Abstract: Background: There are no studies that define the basic epidemiology of pulmonary embolism (PE) and deep venous thrombosis (DVT) in the elderly. This project was undertaken to provide that information. Methods: We obtained all Medicare claims during the period 1986 through 1989 from a random 5% sample of US Medicare enrollees. By selecting codes used for diagnoses and treatment, we identified 7174 cases of PE and 8923 cases of DVT. These cohorts were analyzed to provide incidence by age, race, sex, and geographic location; frequency of invasive treatment; frequency of PE after treatment for DVT; frequency of recurrence of PE; and survival after diagnosis. Results: Annual incidence rates per 1000 at age 65 to 69 years for PE and DVT were 1.3 and 1.8, respectively. Both rates increased steadily with age to 2.8 and 3.1 by age 85 to 89 years. For PE, women had lower rates than men (adjusted relative risk, 0.86; 95% confidence interval, 0.82 to 0.90), and blacks had higher rates than whites (adjusted relative risk, 1.25; 95% confidence interval, 1.15 to 1.36). For DVT, the associations with gender and race were weaker and in the opposite direction. Pulmonary embolectomy was done in 0.2% of cases of PE; interruption of the vena cava was done in 4.4% of cases of PE and in 2% of cases of DVT. Thrombectomy was done in 0.3% of all cases. Pulmonary embolism occurred in 1.7% of patients with DVT within 1 year of hospital discharge for initial treatment. The 1-year recurrence rate for PE was 8.0%. In-hospital mortality associated with PE and DVT was 21% and 3%, respectively. One-year mortality was 39% and 21%. respectively. Conclusions: Pulmonary embolism and DVT are common problems in the elderly. Both increase with age, but the effects of race and sex are small. Current treatment patterns appear to be effective in preventing both PE after DVT and recurrence of PE. Both are associated with substantial 1-year mortality, suggesting the need to understand the role of associated conditions as well as the indications for prophylaxis and the methods of treatment. (Arch Intern Med. 1994;154:861-866)

Abstract: Background: To augment available information about the epidemiology of legionnaires' disease, we analyzed data reported to the passive surveillance system at the Centers for Disease Control and Prevention, Atlanta, Ga, from 1980 through 1989. Methods: Risk of disease associated with specific demographic characteristics and health conditions was calculated by comparing the surveillance group with the US population. Risk of death was calculated using multivariate logistic regression models. Results: A diagnosis of legionnaires' disease was confirmed on the basis of clinical and laboratory criteria for 3254 patients. Disease rates did not vary by year, but were higher in the northern states and during the summer.Legionella pneumophila, serogroup 1, constituted 71.5% of fully identified isolates. This study confirmed previously identified risk factors for legionnaires' disease. In addition, a markedly elevated risk was identified for persons with acquired immunodeficiency syndrome (rate ratio, 41.9; 95% confidence interval, 12.9, 71.0), or hematologic malignancy (rate ratio, 22.4; 95% confidence interval, 19.0, 25.9). Likelihood of death was increased in patients who were elderly or male; those with hospital-acquired infection, renal disease, malignancy, or immunosuppression; and those from whomL pneumophila, serogroup 6, was isolated. Conclusions: Infection withLegionellaremains an important cause of disease and death in the United States. Diagnosis and treatment of legionnaires' disease should be targeted at patients at increased risk for illness and complications due toLegionellainfection. Diagnostic tests for legionnaires' disease based on species other thanL pneumophila, serogroup 1, should be developed and tested. Recommendations for prevention of legionnaires' disease should be focused on settings where there are persons at greatest risk for illness or serious outcome. (Arch Intern Med. 1994;154:2417-2422)

Abstract: Background: Recent evidence indicates that the prevalence of sleep-disordered breathing is remarkably high (24% for men and 9% for women) and that the public health burden attributable to sleep-disordered breathing is substantial. This investigation examines current and former cigarette smoking as potential risk factors for sleep-disordered breathing. Methods: Data were from 811 adults enrolled in the University of Wisconsin Sleep Cohort Study, Madison. The Sleep Cohort Study is a longitudinal, epidemiologic study that uses nocturnal polysomnography to investigate sleep-disordered breathing and other disorders of sleep. The presence and severity of sleep-disordered breathing was quantified by the frequency of apneas and hypopneas per hour of sleep. Results: Logistic regression analyses were used to control for potential confounding factors. Compared with never smokers, current smokers had a significantly greater risk of snoring (odds ratio, 2.29) and of moderate or worse sleep-disordered breathing (odds ratio, 4.44). Heavy smokers (≥40 cigarettes per day) had the greatest risk of mild sleep-disordered breathing (odds ratio, 6.74) and of moderate or worse sleep-disordered breathing (odds ratio, 40.47). Former smoking was unrelated to snoring and sleep-disordered breathing after adjustment for confounders. Conclusions: Current cigarette smokers are at greater risk for sleep-disordered breathing than are never smokers. Heavy smokers have the greatest risk while former smokers are not at increased risk for sleep-disordered breathing. Thus, smoking cessation should be considered in the treatment and prevention of sleep-disordered breathing. (Arch Intern Med. 1994;154:2219-2224)

