Evidence lost to treatment of critically-ill patients? (original) (raw)

Evidence lost to treatment of critically-ill patients?

M. J. SCHULTZ 1,2,3{ }^{1,2,3}, P. E. SPRONK, 1,3,4{ }^{1,3,4}, B. AFESSA 5,6{ }^{5,6}, O. GAJIC 5,6{ }^{5,6}
1{ }^{1} Department of Intensive Care Medicine, Academic Medical Center, University of Amsterdam, Amsterdam, The Netherlands; 2{ }^{2} Laboratory of Experimental Intensive Care and Anesthesiology (L-E-I-C-A), Academic Medical Center, University of Amsterdam, Amsterdam, The Netherlands; 3{ }^{3} HERMES Critical Care Group, Amsterdam, The Netherlands; 4{ }^{4} Department of Intensive Care Medicine, Gelre Hospital - Location Lukas, Apeldoorn, The Netherlands; 5{ }^{5} Division of Pulmonary and Critical Care Medicine, Mayo Clinic, Rochester, MN, USA; 6{ }^{6} Mayo Epidemiology and Translational Research in Intensive Care (M-E-T-R-I-C), Mayo Clinic, Rochester, MN, USA

Abstract

Treatment strategies for critically-ill patients can and should never be excluded from grading processes that classify the evidence and provide decision support for health care workers involved in the care of these patients. Along with grading the available evidence, implementing new therapies and strategies in daily practice is another important but frequently forgotten step in improving care for critically-ill patients. Explanations for why some trials show benefit while other trials do not or even show harm include differences in the timing and the dose of the studied interventions, differences and heterogeneity of study populations and differences in trial protocols. Potential factors that may hamper the implementation of new therapies and strategies include translational problems, potentially biased expert opinions, concerns about side-effects and costs and problems with the recognition of critically-ill patients who might actually benefit from a new therapy or strategy. We discuss difficulties with grading the evidence for and the implementation of hang-protective mechanical ventilation in acute respiratory distress syndrome, glucocorticosteroid therapy in refractory septic shock, glucocorticosteroid therapy in acute respiratory distress syndrome, goal-directed fluid therapy in shock, activated protein C in severe sepsis and intensive insulin therapy in critical illness.

(Minerva Anestesiol 2009;75:715-29)
Key words: Evidence-based medicine - Health plan implementation - Shock, septic - Fluid therapy.

In recent years, the care of critically-ill patients has moved forward with several institutions reporting improved patient outcomes. These improvement trends have been largely attributed to the implementation of certain therapies and strategies tested in clinical trials. However, the evidence supporting the use of these therapies and strategies was often refuted by other, usually multicenter trials. Consequently, enthusiasm for implementation of the studied interventions has remained low, with understandable cynicism hampering the translation of these therapies and strategies to the daily care of critically-ill patients.

Evidence-based medicine attempts to accurately assess and integrate the weight carried by the
various levels of available evidence to certain aspects of medical practice. Specifically, evidence-based medicine seeks to apply judgments about the quality of evidence to those therapies and strategies that depend on rational assessments of benefits and risks. 1{ }^{1} Evidence-based medicine uses grading systems to classify the evidence and to provide decision support for health care providers. 1,2{ }^{1,2} The evidence from large, multicenter randomized clinical trials is ultimately assigned the highest grade, leading to strong recommendations. 1{ }^{1} Although some investigators have challenged this approach, 3,{ }^{3,} 4 intensive care societies have adopted a modified, evidence-based medicine grading system into recent practice guidelines. 2{ }^{2} Along with grading

the available evidence, implementing new therapies or strategies in daily practice is another challenge. 5−7{ }^{5-7} Unfortunately, the implementation part is often inadequately addressed, resulting in imperfect use of several new therapies or strategies in critically-ill patients. 8−11{ }^{8-11}

In this manuscript, we discuss a selection of therapies and strategies that may have the potential to improve the survival of critically-ill patients. We focus on therapies and strategies, the benefits of which are supported by some but not all trials. We critically appraise the differences between the trials, which may account for divergent study results. Are there alternative explanations why confirmatory trials (i.e., trials that were set up to support or establish the validity of previous trials) do not show beneficial effects, apart from the possibility that the studied intervention may indeed not benefit critically-ill patients? We then discuss potential factors that hamper the implementation of these therapies and strategies into daily intensive care practice. If intensive care physicians are not using the new therapies and strategies, what are the potential barriers to implementation? Finally, we propose solutions for better trials (i.e., trials without flaws, as outlined in this review) in criti-cally-ill patients, more appropriate grading systems and better knowledge translation into daily practice.

Therapies and strategies that may have the potential to benefit critically-ill patients

Several recently tested interventions in critical-ly-ill patients resulted in reduced mortality in one or more randomized controlled trials but had no effect on mortality in other trials. These interventions include the use of lung-protective mechanical ventilation in acute lung injury/acute respiratory distress syndrome (ALI/ARDS), glucocorticosteroid therapy in refractory septic shock, glucocorticosteroid therapy in ARDS, goal-directed therapy in shock, activated protein C in severe sepsis and intensive insulin therapy in critical illness. Details of clinical trial design, including the populations studied, the interventions tested, and the outcome data (merely mortality), are given in Table I. Rationales for these therapies and strategies are described below.

