Effect of norepinephrine versus vasopressin weaning on incidence of hypotension in septic shock patients: a systematic review and meta-analysisABSTRACT
Objective: To evaluate the influence of weaning vasopressors – norepinephrine or vasopressin - on deleterious effects; the main objective was to evaluate the incidence of hypotension.
Methods: We performed a systematic review and meta-analysis of 11 studies - 2 randomized controlled trials and 9 observational studies - involving 2,280 patients in whom norepinephrine (n = 1,254) or vasopressin (n = 1,026) was withdrawn first to compare the risk of hypotension and other adverse effects.
Results: Hypotension occurred in 33.4% (420/1,254) of patients in the Norepinephrine Group and in 52,4% (538/1,026) of patients in the Vasopressin Group. There was no difference between groups regarding the incidence of hypotension: OR 0.43 (95%CI 0.18 - 1.03). The subgroup analysis of observational studies suggests a lower incidence of hypotension when norepinephrine is weaned first (OR 0.26; 95%CI 0.13 - 0.53), with high heterogeneity between studies (I2 = 88%; p < 0.01). On the other hand, the subgroup of randomized controlled trials indicates a higher incidence of hypotension when norepinephrine is weaned first (OR 4.04; 95%CI 1.24 - 13.15; I2 = 65%, p = 0.09). The weaning strategy was not associated with other outcomes, including intensive care unit length of stay, hospital length of stay, time on vasopressors, arrhythmias, renal injury, or SOFA.
Conclusion: No differences in hypotension incidence were observed when weaning was initiated with either norepinephrine or vasopressin. However, the high heterogeneity observed across studies warrants caution in drawing definitive conclusions.
Keywords:
Shock; Hypotension; Sepsis; Vasoconstrictors; Norepinephrine; Vasopressin
INTRODUCTION
Norepinephrine (NE) is the preferred pharmacological agent for managing septic shock, as recommended by the Surviving Sepsis Campaign.(1) The concomitant use of arginine vasopressin (AVP) is advised, albeit supported by evidence of moderate quality.(2) Although guidelines exist for the initiation of vasopressors, there is a lack of consensus regarding their tapering. The optimal strategy for vasopressor weaning is currently a research priority in sepsis management.(3)
The observational study conducted by Song et al.,(4) which represents the largest cohort examined to date, reported a higher incidence of hypotension when AVP was discontinued first. Conversely, the initial randomized clinical trial on this subject reported findings that were contrary.(5) The shorter half-life of NE may predispose to hypotension. Conversely, there is speculation of a relative AVP deficiency in the early phase of shock.(6) Furthermore, the clinical implications of hypotension during the stabilization phase remain ambiguous.
Norepinephrine is typically tapered in a titrated manner. However, a review of observational studies indicates that abrupt AVP withdrawal is a common practice.(7) The method of withdrawal may affect the incidence of hypotension observed.
This systematic review and meta-analysis aimed to evaluate the influence of weaning vasopressors – NE or vasopressin - on deleterious effects; the main objective was to evaluate the incidence of hypotension.
METHODS
This systematic review was conducted to evaluate the effect of discontinuing NE first versus AVP first in adult patients with septic shock who were receiving both vasopressors and were eligible for weaning. A protocol for this review was registered on the PROSPERO website (ID CRD42024624101). This manuscript follows the latest Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines.
Eligibility criteria
Eligible studies included original research on adult patients with septic shock who were receiving both NE and AVP and were eligible for vasopressor weaning. Studies were included if they compared a strategy of discontinuing NE first to a strategy of discontinuing AVP first. There were no restrictions on the type of weaning strategy, including abrupt versus gradual discontinuation. Both randomized controlled trials (RCTs) and comparative observational studies were included. Conference abstracts were excluded, and there was no restriction on publication year. Studies had to clearly differentiate between groups receiving NE-first and AVP-first weaning approaches. Studies without a comparator group, those published in languages other than English, Portuguese, or Spanish, and those that did not report relevant outcomes related to vasopressor weaning were excluded. The PICO strategy was:
-
–
Population: adult septic shock patients who were receiving both NE and AVP and were eligible for vasopressor weaning.
-
–
Intervention: a strategy of discontinuing NE first - "NE first".
-
–
Comparative: a strategy of discontinuing AVP first - "AVP first".
-
–
Outcome: incidence of hypotension, as defined in each study.
