Resuscitaion global vs_regional


Published on

Published in: Health & Medicine, Business
  • Be the first to comment

  • Be the first to like this

No Downloads
Total views
On SlideShare
From Embeds
Number of Embeds
Embeds 0
No embeds

No notes for slide

Resuscitaion global vs_regional

  1. 1. Monitoring global volume-related hemodynamic or regionalvariables after initial resuscitation: What is a better predictor ofoutcome in critically ill septic patients?*Martijn Poeze, MD, PhD; Barbara C. J. Solberg; Jan Willem M. Greve, MD, PhD;Graham Ramsay, MD, PhD, FRCS Objective: Regional variables of organ dysfunction are thought dioxide pressure, gastric mucosal pH, mucosal-end tidal PCO2to be better monitoring variables than global pressure-related gap, indocyanine green blood clearance, indocyanine greenhemodynamic variables. Whether a difference exists between plasma clearance, and plasma disappearance rate. Multivari-regional and global volume-related variables in critically ill pa- ate analysis identified lactate, gastric mucosal pH, mucosal-tients after resuscitation is unknown. end tidal PCO2 gap, mucosal-arterial PCO2 gap, indocyanine Design: Prospective diagnostic test evaluation. green plasma clearance, and plasma disappearance rate of dye Setting: University-affiliated mixed intensive care unit. as nondependent predictors of outcome. Patients who subse- Patients: Twenty-eight critically ill patients. quently died had a significantly lower gastric mucosal pH, Interventions: Using standardized resuscitation, hemodynamic higher intramucosal carbon dioxide pressure and mucosal-endoptimization was targeted at mean arterial pressure, heart rate, tidal PCO2 gap, and lower indocyanine green blood clearance,occlusion pressure, cardiac output, systemic vascular resistance, indocyanine green plasma clearance, plasma disappearanceand urine output. Primary outcome variable was in-hospital mor- rate, and right ventricular end-diastolic volume index, of whichtality. gastric mucosal pH, mucosal-end tidal PCO2 gap, and indocya- Measurements and Main Results: During resuscitation,global volume-related hemodynamic variables were measured nine green blood clearance were the most important predictorssimultaneously and compared with regional variables. At ad- of outcome.mission no variable was superior as a predictor of outcome. Conclusions: Initial resuscitation of critically ill patientsDuring resuscitation, significant changes were seen in mean with shock does not require monitoring of regional variables.arterial pressure, central venous pressure, oxygen delivery, After stabilization, however, regional variables are the bestsystemic vascular resistance, total blood volume, right heart predictors of outcome. (Crit Care Med 2005; 33:2494–2500)and ventricle end-diastolic volume, right ventricle ejection KEY WORDS: sepsis; monitoring; predictor; outcome; tonometry;fraction, right and left stroke work index, intramucosal carbon indocyanine greenT he ability to recognize the fea- tent of shock has been identified as an Therefore, there is a need for other tures of shock is crucial in the important contributory factor leading to indicators of tissue oxygenation in pa- management of critically ill a considerable number of “preventable” tients with shock, which are more sensi- patients. In trauma patients, deaths (1). The standard resuscitation tive and specific than the global pressure-failure to recognize the presence or ex- during therapy for shock is aimed at cor- related variables. Two techniques have recting global pressure-related variables. been introduced that may be able to de- However, correcting these variables pro- tect occult hypovolemia. *See also p. 2691. duces variable outcomes in critically ill It has been suggested that measuring From the Department of Surgery (MP, JWMG, GR) patients. There are several possible rea- regional variables of splanchnic perfusionand Intensive Care Medicine (BCJS), University Hospi- sons for this variability. The recognition is a better predictor of the presence oftal Maastricht, The Netherlands. of shock may be hampered by the pa- uncompensated shock than markers of Supported, in part, by the University Hospital tients’ compensatory capacities. A patient global perfusion (5, 6). Splanchnic perfu-Maastricht Research Fund, Maastricht, The Nether-lands. may be in shock despite having a normal sion seems to play an important role dur- There are no financial disclosure for any of the heart rate and blood pressure. Moreover, ing shock and resuscitation. Recent stud-authors. the therapy used to treat shock can nor- ies suggest the presence of a disturbed Address requests for reprints to: M. Poeze, MD, malize variables by which shock is as- splanchnic circulation in apparentlyPhD, Department of Surgery, University Hospital Maas-tricht, P. Debyelaan 25, NL-6202 AZ Maastricht, The sessed clinically, even though defective compensated shock (7–9). Critically illNetherlands. tissue oxygenation may still exist (2– 4). patients with a persistently inadequate Copyright © 2005 by the Society of Critical Care Global pressure-related variables may be splanchnic perfusion are at increased riskMedicine and Lippincott Williams & Wilkins too inaccurate to detect these abnormal- of developing multiple organ failure dur- DOI: 10.1097/01.CCM.0000185642.33586.9D ities. ing their intensive care stay (10, 11).2494 Crit Care Med 2005 Vol. 33, No. 11