Abstract: Background: Cardiovascular risk factors are known to persist over time and to cluster both in childhood and adulthood. Less is known about the persistence of clustering of multiple cardiovascular risk factors comprising adverse levels of systolic blood pressure, the ratio of total cholesterol to high-density lipoprotein cholesterol, and plasma insulin from childhood to young adulthood. Methods: In a community study of cardiovascular risk, 1176 individuals (52% female, 44% black) aged 5 through 17 years at baseline were followed up for 8 years. Results: Calculated as sum of the race-, sex-, and age-specific rankings of systolic blood pressure, insulin level, and total to high-density lipoprotein cholesterol ratio, the multiple risk index was shown to track in all four race-sex groups (year 1 vsyear 8 correlations,.54 to.67). The magnitude of the overall multiple risk index tracking correlation (r=.64) was significantly stronger than that noted for individual risk factors (r=.34 to.57). Among subjects who were initially in the highest quintile of the multiple risk index, 61% remained there 8 years later. Tracking of the multiple risk index increased progressively with age and ponderal index (weight-[height3]). In a stepwise regression analysis, baseline multiple risk index score, baseline ponderal index, change in ponderal index, and change in height were predictive of the multiple risk index score on follow-up. These predictors explained 45% to 60% of the variability in multiple risk index scores among the race-sex groups. Conclusions: The persistence of multiple cardiovascular risk clustering from childhood to adulthood and the impact of obesity in this regard point to the need for preventive measures aimed at developing healthy lifestyles early in life. (Arch Intern Med. 1994;154:1842-1847)

Abstract: Background: Because of the prevalence of pneumococcal pneumonia, the substantial morbidity and mortality associated with many pneumococcal infections, and an increase in the incidence of antibiotic resistance among pneumococcal isolates, considerable efforts for disease prevention have been made using a polyvalent polysaccharide pneumococcal vaccine. Despite numerous clinical trials of the vaccine, its efficacy in the prevention of pneumococcal infections and other clinically relevant medical outcomes in adults remains uncertain. Methods: To assess quantitatively the efficacy of pneumococcal vaccination, a MEDLINE literature search, manual reviews of article bibliographies, and communications with pneumococcal vaccine investigators were used to identify randomized controlled trials of the pneumococcal vaccine. Independent review of 594 articles revealed nine randomized trials with 12 vaccine and control study groups that evaluated clinically relevant outcomes in adults. To estimate a summary effect size for all outcomes, Mantel-Haenszel odds ratios (ORs) and Dersimonian and Laird rate differences (RDs) and their associated 95% confidence intervals (CIs) were computed. Results: Summary ORs demonstrated a statistically significant protective effect of the vaccine for four pneumococcal infection—related outcomes: definitive pneumococcal pneumonia (OR=0.34; 95% CI=0.24 to 0.48), definitive pneumococcal pneumonia for vaccine-containing pneumococcal antigen types only (vaccine types only) (OR=0.17; 95% CI=0.09 to 0.33), presumptive pneumococcal pneumonia (OR=0.47; 95% CI=0.35 to 0.63), and presumptive pneumococcal pneumonia (vaccine types only) (OR=0.39; 95% CI=0.26 to 0.59). The summary RDs, which account for heterogeneity among studies, confirmed a statistically significant protective effect for two of these same outcomes: definitive pneumococcal pneumonia (RD=4/1000; 95% CI=0/1000 to 7/1000) and definitive pneumococcal pneumonia (vaccine types only) (RD=8/1000; 95% CI=1/1000 to 16/1000). Summary ORs and RDs failed to demonstrate a protective effect for pneumonia (all causes), bronchitis, and mortality (all causes) or mortality due to pneumonia or pneumococcal infection. Subgroup analyses showed that for all four pneumococcal infection—related outcomes, vaccine efficacy differed for high- and low-risk subjects, demonstrating efficacy for low-risk subjects and lack of efficacy for high-risk subjects. Conclusions: Pneumococcal vaccination appears efficacious in reducing bacteremic pneumococcal pneumonia in low-risk adults. However, evidence from randomized controlled trials fails to demonstrate vaccine efficacy for pneumococcal infection-related or other medical outcomes in the heterogeneous group of subjects currently labeled as high risk. (Arch Intern Med. 1994;154:2666-2677)

Abstract: Background: Noncompliance with medications is one of the most serious problems facing health care today However, methods to measure compliance have many limitations Methods: To measure specific drug compliance and dosing frequency of two asthma medications, we used medical records data and pharmacy claims data from 276 patients who had concurrent prescriptions for inhaled anti-inflammatory agents and oral theophylline Patients were randomly selected from the pharmacy claims data files of a health maintenance organization The patients' medical records were reviewed, and records that did not contain clear documentation of the medication, dose, and dosing frequency were excluded Data from the remaining 119 medical records were compared with data from pharmacy claims to calculate compliance rates for each medication Results: Our calculations showed that patients were significantly more compliant with prescribed theophylline medication than with two inhaled anti-inflammatory medications (P=0001) No significant differences in compliance were found relative to prescribed dosing frequency (twice daily or less compared with three times daily or more) for either medication (P=6517) Conclusions: Comparison of medical record data with pharmacy claims data is an effective indirect measure of patients' compliance with prescribed oral theophylline and inhaled anti-inflammatory agents Additional interventions must be pursued for patients with asthma regarding adherence to regimens for their prescribed inhaled anti-inflammatory agents (Arch Intern Med 1994;154:1349-1352)

Abstract: In recent years, the importance of gram-negative organisms in the genesis of sepsis has been emphasized. However, this emphasis may no longer be correct; recent studies show an increasing incidence of gram-positive sources of sepsis, and it is possible that these cases may predominate in the coming years. This increase results from more than just a greater prevalence of infection—it appears that gram-positive organisms may also be more virulent in fomenting the disease, as can be evidenced by the emergence of streptococcal toxic shock syndrome and the resurgence of acute rheumatic fever. This may result from the ability of gram-positive organisms to produce more inflammation-causing cell wall constituents, as well as unbound exotoxins. Despite the recent emphasis on gram-negative causes, sepsis resulting from gram-positive sources is increasingly common. Research on these causes of sepsis should be encouraged. (Arch Intern Med. 1994;154:26-34)

