Predicting Postinduction Hypotension in Patients for Elective Surgery under General Anesthesia using a Preanesthesia Ultrasonography of the Subclavian Vein

Abstract

Background:

Estimating fluid status before surgery is not routine, but ultrasound can be used for this purpose. Previous studies have shown a correlation between collapsibility indices of the subclavian vein (SCV) and inferior vena cava (IVC). There is a limited data regarding the suitability of studying SVC for predicting hemodynamic events like hypotension during induction anesthesia. By studying SCV width variations prior to surgery, we hope to get an indirect estimate of the fluid status of the patient, which may predict the occurrence of postinduction hypotension during surgery.

Methodology:

This was a prospective, observational, single-center study registered with the Clinical Trials Registry of India. Patients aged 18 years and above, with ASA physical status 1 or 2, undergoing elective surgery with general anesthesia (GA) and endotracheal intubation were enrolled as per inclusion criteria. There were two observers performing the procedure: observer 1 measuring the blood pressure every minute for 10 min and observer 2 doing the measurements of the SCV ultrasonographically. The measurements included SCVDmax, SCVDmin, and SCVCI, which indicate maximum and minimum diameters of the SCV and its collapsibility index. Measurements were taken both during spontaneous breathing, deep breathing, and after passive leg raising.

Results:

In total, 80 patients were recruited for the study, and these data were analyzed. Patients were analyzed into two groups: a hypotension group (defined by absolute mean arterial pressure (MAP) <65 mmHg or a drop in MAP by more than 30% from baseline) and a no hypotension group. There were no significant differences between the two groups for baseline heart rate (P = 0.0673), SCV diameters (maximum and minimum diameters measured in spontaneous, deep, and during PLR), and collapsibility indices (for spontaneous, deep, and during PLR).

Conclusion:

Preoperative measurements of SCV diameter and its collapsibility index during spontaneous breathing, deep breathing, and passive leg raising are not reliable to predict the occurrence of hypotension after induction of GA.

INTRODUCTION

Postinduction hypotension during general anesthesia (GA) is a common yet unpredictable complication. About 10.3% of patients experience a reduction in blood pressure (BP) after induction, termed as postinduction hypotension of GA, but this may range up to 66.96% of patients.[1,2] The clinical importance of this phenomenon includes an increased incidence of adverse postoperative events, including cardiovascular morbidity and acute kidney injury.[3] A multitude of factors and their interplay are responsible for this phenomenon, including the use of antihypertensives, patient-related factors like age, comorbidities, or volume status of patient, which ultimately affect either the cardiac output or systemic vascular resistance.[4] While many of these factors are nonmodifiable, a few, like hypovolemia, can be identified and modified. Estimation of fluid status may be carried out sonographically. The inferior vena cava (IVC) diameter and its collapsibility index have been studied to predict postinduction hypotension after both general and spinal anesthesia.[,6,7]

The subclavian vein (SCV), being more closer to the head end of the patient, may offer a more convenient approach to the anesthetist for fluid assessment. In most people, a linear probe can be used to find the SCV with relative ease. Previous studies in surgical and intensive care patients showed a reasonable correlation between the collapsibility indices of the SCV and IVC.[8,9] Nevertheless, evidence supporting the direct predictive value of SCV assessment for hypotension following induction remains limited, with prior studies showing conflicting results.[10,11] The primary aim of this prospective observational study was therefore to assess the utility of preoperative SCV diameter and collapsibility index measurements – obtained during spontaneous breathing, deep breathing, and passive leg raising – as predictors of postinduction hypotension in patients undergoing GA.

The novelty of this study lies in evaluating the predictive value of SCV measurements under different physiological conditions (spontaneous breathing, deep breathing, and passive leg raise), specifically in the setting of elective surgeries requiring endotracheal intubation under GA. This approach attempts to clarify existing uncertainties from prior literature and further explore SCV as a practical bedside tool in predicting anesthetic-induced hypotension.

