Critical Appraisal of:
Maitland K, Kiguli S, Opoka RO, et al. Mortality after fluid bolus in African children with severe infection. N Engl J Med 2011;364:2483-95
Research in International Emergency Medicine: Scope, Impact and Challenges
EBM Topic: Subgroup Analysis
Journal Club - Mortality after Fluid Bolus in African Children with Severe Infection - The FEAST Trial
1. Journal Club
The FEAST Trial: Mortality after Fluid
Bolus in African Children with Severe Infection
Farooq Khan MDCM
PGY-4 FRCP-EM
McGill University
August 22nd 2012
Salvatore Mottillo MD
PGY-1 FRCP-EM
McGill University
2. Objectives
• To critically appraise a high quality, high impact
article conducted in a low-income country on a
controversial topic in Emergency Medicine
internationally
• To understand the importance of research and
evidence-based medicine in low and middle income
countries, in particular with respect to EM
• To appreciate the challenges of conducting research
under austere conditions and how they can be
overcome
3. Outline
• Lecture on research in resource-poor settings
• 1st half of FEAST trial video
• Overview of the article
• Appraisal
• EBM topic: Subgroup analysis
• 2nd half of video
• Concluding remarks
5. Outline
• Global Health Research on the WHO agenda
– Why it is important
– Challenges and how to overcome them
– Example of its impact – TDR
• International EM research
– Why it is important
– Challenges and how to overcome them
– Example of high-quality IEM research and its impact
6. No Health Without Research
“It seems astonishing that in the 21st century decisions on health care
can still be made without a solid grounding in research evidence.
This is true even in clinical research, whether for simple or complex
interventions1, where systematic reviews time and time again
conclude that the evidence base is inadequate.2
It is even more true in the areas of health policy and health systems,
where quality research is hampered further by a lack of shared
definitions, a lack of consensus on guiding principles, poor capacity
(especially in low-resource regions), and methodological
challenges.3,4”
Pang T, Terry RF, The PLoS Medicine Editors (2011) WHO/PLoS Collection ‘‘No Health Without
Research’’: A Call for Papers. PLoS Med 8(1): e1001008. doi:10.1371/journal.pmed.1001008
7. WHO – World Health Report 2013
• To show that research is important for meeting
health needs and improving health outcomes
• To encourage countries to therefore invest more
resources in developing and strengthening their
national health research systems
• To argue that countries should not see research as
an expense or as an afterthought, but as an
investment for a better, healthier future
8. WHO Research Policy
• Capacity
• Priorities
• Standards
• Translation
• Organization
WHO Research for Health Strategy - 63rd World Health Assembly May 2010
9. TDR – At a glance - Video
• http://video.who.int/streaming/tdr/TDR_Gate
s_Award_video.wmv
10. TDR – An example of the impact of
global health research
• Take home messages
– Ownership of research by communities
leads to faster translation into action in
those countries
– Empowering countries to develop own
priorities, own research capacities and own
citizens to support healthcare systems leads
to sustainable development
TDR at a glance: Fostering an effective global research effort on diseases of poverty
WHO 2010
11. EM Research in LMIC
“Emergency medical systems are part of the
10/90 gap in health research whereby less
than 10% of global research investment is
spent on problems affecting 90% of the
world’s population”
Global Forum for Health Research. 10/90 report on health research 2002. Geneva:
Global Forum for Health Research; 2002.
