Critical Care

Effect of early vasopressin combined with norepinephrine on short-term mortality in septic shock: A retrospective study based on the MIMIC-IV database

a b s t r a c t

Background: Septic shock is a leading cause of death in Intensive care units , with short-term mortality rates of 35-40%. Vasopressin (AVP) is a second-line vasoactive agent for septic shock, and recent studies suggest that early AVP use can be beneficial. However, differences between early initiation of AVP combined with norepi- nephrine (NE) and nonearly AVP with NE are unclear. A retrospective cohort research was designed to explore the effects of early AVP initiation versus nonearly AVP initiation.

Methods: This retrospective single-center cohort study included adult patients with septic shock from the MIMIC (Medical Information Mart for Intensive Care)-IV database. According to whether AVP was used early in the ICU , patients were assigned to the early- (within 6 h of septic shock onset) and non-early-AVP (at least 6 h after septic shock onset) groups. The primary outcome was 28-day mortality. The secondary out- comes were ICU and hospital mortality, the numbers of vasopressor-free and ventilation-free days at 28 days, ICU length of stay (LOS), hospital LOS, Sequential Organ Failure Assessment score on days 2 and 3, and Renal replacement therapy use on days 2 and 3. Univariate and multivariate cox Proportional-hazards regression, propensity-score matching were used to analyze the differences between the groups.

Results: The study included 531 patients with septic shock: 331 (62.5%) in the early-AVP group and 200 (37.5%) in the non-early-AVP group. For 1:1 matching, 158 patients in the early-AVP group were matched with the same number of patients with nonearly AVP. Regarding the primary outcome, there was no significant difference between the early- and non-early-AVP groups in 28-day mortality (hazard ratio [HR] = 0.91, 95% confidence interval [CI] = 0.68-1.24). For the secondary outcomes, there were no differences between the early- and non-early-AVP groups in ICU mortality (HR = 0.95, 95% CI = 0.67-1.35), hospital mortality (HR = 0.95, 95% CI = 0.69-1.31), the numbers of vasopressor-free and ventilation-free days at 28 days, ICU LOS, hospital LOS, SOFA score on days 2 and 3, and RRT use on days 2 and 3.

Conclusions: There was no difference in short-term mortality between early AVP combined with NE and nonearly AVP with NE in septic shock.

(C) 2023 Published by Elsevier Inc.

  1. Introduction

Septic shock is a life-threatening multiple-organ dysfunction syn- drome caused by infectious disease [1]. It is a more-severe subtype of sepsis, with short-term mortality rates as high as 40% [2], which causes

* Corresponding authors.

E-mail addresses: [email protected] (H. Yin), [email protected] (J. Lyu).

1 Co-first author: Dan He, Luming Zhang, Hai Hu.

enormous health and Economic burdens on both patients and healthcare systems.

fluid management and vasoactive agents are two key parts of resus- citation in septic shock. Norepinephrine (NE) was recommended as the first-line vasoactive agent by the Surviving Sepsis Campaign in 2021 [3]. However, a high NE dose is associated with extreme vasoconstriction, tissue hypoperfusion, and increased mortality risk [4]. Therefore, when an adequate mean arterial pressure (MAP) is difficult to maintain in patients with septic shock who receive high NE doses, there is 0735-6757/(C) 2023 Published by Elsevier Inc.

moderate-quality evidence for the combination of vasopressin (AVP) and NE to reduce the side effects of these high doses [5].

Several studies have demonstrated that septic shock has a transient high plasma AVP concentration in its early stage, but the plasma AVP decreases to very low levels in established septic shock [6]. A study of 19 patients further found that AVP deficiency was associated with the hypotension in vasodilatory septic shock [7]. Several recent studies have explored the association between early AVP use and septic shock outcomes. The VANISH trial compared the early use of AVP with NE and did not find a difference in the number of kidney-failure-free days; however, the confidence interval (CI) included a potential clinically important benefit for early AVP use [8]. A meta-analysis found that early AVP initiation within 6 h in septic shock was associated with reduced Renal replacement therapy use. These previous studies compared early AVP and/or NE with NE alone within 6 h of a septic shock diagnosis [9]. However, it is not clear whether the early initiation of NE combined with AVP is superior to nonearly combined use in septic shock. Therefore, a retrospective cohort study was designed to compare the effect of early AVP initiation (within 6 h of septic shock onset) with nonearly AVP (at least 6 h after septic shock onset).

