Marginal analysis of multiple outcomes with informative cluster size.

In surveillance studies of periodontal disease, the relationship between disease and other health and socioeconomic conditions is of key interest. To determine whether a patient has periodontal disease, multiple clinical measurements (eg, clinical attachment loss, alveolar bone loss, and tooth mobility) are taken at the tooth-level. Researchers often create a composite outcome from these measurements or analyze each outcome separately. Moreover, patients have varying number of teeth, with those who are more prone to the disease having fewer teeth compared to those with good oral health. Such dependence between the outcome of interest and cluster size (number of teeth) is called informative cluster size and results obtained from fitting conventional marginal models can be biased. We propose a novel method to jointly analyze multiple correlated binary outcomes for clustered data with informative cluster size using the class of generalized estimating equations (GEE) with cluster-specific weights. We compare our proposed multivariate outcome cluster-weighted GEE results to those from the convectional GEE using the baseline data from Veterans Affairs Dental Longitudinal Study. In an extensive simulation study, we show that our proposed method yields estimates with minimal relative biases and excellent coverage probabilities.

Diabetogenic effect of pentamidine. In vitro and in vivo studies in a patient with malignant insulinoma.

Pentamidine can cause hypoglycemia followed by hyperglycemia. The mechanism of this biphasic response is not known but has been reported to be similar to that of streptozotocin and N-3-pyridyl-N-p- nitrourea (Vacor). Pentamidine (4 mg/kg per day for 12 days) was used in a patient with malignant insulinoma after several unsuccessful debulking procedures and chlorozotocin therapy. Mean glucose and immunoreactive insulin levels (+/- SEM) before and after therapy were 80 +/- 40 mg/dl versus 70 +/- 50 mg/dl and 216 +/- 12 microU/ml versus 198 +/- 22 microU/ml, respectively. These were not significantly different. The patient's five-month-old malignant insulinoma monolayer cell culture was incubated with pentamidine (60 micrograms/ml) in the presence or absence of supplemented stimulatory medium consisting of glucose (300 mg/dl) and theophylline (20 micrograms/ml). Chloroquine (60 micrograms/ml) was added to inhibit lysosomal degradation of immunoreactive insulin. Aliquots of media for immunoreactive insulin determination were obtained at 30 minutes, 20 hours, 72 hours, and three weeks. The cells were examined by high-power light microscopy at each time interval. At 30 minutes, pentamidine alone caused passive release of immunoreactive insulin, 23 percent higher than control (p less than 0.01). Stimulatory medium increased immunoreactive insulin 45 percent greater than control (p less than 0.01). Pentamidine plus stimulatory medium had no additive effect on immunoreactive insulin released within 30 minutes. At the end of 20 hours, immunoreactive insulin was no different with pentamidine and/or stimulatory medium. However, the addition of chloroquine increased immunoreactive insulin by 35 percent above the medium with pentamidine and stimulatory medium (p less than 0.01). At 72 hours, pentamidine suppressed immunoreactive insulin by 100 percent in all the media, irrespective of the presence or absence of stimulatory medium and/or chloroquine. At the end of three weeks, there was 50 percent suppression of immunoreactive insulin in the control medium, but pentamidine again completely suppressed immunoreactive insulin. High-power microscopy demonstrated intact cells in the control medium, whereas no cell structure could be detected in the media containing pentamidine at three weeks. In summary, pentamidine had no acute in vivo effect in a patient with malignant insulinoma. However, when used in an in vitro monolayer system, pentamidine caused (1) acute immunoreactive insulin release followed by inhibition of immunoreactive insulin secretion and (2) cytolysis of human malignant insulinoma cells in vitro.

Efficacy of heavy-resistance training for active women over sixty: muscular strength, body composition, and program adherence.