12. EM Research in LMIC
• EM research from developed countries
– May not be relevant to patients LMIC
• Variation in genetics, lifestyle and environment
– May not be applicable to LMIC
• Variation in social, economic and political pressures
– May not be easily translated into practice in LMIC
• Lack of resources for implementation, cultural barriers
V. Anantharaman. Scope of Emergency Medicine research in Low and Middle income Countries
EuSEM, AAEM, IV Mediterranean EM Congress 2007
13. Emergency Care in LMIC
Challenges
– Lack of pre-hospital care in
general
– Scarcity of trained personnel at
all levels
– Difficulty utilizing specialized
equipment
– Weak telecommunications
infrastructure
– Barriers in transportation
– Health facilities lacking expertise
– System lacking organization
– Low levels of knowledge
translation
Interventions
– Recruitment of lay responders
– Training programs for nurses,
doctors and first responders
– Equipment matched to level of
training
– Harnessing cellular phone
technology and radio receiver
sets
– Use of existing private motorized
vehicles and bicycle transport
– Implementation of proper triage
guidelines
– Organizing and financing system
development
14. Challenges faced by IEM researchers
• Time
– High service demand
• Personnel
– Lack of dedicated, trained researchers
– Brain drain
• Finances
– EDs are not profitable business units
– Not prioritised by Pharma
– Lack of government funding
– High levels of indirect costs
– V. Anantharaman. Scope of Emergency Medicine research in Low and Middle income Countries
EuSEM, AAEM, IV Mediterranean EM Congress 2007
15. Challenges faced by IEM researchers
• Infrastructure
– Unreliable documentation
– Basic data collection systems may be absent or
poorly implemented
• Credibility
– Negative publication bias in HIC journals
– Paucity of local journals in LMICs
16. Challenges faced by IEM researchers
• Ethical issues
– Lack of valid oversight bodies
– Confidentiality
– Informed consent
– Cultural differences (researcher vs participants)
– Double standards and exploitation of subjects
• Fly-in, fly-out research
Iserson et al Challenges in International Medicine:Ethical Dilemmas, Unanticipated Consequences, and
Accepting Limitations, SAEM Journal 2012
Ford et al, Ethics of conducting research in conflict settings, Conflict and Health 2009, 3:7
17. Strategies to circumvent challenges
• Capacity building on National/regional level
– WHO, TDR, CIDA, etc.
• Capacity building on a local level
– NGOs, University partnerships, leading by
example
• Cost-saving endeavors
– Better accountability, better prediction of direct
and indirect costs, recycling of costs
saved, funding application strategies
18. Strategies to circumvent challenges
• Ethical principles to follow
– Independent external DSMB/IRB
– Collaborative partnerships
– Community engagement
– Social value
– Scientific validity
– Fair selection of participants
– Harm-benefit ratio
– Respect for participants and consent
Ford et al, Ethics of conducting research in conflict settings, Conflict and Health 2009, 3:7
19. Priority areas of research in LMIC
• Epidemiology of conditions that need to be
addressed by emergency systems in these
countries (e.g. trauma, infectious diseases)
• Economic analysis and cost-effectiveness in
areas of low funding (e.g prehospital care,
injury prevention)
• Well designed, locally appropriate
intervention trials that establish effectiveness
20. The FEAST Trial - Video
• http://www.youtube.com/watch?v=hK9VUkL-
DqU
22. Background
• Malaria, sepsis, & other infections associated with
early mortality in African children
• Guidelines recommend fluid resuscitation for shock
• P.I.C.O. Question
– In children presenting with severe febrile illness and
impaired perfusion, does administration of 20 ml boluses
of 5% albumin solution or 0.9% saline solution per kg of
body weight diminish mortality at 48 hours?
23. Methods
• Article type
– Sub-Saharan Africa, multicenter, open-treatment, RCT
• 2 stratum
– Stratum A: children without severe hypotension (n=3,141)
• Group 1: saline-bolus group
• Group 2: albumin-bolus group
• Group 3: control group
– Stratum B: children with severe hypotension (n=29)
• Same groups as A
• Initial boluses: 20 ml/kg
• Primary end-point
– mortality at 48hrs
24. Results
• Total patients: 3,170
– Median age: 24mo.
– Malaria: 57% of patients
– Other parameters
• Mortality at 48hrs
– Saline-bolus group: 10.5%
– Albumin-bolus group: 10.6%
– Control group: 7.3%
27. Critical Appraisal
• Worksheet for a critical review of an article on
therapy+
+Guyatt G, Rennie D, Meade M, Cook D. Users' Guides to the Medical Literature: essentials of Evidence-
Based Clinical Practice. 2nd edition (Jama & Archives Journals). McGraw-Hill. Professional, 2008.
28. Critical Appraisal
GROUP 1
• I. Are the results valid?
– 1. Were patients randomized?
– 2. Was randomization concealed?
– 3. Were patients in the study groups similar with respect
to known prognostic factors?
– 4. To what extent was the study blinded?
29. Critical Appraisal
GROUP 2
• I. Are the results valid? (continued)
– 5. Was follow-up complete?
– 6. Were patients analyzed in the groups to which they
were randomized?
