Efficacy and Outcomes After Vasopressin Guideline Implementation in Septic Shock

669163AOPXXX10.1177/1060028016669163Annals of PharmacotherapyWu et al research-article2016 Research Report Efficacy and Outcomes After Vasopressin Guideline Implementation in Septic Shock Annals of Pharmacotherapy 2017, Vol. 51(1) 13 20 The Author(s) 2016 Reprints and permissions: sagepub.com/journalspermissions.nav DOI: 10.1177/1060028016669163 aop.sagepub.com Janet Y. Wu, PharmD 1, Joanna L. Stollings, PharmD, FCCM, BCPS, BCCCP 1, Arthur P. Wheeler, MD 2,, Matthew W. Semler, MD, MSc 2, and Todd W. Rice, MD, MSc 2 Abstract Background: In septic shock, the dose of norepinephrine (NE) at which vasopressin (AVP) should be added is unknown. Following an increase in AVP price, our medical intensive care unit (MICU) revised its vasopressor guidelines to reserve AVP for patients requiring greater than 50 µg/min of NE. Objective: The purpose of this study is to compare efficacy and safety outcomes for patients admitted before the guideline revision with those for patients admitted after the revision. Methods: This was a single-center, retrospective cohort study of patients admitted to Vanderbilt University Medical Center from November 1, 2014, to November 30, 2015. Before June 1, 2015, the vasopressor guidelines recommended initiation of AVP for patients requiring 10 µg/min of NE or greater. After June 1, 2015, the guidelines recommended initiation of AVP at a NE dose of 50 µg/min or greater. Results: Time to achieve goal mean arterial pressure (MAP) was shorter in the postintervention group (2.0 vs 1.3 hours; P = 0.03) in univariate analysis but not after adjusting for prespecified confounders. Incidence of new arrhythmias was similar between the 2 groups (14.9% vs 10.9%; P = 0.567). In multivariable analysis accounting for baseline severity of illness, admission after the revision was associated with decreased 28-day mortality (odds ratio = 0.34; 95% CI = 0.16-0.71; P = 0.004). Conclusions: Use of a vasopressor guideline restricting AVP initiation in septic patients to those on at least 50 µg/min of NE appeared to be safe and did not affect the time to reach goal MAP. Keywords adrenergic agonists, critical care, sympathomimetics, sepsis, clinical research Background Norepinephrine (NE) and vasopressin (AVP) are commonly used vasopressors for patients in septic shock. The 2012 Surviving Sepsis Campaign guidelines recommend NE as the vasopressor of choice and adding AVP to increase mean arterial pressure (MAP) or to decrease NE dosage. 1 In a study of patients with catecholamine-resistant vasodilatory shock, AVP was initiated in patients who were volume refractory and required NE doses exceeding 0.5 µg/kg/min. The authors found that patients receiving both AVP and NE performed better hemodynamically compared with patients receiving only NE. 2 In contrast, Micek et al 3 showed that patients who received NE and AVP had significantly higher mortality rates than patients on NE alone. To complicate matters further, a maximum dose of NE has not been determined. Doses ranging from 0.2 to 5 µg/kg/min have been used in studies, although higher doses (1.4-3.8 µg/kg/min) are associated with a risk of extreme vasoconstriction and tissue hypoperfusion. 4-6 The Vasopressin versus Norepinephrine Infusion in Patients with Septic Shock (VASST) trial compared NE with NE plus AVP and found no significant difference in 28-day mortality. The researchers found that in the defined stratum of less-severe septic shock, patients who received AVP had a lower mortality rate compared with patients who received NE (26.5% vs 35.7%, P = 0.05). 7 However, the study was not powered to detect this outcome, and some clinicians have called into question whether AVP should be initiated in patients receiving 15 µg/min or less of NE. 1 Vanderbilt University Medical Center, Nashville, TN, USA 2 Vanderbilt University School of Medicine, Nashville, TN, USA Deceased. Corresponding Author: Joanna L. Stollings, Department of Pharmaceutical Services, Vanderbilt University Medical Center, 1211 Medical Center Drive BUH-131 Nashville, TN 37232, USA. Email: joanna.stollings@vanderbilt.edu

