Immunogenicity of Hydrolysate Formulas in Children (Part 1). Review of 202 Reactions

CM allergy (CMA) is a disease of infancy with onset in the first 3 months of life with a 58% incidence and lowers to 39% in the 4th-6th month [1] and to 24,9% (mean) in food-allergic chil­dren aged 1-18 [2]. Since the turn of the century, CM formulas have be­come pro­gressi­vely more com­mon as breast milk (BM) substitutes when BM is unavail­able, and CMA has thus grad­ually become a more com­mon disorder [3]. The evaluation of infants for possi­ble CMA is thus one of the more com­mon problems en­countered by pe­diatri­cians and CMA manage­ment in in­fants and children confronts pediatricians and allergists with one of the most deman­ding chal­lenges. Un­fortunately, both over- and under-dia­gnosis are freq­uent­ly seen.

The ideal CM substitute should be hypoallergenic, easily available, inexpensive, and palat­able in order to obtain a good compli­ance, and with an adequate nutritional value appropria­te to the infant's age. As yet, the CM available sub­sti­tut­es are SPF (soy-pro­tein formu­las), home-made, meat-based formulas (HMMBF) [2], HFs and elemental diets. HFs have been developed with the aim of de­creas­ing or eliminating the allergenic­ity of CM proteins, of re­ducing the antigenic load, and the risk of sensiti­zation. Thus, pro­tein HFs, classified as HA, have been used for feeding ba­bies with CMA [4,5] and/or for the prev­ention of atopic dis­ease in babies of atopic parents [6,7]. In this paper we rep­ort and discuss 202 re­ac­tions to dif­ferent HFs.

HFs available for infant nutrition

Accor­ding to the source of proteins there are four types of HFs, reduced to two types based on the degree of hydrolysis: highly and partially hy­drolyzed (Table 1) [8].

Briefly, these formulas are processed using heat denatu­ration and en­zymatic hydrol­ys­is to reduce the MW of the pep­tides. The reduc­tion of antigenicity is associ­ated with a red­uction of the pala­ta­bility. The composition of HFs is de­pendent on several factors in­cl­ud­ing the degree of digestion, post-hydro­lysis process­ing, elimi­nation of the enzy­m­es used for the hy­drolysis and protein source. Extensively HFs are considered the more HA among the HFs, whereas partial­ly HFs are con­sidered less HA and even dan­gerous in children with CMA [9,10]. However all HFs contain vari­able profile peptides with very high MW, even greater than 6 kD (kilodaltons) [8], which are an in­dex of the ex­tent of their immunogenicity. These formulas are integrated with vegetable lipids, and Alfa-Rè, Ali­mentum and Pregestimil in addition contain medium chain triglycerides (Table 2). All HFs, ex­cepted HA, are lactose free, and contain a small amount of carnitine. They are rather unpalat­able (except HA) and compli­ance is there­fore poor.

Studies in healthy children

Care should be taken when HFs are given for prolonged periods since no data is avai­l­able on the nutritional assessment of infants fed exclu­sively HFs for several months. Rigo et al. [11] showed that a partly whey HF (PWHF) in­duced in full-term newborns fed this product for 6 days a significant increase in plasma con­cen­tration of sev­eral essen­tial amino acids (AA), especially threo­nine and bran­ched-chain amino acids. The total es­sen­tial AA concentration and the ratio of essential to total AA con­centration were higher in the babies fed the PWHF than in the babies fed BM or a whey hydrolysed for­mula (WHF). In a sub­se­quent paper [12] the authors noted that at age 33 days the plasma threo­nine concentration re­mained twice as high and the pla­sma tyrosine, phenylal­an­ine and proline levels were significantly lower in the PWHF group than in the BM-fed infants. Growth and most of the biological indices of protein metabolism were alike in the two groups. Finally the authors [13], while con­firming the AA level alterations, ob­s­erved a drastic reduction in fat Ca and P absorption with the use of a whey-ca­sein HF. In preterm in­fants, compar­ed with the standard preterm formu­las, HFs led to a signifi­cant increase in plasma threonine, a decr­ease in tyrosine and pheny­la­lanine concen­tra­tions, and a reduction in pla­sma histidine, valine, leucine, cystine, methi­onine and/or tryptophan [13].

Vandenplas et al. [14] studied from birth to 3 months of age 45 healthy infants, 20 re­ce­­iv­ing a WHF and 25 a PWHF. They found that except for the iron-binding capacity (IBC), zinc, and urea, which were higher in the PWHF-fed babies, at the end of the study the nutritional status was adequate in both groups. Apparently there was a 16% dropout rate only in the PWHF-fed infants. The in­creased IBC in the PWHF children (351 ± 58.4 mg/dl versus 301 ± 47.4, p = 0.006), was accompanied by similar Hb, Hct, and iron levels val­ues which suggests an impaired iron absorption. Zinc levels were also signifi­cantly differ­ent in the two groups (79 ± 12.9 mg/dl in the PWHF babies versus 67.6 ± 8.9, p = 0.002 in the controls). The most intriguing issue is the statistically signif­icant dif­fer­ence in plasma urea levels (20.5 ± 6.53 mg/dl versus 15.9 ± 3.45, p = 0.009) and its in­creased uri­nary output with an even higher statistical difference (p = 0.001). The au­thors discuss the data of an earlier study [15] in four adults with malab­sorption who ex­peri­enced a high retention of ab­sorbed N when fed an ele­mental diet. However these patients un­derwent N balance studies in a crossover fashion, and were fed in addition solid food, and a diet partly containing a casein hydrolysate formula (CHF): following such diets plasma and urinary urea N levels declined. There­fore this study fails to sup­port the thesis set forth by Vandenplas et al. In addition, the mean daily volume intake was smaller with the PWHF compared with the WHF (590 ver­sus 680 ml/day, p = 0.002), while the protein equiv­alents/dl were almost similar [14]. These findings clear­ly show that the net re­ten­tion of N appears to be lower in the PWHF-fed babies.