– 7. Was the trial stopped early?
30. Critical Appraisal
GROUP 3
• II. What are the results?
– 1. How large was the treatment effect?
– 2. How precise was the treatment effect?
– 3. Secondary endpoints studied?
31. Critical Appraisal
GROUP 4
• III. How can I apply the results to my patient care?
– 1. Were all patient-important outcomes considered?
– 2. Were the study patients similar to my patients?
33. Critical Appraisal - Answers
GROUP 1
• I. Are the results valid?
– 1. Were patients randomized?
• Yes, Block randomization, stratified
– 2. Was randomization concealed?
• Yes, Sealed envelopes, opened in numerical order
– 3. Were patients in the study groups similar with
respect to known prognostic factors?
• Yes, large number of patients recruited (n=3,170)
• Similar baseline characteristics between groups.
37. Critical Appraisal - Answers
GROUP 1
• I. Are the results valid?
– 4. To what extent was the study blinded?
• Open-treatment trial
• Blinded end-point review committee
38. Critical Appraisal
GROUP 2
• I. Are the results valid? (continued)
– 5. Was follow-up complete?
• Few pts lost to f-u, and losses similar between groups
– Saline-bolus group: 8/1047 pts lost to f-u by 48hrs
– Albumin-bolus group: 7/1050 pts lost to f-u by 48hrs
– Control group: 2/1044 pts lost to f-u by 48hrs
39. Critical Appraisal
GROUP 2
• I. Are the results valid? (continued)
– 6. Were patients analyzed in the groups to which they
were randomized?
• Almost all patients received intervention they were assigned to.
– Saline-bolus group: 1041/1047 patients
– Albumin-bolus group: 1045/1050 patients
– Control group: 1043/1044 patients
• ITT analysis
40. Critical Appraisal
GROUP 2
• I. Are the results valid? (continued)
– 7. Was the trial stopped early?
• Yes, unlikely that superiority of the bolus strategy over the control
strategy would be shown
41. Critical Appraisal
GROUP 3
• II. What are the results?
– 1. How large was the treatment effect?
– 2. How precise was the treatment effect?
– Mortality at 48 hours
• Saline-bolus group: 110 deaths /1047 patients (10.5%)
– RR vs control: 1.44 (95% CI, 1.09 to 1.90)
– ARI vs control: 3.22%
– NNH vs control: 31.06
– OR vs control: 1.50
42. Critical Appraisal
GROUP 3
• II. What are the results?
– Mortality at 48 hours
• Albumin-bolus group: 111/1050 (10.6%)
– RR vs control: 1.45 (95% CI, 1.10 to 1.92)
– ARI vs control: 3.29%
– NNH vs control: 30.40
– OR vs control: 1.51
43. Critical Appraisal
GROUP 3
• II. What are the results?
– Mortality at 48 hours
• Bolus therapy (saline + albumin groups)
– RR vs control: 1.45 (95% CI, 1.13 to 1.86)
– ARI vs control: 3.26%
– NNH vs control: 30.67
– OR vs control: 1.50
• No bolus (control group): 76/1044 (7.3%)
46. Critical Appraisal
GROUP 4
• III. How can I apply the results to my patient care?
– 1. Were all patient-important outcomes considered?
• Yes, mortality at 48 hours
– Secondary endpoints
» Mortality at 4 weeks
» Pulmonary edema
» Severe hypotension
» Increased ICP
» Neurologic sequelae
47. Critical Appraisal
GROUP 4
• III. How can I apply the results to my patient care?
– 1. Were the study patients similar to my patients?
• Inclusion criteria
– Children b/w 60d and 12y of age
– ‘Severe febrile illness’
48. Critical Appraisal
GROUP 4
• III. How can I apply the results to my patient care?
– 1. Were the study patients similar to my patients?
• Exclusion criteria
– Severe malnutrition
– Gastroenteritis
– Non-infectious shock
49. Critical Appraisal
GROUP 4
• III. How can I apply the results to my patient care?
– Differences between the study pop. and my patients?
• Severe malnourishment
» Pts with acute severe malnutrition excluded
• 2% (n=70) mid-upper-arm circumference of <11.5 cm.
50. Critical Appraisal
GROUP 4
• III. How can I apply the results to my patient care?
– Differences between the study pop. and my patients?
