After adjusting for multiple confounders, patients who received rhAPC had statistically significantly lower in-hospital mortality according to a multivariable logistic regression analysis (odds ratio, 0.76; 95% confidence interval, 0.66-0.86; P<0.001). Propensity score matching was used to validate the risk adjusted multivariate regression model, and the results were similar. Patients with thrombocytopenia and coagulopathy also were found to have lower, and statistically significant, adjusted risks of hospital mortality.

In October 2011, drotrecogin alfa (activated) [Xigris] manufacturer Eli Lilly and Company withdrew rhAPC from the market and announced the discontinuation of all other ongoing clinical trials. While further discussions about rhAPC might be considered moot, it is worth noting that previous rigorously conducted clinical trials occasionally have yielded discordant results compared to observational, “real-life” investigations. This particular study is hampered by several limitations, including confounding by indication, since patient demographic data were not available. Moreover, data about bleeding complications were not available. As with any observational trial, unmeasured confounders might also explain differences in measured outcomes.

The recent moratorium on the use of rhAPC appears to represent yet another failure of immunomodulatory agents to impact outcomes for critically ill patients with sepsis; however, the infrequent use of rhAPC, in selected populations, was found to be associated with a significant improvement in adjusted hospital mortality in this study. One important finding from this trial is that implementation programs, such as the Surviving Sepsis Campaign, have the potential to change clinical practice. Though rhAPC was rarely used overall, its use did increase over time once rhAPC was endorsed by Surviving Sepsis Campaign. 

Concise Critical Appraisal is a regular feature authored by SCCM member Samuel M. Galvagno Jr., DO. Each installment highlights journal articles most relevant to the critical care practitioner.

The authors conducted two parallel phase 3, multicenter, randomized controlled trials comparing dexmedetomidine with propofol (PRODEX trial) or with midazolam (MIDEX trial) for patients requiring prolonged mechanical ventilation. The combined studies involved 998 patients in 75 ICUs throughout Europe andRussia. Included were patients 18 years or older receiving mechanical ventilation with an expected duration of 24 hours or longer and in need of light to moderate sedation, Block randomization was employed and all patients and study personnel were masked to treatment allocation with a double-dummy design. The primary efficacy outcomes measured were the proportion of time in target sedation range (Richmond Agitation Sedation Scale [RASS] score 0 to -3) without use of rescue therapy and the duration of mechanical ventilation from randomization until patients were liberated. Secondary outcomes included length of ICU stay, ability to cooperate with care and ability to communicate pain.

Both dexmedetomidine/midazolam and dexmedetomidine/propofol were not statistically different in terms of time to desired sedation level. Median duration of mechanical ventilation was shown to be shorter with dexmedetomidine (123 hours; interquartile range [IQR], 67-337) compared to midazolam (164 hours; IQR, 92-380; P = 0.03), but not with dexmedetomidine compared to propofol (P = 0.24). Among the dexmedetomidine group, patient interaction was also improved when compared to both the midazolam and propofol groups. There were no significant differences between groups regarding length of ICU and hospital stays as well as mortality rate. The use of dexmedetomidine was associated with more episodes of hypotension and bradycardia.

The results of this study are consistent with those of the Safety and Efficacy of Dexmedetomidine Compared with Midazolam (SEDCOM) trial. This is the first large-scale study to compare dexmedetomidine with propofol in long-term sedation. Non- inferiority in maintaining the target sedation level with dexmedetomidine compared with propofol and midazolam was achieved. Dexmedetomidine also appeared to reduce the duration of mechanical ventilation compared with midazolam. The fact that patients who received propofol and midazolam were more sedated than those receiving dexmedetomidine (though all were in a broad range of target RASS) might have resulted in improved outcomes stemming from the decreased sedation level and not due to the inherent properties of the agent used. While this well-designed and methodologically rigorous study provides additional evidence to support the use of dexmedetomidine as a first-line sedative in the ICU, further tangible long-term outcome data, as well as confirmation that the benefits are due to choice of sedative, are needed to justify the substantially higher cost of treatment and to consolidate dexmedetomidine’s place in the intensivist’s armamentarium. Dexmedetomidine is scheduled to be available as a substantially cheaper generic preparation within the next year. 

