Preterm newborn readiness for oral feeding : systematic review and meta ‐ analysis

Purpose: To identify and systematize the main studies on the transition from enteral to oral feeding in preterm infants. Research strategy: Articles that describe the transition from oral to enteral feeding in preterm infants were located in MEDLINE, LILACS, and SciELO databases. Selection criteria: Original studies, with available abstract, published in the last 10 years were included. Data analysis: Analysis of the methodology and the main results of the studies, and meta-analysis of the effects of sensory-motor-oral stimulation at the time of transition to full oral feeding and duration of hospitalization were conducted. Results: Twenty-nine national and international publications were considered. Most studies were clinical trials (44.8%) and did not use rating scales to start the transition process (82.7%). In the meta-analysis, positive effect of stimulation of the sensory-motor-oral system was observed with respect to the transition time to oral diet  (p=0.0000), but not in relation to the length of hospital stay  (p=0.09). However, heterogeneity between studies was found both in the analysis of the transition time to full oral feeding (I=93.98) and in the length of hospital stay (I=82.30). Conclusion: The transition to oral feeding is an important moment, and various physical and clinical characteristics of preterm infants have been used to describe this process. Despite the impossibility of generalizing the results due to the heterogeneity of the studies, we have noted the importance of strategies for stimulation of sensory-motor-oral system to decrease the period of transition to full oral feeding system. DOI: 10.1590/2317-1782/20152014104 102 Lima AH, Côrtes MG, Bouzada MCF Friche AAL CoDAS 2015;27(1):101-7 INTRODUÇÃO A prematuridade pode ser entendida como uma condição de risco para o recém-nascido (RN), pois a imaturidade anatomofisiológica associada ao nascimento prematuro predispõe o RN a uma série de dificuldades de adaptação à vida pós-natal. É comum que, no início da vida, o recém-nascido prematuro (RNPT) não seja capaz de se alimentar por via oral (VO), sendo necessário o uso de métodos alternativos de alimentação até que estejam aptos a iniciar a alimentação por via oral. Algumas características são descritas como associadas à dificuldade no estabelecimento da alimentação oral no RNPT, tais como: imaturidade fisiológica e neurológica, hipotonia muscular, hiperreatividade aos estímulos do meio ambiente, instabilidade no controle da temperatura corporal, distúrbios respiratórios, curtos períodos de alerta, síndromes e alterações cardiorrespiratórias. Além disso, é comum a imaturidade do sistema sensório motor oral, que pode levar a uma incapacidade de sugar e/ou coordenar as funções sucção, deglutição e respiração, com impacto negativo nas habilidades de alimentação oral. Como resultado, os RNPT são tradicionalmente alimentados por gavagem até estarem aptos a se alimentar por via oral com sucesso. O processo de transição da alimentação gástrica para via oral é um dos grandes desafios que o RNPT deve superar depois de obtida sua estabilidade fisiológica. A dificuldade no processo de transição da alimentação enteral para alimentação oral ocorre em mais de 30% dos prematuros. Na prática clínica, percebe-se que existe dificuldade em se precisar o momento ideal para o início da transição da alimentação gástrica para VO. O processo de transição para alimentação oral se efetiva de forma bastante distinta entre os serviços de cuidado ao RNPT, e pouco se sabe sobre as técnicas e vias de administração de dieta utilizadas durante esse período de transição. Observa-se que, muitas vezes, os principais indicadores utilizados para transição da alimentação gástrica para VO são critérios físicos e isolados, como a idade gestacional (IG), a corrigida ou o peso. No entanto, esses critérios, isoladamente, podem ser incompletos para indicar a prontidão para alimentação oral, sendo sugerido que outros fatores como saturação de oxigênio e frequência cardíaca durante a alimentação, grau de estabilidade clínica, maturação e desempenho do RN na sucção não nutritiva (SNN) sejam