• High prevalence (57%) malaria
» Subgroup: bolus vs no bolus in pts without malaria (n=1,330)
• RR death at 48 hrs: 1.43 (1.01-2.04)
• Adjusted RR: 1.51 (1.17-1.94)
51. Critical Appraisal
GROUP 4
• III. How can I apply the results to my patient care?
– Differences between the study pop. and my patients?
• Severe anemia
» Subgroup: bolus vs no bolus in pts with less severe anemia or
>5g/dl (n=1,384)
• RR death at 48 hrs: 1.31 (0.93-1.84)
• Adjusted RR: 1.47 (1.14-1.90)
52. Critical Appraisal
GROUP 4
• III. How can I apply the results to my patient care?
– Differences between the study pop. and my patients?
• Pts did not fit WHO definition of shock
» Subgroup: bolus vs no bolus in pts with WHO definition of
shock (n=65)
• ARI death at 48 hrs: 28.0 (3.4-52.5)
54. EBM topic: subgroup analysis
• Definition
– Any evaluation of treatment effects for a specific endpoint
in subgroups of patients defined by baseline
characteristics
• Why use subgroup analysis?
– Differences in treatment effect between subpopulations
– Consistency of trial conclusions for subpopulations
55. EBM topic: subgroup analysis
• Evaluating variability in subgroup analyses
– Visual evaluation of variability
• How similar are the point estimates?
• To what extent to the confidence intervals overlap?
• Statistical tests evaluating variability
– Yes-or-no tests for heterogeneity that generate P-value
– I2 test that quantifies the variability explained by
differences in the results
• Be careful for possible confounders
– Can be adjusted in multivariate analysis
57. EBM topic: subgroup analysis
• Guidelines for Deciding Whether Apparent
Differences in Subgroup Response Are Real+
– Did the hypothesis precede rather than follow the
analysis?
– Was the subgroup difference one of a small number of
hypothesized effects tested?
– Is the subgroup difference suggested by comparisons
within rather than between studies?
+Guyatt G, Rennie D, Meade M, Cook D. Users' Guides to the Medical Literature: essentials of Evidence-
Based Clinical Practice. 2nd edition (Jama & Archives Journals). McGraw-Hill. Professional, 2008.
58. EBM topic: subgroup analysis
• Guidelines for Deciding Whether Apparent
Differences in Subgroup Response Are Real
– Is the magnitude of the subgroup difference large?
– Is the subgroup difference consistent across studies?
– Was the subgroup difference statistically significant?
– Does external evidence support the
hypothesized subgroup difference?
61. Conclusion
• Fluid resuscitation increased the abs. risk of death
at 48hrs by 3.3% in African hospitals
• Importance of fluid resuscitation as a lifesaving
intervention in poor settings for children with
shock?
• Questions raised regarding fluid-resuscitation
guidelines in other settings as well
FEAST trialPublished in NEJM in june last year But continued to make waves in the international EM community, and sparked editorials, commentaries and correspondence in the Lancet and other high impact journals
To showcase a side of research that we don’t get to see here in our ivory towers of the developing worldAnd not take basic research infrastructure for granted
It seems astonishing that in the 21st century decisions on health care can still be made without a solid grounding in research evidence. This is true even in clinical research, whether for simple or complex interventions [1], where systematic reviews time and time again conclude that the evidence base is inadequate [2]. It is even more true in the areas of health policy and health systems, where quality research is hampered further by a lack of shared definitions, a lack of consensus on guiding principles, poor capacity (especially in low-resource regions), and methodological challenges [3,4].
The next World Health Report will discuss the contributions of research to universal health coverage. This flagship report from WHO will, for the first time in its history, focus on research for better healthThe World Health Report will further highlight the significance of conducting and translating health research to help countries move towards universal health coverage
Research policy involves the strengthening of health research systems. It aims to contribute to health system development and health improvement particularly in poorer countries by:the dissemination and translation of valuable knowledge or research;the creation of ethical and evidence- based research policies, including norms and standards;the promotion, monitoring and implementation of high quality health research evidence. - building capacity to strengthen health research systemssupporting the setting of research priorities that meet health needs particularly in low and middle income countriescreating an environment to create good research practice and enable the greater sharing of research evidence, tools and materials.ensuring quality evidence is turned into products and policy.action to strengthen the research culture within WHO and improve the management and coordination of WHO research activities
5 Elimination campaignsLeprosy, onchocerciasis, chagas disease, lymphatic filariasis, visceral leishmaniasisDevelopment of 12 new drug therapiesIvermectin, antimalarials, etc.Documented effectivenessArtemesinin-combination therapies, insecticide-treated bednetsTrained thousands of researchers in developing countries
Cannot talk about the challenges faced by EM researchers without first talking about the challenges of simply practicing EM in these countries.
