Background

Sepsis is a disease defined as systemic and deleterious host immune response to infection, and it is associated with highest mortality in the Intensive Care Unit (ICU) worldwide [1]. Infection with antibiotic-resistant microorganisms is attributable to a high mortality of septic patients. One of the opportunistic pathogens is Pseudomonas aeruginosa, less frequently found as human microflora in healthy individuals. However, it can lead to sepsis in patients with a broad spectrum of diseases, including burns and respiratory pathologies [2]. In this study, we utilized clinically relevant ovine model for sepsis induced by Pseudomonas aeruginosa, because this pathogen was recognized as a main cause of ventilator-associated pneumonia in the ICU [3] and contributes to higher mortality rate of the patients. Previously, we have shown that ovine and human immune responses were comparable to the TLR-4 agonists, lipopolysaccharide and monophosphoryl lipid ALPS [4]. Also, it has also been shown that cardiovascular hemodynamic changes to the LPS were similar in two species [5, 6]. Current standard treatment for sepsis is mostly limited to use of antibiotics, vasopressors, and fluid resuscitation [7, 8]. However, increasing risk of antibiotic resistant pathogen is in concern particularly for hospital acquired pathogens.

More than a hundred phase II and III clinical trials have been performed testing various therapeutic agents including anti-TNF, anti-IL-1R, and anti-LPS agents, corticosteroids, immunoglobulin, and activated protein C in septic patients [9]. However, none of them proved efficacy of the testing compound. Despite many years of intensive research, no drug has been translated to clinical practice for the treatment of sepsis [10,11,12,13]. In 2017, Paul et al. reported that early administration of intravenous vitamin C, hydrocortisone, and thiamine (further referred to combined therapy) attenuated multiorgan failures and improved survival of septic patients greater than 30% [14].

Vitamin C is a scavenger for free radical and inhibits production of reactive oxygen species [15]. Due to its antioxidant property, vitamin C usage has been considered for sepsis treatment and its positive effect for septic patients have been previously reported [16, 17]. At a low dose, hydrocortisone is used in sepsis mainly to support depleted endogenous steroid function [18, 19]. Thiamine is one of the essential sources of energy production from sugar and indirectly involved in the citric acid cycle [20], and its deficiency leads to lactic acidosis by altering aerobic metabolism. Depletion of thiamine was reported approximately in 20%—70% of the septic patients, depending of criteria to define thiamine deficiency [20, 21]. Therapy with thiamine has previously been reported to be beneficial in sepsis [20, 22].

In contrast to work by Paul et al., recent work by Fujii et al. have reported that combined therapy with vitamin C, hydrocortisone, and thiamine did not lead to a rapid resolution of septic shock and was not superior to hydrocortisone alone [23].

Therefore, to shed some lights into the debated issue, we aimed to investigate the efficacy of the combined therapy (vitamin c, hydrocortisone, and thiamine) in a well-characterized ovine model of sepsis induced by Pseudomonas aeruginosa.

Methods

Care and use of animals

All animals were cared for according to the approved protocol by the Institutional Animal Care and Use Committee of the University of Texas Medical Branch, and studies were conducted in compliance with the guidelines of the National Institute of Health, and the American Physiological Society for the care and use of the laboratory animals.

Reagents

Vitamin C (Aspen, #V-0348-05), hydrocortisone (Pfizer, #0009-0825-01), thiamine (Fresenius Kabi, #45816G), and 0.9% sodium chloride (Baxter, #2B1324X), and Pseudomonas aeruginosa were purchased (ATCC, #27317).

Surgical preparation

Female Merino sheep were surgically instrumented a week before the study as previously described [4]. Briefly, under isoflurane anesthesia via endotracheal tube, a Swan-Ganz thermodilution catheter was inserted to the pulmonary artery through the right jugular vein. A polyvinylchloride catheter was placed in the descending aorta via the right femoral artery. A silastic catheter was positioned in the left atrium through the 5th intercostal thoracotomy. Then, sheep were allowed to recover in the Intensive Care Unit.

