loirrrial of Interrial Medicine 1995: 2 3 7 : 591-597

Liberalized diet in patients with type 1 diabetes

I. MUHLHAUSER, U. BOTT, H. O V E R M A N N , W. WAGENER, R. BENDER, V. JORGENS
& M.BERGER
Frorri the Departrnerit of Nutrition and Metabolic Diseases ( WHO-Collaborating Centre for Diabetes), Heinrich-Heine Universitu. Diisseldorf,
Gcrrtiarig

Abstract. Muhlhauser I, Bott U, Overmann H,

Wagener W, Bender R, Jorgens V, Berger M
(Department of Nutrition and Metabolic Diseases
(WHO Collaborating Centre for Diabetes), HeinrichHeine University, Dusseldorf, Germany). Liberalized
diet in patients with type 1 diabetes. 7 Intern Med
19 95; 2 3 7 : 591-7.
Objective. To document that strict dietary regimen
are not necessary in the context of intensified insulin
therapy.
Design. German multicentre, prospective cohort
study: 6 years follow-up.
Setting. Ambulatory examination using a mobile
ambulance.
Subjects. A total of 636 type 1 diabetic patients (age
3 3 f 7 years, diabetes duration 1 5 f7 years ;
mean fSD), who had participated in a structured,
5-day, in-patient, group treatment and teaching programme for intensification of insulin therapy and
liberalization of the diabetes diet 6 years prior to
follow-up.
Main outcome measures. Relations between the
extent to which patients practise a liberalized diet,
the degree of metabolic control (HbAlc, severe
hypoglycaemia, body mass index, cholesterol), and
the patients' perceived burden through dietary treatment.

Introduction
Traditionally, diet is considered a cornerstone of
insulin therapy [ l , 21. Poor diabetes control is
frequently attributed to either inadequate dietary
advice [3] or poor adherence to dietary prescriptions
[4]. Dietary recommendations for insulin-treated
patients include regimens of caloric, carbohydrate,
fat, protein, fibre, sodium and alcohol intake, the ban
of sugar and sugar-containing foods such as biscuits
0 1995 Blackwell Science Ltd

Results. In the total patient group, HbAlc was

7.9+1.6%, and the incidence of severe hypoglycaemia was 0.17 cases per patient during the
preceding year: 31% patients injected insulin < 3
times per day, 58% 4-7 times per day, and 11%used
insulin pump therapy. Only 11% patients reported
following a meal plan, whereas 89% continually
changed timing and amount of carbohydrate intake ;
only 5% had the same number of meals every day,
whereas as many as 20% varied the number of meals
per day by four or more; 53% skipped main meals;
8 5% habitually consumed sugar or sugar containing
foods. Patients with a higher degree of diet
liberalization injected insulin or used an insulin
pump therapy more frequently, and perceived their
dietary treatment to be less burdensome. No clinically
significant associations were found between the
extent of diet liberalization and metabolic control.
Conclusions. Under the conditions where type 1
diabetic patients have the opportunity to participate
in a n intensified insulin treatment and teaching
programme, liberalization of the diabetes diet is not
associated with adverse effects on glycaemic control,
but is associated with less perceived burden through
dietary treatment.
Keywords : diabetes, diet, education, hypoglycaemia,
insulin therapy, quality of life.

and confectionery and, finally, to maintain a constant

daily pattern of food intake with main meals and
snacks [1, 21. Until recently, the American Diabetes
Association (ADA) additionally proposed a complex
and rigid system of food exchange lists [ l ] - potentially useful for the compulsive, motivated, and well
educated patient [51. A scientific justification for its
use was lacking [5]. This also holds true for the
general nutritional recommendations for individuals
with diabetes [ l , 21. Even the recommendations
591

592

I. MUHLHAUSER et al.

concerning the carbohydrate content of the diabetes

diet are still controversial [6]. It is not surprising that
Kelly West entitled his famous review on diabetes
diet, An analysis of failure [7].
Persons with diabetes regard diet as one of the
biggest problems of their disease [8]. especially when
they have to reject food offers and their daily routines
are disrupted as they have to eat regularly [9].
However, type 1 diabetic patients on intensified
insulin therapy can be released from complex and
rigid dietary rules without adverse effects on metabolic control [ 10-161. Liberalization of the diabetes
diet is associated with better quality of life and
enhanced treatment satisfaction [9]. Nevertheless,
diet liberalization is still withheld from most diabetic
patients. It is argued that it is feasible only under
research conditions and in selected patients.
In this study we investigated in a large cohort of
type 1 diabetic patients, the relations between the
extent to which patients practise a liberalized diabetes
diet, the degree of metabolic control and the patients
perceived burden through dietary treatment.

