Hubert J.A. Schouten

Primary prevention of arterial thromboembolism in non-rheumatic atrial fibrillation in primary care : randomised controlled trial comparing two intensities of coumarin with aspirin

Abstract Objective: To investigate the effectiveness of aspirin and coumarin in preventing thromboembolism in patients with non-rheumatic atrial fibrillation in general practice. Design: Randomised controlled trial. Participants: 729 patients aged ≥60 years with atrial fibrillation, recruited in general practice, who had no established indication for coumarin. Mean age was 75 years and mean follow up 2.7 years. Setting: Primary care in the Netherlands. Interventions: Patients eligible for standard intensity coumarin (international normalised ratio 2.5-3.5) were randomly assigned to standard anticoagulation, very low intensity coumarin (international normalised ratio 1.1-1.6), or aspirin (150 mg/day) (stratum 1). Patients ineligible for standard anticoagulation were randomly assigned to low anticoagulation or aspirin (stratum 2). Main outcome measures: Stroke, systemic embolism, major haemorrhage, and vascular death. Results: 108 primary events occurred (annual event rate 5.5%), including 13 major haemorrhages (0.7% a year). The hazard ratio was 0.91 (0.61 to 1.36) for low anticoagulation versus aspirin and 0.78 (0.34 to 1.81) for standard anticoagulation versus aspirin. Non-vascular death was less common in the low anticoagulation group than in the aspirin group (0.41, 0.20 to 0.82). There was no significant difference between the treatment groups in bleeding incidence. High systolic and low diastolic blood pressure and age were independent prognostic factors. Conclusion: In a general practice population (without established indications for coumarin) neither low nor standard intensity anticoagulation is better than aspirin in preventing primary outcome events. Aspirin may therefore be the first choice in patients with atrial fibrillation in general practice. Key messages Studies have shown that patients with non-rheumatic atrial fibrillation may benefit from anticoagulation This study in a general practice population found no benefit of standard or low dose anticoagulation on risk of stroke, systemic embolism, major haemorrhage, or vascular death when compared with aspirin Hypertension and age were prognostic factors for event occurrence Aspirin is the treatment of choice for preventing thromboembolism in primary care patients at low risk

Tumor angiogenesis: An independent prognostic parameter in cervical cancer

OBJECTIVE The purpose of this study was to investigate whether the intensity of tumor angiogenesis, expressed as microvessel density, is a parameter predicting the probability of lymph node metastasis and survival in patients with cervical cancer stages IB and IIA. STUDY DESIGN In a retrospective study of 114 patients with cervical cancer stages IB and IIA, microvessel density, lymph node status, and several other prognostic parameters were correlated with disease-free survival by a multivariate analysis according to Cox proportional-hazards model. RESULTS There was a significant difference in mean microvessel density between tumors with and without pelvic lymph node metastasis (p = 0.002). Both microvessel density (vessels per square millimeter of stroma). (p = 0.05) and pelvic lymph node metastasis (p = 0.007) correlated significantly and independently with disease-free survival. CONCLUSION This study demonstrates that microvessel density is an independent prognostic parameter for disease-free survival in patients with cervical cancer stages Ib and IIa.

Interobserver variation in histopathological grading of cervical dysplasia

In order to assess the variability among histopathologists in grading cervical dysplasia, four experienced histopathologists examined the same set of 106 biopsy specimens and assigned them to one of five diagnostic categories. These were: no dysplasia, mild dysplasia, moderate dysplasia, severe dysplasia and carcinoma in situ. The histopathologists did not discuss the grading criteria beforehand. There was considerable disagreement among the pathologists: unweighted group kappa 0.28, weighted group kappa 0.56. It appeared that all grades of dysplasia were equally difficult to distinguish from adjacent categories. Various explanations for this interobserver variation are put forward.

The effect of beta-carotene on the regression and progression of cervical dysplasia: a clinical experiment.

In order to gain insight into the causality of the relation between beta-carotene and cancer, we performed a randomized placebo-controlled trial in which the effect of beta-carotene on the regression and progression rates of cervical dysplasia were examined. The experimental group (n = 137) received a supplemental dose of 10 mg of beta-carotene daily for 3 months. The control group (n = 141) received placebo capsules. As the outcome parameter, two definitions of regression and progression were used, which were based on the degree of dysplasia before and after the medication period. The number of patients who showed progression was too small to allow conclusions. No effect of beta-carotene on the regression percentages was observed: OR = 0.68 (95% CI: 0.28-1.60) using the broad definition; and OR = 1.22 (95% CI: 0.43-3.41) with the strict definition. A secondary analysis, in which the effect of the total intake of beta-carotene (diet + medication) on the regression percentages of cervical dysplasia was studied, did not show a positive effect either. The paper discusses to what extent issues in the study design may have masked a potential effect and how our results affect the evidence for a causal relation between beta-carotene and cancer.

Sample size formula with a continuous outcome for unequal group sizes and unequal variances.

Guenther derived a simple formula for the accurate computation of equal sample sizes when the two-sample t-test with pooled variance estimate is used. In the present paper, the Guenther formula is generalized to the case of unequal sample sizes. In addition, the case of unequal variances is considered.

