Effect of treatment applied in sepsis on intensive care unit and hospital stay: how effective are albumin/steroid/vasopressor agents?

The incidence and prevalence of sepsis have increased in recent years and it is the most common cause of intensive care admission. The aim of this study was to determine the effects of albumin, steroid, and vasopressor agents and other possible factors on the duration of intensive care unit and hospital stay in sepsis patients. Open access data set obtained from Tohoku Sepsis Registry database was used. Four hundred sixty-two patients admitted to intensive care unit with the diagnosis of sepsis were divided into four groups according to their intensive care unit (≤ 5 or > 5 days) and hospital length of stay (≤ 24 or > 24 days). Demographic data, vital signs, laboratory values, mechanical ventilation requirement, and treatment protocols such as albumin, steroid, and vasopressor agent use were used in the evaluation of the groups. The use of albumin (odds ratio [OR] = 3.76 [95% confidence interval (CI), 2.16–6.56]; p < 0.001), steroids (OR = 2.85 [95% CI, 1.67–4.86]; p < 0.001), and vasopressor agents (OR = 3.56 [95% CI, 2.42–5.24]; p < 0.001) were associated with an increasing risk of prolonged intensive care unit length of stay. Also, it was found that the use of albumin (OR = 3.43 [95% CI, 2.00–5.89]; p < 0.001), steroids (OR = 2.81 [95% CI, 1.66–4.78]; p < 0.001), and vasopressor agents (OR = 4.47 [95% CI, 3.02–6.62]; p < 0.001) were associated with an increasing risk of prolonged hospital length of stay. In addition, prognostic scoring systems, body temperature, mean arterial pressure, pH, PaO2/FiO2 ratio, and mechanical ventilation requirement in the first 24 h were also found to be associated with length of stay in intensive care unit and hospital. There was a significant relationship between platelet count, creatinine, Na, lactic acid, and time between diagnosis of sepsis and source control and intensive care unit length of stay, and between hematocrit and C-reactive protein and hospital length of stay. The use of albumin, steroid, and vasopressor agents has been found to be significantly correlated with both intensive care unit and hospital length of stay. Further studies are needed to determine in what order or at what dosage these agents will be administered in sepsis treatment.


Background
The incidence and prevalence of sepsis, the most common cause of admission to intensive care unit (ICU) for critically ill patients, are increasing globally (Perner et al. 2016). However, sepsis mortality has decreased by 20-30% with advances in sepsis treatment (Gaieski et al. 2013;Kaukonen et al. 2014). This issue has been the concern of the studies conducted in recent years. The studies investigating the long-term outcomes of sepsis are of great interest because of the decrease in mortality and insufficient sensitivity to demonstrate the effect of acute interventions.
The predictability of intensive care unit length of stay (ICULOS) and hospital length of stay (HLOS) is important for both the intensive care unit physicians and patients and their relatives. ICULOS/HLOS has been used in the evaluation of various diseases or surgical procedures in many publications in recent years. ICULOS is affected by the severity of the disease and the rapid and on-site interventions (Knaus et al. 1993). In the literature, there are limited studies investigating the factors affecting ICULOS and HLOS in sepsis. Especially in the treatment of sepsis, the effect of albumin use, vasopressor agents, or steroid use on prognosis is still controversial. Knowing these factors will help to reduce the length of stay in the ICU. Therefore, the aim of this study is to determine the effects of albumin, steroid, and vasopressor agents and other possible factors on ICULOS and HLOS in 462 patients followed up with the diagnosis of sepsis in the ICU.

Methods
The data set (Kudo et al. 2018b) obtained from Tohoku Sepsis Registry database (UMIN000010297) which includes data from patients with sepsis in ICUs of 3 university and 7 community hospitals in Tohoku region in the northern part of Japan and utilized by Kudo et al. (2018a) was used in our study. The study of Kudo et al. (2018a) has been approved by the institutional review board of each institution. All institutional review boards concluded that there was no need for patient information and consent form, as it was an observational study that did not require any treatment other than treatment administered in daily clinical routine according to Japanese guidelines (Kudo et al. 2018a).
