A recommendation development committee was created including neonatologists, paediatricians and nutrition experts. Parent representatives were also surveyed at multiple points during the process.

The target population was determined to be preterm neonates (GA < 37 weeks) and committee members were assigned topics based on expertise.

For each topic, screening was performed according to Preferred Reporting Items for Systematic Review and Meta-Analysis (PRISMA) guidelines [20]. The following keywords and Mesh terms were employed: complementary food; complementary feeding; weaning; introduction; timing; preterm newborn; premature; preterm infants; health outcome; development; adiposity rebound; paediatric obesity; body mass index; nutrition; post-discharge formula; macronutrients; oral dysfunction; allergy; “Weaning”[Mesh]; “Infant, newborn”[Mesh]; “Diet, vegetarian”[Mesh]; “Diet, vegan” [Mesh]. Proper Boolean operators “AND” and “OR” were also included to be as comprehensive as possible. Search limits were set for studies published up to 31st August 2021 in English language. Eligible studies were retrieved using the PubMed, Embase, Cochrane Library and Web of Science databases. Additional studies were identified from conference proceedings, trial registries and the reference lists of the selected papers. As a result, 62 manuscripts were selected for this position paper, including 8 systematic reviews [3, 18, 21,22,23,24,25,26], 8 narrative reviews [27,28,29,30,31,32,33,34], 27 observational studies [19, 35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,51,52,53,54,55,56,57,58,59,60], 4 controlled trials [61,62,63,64], 1 case report [65], 3 commentaries [66,67,68], 1 operational protocol [69], 3 reports [70,71,72], 1 consensus [73], 2 recommendations [74, 75], 2 guidelines [76, 77] and 3 nutritional reference values [78,81,80]. One or more recommendations/statements were drafted for each topic. Grading of Recommendations, Assessment, Development and Evaluation (GRADE) approach [81] was used to assess the quality of evidence (i.e., high, moderate, low or very low) and to define the strength of the recommendations (i.e., weak or strong) according to potential desirable and undesirable consequences of the recommendation. Final recommendations and statements were reviewed by experts and future guideline users to ensure feasibility. Based on available data, recommendations and statements were proposed, discussed and rephrased until a consensus of 90% or more was reached.

When should complementary feeding be started?

The timing for introduction of solid foods in preterm infants is still a matter of debate. Different timeframes were suggested in the past such as 3–6 months of postnatal age (PA) [70, 71], 5–8 months of PA [72, 76] or more recently from 3 months of corrected age (CA) [18].

The majority of data on CF in preterm infants were derived from observational studies, thus reducing the robustness of the recommendations. Only few randomized controlled trials have assessed differences between timings of CF introduction (Table 1). Marriott et al. divided 68 preterm infants in two groups: “preterm weaning strategy (PWS)” group or control group. The PWS group was weaned at 13 weeks of age and at least 3.5 kg body weight compared to 17 weeks of age and at least 5 kg. The PWS group also received advice regarding quality of foods, encouraging the consumption of high-energy and high-protein foods, and a mixture of dried cereals and home-prepared foods with preterm infant formula. Their results show that the PWS featured greater length at 12 months of age, with no differences in weight or head circumference, compared to the control group [61]. A prospective cohort study by Spiegler et al. [35] showed in a regression analysis that length and weight of VLBW infants at 24 months were positively influenced by early introduction of CF: VLBW infants at 24 months of age were on average ~ 0.4 cm taller and 100 g heavier for each month of earlier introduction of CF. Also Rodriguez et al. found a beneficial effect of weaning before 4 months of CA with higher weight gain at 18–24 months of CA in very preterm infants [46].

Table 1 Main features of RCTs and observational studies assessing timing for CF introduction in preterm infants

Conversely, an RCT conducted in India to compare two different timings of CF (4 vs. 6 months CA) in ex preterm neonates with GA < 34 weeks revealed that weight-for-age z score at 12 months CA did not differ between groups, but the 4-month CA group experienced a higher rate of hospital admission primarily due to infectious disease [62]. Hence authors recommend to delay CF until 6 months CA in this population, however generalisability of their findings is uncertain due to the important differences between low and high income countries, including higher mortality rate, environmental conditions, and predominantly vegetarian dietary regimens [68].

