点击显示 收起
1 From the Division of Nephrology and Dialysis, Department of Medicine III (AV, PK, and GS-P), the Institute of Medical and Chemical Laboratory Diagnostics (MF, SS, and CR), and the Departments of Clinical Pharmacology (JP and MW) and Emergency Medicine (MM), University of Vienna.
2 Address reprint requests to A Vychytil, Division of Nephrology and Dialysis, Department of Medicine III, University Hospital Vienna, Währinger Gürtel 18-20, A-1090 Vienna, Austria. E-mail: andreas.vychytil{at}univie.ac.at.
ABSTRACT
Background: Oral ingestion of proteins or amino acids is associated with endothelial dysfunction. The effect of commercial amino
acid peritoneal dialysis solutions on vascular function is unknown.
Objective: We compared the acute effect of intraperitoneal amino acid administration with that of intraperitoneal glucose administration on vascular function in peritoneal dialysis patients.
Design: In an open-label randomized, controlled, crossover and observer-blinded trial, we examined the acute effect of an intraperitoneal application of 2 L commercial 1.1% amino acid solution compared with that of a 2.27% glucose solution in 13 peritoneal dialysis patients. The primary endpoint was the change in forearm reactive hyperemia 6 h after instillation of either dialysis solution.
Results: After 6 h of dwell time, reactive hyperemia was substantially impaired after administration of the amino acid solution compared with the glucose solution (median difference: 202%; 95% CI: 57%, 368%; P = 0.007). In a comparison of differences between values at 6 h and those before treatment, reactive hyperemia significantly decreased during the dwell with the amino acid dialysis solution compared with that with the glucose dialysis solution (median difference: 242%; 95% CI: 53%, -457%; P = 0.013). In an analysis of smoking and nonsmoking patients separately, the difference in forearm blood flow between the 2 treatments was still statistically significant.
Conclusions: One 6-h dwell with a commercial amino acid dialysis solution acutely impairs forearm reactive hyperemia in smoking and nonsmoking peritoneal dialysis patients. Because endothelial dysfunction is associated with increased morbidity and mortality, the long-term use of these solutions may increase the risk of cardiovascular disease.
Key Words: Endothelial function • reactive hyperemia • peritoneal dialysis • amino acids • homocysteine • plethysmography
INTRODUCTION
Peritoneal dialysis is usually performed with glucose-containing dialysis solutions. However, amino acids can serve as
an osmotic agent, instead of glucose. In clinical studies, anabolic effects have been shown in malnourished patients after a
once or twice daily use of an amino acid-containing peritoneal
dialysis solution (1). Commercially available dialysis solutions
contain several amino acids, including methionine. This is of
particular interest because oral ingestion of protein or methionine impairs endothelial function by several mechanisms, including an increase in plasma homocysteine (Hcy), a decrease
in serum folate, or an increase in asymmetric dimethyl arginine
(ADMA) (2-6). Endothelial dysfunction contributes to cardiovascular complications, which are the main causes of death in
dialysis patients (7). We therefore compared the acute effect of
a commercial amino acid dialysis solution with that of a glucose dialysis solution on forearm reactive hyperemia in patients
with end-stage renal disease treated with peritoneal dialysis.
SUBJECTS AND METHODS
Subjects
This open-label, randomized, controlled, crossover, and observer-blinded trial was performed at the University Hospital of
Vienna. Clinically stable patients (older than 18 y and treated
with peritoneal dialysis for =" BORDER="0"> 3 mo) without peritonitis within
2 mo before the study began were included. None of the
patients received an amino acid peritoneal dialysis solution,
folic acid, or vitamin B or C supplements in the period between
the start of peritoneal dialysis and study inclusion. Patients
with serum hemoglobin concentrations < 9 g/dL or with acute
illnesses were excluded. The Ethical Review Board at the
University of Vienna approved the study. All patients provided
written informed consent.
Objectives
We hypothesized that intraperitoneal amino acid administration acutely impairs vascular function in end-stage renal disease. We examined the effect of a 6-h dwell with 2 L of a
commercial 1.1% amino acid dialysis solution (Nutrineal; Baxter Healthcare Corp, Norfolk, United Kingdom) compared with
2 L of a conventional 2.27% glucose dialysis solution (Dianeal;
Baxter Healthcare Corp) on vascular function in peritoneal
dialysis patients.
