Respiratory stroke volume variation assessed by oesophageal Doppler monitoring predicts fluid responsiveness during laparoscopy

Transcription

1 British Journal of Anaesthesia 112 (4): (2014) Advance Access publication 22 December doi: /bja/aet430 CARDIOVASCULAR Respiratory stroke volume variation assessed by oesophageal Doppler monitoring predicts fluid responsiveness during laparoscopy P.-G. Guinot 1 *, B. de Broca 1, E. Bernard 1, O. Abou Arab 1, E. Lorne 1,2 and H. Dupont 1,2 1 Anaesthesiology and Critical Care Department, Amiens University Hospital, Place Victor Pauchet, Amiens 80054, France 2 INSERM ERI12, Jules Verne University of Picardy, 12 rue des Louvels, Amiens 80000, France * Corresponding author. guinotpierregregoire@gmail.com Editor s key points This study demonstrates that stroke volume respiratory variation measured byoesophageal Doppler monitoring accurately predicts fluid responsiveness during laparoscopy with a grey zone ranging between 13 and 15%. The respiratory variation of the peak velocity was less predictive of fluid The corrected flow time was unable to accurately predict fluid Background. This study was designed to assess the ability of the stroke volume respiratory variation (DrespSV) determined by oesophageal Doppler monitoring (ODM) to predict the response to volume expansion (VE) during pneumoperitoneum. The predictive value of DrespSV was evaluated according to the concept of the grey zone. Methods. Patients operated on laparoscopy and monitored by ODM were prospectively included. The exclusion criteria were frequent ectopic beats or preoperative arrhythmia, right ventricular failure, and spontaneous breathing. Haemodynamic parameters and oesophageal Doppler indices [stroke volume (SV), peak velocity (PV), cardiac output (CO), corrected flow time (FTc), respiratory variation of PV (DrespPV) and SV (DrespSV)] were collected before and after VE. Responders were defined as a 15% increase in SV after VE. Results. Thirty-eight (64%) of the 59 patients were responders. A cut-off of.14% DrespSV predicted fluid responsiveness with an area under the ROC curve (AUC) of 0.92 [95% confidence interval (CI): , P,0.0001]. The grey zone of DrespSV ranged between 13 and 15%. With an AUC of 0.71 (95% CI: , P¼0.005), DrespPV fairly accurately predicted fluid FTc was unable to accurately predict fluid Conclusions. DrespSV and DrespPV predicted fluid responsiveness during laparoscopy under strict physiological conditions. FTc was not predictive of fluid responsiveness during laparoscopy. Keywords: anaesthesia; Doppler, intraoperative; laparoscopy; monitoring; stroke volume Accepted for publication: 6 October 2013 Oesophageal Doppler monitoring (ODM) allows non-invasive continuous monitoring of cardiac output (CO) during surgery. 1 Dynamic preload indicators based on interactions between circulatory and respiratory functions under positive mechanical ventilation, such as respiratory variation of stroke volume (DrespSV) and respiratory variation of pulse pressure (DrespPP), predict fluid responsiveness during surgery and in intensive care unit (ICU). 2 5 We have recently shown the ability of DrespSV measured by ODM to predict fluid responsiveness in a mixed population of surgical patients. 6 However, some studies have highlighted the effect of intra-abdominal pressure (IAP) on the reliability of these indices. 7 9 Elevation of IAP up to 30 mm Hg can affect the accuracy and cut-off values of dynamic preload indices. 8 9 In the surgical setting, these results may be affected by the fact that IAP is maintained lower than values usually observed during intra-abdominal hypertension. A recent study investigated the effect of pneumoperitoneum during laparoscopy on dynamic preload indices. 7 Using an uncalibrated pressure waveform device, these authors established that DrespPP and DrespSVdid not change during laparoscopy, but had a relatively poor capacity to predict fluid However, haemodynamic changes associated with increased IAP may have affected the validity of the CO monitor used in this study. 10 To the best of our knowledge, ODM respiratory indices have not been studied during laparoscopy. We hypothesized that DrespSV measured by ODM would be a good indicator of fluid responsiveness during laparoscopy with pneumoperitoneum. The primary objective of this study was to demonstrate that DrespSV monitored by ODM can accurately predict fluid responsiveness during laparoscopy. We also assessed the capacity of DrespPV and corrected flow time (FTc) to predict fluid & The Author [2013]. Published by Oxford University Press on behalf of the British Journal of Anaesthesia. All rights reserved. For Permissions, please journals.permissions@oup.com