METHODOLOGY

This was a single-center, prospective, observational study in a tertiary care teaching hospital conducted between January 2022 and December 2022. Inclusion criteria included (1) all patients aged 18 years and above of either sex, (2) American Society of Anesthesiologists (ASA) physical status classification 1 or 2, and (3) scheduled for elective surgery under GA with endotracheal intubation. Exclusion criteria included patients (i) who were hemodynamically unstable requiring vasopressors to maintain mean arterial pressure (MAP) >65 mmHg, (ii) were on antihypertensive medications in the form of angiotensin-converting enzyme inhibitors (ACEi) and angiotensin receptor blockers (ARBs), (iii) had peripheral vascular disease, (iv) had increased intra-abdominal pressure due to any cause, and (v) had an anticipated difficult airway. We included all elective general surgeries and gynecological surgeries. All patients were enrolled after obtaining informed consent.

Measuring the subclavian vein and its collapsibility index

Two observers were present during the study.

• Observer 1: Examined the patient preoperatively on the day before surgery, ensured that the inclusion criteria were met, and none of the exclusion criteria were present, obtained written informed consent from patients and monitored BP readings before and at each min for 10 min after induction of GA

• Observer 2: Consultant anesthesiologist with experience in ultrasonography performed the SCV measurements prior to the induction of GA.

Standard fasting instructions were followed. All patients received oral pantoprazole 40 mg and oral metoclopramide 10 mg at night and morning of surgery as premedication. Patient was shifted to the operating room. Standard ASA monitors (electrocardiogram, noninvasive BP, and pulse oximeter) were attached, and baseline vitals were noted down. Before the ultrasound evaluation was performed, the patient was made to lie supine and breathe spontaneously for 5 min. Observer 2 performed the ultrasound measurements. The SCV was first visualized using a high-frequency linear array ultrasound probe (6–13 MHz). The transducer was placed beneath the middle third of the clavicle and scanned along the clavicle to the lateral third to obtain a short-axis view of the SCV. The point at which the outline of the short-axis view of the vein was most clearly visible was used for measurements. Then, an M-mode scan was applied, with the cursor in the middle of the vein. The image obtained was frozen to include the maximum and minimum diameters, and then, callipers were used to measure the diameters. To overcome the possibility of foreshortening since the short-axis view was used, observer 2 used a few precautions. The linear high-frequency probe was placed perpendicular to the clavicle, and scanning was done from the middle to lateral third, as per established protocols. The window where the SCV appeared, round and symmetrical with sharply defined anterior and posterior walls, was selected, suggesting perpendicular insonation and minimal foreshortening. For each measurement (spontaneous, deep breathing, passive leg raise [PLR]), observer 2 chose the site along the clavicle where the SCV was most prominently visualized without shadowing or distortion, and where respiratory variation was clearly appreciated. The same anatomical window was used for all three sets of measurements in each patient to avoid angular variation between scans.

The diameter of the right SCV was assessed initially with the patient instructed to breathe spontaneously. A suitable segment was chosen, and the maximum diameter (SCVDmax) and the minimum diameter (SCVDmin) were noted down. After this, the subclavian collapsibility index (CI) was calculated using the following equation:

Three readings were taken with the patient breathing spontaneously, and two closest readings were considered. The diameters and collapsibility index for spontaneous breathing were denoted by SCVDmax-Spont, SCVDmin-Spont, and SCVCI-Spont, respectively. The same was repeated with the patient taking deep breaths. The patient was then given a passive leg raise to 30°–45°, and after 1 min, vitals were noted down. The patient was asked to breathe spontaneously, and the SCV diameter was measured in this position, and collapsibility index was calculated. Observer 1 was blinded to the ultrasound measurements. The ultrasound measurements were done with the machines – Sonosite SII (Fujifilm Sonosite Inc., USA) and GE LOGIQ e portable (GE HealthCare, USA).

Intravenous (IV) line was secured using an 18G cannula (if no access was present), and IV infusion of Ringer’s lactate was started at 10 ml/kg/h.