OBJECTIVE: To determine the effects of resistance training on muscular strength, body composition (percent fat and lean tissue mass), and program adherence in active women over 60 years of age. DESIGN: This study was designed as a stratified, randomized, non-blinded trial. Subjects were stratified into rank-ordered pairs by level of physical activity according to the Blair Seven-Day Recall, then randomly assigned into either a weight-training (WT, n = 18) or control (CON, n = 18) group. PARTICIPANTS AND SETTING: Thirty-six women over the age of 60 (67.1 +/- 1.5, chi +/- SE) were recruited from the San Diego community. All subjects had to be engaging in some form of aerobic exercise at least 3 days each week for a minimum of 6 months. MEASUREMENTS AND INTERVENTION: Subjects engaged in isotonic training 3 days/week on Polaris machines at a target intensity of 80% of the 1 repetition maximum (1 RM) for three sets of seven exercises selected to train major muscle groups of the trunk and upper and lower body. The 1 RM was re-tested every 6 weeks in WT and the workload adjusted to maintain target intensity. The 1 RM was tested at 0, 12, and 24 weeks in CON. Body fatness and lean tissue mass, excluding skeletal mass, were assessed at 0 and 24 weeks by dual energy radiography using a total body scan. RESULTS: Increases in muscle strength of the weight trainers were significant for all seven exercises (5%-65%), with the greatest gains in the shoulder and trunk muscles. Percent body fat of weight trainers decreased significantly (from 38.8% to 37.9%, P < 0.05), while lean tissue mass increased by 1.5 kg (P < 0.01). No changes in body composition were observed in control subjects. Overall program adherence was 83% (15 weight trainers and 15 control women completed the study). Monthly attendance averaged 86.8 +/- 3.3%, with no incidence of injury during the training sessions. CONCLUSIONS: These data indicate that heavy-resistance weight training is safe and enjoyable for older women and that meaningful gains in muscular strength and body composition can be made even in women who are already highly active.

Behavior of adult and aged mice before and after central injection of interleukin-1beta.

The level of locomotor activity, body temperature (T(B)), and feeding for adult (3-5-month old) and aged (22-24-month old) male BALB/c mice was determined and the sensitivity of the two age groups to the anorectic, febrile, and behavioral properties of interleukin-1beta (IL-1beta) in the brain was examined. Baseline locomotor activity and T(B) were markedly lower in aged mice than in adults and the circadian rhythm for both activity and T(B) were disrupted in the aged. Adult and aged mice consumed similar amounts of food during the daytime and nighttime, but aged mice made longer, less frequent visits to the feed cup. To determine if aging affects the responsiveness to central IL-1beta, adult and aged mice were injected intracerebroventricularly with PBS or IL-1beta. Compared to age-matched PBS controls, IL-1beta increased T(B) in both adult and aged mice. The peak deltaT(B) was greater in aged mice than in adults, but because of a lower baseline T(B) in aged mice, peak T(B) after IL-1beta was not different between groups. Locomotor activity of aged mice receiving PBS was about half that of PBS-injected adults and was not depressed further by IL-1beta. However, compared to age-matched PBS controls, centrally administered IL-1beta depressed food intake more in aged mice than in adults. These data indicate that even though feeding, locomotor activity, and T(B) are affected by aging, the central component of the inflammatory response mediated by IL-1beta is retained.

Women in low income groups smoke more: Canterbury 1976-92.

AIM: To examine the relationship between income and levels of smoking in women of child-bearing age. METHODS: Census area units within the Christchurch region were divided into five groups according to average income for each of the census years 1976, 1981 and 1991 respectively. Smoking rates were obtained from census data for 1976 and 1981. For 1992, data from a questionnaire on smoking during pregnancy and obstetric records were used. RESULTS: For all years, higher proportions of women from low income groups smoked. The overall percentage of female smokers in Canterbury fell slightly from 31.6% in 1976 to 27.5% in 1981. There were 27.1% of pregnant women smoking in 1992. There were marked changes by economic groups. The percentage of smokers in the highest income group fell from 24.8% in 1976 to 21.0% in 1981: an 18% fall. Smokers in the lowest income group increased from 36.8% in 1976 to 39.8% in 1981. The same patterns were seen for smoking in pregnancy in the 1992 data. CONCLUSION: While the downward trend for smokers in the high income groups is good news, it is of concern that the proportion of smokers in the lower income groups remains so high. Young women have the highest rates of smoking, which is of special concern when they become pregnant. Smokefree intervention programmes need to be specifically targeted at these groups.

Prevalence of smoking in early pregnancy by census area: measured by anonymous cotinine testing of residual antenatal blood samples.