Concise Critical Appraisal is a regular feature authored by SCCM member Samuel M. Galvagno Jr., DO. Each installment highlights journal articles most relevant to the critical care practitioner.

Special thanks to Ayal Romen, MD, who contributed to this installment of Concise Critical Appraisal. Romen is a fellow in pulmonary critical care at the University of Maryland School of Medicine.

Among cardiac arrest patients, 15,030 received epinephrine versus 402,158 patients who did not. In the adjusted propensity-matched analysis, epinephrine use was associated with better return of spontaneous circulation (hazard ratio [HR], 2.5; 95% confidence interval [CI], 2.03 to 2.48), but worse one-month survival (HR, 0.54; 95% CI, 0.43 to 0.68) and worse functional outcomes (HR, 0.21; 95% CI, 0.10-0.44). In all other adjusted analyses, including various sensitivity analyses, epinephrine was found to have a negative association with survival at one month and with functional outcomes, despite an improvement in immediate return of spontaneous circulation. 

The authors performed a rigorous analysis using state-of-the-art statistical techniques to adjust for known confounders, but because epinephrine was not assigned in a random fashion, it is possible that selection bias due to unmeasured confounders was not accounted for.  Should clinicians consider rejecting the current standard of care that strongly endorses epinephrine as a frontline medication in cardiac arrest? Although this is the largest observational study on this topic to date, these results are only applicable for cardiac arrest patients outside the hospital. The study’s findings must be confirmed by additional efforts that include in-hospital resuscitation data. The authors’ data create equipoise for potential future randomized studies that would have previously been considered unethical.

Concise Critical Appraisal is a regular feature authored by SCCM member Samuel M. Galvagno Jr., DO. Each installment highlights journal articles most relevant to the critical care practitioner.

A total of 184 patients were enrolled (n=87 in the amantadine group) at 11 clinical sites. Eligible patients were 16 to 65 years old who had sustained a non-penetrating TBI. Patients were enrolled four weeks after injury. Amantadine was started at 100 mg twice daily and gradually increased to 200 mg twice daily by week four of the treatment protocol. All patients were treated in a rehabilitation unit, where an intention-to-treat analysis was performed.

In a comparison of the two groups, more patients in the amantadine group had favorable outcomes according to the Disability Rating Scale score (DRS), and fewer remained in a vegetative state. Recovery was faster in the amantadine group as measured by the DRS score (slope, 0.24 points per week; P = 0.007). No significant difference in the incidence of adverse events was reported between the amantadine and placebo groups.

The authors concluded that administration of amantadine between four and 16 weeks after brain injury significantly improved the rate of functional recovery over a four-week treatment period. The results of this study might be confounded by the effect of concomitant psychoactive drugs and individual responses to rehabilitation therapy, as well as other factors. Although amantadine was not given until four weeks following injury –a time when patients had already been discharged from the intensive care unit — this is an important study because, to date, no intervention has been shown to accelerate the rate of recovery and functional outcome in severe TBI and the vegetative state that ensues in 10% to15% of patients.

Concise Critical Appraisal is a regular feature authored by SCCM member Samuel M. Galvagno Jr., DO. Each installment highlights journal articles most relevant to the critical care practitioner.

Of the 896 patients enrolled in the study, 293 (32.7%) received TXA. These patients showed a 6.5% absolute reduction in in-hospital mortality despite having statistically significantly higher Injury Severity Scores. Patients in the massive transfusion cohort who received TXA showed a 13.7% absolute reduction in in-hospital mortality (relative reduction of 49%). A statistically significant correction of coagulation markers was observed in the patients who received TXA.

As with many observational, retrospective studies, there are several inherent limitations in this study, most notably the variability in the dosage and timing of TXA. The cause and time of death also could not be determined based on the nature of the study, introducing a possible survivorship bias. The 30-day outcome was difficult to ascertain for host-nation patients because of the nature of their discharge. Finally, patients who received TXA had a statistically significantly higher rate of thromboembolic events, although the number of events was too small to make any independent risk assessments.

This study demonstrates a positive impact of TXA for blood component-resuscitated combat casualties. The administration of TXA helps to better correct trauma-associated coagulopathies and appears to improve survival, especially in those patients who require massive transfusion.