14 Annals of Pharmacotherapy 51(1) In 2013, a vasopressor guideline was implemented in the medical intensive care unit (MICU) at our institution that recommended (1) no maximum dose of NE and (2) that AVP 0.04 U/min be added once patients reached a NE dose of 10 µg/min. This guideline also recommended discontinuing NE first. A study comparing the preguideline implementation group with the postguideline implementation group found no difference in the time to reach goal MAP or the incidence of arrhythmias. However, the mean dose of vasopressor at which another catecholamine vasopressor was added increased significantly between the control and intervention groups. 8 With a large increase in the price of AVP, new guidelines were implemented in the MICU in 2015 to supplement NE with AVP 0.04 U/min only when patients required doses of 50 µg/min of NE. Contrary to the original guideline, the new one recommended discontinuing AVP first. Given the large increase in price of AVP and lack of robust data to support addition of AVP at low doses of NE, we hypothesized that delay in addition of AVP until patients were receiving 50 µg/min of NE would not result in a change in time to reach goal MAP. Methods Study Design and Participants We conducted a quasiexperimental cohort study using a pre-post intervention design to evaluate the effectiveness of the AVP-sparing guideline revision. This study took place between November 1, 2014, and November 30, 2015, in the MICU. Patients were identified through retrospective chart review and included if they were admitted to the MICU and received NE therapy for septic shock for at least 6 hours. Exclusion criteria were pregnancy, other concomitant vasopressors (epinephrine, phenylephrine, or dopamine), receipt of vasopressors at an outside hospital prior to transfer, NE or AVP initiated outside of the MICU, receipt of NE at doses 50 µg/min, and receipt of vasopressors for an indication other than septic shock. The Vanderbilt University Institutional Review Board approved this study with a waiver of informed consent. Interventions Vasopressor guidelines at our institution were developed in April 2013, which recommended initial NE use without a maximum dose followed by AVP once patients reached NE 10 µg/min. These guidelines recommended discontinuing NE first before stopping AVP and did not place any restrictions on the use of other vasopressors such as epinephrine and phenylephrine. Patients in the preintervention group were managed using the 2013 vasopressor guidelines and were admitted from November 1, 2014, to June 30, 2015. Patients included in the preintervention group received NE <50 µg/min with AVP 0.04 U/min. With the large increase in the price of AVP, these vasopressor guidelines were revised in June 2015 to recommend initial NE use without a maximum dose, AVP initiation once NE reached 50 µg/min, and discontinuation of AVP before NE. Patients in the postintervention group were managed using the 2015 vasopressor guidelines, the AVP-sparing strategy, and were admitted from July 1 to November 1, 2015. Patients included from the postintervention group received NE < 50 µg/min without AVP. The primary objective of this study was to determine if delay in the start of AVP until patients were receiving 50 µg/min of NE changed time to reach goal MAP of 65 mm Hg. Statistical Analysis Data were extracted from the electronic medical records for all patients who met inclusion criteria during the 1-year study period. Flowsheet data were used for the majority of data collection, including doses of NE. Categorical data were analyzed using the χ 2 test and are reported as numbers and percentages. Continuous data were analyzed using the Mann-Whitney U test and reported as median and interquartile range. A linear regression was conducted to examine the association between the pre-post intervention and the outcomes of hospital length of stay and ICU length of stay after accounting for baseline APACHE (Acute Physiologic Chronic Health Evaluation) II score and NE dose. A logistic regression was conducted with the same variables for the outcome of 28-day mortality. A Cox proportional hazard model was developed to examine the relationship between the pre-post intervention and the time until goal MAP was achieved after accounting for prespecified confounders. Statistical analyses were performed with IBM SPSS Statistics, version 23. Results Over the study period, data from 635 patients receiving NE and AVP were reviewed. A total of148 patients were included in the study after 487 patients met the exclusion criteria. Figure 1 illustrates reasons for exclusion. Baseline Demographics Patients in the preintervention and postintervention periods were similar at baseline except for higher APACHE II score (23 vs 18; P = 0.001) and increased prevalence of cardiogenic shock (10.8% vs 2.7%; P = 0.049) in the preintervention group (Table 1).