observados. A decisão pelo momento de iniciar o desmame da sonda ainda não apresenta respostas unânimes, sendo necessária a participação de toda a equipe na definição de critérios. Protocolos de avaliação foram criados para auxiliar os profissionais de saúde na decisão de quando iniciar a alimentação por VO. A Academia Americana de Pediatria define a habilidade de alimentação oral plena como um dos critérios para a alta hospitalar de RNPT. A antecipação do início da alimentação oral em prematuros traz benefícios econômicos e para a saúde física e emo cional dos RN, destacando-se, dentre eles, a redução no período de internação hospitalar, a antecipação do aleitamento materno, a redução do tempo de uso de sonda gástrica e, consequentemente, das complicações advindas do seu uso. Algumas estratégias são utilizadas para melhorar as habilidades motoras orais do RNPT, tais como a SNN e programas de estimulação oral. Essas técnicas são descritas como importantes por contribuírem para que o RNPT atinja a alimentação oral independente mais precocemente, diminuindo o tempo de hospitalização, reduzindo custos médicos hospitalares e permitindo maior interação e o vínculo mãe-bebê. Dessa maneira, compreender o comportamento do RNPT no momento da transição da alimentação enteral para alimentação oral plena e promover a alimentação oral em momento adequado e seguro devem ser um dos objetivos das equipes que prestam assistência ao RNPT. Esta revisão sistemática foi realizada com o objetivo de identificar e sistematizar os principais estudos que descrevem o processo de transição da alimentação enteral para alimentação oral plena em RNPT e verificar o impacto de intervenções no sistema sensório motor oral sobre o tempo de transição para alimentação oral plena e sobre o tempo de permanência hospitalar. ESTRATÉGIA DE PESQUISA Trata-se de revisão sistemática da literatura, sendo que a metodologia adotada foi baseada em estudos anteriores e em recomendações sobre o tema. Foram realizadas buscas na biblioteca virtual BIREME para acesso à base de dados científicos LILACS, no portal PubMed para acesso aos periódicos indexados no MEDLINE, além de pesquisa na base SciELO. As estratégias de busca utilizadas para cada base de dados estão na Tabela 1. Para gerenciamento das referências foi utilizada a ferramenta EndNote Web 3.1. Após a localização dos estudos foi realizada triagem a partir do título e resumo dos artigos, processo realizado por dois avaliadores fonoaudiólogos, de forma independente. CRITÉRIOS DE SELEÇÃO Foram incluídos estudos que preenchessem os seguintes critérios: ser original com resumo disponível; ter sido publicado em língua portuguesa, espanhol ou inglês; descrever o momento de transição da alimentação enteral para alimentação VO em RNPT e ter sido publicado nos últimos dez anos (2003 a 2012). Foram excluídos artigos teóricos e de revisão, além de artigos que avaliavam RNPT que já haviam iniciado a alimentação oral (Figura 1). Os artigos deveriam preencher os critérios de inclusão, sendo que cada questão para inclusão ou exclusão do artigo permitiu, por parte dos avaliadores, três respostas: sim, não e talvez. Quando o artigo obteve apenas respostas “sim” e/ou “talvez” dos dois avaliadores, a publicação foi incluída. Por outro lado, a existência de respostas “não” pelos dois avaliadores excluiu o artigo da análise. Os resultados dos dois avaliadores foram comparados e, nos casos em que houve divergência quanto à inclusão ou exclusão do artigo, foi realizada uma reunião de consenso. Caso não houvesse consenso, era solicitado que o artigo fosse avaliado por um terceiro avaliador. Para avaliar a concordância entre os investigadores no momento da seleção dos artigos, foi realizada análise 103 Prontidão do prematuro para alimentação oral CoDAS 2015;27(1):101-7 de confiabilidade (Kappa). Para os artigos que foram selecionados na triagem prosseguiu-se com a leitura completa e coleta de dados. Finalmente, foi realizada a análise e discussão dos dados obtidos, além de metanálise dos ensaios clínicos em relação ao efeito da estimulação sensório motora oral nos dias de transição para alimentação oral plena e no tempo de permanência hospitalar. Foram coletadas e registradas