In 205 healthy term infants enrolled shortly after birth, and followed-up for 8 months, 72 of whom receiving a PWHF, 68 BM, and 65 a WHF [16], the growth rates were measur­ed at one-month interval until the 4th month. The laboratory studies included the mea­s­urement of IgG and IgE anti-CM an­tibodies in 30.7% in­fants, and several common gas­tro­in­testinal (GI) symptoms were evaluated. The dropout rates for reasons related to the feeding regimen were 24% in the PWHF, and 21% in the WHF groups, higher than in the study by Van­denplas et al. [14]. To demonstrate that feeding an HF in the very first days of life does not influence the development of atopy, 128 infants were fed in a dou­ble-blind way a CM formula or a PWHF as a supplement to BM during the first 5 days. When the amount of early post natal CM supplementation was correlated with subse­quent IgE levels, it was found that the total vol­ume of sup­plements and the frequency with which they were re­ce­iv­ed by the neonates before the start of breast feeding re­sul­ted in a statistically significant increase in the total IgE levels (p = 0.02). The dropout rates were 4% at day 150, and 26% at day 365 [17]. However, feeding a CHF in the neona­tal pe­riod seems to influence the absorption of macro­molecules. In 130 healthy term neonates randomly assigned at birth one of three feed­ing regim­ens for the first 3 days of life, and then exclu­sively breast­fed, only those CHF-fed had at 2 months a signifi­c­ant­ly high­er a-lactal­bumin serum content in compari­son to con­trols, in addi­tion two ba­bies of the CHF-fed group had CMA symptoms at 7-8 weeks of life [18].

So the results at our disposal do not allow a conclusion: on the one hand the clinical im­pact of variations of AA serum content in healthy infants is presently largely un­known, on the other in the chil­dren so far evaluated there are no visible differences in both growth and symptoms.

Clinical properties of HFs: Studies in atopic children

Use of HFs in treating CMA in HR children: Considering the pertinent studies, it is evident that less or more severe reac­tions were observed employing formulas with either whey or casein pro­teins more or less exten­sively hydrolyzed. We discuss first the “programmed” studies which demonstrated the al­lergenicity of HFs in vitro and in vivo in children with IgE-mediated CMA, using skin prick tests (SPT), RAST, and/or challenge studies:

I total, we report 202 reactions to HFs [4,5,24-54]. Reactions to extensively CHFs are shown in (Table 3) [4,5,24-44,50-54], and to par­tially and ex­tensively WHFs in (Table 4) [23,24,26,27,29,39,40,45-50,52]. In particular, we first reported [50] 5 exclusively BM fed infants aged 3-8 months (median 5 months) with IgE-mediated CMA, who expe­ri­enced ana­phylactic reactions when first fed a small amount of a CHF (Table 3). All in­fants had AD (atopic dermatitis) during BM feed­ing, positive SPTs and RAST to CM pro­teins and to the HF; total IgE levels ranged from 45 to 2,990 U/ml. Subsequently the in­fants were suc­ces­sfully fed a SPF.

A new ultra-filtrated WHF was investigated in 66 children with CMA (mean age 1.9 years), who tol­erated it on open food chal­lenge (OFC), except four, only one of whom was positive at the first rechallenge and negative at the final one. How­ever of the 35 sub­jects with IgE-mediated CMA 11% (3/28) had po­sitive SPTs and 6% (2/35) IgEs to the WHF. The authors conclude that this for­mula is safe for chil­dren with CMA, and suggest to per­form a rigidly cont­rolled OFC in chil­dren with im­mediate reactions to CM before starting whatever HF [55].

HFs can induce, in ad­dition to im­mediate re­actions, also intestinal lesions which both SPT and DBPCFC fail to detect, while they can easily docum­ented with light micro­s­co­py [39]. Two babies aged 15 days-10 weeks were fed an EWHF being af­fected with diar­rhoea Æ remission for 10 Æ days recurrence Æ reso­lu­tion when put on a BM diet. The intesti­nal morphol­ogy im­proved, however follow­ing an OFC with the EWHF showed con­ge­s­tion and infl­ammatory in­fil­tration of the corion of a lym­phoplasmo­cytic type, and in­cr­ea­sed num­bers of intraepithe­lial lym­phocytes. Such le­sions are some­times dis­cre­te, and use­ful for di­ag­nostic purposes only when they are compa­red with a previ­ous­ly regular biop­sy (which is ra­rely per­formed in normal condi­tions). Therefore we are confront­ed with elusive forms, be­cause their diagnosis is dif­ficult with the presently available tests, in addition to making necessary the in­testinal biopsy, and the dif­fer­ential dia­g­no­sis with all possible causes of diar­rhoea in little infants.

