Adherence to protein restriction in patients with type 2 diabetes ...

European Journal of Clinical Nutrition (2000) 54, 347±352 ß 2000 Macmillan Publishers Ltd All rights reserved 0954±3007/00 $15.00 www.nature.com/ejcn

Adherence to protein restriction in patients with type 2 diabetes mellitus: a randomized trial LTJ Pijls1*, H de Vries1, JThM van Eijk2 and AJM Donker3 1

Vrije Universiteit Amsterdam, Institute for Research in Extramural Medicine, Amsterdam, The Netherlands; 2University of Maastricht, Department of Health Care Studies, Section of Medical Sociology, Maastricht, The Netherlands; and 3Academisch Ziekenhuis Vrije Universiteit Amsterdam, Department of Internal Medicine, Amsterdam, The Netherlands

Objective: To describe the extent to which diet counselling can decrease protein intake, and to identify predictors of adherence. Design: (1) Randomized trial; (2) observational longitudinal study. Subjects: (1) 125 type 2 diabetic patients in primary care; (2) 59 patients in the experimental group. Intervention: For a period of 12 months, dieticians provided guidance on protein restriction (experimental group, n 59) or the usual dietary advice (control group, n 66). Outcome measures: Adherence was estimated primarily from urinary urea excretion (UUE), but also from foodfrequency questionnaires (FFQ). Results: After 6 months protein intake was, according to the UUE and the FFQ, respectively, 8 g=day (95% CI ÿ2, 13) (8%) and 15 g=day (95%ÿCI 9, 22) (16%) lower in the experimental than in the control group. After 12 months these differences were smaller. Linear regression analysis indicated that protein restriction was greater in patients who were well satis®ed with their pre-existing diet (r 0.32, bper 1=10 3.6 (1, 6) g), in patients who were less overweight (r 0.32, bper kg.mÿ2 1.1 (0.2, 2.0) g), and in patients living alone (r 0.22, b 7.7 (ÿ2, 17) g). These combined factors explained only 11% of variation in adherence. Adherence was not predicted by the number of barriers reported by the patients or by coinciding changes in diet satisfaction. Conclusions: The diet counselling resulted in a very moderate degree of protein restriction only. Predictors of adherence could be identi®ed, but only a few, and their predictive power was limited. Descriptors: adherence; blood pressure; communication; compliance; type 2 diabetes mellitus; patient education; protein-restricted diet; questionnaire; randomised controlled trial; smoking; urea. European Journal of Clinical Nutrition (2000) 54, 347±352

Introduction Diet plays a key role in the management of diabetes mellitus and in the prevention of its complications (Diabetes and Nutrition Study Group of the European Association for the Study of Diabetes, 1995; American Diabetes Association, 1996). Restriction of diet protein bene®cially in¯uences long-term renal functioning in type 1 diabetic patients with microalbuminuria or clinically obvious nephropathy (Pedrini et al, 1996; Waugh & Robertson, 1997). Studies in type 2 diabetes mellitus also suggest bene®cial effects (Sugimoto et al, 1991; Pijls et al, 1999a). For patients with microalbuminuria, European and American consensus statements recommend moderate protein restriction, and for normoalbuminuric patients protein intake should not be excessively high (Diabetes and Nutrition Study Group of the European Association for the study of Diabetes, 1995; American Diabetes Association, 1996). *Correspondence: LTJ Pijls, Health Council of the Netherlands, PO Box 16052, 2500 BB The Hague, The Netherlands. E-mail: [email protected] Guarantor: LTJ Pijls. Contributors: LP prepared and coordinated the ®eld work, analysed the data, and wrote the paper. HdV cooperated in the preparation and organization of the ®eld work, and in writing the paper. JvE contributed to the writing of the paper. AD and all other investigators were involved in designing the study. Received 2 July 1999; revised 17 November 1999; accepted 25 November 1999