informações necessárias para análise dos artigos, como base de dados pesquisada, nome do artigo e primeiro autor, ano de publicação, desenho do estudo, objetivo do estudo, local de realização e utilização de escalas de avaliação. Quanto aos resultados encontrados nos artigos, foram coletadas informações sobre o tamanho amostral e informações sobre o perfil dos RN, tais como sexo, idade gestacional (IG) de nascimento, peso de nascimento, nota de Apgar 5° minuto e número de dias em ventilação mecânica (VM). Além disso, foram coletados dados sobre o processo de transição da alimentação enteral para alimentação oral plena, tais como IG corrigida e peso no momento da primeira VO, indicadores das habilidades de alimentação oral, como proficiência (quantidade de leite ingerida nos primeiros cinco minutos/volume total de leite), eficiência (mL/min) e consumo (quantidade de leite ingerida/quantidade de leite prescrita), IG corrigida e peso na dieta oral plena, IG corrigida e peso na alta hospitalar, número de dias para a transição completa até alimentação VO plena e tipo de alimentação na alta, além do ganho ponderal durante todo o processo de transição. Tabela 1. Estratégias de busca por base de dados, realizadas nos meses de abril e maio de 2012 Base de dados Estratégia de busca MEDLINE (“Sucking Behavior”[Mesh] OR “Feeding Behavior”[Mesh:noexp] OR “Deglutition”[Mesh] OR “Bottle Feeding”[Mesh] OR “Breast Feeding”[Mesh] OR “Infant Nutritional Physiological Phenomena”[Mesh:noexp] OR “Sucking Behaviors” OR “Feeding Patterns” OR “Eating Behavior” OR “Feeding Behaviors” OR Deglutitions OR Swallowing OR Bot


INTRODUCTION
Prematurity may be understood as a risk condition for the newborn (NB), once the anatomical and physiological immaturity associated to preterm birth predisposes the NB to a series of difficulties in adapting to their postnatal life (1) .It is common that, in the beginning of life, the preterm newborn (PTNB) is not able to feed orally, being necessary the use of alternative feeding methods until they are able to take up oral feeding (2) .Some characteristics are described as associated to difficulties in establishing oral feeding for PTNB, such as physiological and neurological immaturity, muscle hypotonia, hyperreactivity to environmental stimuli, instability controlling body temperature, respiratory disorders, short alertness periods, cardiorespiratory syndromes, and alterations.Besides, the immaturity of the sensory-oral-motor system is common, which may lead to an inability to suck and/or coordinate the functions of sucking, swallowing, and breathing, with a negative impact on the oral feeding skills (3)(4)(5) .
As a result, the PTNB are traditionally fed by gavage until they are able to successfully feed orally (3,6) .The process of transition from gastric to oral feeding is one of the great challenges the PTNB must overcome after achieving their physiological stability.The difficulty in the process of transition from enteral to oral feeding occurs in more than 30% preterms (7) .
In clinical practice, it is noticeable that there is a difficulty in pinpointing the ideal moment to start the transition from gastric to oral feeding.The process of transition to oral feeding is carried out in a rather distinct way between the services of care to PTNB, and little is known about the techniques and dietary intake routes used during the transition periods (8) .It is observed that, many times, the main indicators used for the transition from gastric to oral feeding are physical and isolated criteria, such as gestational age (GA), the corrected one, and the weight (9) .However, these criteria alone may be incomplete to indicate the readiness for oral feeding, suggesting other factors such as oxygen saturation and heart rate during feeding, clinical stability degree, maturity, and the performance of the NB in nonnutritive sucking (NNS) (10) .
The decision about the moment when to start weaning from the probe still does not show unanimous answers, being necessary the participation of the whole team in the definition of the criteria.Evaluation protocols were created to aid professionals in taking the decision when to start oral feeding (11)(12)(13)(14)(15) .