  1. Methods
    1. Data source and study population

This study followed the guidelines of the STROBE (Strengthening the Reporting of Observational Studies in Epidemiology) initiative [10]. The study was exempt from approval by our Institutional Review Board (IRB) because it used a public, deidentified database with prior IRB ap- proval. The author passed the relevant training course for accessing and extracting the database and obtained the relevant certificate (No. 47907567). Patients diagnosed with septic shock during 2008-2019 in version 2.0 of the Medical Information Mart for Intensive Care (MIMIC)-IV database were selected [11,12]. Septic shock was defined as patients who were resuscitated with the appropriate fluid, but vaso- pressors were still required to maintain MAP at >65 mmHg, and with serum lactate higher than 2.0 mmol/L based on the Sepsis-3 definition [1]. ICD9 and ICD10 code were applied to identify patients with septic shock (Table S1). The inclusion criteria were as follows: first intensive care unit admission, NE and AVP initiation during ICU stay, older than 18 years, ICU stay of at least 1 day, and an initial NE dose higher than 0.1 ug/kg/min. The exclusion criteria were first care unit in cardiac intensive care unit (CCU) or cardiovascular intensive care unit (CVICU), and no height record. As shown in Fig. 1, we finally included 531 pa- tients: 331 (62.3%) in the early-AVP group and 200 (37.7%) in the non-early-AVP group.

    1. Data extraction

Patient information was extracted using Structured Query Language and Navicat Premium software (version 15.0). Demographic informa- tion included age, sex, body mass index (BMI), ethnicity, first care unit, admission type, and insurance status. Mean values of vital signs on the first day in the ICU were recorded, including heart rate, mean blood pressure (MBP), respiratory rate, temperature, and oxygen satu- ration. The Illness severity scores included Sequential Organ Failure Assessment score, Acute Physiology Score III (APSIII), Systemic Inflammatory Response Syndrome score, and Charlson Comor- bidity Index. Comorbidities included myocardial infarction, congestive heart failure (CHF), atrial fibrillation, hypertension, peripheral vascular disease, cerebrovascular disease, chronic pulmonary disease, diabetes, renal disease, malignant cancer, and severe liver disease. Laboratory test results from the first record after ICU admission were selected, which included white blood cell count, hemoglobin, platelets, creati- nine, blood urea nitrogen, glucose, lactate, pH, bicarbonate, total

Image of Fig. 1

Fig. 1. Flow diagram for patients included in the study.

calcium, potassium, magnesium, sodium, and international normalized ratio. Medications included hydrocortisone, phenylephrine, epineph- rine, dobutamine, and dopamine from the first day of ICU stay. Initial NE and AVP doses were selected from the first infusion rate after ICU admission. Organ-support therapies included RRT and Invasive mechanical ventilation from the first day of ICU stay.

    1. Outcomes

The primary outcome was 28-day mortality. The secondary out- comes were ICU, and hospital mortality, the numbers of vasopressor-free and ventilation-free days at 28 days, ICU length of stay (LOS), hospital LOS, SOFA score on days 2 and 3, and RRT use on days 2 and 3. The 28-day mortality was defined as death between ICU admission and 28 days thereafter. Hospital and ICU mortality were defined as death during ICU and hospital stays, respectively. The numbers of vasopressor-free and ventilation-free days at 28 days were defined as the numbers of days with no vasopressor (i.e., NE, epinephrine, phenylephrine, AVP, dopamine, dobutamine, or milrinone) and no IMV support between ICU admission and 28 days thereafter, respectively.

    1. Statistical analysis

Missing data were listed in Table 1, and the missing proportion of data for univariate variables did not exceed 2.1%. Variables with missing data were estimated using the multiple-imputation method with Random forests in the mice package of R software [13]. Continuous variables were expressed as mean +- standard deviation or median and interquartile-range values, and differences between the groups were analyzed using nonparametric tests. Categorical variables were expressed as numbers and percentages, and differences between the groups were analyzed using Chi-square or Fisher’s exact tests.

A Cox proportional-hazards model was established to explore the association between early AVP exposure and 28-day mortality. We first constructed a univariate model to analyze the independent associ- ation between early AVP exposure and 28-day mortality. To reduce the influences of confounding factors, we further constructed a multivariate Cox regression model to explore the associations between early AVP

Table 1

Baseline characteristics of the original cohort between non-early-AVP and early-AVP groups.





