To the Editor:
Martin et al.1 described a meta-analysis of 17 randomized controlled trials evaluating renal function in 1,230 total surgical patients allocated to hydroxyethyl starch (HES) 130/0.4 or control fluid. At baseline, mean serum creatinine (SCr) was lower in HES 130/0.4 recipients than the control group, and after surgery, the most extreme mean SCr values were higher in the patients receiving HES 130/0.4. However, these differences were not statistically significant. Limited data on acute renal failure and renal replacement therapy available among the trials included in the meta-analysis also showed no significant differences. The investigators concluded that there was no evidence of adverse postoperative renal effects due to HES 130/0.4.
A major weakness of this meta-analysis was short follow-up. In six of the included trials, renal function was evaluated only for 24 h or less after surgery and in two more trials only up to 48 h. SCr is neither a specific nor a sensitive marker for renal tubular injury, and threshold SCr increase for diagnosis of incipient acute kidney injury (AKI) is not typically observed until 48 h or more after surgery.2 This temporal pattern is nicely demonstrated by the Crystalloid versus Hydroxyethyl Starch Trial of HES 130/0.4, which included 2,876 surgical patients comprising 43% of the overall trial population (fig. 1).3 HES 130/0.4 significantly increased SCr in that trial but the effect was not clearly evident until after 48 h, and the SCr peak was not reached until after 4 days. Thus, in eight trials of the meta-analysis, the SCr peak is likely to have been missed because of short follow-up. In the Crystalloid versus Hydroxyethyl Starch Trial, SCr increase was accompanied by increased recourse to renal replacement therapy in the HES 130/0.4 group with a relative risk of 1.21 and 95% CI of 1.00–1.45 (P = 0.04). These adverse renal effects were encountered despite a low average daily HES 130/0.4 dose (526 ml).
Furthermore, five of the nine trials in the meta-analysis with more than 48 h follow-up compared HES 130/0.4 with other artificial colloids known to cause renal impairment.4 Lack of difference in those trials would if anything suggest deleterious renal effects of HES 130/0.4.
The most extreme mean postoperative SCr value is an endpoint of convenience rather than a validated indicator of AKI in individual patients. In reality, the trials of this meta-analysis were not designed to assess HES 130/0.4-induced AKI. SCr data were often reported simply as part of a laboratory test panel. No included trial assessed AKI in accordance with a validated classification system such as risk of renal dysfunction, injury to the kidney, failure of kidney function, loss of kidney function, and end-stage kidney disease or Acute Kidney Injury Network.2 Only one included trial report even specified any definition for acute renal failure.
Beyond these threats to the validity of the meta-analysis, its applicability to routine clinical practice is limited. In 16 of the 17 included trials, patients with renal dysfunction at baseline were excluded. Hence, it can only be concluded that evidence was not found of adverse renal effects among surgical patients at low risk of AKI.
This meta-analysis is based on two premises that: (1) HES 130/0.4 may pose less renal risk than other HES solutions and (2) surgical patients may be less susceptible to HES-induced AKI than other critically ill patients. Neither premise was supported by a systematic review on the comparative safety of colloids encompassing 69 clinical studies, including 42 randomized trials with 10,382 total patients.4 Martin et al. wrongly suggest that the systematic review was incomplete. In fact, all studies fulfilling prespecified selection criteria were included in the systematic review without exception, as elsewhere detailed.5 Moreover, the systematic review has recently received confirmation from a new meta-analysis showing that HES increases mortality (relative risk, 1.09; 95% CI, 1.02–1.17), acute renal failure (relative risk, 1.27; 95% CI, 1.09–1.47), and renal replacement therapy (relative risk, 1.32; 95% CI, 1.15–1.50).6 The effects of various HES solutions were found to be similar in that meta-analysis, as were the effects of HES in assorted populations of critically ill patients.
Additional evidence also raises concern about the perioperative infusion of HES 130/0.4. In a retrospective study of 6,553 patients undergoing cardiopulmonary bypass surgery, HES 130/0.4 exposure was found to be an independent risk factor for renal replacement therapy (odds ratio, 1.708; P = 0.004).7
The risk of excessive bleeding due to HES 130/0.4 also needs to be considered.4 A meta-analysis of randomized cardiopulmonary bypass surgery trials demonstrated increases in postoperative blood loss and blood product transfusion as well as reoperation for bleeding among patients receiving HES.8 There was no evidence that those risks differed among HES solutions. A systematic review of viscoelastic device studies indicated impairment of coagulation specifically by HES 130/0.4.9
On the basis of available evidence, it has been recommended that clinical use of HES for acute volume resuscitation be avoided because of serious safety concerns.6 Whether there exist specific patient groups who might benefit from HES 130/0.4 would have to be demonstrated in future randomized trials specifically designed and adequately powered to detect differences in renal outcomes. Current data are not adequate to establish the perioperative safety of HES 130/0.4.