Sources of interobserver variation in histopathological grading of cervical dysplasia

The present study aimed to assess where the interobserver variation in grading cervical dysplasia stems from. Four experienced pathologists examined 93 histological slides, after they agreed on which morphological characteristics should be considered relevant for grading. They scored 6 morphological characteristics for each slide and assigned it to a degree of dysplasia. Compared to a previous study, the interobserver variation showed a statistically significant improvement: the weighted group kappa value increased from 0.55 to 0.69. For the scores of the individual characteristics considerable interobserver variation was observed: weighted group kappa values ranged from 0.28 to 0.49. The pathologists slightly differed in which characteristics they considered most important for their grading. The agreement on the degree of dysplasia turned out to be better than the agreement on the morphological characteristics on which this diagnosis is based. In the discussion, a few explanations for this paradoxical finding are put forward.

A simple analysis of a simple crossover trial with a dichotomous outcome measure.

For the two-treatment two-period crossover trial with a dichotomous outcome measure a simple and straightforward analysis is proposed: compute the treatment difference for each of the treatment sequences AB and BA, and then average over the sequence groups. The mean difference is used in the statistical test and in the computation of the 95 per cent confidence interval. The method is generalized to the case of a crossover trial including more than two treatments.

The Diagnostic Value of Scoring Models for Organic and Non-organic Gastrointestinal Disease, Including the Irritable-bowel Syndrome

Scoring models based on history and physical examination have been developed to dis criminate patients with non-organic gastrointestinal disease from those who have organic disease. The application of these models may lead to more efficient diagnosis and prevent somatization. Although the models have high diagnostic accuracy in the population in which they have been developed, their value in other populations has not been established. In this study previously developed models were tested in validation populations defined by the original selection criteria from the studies in which the models were developed and in un selected general practice and outpatient populations. The diagnostic performances of the models are expressed in terms of odds ratio and sensitivity and specificity for the classification of patients as having organic and non-organic disease. The diagnostic performances of all the models were rather low in the validation populations. Relatively few elements of the models had independent diagnostic value. In addition, the correlation between the scoring models, expressed in Cohens kappa, was extremely low. The diagnostic values of the scoring models were not reproduced in comparable and unselected populations. Therefore, it is concluded that the diagnostic value of such a model has little external validity. Key words: scoring models; gastrointestinal disease; validation; diagnosis; predictive validity. (Med Decis Making 1994;14:208—216)

Is a Series of Blood Pressure Measurements by the General Practitioner or the Patient a Reliable Alternative to Ambulatory Blood Pressure Measurement?* A Study in General Practice With Reference to Short-term and Long-term Between-Visit Variability

We studied the reproducibility of a series of blood pressure measurements by general practitioner (GP) and patient in comparison with that of ambulatory blood pressure measurement (ABPM), with reference to short-term and long-term between-visit variability using a prospective, comparative diagnostic study. The study group was 88 potentially hypertensive primary care patients (initial systolic blood pressure [SBP] between 160 and 200 mm Hg or with diastolic blood pressure [DBP] between 95 and 115 mm Hg). ABPMs were measured on 2 separate days (at a 6 month interval). Two series of measurements by the doctor (at 1 to 6 month intervals), and the patient (at a 1 week interval) were measured. Mean differences and standard deviations of mean differences (SDD) between two successive series of measurements, and between two ABPMs were computed. The Wilcoxon signed-ranks test was used to compare these standard deviations. Mean initial office-blood pressures were 161 (SBP) and 102 (DBP) mm Hg. Long-term between-visit variability (measurements by GP) was larger than short-term between-visit variability: SDDs were 16 v 11 mm Hg (SBP), and 10 v 8 mm Hg (DBP). The differences in average SBP and DBP between successive ABPMs and between successive series of office measurements by GP and home measurements by patient were not statistically significant. Mean differences between two series of measurements by GP and patient, and between two ABPMs, were 0 +/- 1 mm Hg. SDDs between successive ABPMs and series of measurements by GP and patient ranged from 8 to 11 mm Hg (SBP), and were 6 mm Hg (DBP). No statistically significant differences were found between the SDDs of the studied measurement procedures (SBP and DBP). In our study the reproducibility of ambulatory blood pressure measurement was not found to be better than that of a series of four duplicate measurements by GP or patient. Long-term (6 months interval) between-visit variability was larger than the short-term (1 week interval) between-visit variability.

Planning group sizes in clinical trials with a continuous outcome and repeated measures

Simple formulae have been derived for the accurate computation of sample sizes when the two-sample t-test is used. Required sample sizes may be reduced by taking repeated measures, especially if correlations are about equal. In a trial with two parallel groups, repeated measures may be summarized by a pre-treatment mean and a post-treatment mean. In a cross-over trial measures may be repeated within each treatment period. In case the difference between two independent groups increases in time, individual slopes may be used if the correlation between two measures hardly depends on distance in time between these measures. However, it may be more efficient to use only the first and last measure in the statistical test if correlations strongly depend on time distance.