The composition of the amino acid solution was as follows: 132 mmol Na/L, 105 mmol Cl/L, 1.25 mmol Ca/L, 0.25 mmol Mg/L, 40 mmol lactate/L, 0.510 g serine/L, 0.850 g isoleucine/L, 1.020 g leucine/L, 0.955 g lysine hydrochloride/L, 0.850 g methionine/L, 0.570 g phenylalanine/L, 0.646 g threonine/L, 0.270 g tryptophan/L, 1.393 g valine/L, 0.951 g alanine/L, 1.071 g arginine/L, 0.510 g glycine/L, 0.714 g histidine/L, 0.595 g proline/L, and 0.300 g tyrosine/L. The 2.27% glucose solution contained 132 mmol Na/L, 95 mmol Cl/L, 1.25 mmol Ca/L, 0.25 mmol Mg/L, 40 mmol lactate/L, and 22.7 g glucose/L.
Outcomes
The primary endpoint was endothelium-dependent forearm
reactive hyperemia after 6 h of an intraperitoneal instillation of
amino acid dialysis solution (treatment A) compared with an
intraperitoneal instillation of a conventional glucose solution
(treatment G). After both 6-h dwells, each patient continued
peritoneal dialysis with his or her usual regimen. Forearm
reactive hyperemia was measured before dialysate instillation,
6 and 24 h after instillation of peritoneal dialysate. Endothelium-independent vasodilatation was documented at the same
time points. The course of total Hcy (tHcy) concentrations in
plasma and dialysate was also assessed. The interval between
the treatments was 4 wk. Multiple measurements were made by
a trained assessor to ensure the quality of measurements. Biochemical analyses were performed in an ISO 9001 certified
laboratory. The correct entry of the data in the case report
forms and in the electronic database was ensured by double-checking with the use of study monitors. Treatment with antihypertensives or other vasoactive agents was not changed
during the study period.
Sample size
To determine the sample size calculation, a pilot study was
conducted with 5 patients. We observed a decrease in reactive
forearm hyperemia from 600% to 400% after a 6-h dwell
with a 1.1% amino acid dialysis solution. Assuming a type I
error of 0.05, a power of 0.80, and a dropout of 1-2 patients,
the sample size was estimated to be 15 patients.
Randomization
The random allocation sequence was generated by a clinical
epidemiologist (MM) using the software package STATA,
release 6 (Stata Corp, College Station, TX). The sequence of
the studies (treatment A then treatment G or treatment G then
treatment A) for each individual patient was enclosed in a
separate opaque envelope that was opened immediately before
the study began. Patients were enrolled by one of the investigators. Treatments were administered by study nurses.
Neither the investigator who performed the measurements of vascular function nor the laboratory personnel who analyzed the blood and dialysate samples or the biostatistician were aware of the treatment code. Patients were asked not to inform the investigator who performed the measurements of vascular function about the sequence of the studies.
Measurements
Forearm vascular function
Venous occlusion plethysmography was used to examine
reactive hyperemia in the forearm as a measure of systemic
vascular responsiveness (8, 9). Patients were asked to fast for
10 h before both treatments. After the first plethysmography,
the intraperitoneal dialysate was instilled. A blood sample was
taken simultaneously. Patients received a standardized breakfast and lunch. Outflow of the test solution and the second
plethysmography were performed at 6 h. Smokers were allowed to smoke during the dwell time, but they were asked to
maintain their smoking pattern during both treatments. After
24 h, a third plethysmography was performed.