Preparation of Pseudomonas aeruginosa

Previously prepared glycerol stock of Pseudomonas aeruginosa was used throughout the study. After overnight incubation, optical density of the bacterial culture was determined at 550 nm (OD550) and colony forming units (CFU) were calculated using the previously obtained formula. Then, the required volume of bacterial culture was spun to obtain the bacterial pellet, that was washed twice with PBS. Finally, 1 × 1011 CFU of live Pseudomonas aeruginosa was suspended in 30 mL of 0.9% sodium chloride solution per sheep.

Bacterial instillation to the sheep lungs

On the day of study, randomly chosen sheep were initially anesthetized with intravenous injection of ketamine (500 mg) and isoflurane inhalation (2–5%) via mask and a tracheostomy tube was placed. Under continued anesthesia via tracheostomy tube, thirty milliliters of bacterial suspension were instilled via bronchoscopy to the lungs (10 mL in lower lobe and 10 mL in middle lobe of the right lung and 10 mL in left lung). To ensure effective delivery of bacteria to the lungs, sheep were kept 10 min under anesthesia following instillation.

Groups allocation and treatment

At each time, a paired (control and treatment) sheep study was performed to ensure the comparable degree of injury. After the bacterial instillation, sheep were randomly allocated into Control (received only saline) and Treatment groups (combined therapy). Treatment was provided as follows: vitamin c, (0.75 g every 6 h), hydrocortisone (25 mg every 6 h), and thiamine (100 mg every 12 h). To enhance the translational aspects, the first dose was given 9 h after injury (approximate time of septic shock onset), where the mean arterial pressure (MAP) dropped 20 points below from baseline value. Treatment doses were calculated and scaled down based on the ratio between the average body weight of human to sheep.

Cardiopulmonary function assessment

Cardiopulmonary function was assessed by recording hemodynamic variables, such as mean arterial blood pressure (MAP), heart rate (HR), cardiac output (CO), left atrial pressure (LAP), pulmonary arterial pressure (PAP), and central venous pressure (CVP) every 3 h after injury. Systemic (SVRI) and pulmonary (PVRI) vascular resistance indexes were calculated using a standard formula. Pulmonary function was evaluated by blood gas analysis (i.e., arterial and venous PO2, PCO2, and saturation) (RAPIDPoint 500; Siemens Healthcare, Erlangen, Germany) and pulmonary mechanics variables (i.e., peak and pause airway pressures, and lung compliance) every 3 h. PaO2/FiO2 ratio and pulmonary shunt fraction was calculated. Arterial blood lactate levels were measured using a blood gas analyzer, as well.

Blood cell count

A complete blood cell count was determined in arterial blood collected at baseline (BL), and every 3 h thereafter, using a complete blood cell count analyzer (ADVIA120 hematology system, Malvern, PA, USA).

Determination of bacterial numbers in the circulation

To determine the bacterial number in circulation, 1 mL of venous blood was collected 12 h after injury. Collected samples were cultured on the Trypticase soy-agar plate at 37 degrees, and colony forming units (CFU) were calculated per milliliter of blood with triplicate. At least five independent samples were analyzed.

Postmortem assays

At the end of the study or when sheep met euthanasia criteria, sheep were humanely euthanized, and half of the right lung tissue was harvested for evaluation of lung water content by the previously described method [24]. The left lung was used for the bronchial alveolar lavage (BAL) to determine bacterial numbers. BAL fluid was harvested after washing the lung with 50 mL of sterile PBS. From the recovered lavage fluid, serial dilutions were made to obtain appropriate colony numbers with triplicated culturing on a trypticase soy agar plate. At least 5 independent samples were analyzed per group.

Data analysis

Statistical analysis was performed with GraphPad Prism 8. Significance was determined when appropriate with either two-way ANOVA, one-way ANOVA multiple group post hoc test or a two-tailed unpaired Student’s t test. p < 0.05 was considered to be statistically significant.

Results

Administration of the combined drugs

During the administration of the combined therapy no adverse effects were observed. Two out of 6 sheep (33.3%) survived in treatment group vs. 1 out of 6 (16.7%) in the control group survived throughout 24 h study period (p = 0.7) (Fig. 5D).

