Alimentación complementaria a los 4 versus 6 meses de ...³n_complemen… · alimentación...

Alimentación complementaria a los 4 versus 6 meses de edad para

niños prematuros nacidos con menos de 34 semanas de gestación:

un trabajo randomizado, ensayo abierto, multicéntrico Shuchita Gupta, Ramesh Agarwal, Kailash Chandra Aggarwal, Harish Chellani, Anil Duggal, Sugandha Arya, Sunita

Bhatia, Mari Jeeva Sankar, Vishnubhatla Sreenivas, Vandana Jain, Arun Kumar Gupta, Ashok K Deorari, Vinod K

Paul, e investigadores del estudio CF

Resumen

Fundamento: Hay escasa evidencia acerca del momento óptimo para inciar la alimentación

complementaria en infantes prematuros. Examinamos el efecto de la iniciación de la

alimentación complementaria a los 4 meses versus los 6 meses de edad corregida en el peso

para la edad a los 12 meses de dad corregida en infantes prematuros nacidos con EG<34

semanas.

Métodos: en este trabajo abierto, randomizado, incluimos infantes nacidos < 34 semanas de EG

sin malformaciones mayores de tres hospitales públicos en India. Los infantes elegibles fueron

seguidos desde el nacimiento y asignados randomizadamente (1:1) a los 4 meses de edad

corregida (EC) para recibir alimentación complementaria a los 4 meses de EC (Grupo 4 m), o

continuar con alimentación con leche e iniciación de la alimentación complementaria a los 6

meses de EC (Grupo 6 m), empleando un esquema de randomización de tamaño de bloque

variable, generado por computadora, estratificado por EG (≤30 semanas, y 31-33 semanas). Se

suplementó hierro como estándar. Los participantes y el equipo de implementación no podían

estar enmascarados para el grupo de asignación, pero si los fueron los investigadores de

resultados. El resultado primario fue score Z Peso/edad a los 12 meses de EC (WAZ12) basado en

los estándares de crecimiento del Estudio de Referencia Multicéntrico de Crecimiento de la

OMS. Los análisis se hicieron según intención de tratar. El estudio está registrado con Clinical

Trials Registry of India, CTRI/2012/11/003149.

Hallazgos: Entre el 20 de Marzo, 2013 y 24 de Abril, 2015, 403 infantes fueron asignados

randomizadamente: 206 para recibir alimentación complementaria desde los 4 meses y 197 para

recibir alimentación complementaria desde los 6 meses. 22 infantes en el Grupo 4 m (4

fallecidos, 2 abandonos, 16 perdidos al seguimiento) y 8 infantes en el Grupo 6 m (2 fallecidos, 6

perdidos al seguimiento) fueron excluidos para el análisis del resultado primario. No hubo

diferencia en WAZ12 entre ambos grupos: -1.6 (DS 1.2) en el Grupo 4 m versus -1.6 (DS 1.3) en el

Grupo 6 m (diferencia media 0.005, IC 95% -0.24 a 0.25; p=0.965). Hubo más admisiones

hospitalarias en el Grupo 4 m comparado con el Grupo 6 m; 2.5 episodios por 100 meses-

infante en el Grupo 4 m versus 1.4 episodios por 100 meses-infante en el Grupo 6 m (relación

de tasa de incidencia 1.8, IC 95% 1.0-3.1, p=0.03). 34/188 (18%) infantes en el Grupo 4 m

requirieron admisión hospitalaria, comparado con 18/192 (9%) infantes en el Grupo 6 m.

Interpretación: Pese a que no hubo evidencia de efecto para el resultado primario WAZ12 , la

tasa de ingreso hospitalario más elevada en el Grupo 4 m sugiere una recomendación de iniciar

la alimentación complementaria a los 6 meses en lugar de 4 meses de EC en niños <34 semanas

de EG.

Introducción

La alimentación exclusiva con leche materna durante 6 meses seguida de alimentación

complementaria para infantes de término, es la recomendación estándar de la OMS,

ampliamente apoyada y aceptada por la comunidad global. Sin embargo, ninguna organización

incluida OMS (1), ESPGHAN (2), o la Academia Americana de Pediatría (3) proveen guías basadas

en la evidencia con respecto al momento óptimo de iniciación de la alimentación

complementaria en infantes prematuros que se encuentran en un riesgo muy superior de

restricción del crecimiento postnatal (4, 5). Las únicas guías disponibles vienen del Reino Unido,

y están basadas en una sola revisión no-sistemática de estudios primariamente fisiológicos.

Contexto de investigación Evidencia antes de este estudio Buscamos en MEDLINE y en Cochrane Central Register of Controlled Trials (CENTRAL) hasta el 6 de Febrero de 2016, limitando a estudios humanos, sin restricción de lenguaje o fecha de publicación, empleando términos “alimento complementario” O “alimentos complementarios” O “alimentación complementaria” O “alimentación suplementaria” O “destete” O “alimentos para destete” O “dieta suplementaria” O “semisólidos” O “comidas semisólidas” O “comida semisólida” O “alimentaciones semisólidas” O “alimentación infantil” O “dieta infantil” O “comida de infante” O “comidas de infante”. Encontramos 22723 artículos en MEDLINE y 3667 en CENTRAL. Después de filtrar título y abstracts y excluir duplicados, revisamos texto completo de 50 estudios posiblemente relevantes. Criterios de inclusión para la revisión: todo estudio randomizado o quasi- randomizado u observacional que investigue diferente momento o edad de iniciación o introducción de alimentación complementaria en infantes prematuros. Se definió alimentación complementaria como la iniciación o introducción de comida semisólida, ligera o sólida diferente a leche materna, fórmula o animal. Por tanto, excluimos los estudios que incluían infantes de término o aquellos donde el término “destete” fue empleado para indicar transición de seno materno a leche de fórmula o animal en lugar de comidas semisólidas ligeras o sólidas. Identificamos tres estudios elegibles para esta revisión- un estudio randomizado, uno prospectivo observacional y otro que era un análisis secundario de datos de estudios originales no relacionados. El primer estudio asignó randomizadamente todos los infantes prematuros <37 sem EG (n=68) para recibir alimentación complementaria después de las 13 semanas de edad si pesaban al menos 3.5 kg; o después de las 17 sem de edad si `pesaban al menos 5 kg, La intervención, iniciación de alimentación complementaria en ambos grupos, fue también sujeta a la percepción de parental de que el bebé estuviera listo para aceptarla. El Grupo 13 sem también recibió una co-intervención en forma de energía y alimento concentrado en proteína. El estudio no encontró ninguna diferencia significativa en los DS de longitud (0·2, 0·2 vs –0·1, 0·3), peso (–0·7, 0·2 vs –0·8, 0·2) y circunferencia craneana (–0·5, 0·2 vs –0·5, 0·2) entre ambos grupos. Sin embargo, hubo una mejora en la ganancia de longitud semanal desde el nacimiento hasta los 12 meses de EC (cm media por semana, DS: 5·1, 0·07 vs 4·9, 0·10; p=0·04) y cambio en los DS de media de longitud en el grupo de destete temprano (media, DS: –1·1, 0·2 al nacer hasta 0·2, 0·2 a los 12 meses EC) al comparar con el grupo destete tardío (media, DS: –1·0, 0·2 al nacer hasta –0·1, 0·3 a los 12 meses EC).

El segundo estudio siguió infantes prematuros desde el nacimiento hasta los 12 meses de EC, y registró el momento de iniciación de la alimentación complementaria basados en la información de la introducción de 12 alimentos pre-registrados. La introducción de 4 o más ítems (n=203 infantes) comparados con menos de 4 ítems (n= 54 infantes) a las 17 sem EC estuvo asociada a riesgo 3-5 veces superior de desarrollar eccema a los 12 meses EC. El tercer estudio fue un análisis secundario de la introducción de alimentos sólidos en ≤12 sem o >12 sem en los parámetros de crecimiento de infantes prematuros, de dos trabajos previos- uno sobre fórmula post-alta vs fórmula de término (publicado), y otro no publicado. El estudio no encontró ninguna diferencia en cuanto a la introducción de alimentos sólidos ≤12 sem o >12 sem sobre la ganancia de peso en Kg (media, SE: 8·25, 0·05; n=365 vs 8·27, 0·07; n=102), longitud en cm (70·3, 0·16; n=362 vs 70·6, 0·2; n=102) y CC en cm (45·8, 0·08; n=364 vs 45·8, 0·12; n=102) entre las 12 sem y los 9 meses EC, o entre 12 semanas y 18 meses EC (datos no reportados). No hubo diferencia en la prevalencia de atopía, infección respiratoria baja, gastroenteritis, duración de sueño, o despertar en la noche entre ambos grupos.

Valor agregado de este estudio Nuestro estudio provee evidencia nivel-1 respecto al momento (4 vs 6 meses EC) para la iniciación de la alimentación complementaria en infantes prematuros < 34 sem EG, un grupo muy susceptible a la restricción de crecimiento postnatal. Muestra que la iniciación temprana de alimentación complementaria a los 4 vs 6 meses de EC no mejora el crecimiento de los 12meses EC. No resulta en diferencia en los resultados del neurodesarrollo, composición corporal, mineralización ósea, y otros marcadores de síndrome metabólico como insulino-resistencia, perfil lipídico, y presión sanguínea en la infancia. En cambio, la iniciación temprana a los 4 meses EC aumenta el riesgo de ingreso hospitalario debido a enfermedades concurrentes, predominando diarrea e infecciones respiratorias bajas. En ambos grupos, los patrones dietarios persisten pobres y los depósitos de hierro deplecionados pese a suplementar hasta 12 meses de EC Interpretación: nuestro estudio sugiere que debiera preferirse los 6 meses EC para la iniciación de la alimentación complementaria en PT <34 EG. El hallazgo de deficiencia de Fe y pobres patrones dietarios deberán guiar la investigación futura.

Extrapolar la recomendación de iniciar la alimentación complementaria a los 6 meses de edad

en niños de término a infantes prematuros depende de dos grandes cuestiones: ¿a qué se

refiere “6 meses” en un infante prematuro- edad cronológica (postnatal) o edad corregida? Los

proveedores de cuidado de salud generalmente emplean EC para monitorear el crecimiento

físico y el desarrollo de los infantes prematuros. Segundo, si asumimos que “6 meses” se refiere

a EC, ¿no debieran estos infantes iniciar la alimentación complementaria más temprano que sus

pares de término- por ejemplo, a los 4 meses de EC en lugar de los 6 meses de EC (apéndice p

1)?. Los infantes prematuros tienen mayores requerimientos de energía comparados con los

infantes de término (6, 7), y no se sabe durante cuánto tiempo sólo la lactancia (leche materna o

fórmula) es suficiente para cubrir estos requerimientos. La mayoría de las comidas

complementarias proveen mayor densidad calórica comparada con la leche, y puede cubrir la

brecha de energía entre los requerimientos aumentados del infante prematuro, y la provisión

limitada de la alimentación láctea. Por tanto, una introducción más temprana en infantes

prematuros de la recomendada para niños de término, de la alimentación complementaria

podría ayudar a mejorar su crecimiento.

Hicimos una revisión sistemática para responder a esta pregunta crucial pero encontramos sólo

un trabajo randomizado de baja calidad investigando este aspecto (8). Por lo tanto elegimos

testear la hipótesis de que la iniciación de la alimentación complementaria (definida como

alimentos semisólidos, blandos, o sólidos diferentes a leche humana, fórmula o leche animal) a

los 4 meses comparado con los 6 meses EC en infantes <34 sem EG aumentarán su score Z de

peso para la edad a los 12 meses EC (WAZ12) en 0.5 de score de desviación estándar (SDS;

alrededor de 500 g, basado en los estándares del Estudio de Referencia Multicéntrico de OMS-

MGRS) (9). También investigamos neurodesarrollo, composición corporal, densidad ósea, y

marcadores tempranos de síndrome metabólico por haber poca evidencia en los nexos entre la

alimentación del infante y la salud futura, y buscamos caracterizar potenciales efectos de la

intervención nutricional en algunos marcadores.

Métodos

Diseño del estudio

Realizamos este trabajo abierto, randomizado, de grupo paralelo en tres hospitales públicos; el

All India Institute of Medical Sciences (AIIMS), el Hospital Safdarjung asociado al Vardhman

Mahavir Medical College, y el Hospital Katurba en Nueva Delhi, India. Los tres lugares proveen

cuidado neonatal terciario a neonatos nacidos en la institución. Cuidado Intensivo Neonatal es

provisto a todos los infantes nacidos <34 sem de EG, y se promueve activamente la

alimentación exclusiva al pecho materno (apéndice p 2). Se obtuvo la aprobación ética de los

comités institucionales de ética de los tres sitios. El estudio está registrado con Clinical Trials

Registry of India, CTRI/2012/11/003149.