Concise Critical Appraisal is a regular feature authored by SCCM member Samuel M. Galvagno Jr., DO. Each installment highlights journal articles most relevant to the critical care practitioner.

Special thanks to Stephanie Luekel, MD, who contributed to this installment of Concise Critical Appraise. Luekel is a fellow in trauma and critical care medicine at the R Adams Cowley Shock Trauma Center at the University of Maryland School of Medicine.

According to the study, published in the Journal of the American Medical Association, 508 patients received trophic feeding and 492 patients received full feeding. Both groups were comparable at baseline. There was no difference between groups with regard to the primary endpoint of ventilator-free days; the trophic group averaged 14.9 days, the full group 15.0 days. Sixty-day mortality was not statistically different between the groups (23.2% vs. 22.2% in the trophic vs. full-feeding groups). No significant differences were found between groups in other secondary endpoints, including infectious complications and organ failures. Trophic feedings were associated with less gastrointestinal intolerance. The authors concluded that in patients with ALI, initial trophic feeding for up to six days did not improve outcomes compared with full enteral feeding.

Limitations in this study include possible reporting bias due to the open-label design, and the fact that most patients were treated in medical intensive care units. Underweight patients (i.e., malnourished) were excluded, and 60-day mortality might have been underestimated because patients discharged home were presumed to still be alive. Nevertheless, this rigorously conducted study showed that gastric feeding is feasible early in the course of treatment for ALI and is associated with fewer gastrointestinal complications, although no difference in outcomes is evident.

Concise Critical Appraisal is a regular feature authored by SCCM member Samuel M. Galvagno Jr., DO. Each installment highlights journal articles most relevant to the critical care practitioner.

Several other outcome measures were also assessed. There were no significant differences between the prone and supine patient groups in terms of pulmonary function tests, days of mechanical ventilation (p = 0.22), and intensive care unit length of stay (p=0.15). The mortality rate was 64.1% in the prone group vs. 56.8% in the supine group. Arterial oxygenation, and the percent of well-aerated lung tissue, was slightly higher in the supine group.
There were no statistically significant differences in SF-36 scores between the groups.

While this study consisted of a very small number of patients, only a few studies have examined long-term outcomes for ARDS patients. Importantly, despite recommendations in 1994 suggesting that HRQL be considered for future outcomes studies, this study is one of the few critical care studies attempting to measure it. The results from this study showed that the mortality rate for ARDS was high at one year and that lung function — as well as HRQL — was not statistically different among patients ventilated in the supine vs. prone position. While the results from this study may not be particularly robust based on the small sample size, the methods stand as an example of how future critical care outcomes studies might be conducted.

Concise Critical Appraisal is a regular feature authored by SCCM member Samuel M. Galvagno Jr., DO. Each installment highlights journal articles most relevant to the critical care practitioner.

There appears to be much room for improvement in terms of better antibiotic stewardship with hospitalized patients. In previous studies, up to 50% of antimicrobial use is believed to be inappropriate. (1,2) In this study, a significant number of patients were both under- and over-treated. Moreover, the class of antibiotics most commonly prescribed – quinolones – is known to have limited activity against enterococci. Misclassification bias and reviewer bias were two potential limitations in this retrospective study. Moreover, co-infection with other bacteria may have partially confounded the results. Both the authors and the invited commentators conclude that, due to the rarity of infectious complications associated with Enterococcus, ABU ought not to be treated.

The results from this study support current efforts to curtail inappropriate antimicrobial use through stewardship programs, focused antimicrobial use campaigns, and mindful practice.

1. Davey P, et al. Interventions to improve antibiotic prescribing practices for hospital inpatients. Cochrane Database Syst Rev2005; (4): CD003543.

2. Flanders SA & Saint S. Enhancing the safety of hospitalized patients: Who is minding the antimicrobials? Arch Intern Med 2012; 172(1): 38-40.

Concise Critical Appraisal is a regular feature authored by SCCM member Samuel M. Galvagno Jr., DO. Each installment highlights journal articles most relevant to the critical care practitioner.

