Echocardiography Evaluation of the Effects of Midazolam on Passive Leg Raising Test in Critically ill Patients in the Intensive Care Unit, diagnosed with Sepsis, determined to be hypovolemic and responding to uid treatment

This study used echocardiography parameters to investigate the effects of midazolam sedation on intravascular volume in intubated patients who are diagnosed with sepsis and treated with invasive mechanical ventilation in continuous positive airway pressure mode. This study included 152 intensive care unit patients aged 30–50 years with spontaneous breathing; all were intubated, had their lungs ventilated with a positive end-expiratory pressure of 5 cmH 2 O in continuous positive airway pressure mode via invasive mechanical ventilation, had a Ramsey sedation scale score of 5–6 at 5 min after midazolam administration, and exhibited a uid decit (i.e., inferior vena cava collapsibility index > 42% and > 12% systolic arterial pressure increase in the passive leg raising test). Cardiac index, cardiac output, and velocity time integral measurements were taken after the passive leg raising test, before and after midazolam sedation, in patients with hypovolaemia who responded to uid treatment. Changes in passive leg raising test results were compared before and after administration of midazolam. Cardiac output > 15%, cardiac index > 10%, and > 15% increase in velocity time integral during the passive leg raising test before midazolam administration indicated that patients exhibited hypovolaemia and responded to uid therapy. Cardiac output < 15%, cardiac index < 10%, and < 15% increase in velocity time integral during the passive leg raising test after midazolam administration indicated that patients did not exhibit hypovolaemia. before and inferior collapsibility [VCI-CI], variation, and the passive leg test [PLRT]) are used used to assess performance of the PLRT. the on the PLRT. in the Relative hypovolaemia induced by midazolam can the axis of uid therapy regulated through the PLRT; may uid treatment in patients who exhibit a uid decit, as determined using the PLRT, with normal response to uid treatment. In this investigated the effects of midazolam sedation on PLRT using ECHO parameters in intubated ICU patients diagnosed with sepsis and a uid decit, based on VCI-CI and PLRT results, and who responded to uid treatment, as measured by the PLRT. provided through invasive mechanical ventilation in continuous positive airway pressure (CPAP) mode at positive end-expiratory pressure (PEEP) of 5 cmH 2 O. for 1 of 0.05, Type 2 of 0.20, effect of 0.20, and standard deviation of the change in outcome of 0.88; sample size calculations performed using mean CO of 8.71 ± 3.37 L/min following the PLRT, before midazolam administration, and 7.83 ± 3.78 L/min following the after midazolam administration. Patients in the pilot study included in main study. study, CI and SV were strongly correlated with uid responsiveness. (37) In the present study, we performed the PLRT twice. We performed the rst assessment before midazolam administration; we evaluated the outcome by means of SAP, CI, CO, and VTI measurements. We detected signicant increases in SAP, CO, CI, and VTI after the PLRT. We found that SAP in the supine position was 128.32 ± 25.30 mmHg before midazolam administration and before the PLRT, whereas it was 154.03 ± 18.58 mmHg after the PLRT (p < 0.05). A 15.3984 mmHg increase in SAP (12% increase compared to basal value) before PLRT is an indication that patients exhibit hypovolaemia; an increase of 25.71 mmHg was observed in the present study, indicating that our patients exhibited hypovolaemia and responded to uid therapy. When we evaluated the effect of the PLRT on VTI before midazolam administration, we found that VTI was 19.11 ± 4.05 cm before the PLRT and 22.52 ± 4.42 cm after the PLRT. The increase in mean VTI after the PLRT was statistically signicant. An increase of 2.8665 cm in VTI (15% increase compared to basal value) before the PLRT and before midazolam administration is required for diagnosis of hypovolaemia; in the present study, there was an increase of 4.0184 cm, which indicated that our patients exhibited hypovolaemia and responded to uid therapy. We found that CO was 6.71 ± 1.40 L/min in the supine position before midazolam and before the PLRT, whereas it was 8.37 ± 1.60 L/min after the PLRT (p < 0.05). We considered an increase of 1.0065 L/min in CO (15% increase compared to basal value) before the PLRT to be an indicator of hypovolaemia; the CO increase in PLRT before midazolam administration was 2.0706 L/min in our study. This indicated that our patients exhibited hypovolaemia and responded to uid therapy. We