Comparison of metaraminol versus no metaraminol on time to resolution of shock in critically ill patients

Abstract

Objective

There is limited evidence to support metaraminol use in critically ill patients. Metaraminol is not included as a vasopressor choice in international guidelines for the management of shock. Nevertheless, metaraminol is used in rates up to 42% in this patient population. The objective of this study was to investigate the effectiveness of metaraminol for the treatment of critically ill patients with shock.

Methods

A single-centre retrospective matched observational study was conducted in a 54-bed intensive care unit of a tertiary hospital. Patients aged 16 years or older who were admitted from 2017 to 2019 with shock were included. Patients treated with metaraminol and norepinephrine (MET-NOR) were compared with those treated with norepinephrine without metaraminol (NOR). The primary outcome was the time to resolution of shock defined as the time to cessation of vasopressors. The secondary outcome was vasopressor-free days until 28 days.

Results

There were 286 patients included in this study, including 143 patients in each group. The median time to resolution of shock was 44 hours (IQR 28–66 hours) in the MET-NOR group compared with 27 hours (IQR 14–63 hours) in the NOR group (95% CI of median difference 7 to 19 hours; p<0.01). The Cox regression analysis for the time to resolution of shock showed no significant difference between groups (HR 1.24, 95% CI 0.96 to 1.60; p=0.10). However, the proportional hazards assumption was not met (p<0.01). The median number of vasopressor-free days until 28 days was 26 days (IQR 24–27 days) in the MET-NOR group compared with 27 days (IQR 25–27 days) in the NOR group (95% CI of median difference −0.8 to −0.1 day; p<0.01).

Conclusion

In critically ill patients, metaraminol may be associated with a longer time to resolution of shock compared with those who do not receive metaraminol.

Introduction

Norepinephrine is the recommended first-line vasopressor agent for the management of critically ill patients with shock.^1–3 This is because of improved outcomes with this vasopressor compared with other agents.⁴ International guidelines suggest the addition of vasopressin or epinephrine as secondary vasopressors to increase mean arterial pressure (MAP) to target values or to decrease the consumption of norepinephrine.¹ In some parts of the world, secondary vasopressor selection includes the use of metaraminol, which is a vasopressor that has been available for more than 50 years.⁵ Metaraminol is a sympathomimetic amine that causes vasoconstriction via a direct effect on α₁ adrenergic receptors and indirectly via release of norepinephrine by sympathetic nerve endings.⁶ It also has some positive inotropic activity via effect on β₁ adrenergic receptors.⁶ However, despite the increased use of metaraminol, there remains limited and low-quality evidence to support its use in critically ill patients with shock.^{7 8}

In recent years, pharmacoepidemiology data show that metaraminol use is common and is often used as an initial peripheral vasopressor.⁹ In a prospective observational study of current practice in 70 hospitals in Australia and New Zealand, metaraminol was used in 42% of patients with septic shock.¹⁰ In a pragmatic clinical trial involving 65 intensive care units (ICUs) in the UK, 32% of patients with vasodilatory hypotension were receiving metaraminol at baseline when vasopressor therapy was administered at the discretion of the treating clinician.¹¹ Metaraminol is commonly used in some countries as a peripheral vasopressor in the critically ill due to longstanding clinical norms regarding this route of administration.^{9 12} Patients are then typically transitioned to norepinephrine after central venous access is established.^{9 12} Thus, metaraminol appears to satisfy a clinical niche like phenylephrine use in North America.¹³ Metaraminol is considered to be a safer alternative, especially when needed for short-term use prior to the availability of central intravenous access. However, there is no evidence that it is a safer peripheral vasopressor agent.^14–16 Given its continued use, studies investigating the effect of the addition of metaraminol to vasopressor regimens in the critically ill are needed. Thus, the aim of this study was to investigate the effectiveness of metaraminol when used as a component of vasopressor regimens in critically ill patients with shock.

