A Low-Sodium Diet Potentiates the Effects of Losartan in Type 2 Diabetes

We studied 20 patients with type 2 diabetes, hypertension, and albuminuria of 10–200 μg/min on an ambulatory basis. Participants were recruited from the Austin and Repatriation Medical Center diabetes clinic as well as the surrounding district. Inclusion criteria included a diagnosis of type 2 diabetes, age 30–75 years, seated systolic blood pressure >130 mmHg and/or seated diastolic blood pressure >85 mmHg, AER 10–200 μ g/min, HbA_1c <11.0% (normal 4.3–6.1%), an absence of serious systemic illness, an absence of history of substance abuse, and habitual 24-h urinary sodium excretion >100 mmol/24 h. Exclusion criteria included seated systolic blood pressure >165 mmHg or diastolic blood pressure >100 mmHg, serum potassium >5.5 mmol/l, plasma creatinine >200 μmol/l, long-term use of nonsteroidal anti-inflammatory drugs, a history of recurrent urinary tract infections (more than three per year), BMI >35 kg/m², cardiac failure, nitrate therapy, or intolerance of ACE inhibitors. Antihypertensive or diuretic therapy was stopped for at least 2 weeks before commencing the study. This allowed for a complete washout of prior antihypertensive agents.

The study was approved by the Human Research Ethics Committee at the Austin and Repatriation Medical Center, and all patients gave informed consent before commencement of the study.

The study protocol is outlined in Fig. 1. In this placebo-controlled dietary crossover study, patients were studied on regular- and low-sodium intakes, with each patient acting as his or her own control. The power of the study was based on the assumption that blood pressure would be estimated with a SD of 8 mmHg. To detect a difference between regular- and low-sodium diets of 1 SD, with a power of 80% and an α of 5%, paired data in 10 subjects were required (19). Patients were randomly assigned in a double-blind fashion to receive losartan 50 mg/day (n = 10) or a matching placebo (n = 10). This medication was taken daily for two 4-week phases with a 4-week washout period between phases. There was no crossover in medication assignment. Patients remained on their usual diets during a 2-week run-in period and were then assigned, in random order, to a 2-week dietary period of either restricted sodium intake (target 50–70 mmol/day) or continuation of regular sodium intake (>100 mmol/day). In the second phase, there was a crossover in dietary assignment. Low-sodium diets were conducted on an ambulatory outpatient basis. Patients received advice from a clinical nutritionist and subsequently brought and prepared their own food. They were provided with no-added-salt bread for the low-sodium period.

The terms losartan_RS and losartan_LS were used to refer to the 2-week period, from weeks 2 to 4, in which subjects in the losartan group were assigned to regular- and low-sodium diets, respectively. The terms placebo_RS and placebo_LS were used to refer to the 2-week period, from weeks 2 to 4, in which subjects in the placebo group were assigned to regular- and low-sodium diets, respectively. After the washout period, patients entered the study protocol if sitting systolic and/or diastolic blood pressures (mean of three readings) were >130 and/or >85 mmHg, respectively.

Parameters measured at weeks 2 (after medication run-in) and 4 (after the 2-week dietary period) included 24-h ambulatory blood pressure (ABP), GFR, and effective renal plasma flow (ERPF). Parameters measured at weeks 0, 2, and 4 included body weight, albumin-to-creatinine ratio (ACR) on 24-h urine collection, plasma glucose, electrolytes, plasma renin activity (PRA), ANG-II, and aldosterone. Urinary electrolytes, urea, and creatinine were determined at baseline and weekly during each phase. All biochemical analyses were performed in the morning after patients fasted overnight and before they took the study medication. Measurement of GFR and ERPF was begun 1 h after medication was administered. Measurement of 24-h ABP at week 0 was obtained in a subset of 12 patients.

The 24-h ABP was measured with a portable recording system (Spacelabs 90207; Spacelabs Medical Products, Deerfield, WI) based on an oscillometric method. The 24-h systolic, diastolic, and mean arterial pressures as well as wake and sleep values were recorded. Blood pressure was measured every 30 min from 7:00 a.m. to 11:00 p.m. and every hour from 11:00 p.m. to 7:00 a.m.

