Notes Washed Up in a Bottle

Don. W. MacCorquodale M.D. M.S.P.H.

Periodic Notes From the Field on Philosophy and Science.

REGRESSION TO THE MEAN

George W. Comstock reported on a 1934 study of blood pressure in U. S. Army officers in his “Snippets from the Past:   70 years ago in a recent edition of the American Journal of Epidemiology   (Am J Epidemiology 2005; 161:   38-39).   The author of the 1934 study reported that “in the age group 20-24 years, men with high blood pressures initially tend to have lower blood pressures subsequently, and vice versa.”   Dr. Comstock observed, “Strangely, regression to the mean is not mentioned!”

          COMMENT:   The statistician, Karl Pearson, collected over a thousand records of heights of members of family groups, and he reported his findings in 1903.   He showed that although tall fathers tended to have tall sons, the average height of sons of a group of tall fathers was less than their fathers' height.   The phenomenon became known as regression to the mean.

          I am not sure that regression to the mean has been investigated as thoroughly as it should have been.   I remember a single paper published about twenty years ago in which the authors showed that repeated serum cholesterol determinations on a group of men showed striking evidence of regression to the mean.   Those men with initially high values tended to have lower values when a second determination was done, and vice versa.   If these findings were corroborated by repeated studies, we might be wise to do repeated serum cholesterol determinations, rather than to make decisions based on a single laboratory examination.

RADON AND LUNG CANCER

 A group of European investigators conducted a “collaborative analysis” of individual data from 13 case-control studies of residential radon and lung cancer carried out in nine European countries (Darby S et al.   Radon in homes and risk of lung cancer:   collaborative analysis of individual data from 13 European case-control studies.   British Medical Journal, December 21, 2004 ).   The mean measured radon concentration in the homes of people in the control group was 97 becquerels/m 3 .   For cases of lung cancer the mean concentration was 104Bq/m 3 .   The risk of lung cancer increased by 8.4% (95% CI:   3.0%-15.8%) per 100 Bq/m 3 increase in usual radon, that is, “after correction for the dilution caused by random uncertainties in measuring radon concentrations.”   In the absence of other causes of death, the absolute risk of lung cancer by the age of 75 years at radon concentrations of 0, 100, and 400 Bq/m 3 would be about 0.4%, 0.5%, and 0.7% for life-long nonsmokers and about 25 times greater (10%, 12%, and 16%) for cigarette smokers.

          COMMENT:   When a young woman married a miner who worked in one of the mines of Schneeberg in what is now Germany in the 16 th century, she knew she would be a widow at an early age.   George Bauer a.k.a. Georgius Agricola mentioned a woman of Schneeberg, who had been widowed nine times in his De Re Metallica published a year after his death in 1555.   The miners died of an unidentified lung disease, which was identified as lung cancer three centuries after the publication of Bauer's work, and it has since been shown that lung cancer in the miners was caused by exposure to radon and radon daughters.  

          Miners who worked in the uranium mines in the American Southwest in recent years have also suffered from excess rates of lung cancer, and again, it has been clearly shown that inhalation of radon and radon daughters is the cause of the disease.   The levels of radon in those mines are two to five times higher than those detected in American homes

          Is exposure to radon in the home a cause of lung cancer?   Perhaps, but the lesson from this study is clear.   If you have radon in your home and you're concerned about lung cancer, don't smoke cigarettes.

THYROID NEOPLASIA AND EXPOSURE TO IODINE 131

 The Hanford Nuclear Site in southeastern Washington states was established in 1943 to produced plutonium for atomic weapons.   In 1986 it was revealed that large amounts of radionuclides were released into the atmosphere.   From 1944 through 1957 2.73 x 10 36 Bq of iodine 131 were discharged into the atmosphere (Davis S et al.   Thyroid neoplasia, autoimmune thyroiditis, and hypothyroidism in persons exposed to iodine 131 from the Hanford Nuclear Site.   JAMA 2005 292:   26002613).  