Methods

Study design and setting

A single-centre retrospective observational study was conducted in patients admitted to a 54-bed ICU of a tertiary hospital in Sydney, Australia. The study protocol was approved by the local hospital ethics committee (2019/ETH13182). The manuscript was prepared in accordance with the STROBE (Strengthening the Reporting of Observational Studies in Epidemiology) guidelines for reporting observational cohort studies.¹⁷

Patient selection

Patients aged 16 years or older who were admitted to the ICU from 1 January 2017 to 30 October 2019 and who received metaraminol or norepinephrine continuous infusions for the management of shock were included in the analysis. The patient list was obtained via a database query from the electronic medical record (Philips IntelliSpace Critical Care and Anaesthesia).¹⁸

Shock was defined as per the treating team and according to international guidelines, as the life-threatening condition of circulatory failure that requires the use of vasopressors to maintain a MAP above 65 mm Hg.^{1 2 19} There were no local treatment guidelines for vasopressor selection in the ICU at the study site. Vasopressor selection in the ICU was based on clinician preference. Vasopressors were titrated to a default MAP target of at least 65 mm Hg, unless otherwise specified by the clinician.^{1 2}

Patients were excluded from this study if vasopressors were used for indications other than shock, such as the prevention of neurological deterioration in stroke patients or blood pressure support in brain-dead organ donors. Patients were excluded if metaraminol was only used as bolus doses rather than continuous infusions. However, there was no requirement for minimum duration of infusion as a selection criterion. Patients were excluded if they were transferred from another hospital.

Patients were stratified into two groups based on vasopressor use during their ICU stay: (1) those who received metaraminol and norepinephrine (MET-NOR), and (2) those who received norepinephrine without metaraminol (NOR). A random sample was selected from the MET-NOR group using a random number sequence generated in STATA software (V.16) based on the sample size estimate. This sample was then matched to the NOR group. Patient groups were randomly matched 1:1 via the coarsened exact matching method^{20 21} using the following variables: age, sex, Acute Physiology and Chronic Health Evaluation (APACHE) III score, lowest value of MAP in the first 24 hours of ICU admission, type of ICU admission (elective vs non-elective admission), sepsis diagnosis and requirement for mechanical ventilation. These variables were considered by the investigators to be potential confounders that could affect the primary outcome.

Data collection

Baseline patient characteristics were collected from the electronic medical record manually or utilising the Australian New Zealand Intensive Care Society Registry data extracts embedded within it.²² Variables included age, sex, weight, APACHE III score, source of admission to ICU, type of ICU admission (surgical or non-surgical), type of shock (sepsis or other) and incidence of acute renal failure in the first 24 hours of ICU admission. Acute renal failure was defined as having a 24-hour urine output <410 mL, a serum creatinine >133 µmol/L and not receiving chronic dialysis.²² Data on concurrent vasopressors and inotropes, vasopressor infusion rates and route of vasopressor administration were collected.

Baseline physiological parameters were collected and included the highest value of temperature, white cell count, serum creatinine concentration, heart rate and respiratory rate in the first 24 hours of ICU admission, and lowest value of MAP and systolic blood pressure in the first 24 hours of ICU admission. Data on adverse events were collected from the electronic medical record and included prespecified categories of arrhythmia (atrial fibrillation, bradycardia or tachycardia) or vasopressor extravasation injury.

Outcomes

The primary outcome was the time to resolution of shock, which is defined as the time from initiation to cessation of vasopressors.^{23 24} The primary outcome was reported for the first episode of shock in patients with more than one episode. The initial episode was considered to be concluded when 24 consecutive hours had lapsed without administration of vasopressors, patient was discharged from the ICU or died in the ICU.^{11 23}

The secondary outcome was vasopressor-free days until 28 days. Exploratory outcomes included ICU mortality, hospital mortality, ICU and hospital length of stay, incidence of mechanical ventilation, incidence of renal replacement therapy, duration of mechanical ventilation, duration of renal replacement therapy, ventilator-free days until 28 days, renal replacement therapy-free days until 28 days, recurrence of shock and adverse events (arrhythmia and vasopressor extravasation injury).