The American Academy of Pediatrics defines the ability of full oral feeding as one of the criteria for hospital discharge of the PTNB (16) .The early start of oral feeding in the PTNB brings economic benefits as well as physical and emotional ones to the NB, among which, the reduction of time using gastric prone and, consequently, of complication resulting from its use (17) .Some strategies are used to improve motor and oral abilities of the PTNB, such as the NNS and oral stimulation programs.These techniques are described as important by contributing for the PTNB to achieve independent oral feeding earlier in time, reducing hospitalization time, reducing hospital medical costs, and allowing greater interaction and mother-child bond (18)(19)(20)(21)(22)(23)(24)(25)(26)(27) .
Thus, understanding the behavior of the PTNB at the moment of transition from enteral to full oral feeding and promoting oral feeding at appropriate and safe moments must be the objectives of the teams assisting the PTNB.
This systematic review was carried out with the objective of identifying and systematizing the main studies describing the process of transition from enteral to full oral feeding in the PTNB and verifying the impact of interventions on the motor and oral sensorial system on the time of transition to full oral feeding and on the time of hospital permanence.

RESEARCH STRATEGY
This is a systematic literature review, whose methodology was based on previous studies (28,29) and on recommendations on the theme (30,31) .
Searches were conducted in the BIREME virtual library for access to the scientific database of LILACS, at PubMed for access to indexed journals in MEDLINE, besides researches in SciELO database.The search strategies used for each database are given in Table 1.For the management of references, the EndNote Web 3.1 tool was used.
After locating the studies, a screening process was performed from the title and abstract of the articles, a process carried out by two speech language pathologist evaluators, independently.

SELECTION CRITERIA
The studies that met the following criteria were included: being original with available abstract; having been published in Portuguese, Spanish, or English; describing the moment of transition from enteral to oral feeding in the PTNB; and having been published within the last years (2003 to 2012).
Theoretical and review articles were excluded, along with articles evaluating the PTNB who had already initiated oral feeding (Figure 1).
The articles should fulfill the inclusion criteria, considering that each inclusion or exclusion question of the article allowed, from evaluators, three answers: yes, no, and maybe.When the article obtained only "yes" and/or "maybe" answers by both evaluators, then the publication was included.However, obtaining "no" answers by both evaluators led to the exclusion of the article from the analysis.The results from both evaluators were compared and, in the cases with disagreement as for the inclusion or exclusion of an article, a consensus meeting was carried out.In case of no consensus, it would be requested that the article was assessed by a third evaluator.To assess the agreement among investigators at the moment of selection of the articles, a reliability analysis was performed (Kappa) (12) .For the articles selected in the screening, we continued with their complete reading and data collection.Finally, an analysis and discussion of the data obtained was performed, in addition to the meta-analysis of clinical trials in relation to the motor and oral sensorial stimulation effect in the days of transition to full oral feeding and time of hospitalization.
All necessary information for the analysis of the articles were collected and registered, such as researched database, name of the article and first author, year of publication, design of the study, objective of the study, place where it was conducted, and use of assessment scales.
As for the results found in the articles, we collected information on the sample size and on the profile of the NB, such as gender, GA at birth, birth weight, Apgar 5-minute score, and number of days in mechanical ventilation (MV).Besides, we collected data on the process of transition from enteral to full oral feeding, such as corrected GA and weight at the time of the first oral feeding, indicators of the abilities in oral feeding, such as proficiency (amount of milk consumed within the first 5 minutes/total milk volume), efficiency (mL/min) and consumption (amount of milk consumed/prescribed amount of milk), corrected GA and weight in full oral diet, corrected GA and weight at hospital discharge, number of days until full transition to full oral feeding, and kind of feeding at discharge, besides weight gain during the entire transition process.

