Author + information
- Received June 22, 2005
- Revision received September 30, 2005
- Accepted October 3, 2005
- Published online March 21, 2006.
- Sophia Zoungas, MBBS⁎,
- Barry P. McGrath, MBBS, MD⁎,⁎ (, )
- Pauline Branley, BMed, PhD†,
- Peter G. Kerr, MBBS, PhD‡,
- Christine Muske, BSci(Nurs)†,
- Rory Wolfe, PhD†,
- Robert C. Atkins, MBBS, DSc‡,
- Kathy Nicholls, MBBS, MD§,
- Margaret Fraenkel, BMBS, PhD∥,
- Brian G. Hutchison, MBBS¶,
- Robert Walker, MBChB, MD# and
- John J. McNeil, MBBS, PhD†
- ↵⁎Reprint requests and correspondence:
Prof. Barry P. McGrath, Department of Vascular Sciences and Medicine, Monash University, Dandenong Hospital, Dandenong 3175, Victoria, Australia.
Objectives The Atherosclerosis and Folic Acid Supplementation Trial (ASFAST) aimed to establish whether high-dose folic acid would slow the progression of atherosclerosis and reduce cardiovascular events in patients with chronic renal failure (CRF).
Background Hyperhomocysteinemia is a potential contributor to the high rates of cardiovascular morbidity and mortality in patients with CRF.
Methods A total of 315 subjects with CRF, mean age 57 years (range 24 to 79 years) were randomized to 15 mg folic acid daily or placebo and followed for a median of 3.6 years. The primary intima-media thickness (IMT) and clinical end points were: rate of progression of mean maximum carotid IMT and a composite of myocardial infarction (MI), stroke, and cardiovascular death. Secondary end points included all cardiovascular events and change in pulse wave velocity, systemic arterial compliance and augmentation index. Data were analyzed by intention-to-treat.
Results Plasma total homocysteine was reduced by 19% in the folic acid group. There was no significant difference between the treatment groups in rate of change of IMT or any measure of artery function. Seventy-seven events occurred in the folic acid group (14.9 per 100 patient-years) as compared with 86 in the placebo group (16.3 per 100 patient-years). The rates of the primary and secondary clinical end points at five years were not significantly different after adjustment for baseline differences between the groups (adjusted hazard ratio for MI, stroke, and cardiovascular death: 0.98 [95% confidence interval: 0.66 to 1.47]; p = 0.94; for all cardiovascular events: 0.95 [95% confidence interval: 0.69 to 1.30]; p = 0.75).
Conclusions High-dose folic acid does not slow atheroma progression or improve cardiovascular morbidity or mortality in patients with CRF.
Chronic renal failure (CRF) is a risk factor for accelerated cardiovascular disease. Age-adjusted cardiovascular mortality rates are 10 to 20 times greater in patients with CRF than the general population (1). More than one-half of all dialysis-related deaths are cardiovascular in nature (2–4).
Homocysteine, a nonessential sulphur-containing amino acid, accumulates in CRF owing to both altered metabolism (5) and reduced renal excretion (6). Approximately 85% of patients with CRF have elevated homocysteine levels (6).
Population studies have reported an association between mild hyperhomocysteinemia and occlusive vascular disease, although the findings are inconsistent (7–9). Levels of homocysteine and cardiovascular disease have also been linked to folate status (5).
In CRF, observational studies have suggested a causal association between hyperhomocysteinemia and cardiovascular disease (10,11). Intervention studies are necessary to determine whether the relationship is causal and whether lowering of plasma homocysteine reduces the incidence of cardiovascular events in patients with CRF (12).
Randomized controlled trials of homocysteine-lowering therapy have given inconsistent results, with reports of reduced carotid IMT progression (13,14), improvement in exercise-electrocardiogram responses (15), a decreased rate of restenosis and adverse events after percutaneous coronary intervention (16,17), and an increased rate of restenosis and adverse events after coronary stent placement (18). In patients treated with a multivitamin containing folic acid after ischemic stroke, no benefit in cardiovascular outcome was found (19).
The Atherosclerosis and Folic Acid Supplementation Trial (ASFAST) was a multicenter, randomized, double-blind, placebo-controlled trial of high-dose folic acid therapy in patients with CRF, examining the hypothesis that cardiovascular disease and atheroma progression in CRF is related to hyperhomocysteinemia and can be reduced by lowering homocysteine with folic acid therapy.