found that CI was 4.14 ± 1.26 L/min/m endothelial factors (e.g., nitric oxide), reduce endothelial calcium content, leading to relaxation of vascular smooth muscle cells and subsequent vasodilation. (16,42) Drugs that enhance nitric oxide levels are effective for blood pressure reduction. (16,42) Benzodiazepines have a direct vascular vasodilator effect on arteries and veins. (16,42) In addition, benzodiazepines reduce blood pressure indirectly by regulation of the baroreex system and central inhibition of the autonomic neurocardiac system. (16,42) Benzodiazepines also regulate chloride ion channels, which may mediate benzodiazepine-induced vascular effects. (16,42) In addition, vascular calcium channels sensitive to membrane voltage changes may be inhibited by high concentrations of benzodiazepines. (16,42) At nanomolar concentrations, benzodiazepines bind to non-neurological receptors (i.e., peripheral-type benzodiazepine receptors) in peripheral tissues. (16,42) In one study, systemic administration of the GABA A agonist midazolam to Python molorus caused signicant bradycardia, while the direct cardiac effect of midazolam was not demonstrated; this effect was attributed to midazolam-mediated reduction of cardiac adrenergic tone and enhancement of cholinergic tone. (51) The mean arterial serum concentration of midazolam following intravenous induction (0.2 mg/kg) is 1.36 pg/mL in patients with ischemic heart disease. (18) This is a 95% sucient dose for unconsciousness. (18) Only high doses of midazolam inhibit noradrenaline release during electrical stimulation, via the alpha-1, alpha-2 agonist contraction effect. (18) Clinical concentrations of midazolam do not inhibit noradrenaline release from sympathetic nerve endings or smooth muscle cell vasoconstriction caused by alpha-adrenoceptors. (18) The results of these animal and human studies suggest that our present ndings depended on peripheral vascular vasodilation and the cardiac-depressing effect of midazolam. We presume that the PLRT did not mediate a strong effect after midazolam


Background
Preservation of intravascular volume, vasopressor treatment, and haemodynamic optimisation play important roles in prevention of morbidity and mortality in intensive care unit (ICU) patients with sepsis. (1)(2)(3)(4) Sepsis has been de ned as life-threatening organ dysfunction caused by an irregular host response to infection. (2) Patients with sepsis constitute the majority of critically ill patients hospitalised in the ICU. (5) Airway control also plays an important role in prevention of mortality and morbidity. (2) Provision of sedation is necessary for patients with sepsis who are undergoing orotracheal intubation for airway control. (2,6) Sedative drugs are frequently used in ICU patients. (6) Most of these drugs cause relative hypovolaemia, because they disrupt compensatory mechanisms. (7)(8)(9)(10)(11)(12)(13)(14)(15) Midazolam is a sedative drug frequently used in the ICU. (6) Various animal studies have shown that midazolam causes vasodilation due to its effects on vascular smooth muscle cells and the heart. (16)(17)(18)(19) Static and dynamic parameters (i.e., central venous pressure, vena cava inferior collapsibility index [VCI-CI], vena cava inferior disability index, delta velocity peak, pulse pressure variation, stroke volume variation, and the passive leg raising test [PLRT]) are used to estimate cardiac preload. (2,20−23) Notably, the VCI-CI can be used to assess hypovolaemia without performance of the PLRT. (24)(25)(26)(27) PLRT results and hypovolaemia can also be evaluated by echocardiography (ECHO) parameters. (24,28−37) ECHO is a reliable test to investigate the response to uid therapy in patients exhibit hypovolaemia on the PLRT. (28,29) ECHO is an important tool to identify hypovolaemia and monitor uid resuscitation, particularly in critically ill patients, because of its non-invasive nature, ability to be performed at the bedside, and ability to be repeated frequently. (33) Relative hypovolaemia induced by midazolam can change the axis of uid therapy regulated through the PLRT; it may cause normovolaemia and unresponsive uid treatment in patients who exhibit a uid de cit, as determined using the PLRT, with normal response to uid treatment. In this study, we investigated the effects of midazolam sedation on PLRT using ECHO parameters in intubated ICU patients diagnosed with sepsis and a uid de cit, based on VCI-CI and PLRT results, and who responded to uid treatment, as measured by the PLRT. Respiration was provided through invasive mechanical ventilation in continuous positive airway pressure (CPAP) mode at positive end-expiratory pressure (PEEP) of 5 cmH 2 O.