Lifestyle modi®cation, eg a change of diet, is even more dif®cult to achieve than adherence to drug treatment (Connett & Stamler, 1984; Glasgow et al, 1987; Hanestad & Albrektsen, 1991). The same applies to compliance with preventive, compared to curative regimens (Glanz, 1980; Roderick et al, 1997). Given the important role of diet in the control of diabetes, and maintaining good health in general, one may expect that strategies aimed at improving diet adherence receive substantial scienti®c attention. This appears, however, not to be the case. It is currently impossible to predict with much certainty who is, and who is not likely to achieve good diet adherence (Van Horn et al, 1997). Knowledge of predictors offers health professionals the choice of addressing their counselling mainly to those who are rather amenable to lifestyle modi®cation, or to those who would require additional efforts in order to achieve a worthwhile change. In addition, knowledge of the predictors of diet adherence is essential in order to improve strategies to promote adherence (Dolecek et al, 1986; Coyne et al, 1995; Gillis et al, 1995; Milas et al, 1995). With regard to adherence to drug treatment, it has been demonstrated that even the most effective interventions achieve little improvement (Haynes et al, 1996). The full bene®ts of any regimen will, obviously, not be achieved until efforts are directed towards increasing adherence. The improvement of adherence is expected to pay larger dividends than the development of any new medication or diet regimen.

Adherence to protein restriction in diabetes LTJ Pijls et al

348

The present study ®rst assessed, in type 2 diabetic patients in primary care, the extent to which diet counselling on protein restriction would decrease actual protein intake. The counselling was based on the Health Counselling Model. This model uses psychological concepts, such as self-management, that have shown to be essential in establishing and maintaining desired changes in behaviour (Gerards & Hospers, 1991; Mesters et al, 1999). Second the role of potential predictors of adherence to protein restriction was explored. Patients and methods Design The effect of a diet counselling programme on protein restriction was assessed in a randomized trial. Within the experimental group of this trial, predictors of adherence to protein restriction were identi®ed in an observational longitudinal approach. The study protocol was approved by the Medical Ethics Committee of the Vrije Universiteit in Amsterdam. Patients Details of the selection and randomization procedures have been described previously by Pijls et al (1999a). The patients who entered the randomized trial had type 2 diabetes mellitus (WHO, 1985), were in primary care, were younger than 79 y of age, were not in a period of recovery from severe morbidity, and had no protein-losing enteropathy, venous leg ulcer, pressure ulcer or malignancy, and no psychiatric or serious psychosocial problems. In a quali®cation period that preceded the trial they had had two visits of about 45 min to a dietician during which they received counselling on saturated fat restriction, they had ®lled in a food-frequency questionnaire (Grootenhuis et al, 1995; Pijls et al, 1999b) and had collected two 24 h urine samples. Patients meeting the above-mentioned criteria were included in the trial if they also met the following criteria: (a) microalbuminuria (30 ± 300 mg 24 h, mean of two samples), relatively high albuminuria within the normoalbuminuric range (albuminuria > 20 mg 24 h in at least one sample or urinary albumin concentration > 6.5 mg=l in two samples), or a diabetes duration of at least 5 y; and (b) protein intake 0.80 g=kg per day (according to the urinary urea excretion). Additional criteria to be included in the trial were: to have kept the appointments in the quali®cation period; to seem to be able, according to the dietician, to understand the diet information; and to be not reluctant to try to change their diet. Eventually, the randomized trial included 160 patients. 81 and 79 patients were randomly allocated to the experimental or control group, respectively. A ®rst step in the analysis of potential predictors of adherence was to divide the patients in the experimental group into two sub-groups, according to their degree of adherence. One sub-group consisted of patients who had decreased their protein intake by more than 5 g=day; the other sub-group had had a smaller decrease or an increase in protein intake. Note that a decrease is a desired negative change. The cutoff of ÿ5 g=day more or less equals the median change in protein intake. The 5 g=day cutoff also re¯ects the protein content of, for example, one portion of cheese put on a sandwich.