Organ support-free days until 28 days (vasopressor, ventilator or renal replacement therapy) was defined as the number of days alive and free from intervention until 28 days from ICU admission.²⁵ The number of organ support-free days until 28 days was assigned as zero for patients who died in the ICU.²⁵ Recurrence of shock was defined as the recommencement of vasopressors 24 hours after the initial episode of shock.

Statistical analysis

Categorical variables were reported as proportions and analysed using the Fisher’s exact test. Normally distributed continuous variables were reported as means and SD and analysed using the Student’s t-test. Non-normally distributed continuous variables were reported as medians and IQRs and analysed using the Mann-Whitney U test. The 95% CIs for the median difference were calculated for primary and secondary outcomes. Time to resolution of shock was depicted using Kaplan-Meier curves and compared using the Wilcoxon-Breslow-Gehan test. The Wilcoxon-Breslow-Gehan test was used because it provides a greater weight to the earlier part of the Kaplan-Meier curve, which is pertinent for early vasopressor cessation.²⁶ This is because it is expected that patients in both groups will eventually cease vasopressors during hospitalisation. Patients who died in the ICU were right censored in the primary analysis. A Cox proportional hazards model was conducted to report the primary outcome as a HR with 95% CI. The proportional hazards assumption was evaluated via the time varying covariate method.²⁷ Vasopressor infusions were also depicted graphically for each patient in the MET-NOR group to provide a visual representation of when each vasopressor was used in the cohort during ICU stay.

A sensitivity analysis of the primary outcome was performed by considering death in ICU as a competing risk.^{24 28} A sub-hazard ratio (SHR) of the primary outcome was reported using a competing risk regression.²⁹ A subgroup analysis was also conducted in patients with septic shock and in patients who did not receive concurrent vasopressors or inotropes, other than metaraminol and norepinephrine.

To detect a significant difference in the time to resolution of shock of 1 day between treatment groups, assuming an SD of 3 days in each group, power of 80% and two-sided alpha level of 0.05, it was estimated that 143 patients would be required in each group (286 total sample size).^{24 30}

Simple imputation was used for missing data. Missing data ranged from 1% to 2% for some variables including highest white cell count, highest serum creatinine and total metaraminol duration. All statistical analysis was two-sided with statistical significance level of 0.05 and was conducted using STATA software (V.16) and R software (V.4.0.3).

Results

Patient characteristics

A flow diagram of patient selection is shown in figure 1. A total of 401 patients received MET-NOR during the study period. Of these, 169 patients who received MET-NOR needed to be screened for eligibility to include a target of 143 patients. These patients were matched 1:1 from within a sample of 2750 NOR patients during the same timeframe as described in the Methods section. A total of 189 matched NOR patients needed to be screened for eligibility until a final sample of 143 patients was reached in the NOR group. This resulted in a final sample size of 143 patients in each patient group (286 total) (figure 1).

Figure 1.

Flow diagram of patient selection. Flow diagram of patient selection in the metaraminol and norepinephrine (MET-NOR) group and the norepinephrine without metaraminol (NOR) group.

At baseline, there were no differences in patient characteristics between groups (table 1). The median duration of metaraminol infusion in the ICU was 4 hours (IQR 2–9 hours) in the MET-NOR group. The median metaraminol initial infusion rate in the ICU was 3.0 mg·hour^-1 (IQR 2.3–5.0 mg·hour^-1) and this was titrated up to a maximum of 5.0 mg·hour^-1 (IQR 3.5–8.0 mg·hour^-1) in the MET-NOR group. There was no difference in the initial norepinephrine infusion rate or the maximum norepinephrine infusion rate between groups (table 1). The vasopressor infusions used for each patient are temporally depicted in figure 2. It highlights that in the MET-NOR group, metaraminol was predominantly used as an initial vasopressor prior to transitioning to norepinephrine.

Table 1.