The Society joined HQA in 2008, recognizing a unique opportunity to contribute to the rapidly evolving national discussion around hospital quality measurement and public reporting. Using public hospital data can be a little like turning to TripAdvisor for hotel reviews or to Angie’s List for advice on selecting contractors — anecdotal and akin to the characteristics of Lake Wobegon (all the hospitals have great performance, all the patients are highly satisfied, and all the outcomes are above average). However, in the last five years, the HQA has streamlined the process for reviewing and endorsing structure and outcomes measures for impartial public reporting.  These rigorously chosen data are reported on the U.S. Department of Health and Human Services Hospital Compare, now the nation’s broadest compendium of publicly available, widely accessible, comparable quality measures. Hospital Compare reports on more than 50 performance measures for inpatient and outpatient care, allowing the public and healthcare providers to compare the performance of more than 4,500 hospitals across the nation.

HQA also catalyzed adoption of the Hospital Consumer Assessment of Healthcare Providers and Systems (HCAHPS), the first standardized survey for measuring patients’ perceptions of their hospital care. This survey is now routinely included in all patient satisfaction instruments used by public and private hospitals; it provides a level of rigor and consistency long sought as a benchmark for improving quality. Additionally, HQA endorsed the nation’s first measures for surgical site infections; hospitals that publically reported these infections saw improved outcomes.

A consistent challenge (and gripe) from providers and healthcare institutions is that measurement and reporting consume a disproportionate amount of resources and often overlap due to the use of redundant measures. Further, public reporting of some measures has never been rigorously shown to drive performance improvement or better patient outcomes. HQA, with its partners at the Centers for Medicare & Medicare Services and The Joint Commission, has consistently prioritized streamlined reporting to reduce the burden, identifying high-impact measures that drive improvement and discontinuing reporting on measures that are insensitive or “topped-out” (close to 100% compliance).

So why is HQA closing up shop unlike virtually any other Washington, D.C.-based organization?

The multi-stakeholder HQA measure review model served as the template for other quality alliances and was incorporated into provisions in the Affordable Care Act. Included as part of that transformative law was the outline for a federally mandated quality improvement focus. Now fully realized as the National Quality Forum’s Measures Application Partnership (MAP), the law requires a multi-stakeholder group to identify quality gaps and then specify, validate and endorse public reporting measures that will be used for accountability, performance review and value-based purchasing.

The MAP has kicked into high gear, and HQA is winding down with a measure of satisfaction in transitioning this important work to the National Quality Forum. SCCM Past-President Mitchell Levy, MD, FCCM, serves as an independent content expert for the MAP, ensuring that a critical care perspective will be considered in the selection and approval of measures.

At the final HQA meeting, Carolyn Clancy, MD, Director of the Agency for Healthcare Research and Quality, reflected on the accomplishments of HQA. She recognized the group’s success in advancing public reporting for hospital care and promoting a national conversation about quality in a serious and informed fashion. She noted that while “much work still remains to be done and quality is still deficient,” the HQA has assured that the patients’ perspective and experience are now truly respected.

The Society has a significant, ongoing role to play in achieving the objectives of high-quality healthcare for all. It is my hope that all SCCM members will participate in our national effort to advance the quality and affordability of care for critically ill and injured patients.

During these last two years, it has also been my privilege to chair the Society’s newest committee: the Quality and Safety Committee.  In the future, I will offer comments on that experience and the work of the committee in advancing quality and safety for our vulnerable, critically ill patients, their families and communities.

Ivor S. Douglas, MD, served as an organizational principal to the Hospital Quality Alliance for the Society of Critical Care Medicine. He serves as chair of the Society’s Quality and Safety Committee.  He is Chief of the Division of Pulmonary and Critical Care Medicine and Director of Medical Intensive Care at the Denver Health Medical Center in Colorado. He is associate professor of medicine at the University of Colorado School of Medicine.

Rates of death or an inability to walk without assistance at 60-day follow-up were not statistically significantly different between the restrictive (35.2%) and liberal strategy group (34.7%, P=0.90).  There were no significant differences between groups regarding functional outcomes, including activities of daily living and fatigue. 

The results from this study are consistent with the Transfusion Requirements in Critical Care trial, although increased rates of myocardial infarction or congestive heart failure were not observed in the liberal strategy group.  It should be noted that before surgery, the restrictive strategy group had more transfusions than the liberal group, but this was not statistically significant (P=0.07), and the mean hemoglobin level was 11.3 in both groups before surgery.  Furthermore, the difference between a hemoglobin level of 8 vs. 10 may or may not be clinically significant. This well-designed and methodologically rigorous study provides additional evidence to support a restrictive transfusion strategy in surgical patients, even when cardiac risk factors are present. 