European Journal of Clinical Nutrition

Intervention and control conditions In both study groups, counselling on saturated fat was continued during a 12-month period. In the experimental group only, the additional aim of the dietary guidance was to decrease protein intake to 0.8 g=kg. The strategy applied was to partially replace protein by unsaturated fat and carbohydrates in combination with water-soluble, nondigestible carbohydrates (Diabetes and Nutrition Study Group of the European Association for the study of Diabetes, 1995; American Diabetes Association, 1996). Since efforts to reduce body weight in those patients who were overweight had been undertaken previously, the diet advices were isocaloric with the patients' habitual diets. The counselling was based on the Health Counselling Model (Gerards & Hospers, 1991; Mesters et al, 1999), and the dieticians were trained by Dr Gerards and other experienced tutors. The dieticians tried to identify barriers the patients encountered, tried to involve patients in ®nding solutions, and gave regular feedback. After randomization, patients were seen at 0, 1, 3, 6, 9 and 12 months. If applicable and feasible, the patient's partner also attended the consultations. Consultations lasted for approximately 30 min. The dieticians gave feedback that was based on repeated dietary interviews and on patients' protein intake estimated from urinary urea. Patients were not informed about the association between urea and protein intake. They were invited to contact the dietician if they had any questions between consultations. Individually designed daily menus, based on the patients' habitual diets, were given as a variable basis. These menus were accompanied by lists indicating portions of foods similar in protein content, a food composition table, and recipes. With each study group a separate plenary education meeting was held with the dieticians and researchers. Both groups received the same amount of attention. Measurements The outcome measure protein intake was estimated from urinary urea excretion (BM=Hitachi 747=737, Boeringer Mannheim; Maroni et al, 1985), which is considered to be the most valid reference method available (Bingham et al, 1994; Stamler et al, 1996). In both study groups urinary urea was measured in two 24 h samples at baseline and after 6 and 12 months. In the experimental group it was assessed, additionally, in single 24 h samples at 3 and 9 months. Adjustments were made for deviations from the 24 h collection period, using patient records of the time at the start and at the completion of urine collections. In addition to the objective estimate based on urinary urea, protein intake and its change was estimated subjectively by means of a food-frequency questionnaire (Grootenhuis et al, 1995; Pijls et al, 1999b). Patients in the experimental group were divided into two subgroups, according to their change in protein intake as estimated from urinary urea. A number of potential predictors of adherence were assessed. Education was quanti®ed as the highest level of education attained by the patient or, if higher, that of the partner. Eight levels could be chosen, ranging from less than 6 y of primary school to a university degree. A selfrating of health was obtained by asking in writing: `Do you feel very healthy, healthy, rather healthy, moderately healthy or not healthy?' Hypertension was de®ned as the use of antihypertensive medication, diastolic blood pressure > 95 mmHg or systolic blood pressure > 160 mmHg. The


methods of other potential predictors have been previously described by Pijls et al (1999a). Diet satisfaction was measured before the start of the intervention, and 12 months later, on a visual analogue scale (VAS), ®lled in after breakfast, lunch and dinner on three consecutive days. Patients were asked in writing: `How much did you enjoy the meal you just had?' The extremes of the scale, ranging from an (invisible) 0 to 10, were denoted as ùnpleasant' and `pleasant', respectively. Before the third consultation 3 months later, all patients ®lled in a questionnaire relating to 50 barriers they might experience in trying to comply with the dietary advice given. It was a modi®ed version of a questionnaire developed by the University of Maastricht, The Netherlands. Patients indicated on a ®ve-point scale whether and, if yes, to what extent they were hindered by these barriers. The answers were dichotomized into negative, ie `not at all' or positive, ie à little' to `very much'. The number of positive answers was counted. Data analysis Calculations were made with the aid of SPSS (version 5.02). P-values are given if 0.20. The overall effect of diet counselling on protein restriction was estimated as the difference between the means of changes in protein intake in the experimental and control group. This difference was calculated by means of linear regression analysis with study group and baseline intake as the independent variables, and intake during follow-up as the dependent variable. We then sought to identify, in an observational approach within the experimental group, predictors of the degree of adherence to protein restriction. In the scope of this we ®rstly assessed the differences between the àdherent' and `non-adherent' sub-groups (see `Patients'). Statistical signi®cance was tested by applying two-sided t-tests or w2-tests. Where a variable was not normally distributed, medians with 25 ± 75 percentiles are given, and differences were tested with the Wilcoxon test. Secondly, linear regression analysis was applied in order to assess predictors of adherence on a continuous scale. Protein intake at follow-up was entered as the dependent variable, and baseline intake and potential predictors as the independent variables. In a ®rst step the association between every single potential predictor and adherence was analysed. The variables for which the partial Pearson correlation (r) with change in protein intake exceeded 0.20 and P 0.20 were regarded as predictors, and are listed in Table 4 (`Simple models'). Subsequently, these predictors were entered simultaneously in a multiple regression model (Table 4; `Multiple model'). Baseline intake itself may also be a predictor of adherence. This can be assessed with linear regression analysis of change as dependent variable and baseline intake as independent variable. The result of this analysis will, however, Table 2 Months 0 D3 ± 0 D6 ± 0 D 12 ± 0

be in¯uenced by regression to the mean. We adjusted for this phenomenon, using the Blomqvist method, which is based on the ratio between intra- and inter-subject variation in protein intake and its change (Hayes, et al, 1988; Blomqvist, 1997).