  2. 2. Another option is measuring the Society of Critical Care Medicine consensus the greatest volume available, as has been de-global volume-related variables (such as conference criteria were included (17). In the scribed in detail (19). The following variablesextravascular lung water volume or in- presence of clinically suspected or proven in- were determined: right heart- and right ven-trathoracic blood volume) instead of fection, patients were defined as having sepsis. tricular end-diastolic volume index, right ven- A clinical suspicion of infection was defined as tricular ejection fraction (RVEF), right ven-pressure-related variables. A considerable the presence of abnormalities on radiograph of tricular end-systolic volume index, rightnumber of studies have indicated the lack ventricular stroke work index, left heart end-of relation between the degree of hypovo- the thorax consistent with pulmonary infec- tion, peritonitis confirmed at surgery or clin- diastolic volume index, global end-diastoliclemia and the pressure-related variables volume index, extravascular lung water index, ical examination, or positive Gram-negative(12, 13). Instead, the use of volume- intrathoracic blood volume index, total blood stain for infection. All patients were ventilated.related global hemodynamic variables is volume index, and pulmonary blood volume Patients were included if their Acute Physiol-thought to indicate the presence of hypo- index. ogy and Chronic Health Evaluation (APACHE)volemia (14 –16). Regional Hemodynamic Function Vari- II score was 12 (18). ables. ICG clearance measurements (ICG In view of these observations, we car- Patients with sepsis-induced hypotensionried out a prospective trial in which crit- blood clearance [CBI], ICG plasma clearance or severe sepsis who were expected to die [CPI], and plasma disappearance rate of dyeically ill septic patients with or without within the first 24 hrs were not included. [PDR]) were performed using the ICG bolusovert clinical shock underwent standard- Patients on dialysis at admission were also infusion method. The PDR rate was deter-ized resuscitation to normalize their excluded from participation. mined from the linear elimination curve ofglobal pressure-related hemodynamics. ICG through the liver. CBI and CPI were cal-The responses of both the global volume- culated by multiplying the PDR by the TBVrelated hemodynamic and the regional Measurements of Systemic and (and [1 hematocrit] for the CPI).variables to this resuscitation were eval- Regional Variables Gastric tonometry measurements (PrCO2,uated in terms of their predictive value mucosal-end tidal PCO2 gap [Pr-etCO2-gap], andfor mortality. Our hypothesis was that After inclusion, patients received a 3-Fr mucosal-arterial PCO2 gap) were measured atresuscitation in the intensive care unit thermistor-tipped fiberoptic catheter (PV 10-min intervals using gas-automated capnog- 2024, Pulsion Medical Systems, Munich, Ger- raphy (Tonocap, Datex-Ohmeda) (20). The(ICU) could be aimed at the global vol- many) placed into the descending aorta via a PCO2 and pH values of the blood gases wereume-related variables, but that after nor- corrected for the central blood temperature 4-Fr introducing sheath (Arrow, Reading, PA)malization, optimal resuscitation would measurements, using the formulas provided in the femoral artery. The catheter was thenonly be achieved using regional variables. by the manufacturer (ABL 100, Radiometer, connected to a bedside monitor (COLD Z-021, Pulsion Medical Systems, Munich, Germany). Kopenhagen).PATIENTS AND METHODS In addition, a pulmonary artery thermodilu- tion catheter (7.5-Fr, Baxter) was introduced Resuscitation ProtocolStudy Design and also connected to the COLD bedside mon- itor. A gastric tonometry catheter (14-Fr, Da- Resuscitation was aimed at improving This study was conducted as a prospective tex Ohmeda, Finland) was introduced for mea- global pressure-related hemodynamics asevaluation of the predictive value of global surement of intramucosal carbon dioxide judged by the intensivist on call: in general,pressure- and volume-related hemodynamic pressure (PrCO2) using the gas-automated circulatory support was titrated to increaseand regional variables of splanchnic perfusion capnograph (Tonocap TC-200, Datex-Ohmeda, the MAP 70 mm Hg, systemic vascular re-during the intensive care resuscitation of crit- Finland). Gastric mucosal pH (pHi) was calcu- sistance 1100 dyne/sec/cm5, CI 2.5 L/min/ically ill septic