CarsonJL, et al. for the FOCUS Investigators.  Liberal or restrictive transfusion in high-risk patients after hip surgery.  N Engl J Med 2011; epub ahead of print. 

Concise Critical Appraisal is a regular feature authored by SCCM member Samuel M. Galvagno Jr., DO. Each installment highlights journal articles most relevant to the critical care practitioner.

This large observational trial demonstrated that omission of early thromboprophylaxis is associated with a statistically significantly higher risk of mortality in ICU patients.  As with all observational data sets, residual confounding, and other selection biases, might also explain this association; however, the authors appropriately employed robust statistical techniques to control for differences in baseline characteristics and known confounders.  This study adds compelling evidence in support of the recommendation by the National Quality Forum and the Joint Commission International to use thromboprophylaxis in ICU patients as soon as possible after admission. 

Concise Critical Appraisal is a regular feature authored by SCCM member Samuel M. Galvagno Jr., DO. Each installment highlights journal articles most relevant to the critical care practitioner.

A number of significant results were found in this large cohort study. Overall, new-onset AF was associated with a 7% increase in the adjusted risk of in-hospital death, and 14% of all hospital-associated new-onset AF occurred in the context of severe sepsis (odds ratio [OR] 6.82; 95% confidence interval [CI], 6.54-7.11; P<0.001). In patients with severe sepsis, 2.6% of patients who also had new-onset AF suffered an ischemic stroke, for an average stroke rate of 0.15% per hospital day for patients with new-onset AF. Those with severe sepsis and new-onset AF had a statistically significant greater stroke risk than those with preexisting AF (OR 3.63; 95% CI, 2.51-5.25; P<0.0001). Various risk factors associated with new-onset AF among patients with severe sepsis were identified.

This large observational study is replete with data, and the authors attempted to control for the known limitations involving ICD-9-CD coding. The new-onset AF rate (5.9%) was lower in this study than previously reported rates (6% to 20%), and this may be attributed to the nature of claims data. Furthermore, an immortal-time bias might have falsely lowered the measurable risk for mortality rates associated with new-onset AF. Despite the inherent limitations when using observation data, this study demonstrates significant stroke and mortality risks for patients with severe sepsis and new-onset AF. New-onset AF may be a marker for severity of illness and poor prognosis, or it may directly contribute to mortality. This observational study calls for future investigations to examine the mechanisms responsible for the poor outcomes associated with new-onset AF, as well as possible management options to mitigate the risk for developing AF during severe sepsis.

Concise Critical Appraisal is a regular feature authored by SCCM member Samuel M. Galvagno Jr., DO. Each installment highlights journal articles most relevant to the critical care practitioner.

Fourteen studies met inclusion criteria.  The studies were moderately heterogeneous (I2=38%), and 36.8% of all included patients came from two reports conducted in cardiac surgery ICUs. Overall, patients treated with chlorhexidine had a statistically significantly lower risk of VAP compared to control subjects (RR 0.72; 95% CI, 0.55-0.94, p=0.02).  A subgroup analysis that included studies using 2% chlorhexidine found a 47% lower risk of VAP compared to controls (RR 0.53; 95% CI, 0.31-0.91, p=0.02). The use of antiseptics appeared to have the greatest benefit in cardiac surgery patients (RR 0.41; 95% CI, 0.17-0.98, p=0.05).

This study was methodologically sound, although over a quarter of the data came from cardiac surgery ICUs.  This might have introduced bias since cardiac surgery patients are often electively admitted and intubated under more controlled conditions than trauma or other types of ICU patients. Only two studies evaluated the use of povidone iodine, and different concentrations of chlorhexidine (0.12-2%) were used. Despite the limitations of this review, this meta-analysis provides strong evidence that the use of antiseptics may be beneficial for ICU patients in the prevention of VAP.  The effect appears to be most pronounced in cardiac surgery patients and with 2% chlorhexidine.

