Early, Goal-Directed Therapy for Septic Shock – A Patient-Level Meta-Analysis. PRISM Investigators; Rowan KM, Angus DC, Bailey M, Barnato AE, Bellomo R, Canter RR, Coats TJ, Delaney A, Gimbel E, Grieve RD, Harrison DA, Higgins AM, Howe B, Huang DT, Kellum JA, Mouncey PR, Music E, Peake SL, Pike F, Reade M, Sadique MZ, Singer M, Yealy DM. N Engl J Med. 2017 Jun 8;376(23):2223-2234.
Since the publication of the landmark Rivers trial in 2001, early goal directed therapy (EGDT) for the treatment of severe sepsis and septic shock has dominated the standard of sepsis care. This study changed the way we approach sepsis care around the world. However, many questioned the results based on patient selection, single center and relatively small sample size to be able to extrapolate this across all patients. Recently, several large randomized controlled trials spanning many countries have evaluated EGDT for severe sepsis and septic shock: ProCESS, ARISE and ProMISe. All of these trials demonstrated no difference in mortality with EGDT compared to usual care. The usual care arms, however, already incorporated many of the components of the Rivers’ intervention protocol.
The investigators of this trial have published a meta-analysis in The New England Journal of Medicine that was designed to address the sickest sepsis (septic shock) patient populations which critics argued weren’t adequately powered to address in the large three trials individually. The investigators of these three trials pooled all their data to increase the power to answer this question in a meta-analysis. The mortality was addressed with a control mortality rate of 25-35%, power of 80% with a two sided p value of 0.05 to detect a mortality difference of 4-5% points at 90 days in patients presenting to the emergency department in septic shock.
The study found no difference in 90-day mortality, 24.9% vs 25.4, EGDT vs usual care. The size of the group with the upper third predicted risk of mortality was four times larger than the entire trial size in the Rivers’ trial. Other findings included higher costs, longer ICU LOS (5.3 vs 4.9 days) and more vasoactive drug use (1.9 vs 1.6 days) in the EGDT group. However, it is important to acknowledge that the majority of the patients in these three trials received early antibiotics and aggressive fluid resuscitation prior to enrollment. Finally, it is important to recognize that in all three trials, the patients had an overall relative mortality risk reduction that mirrors the Rivers’ EGDT trial.
The trio of trials of evaluated by the PRISM investigators do not demonstrate a survival benefit in 90-day mortality when utilizing Rivers’ EGDT protocol. These three trials and Rivers’ study all support the use of early identification, early fluid resuscitation and prompt antibiotic therapy as life saving interventions in severe sepsis and septic shock. The remainder of the protocolized cared, such as CVP targets, hemoglobin transfusion goals, ScVO2 are not necessarily needed and a patient centered approach is necessary.
Multi-institutional, prospective, observational study comparing the gastrografin challenge versus standard treatment in adhesive small bowel obstruction. Zielinski MD, Haddad NN, Choudhry AJ, Cullinane DC et al. J Trauma Acute Care Surg. 2017 Jul;83(1):47-54.
The authors prospectively evaluated the use of gastrografin (GG) in the management of adhesive small bowel obstruction (SBO). They hypothesized that “a standardized adhesive SBO protocol enhanced with the GG challenge will have a greater accuracy for predicting the need for operative exploration, and will decrease rates of operative exploration, length of stay (LOS), and complications”.
Their methods employed a prospective, multi-institutional, observational study comparing centers using standardized SBO protocols with and without GG. The diagnosis of SBO was based on computed tomography (CT) evidence of SBO with a transition point coupled with clinical factors supporting SBO including nausea, vomiting, abdominal pain, and distention. Excluded were patients with an internal or external hernia, recent abdominal surgery within 6 weeks, and a history of pelvic or abdominal malignancy. Also excluded were patients requiring emergency surgery for peritonitis, portal venous gas, closed loop obstruction, pneumatosis intestinalis, closed loop obstruction, and hypotension.
Patients without concern for strangulation had a nasogastric tube (NGT) placed with decompression for 2 hours, then 100 cc of GG mixed with 50 cc of water and given down the NGT. Eight hours later, an abdominal x-ray was obtained to look for the presence of the GG in the colon. A successful GG challenge was defined as the presence of a bowel movement or presence of GG in the colon within 8 hours of the GG challenge. If no colon was present then the presence of ileostomy output was the surrogate. Failed GG was defined by anyone where these criteria did not occur within 8 hours.
Statistical analysis for GG predictability utilized the area under the receiver operating characteristic curve (AUROC), with an alpha of 0.05 and a power of 80%. Also, utilizing an operative intervention rate of 25% vs 42% in GG vs non-GG protocol institutions. Therefore 132 patients per group were needed.