349

Results Table 1 presents a trial pro®le. Of the 160 patients included in the trial, 12-month follow-up data were available for 125 of them (78%). Loss to follow-up resulted mainly from comorbidity and patient-related logistic problems; diet problems were reported relatively rarely. In the experimental group (n 59), as compared to the controls (n 66), after 6 and 12 months, respectively, protein intake had become 8 3 (8%, P < 0.001) and 4 3 g=day (4%, P > 0.20) lower (Table 2). The food-frequency data, however, indicated a difference between the study groups of 15 3 g=day (16%) at both 6 and 12 months (data not shown in tables). Based on urinary urea, within the experimental group, 10 patients (17%) had decreased their protein intake by 0.20 g=kg or more; only one patient (2%) had decreased his or her intake to lower than 0.80 g=kg. Table 3 compares characteristics of patients who had achieved at least some degree of protein restriction (àdherence'), with those who had not (`non-adherence'). Listed are characteristics that are generally reported in clinical studies, and characteristics that differed or tended to differ with regard to subjects with `non-adherence' or with àdherence' (P < 0.20). Unadjusted for regression to the mean, protein intake was higher in àdherers' than in `nonadherers' (P < 0.005). Furthermore, among àdherers' compared to `non-adherers', self-ratings of health were higher, and there was heavier smoking, a lower intake of saturated fat, and greater alcohol consumption (all P < 0.15). Patients who had never smoked, former smokers, and current smokers had changed their protein intake according to urinary urea by 1 14, ÿ3 15 and ÿ8 15 g=day, respectively (PANOVA > 0.20, data not shown in tables). Variables we assessed but that did not appear to differ between the two groups (P > 0.20) are not listed in Table 3; Table 1 Inclusion and loss to follow-up in the experimental and control group (n 160, with informed consent) Experimental Random allocation Loss to follow-up: died comorbidity diet-related problems patient-related logistic problems Complete 12 months follow-up Microalbuminuria (30 ± 300 mg=24 h)

Control

81

79

ÿ0 ÿ6 ÿ4 ÿ12 59 17

ÿ2 ÿ4 ÿ0 ÿ7 66 22

Mean changesa in protein intake according to urinary urea excretion in experimental and control group Experimental (E, n 59) 93 19 ÿ3.2 20 ÿ3.3 15 ÿ2.8 18

Control (C, n 66)

Difference between changes in E and C (D E ÿ D C)

93 20 4.4 17 1.1 17

ÿ7.6 2.7* ÿ3.9 3.1

a

g=day; mean s.d.; P-values are listed if P 0.20. *P 0.006. European Journal of Clinical Nutrition


350

this applied to the percentage of men, ethnic origin, family history of diabetes, diabetes duration, HbA1c and hypertension. Subsequently, linear regression analysis was conducted with the potential predictors as the independent variables and the change in protein intake, on a continuous scale, as the dependent variable. A negative regression coef®cient implies that the higher the predictor value, the greater the decrease, ie desired change. It appeared that per 10 g higher baseline protein intake, the subsequent decrease was 4 1 g greater (bper 10 g ÿ4 g, data not included in the tables). When adjusted for regression to the mean, however, it appeared that a higher baseline protein intake was associated with a lower decrease (bper 10 g 3 g). Table 4 presents the variables that met the criteria for prediction, ie r> 0.20 and P 0.20. These were not met by

the number of patient barriers or change in diet satisfaction. However, we observed that the mean decrease in protein intake, according to urinary urea, was relatively large among patients with a greater pre-existing diet satisfaction, among patients with a lower body mass index, and among patients living alone. The ®nding with regard to the body mass index is remarkable, since the simple comparison of àdherence' with `non-adherence' (Table 3) indicated no difference. A closer look at the data revealed that the correlation with body mass index had resulted from a few data that tended to be outliers; the association with body mass index should thus be interpreted with extra caution. Furthermore, the multiple analysis in Table 4 indicated that their predictive properties were independent. The variation explained by the three predictors amounted to only 11%.

