one population to that observed in another. Such relative risks are important and useful, but are expressed differently from the absolute risks discussed here. 

Three additional points need to be mentioned. First, if the observed cancer dose-response relationship derives from epidemiology data, the observed risks are relative, not absolute (the latter are usually reserved for data from animal experiments). Thus, for human carcinogens with reliable dose-response information (e.g., as exists for benzene, arsenic, chromium (+6), asbestos, and several other carcinogens), it is necessary to convert relative risks to absolute risks before extrapolating to low dose. 

Let us assume that enough information is available regarding the levels of benzene in Mr. Z s well, the number of years he consumed the water, and even his water consumption rate, to derive a reasonably accurate estimate of his cumulative exposure from this source. The epidemiologists and biostatisticians carefully evaluate the dose-response data from the published epidemiology studies used as the basis for classifying benzene as a cause of leukemia. Further assume that we learn from this evaluation that Mr. Z incurred a cumulative benzene exposure approximately equivalent to the cumulative exposure that was found to cause a three-fold excess risk of leukemia in the occupational studies of benzene exposure. A relative risk of three. 

The lARC has concluded that epidemiological studies have established the relationship between benzene exposure and the development of acute myelogenous leukemia and that there is sufficient evidence that benzene is carcinogenic to humans. Although a benzene-leukemia association has been made, the exact shape of the dose-response curve and/or the existence of a threshold for the response is unknown and has been the source of speculation and controversy. Some risk assessments suggest exponential increases in relative risk (of leukemias) with increasing cumulative exposure to benzene. At low levels of exposure, however, a small increase in leukemia mortality cannot be distinguished from a no-risk situation. In addition to cumulative dose other factors such as multiple solvent exposure, familial connection, and individual sus-... 

An essential importance for the development of obstructive respiratory diseases, within the scope of cancer mortality of smokers, was indicated by epidemiological studies. It was shown that the relative risk for smokers, with obstructive ventilation parameters [FEV 1% < 60 (Melv)m et al., 1987), respectively 70] (Skillud et al., 1987), to be affected by lung cancer, is significantly higher than that of comparative groups with normal lung-function parameters. 

A prospective study of 17,000 people suggested that the mean relative risk of developing type II diabetes was only 0.5 (0.35-0.72) in those individuals habitually consuming six or more cups of coffee per day compared with those consuming two or less (p = 0.0002). The results of subsequent epidemiological studies on coffee consumption have been in good agreement. 

Several NSAIDs (including aspirin) appear to reduce the incidence of colon cancer when taken chronically. Several large epidemiologic studies have shown a 50% reduction in relative risk when the drugs are taken for 5 years or longer. The mechanism for this protective effect is... 

A major food safety issue addressed by the European Union concerns the increase in the level of circulating insulin-like growth factor (IGF-I) in the target animal and its increased excretion in the milk as a consequence of the administration of rBST. There has been epidemiological evidence for an association between circulating IGF-I levels and the relative risk of breast and prostate cancer (34-36). 

Wollin and Jones (2001) investigated the effects and mechanisms of action of consumption of red wine compared to other alcoholic beverages on the risk of cardiovascular disease. Of particular interest was the form and quantity of alcohol consumed. The relationship between alcohol consumption and mortality is supported by epidemiologic studies suggesting that different forms of alcohol alter the relative risk for mortality. Evidence... 

A meta-analysis of epidemiological studies of ovarian cancer showed a summary estimated relative risk of 0.64 for ever-use of combined oral contraceptives, implying a 36% reduction in ovarian cancer risk (130). This protective effect increased with increasing duration of oral contraceptive use and continued for at least 10 years after discontinuation. Although most of the oral contraceptives reported in these studies were older, higher-dose formulations, the Cancer and Steroid Hormone (CASH) study included users of tablets containing ethinylestradiol 35 pg or less, and this subgroup of women had a reduced risk of ovarian cancer (115). 

Sartwell PE, et al. Oral contraceptives and relative risk of death from venous and pulmonary thromboembolism in the United States an epidemiologic case-control study. Am J Epidemiol 1969 90 365. 

A difficulty with epidemiologic studies at low doses is the inability to adequately control for potentially confounding factors to the extent necessary to exclude spurious observations, either positive or negative. Epidemiologic studies are not capable of detecting increased responses unless the excess relative risk is on the order of 30 to 40 percent or higher. 

A large number of epidemiology and case-control studies have examined the potential association between oral aluminum exposure and Alzheimer s disease. A number of these studies have been criticized for flawed patient selection, poor comparability of exposed and control groups, poor exposure assessment, poor assessment of health outcomes, and weak statistical correlations (Nieboer et al. 1995 Schupf et al. 1989). Studies conducted by Martyn et al. (1989), McLachlan et al. (1996), and Michel et al. (1990) have found an association between oral exposure to aluminum and an increased risk of Alzheimer s disease. In a survey study conducted by Martyn et al. (1989), the incidence of Alzheimer s disease in individuals under the age of 70 was estimated from computerized tomographic (CT) records. The 1,203 subjects lived in 88 county districts within England and Wales. Data on aluminum concentrations in the municipal water over a 10-year period were obtained from water authorities and water companies. The subjects were classified as having probable Alzheimer s disease, possible Alzheimer s disease, other causes of dementia, or epilepsy. The relative risks of Alzheimer s disease were elevated in the subjects living in districts with aluminum water concentrations of >0.01 mg/L. However, the relative risk exceeded unity only in the subjects with aluminum water concentrations of >0.11 mg/L (relative risk of 1.5, 95% confidence interval of 1.1-2.2). 

After a study has been conducted, the data are analyzed to see whether an association is observed. This section briefly defines the two major measures of occurrence (incidence and prevalence) and discusses measures of associations in detail, focusing on the most common measures used in the epidemiological study designs described above, such as correlation coefficient, relative risk, and odds ratio. 

Because case-control and cohort studies are the most commonly used studies in epidemiology, the remaining subsections will focus on these designs. Once an association between exposure and disease has been calculated, the next step is to evaluate the study to determine whether the calculated relative risk could result from error. The error could be random (due to chance), systematic (resulting from the study design) (discussed in Section 26.2.4), or due to confounding (discussed in Section 26.2.5). 

---