Hypoglycaemia at presentation and mortality in a pediatric intensive care unit, Koutiala, Mali

Abstract

International guidelines for critical care in low-resource environments often assume little or no diagnostic capability. This is not always the case in MSF field settings. Optimizing use of diagnostic tools, even simple ones, can potentially improve quality of care. Measuring glycaemia and correcting hypoglycaemia are both simple interventions...

Full text

Hypoglycaemia at presentation and mortality in a
pediatric intensive care unit, Koutiala, Mali
Coldiron M.1, Sagara J.2, Mambula C.3, Umphrey L.4, Schaefer M.4, Grais R.F.1
•International guidelines for critical care in low-resource
environments often assume little or no diagnostic capability
•This is not always the case in MSF field settings
•Optimizing use of diagnostic tools, even simple ones, can
potentially improve quality of care
•Measuring glycaemia and correcting hypoglycaemia are both
simple interventions
METHODS
DISCUSSION AND CONCLUSION
1Epicentre, Paris, France; 2Médecins Sans Frontières, Koutiala, Mali;
3 Médecins Sans Frontières, Paris, France; 4 Médecins Sans Frontières, Sydney, Australia
•Hypoglycaemia is extremely common in this setting and is
associated with very high in-hospital mortality
•Even “moderate” hypoglycemia is associated with a high case
fatality rate
•Standard practice has changed at CSRef Koutiala in light of these
results: all patients admitted to ICU are systematically given
dextrose-containing fluids on admission
RESULTS
INTRODUCTION
OBJECTIVES
•Describe the prevalence of hypoglycaemia and severe
hypoglycaemia on admission to the ICU
•Describe the association of hypoglycaemia with negative clinical
outcomes
Setting
•Reference Health Center of Koutiala, Mali
•More appropriately called a hospital, supported by MSF since
2009, with general paediatrics wards, neonatology unit,
inpatient nutritional rehabilitation unit
•Hosts a suite of 3 intensive care units: a resuscitation room, a
nutritional rehabilitation ICU and a paediatrics ICU
•>11 500 admissions to hospital among children <5yr in 2015
Data collection
•Prospective data collection on all patients admitted to one of
the three ICUs between July 2015 and June 2016
•Data extracted from patient charts during first 24 hours of ICU
stay, at time of discharge from ICU, and at hospital discharge
•Vital signs and clinical symptoms over time
•Anthropometrics
•Physical exam findings
•Admission and discharge diagnoses
•Laboratory data (cultures, CBC, point-of-care haemoglobin
and malaria tests, analysis of CSF)
•Treatments received in first 24 hours
Data analysis
•All data anonymized and described using Stata
•Log-binomial regression used to calculate risk ratios associated
with hypoglycaemia
Analysis of routinely collected data not requiring ERB review.
•3565 children admitted to ICUs over the course of the year
(Figure 1, Table 1)
Pediatric ICU
(N=2253)
Nutr. ICU
(N=752)
Resuscitation
(N=560)
n % n % n %
Sex
Male 1193 53 355 47 298 53
Female 1060 47 397 53 262 47
Age
(months)
2-11 368 16 167 22 95 17
12-23 274 12 224 30 107 19
24-35 519 23 185 25 133 24
36-47 626 28 113 15 125 22
48-59 462 21 63 8100 18
≥60 4 0.2
Time
between onset of illness and presentation to hospital
Same day 173 840 634 6
Next day 235 11 27 451 9
2-3 days 930 42 180 25 211 39
4-6 days 546 25 140 19 148 27
≥7 days 328 15 347 47 101 19
Serum
glucose on admission (mg/dl)
<60 357 16 136 19 220 40
<20 158 756 8162 29
Table 1: Demographic characteristics of patients
•Overall, 20% of children presented with glucose <60 mg/dl
and 11% with glucose <20 mg/dl
•144/376 (38%) children presenting with glucose <20 mg/dl
died in hospital
•76/337 (23%) of children presenting with glucose 20-59 mg/dl
died in hosptial
•After controlling for age, sex, and time since onset of illness,
the risk of in-hospital mortality was significantly higher in
children presenting with hypoglycaemia:
•Glucose <20: Risk Ratio 3.7 [95%CI 3.1-4.3], p<0.001
•Glucose <60: Risk Ratio 3.3 [95%CI 2.8-3.0], p<0.001
1117