Patients and methods

The study is based on a 6-year follow-up of 636
type 1 diabetic patients (81% of the original cohort)
who had taken part in a German multicentre study,
which documented the feasibility of translating an
intensified insulin treatment and teaching programme (TTP) from a specialized university diabetes
centre to general internal medicine departments.
As the outcomes were comparable between the
specialized centre and the nine participating general
hospitals, for the purpose of the present study the
combined group of 636 patients was analysed.
Detailed descriptions of the study population, the
TTP, translation of the TTP, medical care of the
patients after discharge, evaluation protocols, dropouts and results for up to 3 years of follow-up have
been published [17, 181. In short, 784 consecutively
referred type 1 diabetic patients, aged 1 5 4 0 years
and free of advanced diabetic, late complications
(serum creatinine > 177 pmol L- ; blindness] had
taken part in the same, 5-day, inpatient TTP for
intensified insulin therapy in one of the 10 participating hospitals and were re-examined after 1,2, 3 , and
6 years. Patients already treated with continuous
subcutaneous insulin infusion (CSII) on admission
were not considered eligible, although change to CSII

after participation in the TTP was possible. A history

of repeated severe hypoglycaemia was not an exclusion criteria.
The objectives of the TTP were to enable patients
to improve glycaemic control without increasing the
risk of severe hypoglycaemia and to liberalize the
diet. Patients were advised to measure blood glucose
before main meals and at bedtime and to inject NPH
insulin in the morning and at bedtime and regular
insulin before meals. Day-to-day adaptation of insulin
dosages by the patients themselves was considered a
prerequisite for achieving the treatment goals. The
teaching was delivered by a nurse and a dietitjan in
a structured, Monday to Friday inpatient course for
groups of up to 10 patients. Dietary training was
restricted to the estimation of the carbohydrate
content of meals and the calculation of insulin needs
for varying carbohydrate intakes. There were no
regulations of daily energy intake, except for obese
patients, and weight control was self-regulated.
Sugar consumption was not prohibited, although its
prudent use was recommended [12]. Patients were
provided with lists with the carbohydrate contents of
sugar containing foods such as cakes, biscuits and
confectionery. There were no prescriptions with
respect to fat, protein and fibre intake. No meal plan
was provided. The TTP aimed to empower patients to
choose the most appropriate insulin treatment and
dietary regimen for themselves and to carry this out
safely and effectively. Thus, the more liberalized the
diet becomes, the more frequent measurements of
blood glucose, injections of regular insulin, and
adaptations of insulin dosages by the patients are
necessary, whereas a more conservative treatment
regimen with only two daily insulin injections
requires meal planning with respect to timing and
amount of carbohydrate intake. After discharge,
patients were followed-up by their family physicians,
although the consultation of a specialized diabetes
outpatient clinic was possible on referral [17].
According to the German health care system, all
materials for insulin therapy, metabolic self-monitoring and consultations were free of charge for all
patients.
The 6-year follow-up examination was performed
by two investigators, who had not been involved in
the careofthe patients, using a mobile ambulance as
described elsewhere [19]. At the 6-year follow-up, a
standardized interview about eating habits relevant
to the study question was included in the evaluation

