(9) j¿rgen vesti nielsen267-274

Upsala J Med Sci 110 (3): 267–273, 2005 A Low Carbohydrate Diet in Type 1 Diabetes:
Clinical Experience
– A Brief Report
Jørgen Vesti Nielsen, Eva Jönsson, Anette Ivarsson Department of Medicine, Blekingesjukhuset, Karlshamn, Sweden Due to failure to achieve control twenty-two patients with type 1 diabetes with sympto-matic fluctuating blood glucose started on a diet limited to 70-90 g carbohydrates perday and were taught to match the insulin doses accordingly. The caloric requirementswere covered by an increased intake of protein and fat. The purpose was to reduce theblood glucose fluctuations, the rate of hypoglycaemia and to improve HbA1c.
After three and 12 months the rate of hypoglycaemia was significantly lowered from 2.9 + 2.0 to 0.2 + 0.3 and 0.5 + 0.5 episodes per week respectively. The HbA1c levelwas significantly lowered from 7.5 + 0.9 % to 6.4 + 0.7 % after three months and wasstill after 12 months 6.4 + 0.8 %. The meal insulin requirements were reduced from21.1 + 6.7 I.U./day to 12.7 + 3.5 I.U./day and 12.4 + 2.6 I.U./day after three and 12months respectively.
Furthermore the triglyceride level was significantly lowered whereas the levels for total cholesterol and HDL-cholesterol were unchanged.
Conclusion: the present report shows that a 70-90 g carbohydrate diet is a feasible long-term alternative in the treatment of type 1 diabetes and leads to improved gly-caemic control. The Diabetes Control and Complication Trial (DCCT), in which conventional ver-sus intensive insulin treatment was studied, showed that a reduction of the risk ofmicrovascular complications in type 1 diabetes was directly linked to a loweredHbA1c level (1). The incidence of severe hypoglycaemia including coma andseizures, however, was threefold higher in the intensively treated patients, 61.2 ver-sus 18.7 episodes per 100 patient-years (2). The dietary advice in the DCCT includ- Received 25 April 2005Accepted 7 June 2005 ed a carbohydrate proportion of 55 % of the calories. In a 2000 kcal diet thisamount corresponds to about 275 g carbohydrates, mostly consumed as starchesfrom potatoes, bread and pasta, that in the gut swiftly are broken down to glucose.
This quantity of carbohydrates therefore corresponds to about the same amount ofglucose within the body and behaves like ingested glucose. Patients with type 1 diabetes commonly take fixed pre-meal doses of insulin.
Some patients change the dose according to the pre-meal glucose level, but normal-ly the dose is taken without regard to the carbohydrate content of the followingmeal. However, by way of education, patients can learn to estimate the carbohydratecontent of a meal and to match the insulin accordingly. In this way patients withtype 1 diabetes reduced mean HbA1c from 9.5 to 8.5 %. But the high rate of severehypoglycaemia still persisted, 20 episodes per 100 patient-years (3). The lesson from the DCCT is that normal glycaemic control is impossible unless the risk of severe hypoglycaemia can be reduced.
A mismatch between insulin and glucose causes unpredictable glucose swings.
With large insulin doses this may lead to severe hypoglycaemia. The mismatchingis an inherent problem, since first, the assessment of carbohydrates in a meal ishampered by a large error rate (4), and second, the absorption of insulin may varyby up to 30 % (5). The two major determinants of the blood glucose, therefore, can-not be determined with any accuracy. It thus becomes very problematical to matchcarbohydrates with insulin.
Large variations in the input -- carbohydrates and insulin -- may cause large vari- ations in the output -- the glucose levels -- manifested as hypo- and hyperglycaemia.
According to basic calculus a reduction in the size of the input reduces the degree ofvariation in the output.
Consequently, a reduction of both carbohydrates and given insulin ought to lead to a reduced degree of fluctuations and thus allow the patient to reduce the meanglucose level and the HbA1c safely.
