Effects of hydroxyethyl starch in critically ill patients
1 Jena, Germany
2 Créteil- France
* E-mail: frank.brunkhorst{at}med.uni-jena.de
Editor—The Sepsis Occurrence in Acutely ill Patients (SOAP) trial group have published another analysis from their database1 on hydroxyethyl starch (HES) and its effects on renal function.2 However, we are concerned that the methods used were not adequate, and the conclusion drawn ‘that HES had no influence on renal function or the need for renal replacement therapy (RRT)’ must therefore be viewed with caution. Our main concerns are as follows.
- Cohort studies must be planned in such a way that available data about potential confounding factors are of good quality. To ensure this, outcome events must be clearly pre-specified in the protocol and the data which will be collected must specifically address the question.3 4 Unfortunately, this is not the case with the SOAP protocol. Its short case report form is quite condensed and was primarily designed to study the epidemiology of sepsis and related therapeutic measures, not to answer nearly every open question in critical care medicine.
- ‘The "subsequent need for RRT" was defined in the SOAP protocol as the initiation of RRT in the ICU at least 24 h after HES administration or 24 h after admission in patients who did not receive HES’. Thus, the multivariate analyses compared groups with differing definitions for RRT, which is not acceptable.
- Median stay in ICU for the cohort of 3147 patients was only 3 days. Multivariate analysis was done on the 1970/3147 patients who stayed more than 24 h in the ICU. Since the multivariate analysis was undertaken in this subgroup, these patients would need to be characterized in more detail.
- ‘A total of 1287 received only crystalloids’—this means that 41% of patients either did not receive colloids or did not need volume expansion. This does not mean that 41% of the patients received fluid resuscitation with only crystalloids. As crystalloids are often infused as maintenance fluids, this heterogenous patient group cannot serve as a comparator to the HES group.
- ‘The degree of organ failure assessed by the SOFA score, procedures, and the presence of sepsis syndromes on admission in patients who did not receive HES and at onset of HES administration in those who did were also included as independent variables’. Again, adjustment of confounding factors was performed for different time periods in HES and non-HES groups.
- The median amount of HES administered was 555 (IQR, 500–1000) ml per day and the total amount was only 1000 (500–2250) ml per patient. This is an unusually low HES dosage compared with other studies, in which patients with sepsis received median cumulative HES dosages of 4550 ml or 31 ml kg–1 (IQR 19–51), respectively.5 6
- Finally, ‘subsequent need for RRT’ is an inadequate end-point because it is highly dependent on ICU practices, especially in a study with 198 participating ICUs from 24 European countries.
1 Jena, Germany
2 Brussels, Belgium
* E-mail: jlvincen{at}ulb.ac.be
Editor—The SOAP study7 was a collaborative effort involving more than 200 European investigators. The use of HES and its effects on renal function is one of the post hoc analyses of this database.2 We agree with Drs Brunkhorst and Schortgen that cohort studies must be planned in such a way that available data about potential confounding factors are of good quality, and these elements were carefully considered in the SOAP study. Definitions and guidelines were provided to the investigators before data collection. Data management was done centrally with daily plausibility analyses to assure good quality of data. Data to be collected were pre-specified by the steering committee.
If the aim of the SOAP study was solely to investigate the epidemiology of sepsis, we would not have collected data on many therapeutic measures, including fluid therapy. Indeed, the intention was to use the SOAP database to address a number of questions in critical care medicine. Useful data have been reported previously from similar databases and have contributed to our understanding of several aspects in the management of critically ill patients. To give only one example, the SUPPORT study,8 whose aim was to improve end-of-life decision making and care, served as a basis for challenging the use of the pulmonary artery catheter.9
Drs Brunkhorst and Schortgen claim that the multivariate analyses we have presented in our paper2 compared groups with different definitions for RRT, but this is not true. Events should follow predisposition. It was logical to define the ‘subsequent need for RRT’ as the initiation of RRT at least 24 h after HES administration or 24 h after ICU admission in patients who did not receive HES because pre-admission events could have led to renal failure already being present on ICU admission. It is also logical to adjust for the degree of organ failure, procedures, and sepsis syndromes at the onset of HES administration in patients who received HES and at admission in patients who did not receive HES because the latter group has no equivalent time point. The evolution of renal variables in patients with length of stay (LOS) >24 h can be extrapolated from Figures 2 and 3 (patients with LOS <24 h contribute logically only to the first box or column); there was no evidence for an adverse effect of HES on renal function.
Drs Brunkhorst and Schortgen seemed also to have overlooked the data presented in our paper, when they claim that we used patients who received only crystalloids as a comparator to the HES group. Data on the use of other colloids were provided and adjusted for in the multivariate analysis. In addition, the no-HES group included patients who received other colloid solutions. Drs Brunkhorst and Schortgen describe the amount of HES administered in the SOAP study as unusual, but again these are the data. Our study is the first large observational study, to date, to report the type and amount of i.v. fluids in European countries. One cannot compare the amounts of fluids used in an observational study with those used in interventional studies.
Finally, we feel that ‘subsequent need for RRT’ is an adequate end-point in our observational study, especially in view of the large number of centres. We included a large number of factors in our analysis, including the most relevant confounding factors known to influence renal function and outcome of critical illness.
References
1 Accessed on February 10, 2007. Available from http://www.intensive.org/soap/soap_study.doc.
2 Sakr Y, Payen D, Reinhart K, et al. Effects of hydroxyethyl starch administration on renal function in critically ill patients. Br J Anaesth (2007) 98:216–24.
3 Wunsch H, Linde-Zwirble WT, Angus DC. Methods to adjust for bias and confounding in critical care health services research involving observational data. J Crit Care (2006) 21:1–7.[CrossRef][Web of Science][Medline]
4 Hunter D. First, gather the data. N Engl J Med (2006) 354:329–31.
5 Boldt J, Heesen M, Muller M, Pabsdorf M, Hempelmann G. The effects of albumin versus hydroxyethyl starch solution on cardiorespiratory and circulatory variables in critically ill patients. Anesth Analg (1996) 83:254–61.[Abstract]
6 Schortgen F, Lacherade JC, Bruneel F, et al. Effects of hydroxyethylstarch and gelatin on renal function in severe sepsis: a multicentre randomised study. Lancet (2001) 357:911–6.[CrossRef][Web of Science][Medline]
7 Vincent JL, Sakr Y, Sprung CL, et al. Sepsis in European intensive care units: results of the SOAP study. Crit Care Med (2006) 34:344–53.[CrossRef][Web of Science][Medline]
8 The SUPPORT Principal Investigators. A controlled trial to improve care for seriously ill hospitalized patients. The study to understand prognoses and preferences for outcomes and risks of treatments (SUPPORT). JAMA (1995) 274:1591–8.
9 Connors AF Jr, Speroff T, Dawson NV, et al. The effectiveness of right heart catheterization in the initial care of critically ill patients. SUPPORT Investigators. JAMA (1996) 276:889–97.
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||