AIM: To accurately measure the prevalence of smoking in early pregnancy by census area units (CAU) in Christchurch. METHODS: Smoking status in pregnancy was determined by serum cotinine assay for all antenatal blood samples taken over a 6 month period. CAUs in Christchurch were grouped into quartiles according to the proportion of maternal smokers. Social factors from 1991 census data were used to describe the characteristics of each quartile. RESULTS: The overall rate of smoking in pregnancy was 33.0%. Rates ranged from 10.6% to 56.9% for the census area groups. CAUs in the upper quartile (39-57% of women smoking in pregnancy) were clustered together geographically and were associated with lower socioeconomic indices. The strongest correlation was between average income with smoking rates (Pearson correlation coefficient 0.76). CONCLUSION: Smoking rates in pregnancy have remained at around 30% for at least 20 years, with some areas of the city having rates nearly double this. It would seem logical to promote smoke-free pregnancy activities in localities with the highest rates of smoking. Future evaluation of the efficacy of such programmes should be done using objective measurements.

Higher rates of SIDS persist in low income groups.

OBJECTIVE: To examine how changes in the rates of sudden infant death syndrome (SIDS) have varied in different income groups during a 25 year period. METHODOLOGY: Census data were obtained for five census periods (1971, 1976, 1981, 1986 and 1991) on the number of SIDS cases and inevitable deaths. Census area units (CAU) were ranked according to the average income earned by adults over the age of 15 years for each census year. The CAU were then divided into three equal income groups: low, middle and high. RESULTS: The rates of SIDS differed significantly between the three income groups for the 1991 census period with the low income SIDS rate being 4.6/1000 births compared to 1.2/1000 live births for the higher income groups (Chi-squared = 18.3, P < 0.0001). There was no association between rates of inevitable deaths and income groups. CONCLUSION: Currently, low income groups have three times the rate of SIDS compared to those in higher income groups. The reason for this is probably because the disadvantaged groups carry an overall higher burden of risk factors for SIDS. This must be kept in mind as further SIDS educational programmes are developed and implemented.

The febrile stress of routine vaccination does not increase central apnoea in normal infants.

We tested the hypothesis that the febrile stress of routine vaccination would increase central apnoea in normal infants. Twenty-one normal infants had continuous overnight breathing and temperature recorded at home, before and after 58 routine vaccination episodes. Central apnoea, of at least 5 sec duration, was detected by computer algorithm and confirmed by human inspection. The longest recorded apnoea was 16 sec (n = 1) during 3629 h of sleep. Overnight rectal temperature increased after vaccination (median 0.52 degrees C, 95% CI 0.40, 0.65). Apnoea density reduced on 46/53 vaccination nights (median -29%, 95% CI -20, -37) followed by an increase on subsequent nights (median +10%, 95% CI +1%,+21%). Overall, apnoea density was similar during the 3 nights preceding and 4 nights following vaccination (median +1%, 95% CI +9,-6). The febrile stress of routine vaccination did not increase central apnoea in normal infants.

Breathing, sleep state, and rectal temperature oscillations.

Overheating may cause terminal apnoea and cot death. Rectal temperature and breathing patterns were examined in normal infants at home during the first 6 months of life. Twenty one infants had continuous overnight rectal temperature and breathing recordings for 429 nights (mean 20.4 nights, range 7-30) spaced over the first six months of life. Periods when breathing was 'regular' were directly marked on single night records. Sleep state was determined from respiratory variables. 'Regular' breathing was a reliable marker of 'quiet' sleep (specificity 93%). The duration of 'quiet' sleep increased from 6 to 22 minutes from two weeks to three months of age and then remained static, as did the proportion of sleep spent in the quiet phase (9% to 34%). Rectal temperature fell during 66% of quiet sleep and usually rose during rapid eye movement (REM) sleep. The drop in rectal temperature was maximal at the start of quiet sleep, whereas the maximum rise during REM sleep was reached after 10 to 15 minutes. Oscillations in rectal temperature are associated with changes in sleep and breathing state. The maturation of rectal temperature patterns during the first six months of life are closely related to a maturation of sleep state and breathing patterns.

Travel and changes in routine do not increase the risk of sudden infant death syndrome.

We investigated the relationship between travel and changes in routine and the sudden infant death syndrome (SIDS) among 485 SIDS cases compared with 1800 randomly selected control infants. There was no increased risk of SIDS with travel. Special events, such as christenings, were not associated with an increased risk of SIDS. However, visits to and by friends or relatives were associated with a significantly reduced risk of SIDS after controlling for potential confounders (odds ratios = 0.70; 95% confidence interval = 0.52, 0.96). These findings may indicate less social support in SIDS cases.