Low-fructose diet lowers blood pressure and inflammation in patients with chronic kidney disease

Abstract

Background.

Fructose has been strongly linked with hypertension, hyperuricemia and inflammation in experimental models and humans. However, the effect of low-fructose diet on inflammation, hyperuricemia and the progression of renal disease has not yet been evaluated in patients with chronic kidney disease (CKD).

Methods.

Twenty-eight patients (age 59 ± 15 years, 17 males/11 females) with Stages 2 and 3 CKD were switched from a regular (basal) (60.0 g/24 h) to a low (12.0 g/24 h) fructose diet for 6 weeks, followed by a resumption of their regular diet for another 6 weeks. Diet was monitored by a dietician. At the baseline, low- and regular-fructose diet ambulatory blood pressure (BP) was measured and blood sampled for renal function (creatinine), inflammatory markers, fasting glucose and insulin and serum uric acid. Twenty-four-hour urine collections were also obtained for creatinine, uric acid, monocyte chemotatic protein-1, transforming growth factor-beta and N-acetyl-beta-D-glucosaminidase.

Results.

The low-fructose diet tended to improve BP for the whole group (n = 28), while significant reduction of BP was only seen in dippers (n = 20) but not in non-dippers (n = 8). No effects on estimated glomerular filtration rate (eGFR) or proteinuria were observed. Serum uric acid was lowered non-significantly with low-fructose diet (7.1 ± 1.3 versus 6.6 ± 1.0 mg/dL, P < 0.1), whereas a significant decrease in fasting serum insulin was observed (11.2 ± 6.1 versus 8.2 ± 2.9 mIU/mL, P < 0.05) and the reduction persisted after return to the regular diet. A slight but not significant reduction in urinary uric acid and fractional uric acid excretion was observed while the patients were on the low fructose diet. The low-fructose diet also decreased high sensitivity C-reactive protein (hsCRP) (4.3 ± 4.9 versus 3.3 ± 4.5 mg/L; P < 0.01) and soluble intercellular adhesion molecule (sICAM) (250.9 ± 59.4 versus 227 ± 50.5 ng/mL; P < 0.05). The hsCRP returned to baseline with resumption of the regular diet, whereas the reduction in sICAM persisted.

Conclusion.

Low-fructose diet in subjects with CKD can reduce inflammation with some potential benefits on BP. This pilot study needs to be confirmed by a larger clinical trial to determine the long-term benefit of a low-fructose diet compared to other diets in subjects with CKD.

Introduction

Fructose intake from added sugars has increased dramatically over the last century and has recently been implicated as a potential contributor to hypertension, inflammation and kidney disease [1]. Fructose is distinct from other sugars as uric acid is generated during its metabolism [2]. Serum uric acid levels have been found to correlate with the intake of fructose and added sugars [3]. In turn, an elevated serum uric acid has also been shown to be associated with hypertension, inflammation and chronic kidney disease (CKD) [4], and early intervention trials with xanthine oxidase inhibitors such as allopurinol have been found to have benefits on these parameters in both subjects with normal and decreased renal function [5–8].

To date, there have been no published studies to examine the effect of a low-fructose diet in subjects with or without CKD. In particular, the usefulness of a low-fructose diet in subjects with CKD could theoretically be unhelpful given the fact that decreased glomerular filtration rate (GFR) itself may lead to the retention of uric acid and sodium which can cause hyperuricemia and hypertension, respectively [9]. Furthermore, both inflammation and oxidative stress characterize subjects with reduced renal function and is likely to have multiple causes [10].

Given how little is known in this area, we performed an open pilot study to determine the effect of a low-fructose diet on serum uric acid, blood pressure (BP) and markers of inflammation [especially high-sensitivity C-reactive protein (hsCRP) and soluble intercellular adhesion molecule (sICAM) 1] in subjects with stable CKD (Stages 2 and 3). For this pilot study, we estimated fructose intake while patients were on their regular diet along with baseline parameters, followed by a 6-week trial of a low-fructose diet. Subjects then resumed to their regular diet for an additional 6 weeks to determine if the effects of the low-fructose diet were persistent.

