Fluid responsiveness in healthy volunteers: data ... - Wiley Online Library

3 downloads 8240 Views 56KB Size Report
w.ntac.nhs.uk/nmsruntime/saveasdialog. aspx?lID=1225&sID=1010 ... these data are representative of within-patient .... apy within an enhanced recovery proto-.
Correspondence

Anaesthesia 2014, 69, 785–798

es.gov.uk/20130701143131/http://ww w.ntac.nhs.uk/nmsruntime/saveasdialog. aspx?lID=1225&sID=1010 (accessed 09/ 04/2014). 7. NHS England. Commissioning for quality and innovation (CQUIN): 2014/15. http://www.england.nhs.uk/wp-content/ uploads/2014/02/sc-cquin-guid.pdf (accessed 09/04/2014). doi:10.1111/anae.12742

Fluid responsiveness in healthy volunteers: data precision and significance We would like to make some observations about the methodology of the article by Godfrey et al. [1]. Their a priori power analysis was based on 12 consecutive images taken from one of the authors; these data are representative of within-patient variability but not between-patient variability, which plays a much larger role in the reproducibility of Doppler measurements [2]. As a result, there is a large discrepancy between the a priori power analysis which used an estimated mean (SD) of 118.7 (7.3) ml, with a calculated sample size of 6, and what actually happened in the study where the mean was 69.8 (19.6) ml with a retrospective calculated sample size of > 500 patients. The result is that there can be no confidence that a type-2 error has been avoided, because the power analysis was based on incorrect use of pilot data. In this study, there was a high proportion of patients (6/17 (35%)) who had suboptimal views or images, which may indicate a lack of precision of the derived stroke volume; it is unclear why it would be so different to other studies [3].

There is also the assumption that echocardiographic estimation of stroke volume used in this study ‘is the least derived’ and near to a gold standard. However, errors in estimating left ventricular outflow tract diameter and variation in the position and angle of the pulse wave Doppler sample volume will all play a part in reducing the precision. These caveats may explain why, in a review of the use of echocardiography for predicting fluid responsiveness, Doppler estimation of incremental changes in stroke volume was no better than inferior vena caval diameter or trans-aortic Doppler respiratory variation [3]. Finally, caution should be exercised about extrapolating information on the ‘rightness’ or ‘wrongness’ of fluid boluses from observational studies on healthy volunteers, in whom haemodynamic assessment of fluid responsiveness is irrelevant compared with patients having a major operation, in whom fluid status may be adversely affected pre-operatively by their cancer, inflammatory bowel disease or cardiovascular co-morbidities and intra-operatively by fluid shifts, and in whom surgery demands an increase in oxygen delivery to match increased increased oxygen demand. Hypovolaemia is better thought of in terms of outcomes, and the concept of ‘optimal fluid therapy’ has been an important development in modern peri-operative medicine [4]. The presence of ‘functional hypovolaemia’ is better demonstrated if there is a reduction in morbidity following administration of additional fluid volumes. Quite

© 2014 The Association of Anaesthetists of Great Britain and Ireland

rightly, there is some debate about the benefits of goal-directed therapy in elective colorectal surgery in light of recent studies that have failed to reproduce the outcome benefits found historically [5, 6]. In this context, an observational study demonstrating fluid responsiveness to a passive leg raise manoeuvre in healthy volunteers is of limited significance. Furthermore, it should not be assumed that the use of echocardiography to derive stroke volume increases the study’s validity or precision. T. D. Phan V. Nguyen B. Cowie St Vincent’s Hospital Melbourne, Victoria, Australia Email: [email protected] No external funding and no competing interests declared. Previously posted on the Anaesthesia correspondence website: www.anaesthesia correspondence.com.