In brief, the study aims were to determine whether high-dose folic acid supplementation would slow the progression of carotid artery intima-media thickness (IMT) and reduce the incidence of cardiovascular events (a composite end point of myocardial infarction, stroke, and death from cardiovascular cause) in this population. The differences between the treatment groups in all fatal and nonfatal cardiovascular events and indices of arterial function (pulse wave velocity [PWV], systemic arterial compliance [SAC], and augmentation index [AIx]) were also studied.
The double-blind randomized study evaluated daily 15 mg folic acid or identical placebo in 315 patients with end-stage renal disease. The detailed protocol including study design, rationale, organization, measurements, power calculations, and baseline results have been published previously (20).
Men and women aged over 18 years with CRF of any cause, a serum creatinine of 0.40 mmol/l or greater (creatinine clearance <25 ml/min), and either awaiting dialysis or already treated with continuous ambulatory peritoneal dialysis, intermittent peritoneal dialysis, or hemodialysis were eligible for the study.
Three hundred fifteen patients who met the entry criteria were randomly assigned to either 15 mg of folic acid daily or identical placebo. At entry, those being treated with folic acid or multivitamins containing folic acid were required to withdraw treatment; patients with folate deficiency requiring supplementation were excluded. The study medication was dispensed in identical containers with neither study staff nor participant aware of treatment allocation. Study consultations, including carotid IMT measurement, occurred at yearly intervals with treatment scheduled to last five years.
Study conduct and organization
Recruitment commenced on June 30, 1998, and was closed on December 31, 2000. Four renal units in Australia and one in New Zealand participated. The ethics review board at each institution approved the protocol, and the study was conducted in accordance with the National Health and Medical Research Council of Australia Guidelines on Human Experimentation. An independent steering and safety committee supervised the study.
Folate replacement was not allowed during the study. However any participant found to have red cell folate levels below 510 nmol/l on six monthly tests (normal range: 510 to 1,310 nmol/l) were allowed folic acid supplementation, at the smallest dose required to normalize the level, and continued in the study.
The study was terminated on December 31, 2003, and all mortality and morbidity to this date were ascertained for 311 (99%) of the 315 participants randomized. Subjects who underwent renal transplantation during follow-up were censored at the time of transplantation. This was deemed necessary because of the effect of transplantation per se on cardiovascular survival.
The primary IMT end point was change in the rate of progression of mean maximum carotid artery IMT. The main clinical end point was a composite of myocardial infarction, stroke, and death from cardiovascular cause. Both first and subsequent cardiovascular events were included. Secondary end points included all fatal and nonfatal cardiovascular events, including myocardial infarction, stroke, unstable angina, revascularization, and peripheral vascular disease. Changes in PWV, SAC, and AIxwere also assessed. All clinical end points were subject to an independent adjudication process by an endpoint monitoring committee.
Measurement of carotid IMT
The IMT of the common carotid artery was measured using a high-resolution ultrasound machine (DRF-400, Diasonics, Santa Clara, California) with a handheld 7.5-MHz transducer (7.5-SPC mechanic sector transducer, Diasonics). All measurements were performed in the same study laboratory in each major city center, and a single reference laboratory performed all image analyses.
In brief, a region 1.0 cm proximal to the origin of the bulb of both common carotid arteries was identified by B-mode ultrasonography. Three images of each B-mode, taken at three different angles (anterior, lateral, medial), were recorded, digitized, and saved for subsequent analysis. The 60 measurements from the far wall of both common carotid arteries were used in the final data analysis to give mean maximum and mean IMT values.
Measurements of arterial function
A subgroup of subjects (n = 207) from two study centers underwent measurement of indices of arterial stiffness and compliance including: PWV, the speed of travel of the pressure pulse along an arterial segment; SAC, the relationship between change in volume and change in pressure during diastole; and AIx, the percentage contribution of reflected waves to peak central pressure. Details of these methods have been given previously (20).