Information sources and search strategy
The literature was searched by a medical librarian for the concepts of NE, AVP, and septic shock. Search strategies were created using a combination of keywords and standardized index terms. Searches were run on February 5, 2024 in ClinicalTrials.gov (2000+), Ovid Cochrane Central Register of Controlled Trials (1991+), Ovid Embase (1974+), Ovid Medline (1946+ including epub ahead of print, in-process & other non-indexed citations), Scopus (1788+), Web of Science Core Collection (Science Citation Index Expanded 1975+ & Emerging Sources Citation Index 2015+) AND the World Health Organization's ICTRP trial registry (2005+). The whole search strategy is described in table 1S (Supplementary Material). After limiting results to English, Portuguese, and Spanish, and removing animal and pediatric studies based on the exclusion criteria, a total of 4,225 citations were retrieved. A clinical trial by the authors, at the time of the search, which had not yet been published, was included in the analysis. Deduplication was performed automatically in Covidence, leaving 3,069 citations for screening.
Study selection
Pairs of reviewers independently screened titles and abstracts identified by the search strategy to determine potential eligibility. Full-text articles of potentially eligible studies were assessed independently by two reviewers to confirm inclusion. Any disagreements were resolved through discussion and consensus. If necessary, a third reviewer was consulted to resolve disagreements.
Data collection and outcomes
A standardized data extraction form was developed to ensure consistency in data collection. Two investigators initially extracted the data, and a third reviewer cross-checked it for accuracy. Disagreements were resolved by consensus. Extracted data included study design, sample size, study setting, population characteristics, and details of the interventions.
Extracted variables included patient age, sex, setting (intensive care unit [ICU] or other hospital locations), use of mechanical ventilation, and infection source (categorized as respiratory versus non-respiratory).
The outcome of interest was the incidence of hypotension after vasopressor weaning, as defined by each included study.
Secondary outcomes were: length of stay (LOS) in the hospital (in days), LOS in the ICU (in days), time on vasopressor use (in hours), new arrhythmia onset, as defined by each study; kidney injury onset, as defined by each study; and Sequential Organ Failure Assessment (SOFA) score.
Risk of bias assessment
Risk of bias was evaluated at the outcome level for each eligible study following training and calibration exercises. For RCTs, the risk of bias was assessed by using version 2 of the Cochrane risk of bias tool.(8) For observational studies, the modified Newcastle-Ottawa scale tool was used.(9) The assessment of risk of bias was performed by one investigator, and the results were subsequently cross-checked by a methodologist.
Effect measures and analysis
For conducting meta-analyses, the meta package from RStudio (R version 4.2.2) was utilized. The meta-analyses were performed using a random-effects model with Restricted Maximum Likelihood (REML). Compared to the traditional Der Simonian-Laird method, REML typically provides a more conservative estimate of the overall effect, as it often yields a larger estimate of tau,(2) resulting in wider confidence intervals due to its greater accounting for uncertainty. Data described as median and interquartile range were converted to mean and standard deviation, in the methodology described by Wan et al.(10)
The pooled effect estimates for dichotomous outcomes were reported as odds ratios (ORs) with 95% confidence intervals (95%CIs), and those for continuous outcomes were reported as mean differences (MDs) with 95%CIs. We analyzed the data using a random-effects model. Statistical heterogeneity among studies was assessed using the I² statistic, as proposed by Higgins et al.,(11) and visual evaluation was also conducted to assess between-study heterogeneity. To explore a possible cause of heterogeneity, we performed a subgroup analysis, evaluating RCTs and observational studies within each subgroup. We defined a p < 0.05 as statistically significant. In our protocol, we did not plan to perform sensitivity analyses, as this type of analysis cannot provide definitive conclusions regarding the benefit of the intervention.
Certainty assessment
The certainty of the evidence was evaluated using the Grading of Recommendations Assessment, Development, and Evaluation (GRADE) approach. This assessment considered factors such as study design, risk of bias, inconsistency, indirectness, and imprecision of the evidence.(12-14)
Methodological quality
The methodological quality of the included studies varied considerably (Tables 2S and 3S - Supplementary Material). Two RCTs were assessed as having a low risk of bias. Conversely, the nine observational studies were all judged to be at high risk of bias, primarily due to concerns across multiple domains. Specifically, issues were identified in participant selection, group comparability, and outcome assessment. These methodological limitations within the observational studies raise serious concerns about the validity of their findings. The high risk of bias may have led to over- or underestimation of the actual effect and potentially introduced significant confounding, thereby compromising the internal validity of these studies (Table 1). Consequently, the overall quality of the evidence from observational studies is very low, and their results should be interpreted with extreme caution. While the two RCTs offer more robust evidence, the limited number of these studies restricts the overall strength of the conclusions that can be drawn. No conflicts of interest or specific funding sources were identified for any of the included studies.