Baseline vitals were recorded just prior to induction. GA was induced as per standard hospital protocol. The patient was preoxygenated with 100% O₂. IV fentanyl (2 μg/kg) and IV propofol (2–3 mg/kg) were used to induce anesthesia till loss of verbal response. Observer 1 took recordings every minute till 10 min postinduction. After confirming adequacy of bag and mask ventilation, muscle relaxation was achieved using IV vecuronium (0.1 mg/kg). Once 3 min had passed after administering a relaxant, the patient was intubated using an appropriately sized endotracheal tube (ETT). After confirming the correct placement of ETT, anesthesia was maintained using 33% O₂, 67% N₂O, and 0.8%–1.2% isoflurane.

Non invasive blood pressure (NIBP) and MAP were recorded every minute with continuous monitoring of heart rate (HR) and SpO₂ till 10 min postinduction. During this period, observer 1 noted the number of episodes of hypotension. Each episode of hypotension was treated using IV mephentermine 3–6 mg to increase systolic BP near to the baseline value or MAP >70 mmHg. If HR dropped below 50 beats per minute, IV atropine 0.6 mg was given. Till the end of 10 min postinduction, only minimal stimulation, such as surgical preparation of site, was allowed. Once surgery began, no further hemodynamic data were recorded.

The following observations were made [Figure 1]:

Figure 1.

(a) Ultrasound short-axis view of the subclavian vein, (b) M-mode ultrasound of the subclavian vein showing minimum anteroposterior diameter, (c) M-mode ultrasound of the subclavian vein showing the maximun anteroposterior diameter

1. SCV diameter (maximum and minimum) during inspiration and expiration during normal spontaneous breathing (SCVDmax-Spont and SCVDmin-Spont)

2. SCV diameter (maximum and minimum) during inspiration and expiration while taking deep breaths (SCVDmax-Deep and SCVDmin-Deep)

3. SCV diameter (maximum and minimum) during inspiration and expiration after 1 min of passive leg raise (SCVDmax-PLR and SCVDmin-PLR)

4. HR, BP (systolic, diastolic, and MAP), and pulse oximetry recordings at baseline, after 1 min of passive leg raise, prior to induction, and every minute after induction for 10 min

5. Calculations included CI for spontaneous, deep, and PLR (SCVCI-Spont, SCVCI-Deep, and SCVCI-PLR).

Primary outcome

Postinduction hypotension was the primary outcome, defined as the absolute value of MAP <65 mmHg or a decrease in MAP by >30% of the baseline.

Sample size calculation and statistical analysis

The sample size calculation was based on the reference study by Choi et al.[12] Taking a multiple linear regression model with a median effect size based on Cohen’s f² (0.15) to target an α of 0.05 and a β of 0.1 (90% power), a sample size of 73 patients was required. Expecting a dropout rate of 10%, 80 patients were needed. Continuous variables were presented as mean or median after testing for normality using the Kolmogorov–Smirnov test. Categorical variables were represented as numbers and percentages wherever applicable. A pilot study was conducted on 10 participants to assess the intraobserver variability of SCV SCV measurements performed through ultrasound. Each participant underwent three measurements by the same observer under standardized conditions. The intraclass correlation coefficient (ICC [3, 1] and ICC [3, 3]) was calculated using a two-way mixed-effects model with absolute agreement to assess the reliability of single and average measurements, respectively. Baseline characteristics were divided into two groups based on the primary outcome which was the occurrence of postinduction hypotension. Univariate analysis was done using parametric and nonparametric tests wherever applicable to compare the baselines characteristics including the SCV diameters. Scatterplots were constructed by plotting SCV collapsibility index with both lowest MAP postinduction and percentage drop in MAP postinduction. A regression model was applied to get an odds ratio (r²). All statistical analyses were done using SPSS software (version 22.0; IBM Corp., Armonk, New York, USA).

RESULTS

A total of 80 patients were recruited into the study. Among the patients, 51 of them developed postinduction hypotension, which was our primary outcome [Figure 2]. The patients were further analyzed based on the primary outcome. Based on the demographic data summarized in Table 1, we found that there were no significant differences in age, sex, height, weight, body mass index and ASA physical status between the groups.

Figure 2.

Strengthening the Reporting of Observational Studies in Epidemiology flowchart. USG: Ultrasonography

Table 1.