Concise Critical Appraisal is a regular feature authored by SCCM member Samuel M. Galvagno Jr., DO. Each installment highlights journal articles most relevant to the critical care practitioner.

Significant reductions in the use of chloride-rich solutions were found after the washout period.  The incidence of severe hyperchloremia (2.3% vs. 6.2%, p<0.001) and hypernatremia (0.2% vs. 0.9%, p<0.001) were significantly lower in the intervention group; however, a significantly greater incidence of metabolic alkalosis and alkalemia was found in the intervention group (14.7% vs. 10.5%, p<0.001).    Fluid costs decreased from $15,077 to $3,915.  The time-weighted standard base excess increased from 0.5 (+/- 4.5) to 1.8 (+/- 4.7) in the intervention group.   The authors concluded that restriction of chloride-rich solutions was associated with a significant decease in the incidence of metabolic acidosis, hypernatremia, and severe hyperchloremia; however, the incidence of metabolic alkalosis was increased.

There are a few limitations worth noting.  Not all solutions used in this study are available in the U.S. For example, a gelatin colloid, Gelofusine was used in the control group, and this colloid is not widely used in North America (although it is similar to other colloidal preparations).  No exclusion criteria were applied, thus possibly improving the external validity of the study, but the biochemical effects of chloride restriction may not be the same for all patients.  The implications for the biochemical changes in this study remain unknown.  It is not clear if metabolic alkalosis, versus metabolic acidosis, leads to worse outcomes in the critically ill, and the authors plan to undertake future studies to help answer this question. 

Concise Critical Appraisal is a regular feature authored by SCCM member Samuel M. Galvagno Jr., DO. Each installment highlights journal articles most relevant to the critical care practitioner.

The experimental group was reasonably balanced against the control group, although there were some differences.  Patients in the experimental group had higher minute ventilation (p=0.04), greater fluid intake (p=0.09), and higher vasopressor use at the time of enrollment.  Multiple logistic regression was used to control for baseline shock and other mortality-predicting covariates.

The study was terminated early by the data safety monitoring board after an interim analysis indicated worse outcomes in the experimental group.  Patients in the experimental group had fewer ventilator-free days than the placebo group (14.0 [SD 11.1] vs. 17.2 [10.2], p=0.02).  The experimental group also had fewer ICU-free days (14.0 vs 16.7, p=0.04), and higher unadjusted mortality (26.6% vs 16.3 %, p=0.054).  When mortality was adjusted for baseline covariates, no statistically significant different was observed between groups (p=0.11). The experimental group had fewer organ failure-free days (12.3 vs 15.5, p=0.02).

The findings in this study are incongruent with prior work in this area, and this may be attributed to a lower overall ALI mortality (21.7%) than observed in previous n-3 fatty acid studies, a slight imbalance in baseline covariates that might have benefitted the control group, and the use of low tidal volumes in both study groups.  Additionally, the use of different enteral preparations and ventilator protocols in earlier studies might have biased results in favor of n-3 fatty acids.

The authors concluded that enteral supplementation of fatty acids, GLA, and antioxidants in patients with ALI does not improve outcomes and may be harmful.

Concise Critical Appraisal is a regular feature authored by SCCM member Samuel M. Galvagno Jr., DO. Each installment highlights journal articles most relevant to the critical care practitioner.

These results suggest that trainees from hospitals with high- intensity ventilation protocols do not demonstrate knowledge deficits regarding mechanical ventilation as compared to trainees from hospitals with low-intensity protocols.

Potential limitations to this work include the use of examination questions to test knowledge exclusively about mechanical ventilation and the fact that the study population consisted of examinees trained in internal medicine, not surgery, anesthesiology or emergency medicine. Nevertheless, this study has important implications for physician-educators, because the unintended consequences of protocols on education have not been evaluated broadly. Clinical protocols, which are designed to enhance and standardize patient care, may have an impact on medical education, especially in an era where work-hours are limited and care has become compartmentalized.

Concise Critical Appraisal is a regular feature authored by SCCM member Samuel M. Galvagno Jr., DO. Each installment highlights journal articles most relevant to the critical care practitioner.

