Menu
Menu
Menu
Menu
Background: Diarrhoea causes many deaths in children younger than 5 years and identification of risk factors for death is considered a global priority. The effectiveness of currently recommended fluid management for dehydration in routine settings has also not been examined. Methods: For this observational, association study, we analysed prospective clinical data on admission, immediate treatment, and discharge of children age 1–59 months with diarrhoea and dehydration, which were routinely collected from 13 Kenyan hospitals. We analysed participants with full datasets using multivariable mixed-effects logistic regression to assess risk factors for in-hospital death and effect of correct rehydration on early mortality (within 2 days). Findings: Between Oct 1, 2013, and Dec 1, 2016, 8562 children with diarrhoea and dehydration were admitted to hospital and eligible for inclusion in this analysis. Overall mortality was 9% (759 of 8562 participants) and case fatality was directly correlated with severity. Most children (7184 [84%] of 8562) with diarrhoea and dehydration had at least one additional diagnosis (comorbidity). Age of 12 months or younger (adjusted odds ratio [AOR] 1·71, 95% CI 1·42–2·06), female sex (1·41, 1·19–1·66), diarrhoea duration of more than 14 days (2·10, 1·42–3·12), abnormal respiratory signs (3·62, 2·95–4·44), abnormal circulatory signs (2·29, 1·89–2·77), pallor (2·15, 1·76–2·62), use of intravenous fluid (proxy for severity; 1·68, 1·41–2·00), and abnormal neurological signs (3·07, 2·54–3·70) were independently associated with in-hospital mortality across hospitals. Signs of dehydration alone were not associated with in-hospital deaths (AOR 1·08, 0·87–1·35). Correct fluid prescription significantly reduced the risk of early mortality (within 2 days) in all subgroups: abnormal respiratory signs (AOR 1·23, 0·68–2·24), abnormal circulatory signs (0·95, 0·53–1·73), pallor (1·70, 0·95–3·02), dehydration signs only (1·50, 0·79–2·88), and abnormal neurological signs (0·86, 0·51–1·48). Interpretation: Children at risk of in-hospital death are those with complex presentations rather than uncomplicated dehydration, and the prescription of recommended rehydration guidelines reduces risk of death. Strategies to optimise the delivery of recommended guidance should be accompanied by studies on the management of dehydration in children with comorbidities, the vulnerability of young girls, and the delivery of immediate care. Funding: The Wellcome Trust.
For this observational, association study, we analysed prospective routine clinical information on admission, immediate treatment, and discharge, which was collected from 13 first referral-level Kenyan hospitals that constitute the Clinical Information Network (CIN). We excluded one CIN facility from this analysis because it is contextually different from the other CIN facilities; it is staffed by non-physician clinicians (ie, non-degree trained clinicians) and is a health centre with inpatient beds rather than being a first-referral hospital. A detailed description of CIN facilities has been previously published. We screened the database for children meeting the following criteria: age 1–59 months (as guidelines only apply to this group), diarrhoea as a presenting symptom, and a full dataset available. We excluded children with only a minimal dataset or severe acute malnutrition because children with severe acute malnutrition have different fluid treatment guidelines.8 Then we identified children with diarrhoea as a presenting symptom or diagnosis or with dehydration as a diagnosis from the full dataset. Within this group, we included in our analyses only children with both diarrhoea as a presenting complaint or diagnosis and also a primary or secondary diagnosis of dehydration (some, severe, shock, or unclassified). These criteria defined a population eligible for treatment according to WHO and Kenyan guidelines for diarrhoea and dehydration. Diarrhoea was considered only a presenting symptom rather than a diagnosis in children classified as having no dehydration on the standard admission form. Data for HIV status were not recorded comprehensively in this population, but the analysis included children with a diagnosis of HIV. Maternal HIV prevalence in Kenya is 6% and the cumulative 5 year mother-to-child transmission rate is 15%; as such, we expect only a small proportion of children in the dataset to have undiagnosed HIV infection.10 The Kenya Medical Research Institute (KEMRI) Scientific and Ethical Review Committee approved the CIN study enabling use of de-identified data