In detail, venous occlusion plethysmography was performed as follows. During the measurement, each subject was in a supine position in a quiet, air-conditioned, temperature-controlled room at 23 °C. Both forearms were positioned on cushions above the level of the heart. A mercury-in-silastic strain-gauge venous plethysmograph (EC 6; DE Hokanson Inc, Bellevue, WA) was used to measure forearm blood flow (8). After a resting period of 10 min, strain gauges were placed over the widest part of the left forearm 7 cm below the antecubital crease. A small cuff over the wrist to occlude flow to the hands was inflated at suprasystolic pressures 1 min before and continuously throughout the measurement of forearm blood flow. Baseline forearm blood flow was estimated from the rate of increase in forearm volume after venous occlusion of the forearm by inflation of another cuff to 50 mm Hg with a rapid cuff inflator on the upper arm. A 6-s flow recording was repeated every 15 s through 2.5 min. Baseline forearm blood flow was calculated as the mean of these 10 records. Systemic blood pressure was measured noninvasively with an oscillometric measurement device (HP-CMS patient monitor; Hewlett-Packard, Palo Alto, CA) from the right wrist at the end of each set of recordings. Pulse rate was automatically recorded from finger-pulse oxymetry. After the measurement of baseline forearm blood flow, peak reactive hyperemia was calculated. Forearm ischemia was induced by inflation of the cuff on the upper arm to a suprasystolic pressure (250 mm Hg). After 4.5 min of arterial occlusion, the cuff was released, and forearm blood flow was recorded immediately after complete cuff deflation and was repeated every 15 s through 2.5 min. Peak reactive hyperemia was usually achieved between 30 and 60 s, and blood flow returned to baseline after 2.5 min. The percentage increase in blood flow between baseline and peak reactive hyperemia (measured after release of forearm ischemia) was calculated before treatment, 6 and 24 h after dialysate instillation. These values were used for further statistical analyses.
At the same time points, measurement of endothelium-independent vasodilatation in the forearm was performed. These measurements were made 15 min after evaluation of reactive hyperemia with the use of sublingual glyceryl trinitrate (nitroglycerin) as a pharmacologic stimulus (10). Baseline forearm blood flow was recorded for 9 s every 30 s as described above. After 5 recordings (ie, 2.5 min) were made, 0.8 mg glyceryl trinitrate was administered sublingually (Nitrolingual Spray; Pohl-Boskamp, Hohenlockstedt, Germany). Measurements of forearm blood flow were repeated for 10 min. Again, changes in blood flow were expressed as a percentage increase between baseline values and peak values (after administration of glyceryl trinitrate) before treatment and 6 and 24 h after dialysate instillation. After an interval of 4 wk, the exact same procedure was repeated for the other dialysis solution.
Biochemical methods
Complete blood counts as well as blood and dialysate chemistry analyses were performed by using standard procedures.
Concentrations of tHcy were determined by a fluorescence
polarization immunoassay (IMx analyzer; Abbott Laboratories,
Abbott Park, IL). Plasma and dialysate samples from each
patient were batch analyzed. Weekly total Kt/Vurea [where K is
the total urea clearance (urine + dialysate), t is time, and V is
total body water]; characterization of peritoneal transport type;
plasma folate, vitamin B-6, and vitamin B-12 concentrations;
and the presence of 677CT and of 1298AC in the 5,10-methylenetetrahydrofolate reductase (MTHFR) gene were determined as previously reported (11, 12).
Statistical methods
Continuous data are presented as medians within the interquartile range (IQR; from the 25th to the 75th percentile).
Frequencies are given as counts, but we did not calculate
percentages because of the small sample size. We compared
the crude values between treatments A and G after 6 h. We also
calculated differences between the value measured after 6 h
and the value measured before treatment. Because of the crossover design, each patient served as his or her control, and we
used Wilcoxon's signed-rank test to compare groups. All analyses were intention-to-treat. To allow inference from the effect
size, we also calculated the median difference between both
treatments and the corresponding 95% CI of the difference.
Period, interaction, and carryover effects for hyperemia and for
tHcy were assessed with Wilcoxon's rank-sum test. The analysis of the effect of the intervention was also performed by
using a repeated measurement analysis of variance (ANOVA)
that included treatment, smoking status, and the interaction
terms treatment, smoking status, and time. This allowed a
simultaneous assessment of these variables. All calculations
were performed by using the statistical software package
STATA, release 7 (Stata Corp).
RESULTS
Patients
The flow of participants is shown in Figure 1. Recruitment
lasted 4 mo. The demographic and clinical characteristics are
presented in Table 1. Thirteen patients were included in the
final analysis (Figure 1). The biochemical data of these patients
are indicated in Table 2. Treatment-related adverse events
included mild abdominal pain in 2 patients during the dwell
with amino acid dialysis solution and in one patient during the
dwell with the glucose dialysis solution. No adverse events
occurred during measurement of vascular function or after
glyceryl trinitrate application. (Glyceryl trinitrate was not
given to 2 patients during treatments A and G because of a
systolic blood pressure < 100 mm Hg.)
