Low carbohydrate diets – two RCTs and a systematic review
Low carbohydrate diets seems to have replaced house prices as a topic of dinner table conversation. Arguments rage between weight lost (spectacular numbers of stones, pounds or kilograms) and harm done (cholesterol raised, heart attacks suffered). The trouble is that some iron-willed folk would probably make any diet work through rigid calorie restriction, while others are so unfit that an exciting episode of a soap opera on TV might bring on a heart attack.
What is the truth? A difficult one, because until recently anecdote trumped any other real evidence we had. Now though, a systematic review and two randomised trials light up the darkness with a few rays of evidence. What we need to know is whether people can stick to the diet, whether they lose weight, and whether there are any untoward effects arising from the diet.
Systematic review 
A search strategy was used to locate any studies in English that examined the use of low carbohydrate or ketogenic diets. Any study architecture was eligible, and eventually reports of 94 interventions were located. As many as half were very small, with fewer than 30 participants, and some studies had as few as two participants. The largest study had 162 participants, and few had as many as 100.
Some trials were randomised. In most studies the duration of the diet was relatively limited, with few lasting as long as 100 days. Few studies examined daily carbohydrate consumption as low as 60 grams/day or less, and only five studies evaluated such low carbohydrate consumption for more than 90 days.
Given all this, any analysis could be little more than speculative. What emerged, though, was that among obese people, weight loss was associated with longer diet duration and restriction of calorie intake rather than carbohydrate content. There was no significant effect on serum lipids, fasting glucose or insulin levels, or blood pressure.
RCT in 132 people with severe obesity 
Here 132 people with a mean BMI of 43 kg/sq metre were randomised to six months of either a low carbohydrate diet of 30 g/day or less, or a low fat US guideline diet with a 500 calorie daily deficit and fat content of 30% or less of total calories. The diet groups attended weekly teaching groups for a month, and monthly sessions for five months. No specific exercise programme was recommended.
The average age of subjects was 53 years, and their average weight was 131 kg. About 40% had diabetes, 60% were on antihypertensive treatment, and 40% were treated for hyperlipidaemia. Subjects on the low carbohydrate diet reduced their calorie consumption by an average of 460 calories a day to 1,600 calories per day; in the low fat diet the reduction was less, by 270 calories per day, but again to 1,600 calories per day.
More people stayed on the low carbohydrate diet than on the low fat diet (Table 1), though the difference was not significant. The average weight loss was greater in the low carbohydrate (6 kg weight loss) than in the low fat diet (2 kg weight loss, Table 1). A weight loss of 10% or more occurred in 9 of 64 subjects on the low carbohydrate diet, compared with 2 of 68 subjects on the low fat diet. This implies a number needed to treat with low carbohydrate diet for six months of 9 (5 to 58) for a 10% weight reduction, compared with a low fat diet.
Table 1: Low carbohydrate and low fat diets compared – subjects remaining on the diet and mean weight loss over six months
Percent remaining Mean weight reduction (kg)
Month Low carbohydrate Low fat Low carbohydrate Low fat
2 75 62 4.0 1.0
4 73 56 5.5 2.5
6 67 53 5.7 1.8
The major difference in serum lipids was a 20% reduction in serum triglyceride in those on the low carbohydrate diet. With the low fat diet there was an average reduction of 4%. There were no changes in total or lipoprotein cholesterol in either group.
Diabetic subjects on the low carbohydrate diet reduced fasting glucose by 9% in the low carbohydrate diet compared an average reduction of 2% on the low fat diet. Seven subjects in the low carbohydrate group had dose reductions in oral hypoglycaemic drugs or insulin by six months, while on the low fat diet one subject had had an insulin dose reduction, and one started oral therapy. This implies a number needed to treat with low carbohydrate diet for six months of 4 (2 to 10) for a reduction in diabetic therapy, compared with a low fat diet.
There were no changes in blood pressure with either diet. Two patients, both on the low carbohydrate diet, had hospital admissions. One was for chest pain unrelated to cardiac ischaemia, and the other was hyperosmolar coma due to poor compliance with diabetes therapy.
RCT in 63 people with obesity 
Here 63 people with a mean BMI of 34 kg/sq metre were randomised to 12 months of either a low carbohydrate diet (Atkin’s diet), or a conventional diet with calorie restriction deficit and high carbohydrates, with fat content of 25% or less of total calories. Subjects had no counselling or other professional help, and none had diabetes.
The average age of subjects was 44 years, and their average weight was 98 kg. More people stayed on the low carbohydrate diet than on the conventional diet (Table 2), though the difference was not significant. The average weight loss was greater in the low carbohydrate than in the conventional diet at three and six, but not 12 months (Table 2).
Table 2: Low carbohydrate and conventional diets compared – subjects remaining on the diet and mean weight loss over 12 months
Percent remaining Mean weight reduction (kg)
Month Low carbohydrate Conventional diet Low carbohydrate Conventional diet
3 85 70 8.0 3.7
6 73 60 9.6 5.2
12 61 57 7.2 4.4
The low carbohydrate, but not the conventional diet, led to a 20% reduction in serum triglyceride, and about a 20% increase in HDL cholesterol, but not total or LDL cholesterol. There was no change in blood pressure in either group.
What are we to learn from all this? First, perhaps, that there is not a huge amount of evidence. What there is suggests that low carbohydrate diets are likely to deliver a larger weight reduction, certainly at about six months. Moreover, more people on low carbohydrate diets stay on them for longer. Neither of the randomised trials did anything in particular about exercise, a very important component of lifestyle changes needed for sustained weight loss.
The higher proportions of protein and fat do not seem to adversely influence risk factors like total and LDL cholesterol, and probably improve risk factors by reducing triglyceride levels and perhaps increasing HDL cholesterol levels. There may be small changes to insulin resistance.
In the end it may come down to suitably matching a dietary regimen to an individual’s taste. For some of us, a high protein diet, with bacon and eggs, steak, fish and cheese would be heaven. For others, purgatory. Perhaps now people who want or need to lose weight have more of a choice about what might best suit them.
DM Bravata et al. Efficacy and safety of low-carbohydrate diets. A systematic review. JAMA 2003 289: 1837-1850.
FF Samaha et al. A low-carbohydrate as compared with a low-fat diet in severe obesity. New England Journal of Medicine 2003 348: 2074-2081.
GD Foster et al. A randomized trial of a low-carbohydrate diet for obesity. New England Journal of Medicine 2003 348: 2082-2090.