Patients who were admitted to the ICU with the diagnosis of sepsis or who were diagnosed as sepsis after being admitted to the ICU between January and December 2015 were included prospectively in the study (Kudo et al. 2018a). The diagnosis of severe sepsis or septic shock was established according to the international sepsis guidelines published in 2012 (Dellinger et al. 2013). A total of 616 patients were enrolled in the Tohoku Sepsis Registry. In our study, patients younger than 18 years of age, patients who received aggressive treatment for the first 4 days (n = 43), and patients who died during the hospitalization (n = 111) were not included in the present analysis. Data analysis was performed with 462 patients.
These data included demographic data such as age, gender, body mass index (BMI), and comorbidities. Scoring values such as Acute Physiology and Chronic Health Evaluation (APACHE II), Glasgow coma scale (GCS), and sequential organ failure assessment (SOFA) at admission were also included. In addition, vital signs of the first day (body temperature, pulse rate, mean arterial pressure (MAP), and respiration rate), laboratory results (leukocyte, platelet count, hematocrit (HCT), blood glucose, creatinine, Na, K, lactic acid, C-reactive protein (CRP), procalcitonin, bilirubin, pH, and PaO 2 / FiO 2 ratio) were included. The patients' need for mechanical ventilation (MV) in the first 24 h, and such treatment data as the time from the diagnosis of sepsis to the source control or the start of antibiotics were present in the recorded sets. Treatment information data include the use of albumin during fluid resuscitation, vasopressor agents, and steroid for hypotension. Total ICULOS and HLOS data were obtained from the records.
In our study, the median value was used instead of the mean for the cut-off value, since the ICULOS and HLOS were not normally distributed. The median time for ICULOS was 5.5 days and for HLOS 24 days. Four hundred sisty-two patients were divided into four groups according to ICULOS (≤ 5 or > 5 days) and HLOS (≤ 24 or > 24 days). The flow chart of our study was presented in Fig. 1. Our study was conducted in accordance with the Helsinki Declaration Principles.

Statistical analysis
Mean, standard deviation (SD), median, minimum, maximum, frequency, and ratio values were used in the descriptive statistics of the data. Distribution of variables was measured by Shapiro-Wilk test. T test and Mann-Whitney U test were used for the analysis of quantitative independent data; chi-square test, Fisher's exact test when chi-square test conditions were not met, was used for the analysis of qualitative independent data. Univariate binary logistic regression analysis was used to determine risk factors. Data with missing values were excluded from the analysis. Significance level was accepted as p < 0.05 in 95% confidence interval for all analyses. SPSS 22.0 (Statistical Package of Social Sciences Inc., Chicago, IL, USA) software was employed to analyze the data.
Univariate binary logistic regression analysis was performed for each of the risk factors and the coefficients of the risk factors that were significant as a result of the analysis were given in Table 2.

Discussion
In this study, we aimed to determine the relationship between ICULOS/HLOS and albumin, steroid, and vasopressor agent use and to identify possible risk factors for prolonged ICULOS/HLOS in 462 patients followed up with the diagnosis of sepsis in the ICU. As a result of our study, there was a significant relationship between application of albumin, steroid, and vasopressor agent and ICULOS/HLOS. In addition to these factors SOFA, APACHE II, and GCS scoring systems, body temperature, MAP, pH, PaO 2 /FiO 2 ratio, and need for MV in the first 24 h were found to be significantly related with both ICULOS and HLOS. Moreover, platelet count, creatinine, Na, lactic acid, and time between diagnosis of sepsis and source control were found to be correlated with the prolonged ICULOS, and HCT and CRP were found to be correlated with prolonged HLOS.