Hemodynamics

During the administration of the combined therapy no adverse effects were observed. There were no significant differences in the cardiovascular baseline values between the control and treatment groups. In the control group, MAP was slightly increased at 1 h post injury, and gradually decreased reaching its lowest level at 9 h (75.5 ± 9.7) vs. baseline (92.5 ± 7) (Fig. 1A). CVP gradually increased from the baseline (5.2 ± 2.8) peaking at 15 h (11.7 ± 1.1) in control group (Fig. 1B). Cardiac output tended to increase in both groups compared to the baseline values. No statistical difference was noted (Fig. 1C). SVRI gradually decreased at 12 h (751 ± 220) from the baseline (1210 ± 199) (Fig. 1D) in control group. These hemodynamic changes seen in animals of the control group were not affected by the combined therapy, except for the heart rate that was significantly (p = 0.002) higher in the treatment group (Table 1).

Fig. 1
figure 1

Cardiovascular hemodynamics changes. A Mean arterial pressure, B central venous pressure, C cardiac output, and D systemic vascular resistance index. Data are expressed as value ± SD

Full size image
Table 1 Cardiopulmonary hemodynamics
Full size table

All baseline values of the pulmonary hemodynamic variables were comparable between the control and treatment groups. In the control group, PAP gradually increased, and the statistically significant increase was observed at 9 h (28 ± 5), and at 18 h (31 ± 5) compared to its baseline (18 ± 1). A similar trend was observed in the treatment group, except the combined therapy significantly reduced (p = 0.04) the elevation of PAP at 15 h, which was 26 ± 4 in the treatment and 38.7 ± 4 in control groups (Fig. 2A). In the control group, peak airway pressure gradually increased reaching statistical significance at 15 h (33 ± 9) compared to the baseline (16 ± 2), and a similar trend was observed in the treatment group (Fig. 2B).

Fig. 2
figure 2

Pulmonary hemodynamics and mechanics. A pulmonary arterial pressure, B peak airway pressure, C mean airway pressure, D plateau airway pressure, E static compliance, and F pulmonary vascular resistance index (PVRI). Data are expressed as value ± SD. *p < 0.05 indicates significant difference between the groups, †p < 0.05 indicates significant difference compared to baseline

Full size image

Mean airway pressure gradually increased in both groups, reaching statistical significance at 15 h (16 ± 3) and kept higher until 24 h, compared to the baseline (9 ± 1), (Fig. 2C). In the control group, the plateau airway pressure gradually increased, reaching statistical significance at 18 h (40 ± 0) compared to the baseline (15 ± 2). While in the treatment group, it significantly elevated at 15 h (33 ± 5), 21 h (33 ± 1) and 24 h (34 ± 5) compared to the baseline (16 ± 5) (Fig. 2D). There were no statistical differences found between the control and treatment groups.

The static compliance gradually decreased in the control group reaching statistical significance at 15 h (27 ± 19) compared to baseline (51 ± 15), which had a similar trend in the treatment group (Fig. 2E). In the control group, pulmonary vascular resistance index (PVRI) was elevated at 15 h (423 ± 411) and reached statistical significance at 18 h (263 ± 151) compared to the baseline (116 ± 35) (Fig. 2F), while in the treatment group, it stayed at a comparable level to the baseline.

Pulmonary gas exchange

Pulmonary gas exchange gradually deteriorated in both groups throughout the study period. The significant reduction of PaO2/FiO2 (ratio of the partial pressure of oxygen in the arterial blood and the partial pressure of inspired oxygen) was observed in both the control (267 ± 114) and treatment (284 ± 135) groups at 9 h compared to the baseline (514 ± 25 in control and 507 ± 32 treatment) (Fig. 3A). Administration of combined therapy showed no beneficial effects on pulmonary gas exchange on both PaO2/FiO2 ratio and oxygenation index (OI) which were comparable in both control and treatment groups (Fig. 3B).