Missing data(%)

n = 200

n = 331

Age, (yrs)

67.00 (57.00, 76.00)

65.00 (56.00, 75.00)



Sex, n (%)




106 (53.0)

186 (56.2)



94 (47.0)

145 (43.8)



28.48 (24.23, 34.25)

28.06 (24.34, 34.05)



Ethnicity, n (%)




123 (61.5)

205 (61.9)



20 (10.0)

26 (7.9)



57 (28.5)

100 (30.2)


First careunit, n (%)




88 (44.0)

128 (38.7)



23 (11.5)

37 (11.2)



89 (44.5)

166 (50.2)


Admission type, n (%)




117 (58.5)

184 (55.6)



60 (30.0)

92 (27.8)



23 (11.5)

55 (16.6)


Insurance, n (%)




96 (48.0)

153 (46.2)



11 (5.5)

21 (6.3)



Vital signs

93 (46.5)

157 (47.4)


Heart rate, (bpm)

96.89 (84.03, 110.03)

99.39 (84.30, 111.81)



MeanBP, (mmHg)

70.94 (66.74, 75.82)

72.76 (68.95, 77.37)



Respiratory rate, (bpm)

22.48 (19.27, 25.40)

23.40 (19.50, 26.65)



SpO2 (%)

96.49 (95.02, 98.04)

96.74 (94.30, 98.36)



Temperature, (?C)

36.94 (36.61, 37.33)

36.95 (36.60, 37.49)



SOFA score

12.00 (9.00, 15.00)

13.00 (10.00, 15.50)




90.00 (68.75, 113.25)

89.00 (69.00, 112.00)




3.00 (3.00, 4.00)

4.00 (3.00, 4.00)



Charlson Comorbidity Index Comorbidity, n (%)

Myocardial infarction

6.00 (4.00, 8.00)

35 (17.5)

6.00 (4.00, 8.00)

61 (18.4)






63 (31.5)

89 (26.9)



Atrial fibrillation

92 (46.0)

106 (32.0)




121 (60.5)

181 (54.7)



Peripheral vascular disease

20 (10.0)

28 (8.5)



Cerebrovascular disease

23 (11.5)

29 (8.8)



Chronic pulmonary disease

52 (26.0)

87 (26.3)




63 (31.5)

96 (29.0)



Renal disease

44 (22.0)

67 (20.2)



Malignant cancer

36 (18.0)

61 (18.4)



Severe liver disease Laboratory tests

WBC, (k/uL)

30 (15.0)

15.05 (9.38, 20.60)

36 (10.9)

13.70 (7.20, 22.85)





Hemoglobin, (g/dL)

10.25 (9.00, 12.03)

10.50 (8.90, 12.10)



Platelet, (k/uL)

166.00 (112.00, 241.25)

171.00 (99.00, 254.50)



Creatinine, (mg/dL)

1.80 (1.00, 3.10)

1.70 (1.10, 2.65)



Urea nitrogen, (mg/dL)

34.50 (23.00, 56.50)

32.00 (21.00, 51.00)



Glucose, (mg/dl)

133.50 (102.00, 177.25)

139.00 (105.00, 197.50)



Lactate, (mmol/L)

2.55 (1.60, 3.92)

3.20 (2.00, 5.50)




7.29 (7.21, 7.36)

7.24 (7.15, 7.33)



Bicarbonate, (mmol/L)

18.50 (16.00, 22.00)

17.00 (15.00, 21.00)



Total calcium, (mmol/L)

7.70 (7.20, 8.40)

7.40 (6.90, 8.10)



Potassium, (mmol/L)

4.20 (3.80, 4.90)

4.20 (3.60, 4.80)



Magnesium, (mmol/L)

1.90 (1.60, 2.20)

1.80 (1.50, 2.10)



Sodium, (mmol/L)

137.00 (134.00, 141.00)

138.00 (135.00, 141.00)




1.50 (1.20, 1.92)

1.60 (1.30, 1.95)



RRT, n (%)

33 (16.5)

69 (20.8)



IMV, n (%)

129 (64.5)

245 (74.0)



Hydrocortisone, n (%)

30 (15.0)

96 (29.0)