without individual patient consent. The hospitals use WHO and locally adapted guidance for management of common conditions.8, 9, 10, 11 In brief, these hospitals have implemented two clinical data collection tools (standard paediatric admission records and discharge forms) and have a dedicated data clerk who enters information about admission, treatment, and discharge, once the patient is discharged, into a non-proprietary electronic tool.12, 13 The clerks are trained and regularly updated on how to abstract data from medical notes and treatment sheets, including fluid prescription sheets, and on how to interpret them. Error checks are done before the data are uploaded and synchronised into a central server, in which further quality checks are done. Any discrepancies noted at this stage are raised with respective clerks who reconcile them. Periodic visits are made to the participating hospitals by the data management team who re-enter a number of randomly selected files to ascertain the accuracy of data entered by the clerks. A minimal dataset, which consists of data required for the routine health information system from all admissions (patient age, sex, diagnoses, and outcome), is collected for a random sample of otherwise eligible admissions in two high-volume hospitals (to reduce the data entry workload) and in all 13 hospitals for surgical or burns cases, admissions younger than 1 month, and admissions during periods when the single clerk is on leave. The randomisation sequence is system generated automatically and not within the clerks’ control. A full dataset on clinical presentation, diagnoses, treatments, and outcomes is collected on all other cases and at all other times. We aimed to investigate clinical signs associated with mortality and whether prescription of recommended fluid guidance is associated with a reduced risk of mortality. We aimed to examine clinical risk factors for in-hospital death and risk modification associated with intended use of WHO fluid treatment guidance in children admitted with diarrhoea and dehydration across the hospitals. Patient characteristics examined include sex, age (≤12 months or >12 months), duration of diarrhoea (≤14 days or >14 days), length of illness (≤2 days or >2 days), history of bloody diarrhoea, malaria status, and abnormal signs obtained on examination of various systems organised as airway, breathing (respiratory system), circulation, hydration status, and disability (neurological system). A child was deemed to have an abnormal system if any sign within the specific system was abnormal (see panel 1 for definitions of abnormal signs). In the case of dehydration, we also created a variable to represent children with clinical signs indicative of severe dehydration (defined in Kenyan guidelines as the presence of both sunken eyes and delayed skin pinch). Correct fluid prescription was defined based on WHO and Kenyan guidance (panel 2). Airway signs Abnormal airway signs* (only stridor analysed this study) . Circulatory signs Presence of any one or more of the following: capillary refill time greater than 2 s (delayed capillary refill time), temperature gradient (cold hands and feet), weak pulse volume, or pallor. Comorbidity Dehydration plus any of the following: malaria, pneumonia, HIV, tuberculosis, anaemia, meningitis, rickets, or asthma. Dehydration signs Presence of either delayed skin pinch (greater than 1 s) or sunken eyes, or both. Impaired circulation Presence of any one or more of the following: weak pulse volume, temperature gradient (skin temperature up to shoulder or elbow), or capillary refill time longer than 2 s. Impaired consciousness AVPU (Alert, Voice, Pain, Unresponsive) score less than A. Malaria endemic zone A hospital was regarded as located in a high malaria endemic zone if malaria diagnoses comprised more than 50% of admission diagnoses. Neurological or disability signs Presence of any one or more of the following: convulsions, neck stiffness, bulging anterior fontanelle, inability to drink or breastfeed, or impaired consciousness. Pneumonia History of cough or difficulty breathing, age older than 60 days, lower chest wall indrawing or tachypnoea-non-severe pneumonia, and any danger sign (oxygen saturation <90%, cyanosis, inability to drink or breastfeed, impaired consciousness, or grunting). Respiratory illness signs Presence of any one or more of grunting, tachypnoea, chest indrawing, acidotic breathing, crackles, or crepitations. Severe acute malnutrition Defined as clinical diagnosis of severe acute malnutrition, mid-upper arm circumference less than 11·5 cm, or weight for