Lack of training in biostats, epi, EBMPharma favor dept with track record and research infrastructure(power/equipment failures and maintenance, insurance, administrative fees, corruption, etc.)
EM care guidelines recommendrapid, earlyfluidresuscitation for children in shockpoorevidence? Whatstudies are there?4. Dellinger RP, Levy MM, Carlet JM, et al.Surviving Sepsis Campaign: internationalguidelines for management of severe sepsisand septic shock: 2008. Crit Care Med2008;36:296-327. [Erratum, Crit Care Med2008;36:1394-6.]Determination of the Quality of Evidence• Underlying methodologyA RCTB Downgraded RCT or upgraded observational studiesC Well-done observational studiesD Case series or expert opinionFactors that may decrease the strength of evidence1.Poor quality of planning and implementation of available RCTs suggesting high likelihood of bias2.Inconsistency of results (including problems with subgroup analyses)3.Indirectness of evidence (differing population, intervention, control, outcomes, comparison)4.Imprecision of results5.High likelihood of reporting biasMain factors that may increase the strength of evidence1.Large magnitude of effect (direct evidence, relative risk (RR) > 2 with no plausible confounders)2.Very large magnitude of effect with RR > 5 and no threats to validity (by two levels)3.Dose response gradientearly goal-directed resuscitation of the septic patient during the first 6 hrs after recognition (1C); 5. Parker MM, Hazelzet JA, Carcillo JA.Pediatric considerations. Crit Care Med2004;32:Suppl:S591-S594Carcillo JA, Davis AL, Zaritsky A: Role of early fluid resuscitation in pediatric septic shock. JAMA 1991Fluid ResuscitationIntravenous access for fluid resuscitation and inotrope/vasopressor infusion is more difficult to attain in children than in adults. The American Heart Association has developed pediatric advanced life support guidelines for emergency establishment of intravascular support (2). On the basis of a number of studies, it is accepted that aggressive fluid resuscitation with crystalloids or colloids is of fundamental importance to survival of septic shock in children (3, 4). There is only one randomized, controlled trial comparing the use of colloid with crystalloid resuscitation (dextran, gelatin, lactated Ringers, or saline) in children with dengue shock (3). All these children survived, regardless of the fluid used, but the longest time to recovery from shock occurred in children who received lactated Ringers. Among patients with the narrowest pulse pressure, there was a suggestion that colloids were more effective than crystalloids in restoring normal pulse pressure. Fluid infusion is best initiated with boluses of 20 mL/kg over 5–10 mins, titrated to clinical monitors of cardiac output, including heart rate, urine output, capillary refill, and level of consciousness. Children normally have a lower blood pressure than adults and can prevent reduction in blood pressure by vasoconstriction and increasing heart rate. Therefore, blood pressure by itself is not a reliable endpoint for assessing the adequacy of resuscitation. However, once hypotension occurs, cardiovascular collapse may soon follow. Hepatomegaly occurs in children who are fluid overloaded and can be a helpful sign of the adequacy of fluid resuscitation. Large fluid deficits typically exist, and initial volume resuscitation usually requires 40–60 mL/kg but can be much higher (4–6).