Materials and methods

Study population

Twenty-eight patients (mean age 59 ± 15 years, 17 males/11 females) with Stages 2 and 3 CKD participated in the study. Inclusion criteria were age 18–70 years and presence of non-diabetic CKD with stable kidney function defined as change in estimated GFR [eGFR, modification of diet in renal disease (MDRD) formula] of <20% in the prior 6 months. Exclusion criteria were body mass index < 18 kg/m², symptomatic gout, a history of renal transplantation or use of immunosuppressants in the last 3 months. Documented acute bacterial and viral infections in the last 3 months and chronic inflammatory states were also excluded. Five patients were on a uric acid-lowering agent (allopurinol), which were not altered during the study. The study was approved by the local ethical committee at the Nicolaus Copernicus University in Toruń in February 2008. Since there are no prior studies using low-fructose diet in this or other populations, it was not possible to conduct a power analysis, so the target was simply to recruit 25–30 individuals.

Nutritional intervention

The study consisted of baseline laboratory studies, studies performed after 6 weeks on a low-fructose diet and then a final assessment after returning to a regular diet for 6 weeks. To establish baseline (regular) dietary intake, subjects recorded daily food intake for 2 weeks based on meal questionnaires and daily interviews with a dietician. The total daily energy intake and fructose intake were then estimated for each participant. Subjects were then instructed to restrict consumption of sucrose-sweetened as well as artificially sweetened drinks and food for 6 weeks. The low-fructose diet was designed to reduce fructose consumption by 80% compared to baseline. During the final period, subjects were allowed to resume their regular diet. Daily food intake was monitored again in the same manner as before and while the low-fructose diet was administered. Dietary compliance was monitored at 2-week intervals during the low-fructose diet and at regular diet periods. All participants reported good compliance. Both diets provided 55% of energy as carbohydrate, 30% energy as a fat and 15% of energy as a protein. At baseline, the regular diet consisted of a mean total daily fructose consumption of 59 g with 63% from added sugars; this was reduced to 12 g with 17% from added sugars during the low-fructose diet period (Table 1). During the 6 weeks when the subjects resumed to their regular diet, the mean fructose intake was not significantly different from the regular baseline diet (53 versus 59 g/day in the resumed versus basal diet, P = NS, Table 1) although the percentage of fructose from added sugars fell from 63 to 51%, respectively (P < 0.01, Table 1).

Table 1.

Effect of low-fructose diet on renal function, blood pressure and uric acid (n = 28)^a

	Regular (basal) diet	Low fructose diet	Regular (resumed) diet	P-value
Fructose intake [g/day]	59 ± 22	12 ± 3	53 ± 23	I versus II: <0.0001
Natural fructose (% of total fructose intake)	37%	83%	49%	I versus II: <0.0001; I versus III: <0.01
Caloric intake [kcal/day]	2310 ± 240	2290 ± 260	2330 ± 245	NS
Weight [kg]	85.8 ± 11.5	84.3 ± 10.9	84.3 ± 11.3	NS
BMI	29.9 ± 4.2	29.4 ± 4.1	29.4 ± 4.2	NS
Hip [cm]	109.3 ± 11.6	108.8 ± 12.4	110.7 ± 10.6	NS
Serum uric acid (mg/dL)	7.1 ± 1.3	6.6 ± 1	6.4 ± 1.2	I versus II, III <0.1
Urine uric acid [mg/24 h]	508 ± 268	470 ± 259	498 ± 303	NS
Fractional uric acid excretion (%)	6.79 ± 2.33	6.45 ± 2.76	6.42 ± 3.05	NS
Systolic BP [mmHg]	131 ± 13	127 ± 11	128 ± 12	NS
Diastolic BP [mmHg]	79 ± 10	76 ± 8	72 ± 5	I versus II <0.1
Mean arterial pressure [mmHg]	96 ± 11	92 ± 8	91 ± 7	I versus II <0.1
Creatinine (mg/dL)	1.57 ± 0.45	1.51 ± 0.42	1.50 ± 0.48	NS
Creat Cl (mL/min/1.73m2)	39 ± 17	45± 23	45 ±22	NS
eGFR (MDRD)	47 ± 13	48 ± 13	50 ± 15	NS
Urine Na excretion (mmol/24 h)	4.69±1.94	4.86 ± 3.02	5.63 ± 2.62	NS
Proteinuria (g/24 h)	0.12 ± 0.19	0.05 ± 0.11	0.11 ± 0.19	NS
NAG (u/24 h)	5.6 ± 3.1	5.4 ± 3.0	6.0 ± 3.5	NS
Glucose (mg/dL)	98.0 ± 12.0	94.1 ± 12.9	94.3 ± 11.0	NS
Insulin (uIU/mL)	11.2 ± 6.1	8.2 ± 2.9	9.0 ± 4.7	I versus II: <0.05
HOMA	0.11 ± 0.04	0.09 ± 0.05	0.10 ± 0.05	NS

^aHOMA, homeostasis model of assessment.