patients. In-hospital mortality lated using the standard Henderson-Hassel- m2, and urine output 0.5 mL/kg/hr. If thesewas used as the primary end point. bach formula (19). At admission to the ICU, criteria were not met and PAOP was 18 mm The study protocol prospectively defined several variables were obtained to evaluate dis- Hg, patients received a fluid challenge of 500that the effects of the resuscitation should be turbances in hemodynamics and organ func- mL or 1000 mL (depending on blood pressure)analyzed according to the prognostic values. tion. These variables could be divided into saline 0.9% and/or Gelofusine (Braun, Mel-After inclusion, a baseline assessment was per- three groups. sungen, Germany). If this caused an increaseformed. Baseline variables were evaluated for Routine Global Hemodynamic Variables. in CI and/or MAP 10% with the PAOP re-their ability to predict outcome. After resusci- Using a central arterial and a pulmonary ar- maining 18 mm Hg, another fluid challengetation, the hemodynamic variables were as- tery flotation catheter, we measured mean ar- could be given. If the increase in CI or MAPsessed a second time and the prognostic value terial pressure (MAP), heart rate, pulmonary was 10%, then fluid challenging wasof these variables was assessed again at this artery occlusion pressure (PAOP), and cardiac stopped, regardless of PAOP. After fluid chal-time point. index (CI). Mixed venous and arterial blood lenging, dobutamine or noradrenaline (21, 22) Patients fulfilling the inclusion criteria gases were used to measure lactate and to or both were given guided by the effects on CIwere included during a period of 2 yrs from a calculate oxygen delivery and oxygen con- and MAP, aiming at CI 2.5 L/min/m2 andmixed ICU at a university hospital. The Insti- sumption and were drawn simultaneously MAP 70 mm Hg. If PAOP was 18 mm Hg,tutional Review Board of the hospital approved with the thermodilution measurement. The dobutamine was given primarily. Whenever global hemodynamic variables were used tothis study. Written informed consent was ob- hematocrit was 0.3, a transfusion of packed calculate the systemic vascular resistance in-tained from each patient’s next of kin, since all red cells was allowed. The intensivists were dex. Urine output was measured using a Foleypatients were ventilated and sedated at the blinded for the results of the regional tono- urinary catheter.time of inclusion in the study. Global Volume-Related Hemodynamic metric and ICG-dilution variables. Variables. Double dilution measurements A stable hemodynamic status was definedStudy Patients were performed using a 10-mL cold bolus in- by the intensivist using the following guide- fusion of 1 mg/kg indocyanine green (ICG) in lines: MAP 70 mm Hg, or a stable blood Critically ill patients with sepsis as defined saline. The principle of these measurements is pressure (variation in MAP 10 mm Hg, heartby the American College of Chest Physicians/ based on the distribution of an indicator over rate 30%) during 2 hrs without the necessityCrit Care Med 2005 Vol. 33, No. 11 2495
  3. 3. Table 1. Patient demographics Variables Total Group Nonsurvivor Survivor p ValueNumber 28 14 14Age, yrs 64 13 68 14 60 12 .1Gender, % female 39 36 43 .5Bacteriologically proven sepsis, n 17/28 9/14 8/14 .5Suspected source of sepsis, n (%) Abdominal 7 (25) 3 (21) 4 (29) Pulmonary 17 (61) 9 (64) 8 (57) Other 4 (14) 2 (14) 2 (14) .8Medical/surgical 8/20 5/9 3/11 .4APACHE II score 17.7 4.9 18.4 5.3 16.6 4.1 .2MOF score 5.5 1.8 6.1 2.1 4.8 1.2 .3 APACHE, Acute Physiology and Chronic Health Evaluation; MOF, multiple organ failure. Data are presented as mean SD or as percentages. The p value indicates the significance level for the comparison between surviving and nonsurvivingpatients.of increasing any vasopressor or inotropic namic variables, with survival as the fixed Admission Variablestherapy or without the need of fluid adminis- factor and time (resuscitation period) as co-tration to treat a decrease in MAP of 10 mm variate, presented as nonstandardized regres- At admission to the ICU, before ICUHg or a change in heart rate of 30%. sion coefficients (B) (with SE). resuscitation was initiated, nonsurviving This definition was used to divide the study To detect the most important predictors of patients had a significantly lower MAPperiod into two periods: a period from admis- mortality, hemodynamic variables were ana- (Table 2) as well as a significantly lowersion to stabilization and a period after stabili- lyzed using a logistic nonparametric multiple RVEF and higher lactate. Moreover, thesezation. Although the periods before and after regression model (backward conditional). The patients had a significantly lower CBI andstabilization were seen as two distinct periods, validity of the different tests was evaluated by CPI, as well as a lower PDR (Table 2). Thethe same goals for resuscitation were actually means of a method for comparing