A randomized crossover trial
OBJECTIVE—To assess the effect of replacing red meat with chicken in the usual diet and the effect of a low-protein diet on glomerular filtration rate (GFR), urinary albumin excretion rate (UAER), and lipid levels in patients with type 2 diabetes.
RESEARCH DESIGN AND METHODS—A randomized, crossover, controlled trial was conducted with 28 patients with type 2 diabetes (seven women; mean age 58.1 years): 15 patients were normoalbuminuric (UAER <20 μg/min), and 13 patients were microalbuminuric (UAER 20–200 μg/min). A chicken-based diet (red meat replaced with chicken) and a low-protein diet were compared with the patients’ usual diet. Patients followed each diet for 4 weeks with a 4-week washout period between. GFR (51Cr-EDTA single-injection technique), 24-h UAER (immunoturbidimetry), apolipoprotein B, total cholesterol, LDL cholesterol, HDL cholesterol, and triglycerides were measured after each diet.
RESULTS—Normoalbuminuric and microalbuminuric patients with diabetes were analyzed separately. In normoalbuminuric patients, GFR after the chicken (101.3 ± 22.9 ml · min−1 · 1.73 m−2) and low-protein diets (93.8 ± 20.5 ml · min−1 · m−2) was lower than after the usual diet (113.4 ± 31.4 ml · min−1 · 1.73 m−2; P < 0.05). In microalbuminuric patients, apolipoprotein B levels were lower after the chicken (113.5 ± 36.0 mg/dl) and low-protein diets (103.5 ± 40.1 mg/dl) than after the usual diet (134.3 ± 30.7 mg/dl; P < 0.05). Only the chicken diet reduced UAER (median 34.3 μg/min) compared with the low-protein (median 52.3 μg/min) and usual (median 63.8 μg/min) diets (P < 0.05). Glycemic control and blood pressure did not change after the diets.
CONCLUSIONS—A normoproteic diet with chicken as the only source of meat may represent an alternative strategy for treatment of patients with type 2 diabetes and microalbuminuria.
- CD, chicken diet
- GFR, glomerular filtration rate
- LPD, low-protein diet
- UAER, urinary albumin excretion rate
- UD, usual diet
Microalbuminuria affects 20% of patients with type 2 diabetes (1). It is a predictor of end-stage renal failure and an independent risk factor for coronary artery disease (2), which is the main cause of death in patients with type 2 diabetes (3). The treatment of microalbuminuric patients with diabetes includes achievement of optimal metabolic and blood pressure control and use of ACE inhibitors (4). Although high levels of serum cholesterol have been considered a risk factor for development of diabetic nephropathy in patients with type 2 diabetes (5), the effect of reducing cholesterol levels on urinary albumin excretion rate (UAER) is still controversial (6,7). ACE inhibitors have a renoprotective effect on microalbuminuric patients with type 2 diabetes because they stabilize plasma creatinine and albuminuria (8). Although ACE inhibitors are associated with few adverse effects, in some cases, cough, hypersensitivity reaction, decrease of renal function, and severe hypoglycemic episodes requiring hospitalization may limit their use (9). Furthermore, the cost of these drugs may prevent their use in a significant proportion of patients, especially in countries in which the health care system does not cover drug expenses. Therefore, in some situations, it is necessary to use additional measures to halt the progression of microalbuminuria. A low-protein diet can reduce microalbuminuria (10,11) and slow the decline of renal function in proteinuric patients with diabetes (12). However, the long-term safety of this diet is not firmly established.
We have previously described that replacement of red meat with chicken and fish in normoalbuminuric patients with type 1 diabetes reduces hyperfiltration (13), a possible risk factor for diabetic nephropathy (14). Therefore, this study was conducted to analyze the impact of replacing the red meat in the usual diet with chicken and the impact of a low-protein diet on renal function in normoalbuminuric and microalbuminuric patients with type 2 diabetes.