Comparison of demographic data and other characteristics of patients developing and not developing hypotension after induction of general anesthesia

Parameter	Hypotension (n=51)	No hypotension (n=29)	P
Age (years)	44.72±6.74	41.38±7.91	0.542
Males (n)	13	11	0.083
Height (m)	1.54±0.41	1.58±0.62	0.545
Weight (kg)	60.15±13.8	62.07±10.37	0.492
BMI (kg/m2)	24.95±4.73	24.62±3.77	0.372
ASA PS 12	38	21	0.213
	13	8
Comorbidities
Hypertension	9	4	0.76
Diabetes mellitus	3	3	0.66
Hypothyroidism	6	2	0.70

BMI=Body mass index, ASA PS=American Society of Anesthesiologists physical status

The SCV diameters observed while breathing spontaneously, during deep breathing and after 1 minute of passive leg raise and their collapsibility indices derived are given in Table 2.

Table 2.

Comparison of haemodynamic data and subclavian vein diameters

Parameter	Hypotension (n=51)	No hypotension (n=29)	P
Baseline HR (bpm)	74.41±14	77.56±12	0.671
Baseline MAP (mmHg)	100.49±12.68	93.48±12.44	0.009
Lowest MAP after induction (mmHg)	63.18±7.48	74.45±7.96	<0.001
Mean % change in MAP postinduction	36.53±7.88	19.85±6.99	<0.001
SCVDmax-Spont (cm)	0.84±0.22	0.85±0.2	0.435
SCVDmin-Spont (cm)	0.59±0.22	0.62±0.22	0.447
SCVDmax-deep (cm)	0.85±0.22	0.9±0.26	0.147
SCVDmin-deep (cm)	0.46±0.22	0.46±0.17	0.436
SCVDmax-PLR (cm)	0.87±0.2	0.89±0.22	0.314
SCVDmin-PLR (cm)	0.61±0.22	0.63±0.22	0.374
SCVCI-Spont (%)	30.08±15.3	29.08±15.03	0.469
SCVCI-deep (%)	46.43±13.16	49.6±14.9	0.163
SCVCI-PLR (%)	29.76±16.15	29.71±13.54	0.494

MAP=Mean arterial pressure, PLR=Platelet lymphocyte ratio, SCVCI=Subclavian vein collapsibility index, SCVD=Subclavian vein diameter

The ICC for a single SCV measurement (ICC [3,1]) was 0.88 (95% CI: 0.76–0.96), indicating good reliability. The ICC for the average of three measurements (ICC [3,3]) was 0.95 (95% CI: 0.89–0.99), indicating good intraobserver agreement.

There were no significant differences between the two groups for baseline HR (P = 0.671), SCV diameters (maximum and minimum diameters measured in spontaneous, deep, and during PLR), and collapsibility indices (for spontaneous, deep, and during PLR). We found that the group which developed hypotension had a higher baseline MAP prior to induction of GA with P = 0.009. Furthermore, we observed that the lowest MAP after induction to be significantly lower in the hypotension group (P < 0.001). The percentage drop in MAP postinduction was also observed to be lower in the hypotension group (P < 0.001). The percentage drop in MAP was plotted against the SCV collapsibility index during spontaneous breathing (SCVCI-Spont), deep breathing, and passive leg raising [Figures 3-5].

Figure 3.

Scatterplot relating % fall in mean arterial pressure after induction and subclavian vein collapsibility index during spontaneous breathing. The line denotes the trend, and R² denotes the slope and odds ratio

Figure 5.

Scatterplots relating % fall in mean arterial pressure after induction and subclavian vein collapsibility index during passive leg raising. The line denotes the trend. R² denotes the slope and odds ratio

Figure 4.

Scatterplots relating % fall in mean arterial pressure after induction and subclavian vein collapsibility index during deep breathing. The line denotes the trend. R² denotes the slope and odds ratio

These graphs show that there is no relationship between the percentage change in MAP after induction and the SCV collapsibility indices while breathing spontaneously, deep breathing, and passive leg raising. Figures 6-8 show scatterplots relating lowest MAP after induction and SCV collapsibility index during spontaneous breathing (SCVCI-Spont), deep breathing, and passive leg raising.

Figure 6.

