The concept of adjustable fortification (AF) was first proposed by Moro et al. (31) and developed by Arslanoglu et al. (32) in a standardized form. The approach is based on the metabolic response of preterm infants, when renal function and access are normal and protein intake is closely related to blood urea nitrogen (BUN) levels (33). BUN was measured twice weekly following the commencement of standard fortification to assess protein metabolism. SF was continued if the BUN level was between 10 and 16 mg/dl, with an additional 0.4 grams of protein fortifier based on SF if BUN was <10 mg/dl, and 0.8 grams of protein if BUN remained <10 mg/dl, up to a maximum of 1.2 g. Conversely, if BUN > 16 mg/dl, the fortification was gradually reduced to a minimum of 1/4 standard fortification (32). The AF method does not necessitate an analysis of human milk composition and is easy to apply clinically. It increases protein intake without increasing total energy or fluid volume, is well tolerated, and improves physical growth and the long-term prognosis of preterm infants during hospitalization (34, 35). Furthermore, it also increases the rate of physical growth after discharge and may also improve long-term neurodevelopmental outcomes (36). Ergenekon et al. demonstrated that infants receiving AF exhibited significantly higher indices of psychosocial and psychomotor development (BSID-III scores) than those in the SF group at postmenstrual age 18 months (37), as well as significantly higher hearing and speech scores (38).

The use of BUN as an indicator of protein metabolic response is a feature of adjustable fortification. However, BUN does not always fully reflect protein intake, particularly in very preterm infants with a high catabolic rate in the first few weeks of life. In such cases, urea levels may be elevated regardless of protein intake or renal function (39). Furthermore, the limited capacity for urea synthesis and/or renal excretion in preterm infants, as well as increased catabolism in some disease states, do not accurately reflect protein metabolism (40). The clinical study has demonstrated that greater protein intake does not necessarily result in an elevated BUN level. Consequently, BUN is not considered a reliable indicator of adequate protein intake (26). Furthermore, BUN measurement necessitates repeated invasive procedures, which result in increased patient discomfort, and the frequent blood collection and analysis also impose a greater burden on healthcare workers.

To address these shortcomings, a non-invasive AF has been proposed. Mathes et al. (41) demonstrated a high positive correlation between plasma BUN concentration and urine urea-creatinine ratio as well as actual protein intake in preterm infants. It is therefore proposed that the use of the urine urea-creatinine ratio as a valid indicator of protein metabolism could help to estimate actual protein consumption in preterm infants.

The protein content of human milk typically declines with the progression of lactation in a mother and may also exhibit significant inter-individual variability (13–15). Therefore, target fortification (TF) is achieved by analyzing human milk composition at regular intervals (daily or twice weekly) to obtain measured macronutrient component data, and targeting guideline-recommended intakes by adding additional proteins, fats, or carbohydrates to standard fortified human milk to meet the nutritional needs of preterm infants (42, 43). The analysis time is brief, and the results are highly accurate, rendering it suitable for use in NICUs.

TF addresses the variability of macronutrients in breast milk through real-time measurement and supplementation, thereby optimizing the P/E ratio and preventing protein deficiency in infants, while ensuring better growth quality. TF has been demonstrated to improve nutrient intake and the quality of growth, including length, head circumference, fat, and fat-free mass, compared with SF (42–46). Furthermore, the osmolality of human milk following TF was within safe limits, and no signs of gastrointestinal or metabolic intolerance were observed (44).

The use of human milk analyzers and ultrasonic homogenizers is required for TF, which are relatively complex pieces of equipment and require frequent sampling and analysis as well as frequent calibration of the analyzers. This increases the workload in the NICU by 10–15 min per patient per day, with high instrumentation and labor costs that are not readily available for widespread use in NICUs (11, 42). Labor costs have been reduced by reducing the frequency of testing, for example, by switching to weekly analysis, with macronutrient intake remaining within ±5% of target levels (47, 48). A recent survey revealed that only 10% of NICUs were able to perform human milk analysis for clinical purpose, while adjustable fortification was practiced in 41.3% (49). It is crucial to acknowledge that TF does not account for individual metabolic and absorptive differences in preterm infants, potentially leading to under- or over-intake of nutrients. Furthermore, the accuracy of human milk analysis methods and the inaccessibility of various single component fortifiers in resource-limited countries represent additional challenges (50–52).