Radioimmunoassay for albumin was performed by a double-antibody method with intra- and interassay coefficients of variation of 1.8 and 4.8%, respectively, for a sample concentration of 27 mg/l. The determination of creatinine in plasma and urine was performed by the modified Jaffé method (kinetic colorimetric assay). Plasma creatinine and electrolytes were measured on a Hitachi 747 Auto Analyzer (Roche Diagnostics, Mannheim, Germany), and urinary creatinine and electrolytes were measured on a Hitachi 911 automatic analyzer (Roche Diagnostics). HbA_1c was determined by boron affinity chromatography on a Primus CLC330 analyzer (Kansas City, MO).

Specimens for PRA and ANG-II were collected in EDTA tubes on ice, centrifuged within 1 h, and stored at −20°C for analysis at a later date. Plasma ANG-II was measured by direct radioimmunoassay. PRA was determined by measuring the rate of generation of ANG-I by radioimmunoassay after incubating plasma at 37° for 1 h. Plasma aldosterone was measured by direct radioimmunoassay (Coat-A Count Aldosterone; Diagnostic Products, Los Angeles, CA).

GFR was measured by the plasma clearance of nonradioactive iohexol after a single bolus intravenous injection. Plasma concentrations of iohexol were measured by capillary electrophoresis (20), and GFR calculations were performed using the Brochner-Mortensen corrected one-compartment model, as previously described (21).

ERPF was measured using ^99mTc-MAG-3. A single dose of 10 MBq of ^99mTc-MAG-3 was injected as an intravenous bolus, and 11 blood samples were taken over the ensuing 90 min (22). The plasma clearance was fitted to a bi-exponential model using an iterative nonlinear regression curve-fitting program (Sigmaplot; Scientific Graphing Software, Jandel Scientific, CA). ERPF is reported as MAG-3 clearance values. All GFR and ERPF measurements were corrected for body surface area and are expressed as milliliters per minute per 1.73 meters squared.

The filtration fraction (FF)—the filtered proportion of the renal blood flow—was calculated by the equation FF = GFR × 100/ERPF. To calculate the FF, the MAG-3 ERPF values were converted to equivalent para-aminohippurate (PAH) ERPF values using the following formula: clearance of MAG-3 = 0.53 × clearance of PAH (23).

Glomerular capillary pressure (P_GC) was estimated indirectly from the pressure-natriuresis relationship by the method of Kimura et al. (24), based on mean arterial blood pressure, total plasma protein, and the FF.

Completeness of urine collections was verified from measurements of urinary creatinine. For each patient, data from 24-h urine collections were accepted if creatinine excretion fell within 2 SDs of the average creatinine excretion for that patient during the entire study period. On the 6 of 198 occasions when creatinine excretion fell outside this range, data for sodium, potassium, and urea were corrected for the mean creatinine excretion in that particular patient.

Data are presented as means ± SEM with 95% confidence intervals (CIs). Because some parameters, including PRA, ANG-II, and ACR were positively skewed, data were analyzed after logarithmic transformation and are shown as geometric mean multiplied/divided by the tolerance factor. Data measured at multiple time points were analyzed by a single-factor ANOVA with repeated measures, followed by Fisher’s least significant difference test for multiple comparisons. Differences between two groups were analyzed using either Student’s paired or unpaired t test or the χ² analysis for proportions, where appropriate. These analyses were performed using Statview V (Brainpower, Calabasas, CA). The effect of a low-sodium diet on a specific parameter (mean difference and 95% CI) was calculated as the difference between losartan_RS and losartan_LS or placebo_RS and placebo_LS for the losartan and placebo groups, respectively. No order effect was found. All analyses were performed according to the intention-to-treat approach.

There were no significant baseline differences in mean arterial blood pressure, urinary sodium excretion, AER, BMI, duration of diabetes, HbA_1c, or pharmacotherapy for diabetes between the losartan and placebo groups (Table 1).

A similar degree of sodium restriction, as measured by 24-h urinary sodium excretion, was achieved in the losartan_LS (85 ± 14 mmol/ day) and placebo_LS (80 ± 22 mmol/day; NS) study groups (Table 2). A change in weight (Δ weight) was observed during both the losartan_LS (Δ weight: −1.9 ± 0.5 kg) and placebo_LS (Δ weight: −1.0 ± 0.4 kg) phases, which was statistically significant when compared to the Δ weight recorded during the losartan_RS (Δ weight: −0.1 ± 0.2 kg; P = 0.006) and placebo_RS (Δ weight: 0.0 ± 0.2 kg; P = 0.05) phases.