Analyses

At the end of baseline, low-fructose and regular-fructose diet periods, fasting blood sample and 24-h urine collection were obtained. Blood samples were assayed for: creatinine, uric acid, glucose, triglycerides (Architect ci8200; Abbott Diagnostics), insulin (AxSYM; Abbott Diagnostics), hsCRP (BNII, limit of detection 0.175 mg/L; Dade Behring), tumour necrosis factor alpha (ELISA R&D Systems) and sICAM-1 (ELISA R&D Systems). Twenty-four-hour urine was assayed for: creatinine, uric acid, sodium (Architect ci8200; Abbott Diagnostics), monocyte chemotactic protein-1 (MCP-1; ELISA R&D Systems), transforming growth factor beta (TGF-β; ELISA R&D Systems) and N-acetyl-β-glucosaminidase (NAG, commercial kit with a colorimetric method Roche, Mannheim Germany). Creatinine clearance was calculated as: 24-h urine creatinine/serum creatinine and fractional uric acid excretion as clearance of uric acid/creatinine clearance. Homeostasis model assessment of insulin resistance (HOMA-IR) was calculated: fasting insulin (mU/L) × glucose (mmol/L)/22.5.

Twenty-four-hour ambulatory blood pressure monitoring (ABPM) was performed at baseline and at 6 and 12 weeks corresponding to the end of the low-fructose diet period and subsequent regular diet periods (Monitor TM-2430; A&D Company, Limited, Tokyo, Japan). Based on the first ABPM, the patients were divided into two groups: 20 dipper patients and 8 non-dipper patients. Patients with a nocturnal reduction in average daytime systolic and diastolic BP of >10% are classed as dippers. Non-dippers were defined as individuals who did not show such a reduction.

Before the low-fructose diet, an oral fructose tolerance test was conducted in each participant. The test consisted of giving 1 g/kg body weight (b.w) of fructose p.o. with blood collected at 0, 30, 60 and 120 min afterwards for serum uric acid determination.

Statistics

Results are reported as mean ± SD. Group differences were assessed by repeated measures analysis of variance (ANOVA) and non-parametric Friedman’s ANOVA. Statistical significance is defined as P < 0.05.

Results

General characteristics

We performed a pilot study in which 28 non-diabetic patients (17 males, 11 females, mean age 59 years) with Stage 2 and Stage 3 CKD (mean eGFR 47 mL/min/1.73m²) were converted from their regular (basal) diet containing ∼60 g fructose per day (with two-thirds from added sugars) to ∼12 g/day for 6 weeks (Table 1). The diet was monitored by a dietician and was well tolerated. Following the 6-week diet, the subjects were allowed to resume to their regular diet and were followed for 6 additional weeks. Fructose intake during the resumption of the regular diet averaged 53 g/day and was not significantly different from the fructose intake during the basal period (Table 1). In addition, there was no difference in overall caloric intake during all three periods of study: baseline regular diet, 2310 ± 240 kcal/day; low-fructose diet, 2290 ± 260 kcal/day and resumed regular diet, 2330 ± 245 kcal/day.

Uric acid

The oral fructose tolerance test showed a rapid increase of serum uric acid concentration at 30 min which persisted at 60 and 120 min after consumption of 1 g/kg b.w. of fructose (baseline: 7.05 ± 1.30 versus 8.01 ± 1.35; 8.03 ± 1.44; 7.82 ± 1.43 mg/dL, respectively; P < 0.05). A primary question was whether the low-fructose diet might lower serum uric acid levels. As shown in Table 1, the low fructose diet resulted in a non-significant reduction of serum uric acid from 7.1 to 6.6 mg/dL (P < 0.1), with a slight but not significant reduction in urinary uric acid and fractional uric acid excretion. Interestingly, serum uric acid tended to remain lower even after return to the regular diet.

Blood pressure

The low-fructose diet resulted in a mild but non-significant effect on diastolic BP (79.3 ± 10.2 versus 75.6 ± 7.5, P < 0.1) with a trend for improvement in systolic and mean BP (Table 1). The trends for improved BP persisted following resumption of the regular diet. These effects were independent of urinary sodium excretion (Table 1).