Table 3 Characteristicsa,b of patients in the protein-restricted group, divided into sub-groups according to the degree of adherence to protein restriction Adherence: decreasec > 5 g=day (n 28) Demographic Gender (percentage men) Age (y) Education (percentage college) Living alone (%) Medical HbA1 (%) Self-rated health (percentage (very) healthy) Cigarette pack years (number) Current smoker (%) Former smoker (%) Body weight (kg) BMI (kg=m2) Diet-related Patient barriers (number, 0 ± 50) Diet satisfaction (visual analogue scale, 0 ± 10) Change in diet satisfaction Food-frequency questionnaire estimate of intake of: Protein (g) Saturated fat (g) (en-%) Energy (MJ) Alcohol (g) Non drinker percentage

Non-adherence: decreasec < 5 g=day, or increase (n 31)

P-valuesb

61 64 7 37 10

0. 15

71 65 10 43 21 7.7 1.4 62 23(3, 37) 21 57 83 15 27.3 3.8

7.6 1.3 39 5(0, 24) 19 45 79 14 27.3 4.4

14 9 8.6 1.0 ÿ0.29 1.1

0. 10 0. 13

15 9 8.1 1.5 ÿ0.46 1.4

100 18 24.1 9 12.4 3.1 7.3 1.9 2(0, 11) 32

86 17 28.3 12 13.8 2.8 7.6 2.5 0(0, 11) 52

0. 004 0. 14 0. 08 0. 13

a

Mean s.d. or median (25; 75%) or percentage; baseline values unless stated otherwise. Only listed if P 0.20. c Mean intake during 12 month follow-up minus baseline intake, estimated from urinary urea excretion. b

Table 4 Associations between adherence to protein restriction and its predictors (n 59) Change in protein intake (g=day), according to urea excretion Simple modelsa

Baseline BMI (per kg=m2) Baseline diet satisfaction (per 1 on Visual analogue scale 1 ± 10) Living alone (vs with partner) DR2 (%)e

Multiple modelb

rc

Dd

rc

Dd

0.32 0.015f ÿ0.32 0.013f ÿ0.22 0.109f

1.1 (0.2, 2.0) ÿ3.6 (ÿ6, ÿ1) ÿ7.7 (ÿ17, 2)

0.27 0.043f ÿ0.23 0.095f ÿ0.22 0.107f

0.93 (0.0, 1.8) ÿ2.4 (ÿ5.3, 0.4) ÿ7.2 (ÿ16, 2) 11

a Linear regression with diet intake during 12 month follow-up as dependent variable, and baseline intake and predictor as independent variables; only listed if partial r > 0.20 and P 0.20. b As (a), but with all predictors in column in one multiple model. c Pearson's partial correlation. d Difference compared to reference category (95% CI). e Variation explained by addition of predictors to regression model with intakefollow-up as dependent variable and intakebaseline as independent variable. f P-values, only listed if P 0.20.



Discussion The present trial has demonstrated that a counselling program for dietary protein restriction in patients with type 2 diabetes could only moderately decrease protein intake. In the subsequent assessment, we could identify only few and weak predictors of adherence. Taking into account the selection criteria, the study population expectedly was, compared to type 2 diabetic patients in general, a rather motivated group. Compared to usual diet counselling in primary care, the intervention was quite intensive, but the achieved degree of protein restriction was disappointingly low. In order to ®nd out whether the intervention lacked adherence-facilitating elements, a comparison was made with the intervention program used in the Modi®cation of Diet in Renal Disease (MDRD) study (Milas et al, 1995). To date, this is the largest trial on protein restriction. The comparison indicated that the elements included in the MDRD intervention that related to, and thus possibly facilitated good adherence, were also included in our study. For instance, the dieticians gave feedback based on the actual protein intake estimated from urinary urea, but did not explain to the patients that the estimate was based on urinary urea. The intention of the latter was to prevent participants from omitting portions of the 24 h urine collection in order to get a lower estimate of protein intake. In this respect, it is important to note that in the present and other studies (Dolecek et al, 1986; Levine et al, 1989; Van Horn et al, 1997; Pijls et al, 1999b), self-reported data tended to over-estimate adherence. This suggests that patients did understand the concept of protein restriction, but seldom put it into practice. However, explicit confrontation of the less compliant patients with the relatively hard reality of the urinary urea estimates might have been an incentive for improving adherence. In future studies, careful consideration should be given to the advantages and disadvantages of explaining the relationship between protein intake and urinary urea. Probably, the main reason for the low degree of protein restriction was that the participants felt quite healthy, hardly threatened by renal disease, and had a low perception of the importance of compliance (Peterson et al, 1984; Cianciaruso et al, 1989; Glanz, 1997; Bakx et al, 1997; Roderick et al, 1997). Nevertheless, various measures may help to improve adherence, eg the provision of low-protein foods. Such novel products, however, might not be readily accepted by elderly people in particular (Shintani et al, 1991); in the present study these were not provided. The MDRD study demonstrated that more time was to be spent on counselling if a greater protein restriction was desired (Dolecek et al, 1995). This suggests that in the present study better adherence might have been achieved if consultations with the dieticians had been held more frequently. In an intensive intervention study with 3 months of follow-up a plant-based diet was established by groups of patients with type 2 diabetes (Nicholson, 1999). Such a diet, which focuses on certain food groups, may be understood more easily than the protein-restricted diet. The latter focuses on a nutrient, and requires `translation' to food stuffs. Apparently, the commonly encountered strong preference for meat and meat products among elderly patients does not necessarily complicate adherence to plant-based diets.