Measurements of parameters of the RAS over the study period are shown in Table 2. An increase in both PRA and plasma ANG-II levels was observed during the 2-week losartan run-in phase. During the losartan_RS phase, there was no additional change in the plasma ANG-II level or PRA (week 4 vs. week 2: NS), but both indexes remained elevated when compared with baseline (week 4 vs. week 0: P < 0.01). During the losartan_LS phase, there was a highly significant further increase in both the PRA (week 4 vs. week 2: P < 0.01) and plasma ANG-II level (week 4 vs. week 2: P < 0.01). No significant changes in either PRA or plasma ANG-II levels were observed during the placebo_RS phase. However, during the placebo_LS phase, there was a significant increase in both PRA (week 4 vs. week 2: P < 0.01) and plasma ANG-II levels (week 4 vs. week 2: P < 0.05). The absolute increase in mean PRA, but not in plasma ANG-II levels, was greater in the losartan_LS than in the placebo_LS group (PRA: 3.07 ± 1.63 vs. 0.92 ± 1.3 ng · ml⁻¹ · h⁻¹, respectively, P = 0.048; plasma ANG-II: 24.71 ± 1.53 vs. 10.8 ± 1.16 pg/ml, respectively, P = 0.09). A significant increase in plasma aldosterone was observed during both the losartan_LS (week 4 vs. week 2: P < 0.01) and placebo_LS (week 4 vs. week 2: P < 0.01) phases. No significant change in plasma aldosterone was observed during the losartan 2-week run-in phase or during the losartan_RS or placebo_RS phase.

ABP fell significantly during the losartan_LS phase, but remained unchanged during the placebo_LS phase (Tables 3 and 4). The change in blood pressure was greater in the losartan_LS compared with the losartan_RS phase: the mean difference between the losartan_LS and losartan_RS phases for 24-h systolic, diastolic, and mean arterial blood pressures was −9.7 mmHg (CI −17.2 to −2.2), −5.5 mmHg (−8.4 to −2.6), and −7.3 mmHg (−11.3 to −3.3), respectively.

When 24-h ABP was analyzed by wake/sleep periods, there were significant decreases in blood pressure during the losartan_LS phase in wake systolic, diastolic, and mean arterial pressure and in sleep systolic blood pressure.

Changes in ABP from baseline were assessed in a subset of 12 patients. Figure 2 shows that the antihypertensive effect of losartan was doubled by the addition of a low-sodium diet. In the losartan group (n = 6), after the 2-week run-in period with losartan therapy on a regular-sodium diet, the change in ABP was −5.2 ± 3.6 mmHg (NS) from baseline. After 2 weeks of low-sodium diet, from week 2 to week 4, the change from baseline was −10.7 ± 3.7 mmHg (weeks 0–2 vs. weeks 0–4: P = 0.02). In the placebo group (n = 6), no significant changes in 24-h ABP from baseline were observed at week 2 or after 2 weeks of a low-sodium diet at week 4 (−0.3 ± 1.2 and −0.5 ± 2.8 mmHg, respectively; NS).

In the losartan group, there was a significant reduction in the ACR in the low-sodium phase that was not observed in the regular-sodium phase: losartan_LS −29% (CI −50 to −8.5%) vs. losartan_RS +14% (−19.4 to 47.9%) (P = 0.02). In the placebo group, there was no significant change in ACR in either the low- or the regular-sodium phase: placebo_LS + 25% (CI −39.3 to 89.3%) vs. placebo_RS − 13.5% (−41.1 to 14.0%) (P = 0.2).

In the losartan group, ACR did not decrease significantly from baseline until a low-sodium diet was added (Fig. 3).

In the losartan group, when the mean percent difference in ACR between low- and regular-sodium diets (losartan_LS - losartan_RS) was compared with the same parameter in the placebo group (placebo_LS - placebo_RS), a significantly greater albuminuria-lowering effect of the low-sodium diet was found in the losartan group: losartan −43.5% (CI −77.5 to −9.6%) vs. placebo + 38.5% (−31.7 to 108.8%) (P = 0.03).