About the case of urticaria to Alimentum reported by Oldæus et al. [25], the HF was ad­ministered in a SPF, however the patient had a positive RAST only to the HF, thus ex­clud­ing the potential responsibility of the SPF.

Thus, considering the cases referred to in the literature, the use of HFs has provoked 202 reactions, many of which IgE-mediated, in 132 chil­dren, aged 20 days-15 years) to CHFs (1 case of shock, 5 of ana­phy­l­axis, 7 of generalized urticaria, 1 apparent life-threatening event) (+ 2 local­ized), and in 70 children aged one month-15 years to WHF (either ex­tensi­vely or partially) (1 case of shock, 10 of ana­phylaxis, 13 systemic reac­tions, 2 apparent life-threatening events), at varian­ce with other stu­dies [56,57]. In addi­tion there is the unspeci­fied number of significant allergic reac­tions to GS re­sult­ing in the removal of the HA des­ig­nation from the label [46], and those ascribed to a CHF [5]. On the other side Buts et al. [58] af­firm that a PWHF tolerated by >50% of babies who are highly allergic to CM can be called HA even if it can po­ten­tially trigger severe ana­phy­lactic re­actions in <50% of them.

Strategies in preventing atopy in HR babies with the use of HFs

Several studies in babies with a high hereditary risk, that is a severe single or dual parental heredity, tried to prevent atopy with the use of HFs. In several studies a highly CHF has been employed, in others a PWHF (Table 5) [6,7,59-73].

In a prospective, randomized study the avoidance of CM, egg, fish and pe­anuts, during the first 6 months of lactation with the supplementation of a CHF, signific­antly re­duced the prevalence of AD and food allergy at the age of 1 and 2 years, how­ever there was no dif­ference at age 4 and 7 years [59-61]. The compli­ance to the avoidance regimen was poor, there be­ing evi­dence of breaks in the diet, in addition the nursing mothers who avoided CM during the period of lacta­tion were allowed to drink an ex­ten­sively CHF [59].

Chandra et al. [62] showed that in babies drinking a PWHF the prevalence of atopic dis­ease significantly decreased com­pared to that found in SPF- or CM-fed in­fants. How­ever atopic disease was not wholly prevent­ed, since the pre­va­lence of allergic disor­ders in the study groups was 18% at 12 and 26% at 18 months. In addi­tion, 4/5 in­fants who deve­loped atopic symp­toms while on PWHF had a posi­tive SPT to CM proteins, against 2/25 SPF-fed (80% versus 8%) [62] (Fisher 0.0026), thus suggest­ing that sensiti­za­tion to CM proteins in in­fants re­ceiving this HF is by 1000% more frequent than in the SPF-fed babies.

In the studies by Vandenplas [6,64,65] per­formed in HR ba­bies fed eith­er a PWHF or a CM for­mula, the preva­lence of CMA in the HF-fed group at 12 and 36 months of age varied from 6 to 25%.

In an Italian study [71] 279 babies received a dietary and environmental prevention programme, and 80 formed the non-intervention group. However, the babies were neither randomized nor obser­ved blinded concerning eval­uation, the SPTs were not al­ways done, or specific IgE measured, neith­er details of the clinical methods em­ploy­ed, nor of the severity score used to dis­tin­guish different forms of AD were given. In addi­tion we note the high prevalence of allergic disor­ders in the controls (> 42%).

Halken et al. [7] studied prospectively from birth to 18 months of age 105 HR in­fants: the study group was recommended BM and/or a HF (Nutramigen or Profy­lac), while the control group consisted of 54 HR ba­bies born in 1985. The au­thors suggested environm­ental controls and diet­a­ry manipula­tions restricted to solid foods avoidance during the first 6 months of life, which were strictly followed by 85% of cases, there­fore the high pre­­valence of atopic diseases in the two groups (32 and 74%, re­spectively) is unex­pect­­ed. In a subse­quent pros­pec­tive and randomized stu­dy of the same group [55] on 141 HR infants fed the same HFs to evaluate their protective ef­fect, the CMA inci­dence was 3.6%, while that of CM-related symptoms was 26-33% (15% in the 20 BM-fed babies); the figur­es of controls are not specified. None of the mothers had dietary restrictions, thus the compliance with the funda­men­tal advice of strict and exclu­sive breast feeding until the 6th month was scarce: 46% for 1-2, and only 14% for 6 months, whereas the moth­ers in our preventive intervention programme were 100% [3].

Mallet and Henocq [68] have studied the effects of a high-degree CHF in 39 in­fants, in addition 53 were BM- (> 2 months) and HF- fed, and 65 infants fed an adapted CM for­mu­la, 33 of whom receiv­ed also BM as above were the con­trols. Only 31% infants of the HF group (12/39) vs. 45% of BM-HF-fed (24/53) had a surely positi­ve family history. Among the HF-fed, at two years 10/78 (13%) had asthma, and 9/78 AD (11.5%) vs. 12/61 (19,6%) and 26/61 (42,6%), at 4 years there were 8/70 (11.4%) cases of asthma and 5/70 (7%) of AD, vs 6/54 (11,1%) and 14/64 (25,9%), respectively. Therefore the results are dif­fi­cult to int­erpret, apart the high prevalence of atopy in both the HF-fed and the control groups.