Methods
This study was carried out between September 2017 and February 2019 at the Gazi Yaşargil Training and Research Hospital, Anesthesiology and Reanimation Clinic ICU. The study used an observational prospective study design, in accordance with the STROBE statement. Study approval was obtained from the local ethics committee; written informed consent to participate was obtained from the rst-degree relatives of the patients included in the study. The study adhered to the tenets of the 2008 Declaration of Helsinki. In a pilot study conducted with 30 patients, cardiac output (CO) measurements were performed. A sample size of n = 152 was calculated for a Type 1 error of 0.05, Type 2 error of 0.20, effect size of 0.20, and standard deviation of the change in outcome of 0.88; sample size calculations were performed using mean CO of 8.71 ± 3.37 L/min following the PLRT, before midazolam administration, and 7.83 ± 3.78 L/min following the PLRT, after midazolam administration. Patients in the pilot study were included in the main study.
Patients were diagnosed with sepsis in accordance with the 2016 sepsis guidelines. (2) Patients with hypoxia or hypercarbia, combined with Glasgow Coma Scale score < 10, were intubated with 1 mg/kg propofol and 1 µg/kg remifentanil administered intravenously. Patients with restored spontaneous breathing were connected to a mechanical ventilator in CPAP mode at 5 cmH 2 O PEEP, and all measurements were made after this procedure.
Inclusion criteria: 1-Patients (30-50 years of age) hospitalised in the ICU 2-Spontaneously breathing intubated patients, whose lungs were ventilated by means of invasive mechanical ventilation at 5 cmH 2 O PEEP in CPAP mode 3-Patients with a Ramsey sedation scale score of 5-6 at 5 min after administration of midazolam 4-Patients with a uid de cit (i.e., > 42% VCI-CI and > 12% increase in systolic arterial pressure [SAP] on the PLRT) 5-Patients who had been hospitalised and taken to the emergency department with a blue code, who presented directly to the emergency department, or who developed sepsis while in the ICU.
Patient age, body temperature, height, weight, duration of intensive care stay, peak heart rate (PHR), peripheral oxygen saturation, intra-arterial pressures, and peripheral body temperatures were recorded before the study procedure was performed. All ECHO (GE Healthcare Vivid S70N Mannheim, Germany) measurements were rst performed by the cardiology specialist; all measurements were repeated by the intensive care specialist. The results were obtained by the intraobserver anaesthesiologist and given to the interobserver anaesthesiologist; all evaluations were then performed by the interobserver anaesthesiologist. Thus, the experts who made the measurements were blinded to each other. The interobserver anaesthesiologist was also blinded to the experts making the measurements. All data were evaluated by recording the average of two measurements.

Passive Leg Raising Test
When the patient was lying in the supine position, the head was raised 45° above the waist and kept in this position for 2 min; SAP was then recorded on the monitor in mmHg. Then, the legs were raised 45° from the waist and the head was restored to its original position for 1 min; SAP on the monitor was then recorded in mmHg. An increase of > 12% in the measured SAP was evaluated in favour of hypovolaemia and the test result was considered positive. (41) Aortic Diameter and VTI Measurements The patient was placed in the supine position and aortic diameter was measured between the adhesion points of the aortic valve from the aortic annulus line, by means of two-dimensional imaging of the parasternal long axis with the ECHO probe (Fig. 3). VTI values of the left ventricular out ow systolic ow velocity  were recorded in the apical window during a single breath cycle with pulsed wave Doppler, 1 cm below the aortic valve.