No significant changes in GFR, ERPF, or FF were observed during losartan_LS or placebo_LS phases (Tables 3 and 4). A significant change in calculated P_GC was found in the losartan_LS phase (−9.0 ± 1.7 mmHg) compared with the losartan_RS phase (−1.4 ± 1.7 mmHg) (P = 0.002). No changes in P_GC were found in the placebo group.

During the period of dietary sodium restriction, there were no significant changes in plasma concentrations of sodium, urea, or creatinine, and also no significant changes in the urinary excretion of potassium, urea, and creatinine (data not shown).

At the beginning of each phase, HbA_1c was measured. No differences in glycemic control between phases was found for losartan (7.7 ± 0.5 vs. 7.5 ± 0.3% for losartan_RS vs. losartan_LS, respectively; NS) or placebo (7.3 ± 0.3 vs. 7.3 ± 0.3% for placebo_RS vs. placebo_LS, respectively; NS). In the losartan_LS group, there was a small but significant decrease in fasting blood glucose, of dubious clinical significance, at week 4 (Table 2).

In both groups, significant correlations were observed between the percent reduction in 24-h urinary sodium excretion and the fall in mean arterial blood pressure (losartan_LS: y = 0.22x + 5.55, r = 0.68, P = 0.03; placebo_LS: y = 0.11x + 4.95, r = 0.64, P = 0.05) (Figs. 4A and B). The 95% CI (−0.11 to 0.33) for the difference between the gradients (0.11) showed there was no significant difference between the regression lines.

In the losartan group, a strong and significant correlation was found between the fall in mean arterial blood pressure and the percent reduction in ACR (y = 5.96x + 0.00, r = 0.70, P = 0.02) (Fig. 5). No significant correlations between the fall in P_GC and percent reduction in ACR or between changes in mean arterial pressure and ERPF were detected.

This study demonstrated the important role that dietary sodium plays in modulating the antihypertensive and antiproteinuric effects of ANG-receptor antagonists in type 2 diabetes. In patients taking losartan, the magnitude of blood pressure reduction that occurred after 2 weeks of low-sodium diet was equivalent to the effects of adding a second antihypertensive agent (25) and led to an approximate doubling of the antihypertensive effect of the drug.

Unlike many studies that have examined the effects of a low-sodium diet, the current study was performed on an ambulatory basis and without pre-prepared diets. Patient dietary education focused on identifying the sodium content of common foods and determining sodium content by reading food labels. This approach was able to achieve significant reductions in mean urinary sodium excretion, to 80–85 mmol/24 h, and was associated with activation of the systemic RAS. Effects of sodium restriction in hypertensive and nonhypertensive subjects to <100 mmol/day have been recently studied by the Dietary Approaches to Stop Hypertension investigators, who found greater blood pressure reductions during sodium restriction from 100 to 60 mmol/day compared with a reduction from 150 to 100 mmol/day (26). The efficacy of dietary sodium reduction on lowering blood pressure in diabetic subjects has not been extensively characterized. Similar to the findings in the present study, one previous randomized study in type 2 diabetes showed significant blood pressure-lowering effects of sodium restriction; however, unlike in the present study, subjects in that study had more severe hypertension (>160/90 mmHg) and some were on concomitant antihypertensive therapy (27). It remains to be determined whether increased exchangeable body sodium and sodium retention alters the magnitude and temporal nature of the responsiveness to dietary sodium restriction. In the present study, the degree of reduction in urinary sodium excretion was correlated to a reduction in mean arterial blood pressure in both the losartan and placebo groups (Figs. 4A and B), with a nonsignificant trend for greater effects of sodium restriction in the losartan group. The effects of losartan and low-sodium diet on plasma ANG-II, aldosterone, and PRA have been well characterized, and the present findings are consistent with those of previous studies (28–30).

Like previous studies, this study demonstrated that the antiproteinuric effects of losartan in diabetic patients is closely associated with reductions in blood pressure (31–33). However, in the present study, a significant decrease in both blood pressure and albuminuria was observed only when a low-sodium diet was added to losartan therapy.

