MPM O'KANE. E J CLOSE. P G WILES. Ulniversity Dcpartment of Chemical ... MARGARET THOROGOOD. LIANE ROE. KLIM McPHERSON. JIM MANN.
Correlationis between plasma lipid concentrations and dietarn fat intake in diabetic patients Patients with type I diabetes (n= 13) r Value
p Value
Patients with type II diabetes (n= 37) r Value
p V'alue
0-169
-0-071
I'lasma cholesterol Fat as %i
energy
Cholesterol intake
0-410 - 0280 0-289
0-354 0 339
-0061 -0034
0-675 0-718 0-842
Fat as °benergv PolIunsaturated:saturated ratio
I'lasma tnrglycerides 0 443 0 129 -0-490 0 089
0-082 0 119
0-630 0-482
Polvunsaturated:saturated ratio
the three other groups (ranges 0 49-0-75 and 201-306 mg/day). We would be interested to know whether the correlations found in this study remain significant when vegans are excluded as these would be more representative of dietary intakes in outpatients and in the general population.
We examined correlations between plasma lipid concentrations and dietary intake in diabetic patients aged 17-81 years by using a seven day food record. We found similar correlations in 13 subjects with type I diabetes (which were not significant because of the low numbers) but not in 37 subjects with type II diabetes (table). Ms Thorogood and colleagues suggest that the reduction in saturated fat intake was compensated for by an increase in unsaturated fat intake rather than carbohydrate intake. This may not be so. All groups reduced their total fat intake compared with the national average of 42.2%2 (male vegans reduced their fat intake to as low as 33 5%). The national average carbohydrate intake is 44 3%, and the vegans (with high polyunsaturated to saturated fat ratios) obtained 52 5% of their energy from carbohydrate -a significant increase. If the general population were to reduce their saturated fat intake but maintain their total fat intake they would have to increase their intake of monounsaturates or polyunsaturates, or both. If these were increased in the same proportion as in vegans the percentage of energy from polyunsaturates would be 14 5%. This has been shown to lower plasma high density lipoprotein concentrations, predispose some patients to gall stones, and, in animals, suppress the immune system and promote development of tumours.' The percentage of energy from monounsaturates would be 19 9% if the general population adopted the vegan diet and 22 4% if this was adjusted to reduce intake of polyunsaturates to 12%-the accepted upper limit. This would prove an impractical diet to recommend. We conclude that it is still necessary to recommend to the general population a diet that obtains less energy from both total and saturated fat. J A LOCKT )N M P M O'KANE E J CLOSE P G WILES
Ulniversity Dcpartment of Chemical P'athology, l)cpartment of D)ietetics, University Dcpartment of Medicine, Lecds General Infirtoar\, Lecds LS2 9J11
[horogood M\, Roe L, AMclPherson K, Mann J. Dietary intake and plasma lipid lsevls: lessons from a study of the diet of health conscious groups. Br Med 7 1990;300:1297-301. 19 Mav. 2 Ministry of Agriculture, Fisheries, and Food. Household food conisumption and expenditure 1987. London: HMSO, 1989. 3 Gruotdy SM. l)ietar\ therapy for different forms of hyperlipoproteinaemia. Circulation 1987;76:523-8.
AUTHORS' REPLY,-We chose the dietary groups that we studied, including the vegan group, to ensure as wide as possible a range of dietary intakes among subjects who were otherwise exposed to a similar social environment. We consider this to be a strength rather than a weakness of our study. We do not claim that these subjects were in any way
44-
representative of outpatients or, indeed, of the general population, and we believe that we made this clear. Dr J A Lockton and colleagues ask whether the correlations that we reported would still be significant if we excluded vegans from the analysis. We excluded the data on vegans and recalculated the simple Pearson coefficient of correlation between the Keys dietary score and total plasma cholesterol concentration. This gave a correlation of 0-22 (p