In conclusion, only two studies [23,55] show that Alimentum and Profy­lac fulfil the Ameri­can Academy of Pedi­atrics (AAP) recommenda­tions [74] for designing a formula as HA. This is stressed by the inclusion of chil­dren with different ages, clinical di­seases, and a different diagn­ostic work-up, in the absence of well-defined and -validat­ed meth­ods [6,14,63,66], as in­stead suggested by ESPACI and AAP [57,74]. Also the ab­sence of specialists may limit standardization of diagnosis [75]. The high preva­lence of aller­gic disorders especially in the control children seems to be a com­mon char­acter­istic [6,7,62-65], and the 36-74% preva­lence re­port­ed in several studies [6,7,62,63,65,66,68] are unex­pected, as well as the CMA preva­lence found in the PWHF-fed babies (21-35%) [7,60,62,65,66,68,73] (Table 5). However most of these studies inc­luded only infants with a very HR of allergic disease, therefore such high prevalences may be just­ifi­able.

In addition, we have investigated the immunogenicity in the IgE system of a PWHF, 400 ml daily of which were given to 39 mothers of HR babies during the lactation peri­od, while 39 control mothers of HR babies consum­ed 400 ml daily of CM. Although there was no significant differen­ce in both the inci­dence and prevalence of CM-induc­ed AD and of CMA in the babies at 0.5 and 1 year of age, ac­cor­ding to the mother's diet, the number of babies with IgE an­tibodies to CM and with total IgE levels more than 2SD for normal values for age were significantly higher in the group of babies whose moth­ers recei­v­ed the HA for­mula (p = 0.02). We may speculate that when a mother drinks this product, a large amount of im­mun­ogenic peptides are easily absorbed through the in­testinal mu­cosa, thus rapidly reaching the breast and then presented to the T and B cells of her baby. This data suggests that such PWHF seems to be more im­muno­genic in the IgE sy­s­tem than CM [76].

Nine babies breast-fed by mothers who strictly limited the assumption of CM experi­enced anaphylaxis when fed a PWHF [77]. The sensitization seems to have occur­red in the very first days of life as a conse­quence of some feeds in the Maternity Hospi­tal. [78] with the PWHF, which was giv­en again at the 6th month for CMA prophylax­is. Although we cannot exclude that the infants were the victims of pirate bottles, nor we analyzed BM samples for the presence of CM proteins, the following points are in favour of this as­sump­tion: 1) The mothers totally avoided CM and dairy products during lactation, there­fore a sen­sitization through BM can be ruled out; 2) The babies were healthy during breast fee­ding and did not show any symptom or sign sug­gestive of CMA; 3) high levels of IgE antibodies and strongly positive SPTs to the HF were present in the babies [77].

Chemical and immunological properties of HFs: a critical evaluation

The more or less severe reactions as yet reported were unexpected by the in­dus­try, that on the contrary tried to prepare HFs in the hope of reducing significantly the preva­lence of CMA, by maintaining an ade­quate nutri­tional value [79]. It is therefore neces­sary to try to explain the al­lergenic ac­tivity of such HFs:

As regards the main technologies employed in processing, commercially available HFs devised for CMA or CM-intolerant infants are pro­cessed us­ing two main technologies: heat treatment or enzymatic cleavage or both to reduce the MW of peptides to obtain alt­ernatives of minor allergenic potential. Several dif­fer­ent techniques are employed, to­gether with enzymes, such as trypsin, chymotrypsin, pep­sin, carboxyp­epti­dase, etc [22]. Briefly, during infant formu­la pro­cessing, heat-in­duc­ed de­naturation, mainly af­fecting whey proteins, changes the protein structure of the al­lergenic molecule, and most con­formational B epitopes are eliminated, thus fa­cili­tat­ing the hydrolysis, but not of sequ­ential T epitopes [9]. An enzy­matic hy­drolysis is ne­c­es­sary to degrade ca­sein proteins, which reduces the anti­ge­nicity and allergenicity, mostly elimi­nating the se­q­uential epitopes, then the ultrafil­tration is necessary to rem­ove peptides of high MW [22]. These different technical pro­cedures are essential for obtaining an accept­able palatability, and a com­bination of these methods is also in use [80]. However the results cannot be considered as definiti­ve: the degree of hydrolysis can be of 26% in WHFs, of 15% in whey HA formulas (range 1,3-32,5), and of 52% in the only CHF tested [81].