6 centersMaintenance fluidEndpointreviewcommittee – blinded to treatmentreceivedAll children in stratum A (without severe hypotension) were randomly allocated to one of three groups: 1) rapid volume expansion over the course of 1 hour with 20 ml of IV 0.9% saline solution per kg (saline-bolus group), 2) 20 ml of IV 5% human-albumin solution per kg (albumin-bolus group), and 3) no bolus (control group). Children in stratum B (with severe hypotension) were randomly allocated to 40 ml of saline bolus or albumin bolus per kg. The initial boluses were increased to 40 ml per kg (60 ml per kg in stratum B) after a protocol amendment in June 2010. The primary end-point was mortality at 48 hours after randomization. Secondary end-points were mortality at 4 weeks, neurologic sequelae at 4 and 24 weeks, episodes of hypotensive shock within 48 hours after randomization, and adverse events potentially related to fluid resuscitation (pulmonary edema, increased intracranial pressure, and severe allergic reaction).severe hypotension (systolic bloodpressure of <50 mm Hg in children younger than12 months of age, <60 mm Hg in children 1 to5 years of age, and <70 mm Hg in children olderthan 5 years of age)
Separate into groups15 -20 minutes at leastGo around and see groups; stimulate discussion
Expand worksheet
Question for all groupsAccording to thisstudy, benefits of fluidbolusseem minimal and are harmful in the sub-SaharanAfrican setting.Harmfuleffects of fluid administration maybe offset by more intensive patient care (i.e. mechanical ventilation, closer observation) in the Western setting.Will this study change our practice?
1. randomizationwasperformed in permuted blocks of randomsizes, and wasstratifiedaccording to clinical center2. trial numberswerekeptinside opaque, sealedenvelopes, whichwerenumberedconsecutively and opened in numericalorder by a studyclinician.
Successful randomizationNo difference between groups
Successful randomization
Open-treatmenttrial, in whichclinicianswereaware of the treatmentbeingprovided to eachchild.However, an end-point reviewcommittee, whosememberswereunaware of the treatmentassignments, reviewed all deaths, neurologicsequelae, and adverse eventsCliniciansshouldbeblinded in order to preventdifferential administration oftherapiesthat affect the outcome of interest, co-intervention.Overallsupportivetreatmentsweresimilar in each groups (transfusions, atb, etc): A total of 1408 children received blood transfusions — 472 (45%) in the albumin-bolusgroup, 487 (47%) in the saline-bolus group, and449 (43%) in the control group.The receipt, and the timing of the receipt, of blood, quinine, and antibiotics were similar across groups; only bolus fluid resuscitation differed between the intervention and control groups.
Some pts did not receive the right intervention, but so few violations would not bias results, also deaths similar in all groupsan intention to treat (ITT) analysis (sometimes also called intent to treat) is an analysis based on the initial treatment intent, not on the treatment eventually administered. ITT analysis is intended to avoid various misleading artifacts that can arise in intervention research
Yes, at the 5th interimreview of data on January 12, 2011, with data availablefrom 2,995 children, the independent DSMB recommendedstoppingenrollmentowing to safetyconcerns in the bolus groups and becauseitwasveryunlikelythatsuperiority of the bolusstrategy over the control strategywouldbeshownAlthough it is becoming increasingly popular, stopping trials early when one sees an apparent large benefit is risky.27 Trials terminated early will compromise randomization if they stop at a “random high” when prognostic factors temporarily favor the intervention group. Particularly when sample size and the number of events are small, trials stopped early run the risk of greatly overestimating the treatment effect (see Chapter 9.3, Randomized Trials Stopped Early for Benefit)
Albumin-bolus group: 111/1050 (10.6%) RR vs control: 1.45 (95% CI, 1.10 to 1.92) ARI vs control: 3.29% NNH vs control: 30.40 OR vs control: 1.51 Bolus therapy (saline + albumin groups) RR vs control: 1.45 (95% CI, 1.13 to 1.86) ARI vs control: 3.26% NNH vs control: 30.67 OR vs control: 1.50 No bolus (control group): 76/1044 (7.3%) There was no difference in end-points between the saline-bolus and albumin-bolus groups.
Albumin-bolus group: 111/1050 (10.6%) RR vs control: 1.45 (95% CI, 1.10 to 1.92) ARI vs control: 3.29% NNH vs control: 30.40 OR vs control: 1.51 Bolus therapy (saline + albumin groups) RR vs control: 1.45 (95% CI, 1.13 to 1.86) ARI vs control: 3.26% NNH vs control: 30.67 OR vs control: 1.50 No bolus (control group): 76/1044 (7.3%) There was no difference in end-points between the saline-bolus and albumin-bolus groups.