RESEARCH DESIGN AND METHODS
Patients with type 2 diabetes (according to the World Health Organization criteria) attending the Endocrine Division’s outpatient clinic (Hospital de Clínicas de Porto Alegre, Brazil) were selected on the basis of the following criteria: age <75 years, BMI <32 kg/m2, good compliance with diabetes treatment, triglyceride levels <4.52 mmol/l, UAER <200 μg/min, normal liver and thyroid function, and absence of urinary tract infection, other renal disease, and cardiac failure. Treatment with antihypertensive and oral antidiabetic agents was maintained during the study. To avoid potential confounding factors on the effect of diets on renal function and lipid profile, none of the patients were using either ACE inhibitors or hypolipidemic agents. Eligible patients entered a run-in period of ∼2 months in which they were oriented to achieve the best possible metabolic control through dietary adjustments and use of oral antidiabetic agents or insulin. The dietary adjustments were designed to meet the recommendations of the American Diabetes Association as closely as possible (15). However, usual protein intake (1.2–1.5 g/kg body wt) was maintained. At the end of the run-in period, a baseline evaluation was performed. BMI was calculated and sitting blood pressure was measured twice to the nearest 2 mmHg, after a 10-min rest, using a standard mercury sphygmomanometer (phases I and V of Korotkoff sounds). Hypertension was defined as blood pressure ≥140/90 mmHg or when a patient was taking antihypertensive drugs.
Study design, diet composition, and prescription
This study followed a randomized, crossover, controlled clinical trial design. The protocol was approved by the Ethics Committee of the Hospital. All patients gave written informed consent. Three types of diets were assessed: the patient’s usual diet (UD), a low-protein diet (LPD), and a chicken diet (CD). The protein content of the LPD was 0.5–0.8 g/kg per day (vegetable and milk protein only). In the CD, the red meat of the UD was replaced with chicken meat (skinless leg quarter). Patient were assigned to follow the three diets in random order for a period of 4 weeks each. Randomization was stratified according to UAER in normoalbuminuric and microalbuminuric patients separately after the run-in period, as shown in Fig. 1, by one of the authors (T.Z.). The executors (C.M., V.D.M.) were blinded to the randomization process. Consecutive eligible patients were assigned to a sequence of diets as follows: 1) UD, LPD, CD; 2) UD, CD, LPD; 3) LPD, UD, CD; 4) LPD, CD, UD; 5) CD, LPD, UD; and 6) CD, UD, LPD. After each diet, the patients underwent a 4-week washout period during which they maintained their usual diet. Compliance was assessed by 2-day weighed-diet records and measurement of urea in 24-h urine samples at the end of the second and fourth weeks, as previously described (16). The criterion standard for the estimation of protein intake was 24-h nitrogen output. Dietary macronutrients and micronutrients were analyzed using the Nutrabase ‘98 Clinical Nutritional Manager software (Cybersoft, Phoenix, AZ) (17). The primary outcome measures were UAER and GFR. Because there were no previous studies examining the role of chicken diet on UAER in microalbuminuric patients with diabetes, the sample size was calculated based on the reduction of albumin clearance corrected by GFR (fractional albumin clearance) after LPD in microalbuminuric patients with type 1 diabetes (18). To obtain a reduction of 3.61 (×10−6) on the fractional clearance of albumin, it was estimated that 10 microalbuminuric patients had to be included for a significance level of α = 0.05 and a power of 1– β = 0.80.
GFR was measured using the 51Cr-EDTA single-injection technique (coefficient of variation 12%, GFR reference range 72–137.5 ml · min−1 · 1.73 m−2). Urinary albumin was measured in 24-h timed sterile urine samples by immunoturbidimetry (Sera-Pak immuno microalbuminuria; Bayer, Tarrytown, NY; mean intra-assay and interassay coefficients of variation 4.5 and 11.0%, respectively).