Changing lifestyle is generally known to be dif®cult (Connett & Stamler, 1984; Glasgow et al, 1987; Hanestad & Albrektsen, 1991), even more dif®cult than adherence to drug treatment (Connett & Stamler, 1984). It has been estimated that one-third of patients adhere to physicians' drug prescriptions, one-third do so partially and one-third do not adhere at all. The percentage of compliance with diet prescriptions has been estimated to be even lower (Glanz, 1980). A regimen will only be followed, in principle, by those who choose to follow it (Schectman et al, 1990); this decision is an essential step in the Health Counselling Model (Gerards & Hospers, 1991; Mesters et al, 1999). The objective of the quali®cation period was to identify and include those patients who were willing to try to modify their diet. Perhaps the patients should have been questioned more explicitly about their intention to modify their diet, and should have been selected more strictly. This would have resulted in a smaller, but probably more compliant study population. In our experimental group few factors were identi®ed that, to a moderate extent only, were related to adherence. Patients with habitually high protein intake were the ones who achieved the lowest degree of restriction. These patients, unfortunately, also happen to be the ones for whom protein restriction is particularly warranted. Furthermore, patients who were well satis®ed with their preexisting diet showed good adherence. This may be explained by the greater potential to decrease protein intake without creating an unpalatable diet. The MDRD study, the only previous study founded on diet satisfaction, also demonstrated that diet satisfaction during follow-up was a predictor of adherence (Coyne et al, 1995). Although the present study's ®ndings con®rm that a diet satisfaction questionnaire at the start of a programme helps to identify people who require extra attention, the ®ndings do not stress the importance of diet satisfaction during the intervention. Less overweight subjects did relatively well in both the present study and the Multiple Risk Factor Intervention Trial (MRFIT), which was concerned with fat restriction (Van Horn et al, 1997). Better diet adherence in patients living alone, as found in this study, has not previously been reported (Glasgow et al, 1987; Crumb-Johnson et al, 1993; Van Horn et al, 1997). Apparently, not taking into account a partner's food preferences outweighed the role of a partner's support in achieving adherence. In the present study, current smokers tended to adhere relatively well to protein restriction; in the MRFIT it was the non-smokers who did relatively well. With regard to age and adherence, previous studies report both positive (Hanestad & Albrektsen, 1991; Radimer et al, 1992; Van Horn et al, 1997) and negative (Gans et al, 1994) associations; the present study found no association with age. No relationship was found with blood pressure or the use of antihypertensive medication. In the MRFIT, however, which was concerned with fat restriction, subjects on antihypertensive medication demonstrated a relatively low level of adherence (Dolecek et al, 1986). An association between education and adherence has been reported (Reid et al, 1984), but was not observed in the present or other studies (Kiley et al, 1993; Van Horn et al, 1997). In the only study found that addressed patient barriers, the number of barriers predicted adherence (Glasgow et al, 1986). In the present study this association was surprisingly not observed.