Study procedure 1-Patients were diagnosed with sepsis in accordance with the 2016 sepsis guidelines. (2) 2-PHR and SAP were measured in monitored patients who had been placed in the supine position. Maximum and minimum VCI diameters were measured and recorded. VCI-CI was calculated and the study was continued in patients with > 42% VCI-CI.
3-VTI was measured by ECHO in patients with > 42% VCI-CI, with patients in the supine position.
4-SAP and PHR measurements were repeated after the PLRT without any medication. The study was continued in patients with > 12% increase in SAP in the supine position. In these patients, VTI measurements were repeated by ECHO and recorded. 5-Midazolam (0.1 mg/kg) was administered according to the ideal weight of the patient.
No procedures were performed for 5 min.
6-SAP and PHR were measured again 5 min after midazolam administration; VTI was measured again in the supine position by ECHO.
7-The PLRT was performed again, and SAP and PHR measurements were recorded. VTI was measured again with ECHO and recorded.
Calculations were made from the recorded data using the following formulas: Statistical Analysis SPSS Statistics for Mac, version 11.5 (SPSS Inc., Chicago, IL, USA) was used to evaluate the data. The normality of the data distribution was evaluated by the Shapiro-Wilk normality test. Normally distributed data were compared with the paired samples t-test; results are shown as means ± standard deviations. The Wilcoxon test was used to compare data that did not t a normal distribution; results are shown as medians (ranges). P-values < 0.05 were considered statistically signi cant in all analyses.

Results
The demographic and clinical data of the patients are given in Table 1. The source of sepsis in the patients and the microorganisms grown in the blood are given in Table 2.
Examination of the effect of the PLRT on SAP after midazolam administration revealed that SAP was 114.62 ± 24.40 mmHg before the PLRT and 117.72 ± 23.55 mmHg after the PLRT (p > 0.05). A signi cant difference in SAP was observed between before the PLRT, before and after midazolam administration (128.32 ± 25.30 and 114.62 ± 24.40 mmHg), and after the PLRT, before and after midazolam administration (154.03 ± 18.58 and 117.72 ± 23.55 mmHg); SAP was lower after midazolam administration (p < 0.05) ( Table 3). A signi cant increase in PHR was observed after the PLRT had been performed, before midazolam administration (p < 0.05). A signi cant difference in PHR was observed between before the PLRT, before and after midazolam administration (113.53 ± 19.18 and 101.80 ± 21.62 beats/min), and after the PLRT, before and after midazolam administration (121.43 ± 16.65 and 106.45 ± 20.65 beats/min); PHR was lower after midazolam administration (Table 3) (p < 0.05).
1.28 L/min/m 2 after the PLRT; the mean increase in CI after PLRT was statistically signi cant (p < 0.01). A signi cant difference in CI was observed between before the PLRT, before and after midazolam administration (4.14 ± 1.26 and 4.02 ± 1.74), and after the PLRT, before and after midazolam administration (5.47 ± 1.28-4.39 ± 1.72); CI was lower after midazolam administration (p < 0.05). While mean CI before the PLRT after midazolam administration was 4.02 ± 1.74 L/min/m 2 , it was 4.39 ± 1.72 L/min/m 2 after the PLRT after midazolam administration; midazolam signi cantly enhanced the mean CI after the PLRT (p < 0.05) ( Table 4).