Participantes

Fueron elegibles infantes EG <34 sem sin malformaciones mayores, residentes dentro de 60 km

de los hospitales del estudio y que no tuvieran previsto mudarse dentro del año de vida. Los

infantes que requirieron internación desde el nacimiento hasta más allá de las 40 semanas de

edad postmenstrual fueron excluidos. El equipo investigador identificó a los infantes

potencialmente elegibles al nacer, mantuvo su seguimiento después del alta y llamó a su

cuidador telefónicamente a los 4 meses de EC para incluirlo en el estudio. Los infantes fueron

excluidos en esta etapa si ya habían iniciado la alimentación complementaria. La iniciación de la

alimentación complementaria fue definida como iniciación intencional de alimentación

complementaria por la familia sin importar la cantidad, iniciación no intencional pero con aporte

de comida durante más de 3 días, o cuando la cantidad total provista representaba más del 1%

de la recomendación dietaria de calorías para el infante. Consentimiento informado escrito se

obtuvo de los padres o representante legal.

Randomización y enmascaramiento

Los participantes fueron asignados en forma randomizada (1:1) al Grupo 4 m, donde los padres

o familia eran aconsejados para iniciar la alimentación complementaria a los 4 meses de EC; o al

Grupo 6 m, en el cual ellos eran aconsejados continuar con alimentación láctea hasta los 6

meses de EC, seguidos por la iniciación de alimentación complementaria a esa edad. Una

persona independiente (MJS) proveyó las secuencias randomizadas generadas por

computadora, estratificadas por lugar y gestación (30 sem o < y 31-33 sem) con un tamaño

variable desconocido para el equipo investigador involucrado en la implementación. La

asignación fue ocultada empleando sobres sellados, opacos, numerados secuencialmente. Las

familias y el equipo involucrado en la implementación no podían estar cegados a la ubicación

de los grupos luego de la asignación, sin embargo, aquellos determinando los resultados

primarios y secundarios estuvieron cegados. Los mellizos y trillizos fueron asignados al mismo

grupo.

Procedimientos

Las descripciones detalladas de los procesos del estudio pueden encontrarse en el apéndice (p

3). A los 4 meses de EC, asignamos randomizadamente los infantes e hicimos las mediciones

antropométricas basales y al Absorciómetro Dual de Energía por Rayos- X (DXA).

Una sola persona aconsejó a las familias para la iniciación de la alimentación complementaria a

la EC de 4 m o 6 m empleando instrucciones de consejería uniformes, pre-grabadas audio-

visuales en idioma local. Las instrucciones estuvieron basadas en las guías de OMS sobre

alimentación complementaria del niño amamantado (1), y especificaban la frecuencia deseable,

cantidad, consistencia, y textura de la comida, y principios de alimentación, higiene,

alimentación durante y después de enfermedad, y mantenimiento de la lactancia. En la

presentación audio-visual, también incluimos una demostración de cocción de recetas comunes

en la región, que eran preparadas y estandarizadas en términos de energía y otras densidades

de nutrientes de acuerdo con las guías de la OMS. Adicionalmente, las dudas de las familias

eran resueltas en consejería individual. Las madres recibieron un folleto de instrucciones y

recetas sugeridas en lengua local, y un equipo de utensilios del hogar uniforme con capacidad

conocida para medir los ingredientes y alimentar al niño.

Les pedimos a las madres que mantuvieran un registro diario de la dieta incorporando

información sobre el tipo, frecuencia y cantidad de comida consumida por el niño, y los

problemas enfrentados, si los hubiera. Las madres mantuvieron este registro durante 4 semanas

comenzando desde la fecha de la recomendación de iniciar la alimentación complementaria. En

cada visita reforzamos la suplementación con Vitamina D (400 UI/día) y hierro elemental (2-3

mg/kg/día), que fue iniciada desde las 2 semanas de edad como parte de las políticas clínicas de

los hospitales del estudio. Subsiguientemente, citamos a los infantes a visitas en el hospital a los

5, 6, 7, 9, y 12 meses de EC para medir antropometría, recabar morbilidad, y registro de la dieta

en las 24 hrs previas (apéndice p 4). Adicionalmente, proporcionamos a las familias de ambos

grupos con apoyo telefónico según necesidad las 24 hrs, 7 días por semana. En la visita de los

12 meses, hicimos evaluación del neurodesarrollo, DXA, y tomamos muestra de sangre en

ayunas como parte de la evaluación del resultado final. Si un infante no podía ser traído al

hospital, el equipo realizaba una visita domiciliaria para evaluar los resultados, excepto muestra

de sangre y DXA. Medimos la adherencia a la intervención a través de llamados telefónicos los

días 2, 7, 14, 21, y 28 a los 4 meses y nuevamente a los 6 meses de EC en ambos grupos, que fue

también corroborado por el registro diario de la dieta llevado por las madres.

Definimos que los infantes recibían la intervención asignada si se les ofrecía la alimentación

complementaria dentro de las 4 semanas de la recomendación de iniciar la comida, sin importar

la cantidad consumida. Un criterio adicional en el Grupo 6 m fue que si la alimentación

complementaria se había iniciado antes de los 6meses de EC, no debía haber iniciado más de

una semana antes de los 6 m de EC. La categoría de crecimiento intrauterino al nacer y los

scores Z para peso, longitud, y CC al nacer, al alta, y a las 40 sem postmenstruales se basaron en

las tablas de Fenton revisadas 2013 (10), calculados empleando el software antropométrico

provisto para datos de investigación. El ingreso de nutrientes fue calculado a partir de los

registros de 24 hrs de dieta empleando el valor nutritivo de las comidas indias (11). También

calculamos importantes indicadores dietéticos relacionados con la alimentación del infante a los

9 y 12 meses de edad basados en la información registrada durante los registros de 24 hrs

basados en las definiciones de la OMS (apéndice p 4) (12). También anotamos la consistencia y

textura de la comida, que estuvo basada en determinación de un experto debido a la ausencia

de criterio estándar. Los scores Z para peso, longitud, CC, peso para talla, e índice de masa

corporal en los puntos de tiempo subsiguientes (4, 5, 6, 7, 9, y 12 meses EC) se calcularon

utilizando los estándares de crecimiento OMS- MGRS (9).

Resultados

El resultado primario fue WAZ12, basado en los estándares de crecimiento OMS- MGRS (9). El

infante fue pesado empleando una balanza electrónica (Seca, Alemania; exactitud 5 g) y el peso

fue convertido al score Z empleando el software Antropométrico de OMS (versión 3.2.2, 2011),

empleando EC. Los resultados secundarios fueron cualquier morbilidad requiriendo ingreso

hospitalario desde el momento de incorporación hasta los 12 meses de EC, neurodesarrollo,

composición corporal, contenido mineral óseo (BMC) y densidad mineral ósea (BMD), resistencia

a insulina en términos de HOMA-IR (HOmeostatic Model Assessment for Insuline Resistance),

perfil lipídico, presión sanguínea, y ferritina sérica a los 12 meses de EC. El ingreso hospitalario

fue definido como internación por 6 o más horas o muerte en el hospital sin importar la

duración de la estadía. Un pediatra verificó el diagnóstico de cada episodio revisando los

registros del caso y con interacción con el médico tratante si era necesario. Las admisiones

hospitalarias repetidas debían estar separadas por una semana para ser consideradas episodios

separados. El neurodesarrollo fue determinado por un psicólogo clínico certificado empleando

la escala de Evaluación del Desarrollo para Infantes Indios (DASII), una adaptación India

validada, del Bayley- II (13). La composición corporal total, BMC, y BMD fueron evaluados con

DXA (Holoic Discovery W, S/N 84879, versión 13.1.1:7; software versión Apex 3.0). HOMA-IR se

calculó usando la fórmula: (14)

Insulina en ayunas (μU/mL)

× glucosa en ayunas (mmol/L)

HOMA-IR =

22·5

La glucosa plasmática en ayunas (mínimo 4 h) fue medida con el Método Colorimétrico

Enzimático (GOD-PAP) sin desproteinización basado en la enzima glucosa oxidasa, empleando

reactivo Glucosa PAP Fluid Mono (Centronic GmBH, Alemania), en un analizador completamente

automático ROCHE Modular P-800. Insulina sérica y ferritina fueron medidas por inmunoensayo

electrolumniscencia (ECLIA) empleando RocheCobase411 (Roche Diagnostics, Alemania). Un

valor de ferritina sérica menor de 12 fue tomado como corte para depósitos de hierro

deplecionados (15). Para perfil lipídico, colesterol total y triglicéridos en ayunas se investigaron

con métodos enzimáticos y lalipoproteína de alta densidad por método directo, con un

analizador completamente automático ROCHE P-800. La lipoproteína de muy baja densidad fue

calculada empleando valor de triglicérido y la lipoproteína de baja densidad empleando los

valores de colesterol total, lipoproteína de alta densidad y triglicérido con la ecuación de

Friedewald (16). La presión sanguínea fue medida empleando un monitor de presión digital

completamente automático (AG-SafeCHEK TM, modelo AG1010; exactitud 3 mmHg), basado en

el método oscilométrico, utilizando un manguito de presión reutilizable (WelchAllyn® Flexiport TM infantil tamaño número 7). Adicionalmente, proteína C- reactiva sérica se midió utilizando

Inmunoensayo Enzimático de proteína C-reactiva de alta sensibilidad (BioCheck, USA). Esto se

hizo para reflejar mejor el valor de ferritina sérica, que también es un reactante de fase aguda.

Análisis estadístico

Basamos nuestra estimación de tamaño muestral en los datos de WAZ12 para niños <34 sem EG

de un estudio de cohorte de bebés nacidos prematuros de nuestra propia unidad, cuyos

resultados de crecimiento fueron medidos empleando los estándares de la OMS-MGRS (16),

prospectivamente en intervalos de 3 meses desde el nacimiento hasta los 18 meses de edad

(WAZ12: 1.7 ±1.5, n=15; comunicación personal, Sharma P) (17). Asumiendo un efecto de

tamaño de 0.5 SDS, ß de 0.9, y α dos colas de 0.5, el tamaño muestral fue 190 en cada grupo.

Calculando un 5% de pérdida al seguimiento, el tamaño muestral total fue calculado en 400. El

tamaño muestral calculado fue corroborado a través de un estudio piloto (los infantes de la fase

piloto no fueron incluidos en este estudio).

Los datos fueron ingresados por duplicado en una base de datos online desarrollada en Visual

Basic como extremo frontal y servidor MS SQL como extremo final con un rango construido

interno y pruebas lógicas, con auditoria. El análisis se hizo con STATA 11.0 (College Station, TX,

USA), por intención a tratar. Las variables continuas se compararon con test t de Student para

las normalmente distribuidas y Wilcoxon Rank-sum test para los no normalmente distribuidos.

Las proporciones fueron comparadas con chi cuadrado. Construimos una curva de sobrevida de

Kaplan Meir para mostrar el primer episodio de ingreso hospitalario en cada grupo durante el

período de estudio incluyendo la información disponible en todos los infantes, y calculamos el

cociente de riesgo entre ambos grupos usando la regresión de riesgo proporcional de Cox.

También empleamos el modelo Anderson- Gill, que es contar la extensión del proceso para la

regresión proporcional de riesgo de Cox para calcular el cociente de riesgo entre los dos grupos

para dar cuenta de múltiples episodios de ingreso hospitalario para cualquier infante.

Un análisis de subgrupo pre-especificado fue hecho por cada sitio, categoría de crecimiento

fetal intrauterino (PEG, AGE), EG al nacer (<28 sem, 28-30 sem y 31-33 sem), y tipo de

alimentación al ser randomizado (Leche humana, no leche humana, alimentación mixta). Un

análisis de ecuación generalizada de estimación fue empleado para comparar tendencia de

antropometría entre los dos grupos en el tiempo. Explicamos el efecto de agrupamiento sobre

el resultado primario debido a la co-randomización de mellizos y trillizos re analizando el

resultado primario después de excluir los nacimientos múltiples (reteniendo sólo el primer

gemelar) en ambos grupos. El estudio fue supervisado por un Comité Doctoral independiente

que revisó los procesos cada 6 meses. Un consejero de seguridad independiente también revisó

los datos de mortalidad y morbilidad.

Rol de la fuente de financiamiento

El Concejo Indio de Investigación Médica no tuvo rol en el diseño del estudio, recolección de

datos, análisis, interpretación, o escritura del reporte. El autor de correspondencia tuvo acceso

completo a todos los datos en el estudio y tuvo la responsabilidad final de la decisión para

remitir la publicación.

Resultados

Entre el 20 de Marzo de 2013, y el 24 de Abril de 2015, identificamos 2135 nacidos vivos de < 34

sem EG. 412 fueron elegibles a los 4 meses EC, nueve de los cuales negaron consentimiento, y

403 fueron randomizados- 206 para el grupo de 4 meses y 197 Para el de 6 meses (figura 1). En

general, un total de 22 infantes en el Grupo 4 m (un par de mellizos quitaron el consentimiento,

16 infantes se perdieron al seguimiento, y 4 fallecieron) y ocho infantes en el Grupo 6 m (6

infantes perdidos a seguimiento y 2 fallecidos) fueron excluidos del análisis del resultado

primario. Sin embargo, las muertes fueron incluidas en el análisis de resultado secundario.

196/203 infantes en el Grupo 4 m (2 retiraron consentimiento, 1 falleció antes de incorporar la

alimentación complementaria) y 184/193 infantes en el Grupo 6 m (2 fallecieron y dos se

perdieron al seguimiento antes de iniciar la alimentación complementaria) recibieron la

intervención adjudicada.