Characteristics of study patients

Variables	Metaraminol and norepinephrine (n=143)	Norepinephrine (n=143)	P value
Baseline patient characteristics
Age (years), median (IQR)	70 (58–79)	67 (56–76)	0.19
Sex (male), N (%)	80 (56)	78 (55)	0.91
Weight (kg), median (IQR)	74 (63–90)	74 (62–86)	0.87
APACHE III score, median (IQR)	69 (54–86)	68 (53–90)	0.94
Source of ICU admission, N (%)
Operating theatre	51 (36)	47 (33)	0.71
Emergency department	50 (35)	47 (33)	0.80
Ward	42 (29)	49 (34)	0.45
Surgical ICU admission, N (%)	48 (34)	48 (34)	1.00
Septic shock, N (%)	67 (47)	64 (45)	0.81
Acute renal failure, N (%)*	11 (8)	19 (13)	0.18
Physiological parameters, median (IQR)*
Highest temperature (°C)	37.2 (36.7–37.9)	37.3 (36.9– 37.8)	0.57
Highest RR (breaths/min)	26 (22–29)	26 (23–31)	0.21
Highest HR (beats/min)	101 (87–117)	106 (92–124)	0.09
Highest white cell count (109/L)	13.3 (9.4–18.9)	14.3 (10.5– 21.2)	0.21
Highest serum creatinine (µmol/L)	109 (73–184)	112 (79–181)	0.55
Lowest MAP (mm Hg)	59 (54–62)	60 (55–63)	0.28
Lowest SBP (mm Hg)	83 (79–92)	86 (80–91)	0.38
Vasopressor and inotrope utilisation in the ICU
Initial norepinephrine rate (µg·kg-1·min-1), median (IQR)	0.05 (0.03–0.08)	0.06 (0.03–0.10)	0.32
Maximum norepinephrine rate (µg·kg-1·min-1), median (IQR)	0.15 (0.08–0.27)	0.14 (0.07–0.32)	0.94
Concurrent vasopressors, N (%)
Epinephrine	5 (4)	14 (10)	0.06
Vasopressin	24 (17)	33 (23)	0.24
Concurrent inotropes, N (%)
Dobutamine	5 (4)	15 (11)	0.04
Milrinone	5 (4)	19 (13)	<0.01
Isoprenaline	1 (1)	1 (1)	1.00
Levosimendan	2 (1)	7 (5)	0.17
Location where vasopressor first started, N (%)
Operating theatre	25 (17)	19 (13)	0.41
Emergency department	25 (17)	17 (12)	0.24
ICU	93 (65)	107 (75)	0.09
Vasopressor use via peripheral line at any time	128 (90)	9 (6)	<0.01

*In the first 24 hours of ICU admission.

APACHE, Acute Physiology and Chronic Health Evaluation; HR, heart rate; ICU, intensive care unit; MAP, mean arterial pressure; RR, respiratory rate; SBP, systolic blood pressure.

Figure 2.

Vasopressor infusions administered in the intensive care unit. The sequence of vasopressor infusions administered for each patient during the first 100 hours in the intensive care unit in the metaraminol and norepinephrine (MET-NOR) group and the norepinephrine without metaraminol (NOR) group.

There was no difference in the use of other concurrent vasopressors between groups (table 1). However, patients in the MET-NOR group were less likely to receive inotropes such as dobutamine (4% vs 11%, p=0.04) and milrinone (4% vs 13%, p<0.01) (table 1).

Primary outcome

The median time to resolution of shock was 44 hours (IQR 28–66 hours) in the MET-NOR group compared with 27 hours (IQR 14–63 hours) in the NOR group (95% CI of median difference 7 to 19 hours; p<0.01) (table 2, figure 3). The Cox regression analysis for the time to resolution of shock showed no significant difference between groups (HR 1.24, 95% CI 0.96 to 1.60; p=0.10) (table 3). However, the proportional hazards assumption was not met (p<0.01).

Table 2.