RESULTS
The flowchart (Figure 1) summarizes the number of manuscripts at each step of the corpus selection process. Eleven studies met the inclusion criteria and were included in the systematic review. The main characteristics of the studies included are summarized in table 2. The 11 studies included in this meta-analysis comprised 2 RCTs(5,22) and 9 observational studies,(4,15-23) with a total sample size of 2,280 patients. All observational studies were carried out in the United States.(4,15-23) Only two RCTs evaluated the incidence of hypotension after weaning from NE or AVP first: one conducted in South Korea(5) and the other in Brazil.(22) All are unicentric studies, with the exception of one observational study that collected data in three hospitals.(20) Besides hypotension, these studies evaluated other outcomes like time on vasopressor,(5,15,17-19,21-22) LOS (ICU, hospital),(4,5,15-17,19-22) mortality (ICU, hospital),(4,5,15-19,21,22) incidence of arrhythmias,(4,17,21,22) SOFA score,(4,5,15,16,18-21,22) kidney injury and hemodialysis,(4,19,21,22) readmission to ICU(4) and incidence of mechanical ventilation.(21)
The proportion of female patients ranged from 32%(15) to 57%,(22) and the mean age of the included patients ranged from 50 years(17) to 75 years.(19) The SOFA score at inclusion varied between 6(20) and 12.(22) Four studies evaluated exclusively clinical patients(5,16,17,22) and another four studies evaluated clinical-surgical patients.(15,18-20) In five studies, the main septic foci were pulmonary.(5,16,19,22,23) Maximal NE dose described in each study varied between 0.2µg/kg/min(19) and 0.7µg/kg/min;(22) maximal AVP dose was 0.04U/min, described in four studies.(4,19,21,22)
Regarding AVP weaning, there is divergence among the methods. In both clinical trials,(5,22) AVP was titrated down. In observational studies, some observed an abrupt withdrawal as the usual practice,(4,17-19,23) while others did not even describe the withdrawal method.(15,16,20,21)
Main outcome
The forest plot (Figure 2) summarizes the results of a meta-analysis comparing two treatment strategies, NE first and AVP first, across multiple studies. There was no difference between groups regarding the incidence of hypotension: OR 0.43 (95%CI 0.18 - 1.03; I2 = 91%; p < 0.01). The subgroup analysis of observational studies suggests a lower incidence of hypotension when NE is weaned first (OR 0.26; 95%CI 0.13 - 0.53), with high heterogeneity across studies. On the other hand, subgroup analysis of RCTs indicates a higher incidence of hypotension when NE is weaned first (OR 4.04; 95%CI 1.24 - 13.15).
The effect of different vasopressor discontinuation orders on the incidence of hypotension in patients with septic shock.
The meta-analysis indicated a significant effect size, suggesting that study design influences effect size. The observational studies show high heterogeneity (I2 = 88%; p < 0.01), suggesting considerable variability across studies, whereas the randomized trials show lower heterogeneity (I2 = 65%; p = 0.09), indicating moderate consistency. There was significant heterogeneity between the subgroups (observational studies and RCTs): p < 0.01 for subgroup differences.
Secondary outcomes
The results of the meta-analysis performed in the secondary outcomes are described in the table 3. The NE first strategy was associated with an increase in hospital LOS, but not in ICU LOS or time on vasopressors. There was no difference between groups regarding SOFA score or the onset of new arrhythmia or renal injury (Figures 1S to 6S - Supplementary Material).
DISCUSSION
The understanding of the deresuscitation phase remains nascent and lacks a robust scientific foundation. In this meta-analysis of patients experiencing septic shock, a trend towards an increased incidence of hypotension was observed upon the initiation of AVP withdrawal. However, stratified analysis reveals a significant discrepancy between findings from observational studies and those from clinical trials. The aggregated analysis of observational studies indicates a preference for primary NE withdrawal, whereas clinical trials suggest a higher incidence of hypotension when NE withdrawal is initiated. In seeking an explanation for this divergence, one notable aspect is the method of AVP withdrawal. Among observational studies, although most do not provide numerical descriptions, abrupt AVP withdrawal is noted. Hammond et al.(7) conducted a questionnaire-based study in North American hospitals, revealing that approximately 70% of physicians abruptly discontinue AVP without dose titration. All observational studies on vasopressor weaning have been conducted in North America. The practice of abrupt AVP withdrawal, rather than titration, may account for the higher incidence of hypotension observed in observational studies, in contrast to RCTs.(5,23)
A key point of this study is that it represents the first meta-analysis to incorporate the results of the recently published RENOVA trial, the second clinical trial addressing vasopressor weaning. It is noteworthy that the results of the RENOVA trial are consistent with those observed in the DOVSS trial, the first clinical trial on this topic.
