Subjects and methods
This cross-sectional study was conducted between August 2017 and June 2018. We invited 1,516 children residents in Quito, aged between 1 month and 12 years old, who sought preventive medical attention at three primary health-care centers (Lucha de los Pobres Health-care Center, Cotocollao Health-care Center and Clínica Pichincha), three elementary schools (Escuela Francisco Salazar, Centro del Muchacho Trabajador Cotocollao and Centro del Muchacho Trabajador La Marín) and nine kindergarten municipal schools (Abdón Calderón, Andalucía, Carapungo, Colibrí, Cotocollao, Empleados Municipales, Ipiales, La Carolina and Santa Clara) to participate. Non-probabilistic and convenience sampling were used because age and gender distribution at the schools and health-care centers was unknown. The centers are located at an altitude between 2,740 and 2,901 m a.s.l (average 2,810 m a.s.l). Ambient temperature throughout the study was, on average, 14.4 oC (11.5- 20.8 oC) and humidity average was 72.2% (52 - 81%) as reported by Ecuador’s National Institute of Meteorology and Hydrology[12].
We included children who resided in the city at least 2 months before the study, similar criteria used in other study [13], and children younger than 2 months must have been born and remain living in the city until they were examined. Exclusion criteria included a registered axillary temperature >37.5 C° at the time of evaluation, history of respiratory symptoms in the two weeks prior to evaluation, any abnormal cardio-respiratory signs during physical examination, history of chronic cardio-respiratory disease, history of neonatal respiratory disease, history of blood component transfusion in the six months prior to evaluation, and the presence of malnutrition, defined as a Z-score less than -2SD for either height for age or weight for height [9, 10].
Children were enrolled in the study after written informed consent obtained from their parents. The study was approved by the Universidad Internacional del Ecuador Ethics Committee, registered code CEU-005-16 and by the health committee at each center participating in the study. Information on the study was provided to the directive councils and medical teams at each institution.
Fifteen students from fourth year of a school of medicine were rigorously trained in anthropometric measurements and pulse oximetry assessment by the standardization method of the Central America and Panama Institute of Nutrition (INCAP) [14]. To manage measurement bias, the students’ measurements were compared against a pediatrician’s reference pattern, establishing a maximum margin of error of 0.2 kg for weight and 0.5 cm for length and height [14].
Variable definition
Weight and height were measured using high fidelity equipment (Health-o-Meter 498KL and 593KL, USA), regulated and previously calibrated by the Ecuadorian Institute of Normalization. Respiratory frequency was obtained through observation in calm and alert children, visually counting thoracic and abdominal movements over one minute. Temperature was assessed with a flexible digital thermometer (Omron MC-343F, Mexico), placed in children’s armpit until a reading signal was obtained. Heart rate and SpO2 were evaluated using automatically calibrated non-invasive pulse oximeters (Huntleigh MP1R Smartsigns® MiniPulse Huntleigh Healthcare Ltd, Cardiff, United Kingdom). The pulse oximeter used measured functional oxygen saturation with a precision range of ±2%. Pulse oximetry was assessed in calm and alert children. Wrap-around style and fold-over-style probes were used, depending on the subject’s age, and placed either on right hand’s index finger or the big toe for infants. Nail polish remover was provided for subjects who had nail polish present at the time of the test. The SpO2 measurements were considered adequate when a plethysmographic waveforms of perfusion levelled-off on its high end and remained on the output screen for at least 2 minutes. Then, SpO2 measurements and pulse rate, were recorded every 10 seconds for a total of three measurements, and the average was used to determine SpO2 for each study subject [6, 15–17].
Statistical Analysis
Descriptive statistical tests were run for clinical measurements, and 2.5th percentile, 5th percentile, 25th percentile (Q1) and 75th percentile (Q3) for SpO2 distribution. The Kruskal-Wallis test was used to compare differences in SpO2 medians by age groups, Mann-Whitney U test was used to compare medians between males and females. Statistical significance was accepted with p<0.05. Smooth lines were designed for percentiles 2.5th and 5th for SpO2 using the Spline method (smooth.spline function in R, with a 7 degree freedom range). All data was registered in the digital survey platform Survey Monkey®, and analyses were performed using SPSS®, version 24. Graphics were designed using R version 4.3.