Study outcomes

Study outcome	Metaraminol and norepinephrine (n=143)	Norepinephrine (n=143)	P value
Time to resolution of shock (hours), median (IQR)*	44 (28–66)	27 (14–63)	<0.01†
ICU mortality, N (%)	19 (13)	23 (16)	0.62
Hospital mortality, N (%)	29 (20)	28 (20)	1.00
ICU length of stay (days), median (IQR)	4 (3–8)	4 (3–7)	0.84
Hospital length of stay (days), median (IQR)	14 (9–23)	16 (9–27)	0.23
Mechanical ventilation, N (%)	74 (52)	83 (58)	0.34
Renal replacement therapy, N (%)	15 (11)	29 (20)	0.03
Duration of mechanical ventilation (hours), median (IQR)	41 (22–83)	25 (13–106)	0.16
Duration of renal replacement therapy (hours), median (IQR)	126 (63–223)	81 (37–198)	0.44
Vasopressor-free days until 28 days, median (IQR)	26 (24–27)	27 (25–27)	<0.01
Ventilator-free days until 28 days, median (IQR)	27 (26–28)	28 (24–28)	0.53
Renal replacement therapy-free days until 28 days, median (IQR)	28 (28–28)	28 (26–28)	0.17
Recurrence of shock, N (%)	13 (9)	14 (10)	1.00
Adverse events, N (%)‡
Atrial fibrillation	43 (30)	38 (27)	0.60
Tachycardia	1 (1)	1 (1)	1.00

*Reported as median (IQR) hours from Kaplan-Meier analysis.

†Wilcoxon-Breslow-Gehan test.

‡Extravasation and tissue injury were not reported for any of the study patients.

ICU, intensive care unit.

Figure 3.

Kaplan-Meier curves for the time to resolution of shock. Kaplan-Meier curves for the time to resolution of shock in the metaraminol and norepinephrine (MET-NOR) group and the norepinephrine without metaraminol (NOR) group.

Table 3.

Sensitivity analyses of the primary outcome: time to resolution of shock

Metaraminol and norepinephrine*	Norepinephrine*	P value†	HR (95% CI)‡	SHR (95% CI)§
Full sample (n=286)
44 (28 to 66)	27 (14 to 63)	<0.01	1.24 (0.96 to 1.60)	1.16 (0.90 to 1.49)
Subset with sepsis (n=131)
50 (31 to 76)	41 (21 to 75)	0.15	1.09 (0.74 to 1.60)	1.04 (0.71 to 1.51)
Subset without other concurrent vasopressors or inotropes (n=201)
41 (26 to 60)	22 (13 to 41)	<0.01	1.55 (1.16 to 2.07)	1.54 (1.14 to 2.09)
Subset with sepsis and without other concurrent vasopressors or inotropes (n=86)
45 (29 to 61)	26 (14 to 58)	0.02	1.32 (0.84 to 2.07)	1.42 (0.90 to 2.23)

*Reported as median (IQR) hours from Kaplan-Meier analysis.

†Wilcoxon test.

‡HR from Cox regression.

§Sub-hazard ratio with death as a competing risk.

Secondary and exploratory outcomes

Patients in the MET-NOR group had a significantly lower cumulative vasopressor-free days until 28 days compared with those treated in the NOR group (median, 26 days (IQR 24–27 days) vs 27 days (IQR 25–27 days)) (95% CI of median difference −0.8 to −0.1 days; p<0.01) (table 2).

There were no significant between-group differences in most of the exploratory outcomes, with one exception (table 2). Patients in the MET-NOR group had a significantly lower incidence of renal replacement therapy compared with those treated in the NOR group (11% vs 20%; p=0.03).

Sensitivity and subgroup analyses

In the sensitivity analysis, where ICU death was considered to be a competing risk, the SHR for the time to resolution of shock showed no significant difference between groups (SHR 1.16, 95% CI 0.90 to 1.49; p=0.26) (table 3). However, the proportional hazards assumption was not met (p<0.01).