Fig. 3
figure 3

Pulmonary gas exchange. A ratio of partial pressure arterial oxygen and inspired oxygen and B oxygenation index. Data are expressed as value ± SD

Full size image

Microvascular hyper-permeability

Administration of combined therapy showed no beneficial effect on lung water content (p = 0.3) evaluated by measuring lung wet-to-dry weight ratio, which was 7.9 ± 2.9 in control and 7.7 ± 2.6 in treatment groups (Fig. 4A). The plasma protein level was significantly decreased at 18 h (3.3 ± 1.0 in control and treatment 3.8 ± 0.6), compared to their baseline (6.5 ± 0.4 in control and treatment 6.5 ± 0.4). No significant differences were noted between the groups (p = 0.9) (Fig. 4C). Accumulated net fluid balance increases seen in the control group was not attenuated by the combined therapy (p = 0.1) (Fig. 4D). At necropsy, the thoracic fluid level was comparable (p = 0.5) between both control (625 ± 256 mL) and treatment groups (726 ± 256 mL) (Fig. 4B).

Fig. 4
figure 4

Microvascular hyper-permeability. A lung wet-to-dry weight ratio, B thoracic fluid accumulation, C plasma protein, and D accumulated net fluid balance. Data are expressed as value ± SD. †p < 0.05 indicates significant difference compared to baseline

Full size image

Blood cell count

Baselines values for the blood cells counts were comparable between the control and treatment groups. In the control group, the number of white blood cells were sharply decreased reaching the statistically significant difference at 3 h (1276 ± 270) versus the baseline (5480 ± 1119) and remained significantly lower throughout the study compared to the baseline. While eosinophil, polymorphonuclear cell percentage and delta neutrophil index remained at comparable level to the respective baselines. Combined therapy did not affect these variables. Starting at 3 h, neutrophil number sharply decreased (408 ± 208) vs. the baseline (2303 ± 783) and remained significantly lower until 18 h compared to the baseline in the control group. In the treatment group, the neutrophil number was significantly lower throughout the study starting at 3 h (393 ± 114), compared to its baseline (2368 ± 863). A similar trend was observed in the neutrophil percentage changes in two groups (Table 2).

Table 2 Blood cell analysis
Full size table

Bacterial and lactate clearance

Twelve hours after injury, a circulating bacterial number was similar (p = 0.9) in both the control (343 ± 190) and treatment (343 ± 191) groups (Fig. 5A). However, bacterial number in BALF was significantly higher (p < 0.01) in treatment group (1.7 ± 0.4) vs. control (1 ± 0.2) group (Fig. 5B). The lactate level was similarly increased in both groups, reaching statistical significance in control (6.6 ± 5.1) and treatment (6.6 ± 3.8) at 12 h vs. baseline (0.6 ± 0.2) (Fig. 5C).

Fig. 5
figure 5

Bacterial clearance and survival. A The bacterial number in the circulation and B bacterial number in bronchoalveolar lavage fluid (BALF). C Lactate level and D survival of septic sheep with or without combined therapy. Data are expressed as value ± SD. *p < 0.05 between treatment and control groups. P.A: Pseudomonas aeruginosa. †p < 0.05 indicates significant difference compared to baseline

Full size image

Discussion

In this study we report that combined treatment with vitamin C, hydrocortisone, and thiamine had no effect on multiorgan dysfunctions and the mortality of septic sheep. We have chosen the ovine model as the sheep physiologic [25], hemodynamic [5, 6] and genomic responses to TLR agonists, such as lipopolysaccharide and Monophosphoryl lipid A were reported to be quite similar to those in humans [4].

Previously, it was reported that the combination of vitamin C, hydrocortisone, and thiamine exhibited beneficial effects in septic patients and improved the survival rate by 30% [14]. The authors showed that 28 percent of the patient’s blood was positive for gram positive bacteria (Escherichia Coli) and procalcitonin level was > 2 ng/ml [14]. Procalcitonin clearance as well as a sequential organ failure assessment (SOFA) score was improved in these patients by 72 h after treatment [14]. Importantly, the mean duration of vasopressor requirement (hour) was significantly (p = 0.001) shortened with the treatment vs. control group.