Las características basales de los infantes en el estudio al nacer y al randomizarse fueron

similares (tabla 1). La edad media de randomización fue 3.9 meses EC en ambos grupos y el

peso medio y WAZ fueron 5117 g en el Grupo 4 m versus 5187 g en el Grupo 6 m y -2.3 en el

Grupo 4 m versus -2.2 en el Grupo 6 m. La proporción de infantes que recibieron algo de LH fue

similar en ambos grupos en todos los momentos (apéndice p 5). Una alta proporción de

infantes estaban recibiendo alimentación complementaria (sin importar cantidad) 1 mes

después de recomendar el inicio de comida (95.4% vs 90.7%; apéndice p 5).

El resultado primario, WAZ12, fue similar entre los dos grupos (media -1.6, DS 1.2 en el Grupo 4

m vs -1.6, 1.3 en el Grupo 6 m, diferencia media 0.005, IC 95% -0.24 a 0.25; p=0.96) a la media

de EC de 12.3 meses (tabla 2). El cambio en WAZ en los infantes del estudio desde el nacimiento

hasta los 12 meses de EC (p= 0.61; figura 2) y específicamente entre 4 y 12 meses de EC fue

similar (valor de p en el análisis de ecuación de estimación generalizada (GEE), p= 0.836). El peso

medio (7794 vs 7846 g, p= 0.65) a los 12 meses EC también fue similar (apéndice p 5). Datos

adicionales de crecimiento son provistos en el apéndice P 6-7).

Seis infantes fallecieron durante el período en estudio; 4/203 en el Grupo 4 m versus 2/197 en el

Grupo 6 m (RR 1.9, IC 95% 0.4-10.5; tabla 2). 34 (18%)/188 infantes en el Grupo 4 m requirieron

re-ingreso hospitalario durante el período de estudio, comparado con 18 (9%)/192 infantes en

el Grupo 4 m (RR 1.9, IC 95% 1.1-3.3; p=0.01), con 2.5 episodios por 100 meses-infante en el

Grupo 4 m versus 1.4 episodios por 100 meses-infante en el Grupo 6 m (cociente de tasa de

incidencia 1.7, IC 95% 1.0-3.0; p= 0.04, tabla 2). La curva de Kaplan- Meier para el tiempo hasta

el primer episodio de ingreso hospitalario mostró que el ingreso medio primero para el Grupo 4

m fue más temprano comparado con el Grupo 6 m (p= 0.02), con el riesgo de admisión

hospitalaria 48% más bajo en el Grupo 6 m comparado con el Grupo 4 m (cociente de

riesgo(CR) 0.52, IC 95% 0.29-0.92; p= 0.025) durante el período de estudio después de ajustar

por grupo de randomización, hospital y gestación. Los resultados se mantuvieron similares

cuando se tomaron en cuenta múltiples episodios de hospitalización para cualquier infante

empleando el modelo Anderson- Gill (0.56, 0.33-0.94; p= 0.029, figura 3).

Los cocientes de desarrollo motor y mental en los dos grupos fueron similares (tabla 2). No

hubo diferencia significativa en la composición corporal, BMC, BMD, perfil lipídico, HOMA-IR,

presión sanguínea, o ferritina sérica entre ambos grupos. La ferritina sérica media fue 5.4 μg/dL

(RIQ 3.2-12.4) en el Grupo 4 m y 5.7 (2.5-13.3) en el Grupo 6 m (p= 0.73), con casi dos tercios de

infantes en ambos grupos obteniendo un resultado <12 μg/dL (tabla 2).

Un análisis de subgrupo pre-especificado por sitio. Categoría de crecimiento fetal intrauterino,

EG al nacer, y tipo de alimentación a la randomización (LH, no LH, mixta) no presentó ninguna

diferencia en el resultado primario (tabla 3).

La diversidad de la dieta fue aceptable en menos de dos tercios de los infantes en cada grupo

(60.1% vs 55.0 %, p= 0.32; tabla 4). La frecuencia mínima de alimentación aceptable estuvo

presenta en la mayoría de los infantes (93.4% vs 92.6%; p= 0.75). Sin embargo, menos de dos

tercios de los infantes estuvieron recibiendo la dieta mínima aceptable. Una elevada proporción

de infantes en ambos grupos se alimentaban con biberón (tabla 4).

Tabla 2- Resultados primarios y secundarios

Discusión

Nuestro estudio muestra que iniciar la alimentación complementaria a una edad más temprana

de 4 meses comparado con 6 meses de EC resultó en scores WAZ similares y otros resultados

del crecimiento, composición corporal, estado de mineralización ósea, depósitos de hierro (con

suplementación de hierro), y marcadores de síndrome metabólico a los 12 meses EC. Sin

embargo, hubo un riesgo más alto de ingreso hospitalario en el con iniciación más temprana,

sugiriendo que la iniciación de la alimentación complementaria a los 6 meses EC podría ser

preferible a los 4 meses EC en infantes nacidos <34 sem de gestación.

Esta evidencia puede ser considerada robusta, porque el estudio tuvo adecuado poder para

detectar el resultado de interés y cumplió todos los criterios para una alta validez interna: fue un

trabajo randomizado con grupos siendo similar en un principio, los grupos fueron similarmente

tratados excepto por la intervención asignada, no se administró co-intervención, la tasa general

de seguimiento fue alta, los análisis fueron por intención de tratar y los resultados fueron

cegados u objetivos.

Hasta donde sabemos, sólo un trabajo publicado antes comparó el efecto de la iniciación de la

alimentación complementaria en el crecimiento de los infantes prematuros. El estudio tuvo

varias limitaciones y concluyó que no hubo diferencias en los parámetros entre los dos grupos

de intervención a los 12 meses de edad.

Dos otros estudios- uno en niños de término, saludables y uno en BPN, infantes alimentados

exclusivamente con pecho materno- compararon el efecto de la iniciación de alimentación

complementario a los 4 meses versus 6 meses de edad (18, 19). Ambos estudios mostraron que

la intervención no resultó en diferencia significativa en peso o longitud a los 12 meses de edad,

similar a nuestro estudio. Desafortunadamente, no pudimos medir la ingesta de leche humana

en nuestro estudio. Sin embargo, tomamos nota de que en los niños amamantados, el ingreso

de energía por kilo de peso corporal (excluyendo leche materna) fue más alto en el Grupo 4 m

comparado con el Grupo 6 m en todos los hitos del tiempo, pese a que fue similar entre los

infantes no amamantados (excepto por una pequeña diferencia a los 6 meses de EC; apéndice p

9). Debido a que no hubo diferencias en los parámetros de crecimiento entre los dos grupos en

ningún momento, es factible que los niños amamantados en el Grupo 6 m aumentaron su

ingesta de LH y derivaron su energía adicional de la LH. Estudios empleando métodos

isotópicos robustos también han mostrado ingesta de LH aumentada en niños de término

alimentados exclusivamente con LH que continúan siendo amamantados hasta los 6 meses de

edad, y disminución de ingesta entre aquellos que comienzan la alimentación complementaria a

los 4 meses de edad (20-21).

La incidencia general de ingreso hospitalario en la población en estudio fue baja, pero los

infantes en el Grupo 4 m tuvieron más episodios de diarrea y de infecciones de la vía aérea baja

hasta los 12 meses de EC. Pese a que estos podrían ser explicados por la potencial

contaminación de los alimentos complementarios debido a inadecuada higiene (22), una razón

más biológica debería ser considerada. Se sabe que la LH confiere beneficios inmunológicos a

los niños que son especialmente importantes para los infantes prematuros (23), y la ingesta de

LH es posible que haya sido menor en los niños del Grupo 4 m, como se dijo antes. Además, el

rol de las exposiciones dietarias en delinear la función inmunológica a corto y largo plazo en

infantes podría ser otro factor (24, 25). Poca diferencia en los patrones de crecimiento en ambos

grupos con poco crecimiento compensatorio en cada grupo podría explicar la composición

corporal similar y los similares marcadores de síndrome metabólico entre los grupos. La masa

grasa media y el % de grasa a los 12 meses EC en cada grupo estuvieron cerca de los

reportados en la literatura, sugiriendo que la biología del crecimiento entre este grupo de

infantes es similar en distintos escenarios (26). Sin embargo, necesitamos ser cautos con el DEXA

corporal como medida ya que no puede establecer exactamente la adiposidad aberrante que

podría ocurrir.

Las malas prácticas dietéticas entre los niños del estudio pese a la consejería son más difíciles de

comprender, especialmente cuando la evidencia disponible sugiere que la consejería ayuda a

mejorar las prácticas de alimentación (27). Postulamos que las prácticas hubieran sido peores sin

la consejería porque los indicadores obtenidos en nuestro estudio fueron mejores que los

reportados en la literatura de espacios similares (28). Parece que para lograr los estándares

dietarios recomendados en esta población requiere un grado de cambio conductual a nivel

familiar difícil de lograr, y requiere enfoques innovadores para suplementar la consejería.

También es importante considerar que factores del infante como el sabor, la preferencia,

habilidad, o interés en comer y el patrón de alimentación han contribuido a los patrones

dietéticos vistos en el estudio.

Otra preocupación importante revelada por el estudio es la depleción de los depósitos de

hierro, pese a que la mayoría de los niños recibían suplementación. Este hallazgo es un enigma,

y dado que nosotros aseguramos estrictamente la adherencia, requiere alguna explicación.

Estudios han mostrado que la biodisponibilidad de hierro de la dieta india habitual es baja,

debido a relación de elevados fitatos y bajo ácido ascórbico con hierro. Adicionalmente, el

efecto de matriz de la comida y sinergias específicas del contexto local pueden ser un tema (29,

30). También presumimos que la co-administración de suplementos de hierro con comida o

leche podría haber resultado en absorción disminuida del hierro. El clampeo tardío del cordón

no era practicado en ninguno de los sitios durante el período de estudio, y esto puede haber

sido una razón. Sin embargo, pese a todas estas especulaciones, este es un resultado crucial

con una carga sobre el neurodesarrollo de estos infantes vulnerables. Por lo tanto, es

importante continuar investigando para determinar si las dosis recomendadas de

suplementación son adecuadas para infantes prematuros, factores que retardan la absorción de

hierro, y posiblemente mejores preparaciones de hierro con mejor disponibilidad.

Este estudio fue hecho en un país de ingresos medio-bajo, pero proponemos que los resultados

podrían tener relevancia aún en países de altos ingresos por dos razones principales. Primero,

observamos que los patrones de crecimiento de este grupo de infantes son similares en

distintos escenarios, con restricción universal de crecimiento postnatal (31, 32). En este estudio,

los parámetros de masa grasa a los 12 meses de edad fueron cercanos a los reportados por la

literatura (28). Segundo, la alimentación o los patrones dietarios en la infancia son también

similares con bajas tasas de amamantamiento exclusivo e inadecuadas prácticas de alimentación

complementaria (33, 34, 35). Sin embargo, otros factores que podrían limitar la generalización

de resultados a otros escenarios y necesitarían otras consideraciones previas incluyen

diferencias entre países en desarrollo y desarrollados con respecto al PN de infantes

prematuros, tasas de prevalencia de RCIU en la población, prácticas con respecto al empleo de

fórmulas postalta, enriquecidas en nutrientes o leche animal, tasas de alimentación con LH,

fortificación de alimentos complementarios, tasas de infección, y factores sociodemográficos

como educación materna y estado socioeconómico, que podrían influenciar la higiene y la

adhesión al consejo dietético.

También reportamos que la edad media postnatal de los infantes del estudio en la intervención

fue 5.7 meses (DS 0.3) en el Grupo 4 m, y 7.9 meses (DS 0.4) en el Grupo 6 meses. Debido a que

el grupo 4 m no mostró mejores resultados pero aumentó las admisiones hospitalarias, se

entiende que los 6 meses de edad postnatal no debe ser preferida a los 6 meses de EC para la

iniciación de la alimentación complementaria en infantes prematuros.

Las limitaciones de nuestro estudio son el diseño de ensayo abierto y diferente pérdida al

seguimiento en los dos grupos. Sin embargo, los investigadores de resultados estaban

enmascarados para los grupos de asignación, y las características basales de los infantes

perdidos al seguimiento en ambos grupos fueron similares. La utilidad diagnóstica de cualquier

evaluación del neurodesarrollo al año de edad es también inadecuada, lo cual podría también

ser considerada una limitación. Hay una necesidad para mayor seguimiento de esta cohorte, y

re-evaluación de su crecimiento, desarrollo, estado de micronutrientes, y marcadores de

enfermedad crónica.

La investigación futura debería enfocar en identificar una ventana apropiada, si existe una, para

mejorar el crecimiento postnatal en infantes prematuros. También hay necesidad urgente para

identificar las razones de pobres patrones dietarios y depósitos de hierro muy deplecionados

pese a la suplementación en infantes prematuros, y probar potenciales intervenciones e instituir

medidas preventivas para apuntar a lo mismo. El seguimiento a largo plazo de tales infantes,

idealmente hasta la adultez, es altamente deseable.