Two operators performed all image reading of IMT and analysis of arterial function indexes. The intraobserver correlation coefficient for IMT was 0.89, the coefficient of variation was 10.4%, and the interobserver correlation coefficient was 0.88. The intraclass correlation coefficients (95% confidence interval [CI], number of observations) were: IMT 0.85 (0.72 to 0.97; n = 40); PWV 0.95 (0.91 to 0.99; n = 46); SAC 0.92 (0.86 to 0.99; n = 42); and AIx0.98 (0.97 to 0.99; n = 42).
The sample size of the study was based on IMT progression. Using a predicted 0.03 mm IMT increase over three years in the placebo group, i.e., 0.01 mm/year progression, with an estimated standard deviation (SD) of 0.01 mm/year, the study was powered to detect a 33% slower rate of progression in the folic acid group, at a significance level (alpha) of 0.05 (two-sided) with 80% power. The number of subjects required in each treatment arm was 143, that is 286 subjects to be followed for three years.
Assuming a cardiovascular event/death rate in the placebo group of 1.5 events per 10 person-years of follow up, a mean follow-up duration of three years per participant (858 person-years of follow-up in total) and a two-sided significance level (alpha) of 0.05, the study had approximately 91% power to detect a 50% reduction in the rate of major cardiovascular events in the folic acid group.
A semiparametric linear regression model was fitted to mean maximum IMT with an interaction between study arm and time on study (years), assuming linear change over time. The method of generalized estimating equations (GEE) (21) with an exchangeable working correlation matrix was used to estimate the parameters of this model while taking into account possible within-person correlation in mean maximum IMT measurements over time. The standard errors calculated by the GEE method are robust to misspecification of the correlation and variance parts of the model (21). Adjustments in the model were made for the following baseline factors found to be related to death/transplant propensity (age, diastolic blood pressure, diabetes, past cardiovascular disease, and antiplatelet agent use) were then included.
The primary analysis used only observed data and assumed that missing data was missing at random (22). To assess the possible impact of this assumption which is governed by patient dropout, mean maximum IMT was re-analyzed with stratification by time of dropout with respect to IMT measurement.
Survival curves were constructed according to the Kaplan-Meier method. Event rates were compared between treatment groups using Cox proportional hazard regression models. Adjusted hazard ratios were calculated by the same method to allow for imbalances in the treatment groups at study entry.
The GEE method described above was used to compare total plasma homocysteine (tHcy) levels between study arms across the follow-up times without assuming a linear change over time, i.e., performing comparisons at each follow-up time point. A Spearman correlation coefficient was used to describe the relationship between plasma tHcy and red cell folate levels. Analysis of covariance was used to compare 12-month PWV, SAC, and AIxbetween study arms with adjustment for baseline values and assuming that missing measurements were missing at random. The 12-month time point was chosen because it was anticipated that any treatment effect should have been evident by that time.
Data were analyzed on an intention-to-treat basis, i.e., according to allocated treatment group and irrespective of compliance with treatment. Statistical significance is confirmed when p values are <0.05. All analyses were performed in Stata Statistical Software, Release 7.0 (Stata Corp., College Station, Texas).
Three hundred fifteen subjects were recruited over a two-year period and followed for a median of 3.6 years, providing a total of 1,044 patient-years of observation. One hundred fifty-six subjects were assigned to 15 mg folic acid daily and 159 to identical placebo. During follow-up, 71 subjects discontinued treatment (41 from folic acid group, 30 from placebo group) and 80 had renal transplantation (37 from folic acid group, 43 from placebo group). As shown in Figure 1,all randomized subjects were included up to the time of death, transplantation, or closure of the study.
The baseline characteristics of the 315 participants are shown in Table 1.At study entry, the proportion of women in the placebo group was greater than in the folic acid group (37.7% vs. 26.9%). With respect to age, mode of dialysis, cardiovascular risk factor profile, medical therapy, and fasting biochemistry, the two groups were closely matched except for use of antiplatelet agents, which was greater in the placebo than in the folic acid group (31.4% vs. 21.8%).
Of the 102 women and 213 men studied, 35% reported a history of cardiovascular disease (angina [32%], acute myocardial infarction [11%], and/or cerebrovascular disease [9%]), 90% hypertension, 23% diabetes mellitus, and 44% hypercholesterolemia.