Wu et al 15 was associated with both hospital (adjusted odds ratio [OR] = 10; 95% CI = 1.42-2.48; P < 0.0001) and ICU length of stay (adjusted OR = 1.99; 95% CI = 1.38-2.86; P < 0.0001). Mortality at 28 days was significantly higher in the preintervention group at 51% compared with 28% in the postintervention group (Table 2). Logistic regression showed an association between the postintervention group and lower 28-day mortality after adjusting for the type of intervention received, APACHE II score, and NE dose at initiation (adjusted OR = 0.34, 95% CI = 0.16-0.71, P = 0.004; Table 2). Finally, given that AVP was not used in the postintervention phase, our institution net a cost avoidance of $42 157.50 compared with the preintervention group. Figure 1. Derivation of the study cohort. Abbreviations: AVP, vasopressin; MICU, medical intensive care unit; NE, norepinephrine. Outcomes In unadjusted analysis, the time to reach goal MAP was longer in the preintervention group than the postintervention group (2 vs 1.3 hours, P = 0.030). After adjusting for age, APACHE II score, receipt of corticosteroids, initiation of dialysis, and the NE dose at initiation in a Cox proportional hazard model, however, there was no difference in the time to reach goal MAP between patients in the 2 groups (Figure 2). The median NE dose at initiation was 10 µg/min (0.106 µg/kg/min) in the preintervention group versus 5 µg/min (0.068 µg/kg/min) in the postintervention group (P < 0.001). The median maximum NE dose was 24 (0.325 µg/kg/min) versus 11 µg/min (0.145 µg/kg/min; P < 0.001). The dose of NE at which AVP was initiated in the preintervention group was 18 µg/min (0.233 µg/kg/min). The duration of NE was a median of 34.3 and 26.0 hours preintervention and postintervention, respectively (P = 0.094). The duration of AVP in the preintervention group was a median of 28.8 hours, and the duration of concomitant NE and AVP was 25.5 hours. More patients in the preintervention group required dialysis compared with the postintervention group (25.7% vs 8.1%, P = 0.004). There were no differences between the 2 groups with regard to receipt of concomitant medications, including corticosteroids and inotropes. Concomitant corticosteroids were administered in 16 patients (21.6%) in the preintervention group compared with 15 patients (20.3%) in the postintervention group (P = 0.840). The maximum lactate levels were 3.2 and 1.3 mmol/l, respectively (P < 0.001), and there was no difference with regard to heart rate or incidence of new arrhythmias (14.9% vs 10.9%, P = 0.567). There was no difference between the 2 groups in regard to hospital and ICU length of stay. Linear regression showed that the starting NE dose Discussion In this retrospective cohort study using a pre-post intervention design, initiation of AVP at a NE dose of at least 50 µg/ min rather than 10 µg/min did not appear to be associated with a difference in the time to reach goal MAP. Reserving AVP for septic shock patients requiring higher NE doses also appeared safe with regard to new arrhythmias, ICU and hospital length of stay, and in-hospital mortality. The Surviving Sepsis Campaign guidelines recommend achieving a goal MAP of at least 65 mm Hg during the first 6 hours of resuscitation. 1 Persistent hypotension, defined as MAP of less than 65 mm Hg, is associated with poor outcomes, such as increased mortality and need for renal replacement therapy. 