Effect of an Early Resuscitation Protocol on In-hospital Mortality Among Adults With Sepsis and Hypotension: The Simplified Severe Sepsis Protocol 2 Trial

Andrews. JAMA 2017;318(13):1233-1240. doi:10.1001/jama.2017.10913

Clinical Question

• In Zambian adults with sepsis and hypotension, does a resuscitation protocol implemented early after presentation to the Emergency Department improve in-hospital mortality compared with usual care?

Background

• The implementation of sepsis protocols as the standard of care in the developed world has contributed to a reduction in sepsis mortality
• The mortality from sepsis remains high in the developing world
• It remains uncertain whether early sepsis resuscitation protocols could improve mortality from sepsis in resource-limited settings
• Patients who receive usual care may be at high risk of under-resuscitation, and therefore a time-based protocol to promote aggressive use of fluids, blood, and vasopressors may improve the quality of resuscitation and reduce mortality
• The FEAST study showed increased mortality with intravenous fluid bolus administration in African children with sepsis
• A randomised clinical trial conducted in Zambia, investigating intravenous fluids and vasopressor support in sepsis, was stopped early due to possible harm in patients with hypoxia and tachypnoea

Design

• Single-centre, randomised controlled study
• Permuted block randomisation was performed by computer
• Patients, treating clinicians, and clinical study personnel were unblinded to participant enrolment but those responsible for data analysis were blinded
• Sealed, opaque envelopes were used to preserve allocation concealment
• Power calculation
  • Statistical power (beta) set at 80% (20% chance of false negative)
  • Statistical significance level (alpha) defined at 0.05 (5% chance of false positive)
  • Assumed a baseline mortality rate of 65% (previous trial data)
  • Expected absolute risk reduction of 20%
  • 212 patients were required

Setting

• 1500 bed national referral university hospital in Zambia
• October 2012 – November 2013

Population

• Inclusion
  • Patients presenting to the Emergency Department aged 18 years or older with sepsis (defined as suspected infection plus ≥2 systemic inflammatory response syndrome criteria) and hypotension (defined as systolic blood pressure ≤90 mm Hg or mean arterial pressure ≤65 mm Hg)
• Exclusion:
  • The presence of hypoxaemia and severe tachypnoea (defined as arterial oxygen saturation <90% and respiratory rate >40 breaths per minute)
  • Additional exclusion criteria included gastrointestinal bleeding in the absence of fever, congestive heart failure exacerbation, end-stage renal disease, elevated jugular venous pressure (JVP), incarceration, or the need for immediate surgery
• 209 patients received the study interventions, completed follow-up, and were included in the primary analysis
• Both groups were similar at baseline
  • Mean age was 36.7 years
  • 89.5% of patients were HIV positive, with a median CD4 count of 66/µL
  • Median albumin level: 22 g/l
  • Median middle arm circumference: 22.1 cm
  • Median lactate level: 4.3 mmol/l
  • Diagnosis on admission:
    • Pneumonia 49.3%
    • Tuberculosis 62.7%
    • Both 38.3%

Intervention

• Patients received haemodynamic management for the first 6 hours after enrolment
  • An initial 2000 ml bolus of intravenous isotonic crystalloid was administered within 1 hour of enrolment, followed by an additional 2000 ml over the subsequent 4 hours to a total of 4 litres
    • Fluid therapy was stopped if the nurse identified any of the following:
      • Oxygen saturation decrease of 3%
      • Respiratory rate increase by 5 breaths per minute
      • JVP reached 3cm above sternal angle
      • 4 litres had been given
  • If mean arterial pressure remained less than 65 mm Hg after completion of the initial 2000 ml fluid bolus, a dopamine infusion was initiated at 10 µg/kg/min
  • Blood transfusion was administered to maintain a haemoglobin level of greater than 7 g/dL or with severe pallor

Control

• Treating clinicians determined intravenous fluid administration, vasopressor use, and blood transfusion
• During usual care for sepsis in the study setting, in the first 6 hours
  • the volume of intravenous fluid administered averages less than 2 litres
  • < 50% of patients received any intravenous fluid bolus
  • < 20% of patients received a blood transfusion
  • < 2% of patients received a vasopressor

Management common to both groups

• Treating clinicians determined the location of care (intensive care unit or medical ward) and antibiotic selection
• A dedicated study nurse measured heart rate, blood pressure, respiratory rate, and oxygen saturation hourly for the 6 hours after enrolment and supervised the administration of all ordered fluids and medications

Outcome

• Primary outcome: In-hospital mortality was higher in the sepsis protocol group
  • Sepsis protocol 48.1% vs Standard care 33.0%
    • Absolute Risk Increase (ARI): 15.1% (95% CI 2.0%-28.3%; P = 0.0344)
    • Relative Risk (RR): 1.46 (95% CI 1.04-2.05)
    • Number Needed to Harm (NNH): 7 patients
  • In the multivariate analysis adjusting for SAPS-3 at enrolment, risk of in-hospital mortality (RR 1.45 [95% CI 1.04-2.02]; P = .03) and 28-day mortality (RR 1.41 [95% CI 1.08-1.84]; P = .01) was greater in the sepsis protocol group than in the usual care group
• Secondary outcome:
  • Median hospital length of stay was 5 days (IQR 3-8 days) in the sepsis protocol group vs 7 days (IQR 4-12 days) in the usual care group (P = .01)
  • Rates of adverse events were similar between groups
  • Patients in the sepsis protocol group received more intravenous fluid (median 3.5 L [IQR 2.7-4.0 L]) compared with patients in the usual care group (median 2.0 L [IQR 1.0-2.5 L]; P < .001) and more frequently received dopamine (Sepsis protocol 14.2% vs Standard care 1.9%; P < .001)
  •  In the usual care group, only about half of the patients received any intravenous fluids (n = 50; 48.3%)
    • More patients in the sepsis protocol group experienced worsening hypoxemia or tachypnea (35.8%) compared with the usual care group (22.3%; P = .03)
    • Only 1 patient was admitted to an intensive care unit (ICU), which is not surprising given that there are only 10 ICU beds in Zambia, a country of 13 million people
  • Patients in the sepsis protocol group had a greater reduction in blood lactate levels after resuscitation (-1.2 mmol/L (-3.4 to 0.3) vs. -0.5 mmol/L (-2.2 to 1.1)]