It is still controversial whether the use of crystalloid solutions or colloid solutions is better suited for both resuscitation and maintenance in sepsis and septic shock. Albumin has been shown to play a critical role in a variety of diseases and has a serious effect due to its oncotic properties (Caironi and Gattinoni 2009). In a randomized controlled study (Caironi et al. 2014), it was shown that albumin use was not associated with a decrease in mortality, but 20% albumin administration was beneficial in achieving the targeted MAP in the first hour and contributed to improving fluid balance over the next 7 days. In The Saline versus Albumin Fluid Evaluation (SAFE) study (The SAFE Study Investigators 2004), 4% albumin and normal saline applications were compared in intensive care patients, and no difference was found between the groups in terms of mortality. In the CRISTAL study    (Annane et al. 2013), there was no significant difference in 28/90 days mortality between 4 and 20% colloid solutions and normal saline applied groups. In the Albumin Italian Outcome Sepsis (ALBIOS) study (Caironi et al. 2014) in which a serum albumin level of 3 g/dl was targeted for 28 days in septic patients, although higher serum albumin levels were obtained in the treatment group (albumin replacement with a 20% solution), there was no difference in 28/90 day mortality. In addition, these studies showed that albumin administration does not benefit organ failure or mechanical ventilation duration. In a meta-analysis (Delaney et al. 2011) evaluating a total of 17 studies, it was reported that albumin use was associated with lower mortality in patients with sepsis. As a result, there is no clear information about the benefit or harm of albumin in sepsis, so the Surviving Sepsis Campaign (SSC) (Rhodes et al. 2017) guideline recommends adding albumin if there is a significant amount of crystalloid requirement. As a result of our study, it was also found that there was a significant relationship between albumin use and ICULOS/HLOS. In patients with sepsis, albumin replacement was associated with approximately 4-fold increase in the risk of prolonged ICULOS; increased the risk of prolonged HLOS approximately 3.5-fold. Vasopressor agents increase blood pressure by increasing peripheral vascular resistance. Vasopressor agents can be utilized in patients who are hypotensive despite adequate fluid treatment and who develop cardiogenic or pulmonary edema (Keeley et al. 2017). Auchet et al. A meta-analysis (Avni et al. 2015) reported an 11% reduction in 28-day all-cause mortality with norepinephrine. Cochrane systematic review (Gamper et al. 2016), the efficacy of vasopressor agents for the treatment of any circulatory failure was evaluated, and the mortality benefit was not demonstrated in all direct comparisons between different vasopressor agents or vasopressor agent combinations. In another review (De Backer et al. 2012), focusing only on the comparison of norepinephrine and dopamine in septic shock, it has been shown that norepinephrine has an advantage over dopamine in 28-day all-cause mortality. Clinical outcomes other than mortality have been rarely reported, so it was not possible to obtain strong evidence for ICULOS/HLOS. Due to the lack of detailed information about the applied vasopressor agents in the data set used in our study, the differences between agents could not be evaluated, but as a result of our study, the probability of prolonged ICULOS/HLOS was found to be significantly higher in patients receiving vasopressor agent compared to patients not receiving it. The 2016 SSC guideline (Rhodes et al. 2017) recommends the use of intravenous hydrocortisone (200 mg/ day) in patients whose hemodynamic stability cannot be achieved with vasopressor agents. When the related literature is inquired, between 1976 and 2018, 24 randomized clinical trials were published examining the association between steroid use and 28-day mortality in sepsis or septic shock. These studies have conflicting results. The use of steroids has been found to be advantageous in terms of survival in some studies, while no survival benefit has been shown in others (Vandewalle and Libert 2020). Annane et al. ( 2002) found in 2002 that the steroid use in septic shock reduced mortality. However, in the CORTICUS (Sprung et al. 2008) study conducted in 2008, it was found that steroid use did not provide any benefit on mortality In 2016, in the HYPR ESS study (Keh et al. 2016), it was found that hydrocortisone did not prevent the development of septic shock in hospital acquired sepsis patients. In a review of steroid use in sepsis (Gibbison et al. 2017), 22 studies were examined, and only 2 studies indicated that steroid use reduced ICULOS, but there was insufficient data regarding HLOS. In addition to these studies, large-scale studies such as Adjunctive Corticosteroid Treatment in Critically Ill Patients with Septic Shock (ADRENAL) (Venkatesh et al. 2018) and Activated Protein C and Corticosteroids for Human Septic Shock (APROCCHSS) (Annane et al. 2018) have been carried out to clarify the use of steroids. While there was no reduction in 90-day mortality in the ADRENAL study with steroid use, a 6.1% decrease was found in the APROCCHSS study. In addition, in these two studies, the duration of shock resolution, weaning time from mechanical ventilation, and hospital stay were found to be lower in patients receiving steroid therapy. In our study, the probability of prolonged ICULOS/HLOS was found to be approximately 3 times higher in patients receiving steroids. This may be attributed to the difference in the patient population (age, concomitant disease, etc.) or the steroid agent/dose difference used.