9,10 Dunser et al 10 conducted a retrospective cohort study to evaluate the association between arterial blood pressure during the first 24 hours and mortality in sepsis. They found that one or more hypotensive episodes (MAP < 60 mm Hg) was associated with an almost 3-fold increased risk of death (OR = 2.96; 95% CI = 1.06-10.36; P = 0.04). Despite extensive study, the best approach to vasopressor use during septic shock remains undefined. In our study, using NE as an initial agent and reserving the addition of AVP for patients requiring at least 50 µg/min of NE appeared to be as safe as earlier AVP initiation with regard to MAP achieved, incidence of arrhythmias, and clinical outcomes. Left untreated, septic shock can lead to organ dysfunction such as renal failure. AVP has been suggested to not only decrease the dose of NE, but also increase urine output and creatinine clearance and reduce the progression to renal failure. 11-13 A retrospective case series evaluated the effects of AVP on hemodynamic and renal function and found that urine output increased by 79% 4 hours after starting AVP (P = 0.002). However, further increases were not significant at 24- or 48-hour time points. 11 A post hoc analysis of the VASST trial, stratifying patients according to the modified RIFLE criteria, found that AVP decreased the worsening of renal failure and need for dialysis by almost 2-fold

16 Annals of Pharmacotherapy 51(1) Table 1. Baseline Demographics. Preintervention (n = 74) Postintervention (n = 74) P Value Age (years) a 57.5 (49.5-67.0) 58 (48.0-68.0) 0.991 Sex (male) 40 (54.1%) 33 (44.6%) 0.250 Weight (kg) a 78.2 (66.9-98.9) 72.7 (60.0-90.3) 0.082 Race 0.374 White 61 (82.4%) 58 (78.4%) Black 12 (16.2%) 11 (14.9%) Asian 0 (0%) 2 (2.7%) Unknown 1 (1.4%) 3 (4.1%) Past medical history Dialysis 1 (1.4%) 6 (8.1%) 0.053 Liver disease 12 (16.2%) 13 (17.6%) 0.826 Arrhythmia 11 (14.9%) 8 (10.8%) 0.461 CAD 15 (20.3%) 13 (17.6%) 0.675 CHF 11 (14.9%) 20 (27.0%) 0.069 COPD 12 (16.2%) 16 (21.6%) 0.401 Cancer 20 (27.0%) 21 (28.4%) 0.854 Immunosuppressed 10 (13.5%) 13 (17.6%) 0.496 Where admitted from 0.060 Outside hospital 39 (52.7%) 28 (37.8%) Emergency department 35 (47.3%) 43 (58.1%) Floor 0 (0%) 3 (4.1%) Infection source b Pneumonia 30 (40.5%) 28 (37.8%) 0.736 Intra-abdominal 14 (18.9%) 13 (17.6%) 0.831 Line infection 1 (1.4%) 3 (4.1%) 0.311 Urinary 16 (21.6%) 13 (17.6%) 0.534 Skin and soft tissue 6 (8.1%) 13 (17.6%) 0.085 Other 4 (5.4%) 2 (2.7%) 0.405 Unknown 8 (10.8%) 8 (10.8%) 1.000 Type of shock b Septic 74 (100.0%) 74 (100.0%) 1.000 Cardiogenic 8 (10.8%) 2 (2.7%) 0.049 Hypovolemic 4 (5.4%) 7 (9.5%) 0.347 Neurogenic 0 (0%) 0 (0%) 1.000 APACHE II a 23 (19-29.5) 18 (14-26) 0.001 Abbreviations: APACHE, Acute Physiologic Chronic Health Evaluation; CAD, coronary artery disease; CHF, congestive heart failure; COPD, chronic obstructive pulmonary disorder. a Median (interquartile range). b Patient could have more than 1 infection source and more than 1 type of shock. (P = 0.03 and P = 0.02, respectively) among patients with mild renal injury. 13 However, the finding of the beneficial effects of AVP on renal function has not been universal. A recent randomized controlled trial evaluated kidney failure free days in sepsis patients receiving either AVP or NE. The results from this trial found no difference in the distribution of kidney failure free days between the 2 groups (P = 0.88) or rate of renal replacement therapy (AVP, 25.4%; NE, 35.3%; P = 0.73). 14 In our study, patients who did not receive AVP had higher urine output compared with patients who did receive AVP, although this was only significant at the time of and 6 hours after NE initiation. Furthermore, significantly more patients in the preintervention group required dialysis than patients in the postintervention group (25.7% and 8.1%, respectively; P = 0.004). Cox proportional hazard analysis adjusting for differences in dialysis requirement was utilized to account for the potential effects of AVP on renal function (Figure 2). Adverse effects of AVP reported in studies include mesenteric ischemia, myocardial ischemia, digital ischemia, skin necrosis, and hyponatremia. 11,12,15-18 However, most of these studies were relatively small and unblended, so the true risks of AVP infusion remain unknown. Because of the retrospective nature of our study, it was not reasonable to

