American Journal of Epidemiology

Kingsley, Lawrence A.; Zhou, Susan Y. J.; Bacellar, Helena; Rinaldo, Charles R.; Chmiel, Joan; Detels, Roger; Saah, Alfred; VanRaden, Mark; Ho, Monto; Muñoz, Alvaro

doi: 10.1093/oxfordjournals.aje.a116094pmid: 1877593

The 5-year temporal trends in human immunodeficiency virus type 1 (HIV-1) seroconversion between 1984 and 1989 among homosexual/bisexual men participating in the Multicenter AIDS Cohort Study (MACS) are reported. Of 3,262 initially seronegative men, 368 (11.3%) had seroconverted by December 31, 1989. Although the incidence of seroconversion declined precipitously during the first 3 years of follow-up (from 4.1% to 0.9% per 6 months), no evidence for a further substantial reduction was noted after mid-1987, since 6-month incidence rates ranged between 0.5% and 1.2%. The Chicago cohort experienced an increase in HIV-1 seroconversion during both semesters of 1989; 2.1% and 1.6% per 6 months, respectively, became newly infected. Other MACS centers did not report such an increase. Center-specific differences were observed by race; black men were at higher seroconversion risk than white men in Baltimore/Washington (relative risk (RR) = 3.4) and Chicago (RR = 2.4), while Hispanic men were at higher risk than white men in Chicago (RR = 3.3). Younger age (<35 years) was also associated with HIV-1 seroconversion (RR = 1.5). It is disturbing to report an overall annual seroconversion rate of 1.2% for the 2 years prior to December 31, 1989, as well as evidence for a sustained recent increase in Chicago during 1989. Long-term maintenance of safe-sex behaviors should be the cornerstone of acquired immunodeficiency syndrome prevention among homosexual/bisexual men.

Demers, Paul A.; Thomas, David B.; Rosenblatt, Karin A.; Jimenez, L. Margarita; McTiernan, Anne; Stalsberg, Helge; Stemhagen, Annette; Thompson, W. Douglas; Curnen, Mary G. McCrea; Satariano, William; Austin, Donald F.; Isacson, Peter; Greenberg, Raymond S.; Key, Charles;

Brown, Linda Morris; Dosemeci, Mustafa; Blair, Aaron; Burmeister, Leon

doi: 10.1093/oxfordjournals.aje.a116096pmid: 1877595

Information from surrogates is increasingly being used in case-control studies to evaluate cancer risks from pesticides; however, little is known about the quality of this type of information. To address this concern, the authors compared interview data collected in 1987 from 95 male lowa farmers and their wives or other surrogates on the use of specific agricultural pesticides. Agreement between direct and surrogate interviews was excellent (83–100%) for responses to dichotomous (yes/no) questions regarding past agricultural use of specific pesticides. Although there were more discrepancies for detailed questions (e.g., the number of days per year on which each pesticide was handled), responses from spouses appear to be adequate for epidemiologic studies of pesticides and cancer.

Knekt, Paul; Aromaa, Arpo; Maatela, Jouni; Alfthan, Georg; Aaran, Ritva-Kaarina; Nikkari, Tapio; Hakama, Matti; Hakulinen, Timo; Teppo, Lyly

doi: 10.1093/oxfordjournals.aje.a116097pmid: 1877596

The associations between serum alpha-tocopherol, beta-carotene, retinol, retinolbinding protein, and selenium levels and the subsequent occurrence of different cancers of low incidence were investigated in a nested case-control study of 39,268 men and women participating in the Social Insurance Institution's Mobile Clinic Health Examination Survey in Finland. During follow-up from the baseline in 1968–1972 to the end of 1977, a total of 115 cancers of the lip, oral cavity, pharynx, larynx, esophagus, liver, gallbladder, kidney, urinary bladder, brain, and skin were reported to the nationwide Finnish Cancer Registry. Alpha-tocopherol, beta-carotene, retinol, retinol-binding protein, and selenium concentrations were determined from stored serum samples collected from these cancer cases and matched controls at baseline. Several sites indicated an elevated risk of cancer at low levels of the serum variables, although only a few of these associations were statistically significant. Only melanoma patients had significantly lower serum alpha-tocopherol and beta-carotene levels than corresponding controls. Since the numbers of cancer cases were small, no firm conclusions can be drawn from these results until they have been confirmed in studies based on larger cohorts or on pooled data from several small samples.

