A concentrate acidified with citric acid instead of the less physiologic acetic acid has been successfully implemented in the United States for bicarbonate haemodialysis over the past 7 years [1–3]. In contrast to traditional regional citrate anticoagulation, the small amount of citrate used in the acid concentrate (0.8 mmol/L; only about one-fifth of the concentration necessary to achieve anticoagulation [1, 4, 5]) affects the calcium concentration and the locally enhanced coagulation activation in a limited way, resulting in approximately 10% reduction in post-dialysis ionized calcium and in no measurable systemic anticoagulation [1]. The absence of significant systemic repercussions is related not only to the low amount of citrate used but also to the rapid conversion of citrate into bicarbonate, which takes place in the liver and muscles and results in a higher post-dialysis bicarbonataemia [1, 6, 7]. Despite the rapid clearance of citrate, the local consequences of removing calcium from the blood clotting cascade have measurable positive effects on the dialyser life-span in the "reuse" modality and on dialysis quality, as quantified by urea Kt/V [1, 3]. The improvement in urea clearance has been correlated with an assumed favourable effect on dialyser fibre permeability mediated by the intradialyser anticoagulant properties of citrate [1, 3, 8]. Considering the importance of limiting the biocompatibility-related coagulation activation taking place in the extracorporeal circuit [9–17], the availability of a simple way to inhibit it without affecting systemic coagulation and bleeding risk [18] is very promising.

Although thousands of patients have been treated in recent years with haemodialysis fluids based on citric instead of acetic acid, the haemodynamic tolerance (the reduction in ionized calcium concentration and the increase in bicarbonateamia could both result in a lower intra-dialytic blood pressure [19–25]) and the amount of systemic coagulation activation related to each one of the modalities, have not been investigated.

The aim of this randomised, controlled, single-blind, cross-over study in single-use dialyser bicarbonate haemodialysis was to detail the consequences on systemic haemodynamics (primary outcome) and on coagulation activation, acid-base status, calcium balance and dialysis efficiency (secondary outcomes) of using citric instead of acetic acid in haemodialysis fluids.

Twenty-five chronic haemodialysis patients (15 male and 10 female) (sample size arbitrarily set in the absence of previous data or studies with an analogous primary outcome), undergoing dialysis in the dialysis unit of the Ospedale la Carità (Locarno, Switzerland) 3 to 4 hours three times a week who were clinically stable and without intercurrent illnesses were enrolled in the study. A single-blind, cross-over design was used in which the patients were initially randomised (one to one beginning with acetic acid dialysate according to the enrollement sequence) into one of two arms of the study i.e. receiving either acetic acid (modality A) or citric acid (modality C) dialysate. In the following 3 weeks the modality was switched weekly to the alternative one. Finally, with the intention to compensate for the reduction in serum calcium induced by citrate binding, both study arms were completed by a week in which patients were administered a citric acid dialysate containing a calcium supplement of 0.25 mmol calcium/L (modality C+) (see Figure 1 for details).

The haemodialyses were performed using a 4008 H machine equipped with a cartridge of bicarbonate Bibag^© and a high-flux single-use modified polysulfone membrane (Helixone^©; FX 80 or 100 with an effective surface area of 1.8 and 2.2 m² and an ultrafiltration coefficient of 59 and 73 ml/h mmHg, respectively), all from Fresenius Medical Care (Bad Homburg, Germany). Dialysis fluids were produced by Bichsel AG (Interlaken, Switzerland); the fluid composition was as following: sodium 138 mmol/L, magnesium 0.5 mmol/L, glucose 5.5 mmol/L, potassium 2 mmol/L (individually adapted according to the pre-dialysis serum potassium), acetate 3.0 mmol/L in A and 0.3 in C and C+, citrate 0 mmol/L in A and 0.8 in C and C+ and calcium 1.25 and 1.50 mmol/L in A and C and 1.50 and 1.75 in C+. Intradialytic heparin anticoagulation was carried out according to standard procedures and continued without dose adaptations in both arms of the study; priming of the dialyser was heparin free. The prescribed dialyser effective surface area, dialysis fluid conductibility, temperature and composition (with the exception of acetate, citrate and calcium), and effective blood flow were recorded at enrolment in the study and were left unchanged for the following 5 weeks. The medications of the patients (including phosphate binders) were also left unchanged.