Another intriguing issue is the evaluation of higher or lesser MW of HFs. Since the HFs contain peptides of lower MW than the native protein, it is sug­gested that a low MW most likely decreases also their inhe­rent sensitizing cap­acity, which is why the AAP Com­mit­tee of Nu­trition rec­om­mended HFs with a MW <1200 D supposedly not aller­genic [82]. How­ever in peptides as small as 1500 to 2000 D in size, or smaller, even when their abili­ty to act as an allergen is elim­inated, the im­munogenicity of the absorbed frag­ments may theore­tically be as strong as or even stronger than the native molecule [80]. Thus, the MW alone can­not suffice to guarantee antigenicity for any given HF, and especially for commercially avail­able products [83], because the MW distribution de­cla­red by a manu­facturer is ap­proxi­mate, and cannot assess a 100% non allergenici­ty of a given HF [84]. Hence the indiscrimi­nate HA labelling, which lit­erally means “less al­lergenic” than normal CM formula, and not “non- allerg­enic”, is un­con­vincing be­cause is not quantifi­able, hence open to each evaluation [57]. Small but detect­able amounts of residual pepti­des, even with MW >6.000 D can be demonstrated in HFs (Table 1), how­ever in CHFs there is 15-20% of pep­tides with MW >3,850 D and 35-42% with MW betwe­en 340 and 3,850 D [86]. After protein enri­chment by trichloroacetic acid (TCA) precipitation, the pre­s­ence of high-MW polypeptides was shown in HFs, such as pro­tein bands visible in SDS-PAGE with a cha­racteristic pattern [87]. Partial hy­drolyzed formu­las show the higher amount of poly­peptides with a diffused area ranging from 6.500 D and 71 kD, while ex­tensively hy­drolyzed products have a lower residue [87]. The study upon 11 WHF, 7 of which HA, one based on hydro­lyzed ca­sein, and one ul­tra­fil­trated [81] shows the pres­ence of a mean 55,5% content of pepti­des with a MW of 3 kD, 27% of 3-5 kD, 13,5% of 5-10 kD and 6,9% >10 kD, incl­ud­ing two HA HFs with a 26,3% and 40,6% con­tent, thus all test­ed products re­tain some residual antigenicity of one or more of the ind­i­vid­ual CM pro­teins [81]. Previously peptides with MW > 5.000 D were found in GS [88], how­ever by tricine-SDS-PAGE and subsequent silver stain­ing were identified residual pro­tein frac­tions be­tween <14 kD and 20 kD in Alimen­tum, Profy­lac and in the CHF Nu­tramigen and none in Alfa-Rè [56], Yet studies done initially in the animal mo­del (see below) suggest­ed that HFs even with peptides <3.400 D were not immunogenic and pepti­des with MW between 3.400 and 6.500 D would induce only weak reactions. The al­lergen­ic and im­mun­o­genic epitop­es which can be “seen” in CM pro­teins by the hu­man are not neces­sarily the same as those seen by a rabbit, or guinea pig, or lamb;

The AA sequence of an epitope can resi­st the physicochemical manipu­la­tions, in spite of extensive break­down [89], thus HFs may contain an amount of resid­ual intact pro­teins able to stimulate immune reactions in predispos­ed children [90]. New epitopes may arise due to structural changes in the ternary struc­ture, or may be created during heat treatment, or be un­masked during the ready-to-feed prepa­ration of CM for­mula, or fol­lowing di­gestion of food peptides during the intestinal passage. The im­mu­no­genicity of new epi­topes may be as high as that of native pro­tein, or even hi­gher [9]. Thereby the residual antigen­ici­ty and thus the potential allergenici­ty is de­pendent on the food-pro­cessing tech­nolo­gies appli­ed [89]. On clinical grounds, the cells se­c­ret­ing IgM against trypsin- and pepsin-digested ßLG (ß-lactoglobulin) sug­gest that some antigenic epitopes may resist such treat­ments [91], and it is known from the Küst­ner case that epitopes not present in the original food protein can be formed or made ac­ces­sible by digestion, unless the seg­ments of pep­tides that re­sist enzyme hydro­lysis are af­fected and in­activated by adeq­ua­te dena­tu­ra­tion, or some kind of filtration is used to re­move the residual large pepti­des [89]. Lipids and carbohy­drates not concerned by the hy­dro­l­ysis, can bind protein structu­res, build­ing up the substra­tum to make new epitopes. A number of small peptides may be­come antigen­ic and even al­ler­genic, and bind IgE an­tibod­ies, by aggr­egat­ing or cross-linking within each other or with other mole­cules, probably through coval­ent or very strong hydro­p­hobic binding, thus binding a cell mem­brane that can be pre­sented to T cells [83]. HFs con­tain protein frac­tions which re­sulted in a spe­cif­ic IgE bind­ing after in­cubation with serum samples from patients with CMA [10]. As previo­usly reported, HFs are not only al­ler­genic in al­ready sensitized ba­bies, but may be immuno­genic in the IgE system due to residual al­ler­genic epito­pes which bind IgE anti­bodies to CM [10]. Hap­tens can combine with al­bumin or other pro­tein car­riers [83]: so a pepti­de not recog­nized by the im­mune sys­tem will probably never be available [9].