Lung-protective mechanical ventilation using lower tidal volumes

Increasing evidence from animal studies suggested that mechanical ventilation may aggravate or initiate pulmonary inflammation, a phenomenon frequently referred to as ventilatorinduced lung injury (VILI). Observational clinical studies reported dramatically improved outcomes with the use of low tidal volumes and permissive hypercapnia. 37{ }^{37} The hypothesis that alveolar overdistension may play a role in VILI brought several investigators to perform randomized controlled trials in which lung-protective mechanical ventilation using “lower” tidal volumes were compared with standard ventilator settings in patients with ALI/ARDS. 12−16{ }^{12-16} While 2 of these trials showed that a lower tidal volume strategy benefited patients with ALI/ARDS, 12,163{ }^{12,16} 3 trials were “negative”, showing no beneficial effects of the studied intervention. 13−15{ }^{13-15}

Glucocorticosteroid treatment in refractory shock

In severe illness, the hypothalamic-pituitaryadrenal axis is activated. Cortisol plays a key role in maintaining vascular tone and enhances the vasoconstrictive effects of catecholamines. Cortisol also plays a role in endothelial integrity and the regulation of fluid movements within the extracellular compartment. Furthermore, glucocorticosteroids modulate inflammation at all levels, protecting the host against an immune overreaction. Randomized controlled trials conducted more than 30 years ago found no benefit of administering supraphysiological doses of glucocorticosteroids for a short course in non-selected patients with sepsis. 38,39{ }^{38,39} Nevertheless, subsequent smaller studies suggested that a subgroup of patients with vasopressor-resistant septic shock might benefit from physiologic (replacement) dose of glucocorticosteroids administered for presumed relative adrenal insufficiency. 40−43{ }^{40-43} Two recent randomized controlled trials studied the effect of glucocorticosteroid treatment in patients with septic shock. 17{ }^{17} 18{ }^{18} One trial showed the mortality benefit of glucocorticosteroid therapy, 17{ }^{17} the other trial was reported to be negative, showing no mortality reduction. 18{ }^{18}

TABLE I.- Randomized controlled trials in intensive care medicine with mortality as one of the main endpoints.

(To be continued)

Table I.- Randomized controlled trials in intensive care medicine with mortality as one of the main endpoints. (Continues).

(To be continued)

TABLE I.- Randomized controlled trials in intensive care medicine with mortality as one of the main endpoints. (Continues).

VT\mathrm{V}_{\mathrm{T}} : tidal volume; ALL/ABDSA L L / A B D S : acute lung injury/acute respiratory distress syndrome; ICU: intensive care unit; APACHE: acute physiology and chronic health evaluation; PaCO2\mathrm{PaCO}_{2} : arterial carbon dioxide pressure. a{ }^{a} see text for details.

Glucocorticosteroid treatment in ARDS

ARDS is characterized by diffuse alveolar damage and excessive pulmonary inflammation and fibroproliferation. Observational studies in the past suggested that glucocorticosteroid therapy may be useful in the management of ARDS. Three randomized controlled trials tested the effect of glucocorticosteroid treatment in ARDS. 19−21{ }^{19-21} While 2 single-center trials showed mortality reduction
with glucocorticosteroid treatment, 19,211{ }^{19,21} 1 multicenter trial showed no beneficial effects of glucocorticosteroid therapy with regard to mortality. 20{ }^{20}

Goal-directed hemodynamic therapy in shock

Prolonged tissue hypoxia resulting from circulatory failure or systemic inflammatory response syndrome is an important determinant of multiple organ dysfunction syndrome and is associated

with increased morbidity and mortality. Preclinical studies suggested that interventions aiming at the correction of the imbalance between oxygen demand and oxygen delivery, so-called goal-directed hemodynamic therapy, would improve outcome. Several randomized controlled trials tested this strategy in critically-ill patients. 22−26{ }^{22-26} Three of these trials showed a strong beneficial effect on survival of this intervention, 22,25,26{ }^{22,25,26} while 2 other trials showed no beneficial effect. 23,24{ }^{23,24}

Activated protein C in severe sepsis

Coagulopathy is intrinsic to sepsis and is characterized by activation of the coagulation pathway and simultaneous attenuation of fibrinolysis. 44{ }^{44} In addition, levels of the endogenous anticoagulants, activated protein CC, antithrombin and tissue factor pathway inhibitor decline due to both decreased production and increased consumption. 44{ }^{44} Intravascular fibrin formation is thought to perpetuate diffuse endothelial damage in multiple organs, contributing to multiple organ dysfunction and eventually death. Three recent randomized controlled trials studied the effect of activated protein C infusion in sepsis. 27−29{ }^{27-29} One trial reported significant mortality reduction with the use activated protein C in patients with sepsis, 27{ }^{27} and two subsequent trials were negative, showing no mortality reduction. 28,29{ }^{28,29} In response to a request from the European Medicines Agency, a new placebo-controlled trial of activated protein C for persistent septic shock has been initiated recently. 45{ }^{45} The results of this trial, however, are to be awaited.

Intensive insulin therapy in critical illness

Critical illness-induced hyperglycemia is associated with adverse outcome. 46{ }^{46} Hyperglycemia has been suggested to be acutely toxic in critically-ill patients because of accentuated cellular glucose overload and the pronounced toxic side effects of glycolysis and oxidative phosphorylation. 47{ }^{47} Intensive insulin therapy, aiming at blood glucose levels of 4.5-6.1 mmol/L, decreased the mortality and morbidity of critically-ill patients in 2 monocenter randomized controlled trials. 30,31{ }^{30,31} Noteworthy, with more “strict” intensive insulin therapy (i.e., blood glucose levels close to normal
levels) more benefit was achieved when compared to less strict regulation. 48,49{ }^{48,49} However, 2 successive monocenter randomized controlled trials, 32,33{ }^{32,33} and 3 multicenter randomized controlled trials 34−36{ }^{34-36} failed to show a benefit of intensive insulin therapy. In fact, one of these trials even indicated that intensive insulin therapy could cause harm. 36{ }^{36}

Potential reasons for discrepant results and the confusion about tested interventions

What are alternative explanations for why some trials did show beneficial effects while other trials on a similar intervention did not, apart from the possibility that the studied therapy or strategy may indeed not benefit critically-ill patients? Differences in the timing and the dose of the intervention, differences and heterogeneity of study populations and differences in trial protocols may all account for observed discrepancies in the results (Table II). Furthermore, although not showing the superiority of the tested intervention with respect to mortality, some trials actually did show beneficial effects, challenging the crudeness of (short-term) mortality as the sole outcome measure in trials of critically-ill patients.