Referencias

1 Pan American Health Organization, Guiding Principles for Complementary Feeding of the Breastfed Child.

Washington DC: Pan American Health Organization, World Health Organization, 2003.

2 Agostoni C, Decsi T, Fewtrell M, et al. ESPGHAN Committee on Nutrition. Complementary feeding: a

commentary by the ESPGHAN Committee on Nutrition. J Pediatr Gastroenterol Nutr 2008; 46: 99–110.

3 AAP policy statement. Breastfeeding and the use of human milk. Pediatrics 2005; 115: 496.

4 Cooke RJ, Ainsworth SB, Fenton AC. Postnatal growth retardation: a universal problem in preterm infants.

Arch Dis Child Fetal Neonatal Ed 2004; 89: F428–30.

5 Horbar JD, Ehrenkranz RA, Badger GJ, et al. Weight growth velocity and postnatal growth failure in infants

501 to 1500 grams: 2000–2013. Pediatrics 2015; 136: e84–92.

6 Bauer J, Werner C, Gerss J. Metabolic rate analysis of healthy preterm and full-term infants during the first

weeks of life. Am J Clin Nutr 2009; 90: 1517–24.

7 Weintraub V, Mimouni FB, Dollberg S. Effect of birth weight and postnatal age upon resting energy

expenditure in preterm infants. Am J Perinatol 2009; 26: 173–77.

8 Marriott LD, Foote KD, Bishop JA, Kimber AC, Morgan JB. Weaning preterm infants: a randomised

controlled trial. Arch Dis Child Fetal Neonatal Ed 2003; 88: F302–07.

9 WHO Multicentre Growth Reference Study Group. WHO Child Growth Standards based on length/height,

weight and age. Acta Paediatr Suppl 2006; 450: 76–85.

10 Fenton RT, Kim HJ. A systematic review and meta-analysis to revise the Fenton growth chart for preterm

infants. BMC Pediatrics 2013; 13: 59.

11 Gopalan C, Rama Sastri BV, Balasubramanian SC. Nutritive Value of Indian Foods (NVIF). Hyderabad:

National Institute of Nutrition, 1989.

12 World Health Organization. Indicators for assessing infant and young child feeding practices:

conclusions of a consensus meeting held 6–8 November 2007 in Washington DC, USA. WHO

2008.http://apps.who.int/iris/bitstream/10665/43895/1/9789241596664_eng.pdf (accessed March 12,

2016).

13 Pathak P. Baroda norms of motor and mental development of Indian babies from 1 to 30 months.

Psychological stud 1993; 38: 142–49.

14 Matthews DR, Hosker JP, Rudenski AS, Naylor BA, Treacher DF, Turner RC. Homeostasis model

assessment: insulin resistance and beta-cell function from fasting plasma glucose and insulin

concentrations in man. Diabetologia 1985; 28: 412–19.

15 World Health Organization and Centers for Disease Control and Prevention. Assessing the iron status of

populations: including literature reviews. http://www.who.int/nutrition/publications/

micronutrients/anaemia_iron_deficiency/9789241596107.pdf (accessed March 25, 2016).

16 Friedewald WT, Levy RI, Fredrickson DS. Estimation of the concentration of low-density lipoprotein

cholesterol in plasma, without use of the preparative ultracentrifuge. Clin Chem 1972; 18: 499–502.

17 Sharma PK, Sankar MJ, Sapra S, et al. Growth and neurosensory outcomes of preterm very low birth

weight infants at 18 months of corrected age. Indian J Pediatr 2011; 78: 1485–90.

18 Cohen RJ, Brown KH, Canahuati J, Rivera LL, Dewey KG. Effects of age of introduction of complementary

foods on infant breast milk intake, total energy intake, and growth: a randomised intervention study in

Honduras. Lancet 1994; 344: 288–93.

19 Dewey KG, Cohen RJ, Brown KH, Rivera LL. Age of introduction of complementary foods and growth of

term, low-birth-weight, breast-fed infants: a randomized intervention study in Honduras. Am J Clin Nutr

1999; 69: 679–86.

20 Wells JC, Jonsdottir OH, Hibberd PL, et al. Randomized controlled trial of 4 compared with 6 mo of

exclusive breastfeeding in Iceland: differences in breast-milk intake by stable-isotope probe. Am J Clin Nutr

2012; 96: 73–79.

21 Nielsen SB, Reilly JJ, Fewtrell MS, Eaton S, Grinham J, Wells JC. Adequacy of milk intake during exclusive

breastfeeding: a longitudinal study. Pediatrics 2011; 128: e907–14.

22 Weisstaub G, Uauy R. Non-breast milk feeding in developing countries: challenge from microbial and

chemical contaminants. Ann Nutr Metab 2012; 60: 215–19.

23 Gregory KE, Walker WA. Immunologic factors in human milk and disease prevention in the preterm

infant. Curr Pediatr Rep 2013; 1: 222.

24 Thompson AL, Monteagudo-Mera A, Cadenas MB, Lampl ML, Azcarate-Peril MA. Milk and solid-feeding

practices and daycare attendance are associated with differences in bacterial diversity, predominant

communities, and metabolic and immune function of the infant gut microbiome. Front Cell Infect Microbiol

2015; 5: 3.

25 Walker WA, Iyengar RS. Breast milk, microbiota, and intestinal immune homeostasis. Pediatr Res 2015;

77: 220–28.

26 Cooke RJ, Rawlings DJ, McCormick K, et al. Body composition of preterm infants during infancy. Arch Dis

Child Fetal Neonatal Ed 1999; 80: F188–91.

http://www.who.int/nutrition/publications/

27 Shi L, Zhang J. Recent evidence of the effectiveness of educational interventions for improving

complementary feeding practices in developing countries. J Trop Pediatr 2011; 57: 91–98.

28 Senarath U, Agho KE, Akram DE, et al. Comparisons of complementary feeding indicators and

associated factors in children aged 6–23 months across five South Asian countries. Matern Child Nutr 2012;

8 (suppl 1): 89–106.

29 Nair KM, Iyengar V. Iron content, bioavailability & factors affecting iron status of Indians. Indian J Med

Res 2009; 130: 634–45.

30 Nair KM, Augustine LF. Food synergies for improving bioavailability of micronutrients from plant foods.

Food Chemistry, Sept 19, 2016. http://www.sciencedirect.com/science/article/pii/

S0308814616315102 (accessed March 12, 2016).

31 Hack M, Schluchter M, Cartar L, Rahman M, Cuttler L, Borawski E. Growth of very low birth weight infants

to age 20 years. Pediatrics 2003; 112: e30–38.

32 Gladstone M, Oliver C, Van den Broek N. Survival, morbidity, growth and developmental delay for

babies born preterm in low and middle income countries - a systematic review of outcomes measured.

PLoS One 2015; 10: e0120566.

33 Shealy KR, Scanlon KS, Labiner-Wolfe J, Fein SB, Grummer-Strawn LM. Characteristics of breastfeeding

practices among US mothers. Pediatrics2008; 122 (suppl 2): S50–55.

34 Lande B, Andersen LF, Baerug A, et al. Infant feeding practices and associated factors in the first six

months of life: the Norwegian infant nutrition survey. Acta Pediatr 2003; 92: 152–61.

35 Grummer-Strawn LM, Scanlon KS, Fein SB. Infant feeding and feeding transitions during the first year of

life. Pediatrics 2008; 122 (suppl 2): S36–42.

www.thelancet.com/lancetgh Vol 5 May 2017 e501

Articles

Complementary feeding at 4 versus 6 months of age for preterm infants born at less than 34 weeks of gestation: a randomised, open-label, multicentre trialShuchita Gupta, Ramesh Agarwal, Kailash Chandra Aggarwal, Harish Chellani, Anil Duggal, Sugandha Arya, Sunita Bhatia, Mari Jeeva Sankar,

Vishnubhatla Sreenivas, Vandana Jain, Arun Kumar Gupta, Ashok K Deorari, Vinod K Paul, and Investigators of the CF trial*

SummaryBackground Evidence on the optimal time to initiation of complementary feeding in preterm infants is scarce. We examined the effect of initiation of complementary feeding at 4 months versus 6 months of corrected age on weight for age at 12 months corrected age in preterm infants less than 34 weeks of gestation.

Methods In this open-label, randomised trial, we enrolled infants born at less than 34 weeks of gestation with no major malformation from three public health facilities in India. Eligible infants were tracked from birth and randomly assigned (1:1) at 4 months corrected age to receive complementary feeding at 4 months corrected age (4 month group), or continuation of milk feeding and initiation of complementary feeding at 6 months corrected age (6 month group), using computer generated randomisation schedule of variable block size, stratified by gestation (30 weeks or less, and 31–33 weeks). Iron supplementation was provided as standard. Participants and the implementation team could not be masked to group assignment, but outcome assessors were masked. Primary outcome was weight for age Z-score at 12 months corrected age (WAZ12) based on WHO Multicentre Growth Reference Study growth standards. Analyses were by intention to treat. The trial is registered with Clinical Trials Registry of India, number CTRI/2012/11/003149.

Findings Between March 20, 2013, and April 24, 2015, 403 infants were randomly assigned: 206 to receive complementary feeding from 4 months and 197 to receive complementary feeding from 6 months. 22 infants in the 4 month group (four deaths, two withdrawals, 16 lost to follow-up) and eight infants in the 6 month group (two deaths, six lost to follow-up) were excluded from analysis of primary outcome. There was no difference in WAZ12 between two groups: –1·6 (SD 1·2) in the 4 month group versus –1·6 (SD 1·3) in the 6 month group (mean difference 0·005, 95% CI –0·24 to 0·25; p=0·965). There were more hospital admissions in the 4 month group compared with the 6 month group: 2·5 episodes per 100 infant-months in the 4 month group versus 1·4 episodes per 100 infant-months in the 6 month group (incidence rate ratio 1·8, 95% CI 1·0–3·1, p=0·03). 34 (18%) of 188 infants in the 4 month group required hospital admission, compared with 18 (9%) of 192 infants in the 6 month group.

Interpretation Although there was no evidence of effect for the primary endpoint of WAZ12, the higher rate of hospital admission in the 4 month group suggests a recommendation to initiate complementary feeding at 6 months over 4 months of corrected age in infants less than 34 weeks of gestation.

Funding Indian Council of Medical Research supported the study until Nov 14, 2015. Subsequently, Shuchita Gupta’s salary was supported for 2 months by an institute fellowship from All India Institute Of Medical Sciences, and a grant by Wellcome Trust thereafter.

Copyright © The Author(s). Published by Elsevier Ltd. This is an Open Access article under the CC BY license.

IntroductionExclusive breastfeeding for 6 months followed by complementary feeding for term infants is a standard recommendation by WHO, widely endorsed and accepted by the global community. However, none of the organisations including WHO,1 the European Society of Paediatric Gastroenterology, Hepatology and Nutrition (ESPGHAN),2 or the American Academy of Pediatrics3 provide evidence-based guidelines with respect to the optimal time of initiation of complementary feeding in preterm infants who are at a much higher risk of postnatal growth restriction than full-term infants.4,5 The only available guidelines are from the UK, and are based

on a single, non-systematic review of primarily physiological studies.

Extrapolating the recommendation for full-term infants to initiate complementary feeding at 6 months of age to preterm infants is dependent on two major questions: what does 6 months refer to in a preterm infant—chronological (postnatal) age or corrected age? Health-care providers generally use corrected age for monitoring the physical growth and development of preterm infants. Secondarily, if we assume that 6 months refers to the corrected age, should these infants not start complementary feeding earlier than their full-term counterparts—for example, at 4 months corrected age

Lancet Glob Health 2017; 5: e501–11

This online publication has been corrected. The corrected version first appeared at thelancet.com/lancetgh on April 10, 2018

See Comment page e470

*A complete list of the investigators of the CF trial is included at the end of the article

All India Institute of Medical Sciences, Delhi, India (S Gupta MD, R Agarwal DM, A Duggal MD, S Bhatia DCH, M J Sankar DM, V Sreenivas PhD, V Jain MD, A K Gupta MD, A K Deorari MD, V K Paul MD); Vardhman Mahavir Medical College and associated Safdarjung hospital, New Delhi, India (K C Aggarwal MD, H Chellani MD, S Arya MD); and Kasturba Hospital , Delhi, India (A Duggal, S Bhatia)

Correspondence to: Dr Ramesh Agarwal, Department of Pediatrics, All India Institute of Medical Sciences, Delhi 110029, India [email protected]

or

Dr Harish Chellani, Department of Pediatrics, Vardhman Mahavir Medical College and associated Safdarjung hospital, Delhi, 110029, India

Descargado para Anonymous User (n/a) en Univ Antioquia de ClinicalKey.es por Elsevier en marzo 28, 2019.Para uso personal exclusivamente. No se permiten otros usos sin autorización. Copyright ©2019. Elsevier Inc. Todos los derechos reservados.

http://crossmark.crossref.org/dialog/?doi=10.1016/S2214-109X(17)30074-8&domain=pdf

Articles

e502 www.thelancet.com/lancetgh Vol 5 May 2017

instead of 6 months corrected age (appendix p 1)? Preterm infants have higher energy requirements compared with full-term infants,6,7 and it is not known how long in infancy milk feeds alone (breastmilk or

formula) are sufficient to meet their requirements. Most complementary foods provide higher calorie density compared with milk feeds, and can make up for the energy gap between increased requirements of preterm

Research in context

Evidence before this studyWe searched MEDLINE and Cochrane Central Register of Controlled Trials (CENTRAL) up to 6 Feb 2016, limited to human studies but with no restriction of language or publication date, using the terms “complementary food” OR “complementary foods” OR “complementary feeding” OR “supplementary feeding” OR “weaning” OR “weaning foods” OR “beikost” OR “semisolids” OR “semisolid foods” OR “semisolid food” OR “semisolid feeds” OR “infant feeding” OR “infant diet” OR “infant food” OR “infant foods”. We found 22723 articles on MEDLINE and 3667 on CENTRAL. After screening the title and abstracts and excluding duplicates, we reviewed the full text of 50 potentially relevant studies. We also searched the reference list of all these 50 studies to identify any other relevant studies.