351



352

In conclusion, the present diet counselling program resulted in a small degree of protein restriction. A protein-restricted diet for type 2 diabetic patients without nephropathy appears to be scarcely feasible. Adherence was relatively low in patients with a higher baseline protein intake, those who were less satis®ed with their pre-existing diet, in patients not living alone, and possibly in those who were overweight. These factors predicted adherence to a very moderate extent only. Nevertheless, knowledge about these predictors offers health professionals at least some opportunity to choose to address their counselling mainly to either those who are rather amenable to diet modi®cation, or to those who would require additional efforts in order to achieve a worthwhile change. References American Diabetes Association (1996): Position statement. Nutrition recommendations and principles for people with diabetes mellitus. Diabetes Care 19, S16 ± 19. Bakx JC, Sta¯eu A, Van Staveren WA, Van den Hoogen HJM & Van Weel C (1997): Long-term effect of nutritional counselling: a study in family medicine. Am. J. Clin. Nutr. 65, 1946S ± 1950S. Bingham SA (1994): The use of 24-h urine samples and energy expenditure to validate dietary assessments. Am. J. Clin. Nutr. 59 (Suppl), 227S ± 231S. Blomqvist N (1997): On the relation between change and initial value. J. Am. Stat. Assoc. 72, 746 ± 749. Cianciaruso B, Capuano A, D'Amaro E, Ferrara N, Nastasi A, Conte G, Bellizzi V & Andreucci VE (1989): Dietary compliance to a low protein and phosphate diet in patients with chronic renal failure. Kidney Int. 36, S173 ± S176. Connett JE & Stamler J (1984): Responses of black and white males to the special intervention program of the Multiple Risk Factor Intervention Trial. Am. Heart J. 108, 839 ± 848. Coyne T, Olson N, Bradham K, Garcon M, Gregory P & Scherch L (1995): Dietary satisfaction correlated with adherence in the Modi®cation of Diet in Renal Disease Study. J. Am. Diet. Assoc. 95, 1301 ± 1306. Crumb-Johnson R, Smith-Banes M, Hatcher L & Hagan DW (1993): Assessment of differences between compliers and noncompliers in outpatient research diet studies. J. Am. Diet. Assoc. 93, 1041 ± 1042. Diabetes and Nutrition Study Group of the European Association for the Study of Diabetes (1995): Recommendations for the nutritional management of patients with diabetes mellitus. Diab. Nutr. Metab. 8, 1 ± 4. Dolecek TA, Milas NC, Horn LV van, Farrand ME, Gorder DD, Duchene AG, Dyer JR, Stone PA & Randall BL (1986): A long-term nutrition intervention experience: lipid responses and dietary adherence patterns in the Multiple Risk Factor Intervention Trial. J. Am. Diet. Assoc. 86, 752 ± 758. Dolecek TA, Olson MB, Caggiula AW, Dwyer JT, Milas C, Gillis BP, Hartman JA & DiChiro JT (1995): Registered dietitian time requirements in the MDRD Study. J. Am. Diet. Assoc. 95, 1307 ± 1312. Gans KM, Lapane KL, Lasater TM & Carleton RA (1994): Effects of intervention on compliance to referral and lifestyle recommendations given at cholesterol screening programs. Am. J. Prev. Med. 10, 275 ± 282. Gerards FM & Hospers H (1991): Health Counselling; a strategy for individual guidance (in Dutch). Ned. Tijdschr. DieÈtisten 46, 132 ± 136. Gillis BP, Caggiula AW, Chiavacci AT, Coyne T, Doroshenko L, Milas C, Nowalk MP & Scherch LK (1995): Nutrition intervention program of the Modi®cation of Diet in Renal Disease Study: a self-management approach. J. Am. Diet. Assoc. 95, 1288 ± 1294. Glanz K (1980): Compliance with dietary regimens: its magnitude, measurements, and determinants. Prev. Med. 9, 787 ± 804. Glanz K (1997): Dietitians' effectiveness and patient compliance with dietary regimens. J. Am. Diet. Assoc. 75, 631 ± 636. Glasgow RE, McCaul KD & Schafer LC (1986): Barriers to regimen adherence among persons with insulin dependent diabetes. J. Behav. Med. 9, 65 ± 77. Glasgow RE, McGaul KD & Schafer LC (1987): Self-care behaviors and glycemic control in Type I diabetes. J. Chron. Dis. 40, 399 ± 412. Grootenhuis PA, Westenbrink S, Sie CMTL, De Neeling JND, Kok FJ & Bouter LM (1995): A semiquantitative food frequency questionnaire for use in epidemiologic research among the elderly: validation by comparison with dietary history. J. Clin. Epidemiol. 48, 859 ± 868.