A total of 396,380 patients were included, with 47% hospitalized and 16,474 (4%) with an ICU stay.  Over 97% of all patients in the three groups had a primary care physician visit within one year of discharge.   Among all patients with an ICU stay, the incidence of medication discontinuation ranged from 22.8% for antiplatelet/anticoagulant medications to 5.4% for respiratory inhalers. Patients with an ICU stay were 1.48 times more likely to have discontinuation of a statin (95% confidence interval [CI], 1.39-1.57) and 2.31 times more likely to have discontinuation of antiplatelet/anticoagulant medications (95% CI, 2.07-2.57). Compared to controls, the adjusted odds ratio for medication discontinuation after an ICU stay varied from 1.11 for statins (95% CI, 1.05-1.18) to 1.29 for levothyroxine (95% CI, 1.17-1.41). In a preplanned secondary outcome analysis, the risk of death, emergent hospitalization or emergency department visit was significantly increased when statins or antiplatelet/anticoagulant medications were discontinued following hospitalization. The authors concluded that treatment in an ICU places patients at greater risk for unintentional discontinuation of long-term medications.

The high incidence of medication discontinuation in this study may not be caused solely by acts of omission.  ICU or hospital admission may lead to deliberate discontinuation of chronic medications for a wide variety of reasons.  Nonetheless, the authors selected 5 drug classes that have been associated with adverse consequences when discontinued.  Different levels of care provided during the admission, varying ICU and hospital lengths of stay, and lack of other clinical details may also explain the findings in this study.

An accompanying editorial by Jeremy M. Kahn, MD, MS, and Derek C. Angus, MD, MPH, FCCM, notes that solutions to better administer the complexities of medication management are much needed, and this study highlights the potential threats to patient welfare during transition through different levels of care within the health system.

Concise Critical Appraisal is a regular feature authored by SCCM member Samuel M. Galvagno Jr., DO. Each installment highlights journal articles most relevant to the critical care practitioner.

Consecutive patients with confirmed or suspected infection, and two or more systemic inflammatory response syndrome criteria, were enrolled. Of the 291 patients assigned to one of two quantitative resuscitation protocols, the median time from triage to initial antibiotic administration was 115 minutes (interquartile range, 65-175).  A multivariate logistic regression model did not reveal any evidence for confounding since adjusted odds ratios were not significantly different from unadjusted odds ratios when multiple independent variables were evaluated. No association was found between in-hospital mortality rates when antibiotics were given within six hours of initial triage in the emergency department (ED). However, patients who received antibiotics after recognition of shock (n = 172) had a higher mortality rate (n = 119, odds ratio 2.35; 95% CI, 1.12-4.53).

The recommendation for administering antibiotics within one hour for severe sepsis and septic shock is based on expert opinion and a previous large retrospective study that demonstrated higher survival rates when antibiotics were given within one hour. In this most recent study, investigators were unable to show increased mortality rates for each hour’s delay to administration of antibiotics after emergency department triage; although, when antibiotics were delayed for patients in shock, mortality rates increased. These findings, which showed no difference in mortality rates when antibiotics were delayed after triage (but not shock), may be partly attributed to a relatively small sample size, the fact that each participating institution had robust resuscitation protocols, and a lower overall mortality rate for septic shock than described in previous studies. The findings in this study may also be credited to the effect of an early goal-directed resuscitation protocol, in which antibiotics are one of several crucial elements of current evidence-based support for patients with severe sepsis and septic shock.
Jing Tao, MD, senior resident in the Department of Anesthesiology at the University of Maryland, contributed to this installment of Concise Critical Appraisal.

Concise Critical Appraisal is a regular feature authored by SCCM member Samuel M. Galvagno Jr., DO. Each installment highlights journal articles most relevant to the critical care practitioner.

1. Dellinger RP, et al.  Surviving Sepsis Campaign: International guidelines for management of severe sepsis and septic shock: 2008.  Crit Care Med 2008; 36: 296-327.
2. Jones AE, et al.  Lactate clearance vs central venous oxygen saturation as goals of early sepsis therapy: A randomized clinical trial.  JAMA 2010; 303: 739-746.
3. Kumar A, et al.  Duration of hypotension before initiation of effective antimicrobial therapy is the critical determinant of survival in human septic shock.  Crit Care Med 2006; 34: 1589-1596.