The relationship between the need for MV and ICU-LOS/HLOS has been shown in some previous studies (Cislaghi et al. 2007;Güler 2009). Cislaghi et al. (2007) showed that, in patients with coronary artery by-pass grafting, prolongation of MV duration was significantly correlated with both ICULOS and HLOS. In another study dealing with patients undergoing coronary artery surgery, MV duration was found to be one of the factors affecting ICULOS (Güler 2009). Unlike these studies, in our study, it was shown that not the duration of MV but the need for MV on the first day of the patients' diagnosis of sepsis had a significant impact on ICULOS/HLOS. CRP is a protein produced in the acute phase of inflammation. Khaled et al. (2014) found that the first day CRP value was significantly higher in the general ICU patients than the group staying in the ICU more than 7 days. Farah et al. (2018) showed that significant decreases in CRP levels in pneumonia patients on day 2 were associated with shorter HLOS and rapid recovery. Similar to these studies, HLOS and CRP levels were found to be correlated in our study as well.
Thrombocytopenia is a common condition with high mortality in ICUs. In a study excluding hematological diseases, sepsis was reported to be one of the most important causes of thrombocytopenia in patients in ICU (Levi and Löwenberg 2008). Coşkun et al. (2016) found, too, that ICULOS was higher in patients with thrombocytopenia than in those without thrombocytopenia, similar to our study.
In a study investigating the prognostic value of HCT and its utility in the decision of erythrocyte transfusion of anemic patients, Mudumbai et al. (2011) detected an increase in long-term mortality in patients with HCT values less than 25% without transfusion. As a result of our study, low HCT values were found to be associated with prolonged ICULOS/HLOS too. Toptas et al. (2018) targeted to determine the factors affecting the ICULOS in patients followed up in the ICU. Similar to our study, they found a negative correlation between HCT levels and ICULOS in their study.
The 2016 SSC guideline (Rhodes et al. 2017) recommends that 30 ml/kg intravenous crystalloid fluid be administered within the first 3 h. Since the data on fluid replacement was not sufficient in the data set we used in our study, the effect of differences in fluid replacement on ICULOS/HLOS could not be evaluated.

Conclusions
As a result of our study, in patients with sepsis followed up in ICU, the use of albumin, steroid, and vasopressor agents causes a significant increase in ICULOS and HLOS. In addition, it was found that the need for MV in the first 24 h and GCS, APACHE II, and SOFA scoring systems can be used in the prediction of prolonged ICU-LOS/HLOS. If it is desired to create scoring systems that allow the calculation of the estimated length of stay, in addition to these parameters, platelet count, respiratory rate (admission to the intensive care unit), and PaO 2 / FiO 2 ratio can be used for prolonged ICULOS and MAP, HCT, and CRP can also be used for prolonged HLOS.