0 1995 Blackwell Science Ltd ]ourrial

OJ

Internal Medicine 2 3 7 : 591-597

LIBERALIZED DIET I N DIABETES

protocol [17, 181 in order to rate the degree of diet

liberalization for the individual patient. The following
eating habits were considered indicative of diet
liberalization.
1 No meal planning with respect to timing and
amount of carbohydrate intake:
2 day-to-day variation of the number of meals
(defined as consumption of any food except for the
treatment of hypoglycaemia) by two or more ;
3 skipping of main meals at least once per week:
4 habitual consumption of sugar or sugar-containing foods.
Based upon these four items, a score system was
construed ranging from 0, in case none of these four
characteristics was applicable (no diet liberalization
at all), to 4, in case all four characteristics were
applicable (highest degree of diet liberalization).
The perceived burden through dietary treatment
was assessed by patients degree of agreement to four
statements using a four-point rating scale. The items
were adapted from a questionnaire on diabetesrelated stress [20] :
1 I have to give up good tasting food:
2 I often cant eat as much as I want to;
3 I have to eat even if I am not hungry;
4 I eat as if I didnt have diabetes).
Because the four items did not constitute a homogenous subscale, single item scores were analysed.
Haemoglobin A l c was measured by the Diamat@
(Biorad, Munich, Germany) HPLC method (reference
range 4.3%-6.1%), and serum cholesterol by the
Reflotronm (Boehringer-Mannheim, Mannheim,
Germany) method as described [19]. Body weight
was measured with patients wearing normal, everyday clothes, but without shoes, coats and jackets.
Severe hypoglycaemia was defined as a hypoglycaemia treated by intravenous glucose or glucagon injection.
For statistical analysis of the association between
the liberalized diet score (LDS) and the demographic
variables, gender, age, and diabetes duration, and
logistic regression analysis (proportional odds model)
[21] was performed with LDS as the dependent
ordinal variable. Relations between LDS and relevant
response variables were analysed using regression
models with LDS and essential covariables as independent variables. For binary response (hypoglycaemia, blood glucose self monitoring, items 1 to
4 on dietary burden [dichotomised item scores])
logistic regression analysis, and for ordinal response

593

(insulin treatment) the proportional odds model [2 13,

and for continuous response (log cholesterol, body
mass index, HbAlc) linear multiple regression were
used. The covariables in all models were gender, age
and diabetes duration supplemented by HbAlc and
body mass index for the model with log cholesterol,
and HbAlc and insulin dose for the model with body
mass index as the dependent variable. The LDS
scores were included in the regressions by using four
dummy variables with level 4 as the reference
category. Results are given as meansfSD or as
frequencies. For computations, the SAS procedures
LOGISTIC [22] and REG [23] were used.

Results
Vital status was available for all 784 patients except
nine, who could not be traced. Thirteen patients had
died (suicide two : motor-cycle accidents two : cancer
two : hypoglycaemia one : cardiovascular three :
schizophrenia, undefined on autopsy one : pneumonia one: alcoholism, undefined on autopsy one).
A total of 1 3 5 patients either declined to participate
(n = 57) or had moved away too far [17] to be reexamined. Compared to the 6 36 re-examined
patients, the 1 3 5 drop outs had higher HbAlc levels
before the TTP [8.8f1.9% vs. 8 .3 f 1 .8 %; P =
0.01, Students t-test] and a lower incidence of severe
hypoglycaemia during the year before the TTP (0.15
vs. 0.28 cases per patient: P < 0.05, Mann-Whitney
U-test). Selected clinical variables of the 636 patients
(48% women) at the time of the present study are
shown in Table 1.
The incidence of severe hypoglycaemia was 0.17
cases per patient during the preceding year: 26% of
the patients had consulted a physician specialized in
diabetes during the preceding year.
As to eating habits, only 11%of patients reported
following a meal plan, whereas 89% continually
changed timing and amount of carbohydrate intake.
The average number of meals per day was two to
three for 20% of patients, four to five for 57%, and
more than five for 23%, respcctively. Only 5%
reported to have the same number of meals every
day, whereas 18% varied the number of meals by
one, 57% by two or three, and 20% by four or even
more. When asked how often per week they skipped
a main meal 17% answered three times or more
often, 26% once or twice, and 57% never skipped a
main meal.