Inclusion criteria for the Review were: any randomised, quasi-randomised trial or observational study investigating different time or age at initiation or introduction of complementary feeding in preterm infants. Complementary feeding was defined as initiation of semisolid, soft or solid foods other than breast, formula, or animal milk. We therefore excluded any studies enrolling full term infants or those where the term ‘weaning’ was used to indicate transition from breastfeeding to formula or animal milk rather than to semisolid, soft, or solid foods.

On full text review and reference list search, we identified three studies eligible for this Review—one randomised trial, one prospective observational study and one study that was a secondary analysis of data from original unrelated studies. The first study randomly assigned all preterm infants of less than 37 completed weeks gestation (n=68) to receive complementary feeding after 13 weeks of postnatal age if weighing at least 3·5 kg; or after 17 weeks of postnatal age if weighing at least 5·0 kg. The intervention, ie, initiation of complementary feeding in both groups, was also subject to parental perception of infant being ready to accept complementary feeding. The 13 week group also received a co-intervention in the form of energy and protein-dense food. The study did not find any significant difference in the SDs of length (0·2, 0·2 vs –0·1, 0·3), weight (–0·7, 0·2 vs –0·8, 0·2) and head circumference (–0·5, 0·2 vs –0·5, 0·2) between the groups. However, there was some improvement in length gain per week from birth to 12 months corrected age (mean cm per week, SD: 5·1, 0·07 vs 4·9, 0·10; p=0·04) and change in mean length SDs in the early weaning group (mean, SD: –1·1, 0·2 at birth to 0·2, 0·2 at 12 months corrected age) when compared with late weaning group (mean, SD: –1·0, 0·2 at birth to –0·1, 0·3 at 12 months corrected age).

The second study followed preterm infants from birth until 12 months corrected age, and recorded the time of initiation of complementary feeding based on information on introduction of 12 prelisted food items. Introduction of 4 or more items (n=203 infants) compared with that of less than 4 items (n=54 infants) by 17 weeks corrected age was associated with 3·5 times higher risk for developing eczema at 12 months corrected age.

The third study was a secondary analysis of data with respect to introduction of solid foods at ≤12 weeks or >12 weeks on growth parameters of preterm infants, from two previous trials—one on post-discharge formula vs full term formula (published), and another unpublished. The study did not find any difference with regard to introduction of solid foods ≤12 weeks or >12 weeks on weight gain in kg (mean, SE: 8·25, 0·05; n=365 vs 8·27, 0·07; n=102), length gain in cm (70·3, 0·16; n=362 vs 70·6, 0·2; n=102) and head circumference gain in cm (45·8, 0·08; n=364 vs 45·8, 0·12; n=102) between 12 weeks and nine months corrected age, or between 12 weeks and 18 months corrected age (data not reported). There was no difference in the prevalence of atopy, lower respiratory tract infection, gastroenteritis, sleep duration, or waking at night between the two groups.

Added value of this studyOur study provides level-1 evidence with regard to the timing (4 vs 6 months of corrected age) for initiation of complementary feeding in preterm infants less than 34 weeks of gestation, a group most susceptible to postnatal growth restriction. It shows that early initiation of complementary feeding at 4 month compared with 6 months of corrected age does not improve growth of preterm infants at corrected age of 12 months. It also does not result in a difference in neurodevelopment outcomes, body composition, bone mineralisation, and any marker for metabolic syndrome like insulin resistance, lipid profile, and blood pressure in infancy. Rather, earlier initiation of complementary feeding at 4 months corrected age increases the risk of hospital admission due to concurrent morbidities, predominantly diarrhoea and lower respiratory tract infections. In both groups, dietary patterns remain poor and body iron stores remain depleted despite iron supplementation until 12 months of corrected age.

InterpretationOur study suggests that 6 months of corrected age should be preferred over 4 months of corrected age for initiation of complementary feeding in preterm infants less than 34 weeks of gestation. Findings of clinically significant iron deficiency despite iron supplementation and poor dietary patterns in infancy will guide further research.

See Online for appendix


Articles


infants, and limited supply from milk feeds. Therefore, an earlier introduction of complementary feeding in preterm infants than is recommended for full-term infants might help improve their growth.

We did a systematic review to answer this crucial question but found only one low quality randomised trial addressing the issue.8 We therefore chose to test the hypothesis that initiation of complementary feeding (defined as semisolid, soft, or solid foods other than breastmilk, formula, or animal milk) at 4 months compared with 6 months corrected age in infants less than 34 weeks of gestation will increase their weight for age Z-score at 12 months corrected age (WAZ12) by 0·5 standard deviation score (SDS; around 500 g, based on WHO-Multicentre Growth Reference Study [MGRS] growth standards).9 We also investigated neuro-development, body composition, bone density, and early markers of metabolic syndrome as there is little evidence on the links between infant feeding and later health, and sought to characterise potential effects of the nutritional intervention on some markers of the same.

MethodsStudy designWe did this randomised, open-label, parallel group trial at three public health facilities; the All India Institute of Medical Sciences (AIIMS), Vardhman Mahavir Medical College associated Safdarjung hospital, and Kasturba hospital in New Delhi, India. All three sites provide tertiary care neonatal services to inborn neonates. Neonatal intensive care is provided to all infants less than 34 weeks of gestation, and exclusive breastmilk feeding is actively promoted (appendix p 2). Ethics approval was obtained from institutional ethics committees at all sites. The trial is registered with Clinical Trials Registry of India, number CTRI/2012/11/003149.

ParticipantsEligible patients were infants less than 34 weeks of gestation with no major malformation, residing within 60 km of study hospitals and not expected to move away from the study region within 1 year of birth. Infants requiring hospital admission from birth to later than 40 weeks postmenstrual age were excluded. Research team identified potentially eligible infants at birth, kept them in follow-up after discharge and called their caregiver by telephone at 4 months corrected age for enrolment into the study. Infants were excluded at this stage if they had already been started on complementary feeding. Initiation of complementary feeding was defined as intentional initiation of complementary food by family irrespective of amount, unintentional initiation but food given for more than 3 days, or that the total amount given provided more than 1% of recommended dietary allowance of calories for the infant. Written informed consent was obtained from the parents or legally acceptable representative.

Randomisation and maskingParticipants were randomly assigned (1:1 allocation) to the 4 month group, wherein parents or family were advised to initiate complementary feeding at 4 months corrected age; or to the 6 month group, in which they were advised to continue milk feeds until 6 months corrected age, followed by initiation of complementary feeding at 6 months corrected age. An independent person (MJS) provided computer generated random sequences, stratified for site and gestation (30 weeks or less and 31–33 weeks) with variable block size unknown to the research team involved in implementation. Allocation was concealed using sealed, opaque, sequentially numbered envelopes. The families and research team involved in implementation could not be masked to the allocation groups following randomisation, however, those assessing primary and secondary outcomes were masked. Twins and triplets were assigned to the same group.

ProceduresDetailed descriptions of the study procedures can be found in the appendix (p 3). At 4 months corrected age, we randomly assigned the infants and did baseline anthropometric measurements and Dual Energy X-Ray Absorptiometry (DXA).

A single person counselled the families for initiation of complementary feeding at corrected age 4 months or 6 months, using uniform, prerecorded audio-visual counselling instructions in local language. The instructions were based on the WHO guidelines on complementary feeding of the breastfed child,1 and specified the desired frequency, amount, consistency, and texture of food, and principles of responsive feeding, hygiene, feeding during and after illness, and maintenance of breastfeeding. In the audio-visual presentation, we also included a demonstration on cooking common recipes of the region, which were prepared and standardised in terms of energy and other nutrient densities in accordance with the WHO guidelines. Additionally, families’ queries were resolved through one-to-one counselling. The mothers were provided a handout of the instructions and suggested recipes in local language, and a set of uniform household utensils with known capacity to measure the ingredients and to feed the child.

We asked mothers to maintain a daily dietary record incorporating information on type, frequency and amount of food consumed by the child, and problems faced, if any. Mothers maintained this record for 4 weeks beginning from date of counselling for initiation of complementary feeding. We reinforced at each visit the supplementation of the infants with vitamin D (400 IU per day) and elemental iron (2–3 mg/kg per day), which was started from 2 weeks of age as part of the study hospitals’ clinical policy. Subsequently, we called the infants for hospital visit at corrected age 5, 6, 7, 9, and 12 months to measure


Articles


anthropometry, morbidity inquiry, and 24 hour dietary recall (appendix p 4). Additionally, we provided families in both groups with 24-h, 7-day per week need-based telephone support. At the 12 month visit, we did neurodevelopmental assessments, DXA, and fasting blood sampling as part of final outcome assessment. If an infant could not be brought to the hospital, the study team made a home visit to measure outcomes, except blood sampling and DXA. We measured compliance to intervention through telephone calls made on days 2, 7, 14, 21, and 28 at 4 months and again at 6 months of corrected age in both groups, which was also corroborated by the daily dietary record maintained by the mothers.

We defined infants as receiving allocated intervention if they were offered complementary feeding within 4 weeks of counselling for initiation of food, irrespective of amount consumed. An additional criterion in the 6 month group was that if complementary feeding was started before the scheduled 6 months corrected age, it should not have been started more than a week before 6 months corrected age. The intrauterine growth category at birth and the Z scores for weight, length, and head circumference at birth, discharge, and 40 weeks of postmenstrual age were based on revised Fenton’s charts 2013,10 calculated using the provided anthropometric software for research data. Nutrient intakes were calculated from 24 dietary recalls using the nutritive value of Indian foods.11 We also calculated important dietary indicators related to infant feeding at 9 and 12 months of age based on the information recorded during the 24-hour dietary recalls based on WHO definitions (appendix p 4).12 We also noted the consistency and texture of food, which was based on expert assessment due to absence of standard criteria. Z scores for weight, length, head circumference, weight for length, and body–mass index at subsequent timepoints, (ie, 4, 5, 6, 7, 9, and 12 months corrected age) were calculated using WHO-MGRS growth standards.9

OutcomesThe primary outcome was WAZ12, based on WHO-MGRS growth standards.9 The infant was weighed using an electronic weighing scale (Seca, Germany; accuracy 5 g) and the weight was converted into respective Z score using WHO Anthropometric software (version 3.2.2, 2011), using corrected age. Secondary outcomes were any morbidity requiring hospital admission from the time of enrolment until 12 months corrected age, neurodevelopment, body composition, bone mineral content (BMC) and bone mineral density (BMD), insulin resistance in terms of HOMA-IR (HOmeostatic Model Assessment for Insulin Resistance), lipid profile, blood pressure, and serum ferritin at 12 months corrected age. Hospital admission was defined as admission to an inpatient facility for duration of 6 h or more or as inpatient death irrespective of duration of admission. A

paediatrician verified the diagnosis for each episode by reviewing the case records and interaction with the treating physician where required. Repeat hospital admissions for an infant should have been separated by more than a week to count as separate episodes. Neurodevelopment was assessed by a single certified clinical psychologist using Developmental Assessment Scale for Indian Infants (DASII), a validated Indian adaptation of Bayley-II.13 Whole body composition, BMC, and BMD were assessed using DXA (Hologic DISCOVERY W, S/N 84879, version 13.1.1:7; software Apex Version 3.0). HOMA-IR was calculated using the formula:14

Fasting (minimum 4 h) plasma glucose was measured by Enzymatic Colorimetric Test Method (GOD-PAP method) without deproteinisation based on enzyme glucose oxidase, using Glucose PAP Fluid Mono reagent (Centronic GmBH, Germany), on ROCHE Modular P-800 fully automatic analyser. Serum insulin and ferritin were measured by electrochemiluminescence immunoassay (ECLIA) using RocheCobase411 (Roche Diagnostics, Germany). A value of serum ferritin less than 12 was taken as cutoff for depleted iron stores.15 For lipid profile, fasting total cholesterol and triglycerides were assessed using enzymatic methods and high density lipoprotein by direct method, on ROCHE P-800 fully automatic analyser. Very low density lipoprotein was calculated using triglyceride value and low density lipoprotein was calculated using total cholesterol, high density lipoprotein, and triglyceride values using Friedewald equation.16 Blood pressure was measured using arm-type fully automatic digital blood pressure monitor (AG-SafeCHEK™, model AG1010; accuracy 3 mm Hg), based on the oscillometric method, using reusable blood pressure cuff (WelchAllyn® Flexiport™ infant size number 7). Additionally, serum C-reactive protein was measured using high sensitivity C- reactive protein Enzyme Immunoassay (BioCheck, USA). This was done to better reflect on the value of serum ferritin, which is also an acute phase reactant.