Their findings were the result of 173 patients in the GG group and 143 in the non-GG group. Baseline characteristics were similar, with the exception of less mesenteric edema in the GG arm. Statistical analysis demonstrated a sensitivity and specificity of 87% and 71%, respectively. The PPV and NPV were 92% vs 59%, respectively with an accuracy of 84%. There were no significant differences in complication rates between GG and non-GG groups. After multivariate analysis, patients’ that didn’t receive GG were less likely to resolve their SBO.
The authors have shown that the use of GG in a standardized SBO management protocol is associated with similar complications rates, lower rates of operative intervention, and shorter durations of hospital stay. Future randomized controlled trials are needed to assign causality; however many strengths of the study design have provided us with safety and apparent efficacy of GG as compared to prior studies on this disease entity.
Early Versus Delayed Cholecystectomy for Acute Cholecystitis, Are the 72 hours Still the Rule?: A Randomized Trial. Roulin D, Saadi A, Di Mare L, Demartines N, Halkic N. Ann Surg. 2016 Nov;264(5):717-722.
Acute cholecystitis is a common surgical disease treated by the acute care surgeon. Timing of operative intervention is of paramount importance to minimize risk to the patient and optimize their outcomes. The authors of this paper sought to evaluate the surgical dogma of early operative intervention, within 72 hours of symptoms, for acute cholecystitis.
This study is a single center randomized trial in Switzerland. Patients age of 16 years or older and duration of symptoms for greater than 72 hours were screened for enrollment and the Tokyo guidelines were applied for the diagnosis of acute cholecystitis. Ultrasound evaluations were performed by trained radiologists. Excluded from the study were patients with severe sepsis, pregnancy, immunosuppression, perforated cholecystitis, biliary pancreatitis, cholangitis, and acute pancreatitis. Enrolled subjects were randomized to either early laparoscopic cholecystectomy (ELC) or delayed laparoscopic cholecystectomy (DLC).
Both groups received antibiotics after diagnosis and laparoscopic cholecystectomy (LC) was performed with a 3-port technique under the supervision of an attending surgeon. ELC was performed during the daytime shortly after admission and antibiotics stopped after LC. The DLC group received IV antibiotics for 10-14 days and the LC planned for at least 6 weeks after the initial diagnosis.
The primary outcome was overall morbidity at any time from the point of diagnosis until the 30th postoperative day. Morbidity was defined as any adverse event including antibiotic failure in the DLC that required emergency surgery, unplanned hospital readmission or evaluation while awaiting DLC, and the presence of any complications in the ELC group. The secondary outcomes included LOS, operative time, conversion to open, in-hospital costs and duration of antibiotic therapy.
A power analysis revealed a need for 466 patient sample size to detect a decrease in morbidity by 4% with a power of 80% and a 2-sided type 1 error of 0.05. A planned analysis after 50 patients was performed to evaluate for potential superiority to re-calculate a new sample size.
Ultimately 86 patients were randomized, 42 to ELC and 44 to DLC. 6 of the patients in the DLC did not received their assigned treatment: 3 converted to emergency LC, 2 declined LC after successful antibiotic therapy, and one developed an anaphylactic arrest during induction of anesthesia and the case was aborted. Intent to treat analysis was performed. There were significantly worse outcomes early in the study, favoring ELC. A repeat power analysis demonstrated a need for 86 patients to be randomized. After the 86th patient was enrolled, recruitment stopped.
The demographics of the patients evaluated were not statistically different. The patients had similar age (mid to late 50s), mostly male, BMI less than 30 in both groups, 30 % had hypertension, 10% diabetes and great majority had an ASA 1-II (95%). DLC group had greater morbidity 38% vs 14 %, with a high rate of unplanned readmission/consultation in the DLC group. Postoperative complications were similar between DLC and ELC. Obviously antibiotic duration, hospital LOS and costs were higher in the DLC group. Operative time and surgery related outcomes were similar between the two and there was only 1 conversion to open, ELC group.
There are a few limitations to consider when interpreting these data. RCTs that are truncated for an observed clinical benefit have been demonstrated to show greater effect sizes compared to those trials that weren’t stopped early, potentially overestimating clinical benefit. The data being observed to consider truncation in this analysis, could have been observed at a random high during accrual and not be representative of a larger sample. Therefore, when evaluating subgroup analyses to stop accrual early, often a tighter probability is selected (i.e. p < 0.01 vs < 0.05), to decrease the likelihood of type I error and preserves statistical power. Finally, the type of methodology used to truncate a trial early should be cited, i.e., Pocock, Peto, or O’Brien-Fleming, was not done in this analysis. This was not done in this study. The original sample size calculation was based on a non-inferiority trial design however the trial was truncated based on superiority calculation. The study also relied on subjective time of symptoms based on patient recalled duration and accrual duration was 6 years. These are potential confounders to the data interpretation.
Early laparoscopic cholecystectomy for acute cholecystitis in minimally ill patients, with mild obesity, low ASA grade, and few comorbid diseases, performed during the day time with at least 72 hours of clinical symptoms prior to admission provides for shorter LOS, lower cost and is safe.