Authors’ Conclusions

• Among adults with sepsis and hypotension, most of whom were positive for HIV, in a resource-limited setting, a protocol for early resuscitation with administration of intravenous fluids and vasopressors increased in-hospital mortality compared with usual care

Strengths

• Management of sepsis in the developing world is a very important and under-researched area so this is an important study
• Allowance was made in the study design for the potential deleterious effects of IV fluids, with hypoxic patients excluded from the study
  • The harmful effects of protocolised sepsis resuscitation may have been even greater had hypoxic patients been included
  • In addition, the study protocol accounted for the cessation of IV fluid therapy if evidence of respiratory compromise from volume overload existed
• The randomisation process ensured that both groups were similar at baseline
• The allocation concealment process was robust, with sealed envelopes and permuted block randomisation by computer
• All analyses were conducted in an intention to treat basis
• A conventional blood transfusion trigger of 70 g/dl was used
• Clear between-group separation was achieved with respect to IV fluid administration in the first six and twenty four hours (median difference 1.5 and 1.0 litres respectively; p < 0.001)
• A clear difference was also found in dopamine administration between the groups (14.2 vs. 1.9%; p < 0.001)
• The study results were consistent with previous randomised controlled trials within the developing world setting
• The primary outcome of mortality at 28 days was relevant and patient-centred
• An adjusted analysis for SAPS-3 score was performed
• The mortality difference persisted in the multivariate analysis adjusting for severity of illness and across pre-specified subgroups defined by preexisting HIV infection and baseline Glasgow Coma Scale score, haemoglobin level, lactate level, severity of illness, and jugular venous pressure
• No patients were lost to follow-up for primary outcome analysis
  • 3 patients were excluded after randomisation and were therefore not included in primary outcome analysis (leading to a modified Intention To Treat analysis)

Weaknesses

• There is limited generalisability to the developed world as most patients were young, malnourished and HIV positive (with median CD4 counts of 66u/L)
  • In addition, Intensive Care support for deteriorating patients was unavailable due to resource constraints
• This was a single-centre study in which research staff and treating clinicians were unblinded to the allocation groups
  • This may have led to clinician bias and a difference in uncontrolled interventions between the study groups
• Convenience sampling was used, with participants being recruited between Monday morning and midday on Friday
• The inclusion criteria for the study was based upon the traditional Surviving Sepsis Campaign/ American College of Respiratory Physicians criteria for septic shock
  • These criteria were based on the clinical characteristics of comparatively older patients, most of whom had chronic health problems
• Usual care involved a dedicated study nurse measuring clinical observations hourly and supervising the administration of all ordered fluids and medications
  • It is unclear whether this is normal practice in a sub-Saharan African hospital
• Young, malnourished patients in the developing world may have been hypotensive at baseline rather than this being as a result of sepsis-induced vasoplegia
• Patients were included with septic shock on the basis of a systolic blood pressure of < 90mmHg or a MAP of < 65mmHg
  • This differs from the recent EGDT studies and SSC criteria which would mandate fluid resuscitation prior to this diagnosis
• Patients in the sepsis protocol group received 2 litres of IV fluids in the first hour
  • A significant number of these patients were malnourished, and this would have been significantly more than 30 mls/kg
  • Almost 40% of patients in the treatment group received 4 litres or more of IV fluid therapy by six hours after study enrollment
• Due mainly to resource limitations, intravenous fluids were not titrated to haemodynamic parameters
  • They were instead discontinued when patients showed evidence of respiratory compromise
  • Administering IV fluids until a patient is clinically volume overloaded is unsophisticated and at variance with developed world practice
• On average, patients in the usual care group received IV antibiotics thirty minutes earlier, however this was a non-significant difference (p = 0.15)
• Dopamine was used for vasopressor support and this is another intervention that limits generalisability to the developed world
• Mortality outcomes were limited to 28 days

The Bottom Line

• Consistent with previous studies in this area, protocolised intravenous fluids and dopamine appears to be harmful in African patients presenting with severe sepsis
• This should be abandoned as a treatment strategy in this setting
• Due to significant differences in patient characteristics and hospital resources, these results cannot be translated to the developed world

External Links

• [article]Effect of an Early Resuscitation Protocol on In-hospital Mortality Among Adults With Sepsis and Hypotension: A Randomized Clinical Trial
• [further reading] The FEAST Trial summary
• [presentation] The SSSP2 trial

Metadata

Summary author: Fraser Magee
Summary date: 5 November 2017
Peer-review editor: Steve Mathieu and Duncan Chambler