Abstract: Background: Several studies carried out in a limited number of patients demonstrated a wide range of overestimation of glomerular filtration rate (GFR) by serum creatinine level and creatinine clearance (Ccr) in liver disease. Methods: We simultaneously evaluated Ccr, inulin clearance, and predicted GFR calculated from serum creatinine level in 56 cirrhotic patients. Inulin clearance was considered the gold standard for GFR evaluation. Results: The sensitivity of serum creatinine level, predicted GFR, and Ccrin detecting renal failure was 18.5%, 51%, and 74%, respectively. On the basis of inulin clearance, patients were divided into two groups: those with normal GFR (mean, 106±34mL/min per 1.73 m2) (group 1,29 patients) and those with reduced GFR (mean, 56± 19 mL/min per 1.73 m2) (group 2, 27 patients). Predicted GFR and Ccrwere accurate markers of GFR in group 1 patients, while both overestimated GFR by about 50% in group 2 patients. An increased tubular secretion of creatinine accounted for the disparity between Ccrand inulin clearance in these patients. Conclusions: Our results indicate that renal failure is greatly underestimated on the basis of serum creatinine level and Ccrin cirrhotic patients. Clinical implications of this observation include excessive dosage of potentially nephrotoxic drugs and failure to recognize renal impairment induced by such medical treatments as diuretic therapy or paracentesis. (Arch Intern Med. 1994;154:201-205)

Abstract: Background: The purpose of this study was to estimate the sensitivity and specificity of diagnostic tests for gallstones and acute cholecystitis. Methods: All English-language articles published from 1966 through 1992 about tests used in the diagnosis of biliary tract disease were identified through MEDLINE. From 1614 titles, 666 abstracts were examined and 322 articles were read to identify 61 articles with information about sensitivity and specificity. Application of exclusion criteria based on clinical and methodologic criteria left 30 articles for analysis. Cluster-sampling methods were adapted to obtain combined estimates of sensitivities and specificities. Adjustments were made to estimates that were biased because the gold standard was applied preferentially to patients with positive test results. Results: Ultrasound has the best unadjusted sensitivity (0.97; 95% confidence interval, 0.95 to 0.99) and specificity (0.95; 95% confidence interval, 0.88 to 1.00) for evaluating patients with suspected gallstones. Adjusted values are 0.84 (0.76 to 0.92) and 0.99 (0.97 to 1.00), respectively. Adjusted and unadjusted results for oral cholecystogram were lower. Radionuclide scanning has the best sensitivity (0.97; 95% confidence interval, 0.96 to 0.98) and specificity (0.90; 95% confidence interval, 0.86 to 0.95) for evaluating patients with suspected acute cholecystitis; test performance is unaffected by delayed imaging. Unadjusted sensitivity and specificity of ultrasound in evaluating patients with suspected acute cholecystitis are 0.94 (0.92 to 0.96) and 0.78 (0.61 to 0.96); adjusted values are 0.88 (0.74 to 1.00) and 0.80 (0.62 to 0.98). Conclusions: Ultrasound is superior to oral cholecystogram for diagnosing cholelithiasis, and radionuclide scanning is the test of choice for acute cholecystitis. However, sensitivities and specificities are somewhat lower than commonly reported. We recommend estimates that are midway between the adjusted and unadjusted values. (Arch Intern Med. 1994;154:2573-2581)

Abstract: Background: Oxidation of low-density lipoprotein is believed to be an important step in the pathogenesis of atherosclerosis. The purpose of the present study was to determine whether antibody against oxidized low-density lipoprotein, reported to be associated with the progression of carotid atherosclerosis, is predictive of cardiac death and nonfatal myocardial infarction. Methods: Serum samples from 135 cases and their controls, drawn at entry from middle-aged dyslipidemic men participating in the Helsinki Heart Study, a 5-year coronary primary prevention trial with gemfibrozil, were tested for immunoglobulin G class antibodies against oxidized low-density lipoprotein by enzyme-linked immunosorbent assay. Results: The mean antibody level, expressed in optical density units, was significantly higher in cases than in controls (0.412 vs 0.356, P =.002). After adjustment for age, smoking, blood pressure, and high-density lipoprotein cholesterol level, there was a 2.5-fold increased risk (95% confidence interval, 1.3 to 4.9) of a cardiac end point in the highest tertile of antibody level vs the lowest tertile ( P =.005 for trend). Conclusions: Elevated levels of antibodies against oxidized low-density lipoprotein were predictive of myocardial infarction. The effect was independent of low-density lipoprotein cholesterol levels, and the joint effect was additive. Elevated antibody levels modified the effects of classic coronary risk factors. (Arch Intern Med. 1994;154:2605-2609)

Abstract: Background: While observational studies have suggested that unopposed estrogens reduce the incidence of coronary disease in postmenopausal women, there are few data on the effect of combined therapy with estrogens and progestins—a regimen adopted in recent years to minimize the risk of endometrial hyperplasia and cancer. In clinical trials, the addition of progestins has an adverse effect on serum lipid levels, and these lipid effects have raised the question of whether combined estrogen-progestin therapy increases the risk of coronary disease compared with the use of estrogen alone. Methods: We conducted a population-based, case-control study among enrollees of Group Health Cooperative of Puget Sound. Cases were postmenopausal women who sustained an incident fatal or nonfatal myocardial infarction in 1986 through 1990. Controls were a stratified random sample of female Group Health Cooperative enrollees frequency matched to the cases by age and calendar year. We reviewed the medical records of the 502 cases and 1193 controls and conducted brief telephone interviews with consenting survivors. The health maintenance organization's computerized pharmacy database was used to ascertain the use of postmenopausal hormones. For the primary analysis of current use, we classified women into one of three groups: (1) nonusers of hormones; (2) users of estrogens alone; or (3) users of combined therapy including both estrogens and progestins. Each group of hormone users was compared with nonusers. Results: After adjustment for potential confounding factors, the risk ratio of myocardial infarction associated with current use of estrogens alone was 0.69 (95% confidence interval, 0.47 to 1.02); and the risk ratio of myocardial infarction associated with current use of combined therapy was 0.68 (95% confidence interval, 0.38 to 1.22). Duration of combined-therapy use was relatively short, averaging less than 2 years in cases and controls. Conclusions: In this case-control study, the reduced risk of myocardial infarction associated with the use of estrogens alone was consistent with previous observational studies. Although the 95% confidence interval only excluded a risk above 1.22, the current use of combined therapy was not associated with an adverse effect on the incidence of myocardial infarction in postmenopausal women. (Arch Intern Med. 1994;154:1333-1339)

