Background
Perioperative hemodynamic state, usually determined by mean blood pressure, urinary output, and heart rate, is highly related to postoperative (POP) morbidity and mortality [
1‐
3], relationship that has been demonstrated in several studies [
4‐
9]. Evidence has shown that an inadequate tissue perfusion during surgery is associated with anastomotic leak during the postoperative period, which is a major morbid complication [
10]. In consequence, objective parameters of tissue perfusion are needed and central venous oxygen saturation (ScvO
2) has been studied since the 1980 [
4,
5].
Tissue hypoperfusion results from circulatory disturbances during hypoxic and hemorrhagic periods, and evidence suggests that ScvO
2 reflects the early stages of hypoperfusion [
6,
11‐
15]. Other hypoperfusion marker is lactic acid, which is elevated during low tissue oxygenation even under normal vital signs. Nevertheless, lactic acid could be elevated by different circumstances like hepatic impairment as well as those treated with Ringer lactate [
15‐
18]. ScvO
2, being an early hypoperfusion marker, could support early and objective decision-making during surgical procedure.
The main purpose of this study was to explore possible associations between tissue hypoperfusion identified through ScvO2 measurement and the development of anastomotic leak during the postoperative period of abdominal trauma patients. Our hypothesis is that low measurements (<65%) may increase the risk of presenting leakages of anastomosis or sutures during POP period.
Methods
Design, setting, and participants
Originally, the study was designed for a case-control study, but in response to the low incidence of cases during the study period, a prospective cross-sectional study approach was preferred. Patients aged 14 years or older, who underwent surgery secondary to abdominal trauma and that required primary gastrointestinal anastomosis, were included in the study. Patients were excluded when they presented hemodynamic or metabolic impairment secondary to previously known pathologies, and patients who died during surgery or in the immediate postsurgical time. As clinical and local conditions for gastrointestinal repair are different after damage control surgery, those that required this procedure before the final repair were also excluded. The study took place at the Hospital Occidente Kennedy, a reference trauma center of high complexity located in Bogotá D.C., Colombia, between 2011 and 2013. Patients underwent surgery according to each surgeon’s personal criteria and to the hospital’s trauma protocols without taking into account the study objectives.
Variables
The following data was collected for every patient: age, gender, anatomic site of trauma, trauma mechanism, blood pressure, transfusion requirements, hemoglobin levels at arrival, comorbidities, abdominal trauma index (ATI), and the ScvO2 measured during the surgery. ScvO2 was measured through a subclavian venous catheter implanted by the anesthesiologist for central venous pressure measurement. As part of the hospital’s protocol, anesthesiologist prepared patients for surgery and implanted the subclavian venous catheter during the first 30 min of entering the surgery room. ScvO2 was measured right after catheter implantation. All remaining data was obtained from clinical records of every patient. Follow-up of patient was done during the hospitalization days until the first postoperative ambulatory control appointment 10 days after discharge or when anastomotic leak was diagnosed.
Anastomotic diagnosis and ScvO2 cut point
Anastomotic leak or suture leakage was diagnosed from clinical manifestations and confirmed through image studies, or surgery during secondary interventions when needed. For this study’s main outcome, a ScvO
2 cut point was defined as low levels previous to analysis. According to literature, normal levels range between 70 and 75% and levels under 60% suggest an oxidative metabolic disturbance [
19]; therefore, the cut point for this study was set at 65%. All patients with ScvO
2 at 65% or lower were classified as low levels.
Sample selection and bias assessment
According to literature, incidence for anastomotic leak is ranged between 2.4 and 15.9% [
1]. Therefore, considering a risk of 10% to develop anastomotic leak in abdominal trauma patients, a leaked/non-leaked relationship of 1:3; a type I error of 0.05 and type II of 0.010; and an overall final sample of 108 were calculated. Ultimately, due to low incidence of patients during the period of time for the study, an overall sample of 41 patients was reached and an approach for a cross-sectional prospective study was decided.
Selection bias was addressed through sequential and continuous selection of patients until a final sample of 41 patients was achieved taking into account inclusion and exclusion criteria. Hospital’s personnel that attended each case were not influenced by the objectives of the study and acted according to usual practice and protocols established in the hospital. Classification bias was addressed by taking into account data obtained from the clinical record of each patient.