Effects of treatment on tHcy and red cell folate
At the end of follow-up, 17 participants (5%) had received additional folic acid therapy (10 [6%] in the placebo group and 7 [4%] in the folic acid group; dose range 0.3 to 5 mg; median 0.7 mg daily) for the treatment of folate deficiency (red cell folate levels below 510 nmol/l; 8 in the placebo group) or on the basis of treating physician request (2 in the placebo group and 7 in the folic acid group).
The tHcy levels over time are shown in Figure 2.At baseline, median tHcy and red cell folate levels were similar between the treatment groups (folic acid group vs. placebo group: 24.6 vs. 25.1 μmol/l tHcy; 1,354 vs. 1,186 nmol/l red cell folate). After 12 and 36 months of follow-up the median tHcy levels were 19.9 and 21.5 μmol/l, respectively, in the folic acid group compared with 24.4 and 23.9 μmol/l, respectively, in the placebo group (normal range: 3 to 13 μmol/l). At 12 months, 10.9% of the folic acid group had achieved normal tHcy levels as compared with 1.6% of the placebo group. Estimated difference in mean tHcy between treatment groups at 12 months was −7.5 (95% CI: −10.4 to −4.6; p < 0.001); at 24 months −7.1 (95% CI: −10.2 to −4.0; p < 0.001); at 36 months −2.4 (95% CI: −6.8 to 1.9; p = 0.27); and at 48 months −4.7 (95% CI: −9.4 to −0.1; p = 0.05).
The median red cell folate levels increased three-fold in the folic acid group and were unchanged in the placebo group (folic acid group vs. placebo group: 3,819 vs. 1,159 nmol/l) and remained elevated in the folic acid group after 36 months (folic acid group vs. placebo group:2,797 vs. 1,509 nmol/l). The tHcy levels was inversely correlated with red cell folate levels (Spearman correlation r = −0.35; p < 0.001).
At baseline, carotid IMT was not significantly different between the treatment groups (mean ± SD, folic acid group vs. placebo group: 1.06 ± 0.23 mm vs. 1.08 ± 0.28 mm; p = 0.75).
Mean change in IMT from baseline to follow-up is shown in Table 2.There was no significant difference in the rate of progression of mean maximum IMT between the treatment groups (0.01 mm/year, 95% CI: −0.01 to 0.03; p = 0.43). Further analysis stratified by time of drop-out also showed no significant difference in overall IMT change per year between the treatment groups (mean change ± SD: 0.03 ± 0.14 mm/year in the placebo group vs. −0.02 ± 0.17 mm/year in the folic acid group; p = 0.82).
For subjects who dropped out early (after visit 1 or 2) because of death, IMT increased (mean change ± SD: 0.05 ± 0.24 mm/year), whereas for subjects followed for longer (drop-out after visit 3 or 4) or complete follow-up IMT decreased (mean change ± SD: −0.02 ± 0.09 mm/year). The difference in IMT progression between those dying early during follow-up and those surviving longer was 0.07 mm/year (95% CI: 0.02 to 0.12 mm/year; p = 0.01).
The rates of all fatal and nonfatal cardiovascular events in the two treatment groups are shown in Table 3.Seventy-seven events occurred in the folic acid group (14.9 per 100 patient-years) and 86 in the placebo group (16.3 per 100 patient-years). Of these events, 44 (57%) were first events in the folic acid group (9.6 per 100 patient-years) as compared with 53 (62%) in the placebo group (11.7 per 100 patient-years).
Regarding the composite end point of myocardial infarction, stroke, and cardiovascular death, the rate of events in the folic acid group was 6.1 first events and 9.3 total events per 100 patient-years as compared with 8.3 first events and 10.2 total events per 100 patient-years in the placebo group. The difference, however, was not significant (hazard ratios for first and total events: 0.71 [95% CI: 0.42 to 1.21]; p = 0.21; and 0.85 [95% CI: 0.58 to 1.26]; p = 0.43; respectively). After adjustment for baseline differences in gender, diastolic blood pressure, and aspirin use, hazard ratio for first myocardial infarction, stroke, and cardiovascular death was 0.93 (95% CI: 0.58 to 1.48); p = 0.75; adjusted hazard ratio for all myocardial infarction, stroke, and cardiovascular death was 0.98 (95% CI: 0.66 to 1.47); p = 0.94 (Fig. 3).