Microalbuminuria, confirmed at least twice in a 6-month period, was considered to be present when UAER was 20–200 μg/min. Fractional clearance of albumin was calculated by dividing albumin clearance by GFR. Serum albumin level was determined by the bromocresol green technique, glucose level was determined by a glucose oxidase method, creatinine level was measured by the Jaffé reaction, HbA1c was determined by an ion-exchange high-performance liquid chromatography procedure (Merck-Hitachi L-9100 glycated hemoglobin analyzer, reference range 2.7–4.3%; Merck, Darmstadt, Germany), and fructosamine was determined by a colorimetric method (normal range 1.87–2.87 mmol/l). Urinary urea was measured by an enzymatic ultraviolet method (mean intra-assay coefficient of variation 3.8%). The 24-h nitrogen output was obtained according to the following formula: nitrogen intake = urinary urea nitrogen (urea/2) + nonurinary nitrogen (0.031 g/kg). Protein intake was calculated as nitrogen intake × 6.25 (19). Triglycerides and total cholesterol were measured by enzymatic colorimetric methods (Merck Diagnostika, Darmstadt, Germany; Boehringer Mannheim, Buenos Aires, Argentina), and HDL cholesterol was measured by a precipitation method using magnesium chloride and heparin. LDL cholesterol was calculated from the Friedewald equation (20). Apolipoprotein B was determined by immunoturbidimetric method (Unimate 3; Roche Diagnostic System, Basel, Switzerland; intra-assay coefficient of variation 4.2%).
The characteristics of normoalbuminuric and microalbuminuric patients were analyzed using the unpaired Student’s t test or Mann-Whitney’s test or Fisher’s exact test and χ2 test, as appropriate. Differences among the three types of diets were tested either by repeated-measures ANOVA and Student-Newman-Keuls post-hoc test or by Friedman’s ANOVA and least significant differences post-hoc test. P values <0.05 were considered statistically significant. Results were expressed as means ± SD unless otherwise stated. SPSS software (SPSS, Chicago, IL) was used for the analyses.
Of the 53 eligible patients, 20 were excluded from the study for various reasons, and 33 (16 normoalbuminuric and 17 microalbuminuric patients) were randomized to follow one of the six diet sequences described in Fig. 1. Five of the randomized patients (one normoalbuminuric and four microalbuminuric patients) did not complete the study protocol; therefore, 28 patients (15 normoalbuminuric and 13 microalbuminuric) were included in the final analysis. One patient was excluded from the GFR analysis due to technical problems in the measurement of GFR, and one patient was excluded from the UAER analysis because the 24-h urine sample collected after LPD was lost. The main clinical and laboratory characteristics of normoalbuminuric and microalbuminuric patients with type 2 diabetes are reported in Table 1. All women were postmenopausal, but none was taking hormone replacement therapy. Most of these patients were treated with calcium channel blockers. One patient in each group was being treated with a combination of thiazide and calcium channel blockers. The oral antidiabetic agents used were sulfonylurea and metformin. Two normoalbuminuric and four microalbuminuric patients were taking metformin.
Characteristics of the diets
The characteristics of the diets followed by the patients, as assessed by the weighed record method, are described in Table 2. The energetic, protein, carbohydrate, total fat, calcium, and fiber intake was the same for patients after either the CD or the UD; however, the intake of saturated fatty acids, phosphorus, iron, and zinc was lower in the CD. Carbohydrate intake was higher in the LPD, but the total intake of energy and fat, and the intake of saturated and monounsaturated fatty acids, cholesterol, phosphorus, iron, and zinc was lower. The calcium intake in the LPD was the same as in the UD.
The total protein intake (grams per kilogram body weight), as assessed by nitrogen output in the UD (1.31 ± 0.3), in the CD (1.35 ± 0.4), and in the LPD (0.62 ± 0.2), was similar (P > 0.05) to that estimated by weighed-diet records (UD 1.43 ± 0.3, CD 1.35 ± 0.3, LPD 0.66 ± 0.2).
Effect of the diets on renal function
Renal function parameters after the diets are shown in Table 3. GFR values in normoalbuminuric patients were lower after the LPD and the CD than after the UD; there was no difference between GFR values after the LPD and the CD. In microalbuminuric patients, the GFR values were lower only after the LPD as compared with the CD and the UD. No difference was observed between UD and CD GFR values.