Since ambulatory BP was measured, it was possible to do a subset analysis on the effects of low-fructose diet on non-dippers (n = 8) versus dippers (n = 20). As shown in Table 2, the reduction in both systolic and diastolic BP was significant in the dippers but not non-dippers.

Table 2.

Effect of low-fructose diet on renal function, BP and uric acid in dippers and non-dippers.

	Regular (basal) diet	Low fructose diet	Regular (resumed) diet	P-value
Dippers (n = 20)
Systolic BP [mmHg]	130 ± 14	124 ± 8	123 ± 11	I versus II 0.066
Diastolic BP [mmHg]	81 ± 11	75 ± 8	71 ± 3	I versus II 0.036
Mean arterial pressure [mmHg]	97 ± 11	91 ± 7	88 ± 6	I versus II 0.041
Non-dippers (n = 8)
RR systolic [mmHg]	132 ± 12	133 ± 15	138 ± 10	NS
RR diastolic [mmHg]	76 ± 8	77± 6	74 ± 7	NS
MAP [mmHg]	94 ± 9	95 ± 9	95 ± 8	NS

Renal function

We observed no benefit of a low-fructose diet on either measured creatinine clearance or eGFR using the MDRD equation (Table 1). Proteinuria (24 h) as well as tubular injury, manifested as urinary NAG excretion, also did not change during the low-fructose diet.

Inflammation

We observed a significant decrease in both hsCRP and sICAM with low-fructose diet, whereas no effect was observed in urinary MCP-1 or TGF-β (Table 3). The reduction in sICAM-1 persisted during resumption of the regular diet, whereas hsCRP returned to basal levels.

Table 3.

Effect of low-fructose diet on serum hsCRP, TNF-α, sICAM-1 concentration and 24-h urine excretion of MCP-1, TGF-β in whole group (n = 28)^a

	Regular (basal) diet	Low fructose diet	Regular (resumed) diet	P-value
hsCRP (mg/L)	4.3 ± 4.9	3.3 ± 4.5	4.2 ± 8.8	I versus II <0.01; II versus III <0.05
TNF-α (pg/mL)	2.7 ± 2.5	2.4 ± 1.7	1.8 ± 1.5	NS
sICAM-1 (ng/mL)	250.9 ± 59.4	227.7 ± 50.5	229.2 ± 45.6	I versus II 0.05; I versus III 0.02
MCP-1 (ng/24 h)	498.9 ± 384.9	496.0 ± 407.2	527.0 ± 562.0	NS
TGF-β (ng/24 h)	44.2 ± 75.6	28.2 ± 21.1	27.7 ± 19.6	NS

^aTNF-α, tumour necrosis factor alpha.

Insulin resistance

As shown in Table 1, fasting serum insulin levels were decreased by the low-fructose diet (P < 0.01). Insulin levels tended to remain low during the regular diet but this was not significant. In contrast, no significant differences were noted in either fasting glucose levels or homeostasis model assessment index (Table 1).

Discussion

Excessive intake of fructose, primarily from added sugars, has emerged as a risk factor for hyperuricemia [3], hypertension [11–13], metabolic syndrome [14] and kidney disease [15]. Furthermore, the administration of fructose to animals can experimentally induce hypertension [16, 17] renal injury [17–19], inflammation [20] and metabolic syndrome [21]. Nevertheless, to date, the use of a low-fructose diet has not been tested for its protective effects in either patients with metabolic syndrome or patients with CKD. As such, we decided to perform an exploratory pilot study to determine if a low-fructose diet can reduce serum uric acid or improve BP, renal function or markers of inflammation in a stable population with Stage 2 and 3 CKD. For this purpose, we identified 28 subjects with non-diabetic stable CKD as noted by minimal change in eGFR during the prior 6 months. A dietician met with each subject and reviewed their intake of fructose from natural sources (primarily fruit) and added sugars. Subjects were then placed on a low-fructose diet for 6 weeks. In order to determine if the benefits of the low-fructose diet would persist, we continued to follow subjects after they had resumed their regular diet for an additional 6 weeks.