Scatterplots relating lowest mean arterial pressure (MAP) after induction and subclavian vein collapsibility index during spontaneous breathing. The red line shows the MAP cutoff used in this study to define hypotension (65 mmHg). The black line denotes the trend, and R² denotes the slope and odds ratio

Figure 8.

Scatterplots relating lowest mean arterial pressure (MAP) after induction and subclavian vein collapsibility index during passive leg raising. The red line shows the MAP cutoff used in this study to define hypotension (65 mmHg). The black line denotes the trend, and R² denotes the slope and odds ratio

Figure 7.

Scatterplots relating lowest mean arterial pressure (MAP) after induction and subclavian vein collapsibility index during deep breathing. The red line shows the MAP cutoff used in this study to define hypotension (65 mmHg). The black line denotes the trend, and R² denotes the slope and odds ratio

After induction of anesthesia, 51 (63.75%) patients developed hypotension. Among the patients who developed hypotension, 20 (39.2%) patients needed a vasopressor, while the other, 31 (60.8%), needed only fluids. Three patients had only one episode, 11 patients had two episodes, 3 had three episodes, and 1 patient had four, five, and seven episodes of hypotension [Table 3]. Whenever used, the vasopressor used was IV mephentermine. The dose of mephentermine used was 6 mg most commonly (12 patients), while seven patients required only 3 mg. One patient required 9 mg to treat the hypotension. Eleven patients required 12 mg, and one patient required a total of 9 mg of IV mephentermine [Table 3].

Table 3.

Number of patients developing hypotension after induction requiring vasopressors and amount of vasopressor used (mephentermine) (n=20)

Number of episodes of hypotension requiring vasopressors	Number of patients
1	3
2	11
3	3
4, 5, and 7	1 each
Total dose of mephentermine used (mg)
3	7
6	12
9	1

DISCUSSION

Point-of-care ultrasound at the bedside is increasingly becoming popular to guide clinical decision-making, especially in the critical care setting.[13] The assessment of IVC is a commonly used method for estimation of fluid status. Taking a cue from this knowledge, similar protocols are also being tried in the perioperative management of patient posted for surgery.[14]

In the current study, more than 63% of patients developed postinduction hypotension. Although it seems to be a large proportion, the incidence varies not only with age, physical status, comorbidities, and medications taken but also with definition of hypotension. The number will be much smaller for a lower permissible MAP values. Choi et al. used the same definition but had only 30% incidence of hypotension in their population.[12] In our study, the baseline BP was the first BP recorded in the theater. The higher incidence of hypotension in our study could be partly explained by the presence of anxiety prior to induction of anesthesia causing a spurious rise in baseline BP readings. Hence, after induction the drop in BP may have seemed excessive (>30% drop) and as a result defined as hypotension. Of the 51 patients who developed hypotension, only 20 needed vasopressors, and mostly, they were all easily treated with fluids and just one or two doses of mephentermine. The adverse consequence of hypotension also depends not only on the degree of hypotension but also to the duration of hypotension, with sustained hypotension likely to cause more adverse effects. We did not note down the duration of hypotension because we treated hypotension as soon as it was recorded with fluids, vasopressors, or both. Even then, it is desirable that hypotension is avoided, especially in vulnerable patient population.

Reduction in preload is the most common reason for hypotension after induction, mostly due to vasodilatation induced by anesthetics and consequent relative hypovolemia. This response will be exaggerated if the patient is low on volume preoperatively. Hence, investigators have tried to predict intraoperative hypotension based on preoperative measurements of preload. Parameters such as central venous pressure and pulmonary capillary wedge pressure require insertion of catheters, and hence, noninvasive methods of assessment of fluid volume are much more attractive.