Timing of the intervention

There is little doubt that the timing of an intervention is of critical importance in emergency and critical care medicine. Overwhelming data and clinical experience support the concept of the “golden hour” in trauma, coronary or cerebral revascularization, or surgery for ischemic bowel. For example, early thrombolytic therapy (i.e., within 3 hours) saves the lives of patients with acute myocardial infarction but has no effect or may even be harmful if instituted late (i.e., after 24 hours). Similarly, early optimization of oxygen delivery is crucial for the treatment of shock. 50{ }^{50} Studies with early goal-directed hemodynamic therapy (i.e., 8-12 hours postoperatively or before the development of multiple organ failure) show better results than those with late goal directed therapy (i.e., >12>12 hours or after the onset of multiple organ failure). Indeed, 2 recent trials that applied early goal-directed hemodynamic therapy demonstrated dramatic effects on mortality in

TABLE II.- Potential reasons for the diversity in outcome of recent studies in intensive care medicine.

[1]septic shock, although the actual goals were different. 25,26{ }^{25,26} The discrepancy between the results of these 2 studies and previous trials that applied the intervention late in the course of illness 23,24{ }^{23,24} suggests that the timing of goal-directed hemodynamic therapy (early, versus [too] late) may be more important than the actual goals or the means to achieve those goals.

1. See text for details ↩︎

Differences in the timing or in the intervention may also in part explain discrepant results in the trials of glucocorticosteroid therapy in sepsis and glucocorticosteroid therapy in ARDS, as well as discrepant results in the trials of intensive insulin therapy. Indeed, in the last trial of intensive insulin therapy, normoglycemia was reached rather late. 36{ }^{36} When the time until the target is so long, the time window for the prevention of toxicity from hyperglycemia may have passed, and irreversible damage may already have occurred. 47,51{ }^{47,51}

Enrollment criteria and differences in patient populations

The complexity of critical illness and the heterogeneity of syndromes studied in critical care medicine maximally challenge the appropriate conduct of clinical trials. Difficulties in enrollment often lead to the inclusion of less severely ill patients in whom any intervention is unlikely to provide benefit. This point is illustrated in the example of the use of glucocorticosteroids in septic shock. The inclusion criteria in the first trial required the presence of both hypotension and vasopressor use (i.e., vasopressor-resistant hypotension); 17{ }^{17} in the second trial, enrollment criteria were relaxed, requiring either hypotension or vasopressor use (i.e., any patient with septic shock). 18{ }^{18} While the first trial demonstrated an outcome benefit of steroid replacement in vasopressorresistant septic shock, the confirmatory trial showed no outcome benefit of steroid use in nonselected patients with septic shock. Poor enrollment into the second study suggests that the bedside clinicians were reluctant to enroll the most severely ill patients in the confirmatory trial as they were “biased” of the beneficial effects of hydrocortisone therapy in the most severely ill with vasopressor-resistant hypotension. The difference in mortality in the control arms of the 2 studies supports this hypothesis.

The second and third randomized controlled trial of activated protein C in sepsis 28,29{ }^{28,29} have often been used in the discussion of the efficacy of this therapy. However, neither of these trials can be seen as confirmatory or “negative” trials for reasons similar to the ones stated above. Study populations were different from the population in the first randomized controlled trial, as reflected by the lower
severity of illness scores and much lower mortality rates in the control arms of the trials. 27,29{ }^{27,29}
“Dilutional” effects of the enrollment of less severely ill patients are also likely to at least in part explain the differences in the results of clinical trials of goal-directed therapy in shock, again reflected by heterogeneity in the inclusion criteria and low mortality in the control arms of the study.

Dosing of the intervention and differences in trial protocols

In addition to the timing of the intervention and the studied patient populations, the dose of an interventional drug or a procedure is of critical importance for its effectiveness. The difference between the 2 positive and 3 negative trials on the effects of lower tidal volume ventilation in ARDS may in part be related to different dosing regiments. While the size of tidal volumes in the control-arms of the 2 positive trials were twice as large as in the lower tidal volume-arms ( 12 versus 6 mL/kg), 12,16{ }^{12,16} differences in the three negative trials were much smaller ( 10 versus 7 mL/kg,13117 \mathrm{~mL} / \mathrm{kg},{ }^{13} 11 versus 7 mL/kg,147 \mathrm{~mL} / \mathrm{kg},{ }^{14} and 10 versus 7 mL/kg7 \mathrm{~mL} / \mathrm{kg} ). 15{ }^{15} The differences in tidal volume size in the negative trials simply might have been too modest to reveal a mortality difference within the sample size limitations of these trials.

Differences in trial protocols may also account for divergent results of trials on glucocorticosteroid therapy in ARDS. While glucocorticosteroid therapy was tapered slowly in the intervention arms of 2 positive trials, 19,21{ }^{19,21} tapering was performed rather quickly in the negative trial. 20{ }^{20} In the latter trial, the pulmonary condition improved with the use of glucocorticosteroids leading to earlier discontinuation of mechanical ventilation. However, many patients required reinstitution of mechanical ventilation soon after the rapid tapering of glucocorticosteroid therapy. The most common reason for resuming mechanical ventilation in this later trial was shock, suggesting that patients on glucocorticosteroid therapy might have developed adrenal insufficiency due to the quick tapering of glucocorticosteroids, thereby potentially masking the beneficial effects of glucocorticosteroids for ARDS.

Different dosing regimens may also account for discrepancies in the results of the clinical trials of glucocorticosteroid therapy in refractory shock

(pharmacological versus replacement doses, use of fludrocortisone 17,18{ }^{17,18} ), goal-directed therapy in shock (very high doses of dobutamine in some trials), 23{ }^{23} and intensive insulin therapy. Indeed, the blood glucose levels achieved in the intervention groups of the negative trials were much higher than in the original 2 positive trials. 52{ }^{52} At the same time, blood glucose levels in the control groups of the negative trials were much lower. Therefore, the negative studies were essentially different from the first studies, as the negative studies were executed in the flat part of the observational blood glucose level-mortality risk curve. 46{ }^{46}

Premature discontinuation of clinical trials before reaching a pre-specified sample size may also account for differences in study results. Unfortunately, this decision may leave us with an underpowered trial. This problem exists with the last trial on glucocorticosteroid therapy for septic shock 18{ }^{18} and two multicenter trial on intensive insulin therapy. 34,35{ }^{34,35} These studies were all stopped prematurely and, therefore, lack adequate power for the conclusion that glucocorticosteroid therapy or intensive insulin therapy does not improve survival.