DATA ANALYSIS
To evaluate the agreement among the investigators, a reliability analysis was performed, resulting in Kappa (k) of 0.59, a reliability considered as moderate (32) .There was a disagreement among evaluators in 85 studies and, after the consensus meeting, 77 articles were included for complete reading and 29 were included in the final analysis.The data were analyzed in two stages.First, a descriptive analysis of the methodology and results was performed.Up next, a meta-analysis of the clinical trial studies was made, with the objective of verifying the impact OR "Mamadeira" OR "Fenômenos Fisiológicos da Nutrição do Lactente" OR "Conducta en la Lactancia" OR "Conducta Alimentaria" OR "Conduta na Alimentação" OR "Padrões Alimentares" OR "Eating Behavior$" OR "Feeding Pattern$" OR "Alimentación Artificial" OR "Lactancia Materna" OR "Amamentação" OR "Biberones" OR "Sucking Behavior" OR "Breast Feeding" OR "Infant Nutritional Physiological Phenomena" OR "Sucking Behaviors" OR "Feeding Behavior$" OR "Swallowing" OR "Bottle feeding") of the interventions in the motor and oral sensorial system on the time of transition to full oral feeding and on the time of permanence in the hospital.The Stata 10 and Comprehensive Meta-Analysis software were used (33) .The model of random effects was adopted for presenting a more conservative estimate than the models of fixed effects.The forest-plot graphic was used to summarize the estimates.The Q test and the I² index were used to evaluate the heterogeneity among studies and their magnitude, respectively.Percentages of the I² index of approximately 25% (I 2 ≤25), 50% (25<I 2 <75), and 75% (I 2 ≥75) were considered, respectively, low, medium, and high heterogeneity.The results presenting significance level ≤0.05 were considered statistically significant associations.
Most studies (82.7%) did not use standardized evaluation scales to determine whether the PTNB was able to take up on oral feeding at the moment of the evaluation.In these studies, the decision of beginning oral feeding was from doctors and legal guardians or defined rules were used to carry out the transition.The most often used scales in the studies (five studies) were the following: Neonatal Oral-Motor Assessment Scale (NOMAS) (35) , Preterm Infant Breastfeeding Behavior Scale (PIBBS) (46) , Clinical Pathway (20) , and the scale of the stages of suction (21) .
Therefore, despite being available in the literature, standardized protocols, with tested discriminatory reliability and validity (11)(12)(13)(14)(15) , in most studies, they were not used.Health services often use strict protocols, with previously established rules, which do not evaluate the particular development of each NB, which many times results in delays in starting oral diets and in the evolution to full oral feeding.It is believed that the use of validated scales provides a transition from prone to oral diets in a safer and sooner way, reducing the risks for the health of the PTNB (9,11,12,14) .

Variables used in the description of the transition process to oral feeding
Different physical and clinical characteristics of the PTNB were described during the process of transition from enteral to oral feeding: GA and weight at the time of the first oral feeding and in full oral feeding, efficiency, proficiency, and consumption in the first oral feeding, type of feeding at the time of hospital discharge, transition techniques into oral feeding, GA, weight, and days of life at the time of hospital discharge, and weight gain during the transition process to oral feeding.
In relation to the characterization of samples of the PTNB included in the studies, it was observed that all studies described the GA and birth weight of the included NB.The average GA at birth varied from 25.6 to 39.1 weeks and the average weight at birth varied from 815 to 3,564 grams (some studies included compared PTNB to regular NB, justifying the variability of GA and birth weight).The number of days in which NB remained in MV was a variable described in few studies (13.8%) (37,46,47,50) , as well as the 5-minute Apgar described only in 17.2% (8,21,27,44,47) .
The use of proficiency and/or efficiency measures was observed in 37.9% studies.The higher prevalence of observations of the abilities of oral feeding in the evaluation of the results in oral feeding may be justified by the ease of use of these criteria.Lau et al. (53) and Lau and Smith (54) defined the levels of oral feeding abilities from the observation of the combination between proficiency and efficiency, contributing to the determination of the development stage of abilities for oral feeding and facilitating the use of appropriate strategies to stimulate the motor and oral sensorial system.Owing to the ease to apply the evaluation, without the need of using specific tools, these evaluation criteria regarding the prematurity in the transition process to full oral feeding have been greatly used.
Due to the differences in relation to the methodology of the studies, some moments it was difficult to compare the results found.Besides, some of the studies included for the analysis of the methodology did not present important data, such as weight at the moment of the beginning of full oral feeding, number of days until full oral feeding, and corrected GA in full oral feeding.