areas under gastric pHi was significantly lower in theused as a continuum in the primary and sec- the receiver operating characteristic (ROC) nonsurviving patients, with a higherondary phases. After stabilization had been curves (AUCs), as previously described in de- PiCO2, Pr-etCO2, and mucosal-arterialachieved, the previously mentioned measure- tail (23, 24). AUCs were calculated for thements were repeated. variables at admission and after stabilization. PCO2 gap. Variables of global perfusion None of the patients were enterally fed The odds ratios for the variables after stabili- and regional function were both differentduring the study period. H2-blockers were zation were calculated using the optimal cut- between the two patient groups. None ofgiven routinely, using ranitidine. Antibiotics off point derived from the ROC curves. The the global or regional variables at admis-were given blindly or directed at identified optimal cutoff value for predicting mortality sion were more important in predictingorganisms. was calculated as the point with the greatest mortality than another variable on ROC combined sensitivity and specificity. For these curve analysis, although gastric pHi andFollow-Up analyses, a p .05 was considered statistically CPI tended to be more important predic- significant. tors of mortality on logistic regression Patients’ follow-up was continued during analysis (p .06 and p .09 for pHi andthe remainder of the ICU admission. After RESULTS CPI, respectively).discharge from the ICU, patients were moni-tored for 28 days or until final hospital dis- Patientscharge, whichever was longer, for morbidity Changes During Resuscitationand mortality. A total of 28 patients with severe sep- sis or septic shock were included in the Before the protocolized resuscitationStatistics analysis and monitored prospectively (Ta- (i.e., before admission to the ICU), pa- ble 1). The presence of infection was con- tients had received a mean of 4.4 L of Data were analyzed using the SPSS (ver- firmed in 17 patients. The mean APACHE saline and 3.3 L of Gelofusine, with nosion 7.5) software program (SPSS, Chicago, significant difference in volume infused II score at admission was 17.7. The me-IL). Data from patient characteristics were as- dian time between the start of the symp- between the patients with and withoutsessed using Student’s t-test and the chi- toms and admission to the ICU was 18 sepsis-induced hypotension. This amountsquare test. Admission variables were assessed hrs. Of the 28 patients, 12 were admitted of fluids was given during the admissionusing one-way analysis of variance. Data dur-ing the resuscitation period were compared in shock (systolic blood pressure 90 days before ICU admission (median 2.0using the Mann-Whitney U test. The use of the mm Hg or sustained drop of 40 mm days, range 0 –95 days).multiple comparisons may have introduced Hg). There were no differences between During this resuscitation phase, MAP,type I errors. The use of a Bonferroni correc- survivors and nonsurvivors with respect oxygen delivery, and RVEF were in-tion for the multiple comparisons leads to a to age, gender, duration of complaints creased significantly for all patients takencorrected alpha value of .002. before inclusion, number of patients with together (p .002, p .05, and p .02 A multivariate procedure was used to pro- sepsis and sepsis-induced hypotension, for MAP, oxygen delivery, and RVEF, re-vide a regression analysis and analysis of vari- APACHE II, scores or multiple organ fail- spectively). Lactate values were signifi-ance for the multiple dependent hemody- ure scores. cantly decreased.2496 Crit Care Med 2005 Vol. 33, No. 11
  4. 4. Table 2. Global and regional hemodynamics at admission in survivors and nonsurvivors dynamic variables. Third, resuscitation aimed at improving global hemodynam- Variables Survivors Nonsurvivors p Values ics did improve global pressure-relatedGlobal pressure-related variables and oxygen transport variables but not MAP 78 13 70 12 .04a volume-related global hemodynamic and CVP 11 3 12 4 .3 regional variables. DO2I 635 197 572 135 .2 Previously, only a few studies have SVRI 1242 413 1217 528 .8 compared the predictive power of gastric PAOP 14 3 15 6 .4Global volume-related variables tonometric variables with systemic he- TBVI 3708 1188 3719 821 1.0 modynamic variables. Bams et al. (25) RHEDVI 467 218 459 170 .9 determined the predictive value of gastric RVEDVI 172 54 189 60 .4 pHi in cardiac surgical patients. In con- RVSWI 9.1 3.8 10.1 3.9 .3 LVSWI 39.0 14.0 33.1 17.6 .2 trast to our study, they found the mean RVEF 25.5 6.3 21.2 8.1 .04a arterial pressure and mean pulmonary ar-Regional variables terial pressure to be more predictive of pHi 7.34 0.08 7.24 0.1 .03a mortality than gastric pHi or PrCO2 when PrCO2 5.7 1.0 6.7 2.3 .03a measured at admission and after 12 hrs. Pr-etCO2 1.9 1.0 2.6 1.8 .05 Pr-aCO2 0.9 0.7 1.5 1.4 .045a The difference between our study