Crudeness of mortality as an outcome measure

Advances in techniques for prolonged organ support, the influence of patient and family preferences and varying ICU and hospital discharge practices challenge short-term mortality as the sine qua non outcome of critical illness. Outcomes that combine mortality and morbidity have long been adopted in oncology and cardiology and have recently been proposed as primary outcomes of clinical trials in intensive care medicine.

Examples are found with the trials on glucocorticosteroid therapy in refractory shock, 18{ }^{18} glucocorticosteroid therapy in ARDS, 20{ }^{20} and intensive insulin therapy in critical illness, 31{ }^{31} in which no significant reduction of in-hospital mortality was found but all led to improved morbidity. Although these trials are frequently referred to as “negative” trials, time to shock reversal was shorter, 18{ }^{18} the number of ventilator-free days was significantly larger, 20{ }^{20} the incidence of newly acquired kidney injury was lower, and length of stay in intensive care unit and hospital was shorter, 31{ }^{31} as compared to patients not receiving the studied intervention.

Barriers to implementation of new therapies and strategies

The disconnect between the evidence and practice is best illustrated in the recent multicenter study from Spain, 53{ }^{53} showing that only a fraction of patients with septic shock received all appropriate interventions for severe sepsis as judged by the application of surviving sepsis campaign guidelines. 54{ }^{54} In addition to the available “confusing evidence”, as stated above, several factors may hamper implementation of interventions that have the potential to benefit critical-ly-ill patients (Table III). Common factors associated with failure of implementing new interventions in intensive care practice include simple translational problems, potentially biased expert opinions, concerns about possible side effects, costs associated with implementation, and finally problems with the recognition of patients who might actually benefit from a new therapy or strategy.

Simple translational problems

Frequently, there is a disconnect between what providers think is done and what is actually done in their institutions. In the absence of close monitoring and feedback, providers often believe that their practice closely follows recommended guidelines. Accurate translation of what was actually done in a trial into what should also be done in daily practice may not occur, however. For instance, while it was clearly stated that tidal volume was to be titrated in patients based on predicted body weight, as a function of height, it was found that instead of using predicted body weight, physicians used actual body weight for the calculation of tidal volume. 55{ }^{55} Only after education on this simple but important rule, tidal volume size declined.

Similarly, while intensive insulin therapy may seem an easy to implement strategy, several aspects of this strategy that might be important are frequently overlooked, including the correct route for insulin administration, types of infusion-pumps used and nutritional strategies aiming for nonfasting states of critically-ill patients at all times, as well as correct sampling sites, accuracies of glucometers, and the levels of expertise of nurses who delicately and exclusively titrate insulin. 56{ }^{56}

Table III.- Barriers to implementation of new therapies and strategies in intensive care practice.

See test for details

Expert opinions, potentially leading to incorrect advice in guidelines
What is advised in guidelines is not always what has been studied or demonstrated. Indeed, it has been argued that certain trials tested a new therapy or strategy against a therapy or strategy that was already known “to harm” critically-ill patients. As such, it was suggested that the conventional ventilation strategy, using tidal volumes of 12 ml/kg, was no longer used in clinical practice. 57{ }^{57} Although this suggestion was untrue (at that time, average tidal volumes were indeed as large as 12 mL/kg\mathrm{mL} / \mathrm{kg} ), the impression arose that we should not comply with the new ventilation strategy using lower tidal volumes.

The new recommendations for septic patients advised to maintain the blood glucose level <8.3<8.3 mmol/L\mathrm{mmol} / \mathrm{L} instead of adhering to the more strict

thresholds as used in the 2 original randomized controlled trials. 54{ }^{54} However, there are just no trials that provide evidence for intensive insulin therapy in sepsis using the higher threshold. When advising to use the 8.3mmol/L8.3 \mathrm{mmol} / \mathrm{L}, upper limit, more septic patients will be in the higher blood glucose level-range, while studies showed more benefit is achieved with more strict intensive insulin therapy. 48,49{ }^{48,49}

Concerns about possible side effects

One reason for not adopting lung-protective mechanical ventilation was the concern that its use would necessitate and/or increase prescription of sedatives and narcotics because of patient intolerance of lower tidal volumes and increased respiratory rate. 8{ }^{8} This, however, was not found in 2 secondary analysis of the ARDS Network trial, 58,{ }^{58,} 59{ }^{59} and one retrospective analysis. 60{ }^{60} These studies clearly showed that lung-protective mechanical ventilation using lower tidal volumes does not increase sedative/narcotic medication requirements in mechanically ventilated patients.

One side effect with the use glucocorticosteroids that is frequently referred to is the risk of infections and gastrointestinal bleeding. Indeed, there were more episodes of super-infection in patients treated with glucocorticosteroids for refractory shock compared to control patients in one trial, 18{ }^{18} but this was not found in the preceding study. 17{ }^{17} No differences in the incidence of gastrointestinal bleeding with the use glucocorticosteroids were found in the 2 trials. 17,18{ }^{17,18}

Infusion of activated protein C inevitably comes with a risk of bleeding. 27−29{ }^{27-29} Indeed, the incidence of serious bleeding was higher in the activated protein C group than in the control group ( 4%4 \% vs. 2%).272 \%) .{ }^{27} This concern, however, may not be completely understandable in view of the overall benefit of the treatment and the relatively low incidence of clinically significant major bleeding, such as intracerebral bleeding.