Abstract: Background: Elderly patients taking inappropriate drugs are at increased risk for adverse outcomes. We investigated the prevalence of inappropriate drug use and its predisposing factors in community-residing older persons. Methods: We conducted in-home interviews with 414 subjects aged 75 years and older living in the community of Santa Monica, Calif. Inappropriate medication use was evaluated using explicit criteria developed through a modified Delphi consensus process. These criteria identified drugs that should generally be avoided in elderly community-residing subjects regardless of dosage, duration of therapy, or clinical circumstances. Results: Based on these conservative criteria, 14.0% of the subjects were using at least one inappropriate drug. The most common examples were long-acting benzodiazepines, persantine, amitriptyline, and chlorpropamide. Subjects using three or more prescription drugs, compared with one or two, were more likely to be taking an inappropriate medication (odds ratio, 3.9; 95% confidence interval, 1.9 to 7.9). Furthermore, subjects with depressive symptoms had a higher risk of receiving inappropriate medications than nondepressive subjects (odds ratio, 2.2; 95% confidence interval, 1.1 to 4.1). Conclusions: Inappropriate drug use is a common problem in community-residing older persons. The risk of inappropriate drug use is increased in patients taking multiple medications and in patients with depressive symptoms. (Arch Intern Med. 1994;154:2195-2200)

Abstract: Background: Little information has been published on the impact of antihypertensive medications on quality of life in older persons. Particular concern has existed that lowering systolic blood pressure in older persons might have adverse consequences on cognition, mood, or leisure activities. Methods: A multicenter double-blind randomized controlled trial was conducted over an average of 5 years' followup involving 16 academic clinical trial clinics. Participants consisted of 4736 persons (1.06%) selected from 447 921 screenees aged 60 years and older. Systolic blood pressure at baseline ranged from 160 to 219 mm Hg, while diastolic blood pressure was less than 90 mm Hg. Participants were randomized to active antihypertensive drug therapy or matching placebo. Active treatment consisted of 12.5 to 25 mg of chlorthalidone for step 1, while step 2 consisted of 25 to 50 mg of atenolol. If atenolol was contraindicated, 0.05 to 0.10 mg of reserpine could be used for the second-step drug. The impact of drug treatment on measures of cognitive, emotional, and physical function and leisure activities was assessed. Results: Our analyses demonstrate that active treatment of isolated systolic hypertension in the Systolic Hypertension in the Elderly Program cohort had no measured negative effects and, for some measures, a slight positive effect on cognitive, physical, and leisure function. The positive findings in favor of the treatment group were small. There was no effect on measures related to emotional state. Measures of cognitive and emotional function were stable in both groups for the duration of the study. Both treatment groups showed a modest trend toward deterioration of some measures of physical and leisure function over the study period. Conclusions: The overall study cohort exhibited decline over time in activities of daily living, particularly the more strenuous ones, and some decline in certain leisure activities. However, mood, cognitive function, basic self-care, and moderate leisure activity were remarkably stable for both the active and the placebo groups throughout the entire study. Results of this study support the inference that medical treatment of isolated systolic hypertension does not cause deterioration in measures of cognition, emotional state, physical function, or leisure activities. (Arch Intern Med. 1994;154:2154-2160)

Abstract: Background: We examined prospectively the relation of time since stopping smoking with risk of coronary heart disease in middle-aged women. Methods: The study was based on 12 years' follow-up data (1976 through 1988) from the Nurses' Health Study, an ongoing prospective cohort study, in which information on smoking habits was updated every 2 years by a mailed questionnaire. The study population consisted of 117 006 female registered nurses aged 30 to 55 years in 1976 who were free of coronary heart disease, stroke, and cancer at baseline. The outcome measure used was incident coronary heart disease, defined as nonfatal myocardial infarction plus fatal coronary heart disease. Results: A total of 970 incident cases of coronary heart disease (215 among participants who never smoked, 214 among former smokers, and 541 among current smokers) occurred during 1.37 million person-years of follow-up. The multivariate relative risk of total coronary heart disease among current smokers compared with participants who never smoked was 4.23(95% confidence interval, 3.60 to 4.96). Risk of coronary heart disease was highest among smokers who started smoking before the age of 15 years (relativerisk, 9.25; 95% confidence interval, 5.27 to 16.23). The relative risk among former smokers was 1.48 (95% confidence interval, 1.22 to 1.79). On stopping smoking, one third of the excess risk of coronary heart disease was eliminated within 2 years of cessation. Thereafter, the excess risk returned to the level of those who never smoked during the interval 10 to 14 years following cessation. Conclusion: Women who stop smoking will experience an immediate benefit as well as a further longerterm decline in excess risk of coronary heart disease to the level of those who never smoked. (Arch Intern Med. 1994;154:169-175)

