Abnormal shock index exposure and clinical outcomes among critically ill patients: A retrospective cohort analysis. Maheshwari K, Nathanson BH, Munson SH, Hwang S, Yapici HO, Stevens M, Ruiz C, Hunley CF. Journal of Critical Care. 2020 Jun;57:5-12.
The shock index (SI) was introduced in 1967 as a quick and effective way to measure the severity of hypovolemic states. The SI can be calculated using a ratio of heart rate to systolic blood pressure. Since the surviving sepsis campaign, significant focus has been placed on volume status and resuscitation in the critically ill patient. The authors of this well done paper set out to determine if an abnormal SI reading could predict mortality in the general ICU population. Secondary outcomes were to determine the association of morbidity and mortality on cumulative exposure to SI ≥0.9 as well as any relationship between SBP ≤100-mmHg exposure and mortality and its comparison with prolonged SI≥0.9 exposure.
This study was a retrospective chart review using a large medical records health database from US hospitals. Data was examined from January 1, 2010 through June 30, 2018. Excluded patient populations included less than 24 hour ICU length of stay, spinal trauma, brain injury or third degree burns. Abnormal SI thresholds were set at ≥0.9 from previously published literature.
In total, 18,197 patients were included in the study from 82 US hospitals. The mean age of included patients was 62.7 with an APACHE III score of 52. Hospital mortality was 7.8%. Mean length of stay for survivors was 4.2(5.3) ICU days. A single SI reading of ≥0.9 predicted a 90.8% mortality and for every 0.1 increase in maximum SI in the first 24 hours of ICU admission, odds of mortality increased by 4.8% [95%CI p<0.001]. Length of SI ≥0.9 exposure was significantly correlated with in-hospital mortality, AKI, MI, and ICU/hospital length of stay with 4 hour exposure to SI ≥0.9 having an increased odds of death by 5.8%. Acute kidney injury and myocardial injury with also increased by 4.3% and 2.1% respectively. Systolic blood pressure of ≤100-mmHg for ≥2hours was associated with in-hospital mortality.
In conclusion, the authors did an excellent job of examining the relationship of shock index and the ability to predict mortality. They were able to demonstrate that a single SI ≥0.9 value is not predictive of mortality, however prolonged exposure of abnormal SI values can increase mortality and other comorbidities. This study does have limitations including the fact that it is a retrospective database review. Further, this study does not compare SI to more recent technologies such as the CHEETAH NICOM system. Further studies are needed to elucidate the clinical management of patients with abnormal SI readings.
The Sequential Clinical Assessment of Respiratory Function (SCARF) score: A dynamic pulmonary physiologic score that predicts adverse outcomes in critically ill rib fracture patients. Hardin KS, Leasia KN, Haenel J, Moore EE, Burlew CC, Pieracci FM. J Trauma Acute Care Surg. 2019 December;87(6):1260-1268.
There is a wide body of literature available on the rib fractured patient with vast improvements in the care of these injuries. Multiple methods to improve the care of these patients have been implemented including locoregional pain control methods and surgical stabilization. However, even with the improvements in the care of these patients, complications such as pneumonia are still common. Multiple scoring systems have been developed over the years including the Chest Wall Organ Injury Scale, Thoracic Injury Score, the Rib Fracture Score and several others. The goal of this paper was to develop a new scoring system that would be dynamic test that could identify high-risk patients and predict those likely to have a complication as well as increased narcotic use.
The new scoring system was termed the Sequential Clinical Assessment of Respiratory Function (SCARF) score and consisted of four clinical parameters: Incentive spirometry (<50% of predicted), respiratory rate, numeric pain score, and strength of cough. Each score value was given a numeric value of 0 or 1 with equal weighting in the scoring system for a total of 4 points. A prospective cohort study was implemented at a large level 1 trauma center for a 1 year period. Scores were collected daily while patients were in the ICU.
Data was collected on 100 patients in total with a mean age of 52 years old and 70% being male. Rib fracture pattern included 23% with ≥6 fractures and 31% of patients with bilateral fractures. The median SCARF score was 2, with median admission score of 2 and median maximum score of 2. For those patients that developed pneumonia, median admission SCARF score was 3 compared to 2 for those that did not. Median admission SCARF score was 3 for patients with ICU LOS longer than 3 days versus 2 days for those with shorter LOS (p<0.01). Increased oxygen requirements also demonstrated a higher median admission SCARF score of 3 vs. 2 compared to those without higher oxygen requirements (p<0.01). A patient with a worsening SCARF score had a likelihood of a threefold increased risk of developing pneumonia (p=0.04) and were more likely to stay in the ICU for longer than three days. Increasing SCARF score also demonstrated an increase in daily narcotic consumption.