The incidence of all fatal and nonfatal cardiovascular events was not significantly different between groups. The risk reduction was not significant after adjustment for baseline differences in gender, diastolic blood pressure, and aspirin use (adjusted hazard ratio for all first cardiovascular events: 0.89 [95% CI: 0.59 to 1.34]; p = 0.57; adjusted hazard ratio for all recurrent cardiovascular events: 0.95 [95% CI: 0.69 to 1.30]; p = 0.75).
After 12 months of treatment there was no significant difference between the groups in mean PWV(a − f) (difference in means: −0.31 m/s [95% CI: −1.20 to 0.57]; p = 0.49), SAC (difference in means: 0.022 U/mm Hg [95% CI: −0.046 to 0.090]; p = 0.52), or AIx(difference in means: 0.1% [95% CI: −5.3 to 5.5]; p = 0.97).
Twenty-five percent of patients in the folic acid group and 18% in the placebo group stopped taking their treatment because of adverse effects (6 in folic acid group; 2 in placebo group) or unwillingness to continue (20 in folic acid group; 12 in placebo group). No serious adverse effects were reported during the study. In 12 subjects (6 in folic acid group; 6 in placebo group), cobalamin deficiency (<160 pmol/l) was detected and led to replacement.
Observational studies have suggested a causal relationship between hyperhomocysteinemia and cardiovascular morbidity in the general population (7,8) and in patients with CRF (11,23). Folic acid can reduce elevated homocysteine levels (24–28) but has not yet been shown to improve atheroma progression or cardiovascular outcome in CRF.
Elevated homocysteine levels in the general population have been related to carotid IMT measurements (29,30). In CRF as well, higher carotid IMT measurements have been associated with high levels of homocysteine (31). Small studies with limited follow-up have reported a decrease in carotid IMT progression with vitamin supplementation with folic acid in renal transplant recipients (13) and in high-risk vascular patients (14).
In the present cohort of CRF patients we did not find a relationship between baseline carotid IMT and homocysteine level (20). Despite modest homocysteine lowering, high-dose folic acid therapy had no effect on atheroma progression, measured by mean maximum carotid IMT, over the five-year period.
Mean maximum IMT increased in both treatment groups for the first two years and thereafter reduced slightly. One explanation is that those with the greatest risk for atheroma progression and clinical events may have dropped out early. In long-term intervention studies the greatest limitation with following IMT is missing data in subjects having fatal events (32). In this study the finding that IMT was increased in those dropping out early because of death, and decreased in those followed for longer, supports this. Another possible explanation is that the increase in IMT over time was attenuated by dilatation of the carotid artery. Carotid diameter was not measured in our subjects.
A small dose-response study of folic acid therapy in CRF suggested no difference in cardiovascular morbidity or all-cause mortality among patients taking between 1 and 15 mg folic acid daily (12). That small study had no placebo arm. In the present study, after adjustment for baseline differences in gender, diastolic blood pressure, and aspirin use, there was no significant difference in cardiovascular events for those treated with high-dose folic acid compared with placebo. The all-event confidence intervals suggest that a reduction of greater than 34% in the event rate owing to folic acid could be excluded.
We observed no effect of treatment with folic acid on any index of arterial stiffness or compliance. This finding is consistent with the observed lack of effect of folic acid on IMT progression and is contrary to the hypothesis that arterial disease in CRF is related to hyperhomocysteinemia and can be improved by homocysteine lowering. It also suggests that folic acid does not improve arterial function in patients with CRF independent of its homocysteine effect.
There are several possible reasons for the negative findings in this study. They include: homocysteine is not related to vascular disease progression in CRF; there was insufficient lowering of homocysteine to be of biologic significance; the dose or duration of treatment with folic acid was insufficient; and the study had insufficient statistical power, a problem aggravated by the high withdrawal and transplantation rates and folate replacement policy.
Arterial vascular disease in CRF was associated with traditional cardiovascular risk factors, but also with systemic inflammation. Malnutrition has also been implicated in low homocysteine levels (5). This has led to the suggestion that homocysteine is only a marker of cardiovascular disease in CRF and not a pathogenic factor.
After 12 months of folic acid treatment, there was a 19% decrease in total homocysteine levels in the folic acid group; levels were normalized in only 10.9%. This was maintained for up to 36 months and associated with a three-fold increase in red cell folate levels. In the placebo group there was no significant change in total homocysteine or red cell folate levels.