Irwin, Kathleen L.; Weiss, Noel S.; Lee, Nancy C.; Peterson, Herbert B.

doi: 10.1093/oxfordjournals.aje.a116098pmid: 1877597

Several hypotheses predict that tubal sterilization and hysterectomy may influence a woman's risk of developing ovarian cancer. To examine the relation between these surgeries and epithelial ovarian cancer, the authors analyzed data from the Cancer and Steroid Hormone Study, a case-control study of women aged 20–54 years. Eight population-based cancer registries in the United States identified women with newly diagnosed epithelial ovarian cancer during 1980–1982 (n = 494). A comparison sample of female residents of these eight areas (n – 4,238) was identified through random digit dialing. Women who had had tubal sterilization (relative risk (RR) = 0.69, 95% confidence interval (CI) 0.50–0.95), a hysterectomy only (RR = 0.55, 95% CI 0.38–0.81), or a hysterectomy with unilateral oophorectomy (RR = 0.60, 95% CI 0.31–1.17) had lower risks of ovarian cancer than did women who had never had any sterilization surgery. However, the negative associations with tubal sterilization and hysterectomy only appeared to wane after two decades. These findings may be partly explained by the creening for occult ovarian pathology that often accompanies pelvic surgery: Women whose ovaries screen as “negative” may be temporarily at low risk of being diagnosed with ovarian cancer. However, because the decreased risks persisted for so long, it is onceivable that hormonal, mechanical, or circulatory sequelae of these sterilization procedures may act to lower ovarian cancer risk.

Sorel, Janet E.; Ragland, David R.; Syme, S. Leonard

doi: 10.1093/oxfordjournals.aje.a116099pmid: 1877598

Previous studies of Mexican Americans have shown mean diastolic and systolic blood pressures and prevalence rates of hypertension which are either lower than or similar to those for non-Hispanic whites despite the predominance of obesity in Mexican Americans. However, those results are based on restricted samples from California and Texas. Using data from the Second National Health and Nutrition Examination Survey (1976–1980) and the Hispanic Health and Nutrition Examination Survey (1982–1984), the authors examined ethnic differences in blood pressure and hypertension. Regression analyses, stratified by sex, were used to compare mean blood pressures and rates of hypertension in Mexican Americans with those for whites and blacks. Mean diastolic and systolic blood pressures, as well as the prevalence of hypertension, were lower in Mexican Americans than in non-Hispanic whites or in blacks, with whom they shared a remarkably similar risk profile. This effect was unchanged after adjustment for age, body mass index (weight (kg)/height (cm)2 × 100), and education, indicating that blood pressure differences between Mexican Americans, whites, and blacks were not explained by the established correlates of high blood pressure. There are several possible reasons for lower blood pressure in Mexican Americans, including genetic, life-style, and cultural factors.

Canner, Paul L.; Borhani, Nemat O.; Oberman, Albert; Cutler, Jeffrey; Prineas, Ronald J.; Langford, Herbert; Hooper, Frank J.

doi: 10.1093/oxfordjournals.aje.a116100pmid: 1877599

Blood pressure variability is an important consideration in hypertension trials for determining required sample size and consequently making accurate outcome statements. The Hypertension Prevention Trial was a randomized controlled trial carried out in 1983–1986 in four US clinics on men and women with high normal diastolic blood pressure. This trial provided an opportunity to assess conditions affecting blood pressure measurement variability. Trained blood pressure observers measured systolic and diastolic blood pressures twice, 30 seconds apart, using a random-zero sphygmomanometer. The quality of blood pressure measurements was assessed by computing the variability of the two readings per participant-visit for each blood pressure observer at each study clinic. Other sources of variability investigated included observer digit preference, time of day, and ambient temperature. On the basis of data from this population, it is estimated that the standard deviation of blood pressure values can be reduced by 5% by taking two measurements per participant-visit. An additional reduction of variability can be effected by having the duplicate blood pressure measurements made by different blood pressure observers. In special instances where the range of blood pressure values is very restricted, use of the random-zero sphygmomanometer can increase or decrease the among-participant variability in blood pressure values, depending upon where the distribution of blood pressure values is centered.