In clinical trial studies, the effect of stimulation to the motor and oral sensorial system and other clinical approaches during the time of hospital permanence was compared, such as for the gain of weight during the transition process, efficiency, proficiency and consumption during feeding, in the stage and extent of suction, in frequency of suction and expression, in the duration of suction groups and feeding, in the number of adverse events during feeding, in the behavioral state, and in the behaviors of readiness to oral feeding, besides the volume of milk wasted during feeding (10,(20)(21)(22)(23)(24)(25)(26)(27)34,41,51,52) .
For the meta-analysis of the effects of stimulation of the motor and oral sensorial system in the number of days for the complete transition to oral feeding, five studies were included (21)(22)(23)(24)(25) .For the meta-analysis of stimulation effects of the motor and oral sensorial system at the time of hospitalization, three studies were included (22,24,34) .
In Figure 2, we observe a forest-plot graphic in which each line represents a study.For the studies that used more than one intervention group, we opted for performing the isolated analysis of each experimental group in relation to the control group of the study and, therefore, some studies are described in more than one line of the graphic due to the existence of more than one intervention group.The squares represent the standard difference between the averages in relation to the number of days for the transition to full oral feeding and the lines, their confidence intervals (CI).In the last line, represented by a rhombus, it is verified that, in the combinations of the results of the studies, there is statistical significance, showing that the NB who received some kind of motor and oral sensorial stimulation had a shorter time of transition to full oral feeding than the ones who did not have any kind of stimulation.However, there was a great heterogeneity among studies (Q=149.69;p=0.000;I²= 93.98).Figure 3 also shows a forest-plot graphic, in which each line represents an included study.Similar to Figure 2, some studies are described in more than one line because they presented more than one experimental group.The squares represent the standard difference among the averages in relation to the time of hospital permanence (days) and the lines, their GA.In the last line, represented by a rhombus, in the combination of results, it is verified that there was no statistical significance because the GA goes through zero and the p-value was 0.09.Besides, a great variability was observed among studies, with values of Q=22.60 and I²= 82.30.
A great heterogeneity among studies was observed, both for the analysis of the transition time to full oral feeding and in relation to the time of hospital permanence.This high variation was due to methodological differences, limiting generalization of the estimates.The studies included used different types of stimuli, with suction and swallowing exercises, motor and oral support, and peri-and intraoral muscle stimulation, with frequency and intensity varying among groups, making the comparison among them impossible (21,22,24,25,34) .However, despite the impossibility of generalization of the results due to the heterogeneity of the studies, the importance of strategies of stimulation of the motor and oral sensorial system was observed to reduce the transition period to full oral feeding.

CONCLUSION
From this systematic review, it was possible to understand the complexity of the process of transition from enteral to oral feeding.From the analysis of the variables described in the studies, the variability of indicators related to the process of transition to oral feeding and the importance of using standardized evaluations for the aid of health care teams in determining the appropriate and safe moment for the NB to take up oral feeding, still underused, were verified.
Besides, despite the impossibility of generalization of results due to the heterogeneity of the studies, in the meta-analysis, we observed the importance of strategies for the stimulation of the motor and oral sensorial system for the anticipation of the transition period to full oral feeding, with consequent decrease of the hospital time of permanence and costs.It is suggested that future studies comparing the effects of similar interventions in relation to the kind of stimuli and stimulation time length are carried out.

Figure 2 .Figure 3 .
Figure 2. Forest-plot of transition time to full oral feeding

Table 1 .
Stages in the selection of the articles Strategies of search in databases, performed in the months of April and May 2012 or da:2006$ or da:2007$ or da:2008$ or da:2009$ or da:2010$ or da:2011$ or da:2012$) AND LA:(ES OR PT OR EN) SciELO ("Comportamento de Sucção" OR "Comportamento Alimentar" OR "Deglu$" OR "Alimentação Artificial" OR "Aleitamento Materno"