Abstract: Background: Several recent randomized clinical trials of anticoagulation in atrial fibrillation have demonstrated significant reduction in stroke rates with a small incidence of bleeding complications. The objective of this study was to determine whether the recommendations resulting from these trials have been implemented into routine practice, and if the anticoagulation control, therapeutic efficacy, and low complication rates achieved in the trials have been matched in community practice. Methods: We analyzed the anticoagulation practices and outcomes obtained for patients in atrial fibrillation at a large staff model health maintenance organization (HMO). We reviewed the medical records of all patients in atrial fibrillation as of April 1990. We compared demographic characteristics and clinical risk factors between HMO patients and those in the clinical trials. We also compared anticoagulation monitoring, adequacy of anticoagulation control, and clinical outcomes at the HMO with those achieved in the clinical trials. Results: Of 238 HMO patients in atrial fibrillation, 198 were without contraindications and therefore eligible for anticoagulation. Of these, 168 were offered anticoagulation (84.8%) and 156 were receiving anticoagulation therapy (78.8% of those eligible). The HMO patients had a greater prevalence of comorbidities than those in the clinical trials. The routine monitoring interval at the HMO was estimated at between 36.3 and 40.9 days (compared with 21 to 28 days reported in the clinical trials). The prothrombin time ratios at the HMO were in the target range on 50% of days compared with 68% of days in the clinical trials. The annual stroke and major bleeding rates in the HMO patients (1.3% and 0.6%, respectively) were not significantly different from the rates in the clinical trials (1.3% and 1.1%, respectively). The annual minor bleeding rate of 13.6% at the HMO was greater than the 7.8% to 8.4% rates in the two trials with better anticoagulation control (Boston Area Anticoagulation Trial for Atrial Fibrillation and Stroke Prevention in Atrial Fibrillation Study) but was not significantly different than the rates of 12.7% and 13.7% of the two trials with poorer anticoagulation control (Canadian Atrial Fibrillation Anticoagulation Study and Stroke Prevention in Nonrheumatic Atrial Fibrillation Study). Conclusions: Anticoagulation practices in this community setting appear to be good in that a large majority of patients were receiving anticoagulation therapy, and there were few major adverse outcomes. However, this study illustrates two common problems in attempting to apply the results of randomized clinical trials to routine practice: (1) differences between community patient populations and those on which the conclusions of clinical trials are based, and (2) less successful application of therapeutic interventions in settings other than that of a controlled clinical trial. The greater prevalence of comorbidities in the HMO patient population appears to convey a greater overall risk of thromboembolism and bleeding complications than in the clinical trials. In addition, the suboptimal anticoagulation control achieved at the HMO may increase the risks and decrease the potential benefits compared with those achieved in the clinical trials. Thus, theefficacydemonstrated in the clinical trials of anticoagulation in atrial fibrillation may not be directly translated intoeffectivenessin practice. (Arch Intern Med. 1994;154:1945-1953)

Abstract: Background: The purpose of this study was to investigate the association of glycemia with cause-specific mortality in a diabetic population. Methods: The study was a cohort design based in a primary care setting. Participants were all younger-onset diabetic persons (conditions diagnosed when they were younger than 30 years old and taking insulin, N=1210) and a random sample of older-onset diabetic persons (conditions diagnosed when they were 30 years or older, N=1780). Glycosylated hemoglobin levels were obtained at baseline examinations in 1980 to 1982 in which 996 younger-onset and 1370 older-onset persons participated. Median follow-up was 10 years in younger-onset and 8.3 years in older-onset persons; four younger-onset and two older-onset persons were unavailable for follow-up. The main outcome measure was cause-specific mortality as determined from death certificates. Results: In the younger-onset group after controlling for other risk factors in proportional hazards models and considering underlying cause of death, glycosylated hemoglobin was significantly associated with mortality from diabetes (hazard ratio [HR] for a 1% change in glycosylated hemoglobin, 1.25; 95% confidence interval [CI], 1.13 to 1.38) and ischemic heart disease (HR, 1.18; 95% CI, 1.00 to 1.40). In the older-onset group, glycosylated hemoglobin was significantly associated with mortality from diabetes (HR, 1.32; 95% CI, 1.21 to 1.43), ischemic heart disease (HR, 1.10; 95% CI, 1.04 to 1.17), and stroke (HR, 1.17; 95% CI, 1.05 to 1.30), but not cancer (HR, 0.99; 95% CI, 0.88 to 1.10). Results for any mention of specific causes of death were similar. Conclusion: These results suggest possible benefit to the control of glycemia with respect to death due to vascular disease and diabetes. (Arch Intern Med. 1994;154:2473-2479)

Abstract: Objective: To test the hypothesis that the regular use of aspirin and other nonsteroidal anti-inflammatory drugs (NSAIDs) is negatively associated with the risk of subsequent colorectal cancer. Design: Case-control study with four age- and sex-matched control subjects for each incident colorectal cancer case. Population and Setting: Patient population of a large municipal teaching hospital in Atlanta, Ga. Main Outcome Measure: Odds of colorectal cancer as a function of aspirin, nonaspirin NSAIDs, and acetaminophen dispensed to the study population in the 4 years prior to incident colorectal cancer diagnosis. Main Results: The risk of colorectal cancer estimated by odds ratios decreased with inceasing days of exposure to aspirin linearly in a dose-dependent fashion (likelihood ratio statistic: for cumulative days, P P P =.001). There appeared to be a threshold above which nonaspirin NSAIDs afforded protection (likelihood ratio statistic: for cumulative days, P =.021; for cumulative dose, P =.019). Acetaminophen conferred no risk reduction. Conclusion: The results of previous experimental animal models, interventional case studies, and some but not all epidemiological investigations and the present data point toward a causal relationship between NSAID use and the prevention of cancer of the large bowel and rectum. Because of the potential gastrointestinal and renal side effects of NSAID use, particularly in the elderly, chemoprevention trials are now needed to allow riskbenefit analysis in populations at high risk for colorectal cancer. (Arch Intern Med. 1994;154:394-399)