height Z score less than −3 standard deviations. Clinical shock Clinical diagnosis of shock made by clinician or fluid bolus given. Tachypnoea Respiratory rate more than 50 breaths per min if aged 12 months or younger, or more than 40 breaths per min if older than 12 months. WHO shock Presence of all of an AVPU score less than A, weak pulse, and capillary refill time longer than 3 s in the presence of diagnosis of dehydration. In this study, either WHO plan B, WHO plan C, or shock management7, 8 were correctly prescribed. Plan B was correct when given to children classified as having some dehydration and who had not been prescribed bolus fluid and received oral fluid (prescribed for a duration of 4 h or prescribed to be given at regular intervals) or prescribed intravenous fluid (not plan C or fluid bolus and volume <200 mL/kg for a duration of 24 h) plus oral fluid. Plan C was correct when given to children with a diagnosis of severe dehydration, who had not been prescribed bolus, and in whom oral fluid was used. Correct volume of plan C was 30 mL/kg for step 1 and 100 mL/kg for step 2. The correct duration of plan C was 6 h or less. Shock management was correct in children indicated as having shock and prescribed fluid bolus plus correct plan C. Intravenous fluid or bolus was correct if normal (0·9%) saline or Ringer's lactate was used. We did multiple imputation using chained equations to deal with missing data. Using Stata version 15.1, we did imputation with 100 iterations to produce ten imputed datasets on the assumption that co-variable data were missing at random.14, 15 The imputation model included all variables to be considered as risk factors, auxiliary variables (fever, history of vomiting, and cough), and outcomes, but we excluded any variable with greater than 30% missingness.16 We studied risk factors for overall in-hospital mortality and early (within 2 days) in-hospital mortality using mixed-effects logistic regression models, with patient-level data (level I) nested within hospitals (level II) and hospital location in malaria zone as a level II fixed effect. Univariable models (unadjusted) were fitted on the imputed datasets for each patient characteristic and malaria zone location as fixed effects, and hospital intercept as a random effect. The multivariable model (model II; adjusted) was constructed using all patient characteristics and a backward variable selection procedure with a p value for exclusion of 0·05, while maintaining the same multilevel structure as the univariable model. A priori, we decided to include age, sex, and malaria diagnosis in the adjusted model. Final model estimates were derived using Rubin rules.14 Final risk factors are variables independently associated with outcome in the multivariable model (Wald test p value <0·05). Because our analysis focuses on a population in whom we have excluded no dehydration, we also investigated the association between signs identified as risk factors in model II and mortality in a broader population of children admitted with diarrhoea using the same approach as done for those with diarrhoea and dehydration to investigate the generalisability of findings. We investigated effect modification of admission fluid prescription on risk of death in children with signs of dehydration, abnormal respiratory signs, impaired circulation, anaemia, and abnormal neurological signs, by including a binary term for correct initial fluid prescription in the model for early death (within 2 days from admission). We calculated the relative excess odds due to interaction, the attributable proportion, and the multiplicative interaction odds ratio (OR).17 Interactions were analysed in the final model (model II) one at a time. Analysis for effect modification was restricted to early deaths because we hypothesised that this is the group whose outcome might be affected by fluid management at admission. Our database collected only fluid prescribed at admission. Patients with no information on fluid management were excluded from the analysis for effect modification. Data for this report are under the primary jurisdiction of the Ministry of Health in Kenya. Enquiries about using the data can be made to the KEMRI-Wellcome Trust Research Programme Data Governance Committee. The funders of the study had no role in study design, data collection, data analysis, data interpretation, or writing of this manuscript. SA, PA, DG, AA, GI, KS, and ME had access to the raw data. The corresponding author had full access to all the data in the study and had final responsibility for the decision to submit for publication.
N/A