With implementation of intensive insulin therapy, the incidence of severe hypoglycemia increases. Reported incidences of severe hypoglycemia (blood glucose level <2.2mmol/L<2.2 \mathrm{mmol} / \mathrm{L} ) rose by 5 - to 10 fold compared to conventional blood glucose control in randomized controlled trials. 30−36{ }^{30-36} However, the frequently assumed association between severe
hypoglycemia and poor outcome has been challenged recently. 61{ }^{61} Recently experimental data on the effects of severe hypoglycemia on the brain showed bolus glucose reperfusion of the depleted brain to cause more damage than the period of severe hypoglycemia itself. 62{ }^{62}

Costs associated with implementation

Another argument not to comply with the evidence might be the costs associated with the implementation and performance of a new strategy. Indeed, costs associated with infusion of activated protein C are considerable. 63,64{ }^{63,64} Prescription rates of activated protein C in European countries are not associated with the incidence of sepsis. 10{ }^{10} However, there seems to be some correlation between the percentage of activated protein C treated patients and the potential for full reimbursement, although this may not completely explain the difference between countries.

Problems with the recognition of patients who might actually benefit from a new therapy or strategy

One of the most important barriers to the implementation of a low tidal volume strategy was the under-recognition of ALI/ARDS. 9{ }^{9} Similarly, the under-recognition of shock syndrome prior to the development of severe hypotension may hamper the timely institution of life-saving therapies. As discussed above, the timing of oxygen delivery optimization is crucial for its beneficial effects. 50{ }^{50} Indeed, the latest trial on hemodynamic therapy showed that early optimization of oxygen delivery was able to decrease mortality in patients with septic shock. 25{ }^{25} However, the location of patients in the early phase of their critical illness varies among institutions. Although, most critically-ill patients should be treated in the intensive care unit, a good number of them may spend a significant portion of this “golden” time in the emergency department or the hospital ward. Obviously, it should be the department where the patient is admitted where early goal directed therapy should be applied. Of note, in Dutch hospitals it was found that patients like those studied in the latest trial on hemodynamic therapy hardly exist in the intensive care unit setting. 65{ }^{65}

Individual therapies and strategies versus bundles of care

Critically-ill patients often have more than one (life-threatening) problem, and frequently these problems are causally related. For example, patients with ALI/ARDS often also suffer from shock or sepsis, and patients with sepsis seldom show signs of coagulopathy. While most therapies or strategies aim at resolving (only) one particular problem, at the same time it may benefit critically-ill patients by preventing or resolving other problems. Indeed, appropriate and early shock treatment may prevent development of ALI/ARDS, and the infusion of activated protein C may benefit patients suffering from severe sepsis but may also benefit patients with ALI/ARDS. While all trials discussed above tested only one single therapy or strategy, in daily practice many strategies and therapies are started next to each other, if not at the same time. We may never know which strategy or therapy benefits a critically-ill patient more (or most) because all trials tested a strategy or therapy against standard care, which (possibly) also involved other already implemented therapies and strategies. Therefore, we should think more of implementing bundles of care 54{ }^{54} instead of trying to implement only one single strategy or therapy.

Proposed solutions and future considerations

For better trials

When considering a good clinical trial design, several issues need to be addressed. Clear equipoise has to exist to prevent potential harm to vulnerable critical care population. The inclusion/exclusion criteria and the characteristics of the study subjects should be well-defined so that the patients with the highest likelihood of benefit from the proposed intervention are enrolled. In critical care medicine, the timing of the intervention is often critical, and the investigators should refrain from enrolling patients beyond the window of time when the proposed intervention has the highest chance of benefit. In order for the study results to be generalizable, the trial should be multicenter and multinational, and the design should be adequately powered for the outcome of interest.

However, internal validity must not be compromised. Both the study and non-study treatment modalities should be similar in all study sites. The trial should be performed within a certain period of time because of rapid changes in the standard of care.

The primary and secondary outcomes of the study should be defined a priori to avoid misinterpretation. Few of the randomized controlled trials conducted in intensive care units and using mortality as a primary outcome show a beneficial impact of the intervention on the survival of crit-ically-ill patients. 66{ }^{66} Indeed, of 72 randomized controlled trials (publications of adult, multicenter randomized controlled trials carried out in the intensive care unit, with mortality as a primary outcome, and including >50>50 patients), 10 reported a positive impact, 7 reported a detrimental effect, and 55 studies showed no effect on mortality. Methodological limitations of some of the randomized controlled trials may have prevented positive results, and some beneficial effects may have been hidden within significant heterogeneity in the patient population.

More importantly, causes of death are complex in critically-ill patients and often include subjective factors such as patient/family and provider preferences for prolonged life support interventions. Rather than restricted to (short-term) mortality, the outcomes should include a broad spectrum of patient-centered issues including but not limited to morbidity (organ failure-free days, symptoms and complications), patient/family satisfaction, costs and quality of life. Computerized simulation models should evaluate potential outcomes in “the best” and “the worst” case scenarios and may greatly help in the planning stage.

Finally, it may not be a good idea to stop a trial prematurely. While at times it may be imperative to stop a trial because of too many or unacceptable side effects, early termination of a trial because of an unexpectedly large effect of a certain strategy or therapy is only allowed when predefined. Trials are performed to test a certain hypothesis, and beforehand how many patients are needed to reject that hypothesis is calculated. Early termination always leaves us with an underpowered study, which cannot be used in the discussion of the potential benefit of a certain strategy.