Wu et al 17 Figure 2. Time required to achieve goal blood pressure in each study group. In a Cox proportional hazard model adjusting for pre-specified covariates of age, APACHE II score, receipt of corticosteroids, dialysis, and NE dose at initiation, there was no difference between the pre-intervention (AVP) and postintervention (no AVP) groups in the time required to achieve a MAP of at least 65 mmhg (P = 0.07). collect data on rates of ischemia. Instead, we used maximum lactate level as a marker of ischemia and found that patients who received NE and AVP had a significantly higher lactate level compared with patients who only received NE (3.2 vs 1.3, P < 0.001). Although NE is associated with arrhythmias, the rate of development of new arrhythmias was low and did not differ between each group (14.9% vs 10.8%, P = 0.567). There was also no difference between median heart rates at any time point throughout the study. Recent studies have suggested that there may be a potential interaction between AVP and corticosteroids and that treatment with the combination may result in higher plasma levels of AVP and improved outcomes. 19-21 A post hoc substudy of the VASST trial found that there was a significant interaction between receipt of AVP infusion and corticosteroid treatment (P = 0.008). There was also an association with decreased mortality in patients who received corticosteroids and AVP compared with NE (35.9% vs 44.7%, P = 0.03). In patients who did not receive corticosteroids, receipt of AVP was associated with increased mortality compared with receipt of NE (33.7% vs 21.3%, P = 0.06). 19 A subsequent small prospective trial was conducted to evaluate the potential interaction between AVP and corticosteroids. Patients in the corticosteroid group required a shorter duration (3.1 days, 95% CI = 1.1-5.1) and lower total dose of AVP (ratio = 0.47; 95% CI = 0.32-0.71) when compared with the placebo group. 22 Although the difference between the receipt of corticosteroids in the 2 arms of our study was not significant (21.6% vs 20.3%, P = 0.840), given the potential interaction between AVP and corticosteroids shown in the aforementioned trials, differences in receipt of corticosteroids were accounted for in the Cox proportional hazard model (Figure 2). With regard to clinical outcomes, results from the VASST trial showed no difference in mortality in patients who were randomized to receive NE or AVP (39.3% vs 35.4%, P = 0.26). In addition, there were no differences in ICU length of stay (16 vs 15 days, P = 0.14) or hospital length of stay (26 vs 27 days, P = 0.23). 7 Whereas the lengths of stay in our study were shorter than those in the VASST trial, no differences were found between our 2 groups. After adjusting for the type of intervention received, APACHE II score, and NE dose at initiation, linear regression showed that the NE dose was the only variable associated with hospital and ICU lengths of stay (P < 0.0001). There was a difference seen in mortality rates between patients who received AVP and patients who did not (51.4% vs 28.4%, P < 0.004), which remained even after multivariate regression analysis. Although these mortality findings are interesting, unlike the VASST trial, our study was not designed or powered for a mortality end point. To our knowledge, this is the only study that has evaluated time to reach goal MAP and safety outcomes after implementation of an AVP-sparing guideline. Moreover, a cost analysis showed that patients in the preintervention group received an average