Statistical analysis
Qualitative variables were analyzed through descriptive statistics using absolute and relative frequencies. For quantitative variables, central tendency (means and median) and dispersion (rate and standard deviation) measurements were used. For assessing differences in systolic and diastolic arterial pressure, Student’s t test was used for two independent groups. Previous analysis of normalized values on arterial pressure and the homogeneity of variants were done using the Shapiro Wilk test and Levene test respectively. If such circumstances were not achieved, an exact non-parametric Mann Whitney test was used.
A Fisher exact test (expected values of <5) was applied for identifying any association between ScvO2 and the development of anastomotic leak. Risk was assessed using relative risk with a Katz exact confidence interval of 95%. Logistic regression and multivariate analysis was done, but due to the low number of patients, it was excluded from the reports.
Discussion
An association between low levels of ScvO2 during surgery with anastomotic leak as a complication in the postsurgical period was found. Significant difference was identified between mean values of ScvO2 for patients that developed anastomotic leakages and those who did not (60.0% ± 2.94%, and 69.89% ± 7.21% respectively; p = 0.010). A risk of developing an anastomotic leak for patients with low levels of ScvO2 was identified (RR = 39.8 CI 95%: 2.35, p < 0.0001, with a Fisher exact test).
These results support the hypothesis that 65% or lower levels of ScvO
2 may constitute a risk factor for the development of major complications related to gastrointestinal anastomosis and sutures in abdominal trauma patients. Other authors have found similar results [
10,
11,
20‐
25]. Hoshking et al. showed that levels of 66.5% or lower of ScvO
2 are associated with a worst outcome despite a stable hemodynamic state [
15]. Pearse et al. described the changes in ScvO
2 after mayor surgery finding an independent association between levels of ScvO
2 (with a cut point of 65%) and the development of POP complications [
11]. Other cases where low values of ScvO
2 were associated with a worst prognosis are trauma, acute myocardial infarction [
5,
26], severe sepsis, and cardiac failure [
11].
In abdominal trauma, anastomotic leak is one of the most morbid complications [
1,
27‐
30] and hypoperfusion is an important factor for its development. Consequently, different strategies have been developed for the prevention and resolution of hypoperfusion such as goal-directed resuscitation [
9,
31] and goal-directed fluid replacement [
25,
32].
We strongly believe that early and objective hypoperfusion markers can contribute to a supported decision on postponing primary repair. Based on the results from this study, ScvO
2 monitoring during surgery could be a potential and easily available hypoperfusion marker. It is important to mention that values of ScvO
2 may reflect a low oxygen extraction rate secondary to mitochondrial impairment indicating poor outcomes and limits low levels of ScvO
2 sensibility [
15,
33,
34].
Another limiting factor is the catheter placement. It has been proved that ScvO
2 levels vary according to the catheter’s point location [
35]; therefore, it is important to note that the sample was taken from the superior vena cava so measurements could be affected by brain metabolism, specially under anesthesia. Other potential marker is serum lactate, but it was not measured for this study as it can vary in different conditions such as fluid replacement with Ringer lactate [
15‐
18], a common practice in trauma patients. Other limitations of this study were the sample size for the study and that ScvO
2 values were not measured continuously.
A significant association with the requirement of red blood cell transfusion (p = 0.019), low values of hemoglobin at arrival (p = 0.004), and a greater ATI (p < 0.001) was also found. The first two variables are indirect markers of inadequate perfusion, further supporting its association with anastomotic leak development. This suggests that other markers may also help or may become cofounding factors for ScvO2 efficacy. Measurement of other markers (serum lactate, BE, CO2 gap, SaO2, and others), bigger samples, and a proper case-control or randomized controlled trials may be needed for answering these questions. Nonetheless, our results are robust enough to emphasize ScvO2 importance as a hypoperfusion marker and as a risk factor for POP anastomotic leakage in abdominal trauma surgery patients.
Acknowledgements
The authors would like to acknowledge the following:
Miguel Ramirez Gómez, MD. MSc.: Surgical department chief from Hospital Occidente Kennedy
Oscar Geovanni Hernandez, MD. MSc, General Surgeon at Hospital Occidente Kennedy
Sandra Marcela Saade-Lemus, MD and Juan Sebastián Botero-Meneses for the proofreading, grammar, and spelling edition.