FIGURE 1.. Flow of study participants and trial profile. Treatment A
was the instillation of a 1.1% amino acid peritoneal dialysis solution;
treatment G was the infusion of a 2.27% glucose peritoneal dialysis
solution.
View this table:
TABLE 1.. Characteristics of 15 randomly assigned peritoneal dialysis patients1
View this table:
TABLE 2.. Vitamin status and creatinine and albumin concentrations in peritoneal
dialysis patients who received both treatments A and G1
Effect of the intervention on forearm reactive hyperemia
At 6 h, reactive hyperemia (the primary endpoint) was
significantly lower in the treatment A group than in the treatment G group (median difference between treatments A and G:
202%; 95% CI: 57%, 368%; P = 0.007; Table 3). It appears
that reactive hyperemia was not fully restored 24 h after instillation of the amino acid dialysis solution. The difference from
values before treatment, however, was not statistically significant (Table 3). In a comparison of differences in values at 6 h
and those before dialysate instillation between treatment
groups, hyperemia also decreased significantly in the treatment
A group compared with the treatment G group (median difference: 242%; 95% CI: 53%, 457%; P = 0.013). There was no
evidence of a period, carryover, or interaction effect (see
below).
View this table:
TABLE 3.. Effect of peritoneal dialysis solutions containing either glucose or amino acids on forearm blood flow responsiveness1
Forearm blood flow in smokers and nonsmokers separately
is shown in Table 4. In the repeated measurement ANOVA,
treatment was associated with endothelium-dependent forearm
reactive hyperemia (P = 0.002). Being a smoker was also
associated with endothelium-dependent forearm reactive hyperemia (P = 0.049), but no interaction between smoking and
treatment (P = 0.13) or time (P = 0.19) was observed. Treatment did not interact with time (P = 0.12), nor was the 3-way
interaction between time, treatment, and smoking status significant (P = 0.30).
View this table:
TABLE 4.. Effect of peritoneal dialysis solutions containing either glucose or amino
acids on forearm blood flow responsiveness in smoking and nonsmoking
patients1
Effect of glyceryl trinitrate on forearm blood flow
Neither during the dwell with the 1.1% amino acid dialysis
solution nor during treatment with the 2.27% glucose solution
was there a change in forearm blood flow responsiveness to
glyceryl trinitrate (Table 3).
Blood pressure and ultrafiltration
Blood pressure and pulse rate before and 6 h after instillation
of the amino acid or glucose dialysis solution as well as the
outflow volumes are shown in Table 5. There was no significant difference in systolic blood pressure, diastolic blood
pressure, or pulse rate between treatments G and A. Dialysate
outflow volume was slightly but significantly greater after
treatment G than after treatment A.
View this table:
TABLE 5.. Blood pressure, pulse rate, and outflow volume of peritoneal dialysis
patients who received both treatments A and G1
Effect of the intervention on tHcy concentrations
At 6 h after the dialysate instillation, plasma tHcy concentrations did not differ significantly between the 2 treatment
groups (P = 0.73; Table 6). There was also no significant
effect after adjustment for values before treatment when a
comparison of the differences in plasma tHcy concentrations
was made between 6 h and those before dialysate instillation.
In a comparison of the effect of the intervention on plasma
tHcy concentrations between the treatment periods, there was
evidence of a period effect. In patients who received treatment
G and thereafter treatment A, the median difference in plasma
tHcy concentration was 4.7 µmol/L (IQR: 0-6.8), whereas in
patients who received their treatments in the reverse order, the
median difference in plasma tHcy concentration was -3.6
µmol/L (IQR: -10 to -2.4). No evidence of a carryover or
interaction effect was observed. As shown in Table 6, there
tended to be a more pronounced increase in dialysate tHcy
concentrations in the treatment A group than in the treatment G
group.
View this table:
TABLE 6.. Effect of peritoneal dialysis solutions containing either glucose or amino acids on plasma and dialysate total homocysteine concentrations1
In the repeated measurement ANOVA, treatment was also
not associated with plasma Hcy concentrations (P = 0.88).