References 1. Godfrey GEP, Dubrey SW, Handy JM. A prospective observational study of stroke volume responsiveness to a passive leg raise manoeuvre in healthy non-starved volunteers as assessed by transthoracic echocardiography. Anaesthesia 2014; 69: 306–13. 2. Robson SC, Murray A, Peart I, Heads A, Hunter S. Reproducibility of cardiac output measurement by cross sectional and Doppler echocardiography. British Heart Journal 1988; 59: 680–4. 3. Mandeville JC, Colebourn CL. Can transthoracic echocardiography be used to predict fluid responsiveness in the critically ill patient? A systematic review. Critical Care Research and Practice 2012; 2012: 1–9. 4. Bellamy MC. Wet, dry or something else? British Journal of Anaesthesia 2006; 97: 755–7. 5. Brandstrup B, Svendsen PE, Rasmussen M, Belhage B, Rodt S A, Hansen B, et al. 789

Anaesthesia 2014, 69, 785–798 Which goal for fluid therapy during colorectal surgery is followed by the best outcome: near-maximal stroke volume or zero fluid balance? British Journal of Anaesthesia 2012; 109: 191–9. 6. Srinivasa S, Taylor MHG, Singh PP, Yu T-C, Soop M, Hill AG. Randomized clinical trial of goal-directed fluid therapy within an enhanced recovery protocol for elective colectomy. British Journal of Surgery 2013; 100: 66–74. doi:10.1111/anae.12758

A reply We thank Drs Phan, Ngui and Cowie for their comments on our study [1]. We agree entirely with many of the study limitations they highlight and believe that we reflected on these in our article. However, it is possible that they have misunderstood the context in which the study was performed. In the UK, the paradigm of stroke volume-guided fluid therapy has been imposed on anaesthetists, with financial incentives being awarded for ‘compliance’, an issue eruditely illuminated by Drs Stevenson and Stoker in their letter and with whom we entirely agree. While we recognise that caution must be applied when extrapolating to patient information gleaned from healthy volunteers, such healthy volunteer studies have been used to provide the physiological justification for stroke volume-guided fluid therapy; we believe our findings counter the previously cited evidence and reinforce that the ‘real life’ situation is far more complex. To study a specific physiological state (i.e. normovolaemia) requires a specific model with as few confounding/influencing factors as possible. To attempt to address such questions in a 790

Correspondence

patient population introduces numerous variables that cannot be adequately controlled and thus the results become more confusing and even easier to criticise. The use of healthy volunteers provides a ‘clean’ methodological environment in which to ask such specific questions, but with the caveat that the ‘answers’ are more difficult to extrapolate. While Dr. Phan and colleagues correctly point out that we used intra-individual measurements to calculate sample size, we also used published data from critically ill patients (i.e. inter-individual measurements) in our calculations. Power calculations are always a challenge for pilot studies where existing data are not available; however, such calculations are essential in order to gain Research Ethics Committee approval in the UK. We believe that our use of published intra- and inter-individual data, together with our own intra-individual pilot data, was the most rigorous approach to estimating sample size. While it is possible that our findings occurred by chance (which we recognise in our article), we have stressed that our findings remain clinically significant, as variations such as this could occur when managing patients. We fully appreciate that echocardiography has limitations in interpreting stroke volume values; however, this is the case with all available stroke volume/cardiac output monitors, further emphasising that using such monitors as the sole guide to fluid management may not be the panacea suggested in current recommendations. We used transthoracic echocardiography as it

was the most accurate and pragmatic tool to measure stroke volume in a cohort of volunteers and, as their citation suggests [2], is highly effective at detecting volume responsiveness. It would seem that Dr. Phan and colleagues fundamentally share our belief that assessing and managing ‘fluid status’ is complex, and we propose that the current recommendations guiding such therapy should be reviewed. We hope that our work will trigger further study into the physiological and pathophysiological volaemic states encountered clinically and manipulated with intravenous fluid therapy. J. M. Handy Chelsea and Westminster Hospital, London, UK Email: [email protected] G. E. P. Godfrey S. Dubrey Hillingdon Hospital, London, UK No external funding and no competing interests declared. Previously posted on the Anaesthesia correspondence website: www.anaesthesia correspondence.com.

References 1. Godfrey GEP, Dubrey SW, Handy JM. A prospective observational study of stroke volume responsiveness to a passive leg raise manoeuvre in healthy non-starved volunteers as assessed by transthoracic echocardiography. Anaesthesia 2014; 69: 306–13. 2. Mandeville JC, Colebourn CL. Can transthoracic echocardiography be used to predict fluid responsiveness in the critically ill patient? A systematic review. Critical Care Research and Practice 2012; 2012: 1–9. doi:10.1111/anae.12762

© 2014 The Association of Anaesthetists of Great Britain and Ireland