The mechanism by which the addition of a low-sodium diet reduced albuminuria in the losartan group appears to be related to blood pressure reduction, with the fall in mean arterial blood pressure correlating with the percent reduction in ACR (Fig. 5). A significant correlation between decreases in albuminuria and blood pressure has been demonstrated by a meta-analysis for both ACE inhibitors and conventional antihypertensive therapy (34). To further determine whether changes in various renal parameters, including glomerular hemodynamics, may be involved, measurements of GFR, ERPF, FF, and calculated P_GC were performed. No significant changes in GFR, ERPF, or FF were observed with low-sodium diet in either group. A fall in calculated P_GC, which occurred with a low-sodium diet in the losartan group, was linked to a decrease in albuminuria. The elegant micropuncture studies performed by Zatz et al. (35) in diabetic rodents have previously suggested a pivotal role for raised intraglomerular pressure in mediating albuminuria. In those studies, the increase in intraglomerular pressure was reduced by blockade of the RAS with an ACE inhibitor, and this was associated with attenuation of albuminuria. Although P_GC was only calculated and is therefore an indirect measurement of intraglomerular pressure, the findings in the present study are consistent with the hypothesis that a reduction in P_GC is closely linked to a reduction in albuminuria.

Other potential confounding factors that could influence GFR or albuminuria, such as changes in glycemic control (36) or protein intake (37), were also evaluated. No change in dietary protein intake, as assessed by urinary urea excretion, was observed during the period of low-sodium diet, nor were there clinically significant changes in overall glycemic control, as assessed by fasting blood glucose and HbA_1c.

Because ANG-II has both hemodynamic and trophic effects, blockade of its receptors may potentially exert effects on albuminuria reduction via nonhemodynamic mechanisms. In a meta-analysis, ACE inhibitors were found to exert specific antiproteinuric effects, with minimal changes in blood pressure (34). In this study, however, no reduction in albuminuria was observed after 4 weeks of losartan therapy while patients remained on a regular-sodium diet, during which urinary sodium excretion was >200 mmol/day. This finding is consistent with the previous observation in streptozotocin-induced diabetic rats that a high-salt diet blocks the antihypertensive and antiproteinuric effects of ACE inhibitors (10).

This study does not support the concept of sodium modulation of proteinuria, independent of blood pressure reduction, that has been previously described in type 2 diabetic patients receiving verapamil (38).

The observation that renal plasma flow did not change between the regular- and low-sodium diets in the placebo group is consistent with a blunted vasodilator renal plasma flow response to a high-sodium diet, which has been previously described in patients with type 2 diabetes (17) and essential hypertension (30,39). In a previous study of the effects of low and high dietary sodium on mean arterial blood pressure and renal hemodynamics in essential hypertension, a rise in blood pressure on a high-sodium diet was associated with a blunted increase in ERPF (30). In our study, no correlation between the changes in mean arterial blood pressure and ERPF was found.

This study demonstrated that a low-sodium diet optimizes the renoprotective effects of the ANG-receptor blocker, losartan. It also showed that a low-sodium diet is achievable on an ambulatory basis in the short term. Combination antihypertensive medication in a single tablet, consisting of a thiazide diuretic and an ACE inhibitor or ANG-receptor antagonist, has recently become widely available. Moderate sodium restriction, as achieved in the present study, has been shown to be as effective as a thiazide diuretic in lowering blood pressure in the presence of an ACE inhibitor in essential hypertension (40). However, a low-sodium diet is a preferred option because, unlike diuretic therapy, it is not associated with potential adverse effects on lipid and glucose metabolism, nor is it associated with potential disturbances of serum potassium and sodium levels (41). We propose that a low-sodium diet (<100 mmol/day) be used in subjects with type 2 diabetes who are receiving monotherapy with an ANG-receptor antagonist when further blood pressure reduction is required. In these circumstances, the addition of a low-sodium diet should be considered as an appropriate alternative to additional pharmacological antihypertensive agents, including combination therapy with a diuretic.

This work was supported by a Merck medical school grant and an Apex Diabetes Australia Research grant. During the period of this work, C.A.H. was supported by grants from the Austin Hospital Medical Research Foundation and the National Health and Medical Research Council. We thank Judy Winikoff and Aysel Akdeniz for their technical assistance and Dr. Con Tsalamandris for statistical assistance.

Data from this paper have been published in abstract form in Circulation 102 SII:869, 2000.