In normoalbuminuric patients, UAER levels were similar after the diets. However, in microalbuminuric patients, UAER levels were lower after the CD as compared with the levels observed after the UD and the LPD. No difference was observed in UAER levels after the UD and the LPD. The same was observed for fractional clearance of albumin (median × 10−6, range): CD 7.35 (1.04–42.04), LPD 12.04 (2.72–42.49), UD 13.25 (6.22–44.16); P < 0.05.
Effect of the diets on serum lipids
Lipid levels after the diets are shown in Table 3. In normoalbuminuric patients, the diets did not affect the lipid profile. In microalbuminuric patients, apolipoprotein B and total cholesterol were lower after the LPD and the CD than after the UD. No difference was observed between the CD and the LPD. The reduction in LDL levels after CD and LPD did not reach conventional statistical significance (P = 0.097). Triglyceride and HDL levels were not different after the diets.
Effect of the diets on glycemic control, blood pressure levels, and body weight
Glucose control, assessed by fasting plasma glucose and fructosamine, was similar (P > 0.05) after the UD (3.22 ± 0.55 mmol/l), CD (3.14 ± 0.49 mmol/l), and LPD (3.24 ± 0.52 mmol/l). Mean blood pressure levels were also comparable (P = 0.299) after the UD (96.8 ± 7.6 mmHg), CD (96.9 ± 9.2 mmHg), and LPD (94.9 ± 8.6 mmHg). Body weight did not differ after the UD (70.8 ± 8.9 kg) and CD (70.9 ± 8.9 kg), but there was a small but significant decrease (P = 0.025) after the LPD (70.0 ± 9.0 kg).
In this study, we observed a significant reduction of 36% (95% CI 15.8–56.4) in the UAER levels of patients with type 2 diabetes and microalbuminuria after the CD when compared with their usual diet. It was also observed that both the LPD and the CD reduced total cholesterol, LDL, and apolipoprotein B levels in microalbuminuric patients and GFR in normoalbuminuric patients. In microalbuminuric patients, only the LPD decreased GFR.
Changes in the amount or source of protein can influence renal function. The reduction of GFR observed in our normoalbuminuric patients after the CD and the LPD confirms our previous observation in patients with type 1 diabetes (13,21). Other authors have also reported reduction of GFR in normoalbuminuric patients with type 1 diabetes after protein restriction (22) or replacement of animal protein with vegetarian protein (23). The mechanism behind this effect is still unknown; it might result from hemodynamic glomerular factors related to reduction of renal plasma flow (24). However, we did not observe a decrease in UAER after the CD and the LPD in our normoalbuminuric patients with type 2 diabetes, as reported by others in patients with type 1 diabetes. This absence of effect could be related to the lower levels of UAER (median ∼2.9–3.9 μg/min) in our patients, which could be masquerading an eventual effect of these diets. Furthermore, patients with type 2 diabetes may be less responsive to the effect of conditions that affect renal function.
The reduction of UAER in microalbuminuric patients after the CD is unlikely to have been caused entirely by hemodynamic factors, because the decline in GFR in these patients was not significant. Furthermore, the decrease of UAER after the CD was still observed when albuminuria was corrected by GFR (fractional albumin clearance). The lower levels of total cholesterol and of apolipoprotein B in microalbuminuric patients may contribute to the reduction of UAER after the CD. A reduction in the levels of these lipid components was also observed after the LPD. Such a reduction was probably related to the lower dietary content of saturated fatty acids. The magnitude of this reduction (∼18%) is clinically substantial and was only observed in patients with microalbuminuria, probably because of the higher levels of serum cholesterol that are characteristic of this group of patients (5). It is well known that the amount of saturated fatty acids in the diet are predictive of serum cholesterol levels and that replacement of saturated fat with unsaturated fatty acids reduces cholesterol levels (25). In a typical North American diet, replacement of saturated fatty acids with monounsaturated fat (derived from macadamia nuts) has been shown to induce a significant reduction (∼5%) of total and LDL cholesterol (26). High levels of serum cholesterol have been considered a risk factor for development of diabetic nephropathy in patients with type 2 diabetes (5), and high ingestion of saturated fat has been associated with the presence of microalbuminuria in patients with type 1 diabetes (27). Nevertheless, the role of reducing cholesterol levels in the progression of microalbuminuria is still controversial. Administration of simvastatin was associated with a reduction of UAER in microalbuminuric patients with type 2 diabetes in one study (6) but not in another (7).