Examination of the effect of the PLRT on CO before midazolam administration revealed that CO was 6.71 ± 1.40 L/min before the PLRT and 8.37 ± 1.60 L/min after the PLRT (p < 0.01). After midazolam administration, CO was 6.22 ± 1.57 L/min before the PLRT and 6.40 ± 1.67 L/min after the PLRT. The PLRT performed after midazolam administration signi cantly enhanced mean CO (p > 0.05). A signi cant difference in CO was observed between before the PLRT, before and after midazolam administration (6.71 ± 1.40 and 6.22 ± 1.57), and after the PLRT, before and after midazolam administration (8.37 ± 1.60 and 6.40 ± 1.67); CO was lower after midazolam administration (

Discussion
Sedative and anaesthetic drugs are frequently used in critically ill patients in the ICU. (42) Baroreceptors play a very important role in the regulation of dynamic blood pressure; (42) notably, they are depressed throughout general anaesthesia. (42) Most anaesthetics act directly on myocardial contractility and vascular resistance. (42) This clinical condition simultaneously contributes to reduced blood pressure, causes haemodynamic instability, and creates hypovolaemia. (42) The sedative agent propofol reduces systemic vascular resistance and CO and enhances venous capacitance, but only causes small changes in PHR. (42) Sedative and anaesthetic drugs prepare patients for relative hypovolaemia by enhancing venous capacitance, leading to reduced blood volume and CO, as well as the inability to meet tissue oxygen demand and potential onset of hypoxia. (42,43) Therefore, when using sedative and anaesthetic drugs, drugs with a broad therapeutic index and the lowest numbers of cardiovascular side effects and known antagonists should be used. (42,43) Midazolam is a benzodiazepine with known effects on cardiac function, which are well tolerated. Midazolam has a broad therapeutic index and a known antagonist, and its levels can be easily adjusted, particularly in patients who will remain intubated for long periods. (44)(45)(46)(47)(48)(49)(50) In the present study, we initially determined that patients exhibited hypovolaemia when VCI-CI was > 42% (signi cant for hypovolaemia), (24) which we determined in the supine position. We performed the PLRT to examine whether a uid response occurred in patients who exhibited hypovolaemia, according to the VCI-CI (50.95 ± 6.77 ) measurement. Based on SAP results in the supine position before the PLRT (128.32 ± 25.30 mmHg) and SAP results after the PLRT (154.03 ± 18.58 mmHg), an increase in SAP of 15.3984 (> 12%) (41) before the PLRT and before midazolam administration indicated that the patients exhibited hypovolaemia and responded to uid therapy. In the present study, the increase in SAP in the supine position after the PLRT was 25.71 mmHg, which indicated that patients exhibited hypovolaemia and responded to uid therapy.
In one study, VCI-CI was measured by transthoracic ECHO, before and 15 min after an infusion of 500 ml (130/0.4) of 6% hydroxyethyl starch. VTI was measured and a 15% increase in VTI after uid administration was regarded as the uid response. In the same study, patients were considered to exhibit hypovolaemia when VCI-CI was > 40%. (35) In another study, CO was measured by transthoracic ECHO before and after an infusion of 500 ml (130/0.4) of 6% hydroxyethyl starch; the PLRT was also performed. In that study, a 10% increase in CO was considered a positive uid response; a strong positive correlation with the PLRT was observed. Notably, VCI-CI > 42% was regarded as the cut-off for hypovolaemia. (24) In one study, CO was measured by transthoracic ECHO before and after an infusion of 500 ml (130/0.4) of 6% hydroxyethyl starch; a 15% increase in CO was considered a positive uid response. (36) In another study, CI and SV were strongly correlated with uid responsiveness. (37) In the present study, we performed the PLRT twice. We performed the rst assessment before midazolam administration; we evaluated the outcome by means of SAP, CI, CO, and VTI measurements. We detected signi cant increases in SAP, CO, CI, and VTI after the PLRT. We found that SAP in the supine position was 128.32 ± 25.30 mmHg before midazolam administration and before the PLRT, whereas it was 154.03 ± 18.58 mmHg after the PLRT (p < 0.05). A 15.3984 mmHg increase in SAP (12% increase compared to basal value) before PLRT is an