Similarly, a pooled analysis of prospective studies by Morgan et al. [21] showed no effects on height and weight at 24 months of age and health outcomes up to 18 months.

What is known is that preterm infants are usually weaned early (before 4 months of age) compared to their term counterpart [3, 54,55,56,57]. Moreover, the first solid food is often nutritionally inadequate, with a low energy and protein content [56], and wide variability in weaning practices and vitamin and iron supplementations [19]. The entity of prematurity influences greatly the timing of weaning: preterm infants born at 22–32 weeks GA show a 9.90 odds of receiving CF before 4 months of age, compared to term peers [58].

However, also late preterm infants are often weaned early, at a mean postnatal age of 5.7 months and a mean CA of 4.6 months [59].

The early introduction of solid foods in preterm infants has been linked to a higher risk of rapid weight gain [46], allergy and anaemia whilst a delayed weaning (after 7–10 months of PA) may increase the risk of avoidance feeding behaviour [18].

It is noticeable not only that preterm infants are introduced early to CF, but also that the attitude of primary care paediatricians is widely variable in terms of timing of introduction and type of suggested foods [19].

This is partly due to the lack of specific guidelines on CF introduction for preterm infants [22]. The COMA report in 1994 suggested weaning preterm infants with a body weight of at least 5 kg, provided they had acquired a few specific developmental milestones [72]. However, these suggestions may lead to a significant delay in some populations of preterm infants (e.g., ELGAN or ELBW) which would reach such criteria well beyond the timeframe (4–6 months of age) recommended by the ESPGHAN to start CF in term infants [14]. Preterm infants starting CF often show defensive behaviours at mealtime, such as refusal to open the mouth, food selectivity and feeding refusal [60]. More recently, it has been recommended that CF in preterm infants should be started between 5 and 8 months of chronological age [76] when neurodevelopmental skills (e.g. good control of the neck, disappearance of the protrusion reflex of the tongue, the reduction of reflexive suck in favour of lateral tongue movements, and the gradual appearance of lip seal) and readiness to explore new textures and flavours should have been reached by the vast majority of ex preemies. Since an adequate motor development is a pivotal requirement for starting CF, it has also been advised to consider CA in the assessment of the optimal timing for weaning preterm infants. In this respect the limit of 3 months CA has been set to ensure the acquisition of developmental skills which allow the consumption of solid foods. Importantly, CA would be a unifying criterion for the heterogeneous population of preterm infants, since it is applicable to babies of all gestational ages, from the lowest to the highest [3].

Although critical, neurodevelopmental readiness is not the only aspect to take into consideration. Difficult transition to complementary food may also be related to comorbidities, or even behavioural issues, which should be carefully assessed with the aid of a multidisciplinary team. Indeed, the multiple and unpleasant procedures undergone during hospital admission (e.g., orogastric/nasogastric tube feeding, suctioning, intubation) may lead to a negative attitude towards CF. Furthermore, parental emotional factors should not be underestimated, particularly in growth-restricted infants, whose growth rate is often concerning for parents [36, 60].

Currently, there is insufficient evidence to draw final conclusions regarding a specific timing for starting CF in preterm infants, due to their extreme variability in achieving neurodevelopmental and oral skills. Hence, we suggest an individualized approach based on the accurate evaluation of the infant development and attitude towards semi-solid foods, employing corrected or postnatal age as an indicative reference rather than a mandatory schedule.

Recommendations/Statements

CF in preterm infants should be started between 5 and 8 months of chronological age.

Consider also the limit of 3 months CA to ensure the acquisition of developmental skills which allow the consumption of solid foods.

Certainty of evidence: Moderate.

Grade of recommendation: Strong.

Are there specific recommendations for preterm infants with oral dysfunction or comorbidities?

Oral dysfunction is not uncommon among infants born preterm, due to the higher occurrence of comorbidities (e.g., bronchopulmonary dysplasia) or neurodevelopmental impairment [27, 37]. Reportedly, over 15% of preterm infants require enteral tube feeding upon discharge [38]. Lower gestational ages at birth (below 30 weeks) and neonatal surgery have been described as risk factors for oro-motor feeding problems at 12 months’ CA [39]. This sub-group of infants often features greater defensive behaviours at mealtime when starting CF, e.g. refusal to open the mouth, feeding refusal and food selectivity [60].