Elimination of peptides in HFs may be arduous: antigenic structures, which are pro­cess­ed and cleaved into peptides of low MW, consist of 11 to 13 AAs. Schematically, when an antigen is captured by B cells via their surface immunoglobulin (Ig) molecules and int­ernalized, and enzy­matically de­graded inside the cell, the resulting peptide fragments with low MW are presented on the B cell sur­face in association with HLA class II molecules [92]. The HLA-pep­tide com­plex at the surface of the APC (Antigen presenting cells) is recognized by the T-cell receptor, an event which result in the act­ivation of both T and B cells [93]. Subsequently, specific pro­cesses lead to IgE synthe­sis [94]: the pro­ces­sa­tion with the un­fold­ing of the nati­ve polypeptide chain and dema­sk­ing epitop­es hidden by the three-di­mensio­nal folding of the molecule, trigger a speci­fi­c IgE reac­tion in indiv­iduals with atopic back­­grounds [95]. In normal subjects, the AA se­quences equipped with the func­tional properties necessary to be com­plexed with HLA molecu­les can bind to an HLA mole­cule only after the un­folding of proteins [96]. Thus it seems im­possible that the food-process­ing proce­dures in vitro abolish all the epitopes pre­sent in the native protein source [89]. So we are total­ly sceptic about the idea of de­stroy­ing or neutraliz­ing the epi­topes of so a pecu­liar struc­tu­re: a breakdown of 90-95% of the pepti­de reac­tivity leaves suf­ficient antigen to fully stimu­late an anti­gen-specific im­mune re­sponse, even if the al­lergen hydrol­ysis can re­duce the im­muno­genicity dra­mat­i­cal­ly. There does not exist a sophisticated te­chn­ique insuring the se­lec­tive dest­r­uc­tion of such spec­ific epi­topes. Such pro­cesses are op­era­tive alre­ady in the neo­nate/little infant [97]: in an infant ingesting CM the IgE re­action to the non-self-proteins always takes place, as shown by dendritic cells active in the ne­onate mice [98]. The ability of anti­gen frag­ments to interact with specif­ic B lym­phocyt­es is related to the retention of epitopes that can bind surface Ig recep­tors: the AA se­quence for a peptide from bovine serum albu­min bound to HLA class was de­scrib­ed [99]. It is easily un­der­stood that a “final prod­uct” con­sisting of two or three AA residues (and one epi­tope) would be more suffi­cient to trigger allergic reac­tions in HR children than the clear­ly recog­nized by T cells seven or eight AA se­quence with ≥ two epitop­es [100].

Casein has been found in whey HFs and vice versa: residual casein epitopes in all the HA formulas Alfa-Rè, LHA and Pregomin were detected [101]. Trace amounts of casein are present in com­mercial whey preparations, accordingly casein IgE epitopes could be demonstrated in a num­ber of such products [81]. In another study [23] Alimentum and Nutramigen had det­ectable casein and whey proteins, how­ever GS had more than 700 times the amount of de­tectable whey proteins, more than 100 times the amount of casein proteins, and more than 700 times the detectable whole CM proteins than the other two HFs. GS was found to have 2.625 mg of casein/g protein [88]. A recent study [56] has shown by in­hi­bi­tion-ELISA that Nutramigen has a 64,7% of whey conten and Alimen­tum about 50%, whereas Profylac haa a 36,6% of casein content and Alfarè of 17,9% [56]. The appar­ent contra­diction of ca­sein in WHFs and vice versa can be easily understood, since employing the TCA precipi­tation and SDS-PAGE to obtain the separa­tion of whey pro­teins from ca­sein [87], aliquots of ca­sein may remain in the whey frac­tion due to ca­sein-derived low-MW pept­ides [95]. Heating casein at a T = 1210C for 15 minutes is not enough, and boiling it for 30 minut­es is neces­sary to reduce its antigen­icity and immunogenici­ty significantly [89]. Most IgE antibod­ies to CM proteins bind short chains of AAs that have mostly se­quen­tial epitopes, T epi­topes [92], in addi­tion to the high allergen city of the casein k frac­tion [90].

Additional effects are seen in nursing newborns. To appreciate how the elimination of all the epitopes of CM pro­tei­ns is challenging [102], we point out that nursing babies can be sensitized to such pro­teins even through BM. Once consumed by the nursing mother, after a first denaturation with cooking, the CM peptides un­dergo in the ma­ter­nal gut an enzy­ma­ti­c hydrolysis and further denaturation and as­similation, pass through bio­lo­gic mem­branes, and go into the blood and to the breast (entero­mam­mary axis). Reaching thus the infant's gut, they are again hy­drolyzed, ab­sorbed, and enter into the APC where they ex­perience their hun­dredth transfor­ma­tion, yet CM pept­ides are still immunogen­i­c [80] despite the pre­s­ence of se­cretory IgA in BM [103]. Such data explain also why babies who appar­ently never received CM native pro­teins, but were surely fed a HF supple­ment in their very first days of life, at the age of few months when still excl­usi­ve­ly BM-fed ex­per­ien­ced ana­phylaxis after a few ml of the same for­mu­la [7] as the babies fed whole CM in the Ma­ternity [78]. How­ever, we face the crude reality that 50-60% if not 100% of HR children receive a CM formula in the nur­sery without our knowledge [78]. In our hospital, and surely others too, also ba­bies who are meant to be BM-fed receive a pi­rate bottle during the night if they cry [2,78].