Abstract: Background: The effectiveness of graduated compression stockings in prophylaxis of postoperative venous thromboembolism is unclear to many physicians. Surveys show there is considerable variability in their use and their perceived effectiveness. We undertook to establish, by a systematic overview of the literature, the effectiveness of graduated compression stockings in the prophylactic setting of postoperative venous thromboembolism. Methods: Studies published between 1966 and June 1992 were identified through the MEDLINE database, with a search in all languages, through reviews of Current Contents , and including references cited in identified articles. Articles were selected for initial analysis if they assessed the use of graduated compression stockings for the prophylaxis of venous thromboembolism, and if the stocking group was compared with an untreated or unconfounded control group. Criteria were established a priori, to select only studies with sound methods. On the basis of these criteria, 12 studies were identified for the combined analysis. Each eligible study was independently analyzed for the risk of development of deep venous thrombosis in the control and stocking groups, and this was expressed as an odds ratio. Odds ratios were combined across studies by means of the Mantel-Haenszel X 2 procedure. The data were analyzed separately for orthopedic surgery because of its high risk for venous thromboembolism. Results: Eleven of the 12 studies were in moderate-risk surgery (abdominal, gynecologic, and neurosurgery); the summary odds ratio was 0.28, which translates into a risk reduction of 68% (95% confidence interval, 53% to 73%), which is statistically significant ( P P =.17). Conclusions: The use of graduated compression stockings for prophylaxis of venous thromboembolism after moderate-risk surgery results in a significant risk reduction. It is unknown whether the use of graduated compression stockings in combination with other forms of prophylaxis results in further risk reduction. The efficacy of graduated compression stockings in orthopedic surgery has been assessed by only one study that used sound methods, hence no definitive conclusions can be made in these high-risk patients. (Arch Intern Med. 1994;154:67-72)

Abstract: Background: Active cigarette smoking has been established as a potent risk factor for carotid atherosclerosis in clinical populations; however, neither the role of active smoking in general populations nor the impact of environmental tobacco smoke has been well described. Methods: The association between carotid artery wall thickness and cigarette smoking was studied in 12 953 black and white men and women, aged 45 to 65 years, examined in the Atherosclerosis Risk in Communities Study. Participants were classified as current smokers (n=3525), past smokers (n=4315), never smokers reporting weekly exposure to environmental tobacco smoke (ETS or "passive smoking") of at least 1 hour (n=3339), or never smokers reporting no weekly exposure to ETS (n=1774). Carotid artery intimal-medial thickness (IMT) was measured by B-mode ultrasound. Results: Increased IMT was observed in each category, in order from smallest to greatest increase: never smokers not exposed to ETS, never smokers exposed to ETS, past smokers, and current smokers. The larger IMT observed in the nonsmoking group exposed to ETS compared with the nonsmokers not exposed persisted after control for diet, physical activity, body mass index, alcohol intake, education, and major cardiovascular risk factors. Among past and current smokers, increased packyears of exposure was associated with increased IMT. Among nonsmoking men exposed to ETS, there was a significant increase in IMT with increasing number of hours per week of ETS exposure. Conclusions: These data confirm the strong relationship between active smoking and carotid artery IMT and provide initial evidence that passive smoking exposure is related to greater IMT. Increasing exposure to cigarette smoke (either pack-years of active smoking or hours of ETS) was significantly related to increased IMT. (Arch Intern Med. 1994;154:1277-1282)

Abstract: Organophosphate insecticides may cause serious poisoning either accidentally or by deliberate ingestion. Toxic symptoms are produced by acetylcholine accumulation at cholinergic receptors. Diagnosis is based on history of exposure or ingestion, symptoms and signs of cholinergic overactivity and a decrease in serum pseudocholinesterase levels. Following diagnosis, grading of disease severity may identify patients with serious poisoning who should receive treatment in intensive care using adequate doses of anticholinergic drugs. Complications, particularly ventricular arrhythmias, central nervous system depression or seizures, and respiratory failure, should be anticipated and treated. Relapse may occur after seemingly successful treatment. Public education with regard to symptoms of toxicity must be encouraged, and physicians must provide skilled treatment for a potentially lethal condition. (Arch Intern Med. 1994;154:1433-1441)

Abstract: There new studies help clarify important clinical issues regarding antithrombotic therapy for stroke prevention in patients with atrial fibrillation. The European Atrial Fibrillation Trial compared the efficacy of oral anticoagulation, aspirin, and placebo for stroke prevention in patients with atrial fibrillation with a recent stroke or transient ischemic attack. The results of the Stroke Prevention in Atrial Fibrillation II trial, which compared the efficacy of warfarin and aspirin, provide new information regarding the risk of intracranial hemorrhage in elderly patients with atrial fibrillation. Finally, an analysis of pooled data from the first five randomized trials identified clinical features that are predictive of stroke risk in individual patients with atrial fibrillation. These studies address unanswered questions regarding atrial fibrillation and stroke and have significant implications for patient management. (Arch Intern Med. 1994;154:1443-1448)

Abstract: Background: Epidemiologic studies on deep-vein thrombosis (DVT) have been mainly confined to the inpatient population. The aim of this study was to investigate the association between DVT and acquired risk factors in a large cohort of outpatients with clinically suspected DVT. Methods: Consecutive outpatients with clinically suspected DVT were enrolled in the study. Before objective testing, all patients were interviewed by a trained physician for the presence of risk factors for DVT development. Subsequently, the presence or absence of DVT was assessed with venography. Results: Approximately 50% of cases of DVT were considered to be secondary to a major risk factor (immobilization, trauma, and/or recent surgery). Among additional risk factors, only increased age (over 60 years), male gender, malignant neoplasm, heart failure, systemic lupus erythematosus, and arteriopathy were independently associated with the risk of acute DVT. Conclusion: Major risk factors for venous thromboembolism are a common cause of DVT among symptomatic outpatients; therefore, the usefulness of extending DVT prophylaxis in the outpatient setting should be tested. The role of additional risk factors in the development of DVT needs to be established by properly designed studies. (Arch Intern Med. 1994;154:164-168)