Statistical analysisWe based our sample size estimates on the data on WAZ12 for infants less than 34 weeks of gestation from a birth cohort study of preterm infants from our own unit, which measured growth outcomes using WHO-MGRS growth standards,16 prospectively at 3 monthly intervals from birth till 18 months of age (WAZ12: 1·7 ± 1·5, n=15; personal communication, Sharma P).17 Assuming an effect size of 0·5 SDS, β of 0·9, and a two-sided α of 0·05, the sample size was 190 in each group. Accounting for 5% loss to follow-up, total sample size was calculated

fasting insulin (μU/mL) × fasting glucose (mmol/L)

HOMA-IR = 22·5


Articles


to be 400. The calculated sample size was further corroborated through a pilot study (infants in pilot phase were not included in this study).

Data were entered in duplicate in an online database developed in Visual Basic as front-end and MS SQL server as back-end with inbuilt range and logical checks, with audit trail. Analysis was done using STATA 11·0 (College Station, TX, USA), by intention to treat. Continuous variables were compared using Student t test for normally

distributed and Wilcoxon rank-sum for non-normally distributed data. Proportions were compared using chi-square test. We constructed a Kaplan-Meir survival curve to depict the first episode of hospital admission in each group during the study period including available information on all infants, and calculated the hazard ratio between the two groups using Cox proportional hazard regression. We also used the Anderson-Gill model, which is counting process extension for Cox proportional hazard

2135 livebirths at less than 34 weeks gestation

206 allocated to receive complementary feeding at 4 months corrected age 197 allocated to receive complementary feeding at 6 months corrected age

184 analysed for primary outcome 189 analysed for primary outcome

1677 potentially eligible at birth

1089 discharged

412 eligible at 4 months corrected age

403 enrolled

403 randomised

429 ineligible 344 residence >60 km from facility 68 major malformation or congenital infection 2 planned to move away 15 required hospital admission for more than 40 weeks post-menstrual age 29 other exclusions 5 unreliable gestation 2 medico-legal case 1 mother psychiatric patient 16 left against medical advice 5 refused follow-up at facility

114 died 35 already started on complementary food 140 permanently moved away 102 unable to follow up at facility 30 not contactable 1 operation for Hirschsprung disease 2 congenital hydrocephalus 1 toxoplasmosis 11 congenital heart disease 1 under work-up for biliary atresia 1 syndromic 3 accident 3 given for adoption 4 admitted to hospital at time of enrolment 1 infant on orogastric feeds 6 family or social reasons 3 missed appointment 1 mother sick at time of enrolment 218 less than 4 months corrected age at end of enrolments

578 died 5 transferred to another ward 5 admitted at the time of concluding enrolments

7 did not receive allocated intervention 5 started complementary feeding between 5 and 6 months 1 started complementary feeding between 6 and 7 months 1 did not start complementary feeding before death, which occurred between 5 and 6 months

9 refused consent

9 did not receive allocated intervention 1 started complementary feeding earlier 2 started complementary feeding between 7 and 9 months 5 started complementary feeding between 9 and 12 months 1 did not start complementary feeding until 12 months

22 excluded from primary outcome analysis 2 withdrew consent 16 lost to follow-up 4 died

8 excluded from primary outcome analysis 2 died 6 lost to follow-up

Figure 1: Trial profile


Articles


regression to calculate the hazard ratio between the two groups to account for multiple episodes of hospital admission for any infant.

A prespecified subgroup analysis was done by site, intrauterine fetal growth category (small-for-gestational-age, appropriate-for-gestational-age), gestational age at birth (<28 weeks, 28–30 weeks and 31–33 weeks), and type of feeding at randomisation (breastfed, non-breastfed, mixed fed). Generalised estimating equation analysis was used to compare anthropometry trend between the two random groups over time. We accounted for the clustering effect on the primary outcome due to corandomisation of twins and triplets by reanalysing the primary outcome after dropping multiple births (retaining first twin only) in both groups. Study was supervised by an independent Doctoral Committee that reviewed the processes every 6 months. An independent safety adviser also reviewed the data on mortality and morbidity.

Role of funding sourceThe Indian Council of Medical Research had no role in study design, data collection, analysis, interpretation, or writing of the report. The corresponding author had full access to all the data in the study and had final responsibility for the decision to submit for publication.

ResultsBetween March 20, 2013, and April 24, 2015, we identified 2135 livebirths at less than 34 weeks gestation. 412 were eligible at 4 months corrected age, nine of whom refused consent, and 403 were randomly assigned—206 infants to the 4 month group and 197 infants to the 6 month group (figure 1). Overall, a total of 22 infants in the 4 month group (a pair of twins had consent withdrawn, 16 infants were lost to follow-up, and four died) and eight infants in the 6 month group (six infants were lost to follow-up and two died) were excluded from the analysis of the primary outcome. However, deaths were included in the analysis as secondary outcome.

196 of 203 infants in the 4 month group (two withdrew consent, one died before complementary feeding could be initiated) and 184 of 193 infants in the 6 month group (two died and two lost to follow-up before complementary feeding could be initiated) received the allocated intervention.

The baseline characteristics of study infants at birth and at randomisation were similar (table 1). Mean age at randomisation was 3·9 months corrected age in both the groups, and mean weight and WAZ were 5117 g in the 4 month group versus 5187 g in the 6 month group and –2·3 in the 4 month group versus –2·2 in the 6 month group. The proportion of infants receiving any breastfeeding was similar between two groups at all timepoints (appendix p 5). A high proportion of infants were receiving complementary feeding (irrespective of amount) 1 month after counselling for initiation of food (95·4% vs 90·7%; appendix p 5).

4 month group (n=204) 6 month group (n=197)

Maternal age, years (SD) 26·9 (4·9) 26·4 (4·1)

Paternal age, years (SD) 30·5 (5·5) 30·1 (4·7)

Maternal education

Professional, graduate, post-graduate 58 (28·5%) 44 (22·3%)

Intermediate, post-secondary diploma, high school 68 (33·3%) 72 (36·6%)

Middle, primary 61 (29·9%) 55 (27·9%)

Illiterate 17 (8·3%) 26 (13·2%)

Family income , thousand Indian rupees per month (IQR)

10 (7–15) 10 (8–15)

Infant characteristics at birth

Gestation, weeks (SD) 31·7 (1·4) 31·5 (1·7)

Gestation category

<28 weeks 4 (2·0%) 6 (3·1%)

28–31 weeks 66 (32·4%) 68 (34·5%)

32–33 weeks 134 (65·7%) 123 (62·4%)

Birthweight, g (SD) 1479 (308) 1492 (344)

Birthweight Z score (based on weight for gestational age at birth)*

–0·84 (0·71) –0·73 (0·76)

Birthweight category

<1000 g 13 (6·4%) 13 (6·6%)

1000–1499 g 79 (38·7%) 77 (39·1%)

≥1500 g 112 (54·9%) 107 (54·3%)

Small for gestational age* 58 (28·4%) 54 (27·4%)

Multiple births 55 (27·0%) 48 (24·3%)

Female sex 109 (53·4%) 93 (47·2%)

Antenatal steroids received 169 (85·4%) 169 (86·2%)

Duration of NICU stay, days (IQR) 7 (4–12) 6 (4–12)

Infant characteristics at randomisation

Corrected age (months) 3·9 (0·1) 3·9 (0·1)

Weight (g) 5117 (906) 5187 (928)

Weight for age (Z score) –2·3 (1·4) –2·2 (1·4)

Length (cm) 58·5 (3·1) 58·6 (3·2)

Length for age (Z score) –2·0 (1·4) –2·0 (1·5)

Head circumference (cm) 38·9 (1·6) 39·0 (1·4)

Head circumference for age (Z score) –1·7 (1·2) –1·7 (1·2)

Weight for length Z score –0·96 (1·1) –0·9 (1·1)

BMI (kg/m²) 8·7 (1·2) 8·8 (1·2)

BMI for age Z score –1·5 (1·2) –1·5 (1·2)

Body composition

Fat mass (g) 1259 (746), n=166 1242 (662), n=149

Lean�+�BMC mass (g) 4346 (836), n=166 4462 (715), n=149

Total mass (g) 5624 (965), n=166 5690 (985), n=149

Percent fat (%) 21·7 (11·2), n=166 21·1 (9·3), n=149

BMC (g) 95·7 (24·6), n=173 93·8 (23·0), n=156

Bone mineral density (g/cm2) 0·170 (0·028), n=173 0·167(0·022), n=156

Method of feeding

Only breastfeeding 104 (51·2%) 98 (49·8%)

Mixed feeding (breastmilk and animal or formula milk)

61 (30·1%) 55 (27·9%)

Exclusively top-fed (no breastfeeding, only animal or formula milk)

38 (18·7%) 44 (22·3%)

Data are n/N(%), or mean (SD), unless otherwise specified. NICU=neonatal intensive care unit. BMI=body–mass index. BMC=bone mineral content. *Using Fenton growth charts 2013.

Table 1: Baseline characteristics


Articles


The primary outcome, WAZ12, was similar between the two groups (mean –1·6, SD 1·2 in the 4 month group vs –1·6, 1·3 in the 6 month group; mean difference 0·005, 95% CI –0·24 to 0·25; p=0·96) at mean corrected age of

12·2 months (table 2). The change in WAZ in study infants from birth until 12 months corrected age (p=0·61; figure 2) and specifically between 4 and 12 months of corrected age was similar (p value on generalised estimating equation

4 month group 6 month group Mean difference or risk ratio (95% CI)

p value

Primary outcome

Weight for age (Z score) –1·6 (1·2), n=184 –1·6 (1·3) n=189 0·005 (–0·24 to 0·25) 0·965

Secondary outcomes

Death 4/203* 2/197 1·9 (0·4 to 10·5) 0·685

Hospital admission

Infants, n (%)† 34/188 (18·1%) 18/192 (9·4%) 1·9 (1·1 to 3·3) 0·014

Episodes per infant-month (IR) 39/1590 (0·025) 23/1606 (0·014) 1·7 (1·0 to 3·0)‡ 0·039

Diagnosis

Diarrhoea (with or without dehydration)

11 6 ·· ··

LRTI 16 11 ·· ··

Both diarrhoea and LRTI 1 1 ·· ··

Sepsis 3 1 ·· ··

Other 7 3 ·· ··

Unclear 1 1 ·· ··

Neurodevelopment

MoDQ50 84·0 (15·4), n=182 83·8 (14·0), n=184 0·2 (–2·9 to 3·2) 0·918

MoDQ50<70 27 (14·8%), n=182 22 (12·0%), n=184 1·2 (0.7 to 2.1) 0·552

MoDQ97 104·3 (18·6), n=182 102·6 (16·8), n=184 1·7 (–1·9 to 5·4) 0·351

MeDQ50 89·0 (12·4), n=181 89·2 (11·5), n=184 –0·3 (–2·8 to 2·1) 0·786

MeDQ50<70 12 (6·6%), n=181 12 (6·6%), n=182 1·0 (0.5 to 2.2) 0·713

MeDQ97 108·7 (18·4), n=181 109·0 (14·9), n=184 –0·2 (–3·7 to 3·2) 0·893

Body composition

Fat mass (g) 2056 (714), n=134 2128 (762), n=135 –72 (–250 to 105) 0·423

Lean + BMC mass (g) 6182 (805), n=134 6265 (922), n=135 –84 (–292 to 124) 0·428

Total mass (g) 8234 (1129), n=134 8427 (1208), n=135 –193 (–474 to 88) 0·178

Percent fat (%) 24·5 (6·8), n=134 25·3 (6·7), n=135 –0·8 (–2·4 to 0·8) 0·329

BMC(g) 186·0 (35·7), n=135 191·8 (32·9), n=135 –5·7 (–13·9 to 2·5) 0·173

Bone mineral density (g/cm2) 0·25 (0·03), n=135 0·25 (0·03), n=135 –0·001 (–0·008 to 0·006) 0·770

Lipid profile (mg/dL)

Total cholesterol 141·4 (33·1), n=161 141·8 (32·9), n=173 –0·4 (–7·5 to 6·7) 0·919

Triglycerides 123·4 (61·6), n=160 125·1 (67·8), n=173 –1·6 (–15·6 to 12·4) 0·818

HDL 37·9 (12·3), n=161 38·3 (12·1) n=173 –0·4 (–3·1 to 2·2) 0·758

LDL 77·3 (27·9), n=160 76·8 (28·0), n=172 0·5 (–5·5 to 6.5) 0·870

VLDL 25·8 (12·1), n=160 26·3 (13·7), n=173 –0·5 (-3·3 to 2·3) 0·731

HOMA-IR* 0·4 (0·3–0·7), n=153 0·4 (0·2–0·7), n=166 .. 0·675

Blood pressure (mm Hg)

Systolic 80·9 (6·6), n=135 80·9 (6·2), n=149 0·05 (–1·4 to 1·5) 0·951

Diastolic 50·1 (5·8), n=135 50·4 (5·3), n=149 –0·3 (–1·6 to 1·0) 0·632

Serum ferritin (µg/dL ) 5·4 (3·2–12·4), n=160 5·7 (2·5–13·3), n=173 ·· 0·732

Serum ferritin <12 µg/dL, n (%) 119/160 (74·4%) 126/173 (72·8%) 1·0 (0·9 to 1·2) 0·750

Data are mean (SD), n/N (%), or median (IQR). p values are from Fisher-exact test, t-test, Wilcoxon rank-sum or χ² test. IR=incidence rate. LRTI=lower respiratory tract infection. MoDQ50=motor developmental quotient–50th centile. MoDQ97=motor developmental quotient–97th centile. MeDQ50 =mental developmental quotient–50th centile. BMC=bone mineral content. VLDL=very low density lipoprotein. HOMA-IR=Homeostatic Model Assessment for Insulin Resistance.*Outcome of death was known for all enrolled infants except for one infant in the 4 month group. †Outcome of hospital admission was known for 188 infants in 4 month group (184 who completed follow-up and 4 who died), and for 192 infants in the 6 month group (189 who completed follow-up, 2 who died, and 1 infant who was lost to follow-up but had a documented episode of hospital admission before being lost). ‡This is incidence rate ratio. Primary outcome and all secondary outcomes except deaths and hospital admissions were measured at 12 months corrected age. Death and hospital admission data is for the entire study duration, ie, from enrolment at 4 months corrected age until final outcome assessment at 12 months corrected age.