In secondary analysis, a NE-first strategy was associated with a brief but increased hospital LOS, though not with prolonged vasopressor use or ICU LOS. At this juncture, no clinical significance for this finding could be determined. Additionally, the possibility that hospitalization times and vasopressor use may not be competing outcomes with mortality across the groups cannot be excluded. The use of AVP in shock is associated with a lower incidence of arrhythmias(24) and potentially a reduced incidence of acute kidney injury;(25) however, no higher incidence of these outcomes was identified when AVP is first withdrawn from shock management, suggesting that withdrawing NE before AVP should not be associated with increased safety in preventing arrhythmias or acute kidney injury.
This study has limitations that warrant careful consideration when interpreting the results. Notably, significant heterogeneity and discrepancies in the description of other outcomes complicate joint analysis. Consequently, the evaluation was confined to the outcome of hypotension. A potential limitation of our study was that we did not search the gray literature, which might have contributed to a more comprehensive assessment of the available evidence.
CONCLUSION
In our meta-analysis, no difference in hypotension incidence was detected when weaning was initiated with norepinephrine or vasopressin. However, the high heterogeneity observed across studies warrants caution in drawing definitive conclusions. The subgroup analysis of observational studies suggests a lower incidence of hypotension when norepinephrine is weaned first. In contrast, randomized controlled trials indicate a higher incidence of hypotension with norepinephrine-first withdrawal. It is plausible to hypothesize that this discrepancy in findings may be attributed to the methodology employed in vasopressin withdrawal, namely abrupt rather than gradual discontinuation. These findings highlight the imperative for more rigorous randomized controlled trials to elucidate the optimal strategy for tapering vasopressors in the management of septic shock.
Availability of data and materials
Data is available on demand from referees.
REFERENCES
- 1 Evans L, Rhodes A, Alhazzani W, Antonelli M, Coopersmith CM, French C, et al. Surviving sepsis campaign: international guidelines for management of sepsis and septic shock 2021. Intensive Care Med. 2021;47(11):1181-247.
- 2 Russell JA, Walley KR, Singer J, Gordon AC, Hébert PC, Cooper DJ, et al.; VASST Investigators. vasopressin versus norepinephrine infusion in patients with septic shock. N Engl J Med. 2008;358(9):877-87.
- 3 De Backer D, Deutschman CS, Hellman J, Myatra SN, Ostermann M, Prescott HC, et al.; Surviving Sepsis Campaign Research Committee. Surviving Sepsis Campaign Research Priorities 2023. Crit Care Med. 2024;52(2):268-96.
- 4 Song X, Liu X, Evans KD, Frank RD, Barreto EF, Dong Y, et al. The order of vasopressor discontinuation and incidence of hypotension: a retrospective cohort analysis. Sci Rep. 2021;11(1):16680.
- 5 Jeon K, Song JU, Chung CR, Yang JH, Suh GY. Incidence of hypotension according to the discontinuation order of vasopressors in the management of septic shock: a prospective randomized trial (DOVSS). Crit Care. 2018;22(1):131.
- 6 Landry DW, Levin HR, Gallant EM, Ashton RC Jr, Seo S, D’Alessandro D, et al. Vasopressin deficiency contributes to the vasodilation of septic shock. Circulation. 1997;95(5):1122-5.
- 7 Hammond DA, Rech MA, Daley MJ, Devlin JW, Hodge EK, Kooda KJ, et al. Perceptions regarding vasopressin use and practices in septic shock, and cost containment strategies. J Am Coll Clin Pharm. 2019;2(3):257-67.
- 8 Sterne JA, Savović J, Page MJ, Elbers RG, Blencowe NS, Boutron I, et al. RoB 2: a revised tool for assessing risk of bias in randomised trials. BMJ. 2019;366:l4898.