Abstract: Background: Whether the association between excess body weight and dyslipidemia is consistent across different age ranges in women has yet to be determined Methods: The relationship between body weight adjusted for height as calculated by body mass index (BMI; kilograms per square meter) and serum lipid and lipoprotein levels in white women was examined using crosssectional data from the Second National Health and Nutrition Examination Survey Mean lipid levels were determined for six different categories of BMI: (1) 210 or less; (2) 211 to 230; (3) 231 to 250; (4) 251 to 270; (5) 271 to 300; and (6) more than 300, and three age groups: premenopausal women, 20 through 44 years; perimenopausal women, 45 through 59 years; and postmenopausal women, 60 through 74 years Results: Compared with BMI category 2, a BMI in category 5 for premenopausal women was associated with 046 mmol/L (18 mg/dL) higher total cholesterol levels, 068 mmol/L (26 mg/dL) higher non—high-density lipoprotein (HDL) cholesterol levels, and 044 mmol/L (17 mg/ dL) higher low-density lipoprotein (LDL) cholesterol levels For perimenopausal women and postmenopausal women the same change in BMI was associated with much smaller differences in total cholesterol of 016 and 016 mmol/L (6 and 5 mg/dL), non-HDL of 024 and 020 mmol/L (9 and 8 mg/dL), and LDL levels of 013 and 003 mmol/L (5 and 1 mg/dL) More impressively, rising BMI was associated with consistently higher triglyceride levels of 054 to 040 mmol/L (48 to 35 mg/dL) and consistently lower HDL levels of 023 to 013 mmol/L (9 to 5 mg/dL), in all three age groups Conclusion: For young women, excess body weight was associated with higher total, non-HDL and LDL-cholesterol levels, higher triglyceride levels, and lower HDL-cholesterol levels In older women, although similar differences in triglyceride levels and HDL-cholesterol levels were observed, excess body weight was associated with smaller differences in total, non-HDL, and LDL cholesterol More striking than the weightassociated differences in total, non-HDL, and LDL-cholesterol levels were the differences in these lipid parameters observed with age alone Specifically, age category differences were twofold to eightfold greater than differences observed between categories of BMI within a given age Nevertheless, because the lower HDL cholesterol concentrations associated with excess body weight were age independent, total cholesterol-HDL cholesterol ratios were highest in obese postmenopausal women Although age and hormonal status are important affecters of lipoprotein risk factors, body weight also worsens the degree of dyslipidemia in white women (Arch Intern Med 1994;154:401-410)

Abstract: Background: Because adjuvant tamoxifen citrate is given to women with early-stage breast cancer for long periods, it is important to know how it affects risk factors for osteoporotic bone fractures, particularly since rates of bone fracture increase rapidly with age in postmenopausal women. In a 2-year randomized placebo-controlled toxicity study in 140 subjects, we demonstrated that tamoxifen was associated with preservation of bone mineral density (BMD), a major risk factor for fractures, in the lumbar spine. Methods: Five years after entry on this study we reexamined 62 of the original subjects with lumbar spine BMD and serum osteocalcin measurements. These were women available for study because they had not suffered major illnesses and had continued to receive (1) tamoxifen or (2) the no-tamoxifen regimen that they had originally been randomized to receive for the entire 5 years. Results: For lumbar spine BMD at baseline, the 30 subjects in the long-term tamoxifen group and the 32 subjects in the long-term no-tamoxifen group were not significantly different ( P =.26). During the first 2 years of follow-up, the 30 subjects in the long-term tamoxifen group showed the same BMD pattern as the entire 70-patient tamoxifen cohort, and similarly the 32 subjects in the long-term no-tamoxifen group showed the same pattern as the entire 70-patient cohort who received placebo. Five-year mean BMD measurements for each long-term follow-up group showed no significant changes from their respective 2-year levels. However, 5-year BMD measurements between the two groups differed (tamoxifen group, +0.8%; placebo group, —0.7%) (P=.06), and the mean regression lines for the changes in BMD over 5 years differed significantly between the two groups ( P =.0005). Adjustment for differences in body mass index, reported exercise, and calcium supplementation between these two groups did not change these results. Osteocalcin levels, also comparable at baseline in the two groups, were significantly lower in tamoxifen-treated subjects at 5 years ( P =.0005). Conclusions: While remodeling of bone may be lower, resorption of lumbar spine bone mineral is also lower, and tamoxifen preserves BMD in the lumbar spine over 5 years of treatment in postmenopausal women. Over longer periods, this preservation of BMD might be expected to be associated with lower fracture rates. (Arch Intern Med. 1994;154:2585-2588)

Abstract: Background: Mycobacterium marinum is a rare cause of skin infections, and its treatment has been based primarily on the personal experience and preferences of individual investigators without the benefit of large studies. Methods: Thirty-one patients with confirmed M marinum infection were identified at 33 Kaiser Permanente Northern California Region medical centers by microbiologic records, and their charts were reviewed. Results: The upper extremity was affected in 90% of cases, and lymphatic or local spread was seen during the initial examination or during observation in 25 patients (81%). Granuloma was present in 22 (63%) of 35 biopsy specimens, and staining for acid-fast bacteria yielded positive results in two of 22 specimens. Cure or improvement occurred in 22 (81%) of 27 patients in whom outcome could be evaluated. Treatment with ethambutol plus rifampin appeared more successful (effective in five [100%] of five cases) than minocycline treatment (effective in 10 [71%] of 14 cases), although not significantly so ( P =.28). Adverse reactions, most of which were gastrointestinal, occurred in five patients (18%). Conclusion: Ethambutol plus rifampin appears more useful than minocycline in treating cutaneous M marinum infection. This result remains to be confirmed by larger clinical studies, which may be difficult because this infection is relatively rare. (Arch Intern Med. 1994;154:1359-1364)