2 Laparoscopy does not modify the predictability of stroke volume variation BJA Methods Ethics This study was approved by the Institutional Review Board(IRB) for human subjects. Informed consent was waived, as the IRB considered the protocol to be part of routine clinical practice. Patients A prospective, observational study was conducted in Amiens University Hospital. Patients over the age of 18 yr monitored by ODM during laparoscopic surgery in whom the anaesthetist decided to perform volume expansion (VE) were included. Indications for VE were: optimization of CO, arterial hypotension, or haemorrhage. Patients with frequent ectopic beats or preoperative arrhythmia, right ventricular dysfunction, spontaneous ventilation, and contraindications to ODM probe insertion were excluded. Anaesthesia Each patient was monitored by pulse oximetry, non-invasive arterial pressure monitoring, and 3-lead electrocardiogram and underwent general anaesthesia. All patients received a tracheal intubation and were ventilated in volume-controlled mode. The choice of drugs was left to the anaesthetist s discretion and comprised either propofol or etomidate and remifentanil or sufentanil. Anaesthesia was maintained with either an inhaled hypnotic (desflurane or sevofurane) or propofol and the same opioid used for induction. Neuromuscular block was systematically induced by rocuronium (0.6 mg kg 21 )or cisatracurium (0.15 mg kg 21 ). Tidal volume was adjusted to ideal body weight to obtain 7 9 ml kg 21 and the ventilatory rate was adapted to maintain end-tidal CO 2 at kpa; a positive end-expiratory pressure (PEEP) of kpa was applied. Measurements The ventilator parameters (end-expiratory pressure, plateau pressure, and tidal volume) were recorded at baseline. Pneumoperitoneum using insufflated intra-abdominal carbon dioxide was maintained with a preset IAP of,15 mm Hg. The IAP was the average of three measured IAP values. Oesophageal Doppler monitoring The oesophageal Doppler probe (CardioQ TM, Deltex Medical, Gamida, France) was positioned to obtain the optimum signal for descending aorta blood velocity. To avoid artifacts related to the precise distinction of the beginning and end of aortic flow with each ventricular beat that may be distorted by wall thump and run-off, respectively, laminar flow was ensured with a narrow frequency range (blunt velocity profile). Stroke volume (SV), FTc, and peak velocity (PV) were recorded continuously by the ODM software (beat by beat) from aortic blood flow velocity, and their mean values were calculated over 10 s. Respiratory variations (Dresp) of ODM values were obtained as previously described, regardless of the respiratory cycle. 6 The respiratory variation of SV (DrespSV) was calculated as DrespSV¼{(SV max 2SV min )/[(SV max +SV min )/ 2]} 100, where SV min and SV max are the minimum and maximum SV values over one respiratory cycle, respectively. The same method was used to calculate the respiratory variation of PV (DrespPV). All values represented the mean of three measurements. All measurements were analysed off-line using a video sequence of the monitor. ODM is routinely used to monitor surgical patients in our centre with good interobserver variability. 