For motivated patients this principle works well. A carbohydrate content of less than 40 g per day with accurately matched insulin doses seems to normalise theblood glucose levels and diminish the risk of hypoglycaemia in patients with type 1diabetes who adhere to the program; the plasma lipids are normalised also(6-8). The objection to the regimen is that a reduction of carbohydrates to less than 40 g per day probably is unrealistic in the general diabetes population. Such a radicalreduction might limit the number of patients willing to try the method.
A reduction to 70-90 g per day, however, seems a feasible long-term choice (9).
Since a reduced amount of carbohydrates with appropriate insulin doses as well asthe ability to count carbohydrates are factors important in regulating the blood glu-cose, we have combined these two principles in the treatment of patients with poorglycaemic control, when treatment according to the guidelines has failed to stabilisethe blood glucose. The purpose is to reduce the glucose fluctuation and HbA1c.
We have done this in the form of educational programs attended by 6-8 patients.
The patients had all failed to achieve an – in their own opinion – satisfying glucose control. All such patients have at our clinic for a couple of years received informa-tion on the theoretical background for the herein described treatment model. It isthen up to the patients themselves to decide whether they wish to make anychanges. All the patients in the present report have actively sought the presentedtreatment model. A clinical chart review has been performed for the first 24 patients attending the program. It is a quality control, and its purpose is to evaluate to what degree initialpositive changes are retained one year later even without close follow-up.
An approach using a carbohydrate quantity of 70-90 g per day has to our knowl- edge not been reported before. The purpose of the present report is to describe themethod and the results after one year.
Table 1. Changes in the rate of hypoglycaemia, in HbA1c, insulin utilization, cholesterol,HDL-cholesterol and triglycerides in 22 patients with type 1 diabetes before, 3 and 12months after a change to a diet restricted to 70-90 g of carbohydrates per day.
* P is for the difference within the group from base line. § n=15. HbA1c < 5.6 % in non- The patients were outpatients with type 1 diabetes and episodes of hypo- and hypergly-caemia that had been treated at our clinic. The first 24 patients, 17 women and 7 men,who attended the program, constitute the present group. The mean duration of diabetes was 18+13 years and the mean age was 51 + 10 years. Seven patients were over-weight and had BMI > 27 kg/m2 (range 27-38). Elevenpatients administered insulin with an insulin pump. Another 13 patients used long act-ing insulin glargin (Lantus) twice or once a day; for meal-insulin they used insulinaspart in a pen device (NovoRapid) that enables delivery of half units. All patients were requested to record the occurrence of hypo- and hyperglycaemia as well as other symptoms of poor diabetes control before they started on the new regimeas well as 3 months later. About 12 months after the start the patients again in a letterwere asked to record the occurrence of symptomatic hypoglycaemia, i.e. not severe butmanageable by the patients themselves without help from others.
The program began with a 6-hour-meeting followed by, over the next two months, five follow-up meetings lasting approximately two hours. The main issues on the firstday were the effect on the blood glucose by the different constituents of the food, thetiming of insulin administered and the timing of blood glucose measurements. Thepatients received a flow sheet wherein they entered the blood glucose levels, the timeof each meal and the amount of carbohydrates. Further, the time and amount of insulinunits taken were entered together with any occurrence of hyper- and hypo-glycaemia.