The authors in this paper have developed an easy to use scoring system that can have reasonable predictive ability for pneumonia, increased oxygen requirements, and ICU length of stay. This tool is designed to be used on a daily assessment, with worsening scores predictive of complications. The authors argue that this tool can help risk stratify patients and help with predicting resource allocation. Limitations to this study include a small sample size and potential for variability in implementation of the tool. This tool, and others like it may allow for earlier intervention in patient care to prevent complications.
Oxygenation Saturation Index Predicts Clinical Outcomes in ARDS. DesPrez K, McNeil B, Wang C, Bastarache J, Shaver C, Ware L. Chest. 2017 December;152(6):1151-1158.
ARDS is the result of alveolar-capillary injury and associated impaired diffusion/transfer of oxygen from the alveoli to the bloodstream. Correspondingly, blood oxygen content (PaO2) for a given content of inhaled/alveolar oxygen (FiO2) is reduced. The P:F ratio (PaO2/FiO2) is clinically used to report the degree of alveolar-capillary membrane injury/dysfunction and is a central to the Berlin classification of ARDS: Mild (200 < P:F < 300), Moderate (100 < P:F < 200) and Severe (P:F < 100). Though the role of the alveolar-capillary membrane is central in this process as is FiO2 with regards to diffusion/transfer of oxygen to the blood stream, diffusion across any membrane is also dependent upon pressure of the gas/entity being diffused. The P:F ratio does not take other variables that effect PaO2, such as mean airway pressure (MAP), into account. MAP, for example, independent of FiO2 administration can be highly variable depending upon physiology and ventilator settings yet directly effect PaO2. The variables not accounted for in P:F calculations in combination with multiple studies showing that P:F is not an independent predictor of mortality in ARDS has resulted in exploration for other measures such as the oxygenation index (OI [FiO2/PaO2 x MAP x 100]) to better assess the severity of alveolar-capillary dysfunction. OI was first investigated in pediatric populations and has since been demonstrated to be an independent risk factor for mortality in ARDS. However, both P:F and OI require ABGs to be performed and recent evidence supports use of SaO2 in lieu of PaO2 (example SOFA scoring). The oxygenation saturation index (OSI [FiO2 x MAP x 100/SaO2]) has evolved as equivalent to OI in mortality prediction in pediatrics and more sensitive for ARDS diagnosis than diagnosis using ABG obtained PaO2 measurements. In the study presented, the authors hypothesized that the OSI is a noninvasive surrogate for OI in adults with ARDS that is associated with increased mortality and reduced ventilator-free days (VFDs).
The study was retrospective and included patients enrolled in the VALID study at a single institution from 2009-2016. 917 patients had ARDS by Berlin Criteria, these were further screened for SaO2 <97% (changes in PaO2 have minimal effect on SaO2’s >97%) data in their electronic medical records that permitted calculation and comparison of OI vs. OSI on the day of ARDS diagnosis resulting in 329 patients with records for inclusion in the study. Outcomes included: hospital length of stay (HLOS), length of ICU admission (ICU-LOS), ventilator free days (VFD) and hospital mortality. OI and OSI were strongly correlated (Spearman rho = 0.862; p , 0.001). There was no statistical difference in P:F between survivors and non-survivors. OSI was statistically lower in survivors than non-survivors; however OI was not significantly different between survivors and non-survivors. In multivariable logistic regression controlling for age, sex and APACHE II score, OSI was independently associate with mortality; whereas, P:F, OI, lung injury score (LIS) and Assessment of the severity of hypoxemia by using non-invasive pulse oximetry (SaO2/FIO2) (used as a substitute for P:F in SOFA scoring) were not independently associated with mortality. Regarding VFDs, increased OSI and OI were independently associated with significantly fewer VFDs; whereas higher LIS, lower P:F and lower SaO2/FIO2) were not. The individual performance of OSI, OI, P:F, SaO2/FIO2, LIS and APACHE II score to predict hospital mortality was tested by calculating the AUC. OSI (AUC, 0.602), OI (AUC, 0.562), APACHE II (AUC, 0.695) and SpO2/FIO2 (AUC, 0.595) had moderate performance for mortality prediction, although the AUC for OI did not reach statistical significance. Because the association between OSI and mortality is reported as stronger in children, sub analysis of OSI restricted to patients < 40 years old was performed and the AUC was substantially better (AUC, 0779). Overall, OSI and OI on the first day of ARDS diagnosis strongly correlated with each other and both were independently associated with fewer VFDs whereas P:F ratio was not. Only OSI was significantly associated with mortality and both OSI and OI were superior to P:F in prognostic performance and based on ROC analysis OSI may have stronger performance than OI as the later did not reach statistical significance for mortality prediction. OSI on the day of ARDS diagnosis performed as well as the OI in predicting clinical outcomes, is simple to calculate and continuously available while offering more prognostic information than the traditional measure of ARDS severity, P:F ration, while avoiding invasive arterial blood gas monitoring.