          Ionizing radiation from external sources has been shown to be associated with increased risk for thyroid neoplasia.   There is however no convincing evidence that persons exposed to diagnostic or therapeutic doses of 131 I are at increased risk of thyroid tumors.   Environmental exposures to 131I have been associated with:   1) excess thyroid nodules among Marshall Island residents exposed to fallout from testing of nuclear weapons, 2) excess thyroid nodules among men and women in Utah to fallout from testing of nuclear weapons, and 3) excess rates of thyroid cancer in children exposed from the Chernobyl accident.  

          The authors conducted a retrospective cohort study of residents of Washington State to determine if the incidence of thyroid disease was increased among men and women exposed to atmospheric 131 I.   “ No attempt was made to identify an unexposed control group.   Instead, the study was designed to estimate variation of diseases risk over a wide range of doses.” (Emphasis:   DWM).   The cohort was composed of persons with a variety of dosages and focused especially on persons who were children at the time of the peak exposures in the 1945-1946 period.   A total of 3,447 men and women were included in the cohort.

          Twelve women (0.7%) and 7 men (0.4%) had diagnoses of thyroid cancer.   One participant had an earlier diagnosis of thyroid cancer, which was not confirmed by histological examination.   The cumulative incidence of thyroid cancer did not increase significantly with increasing dose.   There was no suggestion that the incidence of benign thyroid nodules increased with increasing dose.   Similarly, there was no increase in the incidence of autoimmune thyroiditis or hypothyroidism with increasing exposure.

          COMMENT:   I was quite surprised that no effect regarding thyroid disease was shown among this cohort, but I am not sure what conclusions one can draw.   Given the lack of a control group and the lack of well quantified data on exposure of individuals one, prudence dictates refraining from assuming that exposure to 131 I does not increase the risk of thyroid disease, particularly thyroid neoplasia.  

CORRELATES OF CIGARETTE SMOKING AMONG MEXICAN STUDENTS

Antillo-Santillan and her colleagues conducted a survey of more than 13,000 adolescents and young adults in the state of Morelos in Mexico (Antillo-Santillan E et al.   Associations between individual and contextual factors and smoking in 13,293 Mexican students.   Amer J Preven Medicine 2005; 28:   41-51.   The participants were selected by probabilistic, stratified sampling of public school students.   The school was the sampling unit.   Young people attending private schools and those not attending school were not included in the sample.   A measure of socioeconomic status was constructed based upon housing type, household income, home ownership, car ownership, etc.   

          The prevalence of regular smoking was 13.4% in males and 6.1% in females.   Frequent alcohol intoxication was powerfully associated with regular smoking (females, OR = 68.5, 95% CI; 37.6-125.2; males, OR = 34.5; 95% CI; 22.6-52.7).   Regular smoking was associated with illegal drug use and smoking by both parents in females and with illegal drug use in males.   Regular cigarette smoking was strongly associated with high socioeconomic status, particularly among high school and university students.   Among students in junior high school, 18% of those from upper class families smoked while just 6% of those from lower class families did so.   Almost 28% of upper class high school and university students smoked cigarettes. 14% of lower class high school and university students did so.

          COMMENT:   The associations between socioeconomic status and the prevalence of cigarette smoking might have been very different if those students in Morelos at the lowest socioeconomic level, young people not attending school, and those in the highest level, those attending private school, had been included in the sample.   Nonetheless, I was surprised to find that smoking was far more prevalent among upper class students.   I was curious about the association between social class and cigarette smoking among Hispanic students in this country.   I was unable to find any data on students, but I found the association among adults quite interesting.   There was a powerful inverse association between educational attainment, a reasonable proxy for social class, and cigarette smoking, which is show below:

                              Whites only, 25 years and older, 1999-2001
                            Current cigarette smokers

           On the other hand, there was almost no association between educational attainment and cigarette smoking among Hispanics:

Hispanics or Latinos, 25 years or older
Current cigarette smokers

          Why is there such a striking difference between this association in Whites and Hispanics?   Perhaps the smoking behavior among the parents is a much more powerful determinant than educational attainment in both groups.