James E. Higgins

Expanded Clinical Evaluation of Lovastatin (EXCEL) study results: two-year efficacy and safety follow-up.

The Expanded Clinical Evaluation of Lovastatin study, a randomized, double-blind, placebo- and diet-controlled multicenter trial, evaluated the efficacy and tolerability of lovastatin over 48 weeks in 8,245 patients with moderately severe hypercholesterolemia. During year 1 of follow-up of the full cohort, lovastatin at 20 or 40 mg/day, or 20 or 40 mg twice daily, produced dose-dependent decreases in low-density lipoprotein (LDL) cholesterol (24% to 40%) and triglyceride levels (10% to 19%), and increases in high-density lipoprotein (HDL) cholesterol (6.6% to 9.5%). In all, 977 patients continued their original blinded treatment for an additional year. In year 2, the LDL cholesterol response to lovastatin was maintained, the triglyceride reductions were somewhat less, and the increases in HDL cholesterol were moderately greater than in year 1. Successive transaminase elevations > 3 times the upper limit of normal were observed in only 1 patient in year 2, yielding a cumulative 2-year incidence of from 0.1% (placebo or lovastatin 20 mg/day) to 1.9% (lovastatin 80 mg/day). Myopathy occurred in only 1 patient during year 2, and over the 2-year study was observed rarely and only at lovastatin dosages of 40 and 80 mg/day. This study indicates that lovastatin maintains its efficacy over long-term follow-up, particularly in effectively lowering LDL cholesterol, is generally well tolerated, and has a favorable safety profile.

Expanded clinical evaluation of lovastatin (EXCEL) study results: IV. Additional perspectives on the tolerability of lovastatin

This randomized, double-blind, multicenter, diet-and-placebo-controlled study was designed to clarify the dose-response relationship of lovastatin therapy to lipid-modifying efficacy and drug-related adverse events. Exclusion criteria were minimized so that study patients were representative of the majority of patients with moderate hypercholesterolemia seen in medical practice. After 6 weeks on the American Heart Association Step 1 Diet, a total of 8,245 patients were randomly assigned to 48 weeks of treatment with diet and placebo or lovastatin at dosages of 20 or 40 mg once a day or 20 or 40 mg twice a day. All adverse events were monitored, with particular attention to evaluation of liver and muscle. Liver transaminase elevations suggestive of possible hepatotoxicity, defined as successive elevations in either aspartate transaminase or alanine aminotransferase greater than 3 times the upper limit of normal, occurred in equal numbers of placebo and lovastatin 20 mg/day treated patients (0.1%). The frequencies were higher in lovastatin 40 mg/day and 80 mg/day patient groups (0.9 and 1.5%, respectively). No patient was diagnosed as having clinically symptomatic hepatic dysfunction. Creatinine kinase (CK) elevations above the upper limit of normal occurred frequently in placebo- (29%), as well as lovastatin-treated patients (29-35%), and muscle symptoms were reported with similar frequency in all groups (7-9%). The combination of muscle symptoms with marked CK elevations (greater than 10 times the upper limit of normal) was seen in only five patients: one in a 40 mg/day dose group and four in the 80 mg/day dose group. No patient developed rhabdomyolysis. The incidence of clinical and laboratory adverse events requiring discontinuation was 6% for the placebo group and from 7% (20 mg/day) to 9% (80 mg/day) for lovastatin treatment groups. No new types of adverse experiences related to lovastatin treatment were reported. Lovastatin, as an adjunct to diet for the reduction of elevated LDL cholesterol, was generally very well tolerated.

Expanded clinical evaluation of lovastatin (EXCEL) study : design and patient characteristics of a double-blind, placebo-controlled study in patients with moderate hypercholesterolemia

The randomized, double-blind, placebo-controlled trial described in this report was undertaken to clarify the dose-response relation of lovastatin therapy to lipid-modifying efficacy (lipid/lipoprotein modification) and drug-related adverse events in a population with moderately elevated fasting plasma total cholesterol (240 to 300 mg/dl) and low-density lipoprotein cholesterol (greater than or equal to 160 mg/dl). Men or women (postmenopausal or surgically sterile), aged 18 to 70 years, were entered into the trial with minimal exclusion criteria. After 4 to 6 weeks of an American Heart Association phase I diet or a more stringent diet, 8,245 patients from 362 sites were randomized to 1 of 5 parallel diet and drug treatment groups: placebo (n = 1,663) or lovastatin, 20 mg (n = 1,642) and 40 mg (n = 1,645) with the evening meal, and 20 mg (n = 1,646) or 40 mg twice daily (n = 1,649). The regimen of diet and lovastatin (or placebo) was followed for 48 weeks. The 5 treatment groups were similar at baseline. The total cohort had the following characteristics: 59% were men (mean age 56 years); 92% were white; 59% had completed at least 1 year of education beyond high school; 57% had a history of cardiovascular and associated disease; 40% had a history of hypertension; and 29% had coronary artery disease. Health habits were similar among groups, with 18% of patients reporting cigarette smoking, 14% reporting that they consume greater than 1 alcoholic beverage daily and 67% reporting no strenous exercise. Mean lipid/lipoprotein levels were also similar among groups, with the following average levels: total cholesterol (258 mg/dl), low-density lipoprotein cholesterol (180 mg/dl), high-density lipoprotein cholesterol (45 mg/dl) and triglycerides (median = 155 mg/dl). The large size of this trial, its placebo-controlled, double-blind design and the similarity of treatment groups at baseline should allow clear documentation of the long-term effects of lovastatin treatment and generalization of the results to a substantial portion of patients who may be candidates for lipid-modifying therapy.