and CBI 837 420 550 368 .02a their study may be related to the patient CPI 587 317 388 262 .03a population selected, whereas another ex-PDR 21.0 9.9 14.9 8.0 .03a planation for the differences might be theUP 1.3 1.4 1.2 1.2 .8Lactate 1.2 0.6 2.3 1.9 .007a presence of extrasplanchnic perfusion ab-Base deficit 2.7 7.2 3.6 6.2 .6 normalities, such as decreased bicarbon- ate content. However, this cannot explain MAP, mean arterial pressure (mm Hg); CVP, central venous pressure (mm Hg); DO2I, oxygen the predictive power of the PrCO2 and thedelivery index (mL/min/m2); SVRI, systemic vascular resistance index (dyne/sec/cm5); PAOP, pulmo- Pr-etCO2 in our study. What is perhapsnary artery occlusion pressure (mm Hg); TBVI, total blood volume index; RHEDVI, right-heart more important is the fact that in ourend-diastolic volume index; RVEDVI, right ventricle end-diastolic volume index; RVSWI, right ventri- study, gastric tonometric variables be-cle stroke work index (g/m/m2); LVSWI, left-ventricle stroke work index (g/m/m2); pHi, gastric came predictive after stabilization hadintramucosal pH; PrCO2, regional (gastric) PCO2 (kPa); Pr-etCO2, difference between regional andend-tidal PCO2; Pr-aCO2, mucosal-arterial PCO2; CBI, indocyanine green blood clearance (mL/min/m2); been achieved, usually 12 hrs after ad-CPI, indocyanine plasma clearance (mL/min/m2); PDR, indocyanine plasma disappearance rate (%); mission. This is confirmed by the resultsUP, urine output (mL/kg/hr). obtained by Maynard et al. (26), who a p .05. Data are presented as mean SD. found a high predictive value for the gas- tric pHi after only 3 days. On the other hand, Lorente et al. (27) found that the Patients presenting at the ICU in an tidal PCO2 gap, and PrCO2 (CBI, regres- mucosal-arterial PCO2 gap was an inde-early phase of sepsis (duration of com- sion coefficient B 0.002, SE 0.0007, pendent predictor of outcome in severeplaints 0 –12 hrs) had a significant in- p .01; pHi, B 10.9, SE 4.6, p burn patients only at admission and notcrease in RVEF, unlike patients present- .02; PrCO2, B 1.5, SE 0.7, p .03; after 12 hrs of resuscitation. Anothering in a late phase ( 24 hrs). Pr-etCO2, B 1.5, SE 0.7, p .02; study found that gastric PrCO2 and sys- constant, B 83.9, SE 35.1). The AUCs temic lactate were both good predictorsHemodynamic Variables After of gastric pHi and Pr-etCO2 were signifi- of outcome in patients with severe sepsisStabilization as Predictors of cantly higher after stabilization than at (28). Similarly, Ivatury et al. (29) indi-Mortality admission and were significantly higher cated that gastric pHi was the best pre- than the AUCs of MAP, PrCO2, central dictor of outcome in severely traumatized After patients had been stabilized, the venous pressure, left ventricular stroke patients with organ failure. Thus, severalglobal volume-related hemodynamic and work index, and arterial pH (Table 4). studies have confirmed the importance ofregional variables were compared be- gastric intramucosal tonometry in pre-tween surviving and nonsurviving pa- DISCUSSION dicting outcome during the resuscitationtients. After stabilization, patients who of critically ill patients.died during their hospital stay had a sig- There were three main findings of this The importance of volume-relatednificantly higher PrCO2 and Pr-etCO2 and study. First, no superior predictor of out- global hemodynamic variables in the as-a significantly lower CBI, CPI, PDR, and come was identified at admission, al- sessment of intravascular volume statuspHi (Table 3). though both regional variables and global has been emphasized before. Numerous All variables were subsequently used hemodynamic variables were different be- studies have indicated that intravascularas independent variables in a logistic re- tween surviving and nonsurviving pa- volumes assessed by the thermo-dye di-gression analysis with mortality as the tients, and the mean values were abnor- lution technique give more accurate in-dependent variable, and areas under the mal in both groups compared with formation on the preload dependency ofROC curves were calculated. The most normal values. Second, after stabiliza- cardiac output than pressure-related vari-powerful predictors (in terms of signifi- tion, regional variables were the most ables (15, 30 –34). In addition to this, ourcant regression coefficients) were CBI, important predictors of mortality com- study investigated the predictive power ofgastric intramucosal pHi, mucosal-end- pared with global volume-related hemo- the global volume-related variables in pa-Crit Care Med 2005 Vol. 33, No. 11 2497