7 Study protocol Each patient was included after intra-abdominal insufflation and stabilization of haemodynamic parameters in the absence of any drug injection or changes in ventilatory parameters. The first VE consisted of infusion of 500 ml Ringer lactate over 10 min and was the only volume challenge recorded for the study. Two sets of measurements [IAP, diastolic arterial pressure (DAP), mean arterial pressure (MAP), systolic arterial pressure (SAP), heart rate (HR), DrespPV, DrespSV, PV, FTc, and SV] were performed before and immediately after the fluid challenge. Statistics At least, 40 patients would be sufficient to demonstrate that DrespSV can predict fluid responsiveness with an area under the ROC curve (AUC) of.0.90, for a power of 80%, an alpha risk of 0.05, and a beta risk of 0.2. The distribution of variables was assessed using the D Agostino Pearson test. Taking into account the exclusion criteria and possible loss of ODM signal during pneumoperitoneum, 61 patients were recruited over a 6-month period. Data are expressed as proportion (percentage) or mean [standard deviation (SD)], as appropriate. Nonresponders and responders were defined by the SV variation (expressed as a percentage) after VE. A positive response wasdefinedasa 15% increase in SV. 56 A Student s paired t-test was used to compare within-group changes in haemodynamic variables. Differences between responders and nonresponders were compared by a Student s t-test. Linear correlations were tested by the Pearson-rank method. A receiveroperating characteristic curve (ROC) was established for DrespSV, FTc, and DrespPV. ROC curves were generated by averaging 1000 bootstrapped samples (sampling with replacement) from the original study population. The test previously described by DeLong and colleagues. 11 was used to compare AUC for each variable. The predictive value of DrespSV was evaluated by using a grey zone approach, as previously described. 512 The grey zone approach indicated two cut-offs between which the prediction of fluid responsiveness remained uncertain. In these cases, the physician must confirm fluid responsiveness by additional information. Differences with a P-value of,0.05 were considered statistically significant. IBM w SPSS w Statistics 18 (IBM) was used to perform statistical analysis. Results Sixty-one patients undergoing laparoscopic surgery were included. Two were excluded because of ODM failure. No 661

3 BJA Guinot et al. patients were treated by vasopressor, and no patients presented arrhythmia during the study period. Finally, 59 patients were included in the study. Indications for surgery were gynaecological cancer surgery (n¼27), visceral surgery (n¼23), radical prostatectomy (n¼7), and nephrectomy (n¼2). Thirty-eight of the 59 patients (64%) were considered to be responders. Patient characteristics are described in Table 1. Baseline CO and SV were lower and DrespSV and DrespPV were higher in responders than in non-responders (Table 2). VE increased SAP, PV, SV and FTc and decreased DrespSV and DrespPV only in the responder group (Table 2). DrespSV and DrespPV were correlated (r¼0.76, P¼0.001). With an AUC of 0.92 [95% confidence interval (CI): , P,0.0001] (Fig. 1), DrespSV presented an excellent ability to predict fluid A cut-off value of 14% yielded a sensitivity of 92% (95% CI: 79 98), a specificity of 87% (95% CI: 64 97), a positive predictive value (PPV) of 92% (95% CI: 78 98), a negative predictive value (NPV) of 86% (95% CI: 63 71), a positive likelihood ratio of 6.45 (95% CI: ), and a negative likelihood ratio of 0.09 (95% CI: ) (Fig. 1). The two methods determined a grey zone ranging between 13 and 15%. The predictability of DrespPV was fair with an AUC of 0.71 (95% CI: , P¼0.005), a sensitivity of 93% (95% CI: 78 98), a specificity of 44% (95% CI: 22 69), a PPV of 74% (95% CI: 57 87), a negative predictive value of 80% (95% CI: 42 98), a positive likelihood ratio of 1.68 (95% CI: ), and a negative likelihood ratio of 0.15 (95% CI: ). With an AUC of 0.53 (95% CI: , P¼0.77), FTc was not predictive of fluid The AUC of DrespSV was greater than the AUC of DrespPV to predict fluid responsiveness (P,0.05). Table 1 Patient characteristics of the study population. Values are expressed as mean (SDs) or number (%). BMI, body mass index Age [mean (SD), yr] 51 (16) Gender (female/male) 42/17 Height [mean (SD), cm] 166 (8) BMI [mean (SD), kg m 22 ] 29 (5) Type of surgery, n (%) Gynecology 27 (46) Digestive 23 (39) Urology 9 (15) ASA, n (%) I 7 (12) II 40 (68) III 12 (20) IV 0 Respiratory parameters Tidal volume [mean (SD), ml kg 21 of predicted body 8 (1) weight] Respiratory rate [mean (SD), min 21 ] 13 (1) Plateau pressure (kpa) 