One intriguing aspect of our study was the observation that the LPD reduced GFR and cholesterol levels, but the reduction of UAER did not reach statistical significance. Other authors observed that a moderate protein restriction was associated with a reduction of UAER in patients with type 2 diabetes (10,11). This discrepancy may be explained by the reported improvement in metabolic control (10,11) and blood pressure levels (11) in these studies, or it could be that a longer period of LPD would be necessary to induce a reduction in UAER. The lower effect of the LPD on UAER in comparison to the CD suggests that some component of chicken meat might have a specific beneficial effect on renal function. This hypothesis is supported by a recent study in which patients with type 1 diabetes and high intake of fish fat and protein (white meat) showed a reduction in the risk for microalbuminuria (28).
Although this study was not designed to analyze the effect of individual fatty acids, chicken meat has a lower amount of saturated fat and a higher proportion of polyunsaturated fatty acids than red meat, namely linoleic acid (1.5 vs. 0.21 g/100 g), linolenic acid (0.3 vs. 0.2 g/100 g), and arachidonic acid (0.12 vs. 0.04 g/100 g) (17). Polyunsaturated fatty acids have been shown to have a beneficial effect on endothelial function, and this could reduce UAER (29).
The present study followed a crossover design, and there is a possibility that the effect of one diet carried over into the next diet, thus influencing the results. However, a significant carryover effect is unlikely, because the diets were given in a random order, and the 4-week washout period was enough to wean patients from the previous diet. Plasma lipids and lipoproteins reach a stable level after a period of 3–4 weeks of dietetic modification (30). Outpatient feeding studies may present compliance problems. These difficulties were not remarkable in our study. According to the assessment by weighed-diet records and urea measurements performed during and at the end of each diet period, the diets that were actually consumed by the patients were very similar to those that had been prescribed. We had already observed this compliant behavior in a similar group of patients (16). The interpretation of our results should take into account that the patients included in this study were relatively well controlled and compliant. These patients were selected to avoid the interference of possible changes in blood pressure and metabolic parameters during the study on analyzed variables. Furthermore, the effect of the CD on UAER and lipid levels observed in this short-term study and its acceptability should be confirmed in long-term studies, ideally with assessment of outcomes such as progression to renal failure and cardiovascular mortality.
In conclusion, a diet in which red meat is replaced with chicken might represent an alternative or additional strategy for the treatment of microalbuminuric patients with type 2 diabetes.
This study was partially supported by Fundação de Amparo à Pesquisa do Estado do Rio Grande do Sul, Projeto de Núcleos de Excelência do Ministério de Ciência e Tecnologia, Hospital de Clínicas de Porto Alegre, and scholarships from Fundação Coordenação de Aperfeiçoamento de Pessoal de Ensino Superior (to T.Z., V.M., and M.P.).
We thank Cristiane Jucá for helping in the collection of the data and Luís Henrique Canani for the careful review of the manuscript and useful suggestions.
Address correspondence and reprint requests to Mirela J. de Azevedo, MD, Serviço de Endocrinologia do Hospital de Clínicas de Porto Alegre, Rua Ramiro Barcelos 2350, Prédio 12, 4° andar, 90035-003, Porto Alegre, RS, Brazil. E-mail:.
Received for publication 29 May 2001 and accepted in revised form 29 December 2001.
A table elsewhere in this issue shows conventional and Système International (SI) units and conversion factors for many substances.
- DIABETES CARE