RESUMO
OBJECTIVES: To evaluate the differences between the observed hemoglobin levels and those estimated based on hematocrit in the context of the 2nd National Study of Human Growth and Development of the Venezuelan Population (SENACREDH). MATERIALS AND METHODS: 6,004 individuals were chosen by a probabilistic multistage cluster sampling representing 7,286,781 inhabitants from North Central Coastal area (Vargas, Carabobo, Capital District, Aragua and Miranda). Means of observed and estimated hemoglobin (hematocrit/3) were compared, using t test for related samples and linear regression. RESULTS: Mean difference between the values of observed and estimated hemoglobin was -0.3446 ±0.0002 (p<0.001); significantly overestimating the hemoglobin values. Regression models of hemoglobin on hematocrit showed an r2=0,87. In order to correct the estimation, we propose a new formula for calculating hemoglobin based on haematocrit values: estimated hemoglobin=(Haematocrit/3.135)+ 0.257. CONCLUSIONS: There is an overestimation of hemoglobin levels from hematocrit levels and therefore an underestimation of the prevalence of anemia; however, a high positive correlation between them was found, allowing modeling for achieving a better estimation of the hemoglobin from the hematocrit value.
Assuntos
Anemia/sangue , Anemia/diagnóstico , Hematócrito , Hemoglobinas/análise , Adolescente , Adulto , Criança , Pré-Escolar , Feminino , Humanos , Lactente , Masculino , Pessoa de Meia-Idade , Venezuela , Adulto JovemRESUMO
Objetivos. Evaluar las diferencias entre el valor de hemoglobina observada y el valor estimado a partir del hematocrito en el marco del Segundo Estudio Nacional de Crecimiento y Desarrollo Humano de la Población Venezolana (SENACREDH) en el eje centro norte costero del país. Materiales y métodos. Por medio de un muestreo probabilístico multietápico por conglomerados se seleccionó un total de 6004 sujetos que representan 7 286 781 habitantes del eje Centro Norte Costero (Vargas, Carabobo, Distrito Capital, Aragua y Miranda). Se compararon medias de la hemoglobina observada y hemoglobina estimada (hematocrito/3), usando la prueba t para muestras relacionadas. Se realizaron regresiones lineales entre hemoglobina observada y hematocrito. Resultados. Se observó que el promedio de las diferencias entre la asignadas a la hemoglobina observada y la estimada por el hematocrito fue de -0,3446 ± 0,0002 (p<0,001); sobreestimando significativamente los valores de hemoglobina. Modelos de regresión del hematocrito para la hemoglobina arrojan un valor de r2=0,87. Al corregir la estimación, se propone una nueva forma de calcularla, con la siguiente fórmula: hemoglobina estimada = (hematocrito/3,135) + 0,257. Conclusiones. Existe una sobreestimación de hemoglobina y, por tanto, subestimación de la prevalencia de anemia a partir del hematocrito; sin embargo, se encuentra una alta correlación entre ambos componentes que permite modelar una mejor estimación de la hemoglobina a partir del hematocrito.
Objectives. To evaluate the differences between the observed hemoglobin levels and those estimated based on hematocrit in the context of the 2nd National Study of Human Growth and Development of the Venezuelan Population (SENACREDH). Materials and methods. 6,004 individuals were chosen by a probabilistic multistage cluster sampling representing 7,286,781 inhabitants from North Central Coastal area (Vargas, Carabobo, Capital District, Aragua and Miranda). Means of observed and estimated hemoglobin (hematocrit/3) were compared, using t test for related samples and linear regression. Results. Mean difference between the values of observed and estimated hemoglobin was -0.3446 ±0.0002 (p<0.001); significantly overestimating the hemoglobin values. Regression models of hemoglobin on hematocrit showed an r2=0,87. In order to correct the estimation, we propose a new formula for calculating hemoglobin based on haematocrit values: estimated hemoglobin=(Haematocrit/3.135)+ 0.257. Conclusions: There is an overestimation of hemoglobin levels from hematocrit levels and therefore an underestimation of the prevalence of anemia; however, a high positive correlation between them was found, allowing modeling for achieving a better estimation of the hemoglobin from the hematocrit value.