1.96 (0.3) IAP [mean (SD), mm Hg] 12 (1) Discussion This study demonstrates that DrespSV measured by ODM accurately predicts fluid responsiveness during laparoscopy with a grey zone ranging between 13 and 15%. DrespPV was less predictive of fluid In contrast, FTc was unable to accurately predict fluid Increased IAP during laparoscopy is known to induce circulatory changes Increased IAP leads to redistribution of abdominal venous volume towards the thoracic compartment, potentially resulting in improved SV, CO, and MAP. Simultaneously on the arterial side, IAP increases systemic vascular resistance (SVR) and MAP, potentially resulting in decreased SV and CO Decreased CO and SV have been related to increased afterload. 16 These effects depend on the level of IAP and the patient s preload status. Studies evaluating the Table 2 Cardiovascular variables in responders and non-responders expressed as mean (SD). HR, heart rate; SAP, systolic arterial pressure; DAP, diastolic arterial pressure; MAP, mean arterial pressure; FTc, corrected flow time; SV, stroke volume; PV, peak velocity; CO, cardiac output; DrespSV, respiratory stroke volume variation; DrespPV, respiratory peak velocity variation. *P,0.05 within groups; P,0.05 between groups Baseline HR (beats min 21 ) Responders 70 (17) 69 (15) Non-responders 63 (15) 63 (14) SAP (mm Hg) Responders 107 (18) 117 (18)* Non-responders 108 (19) 114 (19) DAP (mm Hg) Responders 68 (13) 70 (13) Non-responders 61 (14) 65 (12) MAP (mm Hg) Responders 81 (14) 86 (14) Non-responders 76 (14) 81 (13) FTc (ms) Responders 332 (64) 386 (45)* Non-responders 352 (64) 363 (48) PV (cm s 21 ) Responders 72 (25) 86 (36)* Non-responders 85 (21) 84 (23) SV (ml) Responders 68 (16) 90 (19)* Non-responders 95 (16) 97 (19) CO (ml min 21 ) Responders 4.8 (1.2) 6.2 (1.9)* Non-responders 5.7 (1.5) 6 (1.5) DrespSV (%) Responders 21 (7) 13 (9)* Non-responders 10 (5) 9 (3) DrespPV (%) Responders 13 (7) 10 (5)* Non-responders 9 (5) 7 (5) VE 662

4 Laparoscopy does not modify the predictability of stroke volume variation BJA Average true positive rate False positive rate DrespSV: 0.92 ( ) DrespPV: 0.71 ( ) FTc: 0.53 ( ) Fig 1 Average receiver operator characteristics (ROC) curves of FTc, DrespSV, and DrespPV to discriminate responders and non-responders to VE. Area under ROC appears in cartouche with 95% CI. ability of respiratory-derived indices to predict response to VE have reported discordant results. Some authors have reported no changes in response to intra-abdominal hypertension, while others have reported an alteration of the predictability cut-off values of these changes, or both These conflicting results are not surprising and can be explained by two main factors: the level of IAP that alters cardiopulmonary interactions (and derived indices) and the device used to assess these dynamic preload indices. In contrast to intra-abdominal hypertension, IAP during laparoscopy may explain the absence of modification of cut-off values in the study by Hoiseth and colleagues 7 and in our study. Elevation of IAP progressively increases DrespPP and DrespSV independently of cardiac preload changes. Cyclic circulatory changes induced by positive-pressure ventilation are dependent on IAP. Duperret and colleagues 18 conducted a study evaluating the effects of progressively increasing IAP (from 0 to 30 mm Hg) on respiratory-derived indices and concluded that cyclic circulatory changes in DrespSV increased with increasing IAP. These changes were statistically significant for IAP higher than 20 mm Hg. 18 These results are consistent with the results reported by Bliacheriene and colleagues 17 and Hoiseth and colleagues, 7 who did not find any modification of cut-off values for dynamic preload indices. In these studies, IAP was,15 mm Hg, while IAP reached 30 mm Hg in the studies by Renner and colleagues 8 and Jacques and colleagues. 9 Unlike Hoiseth and colleagues 7 and Renner and colleagues, 8 we found a good predictabilityof DrespSVduring pneumoperitoneum. An uncalibrated pressure waveform device (FloTrac 1 Vigileo w, Edwards Lifesciences, Irvine, CA, USA) or calibrated pulse contour analysis (PiCCO Plus w, Pulsion Medical Systems, Munich, Germany) may not accurately measure SV and SV changes with respiration (i.e. DrespSV). Haemodynamic changes during laparoscopy are partly attributable to activation of the neurohumoral vasoactive system [antidiuretic hormone (ADH), sympathetic and renin-angiotensin-aldosterone systems (RAAS)] induced by carbon dioxide insufflation. 