Phenylephrine, n (%)

63 (31.5)

140 (42.3)



Dopamine, n (%)

9 (4.5)

25 (7.6)



Epinephrine, n (%)

12 (6.0)

74 (22.4)



Dobutamine, n (%)

9 (4.5)

19 (5.7)



Initial norepinephrine dose, (mcg/kg/min)

0.20 (0.12, 0.30)

0.28 (0.15, 0.40)



Initial vasopressin dose, (units/h)

2.40 (2.40, 2.40)

2.40 (2.40, 2.40)



48-huor mortality, n (%)

17 (8.5)

48 (14.5)



72-huor mortality, n (%)

26 (13.0)

54 (16.3)



AVP: vasopressin; BMI: body mass index; MICU: medical intensive care unit; SICU: surgical intensive care unit; SIRS: systemic inflammatory response syndrome; SOFA score: sequential organ failure assessment score; APSIII: acute physiology score III; CHF: congestive heart failure; WBC: white blood cell; INR: international normalized ratio; RRT: renal replacement ther- apy; IMV: invasive mechanical ventilation; Continuous covariate were expressed as median and interquartile-range values.

exposure and 28-day mortality. The crude covariates included patient demographic information, vital signs, comorbidities, illness severity scores, laboratory tests, medications, and organ supportive therapy. The final covariates that were included in the multivariate Cox regres- sion model were screened using stepwise regression analysis.

For sensitivity analysis, univariate and multivariate Cox regression model were constructed to analyze the independent association be- tween early AVP exposure and ICU and hospital mortality. Propensity- score matching (PSM) was performed to balance the baseline differ- ences and selective bias between the groups. The propensity scores were estimated using a multivariate logistic regression analysis with a ratio of 1:1 nearest-neighbor matching and a 0.2 caliper width. Vari- ables included in the PSM analysis are shown in Fig. 2. Standardized mean differences (SMDs) were calculated to assess the differences be- tween the original and matched cohorts. When the SMD of a variable is <0.1, it can be considered that balance has been reached between the groups [14].

We also performed a sensitivity analysis to explore the effect of early AVP in different subgroups, which were based on the initial AVP dose, age, sex, BMI, myocardial infarction, CHF, atrial fibrillation, hyperten- sion, diabetes, renal disease, and severe liver disease.

Statistical analysis was performed using R software (version 4.1.0).

All P values were two-sided.

  1. Results
    1. Baseline characteristics

A total of 531 patients with septic shock were enrolled in the study, including 331 patients (62.5%) in the early AVP group and 200 patients (37.5%) in the non-early AVP group. The median age was 66.00 (56.00, 75.00) years old. Forty-five percent were female and 62 % were Cauca- sian. Patients in the early-AVP group were generally more likely to re- ceive IMV, hydrocortisone, phenylephrine, and epinephrine, have a higher initial NE dose, higher lactate level, higher SIRS score; and have lower pH, bicarbonate, total calcium, and magnesium levels. Mean- while, patients in the non-early-AVP group were more likely to have a lower MBP. The basic characteristics of the groups were listed in Table 1.

    1. Primary outcome

The relationships between early AVP and the potentially predispos- ing factors were examined to determine the relative importance of co- variates using learning vector quantization [15]. A propensity-score model was first constructed that employed 50 covariates. The contribu- tions of individual covariates to the propensity score are presented in Fig. 2. The top-contributing covariates were APSIII, Charlson

Image of Fig. 2

Fig. 2. The contributions of individual covariates to the final propensity score.

Table 2

The association between early and non-early AVP exposure on mortality.



95% CI

Primary outcome 28-day mortality




Model2 Secondary outcome

ICU mortality







Hospital mortality









HR: harzard ratio; CI: confidence interval; ICU: intensive care unit; Model1: univatiate model; Model2: multivariate model.

Comorbidity Index, temperature, MBP, SOFA score, and age. These co- variates were common factors that influenced the decisions of physi- cians about whether to perform early AVP. Supplementary Table S2 lists the variance inflation factor scores of each variable, a score lower than 5 can be considered to indicate no multicollinearity.