0 1995 Blackwell Science Ltd journal of Internal Medicine 237: 591-597

594

I. M U H L H A U S E R et a/.

Table 1 Selected variables in relation to liberalized diet score groups

Liberalized diet score groups'

Patients (no.)
Age (years)
Diabetes duration (years)
Body mass index (kg/m2)
Haemoglobin A l c (%)
Patients with SH (%)
Cholesterol (mmol L-l)
lnsulin dose (U kg-l day-')
lnsulin treatment. % patients with
< 3 insulin injections per day
4-7 insulin injections per day
CSII
Patients with BGSM 2 3 times
per day (%)
Items on perceived dietary burden,
patients who agree (%) with:
1 ' I have to give up good tasting
food'
2 ' I often can't eat as much as
I want to'
3 ' I have to eat even if I am not
hungry *
4 ' I eat as if I didn't have
diabetes'

Total group

22
35+8
17&9
23.953.0
7.9 5 1.2
9
6.1 1.7
0.6650.20

74
34+7
15k8
24.7k3.6
7.8 k 1.4
5.4* 1.2
0.64 f0.15

174
34+7
15+7
24.6k3.2
7.8k1.5
14
5.5k1.3
0.65 k 0 . 2 0

202
33k7
16k7
24.6 k 3.6
8.0f1.6
10
5.5k1.2
0.61 kO.19

157
31k6
14f 7
24.5k3.2
7.9k1.6
14
5.5k1.2
0.63 k 0 . 1 7

629t
33+7
15+7
24.6+ 3.4
7.9k1.6
12
5.5k1.2
0.63k0.19

71
29
0
55

36
61
3
70

36
58
6
69

29
57
14
70

18
63
18
77

31
58
11
71

10

48

16

13

11

12

21

19

23

34

35

29

Values are means f SD or as indicated.

SH. severe hypoglycaemia during preceding year: CSII. continuous subcutaneous insulin infusion: BGSM. blood glucose self
monitoring.
* O = no diet liberalization at all; 4 = highest degree of diet liberalization.
TTotal number of patients adds up to 629 rather than to 636 due to missing values.

Table 2 Logistic regression analysis (proportional odds model) of the liberalized diet score*

Variable

Logistic
coefficient (SE)

Standardized
coefficient

Odds
ratio (95% CI)

Age
Gender?
Diabetes duration

-0.032 (0.01)
-0.229 (0.14)
-0,000 (0.01)

< 0.005
NS
NS

-0.121
-0.063
-0.001

0.97 (0.95-0.99)
0.80 (0.60-1.06)
7 .OO (0.98-1.02)

'Order of response variable: LDS 4 = highest degree of diet liberalization: LDS 0 = no diet liberalization.
7 0 = female; 1 = male.
Score test for the proportional odds assumption: P = 0.43.

Only 15% of patients avoided sugar consumption

completely (except for the treatment of
hypoglycaemia), 30% consumed sugar or sugar
containing foods daily or several times per week, and
5 5% less frequently.
The distribution of patients and selected variables
according to LDS are summarized in Table 1. Older
patients had a tendency to lower scores (Tables 1 , 2 ) .
Patients with higher LDS injected insulin or used

CSII more frequently (Tables 1 , 3 ) ,and they perceived

less burden through dietary treatment as reflected by
significant associations between LDS and items 3 and
4 (Tables 1 , 4). Compared to patients with the
highest LDS,patients with no diet liberalization at all
had higher cholesterol levels (P < 0.05), and the
proportion of patients measuring blood glucose 3
times per day was lower (P < 0.05 : Table 1). Because
of multiple testing these findings should not be

0 1995 Blackwell Science Ltd Iournal o/ Interrial Medicine 237: 591-597

LIBERALIZED DtET I N DIABETES

595

Table 3 Logistic regression analysis (proportional odds model) of insulin treatment*

Logistic
coefficient (SE)

Variable

LDs=o

-2.440 (0.52)
- 1.077 (0.29)
-0.915 (0.23)
-0.529 (0.22)
-0.018 (0.01)
-0.535 (0.17)
0.029 (0.01)

LDS=l
LDs=2
LDs=3
Age
Genderf
Diabetes duration

< 0,0001
< 0.0002

< 0.0001
< 0.02
NS
< 0.002
< 0.02

Standardized
coefficient

Odds ratio (95% CI)

-0.246
-0.190
-0.225
-0.137
-0.068
-0.147
0.120

0.09 (0.03-0.24)
0.34 (0.19-0.61)
0.40 (0.25-0.63)
0.59 (0.38-0.91)
0.98 (0.96-1.01)
0.59 (0.42-0.81)
1.03 (1.01-1.05)

LDS. liberalized diet score (0 = no diet liberalization: 4 = highest degree of diet liberalization).
*Order of response variable: continuous subcutaneous insulin infusion : highest, 4-7 insulin injections per day:
per day: lowest: LDS score 4 as reference variable.
t0 = female: 1 = male.
Score test for the proportional odds assumption: P = 0.44.