Table 2: Primary and secondary outcomes


Articles


[GEE] analysis, p=0·836). Mean weight (7794 vs 7846 g, p=0·65) at 12 months corrected age was also similar (appendix p 5). Additional growth data is provided in the appendix (pp 6–7).

Six infants died during the study period; four of 203 in the 4 month group versus two of 197 in the 6 month group (risk ratio [RR] 1·9, 95% CI 0·4–10·5; table 2). 34 (18%) of 188 infants in the 4 month group required hospital admission during the study period, compared with 18 (9%) of 192 infants in the 6 month group (RR 1·9, 95% CI 1·1–3·3; p=0·01), with 2·5 episodes per 100 infant-months in the 4 month group versus 1·4 episodes per 100 infant-months in the 6 month group (incidence rate ratio 1·7, 95% CI 1·0–3·0; p=0·04, table 2). The Kaplan-Meier curve for the time to first episode of hospital admission showed that the mean first hospital admission for the 4 month group was earlier compared with the 6 month group (p=0·02), with the

risk of hospital admission 48% lower in the 6 month compared with the 4 month group (hazard ratio [HR] 0·52, 95% CI 0·29–0·92; p=0·025) during the study period after adjusting for randomisation group, site, and gestation. The results remained similar when multiple episodes of hospital admission for any infant were taken into account using the Anderson-Gill model (0·56, 0·33–0·94; p=0·029, figure 3).

The motor and mental development quotients in the two groups were similar (table 2). There was no significant difference in body composition, BMC, BMD, lipid profile, HOMA-IR, blood pressure, or serum ferritin between the two groups. Median serum ferritin was 5·4 µg/dL (IQR 3·2–12·4) in the 4 month group and 5·7 (2·5–13·3) in the 6 month group (p=0·73), with almost two-thirds of infants in both groups obtaining a result less than 12 µg/dL (table 2).

A prespecified subgroup analysis by site, intrauterine fetal growth category, gestational age at birth, and type of feeding at randomisation (breastfed, non-breastfed, mixed fed) did not present any difference in the primary outcome (table 3).

Dietary diversity was acceptable in less than two-thirds of infants in either group (60·1% vs 55·0%, p=0·32; table 4). Acceptable minimum meal frequency was present in most infants (93·4% vs 92·6%; p=0·75). However, less than two-thirds of infants were receiving minimum acceptable diet. A high proportion of infants in both groups were bottle-fed (table 4).

DiscussionOur study shows that initiating complementary feeding at an earlier age of 4 months compared with 6 months of corrected age resulted in similar WAZ scores and other growth outcomes, body composition, bone mineralisation status, iron stores (with iron supplementation), and markers of metabolic syndrome at 12 months corrected age. However, there was a higher risk of hospital admission in the group with earlier initiation, suggesting that initiation of complementary feeding at 6 months corrected age might be preferable to 4 months corrected age in infants born at less than 34 weeks of gestation.

This evidence can be considered robust, as the study was adequately powered to detect the outcome of interest and met all criteria for high internal validity: it was a randomised trial with groups being similar at the start, groups were treated similarly except for allocated intervention, no co-intervention was administered, the overall follow-up rate was high, analyses were by intention to treat and the outcomes were either blinded or objective.

To our knowledge, only one trial published earlier compared the effect of initiation of complementary feeding on growth of preterm infants. The study had several limitations and concluded that there was no difference in the anthropometric parameters between the two intervention groups at 12 months of age.

–0·8

–2·3–2·0

–1·9 –1·8 –1·7 –1·6–1·9

–2·8

Birth

p between random groups over time = 0·61

Discharge 40 weeksPMA

4 months 5 months 6 months 7 months 9 months 12 months

–4·0

–4·5

–5·0

Wei

ght f

or a

ge Z

scor

e

Corrected age (months)

–3·5

–3·0

–2·5

–2·0

–1·5

–1·0

–0·5

0

–0·7

–2·2–2·1 –1·9 –1·8

–1·7 –1·6–1·8

–2·8

0·5

4 month6 month

Figure 2: Change in weight for age Z score among study infants over time by study group.*Includes data only for infants who completed 12 month follow-up; data are mean (SD). PMA=postmenstrual age.

Number at risk4 month group6 month group

0 2 4 6 8 10 12

204197

191189

173182

163175

120140

22

00

Time to death or hospital admission since randomisation (months)

0

25

50

75

100

Surv

ival

est

imat

es (%

)

4 month group6 month group

Figure 3: Corrected age at first hospital admission among study infants during the study period by study group


Articles


Two other studies—one in full-term, healthy and one in low birthweight, exclusively breastfed infants—have compared the effect of initiation of complementary feeding at 4 versus 6 months of age.18,19 Both studies showed that the intervention resulted in no significant difference in weight or length at 12 months of age, similar to our study. Unfortunately, we could not measure breastmilk intake in our study. However, we did note that in breastfed infants, energy intake per kg bodyweight (excluding breastmilk) was higher in the 4 month group compared with the 6 month group at all timepoints, although it was similar among non-breastfed infants (except for a small difference at 6 months corrected age; appendix p 9). Since there was no difference in growth parameters between the two groups at any timepoint, it is likely that the breastfed infants in the 6 month group increased their breastmilk intake and were deriving the additional energy from breastmilk. Studies using robust isotopic methods have also shown increased breastmilk intake in full-term exclusively breastfed infants who continue to remain breastfed until 6 months of age, and decreased intake among those who start on complementary feeding at 4 months of age.20,21

The overall incidence of hospital admission in the study population was low, but infants in the 4 month group had more episodes of diarrhoea and lower respiratory tract infections until 12 months corrected age. Although these might be explained by potential contamination of complementary foods due to inadequate hygiene,22 a more biological rationale should also be considered. Breastmilk is known to confer immunological benefits to infants that are especially important for preterm infants,23 and breastmilk intake is likely to have been lower among infants in the 4 month group, as discussed earlier. Besides, the role of dietary exposures in shaping both short and long term immune function in infants might also be a factor.24,25 Little difference in growth patterns between the two groups with little catch up growth in either group might explain similar body composition and similar markers of metabolic syndrome between the groups. The mean fat mass and % fat at 12 months of corrected age in either group were close to that reported in literature, suggesting that the biology of growth among this group of infants is similar across settings.26 However, we need to be cautious with whole body DEXA as a measure as it is not able to accurately establish the aberrant adiposity that might occur.

Poor dietary practices among all study infants despite counselling are more difficult to understand, especially as available evidence suggests that counselling helps to improve feeding practices.27 We postulate that the practices would possibly have been worse without counselling as the indicators obtained in our study were better than reported in literature from similar settings.28 It appears that to achieve the recommended dietary standards in this population requires a degree of

behaviour change at the family level that is difficult to achieve, and requires innovative approaches to supplement the counselling. It is also important to consider that infant factors such as taste, preference, ability, or interest in taking feeds and pattern of eating could also have contributed to the dietary patterns seen in the study.

4 month group 6 month group Mean difference (95% CI)

p value

Study site

All India Institute of Medical Sciences

–1·5 (1·3), n=57 –1·21 (1·4),n=54 –0·3 (–0·8 to 0·2) 0·306

Safdarjung Hospital –1·7 (1·1), n=119 –1·75 (1·2), n=122 0·05 (–0·2 to 0·3) 0·743

Kasturba Hospital –1·4 (1·1), n=9 –2·1 (1·0), n=13 0·7 (–0·2 to 1·6) 0·110

Intrauterine fetal growth category

Small for gestational age –2·1 (1·2), n=52 –2·3 (1·2), n=50 –0·2 (–0·3 to 0·7) 0·408

Appropriate for gestational age –1·4 (1·1), n=131 –1·4 (1·2), n=137 –0·03 (–0·3 to 0·2) 0·806

Gestational age at birth

<28 weeks –2·0 (1·1), n=4 –2·2 (1·9), n=6 0·2 (–2·2 to 2·7) 0·836

28–30 weeks –1·7 (1·2), n=58 –1·7 (1·4), n=67 –0·04 (–0·5 to 0·4) 0·856

31–33 weeks –1·5 (1·1), n=122 –1·5±1·1, n=116 0·002 (–0·3 to 0·3) 0·987

Feeding at complementary feeding randomisation

Breastfed –1·7 (1·1), n=95 –1·8 (1·2), n=93 0·1 (–0·2 to 0·4) 0·561

Non-breastfed –1·7 (1·4), n=33 –1·6 (1·4), n=42 –0·06 (–0·7 to 0·6) 0·846

Mixed food –1·5 (1·1), n=55 –1·4 (1·3), n=54 –0·1 (–0·5 to 0·4) 0·672

Data are mean (SD) unless otherwise specified. p values are from Student t test

Table 3: Prespecified subgroup analysis for primary outcome, weight-for-age Z score at 12 months of corrected age

4 month group 6 month group P-value

Bottle feeding 100/183 (54·6%) 110/189 (58·2%) 0·489

Number of food groups offered 3·7 (0·9), n=183 3·5 (1·0), n=189 0·037

Acceptable minimum dietary diversity* 110/183 (60·1%) 104/189 (55·0%) 0·321

Meal frequency

Breastfed infants (includes only non-liquid feeds)

4·2 (1·4), n=123 4·3 (1·6), n=118 0·762

Non-breastfed infants (includes both milk feeds and solid or semisolid feeds)

8·5 (2·1), n=60 9·0 (2·0), n=71 0·138

Acceptable minimum meal frequency

Breastfed infants (≥3) 112/123 (91·1%) 104/118 (88·1%) 0·457

Non-breastfed infants (≥4) 59/60 (98·3%) 71/71 (100%) 0·279

Minimum acceptable diet (%)† 108/184 (58·7%) 103/189 (54·5%) 0·406

Consumption of iron-rich or iron-fortified foods (%)

175/184 (95·1%) 174/189 (92·1%) 0·216

Consistency of complementary food-thick 182/182* (100%) 182/187* (97·3%) 0·085

Texture of complementary food-grainy 166/182 (91·2%) 167/187 (89·3%) 0·538

Data presented as number (%) or mean (SD) unless otherwise specified. P value is from χ² test or Student t test.*Proportion of infants receiving food from four or more food groups. †Breastfed infants who received food from four food groups or more, and solid, semisolid, or soft foods at least three times on previous day; non-breastfed infants who received at least 2 milk feeds, food from four food groups or more (not including milk), and 4 or more meals during the previous day.

Table 4: Dietary indicators at 12 month corrected age


Articles


Another important concern revealed by the study is the greatly depleted iron stores, despite most infants having received iron supplementation. This is a puzzling finding, and since we had strictly ensured the compliance, it requires some explanation. Studies have shown that iron bioavailability from habitual Indian diets is low, due to high phytate and low ascorbic acid to iron ratios. Additionally, food matrix effect and food synergies specific to the local context could have been an issue.29,30 We also presume that coadministration of iron supplements with food or milk could have resulted in decreased absorption of iron. Delayed cord clamping was not being practiced at any of the sites during the study period, and that could also be a reason in part. However, despite all of these postulations, this is a crucial outcome with a bearing on neurodevelopment of these vulnerable infants. Therefore, it is important to do further research to ascertain the adequacy of recommended doses of supplementation in preterm infants, factors that retard iron absorption, and possibly better iron preparations with enhanced bioavailability.