indication that patients exhibit hypovolaemia; an increase of 25.71 mmHg was observed in the present study, indicating that our patients exhibited hypovolaemia and responded to uid therapy. When we evaluated the effect of the PLRT on VTI before midazolam administration, we found that VTI was 19.11 ± 4.05 cm before the PLRT and 22.52 ± 4.42 cm after the PLRT. The increase in mean VTI after the PLRT was statistically signi cant. An increase of 2.8665 cm in VTI (15% increase compared to basal value) before the PLRT and before midazolam administration is required for diagnosis of hypovolaemia; in the present study, there was an increase of 4.0184 cm, which indicated that our patients exhibited hypovolaemia and responded to uid therapy. We found that CO was 6.71 ± 1.40 L/min in the supine position before midazolam and before the PLRT, whereas it was 8.37 ± 1.60 L/min after the PLRT (p < 0.05). We considered an increase of 1.0065 L/min in CO (15% increase compared to basal value) before the PLRT to be an indicator of hypovolaemia; the CO increase in PLRT before midazolam administration was 2.0706 L/min in our study. This indicated that our patients exhibited hypovolaemia and responded to uid therapy. We found that CI was 4.14 ± 1.26 L/min/m 2 in the supine position before midazolam administration and before the PLRT, whereas it was 5.47 ± 1.28 L/min/m 2 after the PLRT; thus, PLRT caused a signi cant enhancement in CI. We considered an increase of 0.414 L/min/m 2 in CI (10% increase compared to basal value) before the PLRT to be an indicator of hypovolaemia; the increase in CI after the PLRT before midazolam administration was 1.33 L/min/m 2 in the present study, which indicated that our patients exhibited hypovolaemia and responded to uid therapy.
No signi cant increase in SAP was observed after midazolam administration; however, there were signi cant increases in VTI, CO, and CI. SAP in the supine position after midazolam administration and before the PLRT was 114.62 ± 24.40 mmHg, whereas it was 117.72 ± 23.55 after the PLRT (p > 0.05). An increase of 13.7544 mmHg (12% increase compared to basal value) in SAP before the PLRT and after midazolam administration was an indication that the patients exhibited hypovolaemia. In the present study, an increase of 2.7045 mmHg was observed. The increase in SAP decreased during the PLRT, compared to the measurement before midazolam administration; notably, SAP values decreased in some patients compared to pre-PLRT levels. When we evaluated the effect of the PLRT on VTI after administration of midazolam, we found that VTI was 18.35 ± 4.28 cm before the PLRT, whereas it was 19.39 ± 4.95 cm after the PLRT; this increase in mean VTI after the PLRT was statistically signi cant. For diagnosis of hypovolaemia, VTI should have increased by 2.7525 cm (15% increase compared to basal value) after the PLRT and after midazolam administration; however, our study revealed an increase of 1.04 cm. This indicated that our patients did not exhibit hypovolaemia. The increase in VTI, caused by the PLRT after midazolam, was smaller than the increase in VTI after the PLRT and before midazolam administration. CO measured in the supine position was 6.22 ± 1.57 L/min after midazolam administration and before the PLRT, whereas it was 6.40 ± 1.67 L/min after the PLRT. An increase of 0.933 L/min (15% increase compared to basal value) in CO in the PLRT after administration of midazolam would be important in terms of hypovolaemia; however, in the present study, the increase in CO after midazolam was 0.18 L/min. This indicated that our patients did not exhibit hypovolaemia. The increase in CO after the PLRT was reduced after midazolam administration; moreover, CO levels decreased in some patients. There was a signi cant increase in CO before and after the PLRT after midazolam administration; this increase was insu cient to indicate hypovolaemia. CI measured in the supine position was 4.02 ± 1.74 L/min/m 2 after midazolam, whereas it was 4.39 ± 1.72 L/min/m 2 after the PLRT, indicating a signi cant increase in CI. An increase of 0.402 L/min/m 2 (10% increase compared to basal value) in CI before the PLRT was an indication of hypovolaemia; however, the increase of CI in the PLRT before midazolam administration in our study was 0.37 L/min/m 2 , indicating that our patients did not exhibit hypovolaemia.