The effect of desogestrel gestodene and other factors on spotting and bleeding.

Spotting and bleeding are among the most common side effects associated with oral contraceptive (OC) use and their occurrence is a prime determinant of whether a new user will continue to use OCs. Desogestrel and gestodene are two new progestins that were developed in part to minimize the occurrence of these side effects. Assessing the effect of these progestins is difficult, however, in part because their effects may be subtle, requiring a large sample size and possibly being overshadowed by other factors. To address these issues, we analyzed data from two comparative multicenter clinical trials that included 15,421 cycles among 2767 women. One study compared 75 micrograms gestodene + 30 micrograms ethinyl estradiol (EE) with 150 micrograms desogestrel + 30 micrograms EE, the other compared the same gestodene preparation with 150 micrograms desogestrel + 20 micrograms EE. Both studies found a higher risk of spotting or bleeding in all cycles among users of the desogestrel-containing preparation, with the differences ranging between 20% and 70% higher for the first study and 40% and 140% in the second. These differences were statistically significant in four of six cycles in each study and persisted after controlling for consistency and recency of OC use as well as smoking. After pooling the data and controlling for estrogen dose, the desogestrel-containing preparation was significantly associated with more frequent spotting or bleeding in five of six cycles. Smoking and consistency and recency of OC use were also independent predictors of spotting or bleeding.

Parity and use-effectiveness with the contraceptive sponge

The results of a randomized United States study indicated that the Today contraceptive sponge was less effective than the diaphragm (1-year cumulative life-table rate of 17.4 versus 12.9 pregnancies per 100 women, p = 0.01). However, this overall comparison is misleading. Using univariate and multivariate analyses to account for the effects of user characteristics we found parity to be the most important single determinant of effectiveness for users of the sponge, but parity was unimportant as a risk factor for pregnancy among diaphragm users. For nulliparous women the sponge was as effective as a physician-prescribed barrier method (13.9 for sponge, 12.8 for diaphragm, p = 0.45); however, parous women using the sponge were twice as likely to become pregnant (28.3 for sponge, 13.4 for diaphragm, p = 0.001). The effect of parity among sponge users is consistent with the results of international studies of the contraceptive sponge.

The effect of birth interval on perinatal survival and birth weight

Abstract Although the effect of a short birth interval on the first child in a pair has received attention in the literature, the effect on the second child has received less. In this article the authors investigate the complex set of relationships between birth interval, maternal age and parity, and their effects on the birth weight and survival of the later-born child. The data consist of 12,995 singleton births to women of parity two or higher during1977 and 1978 in a single hospital. The outcome of the previous pregnancy is controlled by restricting the analysis to women whose previous pregnancy ended in a live infant who is still living at the time of the index birth. The effect of birth interval on birth weight and on survival is examined simultaneously(via logistic regression), with the effects of maternal age and parity. The risks of adverse outcomes as a function of birth interval are estimated by adjusted odds ratios. After adjusting for maternal age and parity, interval was found to be an important precursor of both perinatal mortality and low birth weight. At all levels of maternal age and parity, babies born during a 9- to 12-month birth interval are at greater risk of low birth weight and/or perinatal mortality than babies born after a longer birth interval.

Delivery type and neonatal mortality among 10,749 breeches.

Data on 10,749 breech presentations were analyzed for the effect of delivery type on neonatal mortality. Most of the data are from developing countries, and most of the hospitals have higher mortality than is found in Europe or the United States. The simultaneous effect of type of hospital where the delivery occurred, type of breech, birthweight, and parity were examined. The benefit of cesarean delivery was greater for nulliparae than multiparae, greater for footlings than for frank or complete breeches, and greater for larger babies than smaller ones. This last finding probably reflects the quality of neonatal care in developing country hospitals rather than the value of cesarean section. Maternal mortality and morbidity was higher among women delivered abdominally than among those delivered vaginally.

Effects of Breastfeeding on IUD Performance

The effect that lactation might have on intrauterine device (IUD) performance was investigated by using data from a series of multicenter clinical trials. Life-table methods were applied to compare breastfeeding and non-breastfeeding women with respect to IUD expulsion, accidental pregnancy, IUD removal for various reasons, and overall continuation of IUD use. Results indicate that breastfeeding does not increase the risk of expulsion or other events, whether the device is inserted immediately (within ten minutes) or more than 42 days after delivery. (Am J Public Health 1983; 73:384-388.)

Sterilization in the northeast of Brazil

Results of household surveys carried out in four states in the Northeast of Brazil showed that female sterilization is the most prevalent method among women who want no more children. Many women who indicated an interest in tubal ligation, however, had not been sterilized. Access to sterilization depends on the type of hospital in which the child was delivered and the type of delivery the woman had. Over 60% of the sterilized women reported that they had a tubal ligation at the time they were hospitalized for a cesarean delivery. Compared with unsterilized women, the sterilized women are characterized by relatively higher education levels and a greater likelihood of urban residence and were far more likely to have had cesarean deliveries and, therefore, to have qualified for postpartum sterilization on medical grounds.

Statistical comparison of Pearl rates

Statistical procedures for hypothesis testing and interval estimation of the difference in a pair of Pearl rates are presented. The p value of the test statistic is evaluated exactly from the binomial distribution or approximately from the standard normal distribution. Interval estimation is provided by the test-based method of Miettinen. The procedures are applied to two published data sets, and the conceptual link between the Pearl rate and life-table methods is discussed.