< 3 insulin injections

Table 4 Logistic regression analysis of dietary burden items*

' I have to eat even if I am not hungry'

Variable

Odds ratio
(95% CI)

ms=o

< 0.0001
NS
NS
NS
NS
NS
< 0.05

8.35 (2.97-23.4)
1.54 (0.67-3.56)
1.21 (0.60-2.46)
0.68 (0.32-1.45)
1.02 (0.98-1.06)
0.79 (0.47-1.32)
0.96 (0.93-0.99)

LDS = 1
LDs=2
LDs=3
Age
Gender?
Diabetes duration

' I eat as if I had no diabetes'

Odds ratio
(95% CI)

NS
< 0.02
< 0.02
NS
NS
NS
< 0.1

0.42 (0.13-1.36)
0.41 (0.20-0.82)
0.53 (0.32-0.87)
0.90 (0.57-1.41)
1.30 (0.90-1.88)
1.00 (0.97-1.0%)
1.02 (1.00-1.05)

LDS = liberalized diet score (0 = no diet liberalization: 4 = highest degree of diet liberalization).
'Order of response variable: 1 = agreement, 0 = disagreement: LDS score 4 as reference variable.
t o = female: 1 = male.
Homer-Lemeshow goodness-of-fit test: P = 0.58.

overinterpreted. No significant associations were

found between LDS and gender, diabetes duration,
HbAlc levels, severe hypoglycaemia, insulin dosage,
body mass index, and items 1 and 2 (Tables 1. 2).
Also, at baseline, metabolic parameters, such as
HbAlc levels, severe hypoglycaemia and body mass
index did not differ between LDS groups.

Discussion
The present study shows that almost all patients with
type 1 diabetes who had the opportunity to participate in an intensified insulin treatment and
teaching programme, of which an essential part is to
offer patients a liberalization of the diet, practise a
liberalized diabetes diet, although to a variable extent.
As expected, patients with higher degrees of diet
liberalization injected insulin or used an insulin
pump therapy more frequently than did patients

following a more traditional dietary regimen,

indicating that, according to the goals of the treatment and teaching programme, patients were able to
find an appropriate balance between eating habits
and insulin therapy.
The eating habits of the patients do not conform to
any of the official guidelines for dietary treatment of
type 1 diabetes [1,21. However, diet liberalization
had no adverse effects on metabolic control as
assessed in this study. There were no clinically
relevant associations between the extent of diet
liberalization and HbAlc, severe hypoglycaemia,
body mass index or serum cholesterol. Patients were
fairly well controlled, as reflected by their HbAlc
levels and the incidence of severe hypoglycaemia.
Considering the different study conditions and patient
inclusion criteria, results compare favourably with
those of the intensively treated patient group of the
Diabetes Control and Complications Trial (DCCT)

0 1995 Blackwell Science Ltd Journal of Internal Medicine 2 3 7 : 591-597

596

I. MUHLHAUSER et al.

[24, 251, in which an ADA diet was recommended

[l]. In the present study, HbAlc values were about
0.5% higher than respective values in the DCCT, but
the frequency of severe hypoglycaemia was lower
[17, 18, 24, 251. In addition, in the present study,
neither serum cholesterol levels nor body weight
were higher than respective age-adjusted values of
the general German population [26-281.
Based on the present study, it cannot be excluded
that individual patients who are overweight or who
have unsatisfactory glycaemic control might profit
from following a more regulated dietary regimen. On
the other hand, liberalization of diabetes diet is a
major motivation for our patients to opt for intensified
insulin therapy. In fact, in this study, patients with
the highest degree of diet liberalization perceived less
burden through their diet than patients with lower
degrees of diet liberalization. Nevertheless, diet remains a problem for type 1 diabetic patients. Even
amongst the patients with the most liberalized diet,
only 35% felt that they ate as if they had no diabetes.
Only limited systematic information is available on
the eating habits of the patients before participation
in the intensified insulin treatment and teaching
programme. In general, patients were treated traditionally following the rather strict dietary recommendations of the German Diabetes Association
which are comparable to international recommendations [ l , 21. This is in accordance with the
conventional insulin therapy these patients practised
before participation in the programme, i.e. 86%
injected insulin twice daily or less frequently, only
10% measured blood glucose three times or more
often per day [17, 181, and 73% reported to have six
or more meals per day (data not shown). However, it
is probable that patients did not adhere completely to
the rigid dietary recommendations.
The present study was not intended to perform a
detailed and comprehensive assessment of the nutritional behaviour and the actual composition of
nutrients consumed by type 1diabetic patients. Only
aspects considered to be relevant indicators of
liberalization of the diabetes diet were evaluated. To
this end, a simple scoring system was construed,
because to date, no specific methodology is available
for the quantification of liberalized diabetes diet. The
close association between the liberalized diet score
system and the patients perception of restrictions
through dietary treatment, shows that the scoring
system is useful and valid.