Midazolam is frequently used to sedate critically ill patients during the terminal period and is a drug selected for palliative sedation. (45)(46)(47)(48) Midazolam has a two-phase metabolism. (44) The rst phase is hydroxylation via CYP3A; the main metabolite is alpha-hydroxy midazolam and a small amount of 4-hydroxy midazolam. (44) Alpha-hydroxy midazolam exhibits activity that is 80-100% of the level exhibited by midazolam. (44) After hydroxylation, alpha-hydroxy midazolam is converted to glucuronate via UDP-glucuronyl transferase; this metabolite exhibits minimal activity (10% of the level exhibited by midazolam). (44) Peripheral vascular resistance controls blood pressure; (16,42) it is directly dependent on vascular tone arising from the balance between vasodilator and vasoconstrictor factors that affect vascular smooth muscle cells. (16,42) Vasodilation depends on the endothelium; (16,42) endothelial factors (e.g., nitric oxide), reduce endothelial calcium content, leading to relaxation of vascular smooth muscle cells and subsequent vasodilation. (16,42) Drugs that enhance nitric oxide levels are effective for blood pressure reduction. (16,42) Benzodiazepines have a direct vascular vasodilator effect on arteries and veins. (16,42) In addition, benzodiazepines reduce blood pressure indirectly by regulation of the barore ex system and central inhibition of the autonomic neurocardiac system. (16,42) Benzodiazepines also regulate chloride ion channels, which may mediate benzodiazepine-induced vascular effects. (16,42) In addition, vascular calcium channels sensitive to membrane voltage changes may be inhibited by high concentrations of benzodiazepines. (16,42) At nanomolar concentrations, benzodiazepines bind to non-neurological receptors (i.e., peripheral-type benzodiazepine receptors) in peripheral tissues. (16,42) In one study, systemic administration of the GABA A agonist midazolam to Python molorus caused signi cant bradycardia, while the direct cardiac effect of midazolam was not demonstrated; this effect was attributed to midazolam-mediated reduction of cardiac adrenergic tone and enhancement of cholinergic tone. (51) The mean arterial serum concentration of midazolam following intravenous induction (0.2 mg/kg) is 1.36 pg/mL in patients with ischemic heart disease. (18) This is a 95% su cient dose for unconsciousness. (18) Only high doses of midazolam inhibit noradrenaline release during electrical stimulation, via the alpha-1, alpha-2 agonist contraction effect. (18) Clinical concentrations of midazolam do not inhibit noradrenaline release from sympathetic nerve endings or smooth muscle cell vasoconstriction caused by alpha-adrenoceptors. (18) The results of these animal and human studies suggest that our present ndings depended on peripheral vascular vasodilation and the cardiac-depressing effect of midazolam. We presume that the PLRT did not mediate a strong effect after midazolam administration, because uid responsiveness decreased. To the best of our knowledge, no study has examined the effects of midazolam on intravascular volume status of patients or on the PLRT.

Limitations
We evaluated the effects of midazolam on the PLRT with SAP, CI, CO, and VTI in patients who had been diagnosed with sepsis; all patients exhibited hypovolaemia, according to VCI-CI results, and were responsive to uid therapy, according to PLRT results. We did not independently or objectively evaluate the effect of midazolam on dynamic parameters during the PLRT by thermodilution, and we did not evaluate correlations with ECHO parameters. Further observational prospective studies are needed to investigate correlations between parameters measured by the thermodilution method and dynamic parameters viewed by ECHO. Human studies are needed to investigate the effect of midazolam on vascular smooth muscle.

Conclusions
Our study revealed signi cant increases in SAP, CI, CO, and VTI values in the PLRT after midazolam administration in patients with sepsis who exhibited hypovolaemia (VCI-CI > 42%) and responded to uid therapy (SAP > 12%). The results demonstrated signi cant changes in CO (> 15%), CI (> 10%), and VTI (> 15%) in the PLRT in terms of both hypovolaemia and uid responsiveness before midazolam administration; however, these increases were insu cient for diagnosis of hypovolaemia. Thus, the PLRT may be unnecessary after administration of midazolam. When the PLRT is applied to patients, it should be performed before midazolam administration for sedation.