The number of glucose tablets needed to correct low blood glucose levels were record-ed as well as the extra insulin units needed to correct a too high pre-meal and bedtimeglucose level. The regimen: A carbohydrate restricted diet (70-90 g per day) that excluded potatoes, rice, pasta, bread and cereals, but included hard bread and vegetables. Intensive glu-cose monitoring, > 4 times per day, was required; recording of the time of glucose lev-els, meals, insulin dosage, exercise and correction of too high or too low pre-meal andbed-time glucose levels with insulin or glucose tablets were also required. The targetfor pre-meal blood glucose was 5.6 mmol/l (plasma glucose 6.0 mmol/l). The patients were instructed to eat three meals a day and abstain from any eating between the regular meals; the interval between meals should be at least four hours –the duration of one dose of aspart. To assist them the patients were given samples ofmenus and recipes showing the carbohydrate and protein content. The initial distribu-tion of carbohydrate, protein and fat was 20%, 30% and 50 % respectively. Sugar orother fast acting carbohydrates were not permitted. During this 2 months period the regimen for each patient was individually tailored. Special attention was paid to the occurrence of gastroparesis. Even though this con- dition might be asymptomatic the effects on glucose control usually are obvious verysoon. In such cases the autonomous nervous system was examined and the patient pre-scribed domperidone (Motilium) and if needed switched to insulin human (Actrapid). After finishing the program the patients were able to handle their own treatment without help. Only routine visits to the diabetes nurse every 3-4 months were continuedor according to the individual patient’s preferences. HbA1c and insulin utilization were recorded at start and after three months. About 12 months later they were once more recorded from the chart. Since HbA1c vary wehave calculated average HbA1c from the charts for the previous year before start foreach individual. HbA1c <5.6 % in non-diabetic persons. (chromatography HPLCmonoS column) Total-cholesterol, HDL-cholesterol (HDL) and triglycerides (TG) were measured at start and after 3 months. It was then measured again about 12 months later.
The results are presented as means with Standard Deviation. Paired T-test is used for Blood Glucose
measurements
Figure 1. Blood glucose values in a patient with type 1 diabetes before and a few days aftera change to a carbohydrate restricted diet (70-90 g daily) with adapted insulin doses. Themeasurements cover 12 days with 58 measurements before the change, and 10 days with Figure 1. Blood glucose values in a patient with type 1 diabetes before and a few days after a change to a carbohydrate restricted diet (70-90 g daily) with adapted insulin doses. The measurements cover The mean for the 58 measurements before the change was 12.9 ± 6 mmol/l. The mean for 12 days with 58 measurements before the change, and 10 days with 70 measurements after the change.
the 70 measurements after the change was 5.9 ± 2 mmol/l.
The mean for the 58 measurements before the change was 12.9 ± 6 mmol/l. The mean for the 70 mea-surements after the change was 5.9 ± 2 mmol/l. Two patients left the program after approximately 1 month for personal reasons. Blood glucose: The effect of the diet on the glucose fluctuations is illustrated in Figure 1. The size of the fluctuations before the change clearly forbids any loweringof the mean glucose levels. For this patient average HbA1c the year prior to startwas 7.5 %. One year later it was 4.7 %. Hypoglycaemia: Nine patients had failed to record the occurrence of hypoglycaemia before start. The following figures are calculated from the records of 15 patients. Therate of symptomatic hypoglycaemia among these was reduced by 94 % after 3 monthson the new regimen, and at twelve months by 82 % from base line. (see Table 1) HbA1c and insulin: Table 1 shows that the effect on HbA1c and insulin adminis- tration was stable from 3 to 12 months in the group.
Lipids: There was no change except for a significant 16 % lowering of the Two patients lost weight due to other illnesses. Otherwise, the bodyweight for the normal-weight patients remained stable, while all the overweight patients lostweight. Six patients were diagnosed with diabetic gastroparesis and were prescribeddomperidon (Motilium).
By lowering the carbohydrate amount and corresponding insulin doses the blood glu-cose fluctuations diminished and a better HbA1c was achieved. The rate of hypogly-caemic episodes was reduced without any deterioration in the lipids. It is likely that therisk of severe hypoglycaemia has diminished also. It is remarkable that the effect is retained without close follow-up. In the DCCT the patients visited the clinic in person once a month and were in contact by phone and faxonce a week to get directions from the diabetes team [1]. This in order to maintain atight control of HbA1c.