For more appropriate grading

In intensive care medicine, the “Grades of Recommendation, Assessment, Development and Evaluation (GRADE) criteria” 2{ }^{2} have been used recently for evaluating therapies and strategies for sepsis. 54{ }^{54} The GRADE-system is based on a sequential assessment of the quality of evidence, followed by an appraisal of the balance between benefit, burden and costs. The GRADE-system further classifies recommendations as “strong” or “weak”, which is considered of greater clinical importance than the quality of evidence level. While the GRADE-system is a step in the right direction towards more meaningful practice guidelines, future recommendations need to place more emphasis on the issues of internal and external validity, including the population studied, timing and dosing of the intervention

For better translation into practice

While there is no “magic bullet” for successful implementation of new strategies in daily intensive care unit practice, several strategies have been shown to facilitate translation of best practices to the bedside. Education efforts have to be accompanied with regular feedback to overcome clinical inertia and raise the awareness of bedside providers that their practice may not be as great as they may think. 55{ }^{55} The advances in medical informatics and electronic medical records allow for development of sophisticated decision support tools to facilitate both early recognition of critical care syndromes 67{ }^{67} and the timely application of best practices. 68{ }^{68} Finally, locally developed protocols that involve not only physicians but all bedside providers have been shown to be essential tools for knowledge translation in multiple studies. 69−{ }^{69-} 71{ }^{71} The culture of quality improvement needs to be pushed towards the ultimate goal of the development of multidisciplinary patient-centric rather than provider-centric models of critical care delivery.

Conclusions

We discussed several therapies and strategies that may have the potential to improve the sur-
vival of critically-ill patients. There are numerous alternative explanations of why confirmatory trials do not show beneficial effects while initial trials are “positive” and many potential factors that may hamper the implementation of the discussed interventions. This manuscript deals with only a small selection of therapies and strategies in crit-ically-ill patients. Similar and other problems with grading and implementation certainly exist for several other therapies and strategies for criticallyill patients.

References

1. Evidence-based medicine. A new approach to teaching the practice of medicine. Evidence-Based Medicine Working Group: JAMA 1992;268:2420-5.
2. Atkins D, Best D, Briss PA, Eccles M, Falck-Ytter V, Flottorp S et al. Grading quality of evidence and strength of recommendations. BMJ 2004;328:1490.
3. Grasselli G, Gattinoni L, Kavanagh B, Latini R, Laupacis A, Lemaire F et al. Feasibility, limits and problems of clinical studies in Intensive Care Unit. Minerva Anestesiol 2007;73:595-601.
4. Tobin MJ. Counterpoint: evidence-based medicine lacks a sound scientific base. Chest 2008;133:1071-4; discussion 1074-1077.
5. Groš R. Personal paper. Beliefs and evidence in changing clinical practice. BMJ 1997;315:418-21.
6. Groš R, Grimshaw J. From best evidence to best practice: effective implementation of change in patients’ care. Lancet 2003;362:1225-30.
7. Cook DJ, Montori VM, McMullin IP, Finfer SR, Rocker GM. Improving patients’ safety locally: changing clinician behaviour. Lancet 2004;363:1224-30.
8. Castro R, Regastra T, Aguirre ML, Llanos OP, Bruhn A, Bugedo G et al. An evidence-based resuscitation algorithm applied from the emergency room to the ICU improves survival of severe septic shock. Minerva Anestesiol 2008;74:22331.
9. Rubenfeld GD, Cooper C, Carter G, Thompson BT, Hudson LD. Barriers to providing lung-protective ventilation to patients with acute lung injury. Crit Care Med 2004;32:128993.
10. Schultz MJ, Levi M. Prescription of rh-APC differs substantially among western European countries. Intensive Care Med 2006;32:630-1.
11. Schultz MJ, Royakkers AA, Levi M, Moenisalan HS. Spronk PE. Intensive Insulin Therapy in Intensive Care: An Example of the Struggle to Implement Evidence-Based Medicine. PLoS Med 2006;3:e456.
12. Amato MB, Barbas CS, Medeiros DM, Magaldi RB, Schettino GP, Lorenzi-Filho G et al. Effect of a protective-ventilation strategy on mortality in the acute respiratory distress syndrome. N Engl J Med 1998;338:347-54.
13. Brochard L, Roudot-Thoraval F, Roupie E, Delclaux C, Chastre J, Fernandez-Mondejar E et al. Tidal volume reduction for prevention of ventilator-induced lung injury in acute respiratory distress syndrome. The Multicentre Trail Group on Tidal Volume reduction in ARDS. Am J Respir Crit Care Med 1998;158:1831-8.
14. Stewart TE, Meade MO, Cook DJ, Granton JT, Hodder RV, Lapinsky SE et al. Evaluation of a ventilation strategy to prevent barotrauma in patients at high risk for acute respirato-