Albumin-bolus group: 111/1050 (10.6%) RR vs control: 1.45 (95% CI, 1.10 to 1.92) ARI vs control: 3.29% NNH vs control: 30.40 OR vs control: 1.51 Bolus therapy (saline + albumin groups) RR vs control: 1.45 (95% CI, 1.13 to 1.86) ARI vs control: 3.26% NNH vs control: 30.67 OR vs control: 1.50 No bolus (control group): 76/1044 (7.3%) There was no difference in end-points between the saline-bolus and albumin-bolus groups.
Albumin-bolus group: 111/1050 (10.6%) RR vs control: 1.45 (95% CI, 1.10 to 1.92) ARI vs control: 3.29% NNH vs control: 30.40 OR vs control: 1.51 Bolus therapy (saline + albumin groups) RR vs control: 1.45 (95% CI, 1.13 to 1.86) ARI vs control: 3.26% NNH vs control: 30.67 OR vs control: 1.50 No bolus (control group): 76/1044 (7.3%) There was no difference in end-points between the saline-bolus and albumin-bolus groups.
Albumin-bolus group: 111/1050 (10.6%) RR vs control: 1.45 (95% CI, 1.10 to 1.92) ARI vs control: 3.29% NNH vs control: 30.40 OR vs control: 1.51 Bolus therapy (saline + albumin groups) RR vs control: 1.45 (95% CI, 1.13 to 1.86) ARI vs control: 3.26% NNH vs control: 30.67 OR vs control: 1.50 No bolus (control group): 76/1044 (7.3%) There was no difference in end-points between the saline-bolus and albumin-bolus groups.
FEAST definiton of shock: severe febrile illness complicated by impaired consciousness (prostration or coma), respiratory distress (increased work of breathing), or both, and with impaired perfusion, as evidenced by one or more of the following: a capillary refill time of 3 or more seconds, lowerlimb temperature gradient,19 weak radial-pulse volume, or severe tachycardia (>180 beats per minute in children younger than 12 months of age, >160 beats per minute in children 1 to 5 years of age, or >140 beats per minute in children older than 5 years of age) (Fig. 1)WHO definition of shock:Cold hands or feet with both capillary refill time >3 secand weak and fast pulseInclusion criteriaExclusion criteriaDifferences in populationslatepresentation (social) malariaanemiamalnourishmentgeneticallyenvironmentalSyndromic care: often dont know the diagnosis, dont have the labs, the radiology, often dont have the h&p – applicable to us in thisway
2. Were all patient-important outcomesconsidered?3. Are the likelybenefitsworth the potentialharms and costs?Patients presentlate in infection compared to industrialized nationNo, sub-SaharanAfricanchildrenbetween 60d and 12y old57% pts had malariaInclusion criteriaExclusion criteriaDifferences in populationslatepresentation (social) malariaanemiamalnourishmentgeneticallyenvironmentalSyndromic care: often dont know the diagnosis, dont have the labs, the radiology, often dont have the h&p – applicable to us in thisway
WHO shock: WHO ETATCold hands or feet with both capillary refill time >3 secand weak and fast pulseCold hands or feet with both capillary refill time >3 secand weak and fast pulse
High prevalence (57%) of malaria in the study population means these results are not applicable to the Western setting.
