Limitations of the correlation coefficient in the validation of diet assessment methods. CODIAB-INSERM-ZENECA Pharma Study Group.

Dietary assessment methods usually are validated one against the other, using the correlation coefficient between the methods to assess agreement. The correlation coefficient, however, is only an indicator of relative agreement and depends on the range of intakes. As this dependence can lead to misinterpretation, we propose that the standard deviation of the difference between methods should be used as an indicator of absolute agreement and the standard deviation of their average as an indicator of the range of intakes. We illustrate this method in a validation study of a diet history against a 3-day diet record in 58 non-insulin-dependent diabetes mellitus patients. A re-analysis of studies from the literature validating food frequency questionnaires against diet records showed that the absolute agreement between methods is very consistent and that the differences in the correlation coefficients are mostly due to the selection of subjects.

The relationship between behavioural pattern, overall and central adiposity in a population of healthy French men.

The relationship between obesity, both general obesity (BMI) and central obesity (measured by the ratio of iliac and thigh circumferences), and a behavioural pattern, which includes alcohol consumption, smoking, stress and lack of sporting activity, was investigated in 467 middle-aged French working men. BMI and central obesity were closely correlated (r = 0.34). Alcohol consumption was positively associated with central obesity (P < 0.001) but it did not significantly influence BMI. BMI decreased with cigarette smoking (P < 0.001), but central obesity only increased significantly after adjustment for BMI (P < 0.05); thus for a given BMI, smoking was associated with a greater degree of central adiposity. Resting heart rate (considered as a measure of stress) was positively associated with both BMI and the iliac-thigh ratio (P < 0.01, P < 0.05), the latter association being due to general obesity. Sporting activity and BMI were only marginally associated (P < 0.07) but central obesity decreased with activity (P < 0.01). Jointly, alcohol consumption, cigarette smoking, the resting heart rate and sporting activity explained 8% of the variance in the iliac-thigh ratio; after adjustment for BMI these behavioural variables still explained 6% of the variance. Central adiposity, which has recently been described as a risk factor for coronary heart disease, non-insulin dependent diabetes and hypertension, was significantly associated with a potentially modifiable behavioural pattern.

Hypocaloric diet and antihypertensive drug treatment. A randomized controlled clinical trial.

OBJECTIVE: To compare the amount of drug quantified by a score needed to control blood pressure in two groups of overweight hypertensive patients, receiving or not receiving a hypocaloric diet. DESIGN: Randomized controlled clinical trial. SETTING: Two hospital outpatient hypertension clinics. PATIENTS: Fifty-four subjects with a DBP between 95 and 110 mmHg and a weight > or = 110% of the ideal weight. INTERVENTION: Allocation to either drug treatment (DT) or hypocaloric diet (HD). In the HD group, after 2 months, an antihypertensive drug was administered to the subjects with a DBP > or = 90 mmHg, following the same scheme protocol as in the DT group. Subjects were followed during 10 months by a clinician only in the DT group and by a clinician and a dietician in the HD group. MAIN OUTCOME MEASURES: Score of treatment: hydrochlorothiazide 25 mg [score = 1] with, as needed to obtain a DBP < 90 mmHg, the addiction of enalapril 10 mg [score = 2], 20 mg [score = 3], and nifedipine 40 mg [score = 4]. RESULTS: At the end of the trial, 5 subjects were lost to follow-up in the HD group and 1 in the DT group (p > 0.05). Mean weight loss was 5.9 kg (sd = 1.2) in the HD and 2.3 kg (sd = 0.7) in the DT group (p = 0.02). Mean decrease in DBP was 18 mmHg (sd = 7) and 15 mmHg (sd = 8) in HD and DT groups respectively (p = 0.36). Mean DBP was 84 mmHg (sd = 7.8) in the HD group and 85 mmHg (sd = 7.2) in the DT group. In the HD group, 8 (38.1%) subjects had a DBP < 90 mmHg without any drug treatment. The mean drug treatment score was 0.86 (sd = 0.91) in the HD and 1.52 (0.70) in the DT group (p = 0.01). CONCLUSION: This study shows that in overweight hypertensives, the quantity of drug needed to achieve an acceptable level of BP is nearly reduced by 50% when an efficient hypocaloric diet is prescribed simultaneously.

[The quantitative alimentary questionnaire technique used by the I.N.S.E.R.M. nutrition department (author's transl)]. / La technique du questionnaire alimentaire quantitatif utilisé par la section nutrition de l'I.N.S.E.R.M.

The utilization, during a certain number of years, of quantitative alimentary questionnaire in epidemiologic studies, has allowed: to model a print allowing the interviewer to transcribe the food ingested into an average for each day by a single person, to reflect on the different methods capable of being adaptable to reach this aim and to index the rocks to keep out of, so that this method would be used with chances of reliability.

[Food consumption in relation to size of collectivity. Parallel time trends (author's transl)]. / La consommation alimentaire des Français en fonction de la taille des agglomérations, reflet de l'évolution longitudinale.

The consumption of the French with respect to the size of parishes and its evolution from 1965 to 1974 can be investigated, thanks to the survey of Institut National de la Statistique et des Etudes Economiques on food consumption of a representative sample of households. During this period, with regard to the growing size of parishes and with respect to time, the percentage of lipidic calories has been raising, chiefly from the decrease in the consumption of vegetable foods which carries with it, at the same time, a decrease of the intake of glucids, dietary fibers and of the total energy intake. The shortcomings of data collecting explain only a small part of the discrepancies observed.

Blood cells and alcohol consumption with special reference to smoking habits.

In the course of the Paris study on risk factors of cardiovascular disease in a large professional group, 7710 active and apparently healthy men aged between 48 and 54 were examined. This study measured the relationship between clinical abnormalities suggesting alcoholic liver disease (ALD) and the following blood parameters: white (WBC) and red (RBC) blood cell counts, haematocrit (H), and mean corpuscular volume (MCV), the former computed as H/RBC ratio. A subsequent analysis was performed on a random sample of 485 subjects without ALD who were questioned on their daily average alcohol consumption. Each subject was classified as ;smoker' or ;non-smoker' according to his daily tobacco consumption for the last five years. Analysis of the data confirmed that smoking and alcohol were related to the blood parameters; but, according to smoking habit, different relationships between alcohol consumption (or ALD) and MCV, RBC, or WBC counts were found: for smokers, RBC count significantly decreased and MCV increased with alcohol consumption (or ALD); for non-smokers, WBC count significantly increased with alcohol consumption (or ALD). So, it would be of interest to consider the relation between alcohol and tobacco in interpreting possible changes in blood parameters and in formulating hypotheses on the mechanisms of their specific action.

[Feasibility survey on the reliability of questionning on past eating habits in a group of men (author's transl)]. / Enquête méthodologique testant la validité d'un interrogatoire portant sur l'alimentation passée d'un groupe de sujets du sexe masculin

Among men 50-57 years old surveyed in 1969 and again in 1974 on their food intake, both questionaires of 37 who stated in 1974 that they had changed their eating habits in the preceding 5 years were selected for analysis. The mean reported caloric intake had dropped very significantly, but about 500 calories. When these persons were questioned in 1974 on their 1969 food consumption, the results compared to the 1969 questionary were identical for almost all food catergories and for total calories. Thus, a food consumption survey retrospective over a 5 years span appears no less reliable than a survey on present eating habits.