  5. 5. tients with septic shock. We found that achieved higher oxygen transport and A number of comments need to bevolume-related hemodynamic variables perfusion variables during their stay at made on the interpretation of our obser-were superior to pressure-related vari- the intensive care unit. In addition, vations. The resuscitation principles usedables in differentiating between survivors Moore et al. (1) found that in critically ill in this study were based on standardand nonsurvivors. Previously, Jaarsma et trauma patients, the inability to improve practice at our ICU, and the intensivist onal. (35) found that the predictive power of oxygen transport during resuscitation call judged whether the resuscitation ledleft ventricular stroke work index using aimed at achieving supranormal oxygen to stabilization of the global hemodynam-echocardiography was higher than that of delivery was related to the development ics. Although this practice led to im-PAOP in patients with acute myocardial of multiple organ failure. However, it re- provement of global hemodynamic vari-infarction. In our study, resuscitation in- mained difficult to identify the individual ables, no improvement or even acreased RVEF in both surviving and non- worsening of the regional variables was patient who was prone to develop multi-surviving patients, indicating that resus- noted. It may be questioned whether the ple organ failure or die on intensive care.citation improves global perfusion in all resuscitation used was adequate and Rhodes et al. (37) described a dobutaminepatients. It is interesting to see that the whether other methods of resuscitationextravascular lung water tended to de- stress test that can identify nonsurvivors. Creteur et al. (38) found that this stress should have been attempted. The use ofcrease during resuscitation. This de- test identifies occult hypovolemia in the thermodilution catheters to improve he-crease in extravascular lung water may splanchnic area. In addition, our study modynamics during sepsis is controver-indicate improved fluid balance or may be found that patients with shock at admis- sial, and some studies have indicated thatdue to the increased application of posi- sion had obvious abnormalities in both the use of these catheters to treat hemo-tive end-expiratory pressure ventilation(36). global, gastric tonometric, and ICG- dynamic disturbances during critical ill- Intensive care resuscitation is aimed clearance variables, whereas abnormali- ness may increase mortality rate (39, 40).at improving the global perfusion. Bishop ties in the hepatosplanchnic variables However, our aim was to study the effectset al. (2) described that surviving patients still persisted after stabilization. of the resuscitation techniques normally used at an ICU. The tendency of the global volume-related variables strokeTable 3. Global hemodynamic and regional variables after stabilization in surviving and nonsurviving volume and RVEF to improve and thepatients significant improvement in oxygen deliv-Variables Survivors Nonsurvivors p Value ery during this resuscitation indicate that the resuscitation led to improvement ofpHi 7.31 0.07 7.24 0.09 .03 the global hemodynamics. The questionPr-etCO2 2.1 0.7 3.1 1.5 .04 which resuscitation techniques should bePrCO2 6.0 0.8 7.1 1.5 .04CBI 838 295 572 262 .02 applied in patients with septic shock re-CPI 587 224 411 168 .02 mains difficult to answer. The use of “su-PDR 21.1 7.1 14.6 5.0 .01 pranormal” hemodynamic goals in pa- tients with septic shock and organ failure pHi, gastric intramucosal pH; Pr-etCO2, difference between regional and end-tidal PCO2; PrCO2, has been shown to have no influence onregional (gastric) PCO2 (kPa); CBI, indocyanine green blood clearance (mL/min/m2); CPI, indocyanine outcome (41, 42) and even increasedplasma clearance (mL/min/m2); PDR, indocyanine plasma disappearance rate. Data are presented as mean SD. Only variables that showed a significant difference were mortality rate in some studies (43). In adisplayed. Comparison between variables after stabilization was achieved between survivors and recent trial, the application of these su-nonsurvivors. pranormal goals early in the course ofTable 4. Area under the receiver operating characteristic curves (AUC) and odds ratios Variable Admission AUC After stabilization AUC Cutoff Points Odds Ratio 95% CICBI 0.69 0.08 0.70 0.08a 533 3.9 1.1–13.8bMAP 0.55 0.05 0.57 0.06 78 0.8 0.3–1.8pHi 0.65 0.05 0.75 0.05b,c 7.17 4.8 1.5–14.6bPrCO2 0.56 0.05 0.65 0.06 5.8 2.0 0.9–4.5Pr-etCO2 0.61 0.06 0.75 0.07b,c 1.8 3.0 1.4–6.3bPr-aCO2 0.59 0.06 0.67 0.09 0.83 1.3 0.6–3.0CVP 0.57 0.06 0.56 0.08 12 0.8 0.2–2.9LVSWI 0.62 0.06 0.60 0.08 37 1.8 0.6–5.6Arterial pH 0.63 0.06 0.63 0.08 7.35 2.1 0.7–6.2 CBI, indocyanine green blood clearance (mL/min/m2); MAP, mean arterial pressure (mm Hg); pHi, gastric intramucosal pH; PrCO2, regional (gastric)PCO2 (kPa); Pr-etCO2, difference between regional and end-tidal PCO2; Pr-aCO2, mucosal-arterial PCO2; CVP, central venous pressure (mm Hg); LVSWI,left-ventricle stroke work index (g/m/m2); CI, confidence interval. a p .05 for CBI vs. CVP; bp .05 for the time point after stabilization vs. the time point at admission; cp .05 for comparison of pHi or Pr-etCO2 vs.MAP, PrCO2, CVP, LVSWI, and arterial pH. Values are AUC SE. Data indicate the differences in AUC at admission and after stabilization. The maximumvalue for the AUC is 1. No differences in AUC were present at admission. The AUC values of pHi and Pr-etCO2 were significantly larger after stabilizationcompared to admission values. The odds ratios represent the increased risk of death for an optimal cut-off point of the test defined by analysis of theROC-curves.2498 Crit Care Med 2005 Vol. 33, No. 11