Being a smoker was associated with plasma Hcy concentrations (P = 0.001), but smoking did not interact with treatment
(P = 0.45) or time (P = 0.24). Treatment did not interact with
time (P = 0.84), nor was the three-way interaction between
time, treatment, and smoking status significant (P = 0.23).
DISCUSSION
We showed that a commercial amino acid peritoneal dialysis
solution results in the acute impairment of reactive forearm
blood flow responsiveness compared with a conventional glucose solution. There was no significant change in plasma tHcy
concentrations during either treatment.
Previous studies have shown that oral methionine loading is associated with an impairment of endothelial function in healthy subjects (2-6). In most of these studies, a methionine dose of 100 mg/kg body wt had been used, which is more than the 1.7 g methionine contained in the amino acid dialysis solution used in the present study. However, oral methionine doses as low as 10 mg/kg body wt have been shown to impair endothelial function (2). Furthermore, the bioavailability of amino acids may be higher after intraperitoneal administration than after oral ingestion.
In the present study the increase in flow-dependent vasodilatation after venous occlusion was 600% above baseline before dialysate instillation. During the 6-h dwell with the glucose solution, vascular responsiveness remained stable. By way of contrast, after a 6-h dwell with an amino acid dialysis solution, flow-dependent vasodilatation was substantially impaired compared with that measured before dialysate instillation.
One possible cause of vascular dysfunction resulting from methionine-containing amino acid dialysis solutions is an increase in plasma concentration of tHcy, which may be directly cytotoxic for endothelial cells (13) or increases oxidative stress (4, 14). In this context, an increase in fasting plasma tHcy concentrations during therapy with the 1.1% amino acid peritoneal dialysis solution has been reported (15). In our study there was a 4-µmol/L increase in the median plasma tHcy concentration 6 h after intraperitoneal instillation of the amino acid dialysis solution compared with a slight decrease in tHcy after the glucose dwell. This difference was probably not significant because 2 patients had severe hyperhomocysteinemia (> 100 µmol/L) because of the high prevalence of MTHFR 677CT/1298AC mutated alleles, which resulted in a huge variability in tHcy concentrations during amino acid treatment (data not shown). The increase in plasma tHcy concentrations during amino acid treatment was also reflected by the more pronounced increase in dialysate tHcy concentrations compared with the glucose treatment. The moderate increase in plasma tHcy in our study did not necessarily preclude an influence of Hcy on endothelial function. Small increases in tHcy of 1.6 µmol/L 2 h after a meal of 500 g lean chicken were associated with endothelial dysfunction (2). However, not all studies showed an association between plasma tHcy concentrations after methionine loading and endothelial dysfunction in healthy subjects (3, 5). In peritoneal dialysis patients there was no significant relation between endothelium-dependent vasodilatation and tHcy concentrations (16). No influence of folic acid treatment on endothelial function has been found in patients with chronic renal failure or in dialysis patients, although this therapy significantly reduced tHcy concentrations (17-19). A possible explanation for these findings is that metabolites other than Hcy could impair vascular function in patients with renal failure. Methionine loading is associated with an acute decrease in serum folate concentrations (20). Therefore, an acute folate-lowering effect of methionine loading rather than an increase in tHcy per se may cause endothelial dysfunction. A nonspecific effect of the amino acid dialysis solution on endothelial function in our study was unlikely because brachial artery flow did not change after oral ingestion of a methionine-free amino acid mix (2).
Another possible explanation for amino acid-related changes in vascular function was an increase in ADMA, which is an important endogenous inhibitor of nitrogen oxide synthase. In animals and in humans, the relation between the degree of endothelial dysfunction and plasma concentration of ADMA after methionine loading was found to be stronger than that between endothelial function and plasma tHcy concentration (5, 21).
Changes in plasma volume may influence vascular function. Dialysate outflow volume was slightly higher after treatment with the glucose dialysis solution than after treatment with the amino acid dialysis solution. The difference, however, was < 300 mL. One would expect that higher ultrafiltration during treatment with the glucose dialysis solution results in a more marked decrease in hyperemia, which was not the case in our study. Therefore, the peritoneal ultrafiltration during treatment with glucose dialysis solution most likely did not explain the observed differences in vascular function between the 2 treatments.