The failure of homocysteine levels to normalize with folic acid therapy in patients with CRF, despite markedly increased red cell folate levels, has previously been described (27). Possible explanations include impaired cellular uptake and utilization of folate metabolites (33), restricted intracellular capacity and saturation of the remethylation pathway (34), and reduction in plasma clearance of homocysteine (35). Touam et al. (36) suggested that the folate resistance could be overcome with the use of L-5-formyltetrahydrofolate; their result could not be replicated by others with the use of L-5-methyltetrahydrofolate (37).
In the present study, failure to normalize homocysteine levels was not explained by concurrent cobalamin deficiency, because all our patients were required to have normal levels of cobalamin, which was monitored throughout the trial. Others have studied treatment with pyridoxine (26,38), serine (39), betaine (40,41), and N-acetylcysteine (42) and similarly have shown no additional effect on homocysteine levels in dialyzed patients. The challenge remains to identify a therapy which can normalize homocysteine levels in patients with CRF.
Folic acid at a dose of 15 mg daily had no significant effect on atheroma progression nor did it improve cardiovascular outcome for patients with CRF. There was also no improvement in any index of artery function. Homocysteine levels were lowered but not normalized. It is possible that a treatment effect of <33% was missed. Widespread use of high-dose folic acid for cardiovascular protection in chronic renal failure cannot be advocated on the basis of these findings.
We are indebted to the patients and the health care teams at each study site (in Australia: Monash Medical Centre, Clayton, Victoria; Royal Melbourne Hospital, Parkville, Victoria; Austin and Repatriation Medical Centre, Heidelberg, Victoria; and Sir Charles Gairdner Hospital, Perth, Western Australia; and in New Zealand: Dunedin Hospital, Dunedin), to the members of the safety monitoring committee (Dr. Kelvin Lynn [Chair], Dr. David McGregor, and Prof. Mark Richards) and the end point adjudication committee (Prof. Henry Krum [Chair], Dr. Robyn Langham, and Dr. Pauline Branley), and to the staff at the coordinating study center for their invaluable assistance in the conduct of this study.
This study was supported by grants-in-aid from the National Health and Medical Research Council of Australia and National Kidney Foundation of New Zealand.
- Abbreviations and Acronyms
- augmentation index
- Atherosclerosis and Folic Acid Supplementation Trial
- confidence interval
- chronic renal failure
- generalized estimating equation
- intima-media thickness
- pulse wave velocity
- systemic arterial compliance
- total plasma homocysteine
- Received June 22, 2005.
- Revision received September 30, 2005.
- Accepted October 3, 2005.
- American College of Cardiology Foundation
- Disney A.P.S.
- Dennis V.W.,
- Robinson K.
- Robinson K.,
- Gupta A.,
- Dennis V.,
- et al.
- Verhoef P.,
- Hennekens C.H.,
- Malinow M.R.,
- Kok F.J.,
- Willett W.C.,
- Stampfer M.J.
- Jungers P.,
- Massy Z.A.,
- Khoa T.N.,
- et al.
- Moustapha A.,
- Naso A.,
- et al.
- Wrone E.M.,
- Hornberger J.M.,
- Zehnder J.L.,
- McCann L.M.,
- Coplon N.S.,
- Fortmann S.P.
- Till U.,
- Jentsch A.,
- Rohl P.,
- et al.
- Schnyder G.,
- Roffi M.,
- Flammer Y.,
- Pin R.,
- Hess O.M.
- Liang K.Y.,
- Zeger S.L.
- Little R.J.A.,
- Rubin D.B.
- Arnadottir M.,
- Brattstrom L.,
- Simonsen O.,
- et al.
- Malinow M.R.,
- Nieto F.J.,
- Szklo M.,
- Chambless L.E.,
- Bond G.,
- Atherosclerosis Risk in Communities Study
- Lim P.S.,
- Hung W.R.,
- Wei Y.H.
- Bostom A.G.,
- Shemin D.,
- Bagley P.,
- et al.
- Arnadottir M.,
- Brattstrom L.,
- Simonsen O.,
- et al.
- van Guldener C.,
- Janssen M.J.,
- Lambert J.,
- ter Wee P.M.,
- Donker A.J.,
- Stehouwer C.D.