Positive Fluid Balance and Association with Post-Traumatic Acute Kidney Injury. Hatton G, Du R, Wei S, Harvin J, Finkel K, Wade C, Kao L. JACS. 2019 Nov 14;S1072-7515(19)32169-6.
Critically ill trauma patients frequently require blood and fluid resuscitation to restore intravascular volume. However, fewer than 50% of critically ill, hemodynamically unstable patients are fluid responsive, defined by an increase in stroke volume of 10% to 15% after a fluid challenge. Liberal, in excess of hemodynamic responsiveness, fluid administration and fluid positivity after trauma has been shown to increase mortality and worsen organ dysfunction. Acute kidney injury (AKI) after trauma is most likely multifactorial however acute kidney injury after trauma is independently associated with mortality, multiple organ dysfunction, length of stay, cost of care, and discharge to a destination other than home. The relationship between fluid positivity and renal dysfunction after trauma has not previously been evaluated though this is a factor that clinicians can control. The objective of the study was to explore the relationship between fluid balance and post-traumatic acute kidney injury. The hypotheses evaluated in this study were that positive fluid balance is common and is independently associated with AKI development after severe trauma. To test the authors’ hypotheses, a retrospective cohort study was conducted at a single Level 1 Trauma center and included trauma patients > 16 years old who required ICU admission and did not die in the first 48 hours over a 6 month period (Jan – June 2017). Additionally, patients with pre-existing end-stage renal disease, congestive heart failure, >20% TBSA burns and rhabdomyolysis were excluded.
Demographic, past medical history, injuries, length of stay, ventilator days, and in-hospital mortality were obtained from the trauma registry. Laboratory results, quantity of fluids, and type of fluids administered were extracted from the medical record. Fluid balances were computed by subtracting all outputs from all intakes, without capture of insensible losses. Patients were split into 3 groups based on 48-hour fluid balance: fluid negative = fluid balance below -2 liters, euvolemic = fluid balance between -2 and +2 liters, and fluid positive = fluid balance above +2 liters. The diagnosis of AKI was made, and stages assigned based on Kidney Disease Improving Global Outcomes (KDIGO) criteria via comparison of highest 1st week Cr and baseline Cr. AKI was diagnosed if criteria for AKI stage 1 or higher was met in the first 7 days from admission. A total of 364 patients met criteria and 8 of those had missing data points, so analysis was performed on 356 patients.179 (49%) were fluid positive, 8 (2%) were fluid negative and 177 (49%) were euvolemic at 48 hours after admission. The investigators combined the fluid negative and euvolemic patients into a single group because of similar patient characteristics. Fluid positivity was associated with lower admission blood pressure, worse base deficit and more severely injured than patients who were fluid negative. The fluid positivity was independently associated with an increased risk of AKI with a relative risk of 1.98 by multivariable analysis. Increased age was also independently associated with AKI. With regards to mortality, fluid positivity was associated with a higher, but not statistically significant, relative risk of mortality (1.65, p=0.37), whereas higher ISS and age were independently (p<0.05) associated with increased mortality. The authors further evaluated the reason for fluid administration: hemorrhage or inflammatory response. Though incidence of AKI was higher in the hemorrhage group (64% vs. 18%), among patients with inflammatory response 100% of patients who developed AKI had stage 3 AKI.
In this study the incidence of fluid positivity (>2 liters positive) at 48 hours was common and ~ 50%. Fluid positivity of >2 liters was independently and incrementally associated with AKI, with the severity of AKI being higher if the need for fluid administration was not hemorrhage but inflammatory based. Though fluid administration is based on a myriad of clinical parameters, consideration of fluid responsiveness should be considered and investigated as an end point of volume resuscitation to prevent unnecessary fluid administration and subsequent AKI.