  6. 6. dicts postinjury multiple organ failure. ume as indicators of cardiac preload. J Crit T J Trauma 1992; 33:58 – 65 Care 1996; 11:180 –1808 his study shows 2. Bishop MH, Shoemaker WC, Appel PL, et al: 16. Holm C, Melcer B, Horbrand F, et al: In- Relationship between supranormal circula- trathoracic blood volume as an end point in that during stan- tory values, time delays, and outcome in se- resuscitation of the severely burned: An ob- verely traumatized patients. Crit Care Med servational study of 24 patients. J Trauma dard intensive care 1993; 21:56 – 63 2000; 48:728 –734 3. Dabrowski GP, Steinberg SM, Ferrara JJ, et 17. Bone RC, Balk RA, Cerra FB, et al: Defini- resuscitation in patients al: A critical assessment of endpoints of tions for sepsis and organ failure and guide- shock resuscitation. Surg Clin North Am lines for the use of innovative therapies in with septic shock, variables 2002; 80:825– 844 sepsis. Chest 1992; 101:1644 –1655 4. Wo CC, Shoemaker WC, Appel PL, et al: 18. Knaus WA, Draper EA, Wagner DP, et al: of splanchnic function are Unreliability of blood pressure and heart rate APACHE II: A severity of disease classifica- better predictors of outcome for evaluation of circulatory stability in tion system. Crit Care Med 1985; 13: emergency resuscitation and critical illness. 818 – 828 than global pressure- or vol- Crit Care Med 1993; 21:218 –223 19. Brinkmann A, Calzia E, Träger K, et al: Mon- 5. Brinkmann A, Calzia E, Träger K, et al: Mon- itoring the hepato-splanchnic region in the ume-related variables. itoring the hepato-splanchnic region in the critically ill patient. Measurement tech- critically ill patient. Measurement tech- niques and clinical relevance. Intensive Care niques and clinical relevance. Intensive Care Med 1998; 24:542–556 Med 1998; 24:542–556 20. Gutierrez G, Palizas F, Doglio G, et al: Gas- 6. Santoso JT, Wisner DH, Battistella FD, et al: tric intramucosal pH as a therapeutic indexseptic shock led to a reduced mortality Comparison of gastric mucosal pH and clin- of tissue oxygenation in critically ill. Lancetrate (44). ical judgement in critically ill patients. Eur 1992; 339:195–199 The use of the multiple comparisons J Surg 1998; 164:521–526 21. Reilly PM, Bulkley GB: Vasoactive mediatorsmay have introduced type I errors. The 7. Oud L, Kruse JA: Progressive gastric in- and splanchnic perfusion. Crit Care Meduse of a Bonferroni correction for the tramucosal acidosis follows resuscitation 1993; 21(2 Suppl):S55–S68multiple comparisons leads to a corrected from hemorrhagic shock. Shock 1996; 22. Poeze M, Greve JWM, Ramsay G: Is splanch- 6:61– 65 nic perfusion a critical problem in sepsis? In:alpha value of .002. This means that the 8. Poeze M, Takala J, Greve JWM, et al: Pre- Sepsis and Organ Dysfunction. From Basicslikelihood of finding a significant differ- operative tonometry is predictive for mortal- to Clinical Approach. First Edition. Baue AE,ence by chance after testing all the he- ity and morbidity in high-risk surgical pa- Berlot G, Gullo A, et al (Eds). Milano, Spring-modynamic variables would be consider- tients. Intensive Care Med 2000; 26: er-Verlag, 1999, pp 169 –181able. However, since the results 1272–1281 23. Hanley JA, McNeil BJ: A method for compar-concerning the prospectively analyzed re- 9. Edouard AR, Dergremont A-C, Duranteau J, ing the areas under receiver operating char-gional variables all indicate at the same et al: Heterogeneous regional vascular re- acteristic curves derived from the samedirection (i.e., regional variables are bet- sponses to simulated transient hypovolemia cases. Radiology 1983; 148:839 – 843ter predictors of outcome than global in man. Intensive Care Med 1994; 20: 24. Hanley JA, McNeil BJ: The meaning and usevariables), the use of a Bonferroni correc- 414 – 420 of the area under a receiver operating char-tion analysis is not necessary. Moreover, 10. Kirton O, Windsor J, Wedderburn R, et al: acteristic (ROC) curve. Radiology 1982; 143: Failure of splanchnic resuscitation in the 29 –36we found important differences in predic- acutely injured trauma patient correlates 25. Bams JL, Mariani MA, Groeneveld AB: Pre-tive capability between global hemody- with multiple organ system failure and dicting outcome after cardiac surgery: Com-namic and regional variables, based on length of stay in the ICU. Chest 1998; 113: parison of global haemodynamic and tono-logistic regression analysis and AUC anal- 1064 –1169 metric variables. Br J Anaesth 2000; 82:ysis. However, with the relatively small 11. Poeze M, Ramsay G, Buurman WA, et al: 33–37patient numbers studied, our study Increased hepatosplanchnic inflammation 26. Maynard ND, Bihari D, Beale R, et al: Assess-should be regarded as an initial study for precedes the development of organ dysfunc- ment of splanchnic oxygenation by gastricthe design of a larger trial comparing the tion after elective high-risk surgery. Shock tonometry in patients with acute circulatorydifferent diagnostic modalities. 