The protocol for the FEAST was developed after executing 3 preparatory randomised trials and extensive data gathering. These have provided supportive physiological , efficacy and safety data to justify the protocol and to inform the trial design.
The final protocol has been through extensive consultation and international review. The children entering into the trial will be critically sick and therefore mortality is likely to be high. Considerable attention was paid to address all potential concerns with regard to the justification, design and conduct of the trial.
|Background and Justification
|Why is there a need to conduct this trial?||Across Africa clinical practise with respect to fluid management varies widely: most centres restricting fluids to maintenance or not at all, while others resort to the administration of whole blood, despite cost (30-50 USD/unit) and concerns of safety. Shock is rarely diagnosed or treated. Making significant changes to current practise and thus international policy will require definitive results from a well powered randomised controlled trial.|
|Why will FEAST include both those with malaria and sepsis?
Your previous trials and the supportive background data focus upon volume depletion in children with severe malaria only.
|Globally frontline medical staff faced with a sick child usually initiate life-supportive treatment (including fluid expansion) without supportive diagnostic information. In Africa too most treatment decisions for very sick children also lack supportive diagnosis other than the clinical findings at the bedside. In this trial we have chosen to include both severe malaria and bacterial sepsis since they are clinically indistinguishable at admission. In usual circumstances emergency care for these children proceeds before securing a diagnosis. For the trial results to be contextually relevant to the acute care of sick children we opted to design a pragmatic trial where decision to enrol and treat is based on a rapid structured clinical assessment. This trial therefore supports a vertical or ‘generic’ approach to the emergency management of a sick child rather than the more syndromic based approach which often results in conflicting views and recommendations for treatment of common complications (such a fluid management). and a dilemma for frontline staff over which guideline to follow.|
|Don’t we already know how to treat children with septic shock?||In Europe, Northern America and Australia guidelines that specify how to identify and treat children with septic shock leading to intravascular volume depletion have been widely adopted and resulted in significant improvements in outcome. Nevertheless, the context in which these are applied are very different. Access to well resourced emergency and intensive care units is guaranteed whereas most children in Africa will be treated in hospitals with very limited infrastructure and few personnel. These are major obstacles to administering potential life saving fluid resuscitation where current medical practise differs and where the safety of the intervention remains a major concern.|
|Eligibility criteria and study population
|Do the definitions for trial conflict with current WHO and national paediatric guidelines and thus preclude a control arm (low volume maintenance only)?||WHO currently only recommends rapid fluid resuscitation for the treatment of advanced shock (that is defined as the presence of cold hands or feet plus both a weak rapid pulse and capillary refill time >3 seconds) in children. Validation of these signs using our paediatric surveillance data in Kilifi indicates that shock would rarely diagnosed and mortality extremely high in that group. This was also confirmed when these signs were prospectively validated in Brazil. The WHO definition approximates to children enrolled into Group B arm of the trial (decompensated shock) who all receive fluid boluses at admission. It also approximates to the ‘rescue criteria’ for children allocated to control. We therefore see no conflict with current recommendations. Finally, if the admitting doctor feels that the patient would better off not enrolling than enrolling into the trial (ie clinical equipoise is not supported) then they may chose not to enrol the child into the trial.|
|Some of the clinical signs are not currently familiar to, or in regular use by, the medical teams.
Will this be a problem for enrolment or cause bias?
|Before the start of the trial we will give a training course to all members of the acute paediatric services at the hospitals involved in the trial. This will include both trial and non trial staff. Visual and video training material will be available onsite throughout the trial. A basic life support course will be taught too by our qualified trainers. Prior to the trial the team will familiarise themselves with the clinical signs by screening the general paediatric admission population under supervision of a trainer. These steps should improve sign recognition for all patients and reduce the potential for bias between trial sites.|
|Are the children with rapid breathing in heart failure or at risk of heart failure?||Detailed physiological studies of critically sick children with respiratory distress have show that the features of metabolic acidosis and volume depletion were common is these children. Conversely, confirmatory physiological evidence to suggest heart failure was rarely reported.