Results of the subgroup analysis are presented in table 3. In the subgroup of patients who did not receive other concurrent vasopressors (eg, vasopressin) or inotropes (n=201), patients in the MET-NOR group had a longer median time to resolution of shock 41 hours (IQR 26–60 hours) compared with those in the NOR group of 22 hours (IQR 13–41 hours) (p<0.01). There was no significant difference in the time of resolution of shock in the subgroup of patients with sepsis. However, in the subgroup of patients with sepsis who did not receive other concurrent vasopressors or inotropes (n=86), patients in the MET-NOR group had a longer time to resolution of shock compared with those in the NOR group (median, 45 hours (IQR 29–61 hours) vs 26 hours (IQR 14 –58 hours), p=0.02).

Discussion

The key finding of this study is that in critically ill patients, metaraminol use was associated with a longer time to resolution of shock compared with those treated with norepinephrine without metaraminol. These findings were consistent among the sensitivity and subgroup analyses. However, the use of metaraminol was in a sequential pattern rather than as a secondary or adjunctive agent. This is the first study that compares the role of the recommended first-line standard vasopressor norepinephrine with or without metaraminol on clinical outcomes in critically ill patients with shock.^{1 2 8} The results should be interpreted with caution. Although the groups were matched, there is the potential for unknown confounders that may have affected these results.

Studies investigating metaraminol for shock were conducted several decades ago, consisted of small heterogeneous patient populations or compared effects on haemodynamic parameters rather than clinically relevant outcomes.^{7 31–34} As such, metaraminol is not currently included as a vasopressor therapy in evidence-based guidelines for the management of septic or cardiogenic shock.^{1 2} Data for the utilisation of metaraminol infusions have been predominantly extrapolated from intraoperative use, especially for the short-term prevention and treatment of transient hypotension due to neuraxial anaesthesia in operating theatres.^{35 36} However, patients with transient hypotension in an operating theatre setting are not comparable to critically ill patients with shock because of differences in the underlying pathophysiology and drivers of hypotension.^{6 19} Our study provides some more contemporary insight into the use of metaraminol and its potential association with clinical outcomes.

The cumulative vasopressor-free days until 28 days was slightly lower in those patients treated with metaraminol and norepinephrine. This is in contrast to other vasopressor comparisons in previous trials. For instance, there was no difference in vasopressor-free days until day 28 between critically ill patients treated with norepinephrine versus epinephrine.²³ Additionally, there was no difference in vasopressor duration between critically ill patients who received norepinephrine versus vasopressin in septic shock.³⁷ Those treated with norepinephrine had a higher number of cumulative vasopressor-free days until 28 days versus those treated with dopamine.³⁸ However, dopamine is no longer frequently used in critically ill patients with shock due to an increased risk of arrhythmia.⁸

Baseline characteristics were largely similar in between groups except for the use of concurrent inotropes. Dobutamine and milrinone were used less frequently in the MET-NOR group compared with the NOR group. This may have influenced the time to resolution of shock as inotropes decrease systemic vascular resistance as a compensatory mechanism for the increased cardiac output.^{39 40} This can consequently affect the MAP resulting in increased vasopressor requirements, thus potentially introducing selection bias.^{39 40} Therefore, exploratory sensitivity analyses were conducted to investigate these differences by excluding patients with other concurrent vasopressors or inotropes as reported in the results. However, the results remained consistent with the primary analysis.

Metaraminol infusions were used for a median duration of 4 hours (IQR 2–9 hours). This was slightly shorter than what has been reported in a multicentre observational study involving 70 hospitals in Australia and New Zealand, where the median duration of use was 5.8 hours (IQR 2.7–26 hours).¹⁰ However, the large range suggests that the utilisation in the present study was consistent with this previous study as there is considerable variation between patients. The duration of use could have an impact on outcomes, as it would suggest a greater extent of substitution of first-line therapies such as norepinephrine.