Forty-two mechanically ventilated patients in a mixed medical-surgical ICU were prospectively enrolled in this blinded, non-randomized trial. Enrollment in the trial was triggered by the need for thoracic computed tomography (CT), which was used as the gold standard for all patients.  All patients had a CXR, CT, and ultrasound examination. The primary author performed all ultrasound exams and was blinded to the CT findings; the exams were not reviewed by a blinded radiologist.

Ultrasound had superior sensitivity and specificity for the detection of consolidation, pleural effusion, and interstitial edema when compared to CXR. A positive likelihood ratio of 14.29 was reported for detecting consolidation with ultrasound (100% sensitivity, 78% specificity); a likelihood ratio of 13.4 was reported for diagnosing interstitial edema (94% sensitivity, 93% specificity). Alternatively, CXR had a sensitivity of only 38% for consolidation and a sensitivity of 46% for interstitial edema. Ultrasound identified six of eight pneumothoraces with a sensitivity of 75%. None of the pneumothoraces were clinically significant. The authors concluded that lung ultrasound demonstrated superior diagnostic performance compared to CXR, and may be considered as an alternative to computed tomography (CT) in some instances.

Since patient selection was based on a predefined need for a CT, the study was subject to considerable verification bias.  Moreover, all exams were performed by one investigator, and none of the exams were confirmed by a blinded radiologist or second ultrasonographer.  All patients were positioned laterally for the exam, and this positioning might have changed the localization for some abnormalities, and may not always be feasible or safe for many ICU patients in other settings.  The case mix, plagued by the limitation of a small sample size, was heavily skewed towards trauma patients (n=11) and patients with sepsis (n=18), further limiting the generalizability across diverse ICU patient populations.

Notwithstanding the significant limitations, this work stands as yet another example of how ultrasound might be used in the ICU as a safer and cheaper alternative to other diagnostic modalities. Training, equipment acquisition and quality control remain significant concerns that must be addressed before ultrasound can be reliably used in place of an established “gold standard” such as chest CT.

Concise Critical Appraisal is a regular feature authored by SCCM member Samuel M. Galvagno Jr., DO. Each installment highlights journal articles most relevant to the critical care practitioner.

After a limited search involving PubMed, EMBASE, and the Cochrane registry, six randomized controlled trials were identified. These studies reported the type of arrest (in-hospital versus out-of-hospital), a comparison group (placebo or another drug) and survival. Primary outcomes were return of spontaneous circulation (ROSC), 30-day survival and functional neurological status as defined by a Glasgow-Pittsburgh Cerebral Performance Category (CPC) score of 1 or 2. The authors did not report so-called gray literature (i.e., research not found through conventional publishing channels), abstracts from recent conferences, references from included studies, or other bibliographic databases. Moreover, formal assessments for risk of bias were not described in the methods section of the paper. Validity criteria were unclear, but the authors reported assessing methodological quality in an eSupplement that was not available at the time of this appraisal.

Despite moderate to high degrees of heterogeneity (I2 46-71%), the authors completed a meta-analysis. Neither ROSC, long-term survival or favorable neurological outcomes were statistically significantly associated with the use of vasopressin. In a subgroup analysis of studies reporting long-term survival after asystole and administration of the drug within 20 minutes of arrest, vasopressin was associated with higher odds of survival (odds ratio [OR] 2.84; 95% confidence interval [CI] 1.19, 6.79; p=0.02). ROSC was also higher in two studies reporting the use of vasopressin in asystole when given within 20 minutes of arrest (OR 1.70; 95% CI 1.17, 2.47; p=0.005).

To date, none of the four core advanced cardiac life support drugs (lidocaine, amiodarone, epinephrine, vasopressin) have been shown to affect long-term survival, even though ROSC rates may be increased in both human and animal models. Although these drugs have not been shown to advance long-term outcomes, most experts believe there is little harm and potential short-term benefits. Notwithstanding the methodological limitations of this meta-analysis, the results support a possible role for vasopressin in asystole when the drug is given promptly as a one-time substitution for epinephrine. Whether there is enough equipoise to pursue this finding with additional placebo-controlled studies remains to be seen.

Concise Critical Appraisal is a regular feature authored by SCCM member Samuel M. Galvagno Jr., DO. Each installment highlights journal articles most relevant to the critical care practitioner.