In contrast, a recent clinical trial by Fujii et al. showed that combined therapy with vitamin C, hydrocortisone, and thiamine did not significantly improve the survival of patients with septic shock [23]. In that study by Fujii et al., patients were included with suspected (or documented) infection and exhibited lactate levels greater than 2 mmol/L. The 90-day survival was 71.4% in control vs. 75.5% in the treatment group. The SOFA score was significantly higher in the treatment group than the control. The vasopressor requirement was comparable in both groups [23]. Despite the similar pulmonary origin of sepsis in both studies [26], the exact reason for the different outcomes is not clear.

Here we report our current study results indicating that the combined therapy exhibits no benefits for cardiopulmonary functions and survival of septic sheep. Combined therapy did not attenuate any measured variables i.e., MAP, pulmonary gas exchange, pulmonary mechanics, lung microvascular permeability, and lactate clearance. Our data supports the outcome of studies by Fujii et al. In the present study, the combined treatment did not affect numbers of circulating bacteria in septic sheep.

It was reported that multiple organ failure inversely correlates with concentration of circulating vitamin C in septic patients [26, 27], and its intravenous supplementation reported to be beneficial in multiorgan functions. Vitamin C supplementation has been shown to reduce endothelial hyper-permeability, improve responsiveness to vasoconstrictors, and scavenge reactive oxygen species [28,29,30,31].

Recent evidence indicates the critical role of mitochondrial dysfunction in organ injury during sepsis. As a key cofactor for mitochondrial aerobic respiration and redox status, thiamine has been reported to improve mitochondrial function as evidenced by reduced organ tissue histologic changes and cellular oxygen consumption [32, 33].

Hydrocortisone has been recommended as a useful treatment to decrease mortality in hemodynamically unstable septic shock patients, who are non-responsive to fluids and moderate dose of vasopressor and benefiting from anti-inflammatory and adrenal insufficiency supplementation therapies [34].

To note, in the present study, the bacterial number in the BALF was significantly (p = 0.004) higher in the treatment group (1.8 ± 0.4) than the control group (1.0 ± 0.2). As mentioned, we instilled the same amounts of bacteria into the lungs of both groups of sheep conducted side-by-side (treatment and control sheep study). Although, the exact mechanism of the higher number of bacteria in BALF of the treated group remains unclear, our data may suggest that the bacterial clearance might be reduced by the treatment. This may be related to the significant decrease of circulating neutrophils in the treatment group. Our finding is supported by the previous report that the higher mortality of patients treated with vitamin C, hydrocortisone, and thiamine was observed when white blood cell numbers were less than 15.000/mm3 [35]. Additionally, we speculate that the combined therapy might support the bacterial infection by sustaining the bacterial propagation. We further speculate that use of hydrocortisone may be associated with immune suppression. Previous studies reported immune suppressive properties of the hydrocortisone at higher dose [1, 36].

Thiamine biosynthesis occurs in bacteria, but not in the human body. During the thiamine synthesis pathway, thiamine monophosphate kinase catalyzes the final step of the pathway by phosphorylating thiamine monophosphate to thiamine pyrophosphate, as the essential cofactor for various crucial cellular processes in bacteria [37]. This may suggest that exogenous administration of thiamine might support the bacterial growth. In regard to vitamin C, we speculate that it may have no effect on the number of BALF bacteria in our present study as it has been shown to inhibit that of Pseudomonas aeruginosa biofilm formation without affecting bacterial propagation [38].

Our study has few limitations: (1) Our conclusion on the effects of each tested agents on the bacterial growth and clearance is rather speculative; (2) Efficacy of each individual therapy was not tested; (3) No dose dependent study has been performed for each of the compounds; (4) Effects of the combined therapy was tested only in Gram-negative sepsis; (5) The study duration is relatively short; and (6) Symptomatic therapies, such as antibiotics and vasopressors were not used. Nevertheless, results of our present study demonstrate that a combined therapy failed to attenuate severity of sepsis-associated multiorgan functions and mortality. Our results are supported by results of recent clinical and basic science studies reporting inefficiency of combined therapy with vitamin C, hydrocortisone and thiamine [39,40,41].

Conclusions

Based on the obtained results, we conclude that intravenous administration of the combination of vitamin C, hydrocortisone, and thiamine did not exert benefits in the ovine model of sepsis and septic shock induced by Pseudomonas aeruginosa.