There was no control group in the present study,

i.e. patients performing intensified insulin therapy
but following a strict dietary regimen. Such a study,
especially addressing the possible effects of sugar
consumption, has been performed previously in
selected patients with pump treatment [12], in which
sugar consumption had no adverse metabolic effects.
In addition, several randomized controlled studies
with observation periods of up to 2 years have
demonstrated the efficacy and safety of the whole
package of the intervention programme, consisting
of intensified insulin therapy, comprehensive patient
training and liberalization of the diet [14, 151.
In conclusion, the present study shows that a
majority of patients with type 1 diabetes practise a
liberalized diet 6 years after participation in an
intensified insulin treatment and teaching
programme. Under the condition of frequent blood
glucose self-monitoring and frequent insulin injections, including self-adaptation of insulin dosages,
liberalization of the diabetes diet was not associated
with adverse effects on glycaemic control, but was
associated with less perceived burden through dietary
treatment.

Acknowledgements
This study was supported by Boehringer-Mannheim,
Mannheim, Germany, and by the P. Klockner Stiftung, Duisburg, Germany (grants to Professor M.
Berger).

References
American Diabetes Association. Position statement. Nutrition
recommendations and principles for individuals with diabetes
mellitus: 1986. Diabetes Care 1987: 10: 126-32.
Diabetes and Nutrition Study Group of the European Association for the Study of Diabetes - 1988. Nutritional
recommendations for individuals with diabetes mellitus. Diab
Nutr Metab 1988: 1: 145-9.
Knowles HC. Guest GM, Lampe J, Kessler M. Skillmann TG.
The course of juvenile diabetes treated with unmeasured diet.
Diabetes 1965; 14: 239-73.
Christensen NK. Terry RD. Wyatt S. Pichart JW. Lorenz RA.
Quantitative assessment of dietary adherence in patients with
insulin-dependent diabetes rnellitus. Diabetes Care 1983 ; 6:
245-50.
Nuttall FQ. Carbohydrate and dietary management of
individuals with insulin-requiring diabetes. Diabetes Care
1993: 16: 1039-42.
Vinik AI. Lauterio TJ. Wing RR. Should the bee suck honey or
lard? Diabetes Care 1993: 16: 1045-7.
West K. An analysis of failure. Ann Intern Med 1973: 79:
425-34.