of 4 vials of AVP during the admission, resulting in a cumulative cost of $42 157.50. Use of similar protocols at other institutions might be expected to result in substantial cost savings. However, our study also has some limitations. Given the retrospective nature of this study, it is possible that selection bias may have affected our findings. The preintervention group may have included sicker patients, as seen by the higher APACHE II score of 23 compared with 18. We were also unable to evaluate for guideline adherence in this study. The median dose of NE at which AVP was initiated in the preintervention cohort was 18 µg/min, which is higher than the protocol recommendation of initiating AVP at a NE dose of 10 µg/min. The use of AVP in the preintervention group may have been reserved for more critically ill patients, as observed by differences in variables such as the APACHE II scores. We attempted to account for this potential selection bias through multivariate analysis using the Cox proportional hazard model as well as through multiple linear and logistic regressions by accounting for the type of intervention received. Because this study was conducted in a large academic tertiary medical center, many patients are referrals from outside facilities (52.7% of patients in the preintervention group and 37.8% of patients in the postintervention group). Although patients who received vasopressors at an outside hospital were excluded, we were unable to

18 Annals of Pharmacotherapy 51(1) Table 2. Clinical Outcomes. Preintervention (n = 74) Postintervention (n = 74) P Value Adjusted OR 95% CI P Value Time to reach goal MAP (hours) a 2 (1.0-3.6) 1.3 (1.0-2.2) 0.030 1.412 0.97-2.05 0.07 MAP (mm Hg) a At initiation of NE 56 (52-60) 56 (51-60) 0.667 At initiation of AVP 61 (57-65) 6 hours 66 (60-74) 66 (60-71) 0.914 12 hours 68 (63-74) 68 (63-73) 0.859 18 hours 66 (61-73) 67 (62-74) 0.361 24 hours 68 (62-74) 66 (62-74) 0.323 30 hours 67 (63-73) 65 (62-73) 0.483 36 hours 68 (63-73) 69 (62-76) 0.782 42 hours 65 (62-73) 68 (64-73) 0.487 48 hours 65 (62-68) 65 (61-73) 0.409 Heart rate (bpm) a At initiation of NE 98 (81-114) 92 (79-108) 0.185 At initiation of AVP 104 (84-118) 6 hours 98 (81-112) 91 (76-106) 0.095 12 hours 97 (81-111) 90 (77-106) 0.090 18 hours 95 (79-111) 89 (77-108) 0.389 24 hours 100 (78-108) 94 (75-114) 0.974 30 hours 99 (76-111) 95 (77-115) 0.866 36 hours 98 (74-109) 91 (80-109) 0.803 42 hours 94 (73-107) 95 (84-103) 0.887 48 hours 98 (71-109) 91 (80-107) 1.000 NE doses (µg/min) a Initiation 10.0 (5.0-12.8) 5.0 (4.0-8.0) <0.001 Maximum 24.0 (20.0-33.5) 11.0 (6.0-20.0) <0.001 At initiation of AVP 18.0 (14.0-24.0) NE doses (µg/kg/min) a Initiation 0.106 (0.071-0.195) 0.068 (0.046-0.104) <0.001 Maximum 0.325 (0.231-0.392) 0.145 (0.093-0.257) <0.001 At initiation of AVP 0.233 (0.161-0.330) Duration (hours) a NE 34.3 (21.0-59.1) 26.0 (13.1-61.7) 0.094 AVP 28.8 (9.1-52.4) NE and AVP 25.5 (9.0-43.3) Crystalloid intake at 6 hours a 2.0 (1.0-2.0) 2.0 (1.0-2.6) 0.655 Urine output (ml/kg/h) a At initiation of NE 0.17 (0.04-0.45) 0.29 (0.07-0.75) 0.047 At initiation of AVP 0.27 (0.06-0.47) 6 hours 0.24 (0.05-0.53) 0.41 (0.08-1.08) 0.041 12 hours 0.26 (0.08-0.57) 0.42 (0.05-0.92) 0.168 18 hours 0.31 (0.09-0.90) 0.32 (0.04-0.83) 0.573 24 hours 0.38 (0.06-1.10) 0.38 (0.03-0.74) 0.408 30 hours 0.36 (0.06-0.69) 0.28 (0.02-0.84) 0.912 36 hours 0.35 (0.04-0.68) 0.51 (0.08-0.94) 0.204 42 hours 0.18 (0.03-0.60) 0.49 (0.09-0.91) 0.175 48 hours 0.12 (0.01-0.92) 0.44 (0.09-1.02) 0.338 Maximum lactate (mg/dl) a 3.2 (1.8-6.0) 1.3 (0.6-2.4) <0.001 Concomitant medications Steroids 16 (21.6%) 15 (20.3%) 0.840 Dobutamine 1 (1.4%) 2 (2.7%) 0.560 Milrinone 0 (0%) 1 (1.4%) 0.316 (continued)