Abstract: Exogenous lipoid pneumonia is an uncommon condition resulting from aspirating or inhaling fatlike material, such as mineral oil found in laxatives and various aerosolized industrial materials These substances elicit a foreign body reaction and proliferative fibrosis in the lung Because symptoms are absent or nonspecific and the roentgenographic findings simulate other diseases, exogenous lipoid pneumonia is often unrecognized Yet, appropriate historical inquiries and simple laboratory tests can lead to the correct diagnosis, removal of the offending agent, and, potentially, improvement in lung function before serious complications develop (Arch Intern Med 1994;154:686-692)

Abstract: Background: Counseling patients about the risks and benefits of in-hospital cardiopulmonary resuscitation (CPR) can potentially reduce patient suffering and hospital costs. However, there is currently much disagreement regarding the overall rate of in-hospital CPR survival and characteristics that identify patients more or less likely to survive CPR. Methods: The charts of all adults who were pulseless and received basic CPR at a 720-bed university hospital during 1990 and 1991 were reviewed. Patients were excluded if cardiac arrest occurred outside the hospital or in the emergency department, operating room, recovery room, or cardiac catheterization laboratory. Each patient's chart was reviewed to determine the presence of explicitly defined clinical characteristics. Results: Overall, 50 (16.0%) of 313 patients survived to discharge. Before arrest, only impaired functional capacity and sepsis identified patients unlikely to survive CPR. Of adults suffering cardiac arrest during the study period, only 22% underwent CPR, including 13.0% of those with cancer and 18.1% of those 70 years or older. Conclusions: The use of do-not-resuscitate orders to exclude patients who were inappropriate candidates for CPR may explain why the survival rate reported here is higher than similar reports and why more clinical characteristics were not found to predict CPR survival. Investigators of inhospital CPR should use explicit criteria to describe the conditions studied and report survival for patients who receive basic CPR. The impact of do-not-resuscitate orders on survival rates must be considered. Functional capacity deserves further investigation as a predictor of CPR survival. ( Arch Intern Med . 1994;154:2426-2432)

Abstract: Background: Hypertension is the most common medical disorder associated with obesity. The relationship between dietary weight loss and the reduction of blood pressure is well established. However, the effect of gastric bypass surgery on blood pressure has not been well studied. Methods: We evaluated the relationship between weight loss and blood pressure in patients with diastolic hypertension who had gastric bypass surgery for morbid obesity. Patients were defined as hypertensive if taking antihypertensive medication or if both the preoperative office and mean hospital diastolic blood pressures were greater than 90 mm Hg. Two of the authors (J.L.C.,M.E.R.), blinded to all postoperative weights, classified the follow-up hypertensive status into one of four categories: resolved, improved, no change, or worse. The relationship between postoperative changes in blood pressure status and mean weight loss, percent excess weight loss, and body mass index were examined using a one-way analysis of variance. The relationship between postoperative weight loss and blood pressure was assessed in the baseline normotensive population using linear regression analysis. Results: There were 45 patients with diastolic hypertension; 91% were taking an antihypertensive medication. The mean follow-up was 39 months. The mean pre-operative weight was 137 kg and the mean weight loss at 1, 12, and 24 months following surgery was 13, 21, and 45 kg, respectively. Twelve months after surgery, hypertension had resolved in 22 patients (54%) and had improved in six patients (15%). These findings persisted through 48 months postoperatively. There was a significant relationship between the percentage of excess weight lost and improvement of hypertension at the 6-month and 12-month follow-up visits. There was also a significant relationship between the body mass index and improvement of hypertension at the 1-month, 12-month, 24-month, and 48-month follow-up visits. In the base-line normotensive patients there was not a significant relationship between our weight loss measures and changes in blood pressure. Conclusions: We conclude that postoperative weight loss in patients undergoing gastric bypass surgery was associated with resolution or improvement of diastolic hypertension in approximately 70% of cases. Resolution or improvement of hypertension occurred more often in patients with a lower postoperative body mass index. (Arch Intern Med. 1994;154:193-200)

Abstract: Background: The safety and effectiveness of different dosages and combinations of antihypertensive agents can be efficiently studied using a multifactorial trial design. In consultation with the Cardio-Renal Division of the Food and Drug Administration, we conducted a randomized, double-blind, placebo-controlled, 3×4 factorial trial of bisoprolol, a β1-selective adrenergic blocking agent, and hydrochlorothiazide. Methods: A total of 512 patients with mild to moderate essential hypertension were randomized to once-daily treatment with bisoprolol (0, 2.5, 10, or 40 mg), hydrochlorothiazide (0, 6.25, or 25 mg), and all possible combinations. Diastolic and systolic blood pressures were monitored during this 12-week trial. Results: The effects of bisoprolol and hydrochlorothiazide were additive with respect to reductions in diastolic and systolic blood pressures over the dosage ranges studied. The addition of hydrochlorothiazide (or bisoprolol) to therapy with bisoprolol (or hydrochlorothiazide) produced an incremental reduction in blood pressure. Dosages of hydrochlorothiazide as low as 6.25 mg/d contributed a significant antihypertensive effect. A hydrochlorothiazide dosage of 6.25 mg/d produced significantly less hypokalemia and less of an increase in uric acid levels than a dosage of 25 mg/d. The low-dose combination of bisoprolol, 2.5 mg/d, and hydrochlorothiazide, 6.25 mg/d, reduced diastolic blood pressure to lower than 90 mm Hg in 61% of patients and demonstrated a safety profile that compared favorably with that of placebo. Conclusions: The utility of factorial design trials to characterize dose-response relationships and to test the potential interactions between various antihypertensive agents has been demonstrated. The combination of low dosages of bisoprolol and hydrochlorothiazide may be a rational alternative to conventional stepped-care therapy for the initial treatment of patients with mild to moderate hypertension. (Arch Intern Med. 1994;154:1461-1468)