0 1995 Blackwell Science Ltd journal of Internal Medicine 2 3 7 : 591-597

LIBERALIZED D I E T I N D I A B E T E S

8 Lockwood D. Frey ML. Gladish NA. Hiss RG. The biggest

problem in diabetes. Diabetes Educator 1986: 1 2 : 30-33.
9 Chantelau E. Relation of dietary barriers in patients with
insulin-dependent diabetes mellitus to different modes of
treatment. j Atti Dietetic Assoc 1992: 92: 1129-31.
10 Stoke K. Wolff 1. Die Behandlung der kindlichen
Zuckerkrankheit mit frei gewahlter Kost. Ergeb Inn Med
Kitiderheilkd 1939: 5 6 : 154-93.
11 Chantelau E. Sonnenberg GE. Stanitzek-Schmidt I, Best F.
Altenahr H. Berger M. Diet liberalization and metabolic control
in type 1 diabetic outpatients treated by continuous
subcutaneous insulin infusion. Diabetes Care 1982; 5:
612-16.
12 Chantelau EA. Gosseringer G. Sonnenberg GE. Berger M.
Moderate intake of sucrose does not impair metabolic control
in pump-treated diabetic out-patients. Diabetologia 1985 : 28:
204-7.
13 Chantelau EA. Frenzen A. Gosseringer G. Hansen I. Berger M.
Intensive insulin therapy justifies simplification of the diabetes
diet: a prospective study in insulin-dependent diabetics. A m j
Clin Nutr 1987: 45: 958-62.
1 4 Mulhauser I. Bruckner I. Berger M. Cheta D, Jorgens V.
Ionescu-Tirgoviste C et al. Evaluation of a n intensified insulin
treatment and teaching programme as routine management
of type 1 (insulin-dependent) diabetes. The BucharestDusseldorf Study. Diabetologia 1987: 30: 681-90.
1 5 Starostina EG. Antsiferov M. Galstyan GR. Trautner Ch,
Jorgens V. Bott U et al. Effectiveness and cost-benefit analysis
of intensive treatment and teaching programmes for type 1
(insulin-dependent) diabetes mellitus in Moscow - blood glucose versus urine glucose self-monitoring. Diabetologia 1994:
37: 170-76.
16 Waldhausl WK. The physiological basis of insulin treatment
-clinical aspects. Diabetologia 1986; 29: 837-49.
17 Jorgens V. Grusser M. Bott U. Muhlhauser I. Berger M.
Effective and safe translation of intensified insulin therapy to
general internal medicine departments. Diabetologia 1993 ;
3 6 : 99-105.
18 Bott U. Jorgens V. Grusser M. Bender R. Muhlhauser I. Berger
M. Predictors of glycaemic control in type 1 diabetic patients
after participation in an intensified treatment and teaching
programme. Diabetic Med 1994: 1 1 : 362-71.

597

19 Muhlhauser I. Sulzer M. Berger M. Quality assessment of

diabetes care according to the recommendations of the St
Vincent Declaration: a population-based study in a rural area
of Austria. Diabetologia 1992: 35: 429-35.
2 0 Waadt S, Herschbach P. Duran G. Henrich G. Hillebrand B.
F.
Entwicklung
eines
Fragebogens
zu
Strian
Behandlungsproblemen und Therapiezuweisung bei Patienten
mit Diabetes mellitus. Prax KIin Verhaltensmed Rehab 1992: 5 :
306-12.
21 McCullagh P. Regression models for ordinal data. j R Statist
Soc 1980; 842: 1 0 9 4 2 .
22 SAS Technical Report P-200. SAS/STAT Software: CALIS and
LOGISTIC Procedures, Release 6.04. Cary. NC: SAS Institute
Inc.. 1990.
23 SAS/STAT Guide for Personal Computers. Version 6 Edition.
Cary, NC: SAS Institute Inc., 1987.
24 The DCCT Research Group. Epldemiology of severe
hypoglycemia in the diabetes control and complications trial.
Am Med 1991: 90: 450-59.
25 The DCCT Research Group. The effect of intensive treatment of
diabetes o n the development and progression of long-term
complications in insulin-dependent diabetes mellitus. N Engl j
Med 1993: 329: 977-86.
2 6 Stieber J, Heme HW, Doring A, Filiplak-Pittroff B. Tietze M.
Gostomzyk J, Keil U. KardiovaskulBres Risikofaktorenprofil.
Verlaufsdaten aus der MONICA-Studienregion Augsburg
(1984/85 vs. 1989/90). Munch M d Wschr 1991; 133:
370-75.
2 7 Richter WO, Mohrle W, Schwandt P. Cholesterinwerte in der
Bayerischen Bevolkerung. Munch Mcd Wschr 1992; 134:
361-4.
28 Bergmann KE. Menzel R, Bergmann 13, rietze K, Stolzenberg H.
Hoffmeister H. Verbreitung von Ubergewicht in der
erwachsenen Bevolkerung der Bundesrepublik Deutschland.
Akt Erniihr 1989; 1 4 : 205-8.
Received 19 September 1994: accepted 9 February 1994.

Correspondence: Dr Ingrid Muhlhauser. Medizinische Klinik der

Universitat Dusseldorf, Abteilung Stoffwcchsel und Ernahrung,
MoorenstraBe 5. D-40225 Dusseldorf. Germany.

0 1995 Blackwell Science Ltd journal of Internal Medicine 2 3 7 : 591-597