Minimal levels of ßLG in BM (Table 6) [25,84,103-109] can be al­lerge­nic at the point of triggering anaphylaxis [33], although IgA antibodies to CM proteins can play an im­por­tant role in the exclusion and elimination of ßLG [103]. Accord­ing to data shown in (Table 6) BM has less mean ßLG levels than ready-to-use Nutrami­gen: 7-12 times after drinking 400 ml of CM, and 84 in basal conditions. It fol­lows that the as­sumption of HFs declared to be HA conta­ining little amounts of residual ßLG, however in con­cen­trations much higher than those found in BM, can trigger severe ana­phy­lac­tic reactions in ba­bies with IgE-mediated CMA, elicited by IgE cross-reactive with epi­topes pre­sent de­spite the enzy­matic hydrol­y­sis or even multiplied by it [85]. It is of note that certain extensi­vely HFs contaminat­ed by pro­­teins left intact by enz­ymati­c hy­drol­ysis, con­tain signifi­c­ant amounts of ßLG de­tected with the methods of ELISA and RAST-inh­ibition: in dry powder a CHF con­tains 0,0056 ± 0,0005 mg/g, a WHF 200: in the ready-to-use formu­las the levels varied between 0,84 ± 0,07 and 31.200 ± 3.744 mg/l (Table 6). It has been shown that products of digestion of ßLG retain allergenic activity [91]: Haddad et al de­monstra­ted that 10 CM-aller­gic children with or without IgE antib­odies to undig­ested ßLG all had IgE antibodies to the peptic or pep­tic and tryptic digests of ßLG [110]. In­deed ßLG is stable to digestion for 60 min, thus as other intact proteins is capable of crossing the gut mucosal membrane and of entering the cir­culatory sys­tem, with all likelihood to be ab­sorbed, hence eliciting allerg­enic respon­ses [111].

Several studies have shown in animal models that CM protein HFs failed to elicit an IgG anti­body re­s­ponse to CM protein or induce passive cutaneous anaphylaxis [79,102,112-114]. Subsequently, Boner et al. [115] evaluated the allergen­icity of a PWHF in gui­nea pigs fed ad libitum CM, pasteuriz­ed CM (PCM), or PWHF. On day 37 were IV chal­lenged: one fa­tal reaction was pro­voked by the sequence PWHF-PWHF or PCM-PWHF, and non­fa­tal re­ac­tions by the same se­quences, and the sequence PCM-PWHF [115]. Again in gui­nea pigs even em­ploying the more rigorous challenge schedule, the extensi­vely hydro­lyzed formulas failed to cause evident symptoms of ana­phylaxis [116]. Cordle et al. [117] using the rabbit hyperimmunization model, have measur­ed with the ELISA im­muno­as­say the immunologic active bovine whey pro­tein (IAW) and acti­ve bovine ca­sein (IAC) in HFs demonstrating that ei­ther for­mula conta­in­ing intact, or partially, or ex­ten­sively hy­drolyzed proteins, elicited im­mune resp­on­ses (mg/g pro­tein): the par­tially hydrolyz­ed 1060-1210 IAC and 140,000-210,000 IAW, the highly hyd­­rolyzed 7.68-12.2 IAC and 19.1-25.6 IAW, while the intact CM-for­mulas 400,000-800,000 IAC and 180,000-600,000 IAW).

Consequently, the pro­tein mole­cules, even if hydrolyzed in pep­tides of very low MW, retain all their patrim­o­ny of im­munoreac­tive epi­topes, cap­able either of sti­mu­lat­ing the im­muno­com­pe­tent cells of the baby, or in­ducing the synthesis of IgE antibod­ies. Hence all the epito­pes of a CM protein, na­tive or ma­ni­pulated, are potential­ly im­muno­g­enic, but they can defined as such only with re­fer­ence to the sensitized in­divi­dual. [85].

Laboratory diagnosis of children with HF allergy

Laboratory diagnosis can be done employing in vitro and in vivo tests, use­ful also for clinical testing in at risk infants before the use of HFs, and clinical testing of formulas with standardized procedures [2,57,74].

The most reliable methods are those using the RAST, the ELISA, and their inhibition variants, since their results are not crucially de­pendent on epitope density. In Table 6 are shown the results obtained with the ELISA and RAST in­hibition tests. The very great differences demon­strated be­tween PWHFs and CHFs have been comment­ed ear­lier. Ana­l­yz­ing the dry weight con­cen­trations, the PWHFs ap­pear to be 585 times less aller­genic, and the EWHFs 21x106 less al­ler­gen­ic than CM. Eval­uat­ing the ready-to-use levels, the figures are 128 and 4.760.000, respec­tively.

As an alternative, the FAST inhibition test can be utilized (118). Using sera from pa­tients with CM anaphylaxis and high IgE ac­tivity to CM, casein, ßLG, and a-lac­talbu­min, the authors evaluated Nutr­amigen, Alimentum, GS, and Alfa-Rè. All the HFs bound IgE spe­cific to CM and its allergens, although in general the HFs bound sig­nifi­cantly less IgE than the CM formulas, with the notable excep­tion of GS (Tables 7, 8, 9).

SPTs, RAST and DBPCFC have been employed as detailed be­fore [18-20, 23]. Among 26 children with different forms of CMA, there was a group with more severe reactions to CM, which also had positi­ve SPTs to CHFs (Alimentum, Nutr­amigen), while children with less sev­ere form of CMA had not positive SPTs to these products, but all had posi­tive SPTs to the PWHF [119]. However, both SPTs and RAST failed to correlate with clini­cal react­ivity to Nutramigen [56], while 2 out of 11 CMA children (18,2%) with SPTs positive to the formula experienced a clinical reaction fol­low­ing an OFC [37]. Thus it appears that babies with SPTs positive to HFs cannot ingest HFs with impunity.