Wu et al 19 Table 2. (continued) Preintervention (n = 74) Postintervention (n = 74) P Value Adjusted OR 95% CI P Value Blood transfusions 34 (45.9%) 29 (39.2%) 0.406 Cardiac arrest 7 (9.5%) 1 (1.4%) 0.029 New arrhythmia 0.567 Atrial fibrillation 8 (10.8%) 7 (9.5%) Ventricular fibrillation 0 (0%) 0 (0%) Ventricular tachycardia 3 (4.1%) 1 (1.4%) Cumulative AVP cost $42 157.50 $0 Mechanical ventilation 36 (48.6%) 36 (48.6%) 1.000 Dialysis 19 (25.7%) 6 (8.1%) 0.004 Hospital length of stay a 9 (5-16) 11 (7-19) 0.167 Pre-Post group 10 0.30-341.38 0.20 APACHE II score 0.97 0.79-1.1 0.78 NE dose initiation 1.87 1.42-2.48 <0.0001 ICU length of stay a 6 (4-9) 7 (4-11) 0.474 Pre-Post group 10.45 0.11-1026.59 0.31 APACHE II score 0.93 0.72-1.21 0.60 NE dose initiation 1.99 1.38-2.86 <0.0001 28-Day mortality 38 (51.4%) 21 (28.4%) 0.004 Pre-Post group 0.34 0.16-0.71 0.004 APACHE II score 1.03 0.98-1.07 0.22 NE dose initiation 0.95 0.89-1.01 0.10 Abbreviations: APACHE, Acute Physiologic Chronic Health Evaluation; AVP, vasopressin; ICU, intensive care unit; MAP, mean arterial pressure; NE, norepinephrine; OR, odds ratio. a Median (interquartile range). Multivariate analyses adjusted for the intervention, APACHE II score, and the NE dose at initiation, and values shown are ORs for the specific variables. collect data on fluid status and determine if patients were appropriately fluid resuscitated in those included. However, the median amount of crystalloids received within 6 hours of starting NE therapy was 2 L in both groups, suggesting that patients received the recommended 30-mL/kg crystalloid for septic shock. 1 Furthermore, this study was only conducted in the medical ICU, which limits the generalizability of the results of this study to other, non-medical ICUs. Finally, because of the study design, an a priori power calculation was not performed. However, with the number of patients available in each arm, the study has 80% power to detect a difference of half an hour in time to reach goal MAP, with a 2-sided alpha of 0.05, assuming a baseline mean of 2 hours with a SD of 0.9 hours to reach goal MAP. In conclusion, there is no difference in the time to reach goal MAP between patients who receive NE <50 µg/min with or without AVP. This AVP-sparing guideline showed little difference in efficacy, safety, and clinical outcomes after adjusting for baseline differences. The guideline revision was associated with decreased mortality after the multivariate regression analysis, although our study was not powered for this outcome. Large, prospective, randomized trials are necessary to identify the dosage of NE at which AVP should be initiated. Declaration of Conflicting Interests The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article. Funding The authors received no financial support for the research, authorship, and/or publication of this article. References 1. Dellinger RP, Levy MM, Rhodes A, et al. Surviving sepsis campaign: international guidelines for management of severe sepsis and septic shock: 2012. Crit Care Med. 2013;41:580-637. 2. Dunser MW, Mayr AJ, Ulmer H, et al. Arginine vasopressin in advanced vasodilatory shock: a prospective, randomized, controlled study. Circulation. 2003;107:2313-2319. 3. Micek ST, Shah P, Hollands JM, Shah RA, Shannon WD, Kollef MH. Addition of vasopressin to norepinephrine as independent predictor of mortality in patients with refractory septic shock: an observational study. Surg Infect (Larchmt). 2007;8:189-200. 4. Dopp-Zemel D, Groeneveld AB. High-dose norepinephrine treatment: determinants of mortality and futility in critically ill patients. Am J Crit Care. 2013;22:22-32. 5. Jenkins CR, Gomersall CD, Leung P, Joynt GM. Outcome of patients receiving high dose vasopressor therapy: a retrospective cohort study. Anaesth Intensive Care. 2009;37:286-289.