In the original cohort, there were no differences in 28-day mor- tality (44.1% vs 49.0%, P = 0.314) between the early- and non- early-AVP groups. In the matched cohort, 158 patients exposed to early AVP were matched with the same number of patients with nonearly AVP. As shown in Supplementary Table S3 and S4, all covar- iates were balanced between the groups of the matched cohorts (SMD < 0.1), except for cerebrovascular disease, chronic pulmory disease, epinephrine use, and glucose level. In the matched cohort, we also found no difference in 28-day mortality (48.1% vs 48.7%, P = 1.000) between the two groups.

Cox regression analysis of the univariate model demonstrated that there were no difference in 28-day mortality between the early- and non-early-AVP groups (HR = 0.92, 95% CI = 0.71-1.19). After adjusting for covariates, the multivariate model also showed no difference in 28- day mortality (HR = 0.91, 95% CI = 0.68-1.24) between the groups. The detailed results were listed in Table 2.

    1. Secondary outcomes

We evaluated several secondary outcomes to compare other differ- ences between early and nonearly AVP administration. Several key secondary outcomes were observed. There were no differences in ICU mortality (36.1% vs 42.4%, P = 0.300), and hospital mortality (44.9% vs 48.7%, P = 0.573) between the early- and non-early-AVP groups. The multivariate Cox regression model also showed no differences in ICU mortality (HR = 0.95, 95% CI = 0.67-1.35), and hospital mortality (HR = 0.95, 95% CI = 0.69-1.31) between the two groups. In the

matched cohort, there were no differences in the numbers of vasopressor-free and ventilation-free days at 28 days, ICU LOS, hospital LOS, SOFA score on days 2 and 3, and RRT use on days 2 and 3. The detailed outcomes were listed in Table 3.

    1. Subgroup analyses of 28-day mortality

The results of the subgroup analyses were shown in Fig. 3. There was no significant difference between the early- and non-early-AVP groups in each stratified population, except for that stratified by atrial fibrilla- tion. The forest plot indicated that patients with atrial fibrillation had a higher 28-day mortality risk in the early-AVP group than in the non- early-AVP group.

  1. Discussion

This single-center retrospective study that compared early AVP (within 6 h) with nonearly AVP (after 6 h) after the onset of septic shock found no difference in short-term mortality rates, including in the hospital, ICU, and after 28 days. Literature on the effect of early AVP use on the mortality of patients with septic shock is lacking. Our results on short-term mortality were consistent with the findings of previous retrospective research studies [16], prospective studies [8,17], and meta-analyses [9,18].

Our secondary outcome analysis that included the numbers of vasopressor-free and ventilation-free days at 28 days, ICU LOS, hos- pital LOS, SOFA score on days 2 and 3, and RRT use on days 2 and 3 also found no differences between the early- and non-early-AVP groups. To explore whether using AVP earlier improves patient prog- nosis, we further analyzed AVP infusion within 2 h of septic shock onset compared with 2 h after onset, and similarly found no differ- ences in mortality and secondary outcomes. The results are listed in Supplementary Tables S5 and S6.

Several studies that evaluated how early AVP use affected multiple-Organ functions in septic shock produced inconsistent results. In the VANISH trial, early single AVP use compared with NE did not improve the number of kidney-failure-free days [8]. Ham- mond et al. conducted a retrospective study that compared AVP with NE monotherapy and found that the target MAP was reached sooner and organ dysfunction at 72 h was relieved more effectively than in delayed or no initiation of treatment [16]. A Prospective open-label trial that used AVP plus NE versus NE alone found that it can achieve a MAP of 65 mmHg more rapidly, indicating that vaso- pressor use could be withdrawn earlier [17]. However, in our study the early use of AVP did not decrease the number of vasopressor- free days. AVP plus NE may be more effective in maintaining MAP, but it was not clear whether the benefit can translate into an improve- ment in systemic tissue perfusion.

Table 3

Outcomes of original cohort and matched cohort.