Recent years have seen a rise in the use of bedside point-of-care ultrasonography for determining intravascular volume status. The intravascular volume status and the clinical response to fluid treatment can both be reliably determined by measuring the IVC.[15,16] Nagi et al. studied 58 patients admitted to intensive care unit with sepsis and studied the correlation between the maximum and minimum IVC diameters and the collapsibility. The stroke volume increased significantly with fluid challenge when the IVC collapsibility was more than 0.32 (fluid responders). They concluded that IVC-CI is a good and sensitive predictor of fluid responsiveness.[17] Although Roy et al. found IVC-CI to be an unreliable predictor of postspinal hypotension,[18] Zhang and Critchley found IVC-CI >43% to be good predictor of hypotension after induction of GA.[19] Similarly, Szabó et al. found IVC-CI >50% to be good predictor of hypotension after induction of GA.[20] Kent et al. showed that the correlation between IVC diameters and internal jugular vein (IJV)/femoral veins is weak. They found that IJV-CI overestimates and FV-CI underestimates need for volume.[9]

Kim et al. conducted a meta-analysis comparing variation in the diameters of IVC, IJV, SCV, and femoral vein with regard to fluid responsiveness.[21] IVC collapsibility index was reliable, IJV diameter changes had a sensitivity and specificity nearly 0.8, but they found insufficient evidence regarding SCV. Perhaps SCV is not as easily collapsible as IVC or IJV. In contrast, Giraud et al. found SCV diameter changes with respiration to be a good predictor of fluid responsiveness.[22] However, their patients were mechanically ventilated (VT of 8 ml/kg +), and they measured vein diameter in long axis.

Choi et al. studied the relationship between SCV collapsibility index measured preoperatively, during spontaneous breathing and deep breathing with hypotension after induction of anesthesia.[12] They found that SCV collapsibility index measured during deep breathing but not spontaneous breathing to be predictive of hypotension after induction of anesthesia. We could not demonstrate such relationship even during deep breathing. Zhou and Hu responded on the published article by Choi et al., claiming that the depth of anesthesia was not well defined. They advised that the depth of anesthesia should be monitored, using the bispectral index so that deep anesthesia can be avoided, particularly in elderly patients. High positive end-expiratory pressure may also impair venous return and produce a drop in BP. The study did not mention whether they ventilated all patients similarly. This may also influence the interpretation of the results. They also opined that measurements may be made thrice and average taken to improve accuracy.[23] It has been claimed that SCV evaluation, particularly SCV collapsibility assessment, is a useful supplement to IVC measurement. However, reports of weak correlations between IJV and femoral vein measurements and IVC readings mean that they should not be employed in clinical decision-making.[24] Choi et al. in their study did not exclude patients taking ACE inhibitors or ARBs. They explained that withholding these medications did not impact the development of hypotension, and that their sample size was not adequate to study the association of these medications with the development of hypotension.[12] In our study, we decided to exclude such patients since a systematic review showed such patients to have an increased relative risk of developing hypotension.[25] Furthermore, in our institute, the choice to withhold these drugs prior to surgery is left to the discretion of the anesthetist. We enrolled patients who were scheduled to undergo surgery under GA, and thus, had a heterogeneous group unlike Choi et al. who included patients posted for laparoscopic cholecystectomy. They emphasized that patients with cholecystitis had a positive Murphy sign owing to an inflamed gallbladder and that IVC assessment would be inconclusive in such patients. They stated that for such patients, SCV assessment was more feasible.[12] In our institute, bispectral index is not commonly used to measure the depth of anesthesia. As a result, this may have interfered with the BP readings as a deeper anesthetic state may have led to hypotension.

We had certain limitations in our study: our surgical group was not standardized. We could not standardize the ultrasound machine. We used any one of the following machines: Sonosite and two GE machines. The resolution was higher and measurements easier with Sonosite but was subject to availability. Further, despite taking all precautions, there is still a possibility of foreshortening during the short-axis measurements of the SCV. Being an observational study, the induction sequence could not be standardized. We did not exclude patients based on BP measured on the previous day and prior to surgery. Anxiety may have spuriously increased their BP and the drop seemed drastic after induction of anesthesia. Comparison of SCV collapsibility was not done with IVC collapsibility or central venous pressure (CVP) also. Hence, further studies can be done to explore the best possible predictor of postinduction hypotension.

CONCLUSION

Preoperative measurements of SCV diameter and its collapsibility index during spontaneous breathing, deep breathing, and passive leg raising are not reliable to predict occurrence of hypotension after induction of GA.

Conflicts of interest

There are no conflicts of interest.

Funding Statement

Nil.