-
9 Wells GA, Shea B, O’Connell D, Peterson J, Welch V, Losos M, et al. The Newcastle-Ottawa Scale (NOS) for assessing the quality of nonrandomized studies in meta analysis. 2004. [cited 2023June 13]. Available from: https://www.ohri.ca/programs/clinical_epidemiology/oxford.asp
» https://www.ohri.ca/programs/clinical_epidemiology/oxford.asp - 10 Wan X, Wang W, Liu J, Tong T. Estimating the sample mean and standard deviation from the sample size, median, range and/or interquartile range. BMC Med Res Methodol. 2014;14(1):135.
- 11 Higgins JP, Thompson SG. Quantifying heterogeneity in a meta-analysis. Stat Med. 2002;21(11):1539-58.
- 12 Atkins D, Best D, Briss PA, Eccles M, Falck-Ytter Y, Flottorp S, et al.; GRADE Working Group. Grading quality of evidence and strength of recommendations. BMJ. 2004;328(7454):1490.
- 13 Balshem H, Helfand M, Schünemann HJ, Oxman AD, Kunz R, Brozek J, et al. GRADE guidelines: 3. Rating the quality of evidence. J Clin Epidemiol. 2011;64(4):401-6.
- 14 Murad MH, Mustafa RA, Schünemann HJ, Sultan S, Santesso N. Rating the certainty in evidence in the absence of a single estimate of effect. Evid Based Med. 2017;22(3):85-7.
- 15 Bauer SR, Aloi JJ, Ahrens CL, Yeh JY, Culver DA, Reddy AJ. Discontinuation of vasopressin before norepinephrine increases the incidence of hypotension in patients recovering from septic shock: a retrospective cohort study. J Crit Care. 2010;25(2):362.e7-11.
- 16 Hammond DA, McCain K, Painter JT, Clem OA, Cullen J, Brotherton AL, et al. Discontinuation of vasopressin before norepinephrine in the recovery phase of septic shock. J Intensive Care Med. 2019;34(10):805-10.
- 17 Bissell BD, Magee C, Moran P, Bastin ML, Flannery AH. Hemodynamic instability secondary to vasopressin withdrawal in septic shock. J Intensive Care Med. 2019;34(9):761-5.
- 18 Sacha GL, Lam SW, Duggal A, Torbic H, Reddy AJ, Bauer SR. Hypotension risk based on vasoactive agent discontinuation order in patients in the recovery phase of septic shock. Pharmacotherapy. 2018;38(3):319-26.
- 19 Musallam N, Altshuler D, Merchan C, Zakhary B, Aberle C, Papadopoulos J. Evaluating vasopressor discontinuation strategies in patients with septic shock on concomitant norepinephrine and vasopressin infusions. Ann Pharmacother. 2018;52(8):733-9.
- 20 Bredhold BE, Winters SD, Callison JC Jr, Heidel RE, Allen LM, Hamilton LA. Impact of the sequence of norepinephrine and vasopressin discontinuation in patients recovering from septic shock. Hosp Pharm. 2020;55(1):26-31.
- 21 Taylor A, Jones T, Forehand CC, Smith SE, Dykes H, Newsome AS. Vasopressor discontinuation order in septic shock with reduced left ventricular function. J Pharm Pract. 2022;35(6):879-85.
- 22 Mallmann C, Galiotto TM, de Oliveira MS, Moraes RB. Reduction of norepinephrine versus vasopressin in the stabilization phase of septic shock: RENOVA clinical trial. Med Intensiva (Engl Ed). 2025;49(10):502147.
- 23 Buckley CT, Turner B, Walsh D, Garrett MJ, Ooka VN. Analysis of vasopressor discontinuation and the incidence of rebound hypotension in patients with septic shock. Hosp Pharm. 2021;56(1):47-53.
- 24 McIntyre WF, Um KJ, Alhazzani W, Lengyel AP, Hajjar L, Gordon AC, et al. Association of vasopressin plus catecholamine vasopressors vs catecholamines alone with atrial fibrillation in patients with distributive shock: a systematic review and meta-analysis. JAMA. 2018;319(18):1889-900.
- 25 Nedel WL, Rech TH, Ribeiro RA, Pellegrini JA, Moraes RB. Renal outcomes of vasopressin and its analogs in distributive shock: a systematic review and meta-analysis of randomized trials. Crit Care Med. 2019;47(1):e44-51.
Edited by
-
Responsible editor:
Elisa Estenssoro https://orcid.org/0000-0002-8896-5688