Original cohort

Matched cohort

Non-early AVP

Early AVP


Non-early AVP

Early AVP


n = 200

n = 331

n = 158

n = 158

28-day mortality (n, %)

98 (49.0)

146 (44.1)


77 (48.7)

76 (48.1)


ICU mortality (n, %)

84 (42.0)

116 (35.0)


67 (42.4)

57 (36.1)


Hospital mortality (n, %)

98 (49.0)

141 (42.6)


77 (48.7)

71 (44.9)


Vasopressor free days at 28-day (days)

22.47 (18.40, 25.02)

22.86 (19.27, 25.63)


23.09 (18.68, 25.07)

22.96 (19.27, 25.80)


Ventilation free days at 28-day (days)

24.67 (20.62, 26.66)

24.72 (20.50, 26.44)


24.62 (20.41, 26.62)

24.83 (22.42, 26.71)


ICU LOS (days)

9.03 (4.38, 16.89)

8.00 (3.67, 13.44)


7.92 (4.23, 16.08)

6.96 (3.53, 13.20)


Hospital LOS (days)

15.00 (7.75, 27.00)

15.00 (7.00, 23.00)


15.00 (7.00, 26.75)

15.00 (6.25, 23.75)


SOFA score at day 2

11.50 (9.00, 14.00)

11.00 (9.00, 15.00)


11.00 (9.00, 14.00)

11.00 (9.00, 14.00)


SOFA score at day 3

10.00 (7.00, 14.00)

10.00 (6.00, 13.00)


10.00 (7.00, 14.00)

10.00 (6.00, 13.00)


RRT at day 2 (n, %)

40 (20.0)

83 (25.1)


29 (18.2)

43 (27.0)


RRT at day 3 (n, %)

46 (23.0)

76 (23.0)


35 (22.0)

40 (25.2)


AVP,vasopressin; ICU, intensive care unit; LOS, length of stay; SOFA, Sequential Organ Failure Assessment; RRT, renal replacement therapy.

Image of Fig. 3

Fig. 3. Subgroup analysis of relationship between groups and 28-day mortality.

AVP is a vasoconstriction agent that causes V1-receptor activa- tion in vascular smooth muscle, and has a different mechanism of action than other catecholamines. At the onset of septic shock, large amounts of AVP are released to maintain organ perfusion. The progression of septic shock results in plasma AVP levels decreasing, and so AVP supplementation may be an effective choice for main- taining a stable circulation. However, AVP at normal or high concen- trations has little effect on blood pressure and may induce adverse effects [19]. High-dose AVP (>0.04 U/min) is also associated with potentially deleterious vasoconstriction of the renal, mesenteric, pulmonary, and coronary vasculature. Guerci et al. investigated sep- tic shock and found earlier AVP in the “refractory” profile could counteract harmful high NE doses, whereas AVP was not necessary in the “controlled” profile [20]. The adequate condition and timing of AVP initiation also remains challenging and should be character- ized individually. For the beneficial effect of AVP in septic shock, the selection of AVP may be based on the NE response, plasma AVP levels, the optimal dose, and early application.

There were several limitations to our study. First, it had a retrospec- tive design, and there may be some underlying factors that affected the outcomes. We applied several methods including setting strict inclusion and exclusion criteria, a PSM cohort, and establishing univariate and multivariate models, but other unknown confounders may remain. Sec- ond, the time of AVP initiation in septic shock was based on the onset of NE initiation, and the precise time of septic shock onset was not clear, and this may have caused delays in AVP initiation. Our results must therefore be interpreted with caution.

  1. Conclusion

There were no significant differences between early AVP combined with NE and nonearly AVP combined with NE in short-term mortality, the numbers of vasopressor-free and ventilation-free days at 28 days, ICU LOS, hospital LOS, SOFA score on days 2 and 3, or RRT use on days 2 and 3. The effect of early AVP use needs to be further assessed in large randomized controlled trials.

CRediT authorship contribution statement

Dan He: Writing – original draft. Luming Zhang: Writing – review & editing. Hai Hu: Writing – review & editing. Wan-jie Gu: Conceptualiza- tion. Xuehao Lu: Data curation. Minshang Qiu: Investigation. Chao Li: Validation. Haiyan Yin: Supervision. Jun Lyu: Validation, Supervision, Software, Methodology.


This study was supported by the National Natural Science Foun- dation of China (No. 82072232 and No. 81871585), the Clinical Fron- tier Technology Program of the First Affiliated Hospital of Jinan University, China (No. JNU1AF-CFTP-2022-a01235), the Science and Technology Projects in Guangzhou, China (No. 202201020054 and No. 2023A03J1032), the Hunan Health Commission, China (No. 202104112215).

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influ- ence the work reported in this paper.

Appendix A. Supplementary data

Supplementary data to this article can be found online at https://doi. org/10.1016/j.ajem.2023.04.040.