19 These changes may lead to increased SVR with modification of arterial mechanical properties. 13 The ability of uncalibrated and calibrated pressure waveform devices to accurately measure SV is influenced by SVR When comparing DrespSV measured by a calibrated pulse contour device and DrespSV derived from an ultrasonic flow probe on the aorta, Kubitz and colleagues 20 demonstrated that the limits of agreement widened with increasing SVR. Similarly, Suehiro and colleagues 10 recently demonstrated the influence of SVR on the ability of the Vigileo-Flotrac system to accurately measure CO. SV measured by ODM may be less influenced by SVR than SV measured by devices using peripheral pressure waveform. DrespPV was found to be less accurate than DrespSV. SV is assessed by integration of aortic blood velocity over left ventricular ejection time, whereas the software device automatically measures the peak value of aortic blood flow. Measurement of PV may not be equivalent to measurement of SV. Unlike respiratory changes in aortic blood velocity, DrespPV may not accurately reflect DrespSV, as DrespPV may vary in different proportions from DrespSV, which may explain the discordant results obtained for these two indices. FTc did not predict haemodynamic response to VE. FTc is a complex static indicator influenced by several factors, such as preload, afterload, and inotropic state As described above, circulatory changes during laparoscopy may alter FTc and its ability to predict fluid Although FTc cannot predict response to VE during laparoscopy, it can be used to evaluate the effect of the treatments administered or can be integrated as a limit to optimize CO while avoiding excessive fluid loading. This study has several limitations. As the mean IAP was,15 mm Hg, our results cannot be extrapolated to higher IAP values. In clinical practice, ODM could be affected by various artifacts (electric scalpels, position changes, etc.) and may sometimes require repositioning of the probe. This disadvantage rarely limits bedside use of ODM. Another limitation is that this was an observational study that did not compare the accuracy of dynamic indicators measured by several devices. The incidence of fluid responsiveness depends on the indication for VE, and may affect the sensitivity/specificity analysis. Most of the previous studies evaluating fluid responsiveness in the operating theatre reported an incidence of fluid responsiveness similar to that observed in the present study. 2 4 As DrespSV measured by ODM was not compared with DrespSV measured by pressure waveform devices, no conclusions can be drawn concerning the superiority of ODM or pressure waveform devices. The primary objective of this study was to evaluate the accuracy of DrespSV measured by ODM to predict fluid responsiveness during laparoscopy. Like 663

5 BJA Guinot et al. all respiratory-derived indices, DrespSV is dependent on strict measuring conditions. As patients with cardiac arrhythmia, right ventricular failure, spontaneous ventilation, and altered lung compliance were excluded from this study, and tidal volume was set between 7 and 9 ml kg 21, our results cannot be extrapolated to patients not satisfying these criteria. In conclusion, this study demonstrated that DrespSV monitored by ODM is a reliable indicator of fluid responsiveness during laparoscopy with pneumoperitoneum. The ability of DrespSV to predict fluid responsiveness was not affected by pneumoperitoneum with a target IAP of,15 mm Hg. The grey zone ranged between 13 and 15%. In contrast, DrespPV was unable to accurately predict fluid responsiveness in these conditions and FTc was not predictive. Authors contributions P.