African children are more likely to be severely anemic; hence fluid resuscitation in this population may be harmful
WHO shock: WHO ETATCold hands or feet with both capillary refill time >3 secand weak and fast pulseFEAST definiton of shock: severe febrile illness complicated by impaired consciousness (prostration or coma), respiratory distress (increased work of breathing), or both, and with impaired perfusion, as evidenced by one or more of the following: a capillary refill time of 3 or more seconds, lowerlimb temperature gradient,19 weak radial-pulse volume, or severe tachycardia (>180 beats per minute in children younger than 12 months of age, >160 beats per minute in children 1 to 5 years of age, or >140 beats per minute in children older than 5 years of age) (Fig. 1)
To look for Differences in treatment effect between different patient populationsTo investigate the consistency of trial conclusions for different subpopulations
Fortunately, one can resolve this unsatisfactory situation. Having completed the review, investigators should present the results in a way that allows clinicians to check whether results proved similar from study to study. There are 4 elements to consider when deciding whether the results are sufficiently similar to warrant comfort with a single estimate of treatment effects that applies across the populations, interventions, and outcomes studied (Table 19-5, which also appears as Table 20.3-1 in Chapter 20.3, Making Sense of Variability in Study Results). First, how similar are the study-specific estimates of the treatment effect (that is, the point estimates) from the individual studies? The more different they are, the more clinicians should question the decision to pool results across studies. Second, to what extent are differences among the results of individual studies greater than you would expect by chance? Users can make an initial assessment by examining the extent to which the confidence intervals (CIs) overlap. The greater the overlap, the more comfortable one is with pooling results. Widely separated CIs flag the presence of important variability in results that requires explanation. Clinicians can also look to formal statistical analyses called tests for heterogeneity, which address the null hypothesis that underlying effects are in fact similar across studies and the observed differences in the size of effect between studies are due to chance. When the P value associated with the test of heterogeneity is small (for instance, P < .05), chance becomes an unlikely explanation for the observed differences in the size of the effect (see Chapter 10.1, Hypothesis Testing). A fourth criterion is another statistic, the I2, which describes the percentage of the variability in effect estimates that is due to underlying differences in effect rather than chance.32 Rough guides for the interpretation of I2 suggest that a value of less than 20% represents minimal vaLow p value means there actually is a difference between 2 groups and is not due to chanceMention lower N in subgroup
Is the difference du to a confounding factor other than that in the subgroup analysis?Low p value means there actually is a difference between 2 groups and is not due to chanceHigh p value mean that any difference seen between 2 groups is likely due to chanceThe heterogeneity of treatment effects across the levels of a baseline variable refers to the circumstance in which the treatment effects vary across the levels of the baseline characteristic. Heterogeneity is sometimes further classified as being either quantitative or qualitative. In the first case, one treatment is always better than the other, but by various degrees, whereas in the second case, one treatment is better than the other for one subgroup of patients and worse than the other for another subgroup of patients. Such variation, also called “effect modification,” is typically expressed in a statistical model as an interaction term or terms between the treatment group and the baseline variable. The presence or absence of interaction is specific to the measure of the treatment effect.The appropriate statistical method for assessing the heterogeneity of treatment effects among the levels of a baseline variable begins with a statistical test for interaction.10-13 For example, Sacks et al.8 showed the heterogeneity in pravastatin efficacy by reporting a statistically significant (P=0.03) result of testing for the interaction between the treatment and baseline LDL level when the measure of the treatment effect was the relative risk. Many trials lack the power to detect heterogeneity in treatment effect; thus, the inability to find significant interactions does not show that the treatment effect seen overall necessarily applies to all subjects. A common mistake is to claim heterogeneity on the basis of separate tests of treatment effects within each of the levels of the baseline variable.6,7,14 For example, testing the hypothesis that there is no treatment effect in women and then testing it separately in men does not address the question of whether treatment differences vary according to sex. Another common error is to claim heterogeneity on the basis of the observed treatment-effect sizes within each subgroup, ignoring the uncertainty of these estimates.
Did the hypothesis precede rather than follow the analysis?Hypotheses for secondary endpoints should be mentioned in the abstract or methodsCommonly, study authors carry out subgroup analyses after analyzing the data in an effort to demonstrate a treatment effect in a small portion of the populationWas the subgroup difference one of a small number of hypothesized effects tested?Studies that conduct a large number of subgroup analyses are more likely to be ‘fishing’ for false positive treatment effectsIs the subgroup difference suggested by comparisons within rather than between studies?Subgroup analyses should always be conducted within the same study. Comparing analyses between studies introduces several biases (including different patient populations, different interventions, different assessment of outcomes, etc)
Is the magnitude of the subgroup difference large?How large is the difference in risk between a subgroup and the rest of the population studied?Is the subgroup difference consistent across studies?Have other studies shown similar resultsWas the subgroup difference statistically significant?In order to be statistically significant the confidence interval should not cross unity. Also should have a large number of pts in subgroupDoes external evidence support the hypothesized subgroup difference?There should be a plausible explanation for the results obtained in the subgroup analysis.
2. Were all patient-important outcomesconsidered?3. Are the likelybenefitsworth the potentialharms and costs?According to thisstudy, benefits of fluidbolusseem minimal and are harmful in the sub-SaharanAfrican setting.Harmfuleffects of fluid administration maybe offset by more intensive patient care (i.e. mechanical ventilation, closer observation) in the Western setting.Will this study change our practice?
The results challenge the importance of fluid resuscitation as a lifesaving intervention in poor settings for children with shock?