ry distress syndrome. Pressure- and Volume-Limited Ventilation Strategy Group. N Engl J Med 1998;33:355-61.
15. Brower RG, Shanholtz CB, Fessler HE, Shade DM, White P, Jr., Wiener CM et al. Prospective, randomized, controlled clinical trial competing traditional versus reduced tidal volume ventilation in acute respiratory distress syndrome patients. Crit Care Med 1999;27:1492-8.
16. Ventilation with Lower Tidal Volumes as Compared with Traditional Tidal Volumes for Acute Lung Injury and the Acute Respiratory Distress Syndrome. N Engl J Med 2000;342:1301-8.
17. Annane D, Sebille V, Charpentier C, Bollaert PE, Francois B, Korach JM et al. Effect of treatment with low doses of hydrocortisone and fludrocortisone on mortality in patients with septic shock. JAMA 2002;288:862-71.
18. Sprung CL, Annane D, Keh D, Moreno R, Singer M, Freivegel K et al. Hydrocortisone therapy for patients with septic shock. N Engl J Med 2008;358:111-24.
19. Meduri GU, Headley AS, Golden E, Carson SJ, Umberger RA, Kelso T et al. Effect of prolonged methylprednisolone therapy in unresolving acute respiratory distress syndrome: a randomized controlled trial. JAMA 1998;280:159-65.
20. Steinberg KP, Hudson LD, Goodman RB, Hough CL, Lanken PN, Hyzy R et al. Efficacy and safety of corticosteroids for persistent acute respiratory distress syndrome. N Engl J Med 2006;354:1671-84.
21. Meduri GU, Goldiey A, Freire AX, Taylor E, Zaman M, Carson SJ et al. Methylprednisolone infusion in early severe ARDS: results of a randomized controlled trial. Chest 2007;131:954-63.
22. Shoemaker WC, Appel PL, Kram HB, Waxman K, Lee TS. Prospective trial of supranormal values of survivors as therapeutic goals in high-risk surgical patients. Chest 1988;94:117686 .
23. Hayes MA, Timmins AC, Yau EH, Palazzo M, Hinds CJ, Watson D. Elevation of systemic oxygen delivery in the treatment of critically ill patients. N Engl J Med 1994;330:171722 .
24. Gattinoni L, Brazzi L, Pelosi P, Latini R, Tognoni G, Pesenti A et al. A trial of goal-oriented hemodynamic therapy in critically ill patients. SvO2 Collaborative Group. N Engl J Med 1995;333:1025-32.
25. Rivers E, Nguyen B, Haestad S, Ressler J, Muzzin A, Knoblich B et al. Early goal-directed therapy in the treatment of severe sepsis and septic shock. N Engl J Med 2001;345:1368-77.
26. Lin SM, Huang CD, Lin HC, Liu CY, Wang CH, Kuo HP. A modified goal-directed protocol improves clinical outcomes in intensive care unit patients with septic shock: a randomized controlled trial. Shock 2006;26:551-7.
27. Bernard GR, Vincent JL, Laterne PF, LaRosa SP, Dhainaut JF, Lopez-Rodriguez A et al. Efficacy and safety of recombinant human activated protein C for severe sepsis. N Engl J Med 2001;344:699-709.
28. Abraham E, Laterne PF, Garg R, Levy H, Talwar D, Trasekoma BL et al. Drotrecogin alfa (activated) for adults with severe sepsis and a low risk of death. N Engl J Med 2005;353:133241.
29. Nadel S, Goldstein B, Williams MD, Dalton H, Peters M, Macias WL et al. Drotrecogin alfa (activated) in children with severe sepsis: a multicenter, phase III randomised controlled trial. Lancet 2007;369:836-43.
30. van den Berghe G, Wouters P, Weekers F, Verwaest C, Bruynincks F, Schetz M et al. Intensive insulin therapy in the critically ill patients. N Engl J Med 2001;345:1359-67.
31. Van den Berghe G, Wilmer A, Hermans G, Meersseman W, Wouters PJ, Milants I et al. Intensive insulin therapy in the medical ICU. N Engl J Med 2006;354:449-61.
32. Arabi YM, Dabbagh OC, Tamim HM, Al-Shimemei AA, Memish ZA, Haddad SH et al. Intensive versus conventional insulin therapy: a randomized controlled trial in medical and surgical critically ill patients. Crit Care Med 2008;36:3190-7.
33. De La Rosa GDC, Donado JH, Restrepo AH, Quintero AM, Gonzalez LG, Saldarriaga NE et al. Strict glycaemic control in patients hospitalised in a mixed medical and surgical intensive care unit: a randomised clinical trial. Crit Care 2008;12:R120.
34. Brunkhorst FM, Engel C, Bloos F, Meier-Hellmann A, Ragaller M, Weiler N et al. Intensive insulin therapy and pentastarch resuscitation in severe sepsis. N Engl J Med 2008;358:125-39.
35. Preiser JC, Devos P, Ruiz-Santana S, Melot C, Annane D, Groenewald J et al. A prospective randomised multi-centre controlled trial on tight glucose control by intensive insulin therapy in adult intensive care units: the Glucontrol study. Intensive Care Med 2009;35:1738-48.
36. Finfer S, Chittock DR, Su SY, Blair D, Foster D, Dhingra V et al. Intensive versus conventional glucose control in critically ill patients. N Engl J Med 2009;360:1283-97.
37. Hickling KG, Henderson SJ, Jackson R: Low mortality associated with low volume pressure limited ventilation with permissive hypercapnia in severe adult respiratory distress syndrome. Intensive Care Med 1990;16:372-7.
38. Cronin L, Cook DJ, Carter J, Heyland DK, King D, Lansang MAD et al. Corticosteroid for sepsis: A critical appraisal and meta-analysis of the literature. Crit Care Med 1995;23:14309 .
39. Lefering R, Neugebauer EAM. Steroid controversy in sepsis and septic shock: a meta-analysis. Crit Care Med 1995;23:1294-303.
40. Bollaert PE, Charpentier C, Levy B, Debouverie M, Audibert G, Larcan A. Reversal of late septic shock with supraphysiological doses of hydrocortisone. Crit Care Med 1998;26:64550 .
41. Bollaert PE: Stress doses of glucocorticoids in catecholamine dependency: a new therapy for a new syndrome? Intensive Care Med 2000;26:3-5.
42. Briegel J, Forst H, Haller M, Schelling G, Kilger E, Kuprat G et al. Stress doses of hydrocortisone reverse hyperdynamic septic shock: a prospective, randomized, double-blind sin-gle-center study. Crit Care Med 1999;27:723-32.
43. Chawla K, Kupfer Y, Goldman I, Tessler S. Hydrocortisone reverses refractory septic shock (abstract). Crit Care Med 1999;27:A33.
44. Levi M, Ten Cate H: Disseminated intravascular coagulation. N Engl J Med 1999;341:586-92.
45. Finfer S, Ranieri VM, Thompson BT, Barie PS, Dhainaut JF, Douglas JS et al. Design, conduct, analysis and reporting of a multi-national placebo-controlled trial of activated protein C for persistent septic shock. Intensive Care Med 2008;34:1935-47.
46. Bagshaw SM, Egl M, George C, Bellomo R. Early blood glucose control and mortality in critically ill patients in Australia. Crit Care Med 2009;37:463-70.
47. Van den Berghe G, Wilmer A, Milants I, Wouters PJ, Bouckaert B, Bruynincks F et al. Intensive insulin therapy in mixed medical, surgical intensive care units: benefit versus harm. Diabetes 2006;55:3151-9.
48. Van den Berghe G, Wouters PJ, Bouillon R, Weekers F, Verwaest C, Schetz M et al. Outcome benefit of intensive insulin therapy in the critically ill: Insulin dose versus glycemic control. Crit Care Med 2003;31:359-66.
49. Finney SJ, Zekoeid C, Elia A, Evans TW. Glucose control and mortality in critically ill patients. JAMA 2003;290:20417.
50. Kern JW, Shoemaker WC. Meta-analysis of hemodynamic optimization in high-risk patients. Crit Care Med 2002;30:1686-92.
51. Ceriello A, Ihnat MA, Thorpe JE. Clinical review 2: The “metabolic memory”: is more than just tight glucose control necessary to prevent diabetic complications? J Clin Endocrinol Metab 2009;94:410-5.
52. Schultz MJ, Spronk PE, van Braam Houckgeest F. Glucontrol,