  1. Singer M, Deutschman CS, Seymour CW, et al. The third international consensus def- initions for sepsis and septic shock (Sepsis-3). JAMA. 2016;315(8):801. https://doi. org/10.1001/jama.2016.0287.
  2. Bauer M, Gerlach H, Vogelmann T, Preissing F, Stiefel J, Adam D. Mortality in sepsis and septic shock in Europe, North America and Australia between 2009 and 2019

— results from a systematic review and meta-analysis. Crit Care. 2020;24(1):239.

  1. Evans L, Rhodes A, Alhazzani W, et al. Surviving sepsis campaign: international guidelines for management of sepsis and septic shock 2021. Intensive Care Med. 2021;47(11):1181-247.
  2. Abid O, Akca S, Haji-Michael P, Vincent JL. Strong vasopressor support may be futile in the intensive care unit patient with multiple organ failure. Crit Care Med. 2000;28


  1. Ukor IF, Walley KR. Vasopressin in Vasodilatory shock. Crit Care Clin. 2019;35(2): 247-61.
  2. Holmes CL, Patel BM, Russell JA, Walley KR. Physiology of vasopressin relevant to management of septic shock. Chest. 2001;120(3):989-1002. 1378/chest.120.3.989.
  3. Buijk SE, Bruining HA. Vasopressin deficiency contributes to the vasodilation of sep- tic shock. Circulation. 1998;98(2):187.
  4. Gordon AC, Mason AJ, Thirunavukkarasu N, et al. Effect of early vasopressin vs nor- epinephrine on kidney failure in patients with septic shock: the VANISH randomized clinical trial. JAMA. 2016;316(5):509-18.
  5. Huang H, Wu C, Shen Q, Xu H, Fang Y, Mao W. The effect of early vasopressin use on patients with septic shock: a systematic review and meta-analysis. Am J Emerg Med. 2021;48:203-8.
  6. von Elm E, Altman DG, Egger M, Pocock SJ, Gotzsche PC, Vandenbroucke JP. The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement: guidelines for reporting observational studies. The Lancet. 2007;370 (9596):1453-7.
  7. Wu WT, Li YJ, Feng AZ, et al. Data mining in clinical big data: the frequently used da- tabases, steps, and methodological models. Mil Med Res. 2021;8(1):44. https://doi. org/10.1186/s40779-021-00338-z.
  8. Yang J, Li Y, Liu Q, et al. Brief introduction of medical database and data mining tech- nology in big data era. J Evid Based Med. 2020;13(1):57-69. jebm.12373.
  9. Beesley LJ, Bondarenko I, Elliot MR, Kurian AW, Katz SJ, Taylor JM. Multiple imputa- tion with missing data indicators. Stat Methods Med Res. 2021;30(12):2685-700.
  10. Benedetto U, Head SJ, Angelini GD, Blackstone EH. Statistical primer: propensity score matching and its alternatives+. Eur J Cardiothorac Surg. 2018;53(6):1112-7.
  11. Pourghasemi HR, Gayen A, Lasaponara R, Tiefenbacher JP. Application of learning vector quantization and different machine learning techniques to assessing forest fire influence factors and spatial modelling. Environ Res. 2020;184:109321.
  12. Hammond DA, Cullen J, Painter JT, et al. Efficacy and safety of the early addition of vasopressin to norepinephrine in septic shock. J Intensive Care Med. 2019;34(11- 12):910-6.
  13. Hammond DA, Ficek OA, Painter JT, et al. Prospective open-label trial of early con- comitant vasopressin and norepinephrine therapy versus initial norepinephrine monotherapy in septic shock. Pharmacotherapy. 2018;38(5):531-8. https://doi. org/10.1002/phar.2105.
  14. Sedhai YR, Shrestha DB, Budhathoki P, et al. Vasopressin versus norepinephrine as the first-line vasopressor in septic shock: a systematic review and meta-analysis. J Clin Transl Res. 2022;8(3):185-99.
  15. Abboud FM, Floras JS, Aylward PE, Guo GB, Gupta BN, Schmid PG. Role of vasopressin in cardiovascular and blood pressure regulation. Blood Vessels. 1990;27(2-5): 106-15.
  16. Guerci P, Belveyre T, Mongardon N, Novy E. When to start vasopressin in septic shock: the strategy we propose. Crit Care. 2022;26(1):125. 1186/s13054-022-04001-4.