-G.G. conceived, designed and coordinated the study, and drafted the manuscript. E.B., B.d.B., and O.A.A. participated in coordination of the study. E.L. and H.D. participated in coordination of the study and helped to draft the manuscript. Declaration of interest None declared. References 1 Mark JB, Steinbrook RA, Gugino LD, et al. Continuous noninvasive monitoring of cardiac output with esophageal Doppler ultrasound during cardiac surgery. Anesth Analg 1986; 65: Derichard A, Robin E, Tavernier B, et al. Automated pulse pressure and stroke volume variations from radial artery: evaluation during major abdominal surgery. Br J Anaesth 2009; 103: Michard F, Teboul JL. Using heart-lung interactions to assess fluid responsiveness during mechanical ventilation. Crit Care 2000; 4: Zhang Z, Lu B, Sheng X, Jin N. Accuracy of stroke volume variation in predicting fluid responsiveness: a systematic review and meta-analysis. J Anesth 2011; 25: Cannesson M, Le Manach Y, Hofer CK, et al. Assessing the diagnostic accuracy of pulse pressure variations for the prediction of fluid responsiveness: a gray zone approach. Anesthesiology 2011; 115: Guinot PG, de Broca B, Abou Arab O, et al. Ability of stroke volume variation measured by oesophageal Doppler monitoring to predict fluid responsiveness during surgery. Br J Anaesth 2013; 110: Hoiseth LO, Hoff IE, Myre K, Landsverk SA, Kirkebøen KA. Dynamic variables of fluid responsiveness during pneumoperitoneum and laparoscopic surgery. Acta Anaesthesiol Scand 2012; 56: Renner J, Gruenewald M, Quaden R, et al. Influence of increased intra-abdominal pressure on fluid responsiveness predicted by pulse pressure variation and stroke volume variation in a porcine model. Crit Care Med 2009; 37: Jacques D, Bendjelid K, Duperret S, Colling J, Piriou V, Viale JP. Pulse pressure variation and stroke volume variation during increased intra-abdominal pressure: an experimental study. Crit Care 2011; 15: R33 10 Suehiro K, Tanaka K, Matsura T, Mori T, Nishikawa K. Systemic vascular resistance has an impact on the reliability of the Vigileo-Flotrac system in measuring cardiac output and tracking cardiac output. Br J Anaesth 2013; 111: DeLong ER, DeLong DM, Clarke-Pearson DL. Comparing the areas under two or more correlated receiver operating characteristic curves: a nonparametric approach. Biometrics 1988; 44: Ray P, Le Manach Y, Riou B, Houle TT. Statistical evaluation of a biomarker. Anesthesiology 2010; 112: Joris JL, Noirot DP, Legrand MJ, Jacquet NJ, Lamy ML. Hemodynamic changes during laparoscopic cholecystectomy. Anesth Analg 1993; 76: Harris SN, Ballantyne GH, Luther MA, Perrino AC Jr. Alterations of cardiovascular performance during laparoscopic colectomy: a combined hemodynamic and echocardiographic analysis. Anesth Analg 1996; 83: Vivier E, Metton O, Piriou V, et al. Effects of increased intraabdominal pressure on central circulation. Br J Anaesth 2006; 96: Alfonsini P, Vieillard-Baron A, Coggia M, et al. Cardiac function during intraperitoneal CO 2 insufflation for aortic surgery: a transesophageal echocardiographic study. Anesth Analg 2006; 102: Bliacheriene F, Machado SB, Fonseca EB, Otsuke D, Auler JO Jr, Michard F. Pulse pressure variation as a tool to detect hypovolemia during pneumoperitoneum. Acta Anaesthesiol Scand 2007; 51: Duperret S, Lhuillier F, Piriou V, et al. Increased intra-abdominal pressure affects respiratory variations in arterial pressure in normovolemic and hypovolaemic mechanically ventilated healthy pigs. Intensive Care Med 2007; 33: Koivusalo AM, Lindgren L. Effects of carbon dioxide pneumoperitoneum for laparoscopic cholecystectomy. Acta Anaesthesiol Scand 2000; 44: Kubitz JC, Annecke T, Forkl S, et al. Validation of pulse contour derived stroke volume variation during modifications of cardiac afterload. Br J Anaesth 2007; 98: SingerM, Allen MJ, Webb AR, Bennett ED. Effects of alterations in left ventricular filling, contractility, and systemic vascular resistance on the ascending aortic blood velocity waveform of normal subjects. Crit Care Med 1991; 19: Singer M, Bennett ED. Noninvasive optimization of left ventricularfilling using esophageal Doppler. Crit Care Med 1991; 19: Sinclair S, James S, Singer M. Intraoperative intravascular volume optimisation and length of hospital stay after repair of proximal femoral fracture: randomized controlled trial. Br Med J 1997; 315: Handling editor: M. M. R. F. Struys 664