In the safety and efficacy fluid resuscitation studies we conducted we anticipated that the intervention may have the potential to cause pulmonary oedema (heart failure) as a life threatening complication, especially in children with evidence of mild-moderate malnutrition. We have now reported data from four separate trials including 301 closely monitored children with severe malarial acidosis receiving volume expansion, including children with severe symptomatic malarial anaemia (haemoglobin <5g/dl) and 20-30% had anthropometric indices of moderate malnutrition. Clinical features of pulmonary oedema only developed in two cases – an overall complication rate ~0.7%. Diuretics were neither prescribed routinely nor were they required in these trials. The two cases included severe salicylate toxicity and a child who inadvertently received a rapid whole blood transfusion (for stable anaemia) and so neither were directly related to the fluid intervention.
|Why have you included children with a low haemoglobin (<5g/dl)?||Often children with severe anaemia and respiratory distress die whilst awaiting the arrival of an emergency transfusion. We have shown in a small pilot study that children given pre transfusion management with a bolus of saline or albumin were no worse off than if left waiting for a transfusion without pre-transfusion management. Indeed significant clinical improvements were observed even prior to transfusion. As this is a pragmatic trial and fluid interventions would be given whilst blood test results awaited these cases were also included. Moreover, significant delays would be encountered if enrolment were consequent upon the laboratory results. Volume replacement whilst awaiting emergency blood transfusion is recommended by the WHO for every other indication/group and information obtained about this important subgroup in the FEAST trial would generate important evidence to support the basis for acute transfusion practice guidelines.|
|Justification for choice of trial fluids
|Can you justify the use of albumin in the trial and will it ever be available for use in Africa?
Why haven’t you used a cheaper, synthetic colloid rather than albumin?
|We recognise that in resource poor countries of Africa, the use of albumin or any synthetic colloid as a resuscitation fluid is complicated by both cost and lack of availability. We have recently examined whether a similar survival benefit we have seen with albumin could be achieved using a cheaper non-albumin colloidal solution (Gelofusine). Although we reconfirmed the striking survival benefit of albumin infusion, seen in our previous trials, we failed to identify any clear survival benefits of Gelofusine over crystalloid solutions (saline) and demonstrated an increase risk of allergic reactions to these synthetic compounds. Taken together these suggest the need to explore albumin rather than other synthetic colloids in a definitive trial. Albumin has a range of physiological effects in addition to its colloidal properties and if shown to be effective could be rapidly introduced once local manufacturing from a national transfusion services introduced and established. Cost economic analysis indicate that at current costs to save one life by albumin infusion would be approximately US$ 35 and would be cost-effective over 0.9% saline for a decision-maker prepared to pay US$ 8.83 per life gained.|
|Your previous trial showed a high mortality in the saline arm – can the use of it be justified in this trial?||The design of our Phase II trial did not enable us to determine whether saline administration was preferable to standard treatment alone, or whether saline administration was actually hazardous. In that trial children were enrolled on the basis of a feature of severe illness (respiratory distress and/o impaired consciousness) plus metabolic acidosis. Mortality in children with coma receiving saline boluses was similar to that in published studies from comparable patient groups, which have consistently documented rates of 28-41%. Nevertheless, it is possible that saline may confer benefit over control, particularly if the volume and rate administrated is dictated by additional features of shock (rather than base excess alone). For non-comatose children receiving saline, mortality was lower than previously reported (1/70, ~1%). As saline, is cheap, widely available, and its use is currently recommended by the WHO for children without severe malaria, providing an adequate evidence base for these recommendations is essential, particularly when opinion on the optimal fluid management remains divided.|
|You have chosen to use Saline and not Ringer’ lactate – why?||Both are solutions for fluid resuscitation and are widely available in Tanzania, Uganda and Kenya. However, Saline contains only sodium chloride (salt) and water whereas Ringer’s Lactate contains sodium, lactate (which is metabolised to bicarbonate), and potassium. We have previously shown that children with severe malaria often have dangerously raised levels of potassium at admission so giving extra potassium is not advisable. This precludes the use of Ringer’s Lactate in this trial.|
|Practicalities and logical issues
|Will you need specialist facilities to manage these children?||No, this is a pragmatic trial which addresses the simple question of whether the addition of a bolus of resuscitation fluid to standard management can result in significant improvements to outcome. We have over 10 years of accumulated experience with fluid resuscitation in severely ill Kenyan children and it has not been shown to result in harm. We do not aim to make significant changes to the environment in which the trial is conducted other than the professional training of the trial and non- trial staff to aide identification and treatment of all sick children. The addition of high dependency care or intensive care facilities for the conduct of the trial would mean a further trial in low resource hospitals would be required to assess the generalisablity of the active intervention protocol. This is neither justified on the basis of safety, cost required or timely advancement of medical therapeutic practice.|
|Will you be able to folFrelow children up at one month?||All the investigators have experience working with communities and long term follow up of patients in studies or trials. This shared experience plus that of the local investigators will prove invaluable in the ability to identify non-attendees to follow up. This resource will help design demographic details recorded to trace all families prior to hospital discharge. Furthermore, the benefits to the families and child will include the reimbursement for transport cost after discharge and for follow up visits plus any treatment costs required during the visits.|