The RAST in­hibition may be useful to predict the re­activity or cross-reac­tivity of IgE anti-CM antibodies with infant formu­las. Accordingly, 1-4% of infants with CMA would recognize epitopes of CM proteins in about one of the HFs. Six out of 13 infants with CMA (46%) indeed reacted to four HFs, Alimentum, Alfa-Rè, GS, and Nutramigen [120].

Clinical testing of formulas in children with HF allergy

Clinical testing of HFs with standardized procedures has been recommen­ded by the ES­PACI [57], and the Sub­committee of Nutrition and Allergic Disease of the AAP [74]. The AAP Subcommittee has recommended that DBPCFC stud­ies be conducted with standa­rd­ized procedures, also to exclude the possibility of a type II er­ror in the data. In the case of HFs, in­fants and chil­dren with documented CMA should be studied in conformity. A suffi­cient number of such pa­tients should be enrolled in a prospective study organized in such a way that it can be pro­jected with 95% confi­dence that 90% of chil­dren with CMA will not react to the HF stud­ied. This model was followed by two studies [23,55].

Considering the 90% as a cut-off point, and following the procedures, a group fed HFs should be formed by 28 chil­dren with CMA free of clinical symptoms. If 1 or 2 infants react, the sample should be increased up to 46 or 61 children, re­specti­vely. If the 90% level is judged a too low cut-off, to use a 95% prob­abil­ity level, the sample should con­sist of 120 children with no re­action, but if only one infant reacts to the HF, the group must be en­larged to en­compass approximately 400 infants (plus 400 controls equally representa­tive). In our opinion it would be inconceivable to re­cruit so many infants, who in ad­di­tion should comply with that way of approach­ing the above recommenda­tions. Facit: be­fore em­bark­ing on a new study or reporting an anaphy­lactic reaction one should con­sult an expert in statistics. Halken et al. [55], have recommen­ded to per­form a strictly su­pervised OFC in children with immediate reac­tions to CM be­fore start­ing whatever HF. In two infants the anaphylactic symptoms ap­peared as soon as one drop of CM or Alfa-Ré was admin­istered, and 5 ml to the other three [50].

It might be appropriate that HFs, before they are employed in vivo in in­fants with IgE-me­diated allergies, be studied with adequate methods such as RAST-inhibition, CRIE, im­mu­noblot­ting, ELISA, etc, and studies be organized according to previ­ous spec­ifica­tions [57,74,75]. The lab­oratory a­ni­mal hyper­immuniza­tion model proposed to eval­u­ate the immun­o­genicity of protein HFs (117) could re­duce the costs, the risks and the time in­he­rent in the DBPCFC studies.

Management of children with allergy to HFs

For the treatment of atopic children, the suitable CM sub­sti­tut­es are BM if still availa­ble, SPFs [3], HMMBF [2], and L-amino acid formula (AAF) [37,40,41,75]. As regards the AAF, the above references seem to be suf­ficient for a correct use. HMMBFs were first employ­ed in 1973 [121] and are HA in the strict sense, since reactions to HMMBF were never re­port­ed, are very well ac­cepted by newborns, infants [122], chil­dren aged 8-9 years [123] and even par­ents, also insuring a regular growth [122].

We stress that many concerns regarding the HFs depen­d by the striking lack of scien­tif­ic data provided by the Companies regarding a lot of tech­ni­cal aspects, not useful even to the Companies. However, the real composition of these HFs remains largely un­known. Moreover, varying the tec­hno­logies according to the different Compa­nies, at least this technical information should be gi­ven to the medical special­i­sts or in any case to the doctor prescribing the HF. Not even topics such as the pos­sible contamina­tion of the product during fabrication or packing pro­ced­ures, or the lot-to-lot variabil­ity of HFs have been discussed, and we think that a word of caution is needed.

Although the proteins of HFs have been processed by physico-chemical procedu­res and therefore contain peptides of lower MW than the native protein source, the pepti­des still have allergenic potency and can be recognized by cell-bound IgE of a child with CMA. We point out that the chan­ges in the intestinal mu­cosa [39], which unless careful­ly sought after are likely to be missed. One child who in spite of a bipar­ental history of atopy was not subjected to prophylactic measures was dead 8 hours after the first of two feedings of PHWF (12 hours after the last SPF was given) [124].

On clinical grounds, reactions to casein were first observed by Glaser in 1944 [125]. We have documented >202 allergic reactions induced by HFs, from 17 cases of shock-anaphylaxis (ome every 3.3 years), 3 apparent life-threatening events, and 18 systemic reac­tions (17.8%) to worsening of eczema lesions and GI symptoms (Tables 3, 4). In compa­ri­son, SPFs have provoked one case of anaphy­laxis/22.3 years [3], with a difference of 676%. Sampson has proposed a 10-18% prevalence for HF allergy (in high atopic children) [37,56], but in Milan 86% of neonates are rec­ommen­ded HFs (14% par­tially and 70% ex­tensively), also in the absen­ce of a history of family atopy [126]: how much can incr­ease the prevalence of HF allergy and of ad­verse reac­tions? On scientific grounds, we conclude that neither SPTs nor RAST are rel­i­able, whereas only DBPCFC (OFC in infants) should be considered diag­nos­tic [57,74,75]. Thus, further large trials should be performed using well-de­fined samples of HR in­fants.