However, guidelines regarding CF for preterm infants with oral dysfunction or major comorbidities are still lacking, hence the nutritional strategy for these infants should be tailored and revised regularly (Table 2). Seemingly, a greater amount of food is consumed by preterm infants using a spoon-assisted mode of feeding [63], probably due to the decreased gag reflex elicited by the introduction of food with higher texture [28]. CF may be started at 3–4 months of corrected GA, encouraging the consumption of thicker foods which may be swallowed more easily.

Table 2 Main features of trials and observational studies assessing preterm infants with oral dysfunction or comorbidities

Importantly, preterm infants with oral dysfunctions or comorbidities require a multidisciplinary follow up encompassing nutrition experts, speech therapists, and a behavioural psychologist [28, 29, 69]: oro-motor stimulation should be started early for infants on prolonged tube feeds. Infants with gastrointestinal issues should also be followed up by a paediatric gastroenterologist. Ex preemies with bronchopulmonary dysplasia should be weaned with low salt, limited volume, and high energy diets; these infants usually better tolerate foods given by spoon since they may suffer from mild hypoxic spells when suckling liquids.

Complete foods based on amino-acid mixtures concentrate in small volumes a high macronutrients content: they may be an option to meet the high nutritional requirements of infants with comorbidities or of those infants unable to ingest large quantities of food [3].

Recommendations/Statements

Certainty of evidence: Low.

Grade of recommendation: Weak.

Which type of food should be recommended?

When it comes to solid foods for preterm infants, two critical aspects should be taken in consideration. Firstly, if the acceptance and consumption of semi-solid food is still inadequate, attention should be paid to the intake of micronutrients. In this respect, supplementation with iron and multivitamin products are helpful to ensure the correct supply of micronutrients. Secondly, if catch-up growth has not been reached by the time of weaning, a high protein and energy intake should be promoted by means of the correct formula or specific foods to propose. The choice of the right formula milk (i.e., post-discharge or standard formula) is also dependent on the milk tolerance of the infant, since less mature preterm infants may have immature feeding skills but higher energy requirements [3, 30].

Importantly, several figures are involved in the process of weaning a preterm infant: families, primary care paediatricians and nutrition experts. Each figure plays an important role. The family is pivotal since it represents the main support for the babies and their parents. According to a recent systematic review, nutrition education for families may decrease the risk of undernutrition in term infants [23], hence we may speculate that the same could occur with ex-preemies. The primary care paediatrician should carefully evaluate growth patterns and ensure adequate adherence to prescriptions and nutritional advice. Lastly, the nutritional expert should guide all the weaning process by carefully evaluating the infant nutritional needs and neurodevelopmental and oral skills, in order to provide tailored recommendations.

More specific recommendations for preterm infants regarding type of foods to choose, sequence and speed of introduction are lacking, hence guidelines for term infants currently remain the gold standard [14]. Importantly, the beginning of CF is associated with significant changes in both macronutrients and micronutrients intake, with the risk of nutritional imbalances. The energy requirement differs according to the degree of prematurity. Embleton et al. [40] showed that preterm infants often fail to meet their dietary intake (energy 102 kcal/kg/day; protein 3.0 g/kg/day) since the first days of life and that such deficits are not recovered by the time of discharge.

Recently, Salvatori et al. [24] suggested intake of macronutrients for preterm infants taking into account recommendations conceived by The Italian Society of Human Nutrition with LARNs (Reference intake Levels of Nutrients and energy for the Italian population) of 2014 [78], the Dietary Reference Values for nutrients of European Food Safety Authority (EFSA) of 2017 [79] and the Nutrient Reference Values for Australia and New Zealand Including Recommended Dietary Intakes of 2017 [80] (Table 3).

Table 3 Macronutrients adequate intake for infants

As for micronutrients, iron supply is a matter of concern due to its essential role for brain development. Iron supplementation is recommended for preterm infants until at least 6–12 months of age [66]. However, from 6 months of age the supplementation alone would not be sufficient to provide the adequate amount of iron, hence the consumption of foods rich in iron (e.g., meat, iron-fortified cereals, fish) should be encouraged.

Recommendations/Statements

Recommendations for preterm infants regarding type of foods to choose, sequence and speed of introduction may be considered the same as for term infants currently.