The group presented here achieved and maintained an HbA1c close to the one accomplished in the DCCT without close contact to the diabetes team. This suggeststhat the treatment method presented here is a more reliable tool for the patients. Thismight be due to the improved predictability of the blood glucose as illustrated in thereduced rate of hypoglycaemic episodes. Most patients experience more freedom since they do not have to eat at a fixed time.
With a correct dose of basal insulin a meal can be postponed or even cancelled withoutconsequences for the glucose levels. We did not study compliance to the diet, but since both HbA1c and insulin require- ments stayed low, the findings for the whole group further suggest that the selected ini-tial amount of carbohydrates is feasible in the long term in patients with type 1 dia-betes. It was still possible for individual patients at their own discretion to pursue astricter carbohydrate reduction in order to achieve an even better control, and two per-sons did so. The high carbohydrate dietary advice in type 1 diabetes is based on avoidance of protein and fat in the food, especially saturated fat. The importance of this may how-ever, with respect to the present findings, be reconsidered (10-12). In summary: The approach to therapy reported here is feasible in type 1 diabetes for motivated patients. It improves HbA1c and appears to increase safety by lowering therate of symptomatic hypoglycaemia, and the effect was retained one year later.
We are grateful to Peter Naeser, MD, PhD, for reading the manuscript and for valuable advice. We thank Kamma Willumsen, dietician, for help.
The Diabetes Control and Complications Trial Research Group: (1993). The effect of intensive treatment of diabetes on the development and progression of long-term complications in insulin-dependent diabetes mellitus. N Eng J Med 329:977-986 The diabetes Control and complications Trial Research Group (1997). Hypoglycaemia in the Diabetes Control and Complications Trial. Diabetes 46(2):271-86 DAFNE Study Group (2002). Training in flexible, intensive insulin management toenable dietary freedom in people with type 1 diabetes: dose adjustment for normaleating (DAFNE) randomised controlled trial. BMJ 325:746-749 Waldron S (1996). Controversies in the dietary management of diabetes in childhoodand adolescence. Br J Hosp Med 56(8):450-454 Heinemann L (2002). Variability of insulin absorption and insulin action. DiabetesTechnol Ther 4(5):673-82 Bernstein RK (1980). Virtually continous euglycemia for 5 yr in a labile juvenile-onset diabetic patient under noninvasive closed-loop control. Diabetes Care3(1):140-143 Bernstein RK (1997). Dr. Bernstein’s Diabetes Solution; First edition; Little Brownand Company , New York O’Neill DF, Westman EC, Bernstein RK (2003). The effects of a low-carbohydrateregimen on glycemic control and serum lipids in diabetes mellitus. Metab SyndRel Disord 1(4):291-298 Nielsen JV, Joensson E, Nilsson AK (2005). Lasting improvement of hyperglycaemia and bodyweight: low-carbohydrate diet in type 2 diabetes. – Abrief report. Ups J Med Sci 110:69-74 Manninen AH (2004). High-protein weight loss diets and purported adverse effects:where is the evidence? Sports Nutrition Review Journal (1):45-51 Hooper L, Summerbell CD, Higgins JTP, et al.(2001) Reduced or modified dietary fat for prevention of preventing cardiovascular disease. Cochrane Database System Rev3:CD002137 Ravnskov U (1998). The questionable role of saturated and polyunsaturated fattyacids in cardiovascular disease. J Clin Epidemiol 51(6):443-460.
Jørgen Vesti NielsenConsultant PhysicianDept. MedicineBlekingesjukhuset KarlshamnLänsmansvägen 1, 374 80 Karlshamn, SwedenTelephone: +46 454 731000e-mail: jorgen.vesti-nielsen@ltblekinge.se

Source: http://fhskola.ltblekinge.se/download/18.483745d1110dfd532008000742/low-carbohydrate_diet_in_type_1_diabetes.pdf