Studies have demonstrated that either adjustable or targeted fortification is more effective than SF in promoting early physical growth in preterm infants (10). However, there is a paucity of literature comparing the two approaches, and the results are inconsistent. Kadıoğlu et al. (53) found that the rate of growth in length and head circumference in the TF group was lower than that in the AF group. Conversely, the study by Bulut et al. found that the rate of growth in weight and head circumference in the TF group was significantly better than that in the AF group (54). Only one study compared the incidence of bronchopulmonary dysplasia and metabolic bone disease in preterm infants receiving targeted vs. adjustable fortification. Neither difference was statistically significant (53). A meta-analysis of these two studies was not quantitatively combined due to the limited number of studies and cases within each, and the observed effect remains to be further confirmed (55).

Non-invasive targeted fortification is more aligned with the optimal needs of very preterm infants and represents a promising approach. However, it necessitates the use of specialized equipment and is time-consuming, rendering it impractical in most NICUs (49). Consequently, a practical fortification strategy to optimize protein intake without the necessity for human milk analysis has been proposed: adapted fortification.

Minarski et al. (19) developed an equation for calculating human milk protein content: protein [g/100 ml] = 6.755/postnatal days + 0.852. This was achieved by analyzing the protein composition of human milk from 457 samples taken from 41 mothers during different stages of lactation and performing regression analyses and mathematical modeling of the trend towards a decrease in protein content with increasing lactation days. A validation cohort of 10 mothers with 141 human milk analyses was used to validate the equation, which was followed by target-volume protein supplementation, and the results demonstrated that the target protein supply was achieved in more than 95% of human milk samples. This straightforward, non-invasive fortification is regarded as a modified form of TF. The method compensates for changes in human milk protein content with minimal additional effort and establishes a viable fortification strategy for daily practice in NICUs.

Stepwise corrected fortification (SCF) is also a practical method based on the same fact that the protein in human milk decreases progressively at different stages of lactation. Started with standard fortification, and then progressively increased adding the protein and calories in human milk at different weeks after the birth of preterm infants in order to reach the recommended intake (56).

Pillai et al. (56) observed that the protein content of human milk declines in early lactation and is lower than expected from week 3 onwards. Consequently, standard fortification was introduced for the first 2 weeks of life, followed by an additional 1.27 g of protein fortifier for weeks 3–4 of life, and then 1.57 g of protein for weeks 5–6 of life, and 1.81 g of protein at week 7 of life and beyond. Since DHM was lower in protein and calories, 2.05 g of protein were added. Following the aforementioned fortification procedure, 68% of human milk samples fell within the recommended range for protein and calories, in comparison to only 5% in all human milk samples fortified with SF and 0% in samples after five weeks. It was observed that SCF had a greater impact on protein content compared with calories. Thus, SCF markedly augmented the final protein and calorie content of human milk, thereby offering a promising avenue for enhancing nutritional intake in preterm infants.

In the course of the study of SCF, Pillai et al. observed that the degree of yellow coloration of human milk may be correlated with the measured calories. Consequently, they introduced the bedside color tool, the Human Milk Calorie Guide (HMCG) (57). The objective of this tool is to predict the calories in human milk in resource-limited settings where human milk analyzers are not available.

The color tool comprises nine colors, divided into three rows. The lighter “watery” shade is found in row A, the “normal white” shade in row B, and the “creamy yellow” shade in row C. Each row contains three colors, with the lighter shades located on the left side and the darker shades on the right side. It was assumed that the calorie values of human milk would increase in line with the “yellowness” of the samples, from those in row A to those in row C. The MOM and DHM samples were then color-matched and analyzed for macronutrients. Among DHM samples, the color tool demonstrated the greatest accuracy in predicting low calorie values (<55 kcal/dl, AUC of 0.87 for row A), exhibited average accuracy in predicting high calorie values above 70 kcal/dl (AUC of 0.77 for row C), and exhibited poor prediction performance for MOM (56). The authors concluded that this color tool is a reliable predictor of the lower caloric range of donor milk and has the potential to improve donor milk fortification practices.

Arıkan et al. (58) analyzed the protein content of weekly human milk from 108 mothers of preterm infants in four groups of extremely preterm, very preterm, early preterm, and mid-late preterm during the first five weeks postnatal. The P10, P25, P50, P75, and P90 percentile curves of changes in human milk protein content of preterm infants of different gestational ages and postnatal weeks of age were plotted to create a new practical and individualized fortification guideline, replacing the laborious targeted or adjustable fortification currently in use.

The objective of these novel fortification above mentioned is to address the limitations of standard and adjustable fortification or targeted fortification in providing adequate nutrition for preterm infants, based on diverse clinical settings and different resource accessibility. However, the efficacy and safety of these methods remain unproven, as there is a paucity of clinical applications or comparative studies. It is also worth considering whether these equations or models derived from human milk composition data from a particular population are suitable for use in other populations from different countries and regions. The principle, advantages and disadvantages, and current state of application of the most of fortification methods presented are summarized in Table 1.