2002; 17:451– 458 failure. JAMA 1993; 270:1203–1210 12. Singer M, Bennett D: Pitfalls of pulmonary 27. Lorente JA, Ezpeleta A, Esteban A, et al: artery catheterization highlighted by Dopp- Systemic hemodynamics, gastric intramuco-CONCLUSIONS ler ultrasound. Crit Care Med 1989; 17: sal PCO2 changes, and outcome in critically This study shows that during standard 1060 –1061 ill burn patients. Crit Care Med 2000; 28: 13. Godje O, Peyerl M, Seebauer T, et al: Central 1728 –1735intensive care resuscitation in patients venous pressure, pulmonary capillary wedge 28. Friedman G, Berlot G, Kahn R, et al: Com-with septic shock, variables of splanchnic pressure and intrathoracic blood volumes as bined measurements of blood lactate concen-function are better predictors of outcome preload indicators in cardiac surgery pa- trations and gastric intramucosal pH in pa-than global pressure- or volume-related tients. Eur J Cardiothorac Surg 1998; 13: tients with severe sepsis. Crit Care Med 1995;variables. No superior hemodynamic 533–539 23:1184 –1193variable was identified as a predictor of 14. Chang MC, Meredith JW: Cardiac preload, 29. Ivatury RR, Simon RJ, Havriliak D, et al:outcome before resuscitation is started. splanchnic perfusion, and their relationship Gastric mucosal pH and oxygen delivery and during resuscitation in trauma patients. oxygen consumption as indices in the assess-REFERENCES J Trauma 1997; 42:577–582 ment of adequacy of resuscitation after trau- 15. Lichtwarck-Aschoff M, Beale R, Pfeiffer UJ: ma: a prospective randomised study. 1. Moore FA, Haenel JB, Moore EE, et al: In- Central venous pressure, pulmonary artery J Trauma 1995; 39:128 –136 commensurate oxygen consumption in re- occlusion pressure, intrathoracic blood vol- 30. Brock H, Gabriel C, Bibl D, et al: Monitoring sponse to maximal oxygen availability pre- ume, and right ventricular end-diastolic vol- intravascular volumes for postoperative vol-Crit Care Med 2005 Vol. 33, No. 11 2499
  7. 7. ume therapy. Eur J Anaesthesiol 2002; 19: tensive Care Med 1992; 18:142–147 effectiveness of right heart catheterization in 288 –294 35. Jaarsma W, Visser CA, Eenige van MJ, et al: the initial care of critically ill patients. JAMA31. Bindels AJ, van der Hoeven JG, Graafland AD, Predictive value of two-dimensional echocar- 1996; 276:889 – 897 et al: Relationships between volume and diographic and hemodynamic measurements 40. Hayes MA, Timmins AC, Yau EHS, et al: pressure measurements and stroke volume on admission with acute myocardial infarc- Elevation of systemic oxygen delivery in the in critically ill patients. Crit Care 2000; tion. J Am Soc Echocardiogr 1998; treatment of critically ill patients. New Engl 4:193–199 1:187–193 J Med 1994; 330:1717–172232. Sakka SG, Reinhart K, Meier-Hellmann A: 36. Mitchell JP, Schuller D, Calandrino FS, et 41. Poeze M, Greve JWM, Ramsay G: Oxygen Comparison of pulmonary artery and arterial al: Improved outcome based on fluid man- delivery in septic shock. Chest 1999; 116: thermodilution cardiac output in critically ill agement in critically ill patients requiring 1145 patients. Intensive Care Med 1999; 25: pulmonary artery catheterisation. Am J Re- 42. Boyd O, Hayes M: The oxygen trail: The goal. 843– 846 spir Crit Care Med 1992; 145:990 –998 Br Med Bull 1999; 55:125–13933. Kisch H, Leucht S, Lichtwarck-Aschoff M, et 37. Rhodes A, Malagon I, Lamb FJ, et al: Failure 43. Hayes MA, Timmins AC, Yau EHS, et al: al: Accuracy and reproducibility of the mea- to increase oxygen consumption is a predic- Elevation of systemic oxygen delivery in the surement of actively circulating blood vol- tor of mortality in septic patients. Critical treatment of critically ill patients. New Engl ume with an integrated fiberoptic monitor- Care Med 1999; 27:2361–2366 J Med 1994; 330:1717–1722 ing system. Crit Care Med 1995; 23:885– 893 38. Creteur J, De-Backer D, Vincent J-L: A do- 44. Rivers E, Nguyen B, Havstad S, et al: Early34. Lichtwarck-Aschoff M, Zeravik J, Pfeiffer UJ: butamine test can disclose hepatosplanchnic Goal-Directed Therapy Collaborative Group. Intrathoracic blood volume accurately re- hypoperfusion in septic patients. Am J Respir Early goal-directed therapy in the treatment flects circulatory volume status in critically Crit Care Med 1999; 160:839 – 845 of severe sepsis and septic shock. New Engl ill patients with mechanical ventilation. In- 39. Connors AF, Speroff T, Dawson NV, et al: The J Med 2001; 345:1368 –13772500 Crit Care Med 2005 Vol. 33, No. 11