Correa-Villaseñor, Adolfo; McCarter, Robert; Downing, John; Ferencz, Charlotte

doi: 10.1093/oxfordjournals.aje.a116101pmid: 1877600

Cardiovascular malformations were examined for white/black variation in the Baltimore-Washington Infant Study. In this population-based case-control study, cases (n = 2,087) were live births with cardiovascular malformations ascertained through pediatric cardiology centers and 53 hospitals in Maryland, the District of Columbia, and northern Virginia between 1981 and 1987. Controls (n = 2,721) were a random sample of infants from the live-birth cohort that gave rise to the cases. The proportion of infants that were white was similar for all cases as a group and controls (0.68 and 0.67, respectively). Subgroup analysis, however, revealed an excess of white infants among cases with Ebstein's anomaly (odds ratio (OR) = 3.7, 95% confidence interval (CI) 1.1–12.5), aortic stenosis (OR = 3.6, 95% CI 1.7–7.6), pulmonary atresia (OR = 2.5, 95% CI 1.0–6.1), coarctation of the aorta (OR = 2.2, 95% CI 1.4–3.5), and D-transposition of the great arteries (OR = 1.6, 95% CI 1.1–2.5), and a deficit of white infants among cases with pulmonary stenosis (OR = 0.6, 95% CI 0.4–0.8) and heterotaxia (OR = 0.4, 95% CI 0.3–0.8). These associations remained when cases were stratified by infant's age or by method of diagnosis. Controlling for socioeconomic factors attenuated the white excess for Ebstein's anomaly (OR = 3.0, 95% CI 0.9–10.5), disclosed a white excess among cases of L-transposition of the great arteries (OR = 2.8, 95% CI 1.0–8.0), and revealed that the white excess for aortic stenosis was limited to low and middle socioeconomic strata. These results highlight racial variations in cardiovascular malformations, suggest that socioeconomic factors account for some of this variation, and identify malformation subgroups for which further evaluation of sociocultural, environmental, and familial factors is needed.

Evans, Denis A.; Smith, Laurel A.; Scherr, Paul A.; Albert, Marilyn S.; Funkenstein, H. Harris; Hebert, Liesi E.

doi: 10.1093/oxfordjournals.aje.a116102pmid: 1877601

A random sample of 467 persons over age 65 years from the population of an urban US community, stratified by age, sex, and performance on a brief memory test, underwent clinical evaluation for dementing illness in 1982–1984. Of these persons, 134 had probable Alzheimer's disease, 166 had possible Alzheimer's disease, and 167 had no evidence of Alzheimer's disease. Over a median follow-up period of 4.9 years following evaluation, 165 (35%) died. Overall, persons with probable Alzheimer's disease had a relative risk of death 1.44 (95% confidence interval (CI) 1.05–1.96) times that of the unaffected. Level of cognitive impairment and the presence of cachexia upon physical examination both strongly and independently modified risk of death. Among those with probable Alzheimer's disease, mortality for those with mild or moderate cognitive impairment and no evidence of cachexia was comparable to that of the unaffected. However, among those with probable Alzheimer's disease and either severe cognitive impairmen or cachexia, the risk of death was substantially higher. Persons with probable Alzheimer's disease who had both severe cognitive impairment and clear cachexia had a risk of death 4.60 (95% CI 1.63–13.1) times that of unaffected persons.

Lew, Judy F.; Swerdlow, David L.; Dance, Mario E.; Griffin, Patricia M.; Bopp, Cheryl A.; Gillenwater, Michael J.; Mercatante, Tony; Glass, Roger I.

doi: 10.1093/oxfordjournals.aje.a116103pmid: 1652203

From October 23 to October 27, 1989, an outbreak of gastroenteritis occurred aboard a cruise ship in the Caribbean. The 818 passengers and 518 crew members were surveyed for gastrointestinal symptoms; 72 (14%) of 512 passengers and 12 (3%) of 388 crew members who answered the survey reported having a diarrheal illness. Multiple-antibiotic-resistant Shigella flexneri 4a was isolated from 19 ill passengers and two ill crew members. Thirteen people were hospitalized, and prolonged duration of illness was associated with taking an antibiotic to which the isolated strain of Shigella was resistant. A case-control study of food items implicated German potato salad as the vehicle of transmission. It was prepared and probably infected by a food handler from a country where multiple-antibiotic-resistant Shigella is common. Spread may have been facilitated by the limited availability of toilet facilities for the galley crew. This outbreak demonstrates how antibiotic-resistant strains can be introduced into the United States, where they can pose treatment problems. The continuing problem of foodborne gastrointestinal disease in settings such as cruise ships underscores the need for basic hygienic control for food handlers and food preparation areas. In addition, the availability of adequate working conditions for crew members, including appropriately furnished toilet facilities, may be important issues that must be addressed in order to decrease the frequency of diarrhea outbreaks aboard cruise ships.