20 Annals of Pharmacotherapy 51(1) 6. Benbenishty J, Weissman C, Sprung CL, Brodsky-Israeli M, Weiss Y. Characteristics of patients receiving vasopressors. Heart Lung. 2011;40:247-252. 7. Russell JA, Walley KR, Singer J, et al. Vasopressin versus norepinephrine infusion in patients with septic shock. N Engl J Med. 2008;358:877-887. 8. Stollings J, Mitchell D, Rice T, Wheeler A. An interdisciplinary project to simplify vasopressor use in a medical intensive care unit. Crit Care Med. 2013;41(12): 331. 9. Varpula M, Tallgren M, Saukkonen K, Voipio-Pulkki LM, Pettila V. Hemodynamic variables related to outcome in septic shock. Intensive Care Med. 2005;31:1066-1071. 10. Dunser MW, Takala J, Ulmer H, et al. Arterial blood pressure during early sepsis and outcome. Intensive Care Med. 2009;35:1225-1233. 11. Holmes CL, Walley KR, Chittock DR, Lehman T, Russell JA. The effects of vasopressin on hemodynamics and renal function in severe septic shock: a case series. Intensive Care Med. 2001;27:1416-1421. 12. Patel BM, Chittock DR, Russell JA, Walley KR. Beneficial effects of short-term vasopressin infusion during severe septic shock. Anesthesiology. 2002;96:576-582. 13. Gordon AC, Russell JA, Walley KR, et al. The effects of vasopressin on acute kidney injury in septic shock. Intensive Care Med. 2010;36:83-91. 14. Gordon AC, Mason AJ, Thirunavukkarasu N, et al. Effect of early vasopressin vs norepinephrine on kidney failure in patients with septic shock. JAMA. 2016;316:509-518. 15. Luckner G, Dunser MW, Jochberger S, et al. Arginine vasopressin in 316 patients with advanced vasodilatory shock. Crit Care Med. 2005;33:2659-2666. 16. Luckner G, Mayr VD, Jochberger S, et al. Comparison of two dose regimens of arginine vasopressin in advanced vasodilatory shock. Crit Care Med. 2007;35:2280-2285. 17. Malay MB, Ashton RC Jr, Landry DW, Townsend RN. Lowdose vasopressin in the treatment of vasodilatory septic shock. J Trauma. 1999;47:699-703; discussion 703-705. 18. Knotzer H, Maier S, Dunser MW, et al. Arginine vasopressin does not alter mucosal tissue oxygen tension and oxygen supply in an acute endotoxemic pig model. Intensive Care Med. 2006;32:170-174. 19. Russell JA, Walley KR, Gordon AC, et al. Interaction of vasopressin infusion, corticosteroid treatment, and mortality of septic shock. Crit Care Med. 2009;37:811-818. 20. Bauer SR, Lam SW, Cha SS, Oyen LJ. Effect of corticosteroids on arginine vasopressin-containing vasopressor therapy for septic shock: a case control study. J Crit Care. 2008;23:500-506. 21. Torgersen C, Luckner G, Schroder DC, et al. Concomitant arginine-vasopressin and hydrocortisone therapy in severe septic shock: association with mortality. Intensive Care Med. 2011;37:1432-1437. 22. Gordon AC, Mason AJ, Perkins GD, et al. The interaction of vasopressin and corticosteroids in septic shock: a pilot randomized controlled trial. Crit Care Med. 2014;42: 1325-1333.