This study was done in a lower middle-income country setting, but we propose that the results could hold relevance even for high-income country settings for two main reasons. First, we observe that growth pattern of this group of infants is similar across settings, with universal postnatal growth restriction.31,32 In this study, the parameters of fat mass at 12 months of age were close to that reported in the literature.28 Second, the feeding or dietary patterns in infancy are also similar with low rates of exclusive breastfeeding and inadequate complementary feeding practices.33,34,35 However, other factors that might limit the generalisation of results to other settings and would need prior consideration include differences between developing and developed country settings with respect to the birthweight of preterm infants, prevalence rates of intrauterine growth restriction in the population, practices with respect to use of post-discharge, nutrient-enriched formula or animal milk, breastfeeding rates, fortification of complementary foods, background infection rates, and sociodemographic factors such as maternal education and socioeconomic status, which would influence hygiene and compliance to dietary advice.

We also report that the mean postnatal age of study infants at intervention was 5·7 months (SD 0·3) in the 4 month group, and 7·9 months (SD 0·4) in the 6 month group. Since the 4 month group did not result in improved outcomes but increased hospital admissions, it follows that 6 months of postnatal age might not be preferred over 6 months of corrected age for initiation of complementary feeding in preterm infants.

The limitations of the study are the open-label design and differential loss to follow-up in the two allocation groups. However, the outcome assessors were masked to the allocation groups, and the baseline characteristics of infants lost to follow-up in either group were similar. The diagnostic utility of any neurodevelopmental assessment

at 1 year of age is also inadequate, which might also be considered a limitation. There is a need for longer term follow-up of this cohort, and reassessment of their growth, development, micronutrient status, and markers of chronic disease.

Future research should focus on identifying an appropriate window, if one exists, to improve the postnatal growth of preterm infants. There is also an urgent need to identify the reasons for poor dietary patterns and greatly depleted iron stores despite supplementation in the preterm infants, and test potential interventions and institute preventive measures to target the same. Long-term follow-up of such infants, ideally till adulthood, is highly desirable.ContributorsSG, RA, VKP, AA, VS, VJ, MB, and MJS developed the protocol and study tools. RA, VKP, HC, SA, MJS, VS, SSai, SSap, AKD, SB, KCA, VJ, SKD, and SKG were responsible for study oversight. SG, SSai, MB, AS, RK, FM, KG, FK, GD, AT, PKP, RS, LG, BK, VY, NR, and SL were responsible for study implementation. GS, AS, RK, FM, KG, FK, SB, and GDD did data collection. SSur were responsible for data mangement. VS and CPY did data analysis. RA, HC, SA, KCA, MJS, SSai, SSap, MB, AKD, VKP, VS, AKG, and VJ did quality assurance. CKN, SG, and ML did systematic review for research in context. SG, RA, VKP, HC, MJS, VS, KCA, and SA wrote the manuscript. RA would act as overall guarantor.

Declaration of interestsThe authors declare no competing interests.

Investigators of CF trial (all from AIIMS)Anuja Aggarwala MSc, Savita Saini MSc, Savita Sapra PhD, Madhumati Bose PGD, Chandra Kumar Natarajan DM, Ajay Singh MSW, Reena Kuriakose MSW, Faizan Mujeeb MSW, Kanaklata Gupta PhD, Farah Khan MA, Sukhram Babu MSW, Garima Dhankar MPT, Somi Suresh MSc, Anne Therasa GNM, Pawan Kumar Popli, Ramesh Sharma, Lalit Gupta, Brijesh Kumar, Vikas Yadav MSc, Chander Prakash Yadav MPhil, Pratibha Gupta BCom, Nisha Rani BA, Meenakshi BCom, Sant Lal MBA

Doctoral Committee members (as per the AIIMS institutional norm, DC reviewed the progress of the study every 6 months and ensured scientific rigor and quality)Vinod K Paul, Govind Makharia, Chandrakant S Pandav, Ramesh Agarwal, Kailash Chandra Aggarwal, Harish Chellani, Sugandha Arya, Vishnubhatla Sreenivas, Vandana Jain, Arun Kumar Gupta, Ashok K Deorari, Rakesh Lodha

AcknowledgmentsWe thank Prof Vinod Bhutani (Stanford University, US), Prof Haresh Kirpalani (Children’s Hospital of Pennsylvania, US), and Prof Siddarth Ramji (Maulana Azad Medical College, Delhi) for expert advice on study protocol, Dr Nita Bhandari (Society for Applied Studies, Delhi) for her insightful guidance with respect to study design and development of complementary feeding counselling package and tools thereof, Dr Suparna Ghosh-Jerath and Ms Preeti (Indian Institute of Public Health, Delhi) for expert technical guidance and assistance in the development of CF counselling package, members of Doctoral Committee from AIIMS, Dr Govind Makharia and Prof CS Pandav for study oversight and valuable guidance, Prof R Lakshmy (AIIMS) for overseeing lipid analysis, Prof Mohd Irshad (AIIMS) for glucose estimation, Dr Rakesh Lodha (AIIMS) for being safety adviser, Dr Mercedes de Onis (World Health Organization) for sharing the WHOMGRS anthropometry procedures CD, Dr John Shepherd (UCSF School of Medicine, US) for guidance on DXA, Dr Rachna Sehgal (Safdarjung hospital, New Delhi, India), Dr Meenakshi Salhan (Safdarjung hospital, New Delhi, India), and Dr Shobhana Gupta (Safdarjung hospital, New Delhi, India) for helping with patient care, Mr Arun Kumar and Mr Tapish Pandey (AIIMS) for helping with the samples handling, transport, and processing, Dr Suman Chaurasia


Articles


(AIIMS) for his advice and support at various stages, nursing Tutor Levis (AIIMS) for helping with the systematic review, Dr Pradeep Sharma (Apollo Hospital, Ludhiana) for sharing his study data for sample size calculation, and Ms Rekha Pangal, Ms Geeta Mehto and Ms Tushara Philip for helping data collection. We owe special thanks to Dr K Madhavan Nair (Nutrition Society of India) who helped us with the reviewer’s queries on iron stores.

References1 Pan American Health Organization, Guiding Principles for

Complementary Feeding of the Breastfed Child. Washington DC: Pan American Health Organization, World Health Organization, 2003.

2 Agostoni C, Decsi T, Fewtrell M, et al. ESPGHAN Committee on Nutrition. Complementary feeding: a commentary by the ESPGHAN Committee on Nutrition. J Pediatr Gastroenterol Nutr 2008; 46: 99–110.

3 AAP policy statement. Breastfeeding and the use of human milk. Pediatrics 2005; 115: 496.

4 Cooke RJ, Ainsworth SB, Fenton AC. Postnatal growth retardation: a universal problem in preterm infants. Arch Dis Child Fetal Neonatal Ed 2004; 89: F428–30.

5 Horbar JD, Ehrenkranz RA, Badger GJ, et al. Weight growth velocity and postnatal growth failure in infants 501 to 1500 grams: 2000–2013. Pediatrics 2015; 136: e84–92.

6 Bauer J, Werner C, Gerss J. Metabolic rate analysis of healthy preterm and full-term infants during the first weeks of life. Am J Clin Nutr 2009; 90: 1517–24.

7 Weintraub V, Mimouni FB, Dollberg S. Effect of birth weight and postnatal age upon resting energy expenditure in preterm infants. Am J Perinatol 2009; 26: 173–77.

8 Marriott LD, Foote KD, Bishop JA, Kimber AC, Morgan JB. Weaning preterm infants: a randomised controlled trial. Arch Dis Child Fetal Neonatal Ed 2003; 88: F302–07.

9 WHO Multicentre Growth Reference Study Group. WHO Child Growth Standards based on length/height, weight and age. Acta Paediatr Suppl 2006; 450: 76–85.

10 Fenton RT, Kim HJ. A systematic review and meta-analysis to revise the Fenton growth chart for preterm infants. BMC Pediatrics 2013; 13: 59.

11 Gopalan C, Rama Sastri BV, Balasubramanian SC. Nutritive Value of Indian Foods (NVIF). Hyderabad: National Institute of Nutrition, 1989.

12 World Health Organization. Indicators for assessing infant and young child feeding practices: conclusions of a consensus meeting held 6–8 November 2007 in Washington DC, USA. WHO 2008. http://apps.who.int/iris/bitstream/10665/43895/1/9789241596664_eng.pdf (accessed March 12, 2016).

13 Pathak P. Baroda norms of motor and mental development of Indian babies from 1 to 30 months. Psychological stud 1993; 38: 142–49.

14 Matthews DR, Hosker JP, Rudenski AS, Naylor BA, Treacher DF, Turner RC. Homeostasis model assessment: insulin resistance and beta-cell function from fasting plasma glucose and insulin concentrations in man. Diabetologia 1985; 28: 412–19.

15 World Health Organization and Centers for Disease Control and Prevention. Assessing the iron status of populations: including literature reviews. http://www.who.int/nutrition/publications/micronutrients/anaemia_iron_deficiency/9789241596107.pdf (accessed March 25, 2016).

16 Friedewald WT, Levy RI, Fredrickson DS. Estimation of the concentration of low-density lipoprotein cholesterol in plasma, without use of the preparative ultracentrifuge. Clin Chem 1972; 18: 499–502.

17 Sharma PK, Sankar MJ, Sapra S, et al. Growth and neurosensory outcomes of preterm very low birth weight infants at 18 months of corrected age. Indian J Pediatr 2011; 78: 1485–90.

18 Cohen RJ, Brown KH, Canahuati J, Rivera LL, Dewey KG. Effects of age of introduction of complementary foods on infant breast milk intake, total energy intake, and growth: a randomised intervention study in Honduras. Lancet 1994; 344: 288–93.

19 Dewey KG, Cohen RJ, Brown KH, Rivera LL. Age of introduction of complementary foods and growth of term, low-birth-weight, breast-fed infants: a randomized intervention study in Honduras. Am J Clin Nutr 1999; 69: 679–86.

20 Wells JC, Jonsdottir OH, Hibberd PL, et al. Randomized controlled trial of 4 compared with 6 mo of exclusive breastfeeding in Iceland: differences in breast-milk intake by stable-isotope probe. Am J Clin Nutr 2012; 96: 73–79.

21 Nielsen SB, Reilly JJ, Fewtrell MS, Eaton S, Grinham J, Wells JC. Adequacy of milk intake during exclusive breastfeeding: a longitudinal study. Pediatrics 2011; 128: e907–14.

22 Weisstaub G, Uauy R. Non-breast milk feeding in developing countries: challenge from microbial and chemical contaminants. Ann Nutr Metab 2012; 60: 215–19.

23 Gregory KE, Walker WA. Immunologic factors in human milk and disease prevention in the preterm infant. Curr Pediatr Rep 2013; 1: 222.

24 Thompson AL, Monteagudo-Mera A, Cadenas MB, Lampl ML, Azcarate-Peril MA. Milk and solid-feeding practices and daycare attendance are associated with differences in bacterial diversity, predominant communities, and metabolic and immune function of the infant gut microbiome. Front Cell Infect Microbiol 2015; 5: 3.

25 Walker WA, Iyengar RS. Breast milk, microbiota, and intestinal immune homeostasis. Pediatr Res 2015; 77: 220–28.

26 Cooke RJ, Rawlings DJ, McCormick K, et al. Body composition of preterm infants during infancy. Arch Dis Child Fetal Neonatal Ed 1999; 80: F188–91.

27 Shi L, Zhang J. Recent evidence of the effectiveness of educational interventions for improving complementary feeding practices in developing countries. J Trop Pediatr 2011; 57: 91–98.

28 Senarath U, Agho KE, Akram DE, et al. Comparisons of complementary feeding indicators and associated factors in children aged 6–23 months across five South Asian countries. Matern Child Nutr 2012; 8 (suppl 1): 89–106.

29 Nair KM, Iyengar V. Iron content, bioavailability & factors affecting iron status of Indians. Indian J Med Res 2009; 130: 634–45.

30 Nair KM, Augustine LF. Food synergies for improving bioavailability of micronutrients from plant foods. Food Chemistry, Sept 19, 2016. http://www.sciencedirect.com/science/article/pii/S0308814616315102 (accessed March 12, 2016).

31 Hack M, Schluchter M, Cartar L, Rahman M, Cuttler L, Borawski E. Growth of very low birth weight infants to age 20 years. Pediatrics 2003; 112: e30–38.

32 Gladstone M, Oliver C, Van den Broek N. Survival, morbidity, growth and developmental delay for babies born preterm in low and middle income countries - a systematic review of outcomes measured. PLoS One 2015; 10: e0120566.

33 Shealy KR, Scanlon KS, Labiner-Wolfe J, Fein SB, Grummer-Strawn LM. Characteristics of breastfeeding practices among US mothers. Pediatrics 2008; 122 (suppl 2): S50–55.

34 Lande B, Andersen LF, Baerug A, et al. Infant feeding practices and associated factors in the first six months of life: the Norwegian infant nutrition survey. Acta Pediatr 2003; 92: 152–61.

35 Grummer-Strawn LM, Scanlon KS, Fein SB. Infant feeding and feeding transitions during the first year of life. Pediatrics 2008; 122 (suppl 2): S36–42.


Alimentación complementaria a los 4 versus 6 meses de ...³n_complemen… · alimentación...

Documents

Transcript of Alimentación complementaria a los 4 versus 6 meses de ...³n_complemen… · alimentación...