Hanestad BR & Albrektsen G (1991): Quality of life, perceived dif®culties in adherence to a diabetes regimen, and blood glucose control. Diab. Med. 8, 759 ± 764. Hayes RJ (1988): Methods for assessing whether change depends on initial value. Stat. Med. 7, 915 ± 927. Haynes RB, McKibbon KA & Kanani R (1996): Systematic review of randomised trials of interventions to assist patients to follow prescriptions for medications. Lancet 348, 383 ± 386. Kiley DJ, Lam CS & Pollak R (1993): A study of treatment compliance following kidney transplantation. Transplantation 55, 51 ± 56. Levine SE, D'Elia JA, Bistrian B, Smith-Ossman S, Gleason R, Mitch WE & Miller DG (1989): Protein-restricted diets in diabetic nephropathy. Nephron 52, 55 ± 61. Maroni BJ, Steinman TI & Mitch WE (1985): A method for estimating nitrogen intake of patients with chronic renal failure. Kidney Int. 27, 58 ± 65. Mesters 1, Creer TL, Gerards FM. Self-management and respiratory disorders Guiding patients from health counseling and self-management perspectives. In: Behavioral Sciences and Respiratory Disorders, ed. TL Greer, A Kaptein. Reading: Harvard Academic, 1999. Milas NC, Nowalk MP, Akpele L, Castaldo L, Coyne T, Doroshenko L, Kigawa L, Korzec-Ramirez D, Scherch LK & Snetselaar L (1995): Factors associated with adherence to the dietary protein intervention in the Modi®cation of Diet in Renal Disease Study. J. Am. Diet. Assoc. 95, 1295 ± 1300. Nicholson AS (1999): Effect of a low-fat, unre®ned vegan diet on Type 2 diabetes. Am. J. Clin. Nutr. 70, 624S ± 624S. Pedrini MT, Levey AS, Lau J, Chalmers TC & Wang PH (1996): The effect of dietary protein restriction on the progression of diabetic and nondiabetic renal diseases: a meta-analysis. Ann. Intern. Med. 124, 627 ± 632. Peterson GM, McLean S & Senator GB (1984): Determinants of patient compliance, control, presence of complications, and handicap in noninsulin-dependent diabetes. Aust. N.Z. J. Med. 14, 135 ± 141. Pijls LTJ, Vries de H, Donker AJM & Van Eijk JThM (1999a): The effect of protein restriction on albuminuria in patients with Type 2 diabetes mellitus; a randomised trial. Nephrol. Dial. Transplant. 14, 1445 ± 1453. Pijls LTJ, Vries de H, Donker AJM & Van Eijk JThM (1999b): Reproducibility and biomarker-based validity and responsiveness of a food frequency questionnaire to estimate protein intake. Am. J. Epidemiol. 150, 987 ± 995. Radimer KL, Harvey PWJ, Green A & Orrell E (1992): Compliance with dietary goals in a Queensland community. Austral. J. Public Health 16, 277 ± 281. Reid V, Graham I, Hickly N & Mulcahy R (1984): Factors affecting dietary compliance in coronary patients included in a secondary prevention programme. Hum. Nutr. 38, 279 ± 287. Roderick P, Ruddock V, Hunt P & Miller G (1997): A randomized trial to evaluate the effectiveness of dietary advice by practice nurses in lowering diet-related coronary heart disease risk. Br. J. Gen. Pract. 47, 7 ± 11. Schectman G, McKinney P, Pleuss J & Hoffman RG (1990): Dietary intake of Americans reporting adherence to a low cholesterol diet (NHANES II). Am. J. Public Health 80, 698 ± 703. Shintani TT, Hughes CK, Beckham S & O'Connor K (1991): Obesity and cardiovascular risk intervention through the ad libitum feeding of traditional Hawaiian diet. Am. J. Clin. Nutr. 53, 1647S ± 1651S. Stamler J, Elliott P, Kesteloot h, Nichols R, Claeys G, Dyer AR & Stamler R, for the INTERSALT Cooperative Research Group (1996): Inverse relation of dietary protein markers with blood pressure. Findings for 10 020 men and women in the INTERSALT Study. Circulation 94, 1629 ± 1634. Sugimoto T, Kikkawa R, Haneda M & Shigeta y (1991): Effect of dietary protein restriction on proteinuria in non-insulin-dependent diabetic patients with nephropathy. J. Nutr. Sci. Vitaminol. 37, S87 ± S92. Van Horn LV, Dolecek TA, Grandits GA & Skweres L (1997): Adherence to dietary recommendations in the special intervention group in the Multiple Risk Factor Intervention Trial. Am. J. Clin. Nutr. 65, 289S ± 304S. Waugh NR & Robertson AM (1997): Protein restriction in diabetic renal disease. In: Diabetes Module of the Cochrane Database of Systematic Reviews [updated 1 September], eds. R Williams, A Nicolucci, HMJ Krans & G Ramirez. The Cochrane Library. The Cochrane Collaboration; Issue 4. Oxford: Update Software. WHO (1985): Diabetes mellitus (WHO Technical Report Series no. 727). Geneva: WHO.