Serum BUN, creatinine, potassium, phosphate, calcium and whole-blood pH, bicarbonate and ionised calcium were measured at the beginning and at the end of the third dialysis session of each week. Ionised and total calcium were also measured at the beginning and at the end of the first dialysis session of each week. Prothrombin fragments 1+2 (F1+2) and thrombin-antithrombin complexes (TAT) were determined at the beginning and at the end of the third dialysis session of each week [26]. Blood samples were taken from the arterial limb of the shunt.

Systolic and diastolic blood pressures and heart rate were measured before the start of each session and then repeated at 30-min intervals throughout dialysis with an automated Blood Pressure Monitor 4008 (Fresenius Medical Care, Bad Homburg, Germany) integrated in the dialysis machine. Stroke volumes (integrated mean of the flow waveform between the current upstroke and the dicrotic notch) and peripheral resistances (ratio of mean arterial pressure to stroke volume multiplied by heart rate) were evaluated between 5 and 10 minutes after the start of the session and then every 45 minutes using a finger beat-to-beat monitor Finometer^© (Finapres Medical Systems BV, Arnhem, The Netherlands). The use of isotonic saline infusions (100–200 ml) to treat symptomatic hypotension or symptoms related to intravascular hypovolaemia was registered.

Kt/V was calculated using a second-generation single-pool Daugirdas formula (Kt/V = -ln(R-0.03) + [(4-3.5 × R) × (UF/W)] where R = post-dialysis BUN/pre-dialysis BUN, UF = net ultrafiltration and W = weight).

Citrate accumulation during dialysis was estimated by calculating the calcium gap = |post – predialysis total calcium| - |post – predialysis ionised calcium|[6]. In a post-hoc analysis, 0.2 mmol/L was used as a cut-off value to classify patients as rapid (< 0.2 mmol/L) or slow (= 0.2 mmol/L) metabolisers.

Results were expressed as mean ± SD. Statistical analyses were performed using a statistical software package (SPSS 12.0; SPSS Inc., Chicago, IL, USA). Laboratory and haemodynamic parameters were compared first with an ANOVA and then, if significant, by a paired t-test of the mean of the values obtained in each patient with each modality. Haemodynamic parameters as a function of dialysis time were compared using a trapezoidal estimation of the area under the curves followed by a Wilcoxon signed rank test. To better understand the haemodynamic changes, an evaluation of the maximum increase and decrease in each parameter was added to the data analysis (see Figure 2 for an explanation of the calculation). Percentages were compared using Fisher's exact test. In all cases, p = 0.05 was considered statistically significant; p was expressed as ns (not significant), = 0.05, < 0.05, < 0.01 and < 0.001.

The protocol of the study was approved by the local Ethical Committee. All patients gave written informed consent prior to enrolment in the study.

The use of citrate rather than acetate as a dialysate in bicarbonate haemodialysis decreases peripheral resistances and slightly reduces systolic and diastolic blood pressures. Nonetheless, both the analysis of maximum fluctuations in peripheral resistances during dialysis and data describing subjective tolerance suggest a trend towards improved haemodynamic stability for patients on the citrate schedule. Interestingly, patients developing intradialytic hypertension experienced a larger reduction in blood pressure. Even with single-use dialysers, citrate resulted in better urea and phosphate removal, probably by preventing clotting of the dialyser fibres. To maintain the same intradialytic calcium balance following the switch to citrate, the calcium concentration of the dialysate should be increased of about 0.25 mmol/L. However, considering that steady-state pre-dialysis calcium was not significantly modified by the use of citrate, the need for calcium supplementation of the dialysate, particularly in patients treated with a 1.25 mmol calcium/L regimen, should be individually determined. In the citrate modality, a systemic effect on the coagulation cascade, as evaluated by specific markers, was not detected. This was probably associated with reduced intra-dialyser coagulation activation. Therefore, even if the use of citrate dialysate is recommended, in order to reduce the need for anticoagulants in patients at risk of bleeding, this aspect needs further investigation.

The positive impact of the citrate dialysate on dialyser efficiency, on acid-base status and the haemodynamic pattern as well as the subjective tolerance together suggest that the substitution of citrate for acetate in dialysis fluids improves haemodialysis in selected patients.