Subgroup analyses in patients with sepsis showed no difference in the time to resolution of shock between groups. However, the present study included patients with all aetiologies of shock and was not powered to detect differences in between people with different shock subtypes such as those with septic or cardiogenic shock. Clinical guidelines vary with vasopressor recommendations according to the type of shock.^1–3 Identifying the type of shock is important to tailor the choice of fluid resuscitation and vasopressor therapy to ensure favourable efficacy and safety outcomes.³

There were no significant between-group differences in exploratory outcomes, except for the requirement for renal replacement therapy. Patients in the MET-NOR group had a lower incidence of renal replacement therapy compared with those treated in the NOR group. However, there was no difference in the duration of renal replacement therapy or the renal replacement therapy-free days until 28 days. There is limited information on the mechanism of action or impact of metaraminol on renal indices. A small study in 10 patients with septic shock showed there was no difference in the acid–base status or base excess when treated with metaraminol and norepinephrine.³¹ Furthermore, a single retrospective observational study in 98 patients with septic shock showed no difference in urine output, urinary albumin, urinary β₂ microglobulin, blood urea nitrogen or urinary creatinine with varying maximum dose rates of metaraminol infusion (range 0.1 to >1.0 µg·kg^-1·min^-1), however, there was no comparative vasopressor assessed.⁴¹ Further investigation may be required to further explore this study finding to ascertain the impact of metaraminol compared with other vasopressors on renal parameters in critically ill patients with shock.

Peripheral infusions were used in 90% of the patients in the MET-NOR group compared with 6% in the NOR group. This is because metaraminol was predominantly used as a peripheral vasopressor. There is a perception that it is safer when peripheral administration is needed. This has developed over years of peripheral use and experience with this agent. Thus, it has come to satisfy a clinical niche, although lack of evidence. It is unlikely that worse outcomes in the MET-NOR group in our study was due to the requirement of peripheral vasopressor infusions,^{14 16} which is primarily driven by the presence of central venous catheters at the time that vasopressors are initiated.⁹

Limitations

This study had some limitations. First, this was a retrospective study design and hence limited with the accuracy of electronic medical record documentation. Second, the electronic medical record only captured data during the ICU admission. The time to resolution of shock would have been longer in patients who started vasopressor therapy in the emergency department or operating theatre. However, the rate of admissions that were started on vasopressors in the emergency department and operating theatre was low and similar across groups. Third, the use of inotropes was different between groups. However, sensitivity analyses were conducted to explore the possible impact of this difference. Fourth, patients with all aetiologies of shock were included. Hence, it is difficult to extrapolate the findings to patients with a specific shock subgroup (ie, septic or cardiogenic). Finally, this was a single-centre study evaluating the practice of the sequential use of metaraminol and norepinephrine, which is done due to the perception that metaraminol is safer as a peripheral agent. The study did not evaluate the adjunctive use of metaraminol when used in combination with other vasopressors. The design was based on the patterns of vasopressor use at the study site.⁹ We were unable to identify comparable patients who received norepinephrine monotherapy versus metaraminol monotherapy, which may be the focus of future investigations.

Implications for practice and further research

Metaraminol may be associated with a longer time to resolution of shock compared with those who do not receive metaraminol. Future prospective studies with well-matched cohorts are required before metaraminol is used as a first-line vasopressor in critically ill patients with shock.

Conclusion

Critically ill patients who received metaraminol and norepinephrine during ICU stay had a longer time to resolution of shock compared with those treated with norepinephrine without metaraminol.

What this paper adds.

What is already known on this subject

• Metaraminol is increasingly being used as a first-line vasopressor agent in the management of shock in critically ill patients in rates up to 42%.

• There are limited efficacy and safety data to support the use of metaraminol over norepinephrine in this patient population.

What this study adds

• Metaraminol is usually used as an initial vasopressor prior to transitioning to norepinephrine.

• In critically ill patients, metaraminol may be associated with a longer time to resolution of shock compared with those who do not receive metaraminol.

Data availability statement

Data are available upon reasonable request.

Ethics statements

Patient consent for publication

Not applicable.

Ethics approval

The study protocol was approved by the local hospital ethics committee at Royal Prince Alfred Hospital, Sydney Local Health District, NSW, Australia (2019/ETH13182).