Long-term smoking is a major risk factor for cardiovascular disease. Several clinical studies found endothelial dysfunction after acute smoking as well as after chronic smoking (22, 23). In our crossover trial we included smoking and nonsmoking patients. To exclude a possible influence of smoking on our results, we analyzed smokers and nonsmokers separately using a two-factor repeated measurement ANOVA. There was no significant interaction between treatment and smoking. Compared with nonsmokers, smokers showed a less pronounced increase in forearm blood flow during treatment with both solutions, but both groups responded to the treatment (Table 4).
There were certain limitations to our study. We used an open-label design, which meant that both the patients and the treating physicians were aware of the treatment allocation. However, the assessor of vascular function was blinded. Thus, we did not expect any relevant observer or information bias. The duration of the present study, in which the influence of a 6-h dwell with an amino acid dialysis solution on vascular function was examined, did not allow inferences to clinically more relevant long-term effects of this solution. However, the somewhat incomplete recovery of forearm reactive hyperemia within 24 h and the fact that malnourished peritoneal dialysis patients use this solution once or twice per day indicates a need for long-term studies. Until these studies are underway, the benefit and potential risk have to be weighed against each other in patients who receive amino acid peritoneal dialysis solutions. We did not select the participants on the basis of their comorbidity, because there is a high prevalence of vascular dysfunction related to hypertension and arteriosclerosis in renal failure. Therefore, the potential aggravation of vascular dysfunction found in the present study may be relevant for a significant proportion of dialysis patients. In many studies, the response to intraarterial infusion of vasoactive drugs into the brachial artery is used to assess forearm resistance vessel function. We did not use such techniques because of concerns arising from arterial puncture-related complications and a potential future need of hemodialysis fistulas in our patients. Thus, a single-arm, noninvasive measurement of reactive hyperemia and glyceryl trinitrate-induced vasodilatation of the forearm was chosen for this study, which proved sensitive in our pilot experiments. As opposed to ultrasound measurement of flow-mediated dilatation of the brachial artery, which is only capable of describing an effect in a conduit artery, these measures represent responsiveness of resistance vessels in the forearm circulation as already described in other patient cohorts, including the forearm muscle and skin microvasculature (9, 10). Period, interaction, and carryover effects need special consideration in a crossover trial. We found no interaction or carryover effects. Ideally, the patients' underlying conditions and ability to respond to the intervention remained unchanged from the first to the second treatment period. If this is not the case, there is evidence of a period effect. We found a period effect for the secondary endpoint, plasma tHcy concentration. There is no explanation for this observation, and it remains unclear whether this is a genuine effect or just a random variation, particularly because there was no period effect for the primary endpoint. In any case, it makes the interpretation of the effect of intervention on tHcy concentrations difficult.
In conclusion, we found acute impairment of reactive forearm blood flow responsiveness in smoking and nonsmoking peritoneal dialysis patients after one session of a 6-h dwell with a commercial amino acid dialysis solution. Because endothelial dysfunction is associated with increased morbidity and mortality, the long-term use of these solutions may increase the risk of cardiovascular disease. Ongoing studies should characterize the amino acids involved and address the long-term cardiovascular effects of amino acid-containing peritoneal dialysis solutions.
ACKNOWLEDGMENTS
We thank Mathilde Sector for correcting the manuscript.
AV and GS-P conceived and designed the study. AV, MF, JP, PK, and SS contributed to data acquisition. AV, MF, MM, CR, MW, and GS-P helped analyze and interpret the data. AV, MF, JP, PK, SS, CR, MW, and GS-P helped draft the manuscript. AV, MF, MM, MW, and GS-P critically revised the manuscript for important intellectual content. MM conducted the statistical analyses and wrote parts of the results section and the section on study limitations. AV, MF, JP, MW, and GS-P provided administrative, technical, or material support. No pharmaceutical company was involved in this study. AV and GS-P received honoraria or travel grants from Baxter, Gambro, and Fresenius (manufacturers of dialysis solutions) that were unrelated to the conduct of this trial. The other authors had no financial or nonfinancial conflict of interest to declare.
REFERENCES