no control, or out of control? Intensive Care Med 2009. [Epub ahead of print]
53. Ferrer R, Artigas A, Levy MM, Blanco J, Gonzalez-Diaz G, Gamacho-Montero J et al. Improvement in process of care and outcome after a multicenter severe sepsis educational program in Spain. JAMA 2008;299:2294-303.
54. Dellinger RP, Levy MM, Carlet JM, Bion J, Parker MM, Jaeschke R et al. Surviving Sepsis Campaign: international guidelines for management of severe sepsis and septic shock: 2008. Crit Care Med 2008;36:296-327.
55. Wolthuis EK, Korevaar JC, Spronk P, Kuiper MA, Daziljic M, Vroom MB et al. Feedback and education improve physician compliance in use of lung-protective mechanical ventilation. Intensive Care Med 2005;31:540-6.
56. Van den Berghe G, Schetz M, Vlasselaers D, Hermans G, Wilmer A, Bouillon R et al. Clinical review: Intensive insulin therapy in critically ill patients. NICE-SUGAR or Leuven blood glucose target? J Clin Endocrinol Metab 2009;94:316370 .
57. Eichacker PQ, Gerstenberger EP, Banks SM, Cui X, Natanson C. Meta-analysis of acute lung injury and acute respiratory distress syndrome trials testing low tidal volumes. Am J Respir Crit Care Med 2002;166:1510-4.
58. Cheng JW, Eisner MD, Thompson BT, Ware LB, Matthay MA. Acute effects of tidal volume strategy on hemodynamics, fluid balance, and sedation in acute lung injury. Crit Care Med 2005;33:63-70; discussion 239-240.
59. Kahn JM, Andersson L, Karir V, Polissar NL, Neff MJ, Rubenfeld GD. Low tidal volume ventilation does not increase sedation use in patients with acute lung injury. Crit Care Med 2005;33:766-71.
60. Wolthuis EK, Veelo DP, Choi G, Determann RM, Korevaar JC, Spronk PE et al. Mechanical ventilation with lower tidal volumes does not influence the prescription of opioids or sedatives. Crit Care 2007;11:R77.
61. Vriesendorp TM, DeVries JH, van Santen S, Moeniralam HS, de Jonge E, Roos YB et al. Evaluation of short-term consequences of hypoglycemia in an intensive care unit. Crit Care Med 2006;34:2714-8.
62. Suh SW, Gum ET, Hamby AM, Chan PH, Swanson RA. Hypoglycemic neuronal death is triggered by glucose reperfusion and activation of neuronal NADPH oxidase. J Clin Invest 2007;117:910-8.
63. Manns BJ, Lee H, Doig CJ, Johnson D, Donaldson C. An economic evaluation of activated protein C treatment for severe sepsis. N Engl J Med 2002;347:993-1000.
64. Davies A, Ridley S, Hutton J, Chinn C, Barber B. Angus DC. Cost effectiveness of drotrecogin alfa (activated) for the treatment of severe sepsis in the United Kingdom. Anaesthesia 2005;60:155-62.
65. van Boest PA, Hoñstra JJ, Schultz MJ, Boerna EC, Spronk PE, Kuiper MA. The incidence of low venous oxygen saturation on admission to the intensive care unit: a multi-center observational study in The Netherlands. Crit Care 2008;12:R33.
66. Osipina-Tascon GA, Buchele GL, Vincent JL. Multicenter, randomized, controlled trials evaluating mortality in intensive care: doomed to fail? Crit Care Med 2008;36:1311-22.
67. Herasevich V, Yilmaz M, Khan H, Chute C, Gajic O. Rule base systemfor identification of patients with specific critical care syndromes: The"Seiffer" for acute lung injury. J Am Med Informatics 2007;11:972.
68. Rana R, Afessa B, Keegan MT, Whalen FX, Jr., Nuttall GA, Everson LK et al. Evidence-based red cell transfusion in the critically ill: quality improvement using computerized physician order entry. Crit Care Med 2006;34:1892-7.
69. Ely EW, Baker AM, Dunagan DP, Burke HL, Smith AC, Kelly PT et al. Effect on the duration of mechanical ventilation of identifying patients capable of breathing spontaneously. N Engl J Med 1996;335:1864-9.
70. Yilmaz M, Keegan MT, Iscimen R, Afessa B, Buck CF, Hubmayr RD et al. Toward the prevention of acute lung injury: protocol-guided limitation of large tidal volume ventilation and inappropriate transfusion. Crit Care Med 2007;35:1660-6; quiz 1667.
71. Shapiro NI, Howell MD, Talmor D, Labey D, Ngo L, Buras J et al. Implementation and outcomes of the Multiple Urgent Sepsis Therapies (MUST) protocol. Crit Care Med 2006;34:1025-32.

Funding.-M: Schultz is supported by the Netherlands Organization for Health Research and Development (ZonMW), NWO-VENI grant 2004 (project number 016.056.001).
Received on July 20, 2009. Accepted for publication on October 21, 2009.
Corresponding author: M. J. Schultz, Internist-Intensivist, MD PhD FCCP, Department of Intensive Care Medicine, C3-415, Academic Medical Center, University of Amsterdam, Melbergdreef 9, 1105 AZ Amsterdam, The Netherlands. E-mail: m.j.schultz@amc.uva.nl