A primary interest was whether this diet could reduce serum uric acid levels. Uric acid levels are modulated by numerous factors, including dietary intake of purines and fructose, endogenous generation, and renal and gastrointestinal excretion. Serum uric acid increases in patients with reduced renal function due to impaired renal excretion, and levels of 7.0 mg/dL or higher are present in ∼50% of subjects on dialysis [9]. In the current study, a low-fructose diet reduced serum uric acid by ∼0.5 mg/dL which did not meet statistical significance. It is possible that with a larger number of subjects, or with a longer duration of the diet, significance might have been achieved. However, it seems likely that to reduce uric acid levels in subjects with CKD to 5.0–5.5 mg/dL, which is what correlates epidemiologically with good BP or renal function [4] that additional measures besides diet are likely necessary.

We also found that the low-fructose diet tended to reduce BP, but this effect was most significant in dippers in whom BP falls at night. Fructose may increase BP via multiple mechanisms but one effect may be by stimulation of the sympathetic nervous system [22]. In addition, the effects of fructose to raise BP mostly manifests when animals are ingesting fructose [16]. Indeed, in humans, the ingestion of fructose results in increases in BP occur shortly after ingestion [23] and tend to be most elevated during the day [24]. Hence, the inability for low-fructose diet to reduce BP in non-dippers could be due to the fact that the diet is least effective during the night and that non-dippers have an elevation in nocturnal BP that is not mediated by sympathetic nervous system activation.

Of note, we did not identify any benefit of low-fructose diet on renal function, although a trend was observed for an improvement in serum creatinine and creatinine clearance. It is possible that with longer follow-up or more subjects, a change might have been observed. Our subjects were also selected to have relatively stable renal function, so very little progression was seen in either group. We did note, however, some improvement in markers of inflammation (sICAM-1 and hsCRP) and in serum insulin levels. The significance of these findings remains to be established.

One interesting observation was that the benefits on serum uric acid, BP and inflammation tended to remain even after returning to the regular diet. In this regard, the intake of fructose is known to upregulate the expression of both the transporter for fructose in the gut (Glut5) as well as its key metabolizing enzyme, fructokinase, in the liver [25]. Subjects with non-alcoholic fatty liver disease have a history of high-fructose ingestion from added sugars and show elevated levels of fructokinase messenger RNA in their liver [26]. Furthermore, a pilot study showed that subjects on a high-fructose diet show an enhanced increase in serum uric acid in response to a set dose of fructose, whereas the same subjects on a low-fructose diet show the opposite [27]. Thus, one possible explanation is that the implementation of a low-fructose diet may have led to a down-regulation of Glut5 and fructokinase, resulting in a reduction in fructose absorption and metabolism that persisted despite resuming the regular diet. Another possibility is that characteristics in the resumed regular diet was different from the basal regular diet in respect to fructose content. While total fructose intake was not different, subjects resuming their ‘regular’ diet did reduce the relative percent of added sugars with a relative increase in natural fructose sources (Table 1). Some studies suggest that it is the fructose content from added sugars that confers the risk for hypertension as opposed to fructose from fruits that also contain a host of beneficial substances such as antioxidants [11]. Therefore, it remains possible that some of the benefits in the resumed regular diet may reflect a relative shift in fructose intake from added sugars to natural fruits.

This study has several limitations. Firstly, since a low-fructose diet has never been administered to subjects with CKD, the study was of necessity exploratory and did not include a power analysis. Secondly, for this initial study, there was no placebo group, and data were compared to baseline levels. However, we did ensure these subjects had stable renal function for the prior 6 months, and no alterations in medications were provided during the course of the study. A third problem relates to the persistent effect of the low-fructose diet after reinstitution of the regular diet. Although it could be due to the downregulation of fructokinase and Glut5 as proposed, it is important to have a control group that is followed for the same duration that has not received a low-fructose diet. Finally, it is also possible that some of the benefits may have related to the 1 kg reduction in weight observed, even though this was not significant. However, meta-analyses suggest that a reduction of weight of 1 kg will only lower systolic BP by 1 mmHg, so it seems unlikely that this is a significant factor [28]. Furthermore, there were no changes in total calorie (energy intake) during all three phases of the study.

Despite the limitations, the study provides some very interesting findings. Firstly, we demonstrate the feasibility of using a low-fructose diet in subjects with CKD, and we have found that it can reduce some markers of inflammation. While there were no benefits on renal function in this pilot study, we did observe a reduction in BP in subjects with a ‘dipper’ physiology. Finally, this pilot study will also provide data required to develop a prospective study with the appropriate power calculations. We therefore hope this study will be useful for others who are interested in dietary approaches to improve metabolic and physiological outcomes in subjects with CKD.