Introduction - Gastrointestinal tonometry has been widely used in adult practice for the early detection of shock and multi-organ failure. Its application in paediatrics has been limited by unsuitably large tonometers and doubt about the accuracy of measurements when saline is used as a tonometric fluid / vehicle for carbon dioxide (CO₂) equilibration. Objective - To evaluate the accuracy and reliability of the newly developed saline 5 French (5F) neonatal gastric tonometer. Study Design - (a) Direct in-vivo comparison of the 5F 0.9%saline tonometer (NST) with the recirculating gas tonometer (RGT) [the current reference standard in adult practice] in 10 Paediatric intensive care unit (PICU) patients, measuring tonometric PCO₂ (PtCO₂) and gastric intramucosal PCO₂ (PiCO₂). (b) In-vivo comparison of PiCO₂ measurements from two 5F tonometers in 10 PICU patients in unfed and fed state. (c) In-vitro comparison of reference PCO₂ to PtCO₂ values obtained using 0.9%saline and phosphate buffered saline in SF tonometers, and the RGT. Results - (a) Comparing the SF NST to RGT in 50 paired simultaneous measurements over PtCO₂ range 3.0 - 9.7kPa, the mean bias was -1.44kPa; limits of agreements (LOA) ±1.45kPa. The mean values of PtCO₂- derived gastric intramucosal pH (pHi) and PiCO₂-PaCO₂ difference differed significantly by -.11 and + 1.1kPa respectively (p<0.0001). (b) 100 paired 5F NST measurements (50 fed/ 50 unfed) over PtCO₂ range 2.48-11.1kPa were assessed. No significant difference was observed in PtCO₂: mean difference (standard deviation) - unfed 0.05kPa (0.36) (p=0.36); fed 0.05kPa (0.42) (p=0.43). (c) 20 consecutive measurements of PtCO₂ were obtained from the 5F NST, 5F phosphate buffered saline tonometer (PBST) and RGT at constant reference PCOi's of 2.5, 5.0, 7.5, 10.0kPa. The 5F NST underestimated the reference PCO₂ by a mean bias of 58% (LOA ±20%); the 5F PBST by 6% (LOA ±26%); while the RGT performed best with a mean bias of 5.7% and tight LOA ±1.5%. Conclusion - There are inherent problems in the methodology of the saline tonometry utilised in the 5F neonatal gastric tonometer. The use of the saline SF neonatal gastric tonometer to monitor gut perfusion in neonates and children should be interpreted with caution. Recirculating gas tonometry is the most accurate method of tonometry studied.

A research report submitted to the Faculty of Health Sciences, University of the Witwatersrand, Johannesburg in partial fulfillment of the requirements for the degree of Master of Medicine in the branch of Paediatrics and Child Health 29 May 2017 / Introduction Sepsis, and in particular, severe sepsis, remains a major cause of death in children worldwide. One of the areas where the burden of sepsis is keenly felt is in the paediatric intensive care unit (PICU) setting, contributing significantly to childhood mortality. Fungal organisms have emerged as a major organism contributing to nosocomial sepsis in PICU. No local data regarding nosocomial fungal sepsis in the non-neonatal, PICU population exists regarding this matter. This study describes the characteristics of patients with nosocomial fungal sepsis in the PICU at South Africa’s largest hospital Chris Hani Baragwanath Academic Hospital (CHBAH). Methods This study was a retrospective review of patient records. All patients aged 0-16 years admitted to the PICU at Chris Hani Baragwanath Academic Hospital (CHBAH) from January 2008 through December 2011 were assessed. A total of seventeen patients who developed nosocomial fungal sepsis were included in this study. Results The incidence of candidaemia was reported to be 3.2 per 100 cases. The major age group affected by nosocomial fungal sepsis was the under one age group. The most common diagnoses on admission were lower respiratory tract infection (LRTI) followed by haematology-oncology and acute gastroenteritis cases. ICU factors found to commonly co-­‐exist with proven nosocomial fungal sepsis were presence of a central venous catheter (100%), mechanical ventilation (82%), arterial line (70%), and systemic corticosteroid use (47%). The penicillin class was the most common antimicrobial that patients were found to be on at the time of nosocomial sepsis. The most common fungal organism as a cause for nosocomial sepsis was C. parapsilosis rather than C. albicans. Furthermore, the majority of this study’s isolates were susceptible to voriconazole rather the current empiric antifungal of choice, namely fluconazole. Conclusion The presence of central venous catheters, arterial lines, mechanical ventilation and systemic corticosteroid use is common in paediatric patients with nosocomial fungal sepsis. However, this study was unable to determine statistically significant factors associated with fungal sepsis in a tertiary PICU due to the surprisingly small number of cases (n=35) detected over a four-year period. This perhaps represents the most striking finding of the study together with a concerning pattern of fluconazole resistance (14%) among isolated organisms. / MT2017

Objectives: To describe the characteristics of children who died and their modes of dying in a South African Paediatric Intensive Care Unit (PICU). Design: Retrospective review of data extracted from the Child Healthcare Problem Identification Programme (Child PIP)and the PICU summary system (admission and death records) on children of any age who died in the PICU between 01 January 2013 and 31 December 2017. Setting: Single-centre tertiary institution. Patients: All children who died during PICU admission were included. Measurements and Main Results: Four-hundred and fifty-one (54% male; median (IQR) age 7 (1-30) months) patients died in PICU on median (IQR) 3 (1-7) days after PICU admission; 103 (22.8%) had a cardiac arrest prior to PICU admission. Mode of death in 23.7% (n=107) was withdrawal of life sustaining therapies; 36.1% (n=163) died after limitation of life sustaining therapies; 22.0% (n=99) died after failed resuscitation and 17.3% (n=78) were diagnosed brain dead. Ultimately, 270 (60%) children died after the decision to limit or withdraw life sustaining therapies. There was no difference in the number of deaths during office and after-hours periods (45.5% vs. 54%; p = 0.07). Severe sepsis (21.9%) was the most common condition associated with death, followed by cardiac disease (18.6%).Ninety-four (20.8%) patients were readmitted to the PICU within the same year; 278 (61.6%) had complex chronic disorders. During the last phase of life, 75.0% (n=342) were on inotropes, 95.9% (n=428) were ventilated, 12.0% (n=45) received inhaled nitric oxide and 10.8% (n=46) renal replacement therapy. Only 1.5% (n=7) of children became organ donors and postmortems were done in 47.2% (n=213) of the patients. Conclusions: Most PICU deaths occurred after a decision to limit or withdraw life-sustaining therapy. Severe sepsis was the most common condition associated with death. Referral for organ donation was extremely rare.