8/13/2019 Bilirrubina 31874289[1]

1/4

Acta Pdiatrica ISSN 08035253

R E G U L A R A R T I C L E

Neonatal hyperbilirubinemia increases intestinal protein permeabilityand the prevalence of cows milk protein intoleranceFrancesco Raimondi (raimondi@unina.it)1, Flavia Indrio2, Valeria Crivaro1, Gabriella Araimo1, Letizia Capasso1, Roberto Paludetto1

1.Department of Pediatrics, Division of Neonatology, Universita Federico II, Napoli, Italy

2.Department of Pediatrics, University of Bari, Bari, Italy

Keywords

Bilirubin, Intestine, Intolerance, Neonate,Permeability

Correspondence

Francesco Raimondi, M.D., Department ofPediatrics, Division of Neonatology, UniversitaFederico II, Via Pansini 5, 80131 Naples, Italy.Tel/Fax: +390817462908 |Email: raimondi@unina.it

Received

23 November 2007; revised 25 January 2008;accepted 6 February 2008.

DOI:10.1111/j.1651-2227.2008.00746.x

Abstract

Aims: Bilirubin is a newly discovered modulator of the gut barrier in vitroand in vivo. We studied the

effect of bilirubin on the serosal to mucosal intestinal permeability in vivo. We also investigated the

prevalence of cows milk protein intolerance (CMPI) in infants with moderate hyperbilirubinemia

versus matched controls.

Methods: Faecal alpha 1 antitrypsin (a1AT) was used to monitor intestinal protein loss; a large cohort

was prospectively followed for 12 months for sign and symptoms of CMPI.

Results: Neonates with hyperbilirubinemia had higher stool excretion of a1AT than controls (0.68

0.28 mg/g vs. 0.25 0.11 mg/g; p < 0.01). Faecal a1AT correlates with total serum bilirubin (TSB)

(r = 0.85; p < 0.01). Also, in the first 12 months of life, formerly hyperbilirubinemic infants had an

higher prevalence of CMPI (14/353 vs. 4/339; 2

= 4.018, p = 0.045).

Conclusions: Neonatal hyperbilirubinemia increases stool protein loss and is also a mild risk factor for CMPI.

INTRODUCTION

The intestinal epithelium is the single widest interface of the

human body with the outer world. Far from being a pas-

sive fence function, the mechanisms of selective perme-

ation of the intestinal barrier have been extensively investi-

gated in the past two decades, and recent studies have sheda new light on the complex relation between intestinal per-

meability and human disease (1). For instance, a gut barrier

failure in the neonatal period in a genetically predisposed

individual has been linked to an early food allergen sensiti-

zation (2). New modulators of intestinal permeability havebeen found. Among molecules of different chemical nature,

we have recently demonstrated that unconjugated bilirubin

(UCB) increases, in a reversible fashion, the space between

adjacent enterocytes (the so-called paracellular pathway) in

anin vitromodel of intestinal epithelial barrier (3). We have

also shown that hyperbilirubinemic neonates have an in-creased intestinal permeability investigated with the double

sugar urinary recovery technique (4).

The present paper provides further evidence that bilirubin

is anin vivomodulator of the gut barrier using faecal alpha

1 antitrypsin (a1AT) dosage. Because an enhanced intestinalpermeability could promote allergen sensitization, we have

also tested the hypothesis that neonatal hyperbilirubinemia

increases the risk of cows milk intolerance.

PATIENTS AND METHODS

Noninvasive determination of intestinal permeability

by faecal a1AT determination

Patients

Healthy, full-term, appropriate for gestational age (AGA)

newborns with normal Apgar scores and total serum biliru-

bin (TSB) between 12 and 17.9 mg/dL were enrolled in

the hyperbilirubinemia group. Consecutive healthy, full-term

newborns with peak TSB less than 8 mg/dL were matched

with the cases for gestational age, sex and birthweight.Exclusion criteria were: cephaloematoma, congenital mal-

formation, inborn errors of metabolism, proven sepsis or

infection, maternal-neonatal blood group incompatibility,

glucose-phosphate dehydrogenase deficiency or erythrocyte

membrane defects or formula feeding. All newborns were

studied at days 69 of life, when meconium-free stoolswere passed regularly. A subgroup of 10 hyperbilirubinemic

neonates had a second stool sample taken at days 1012,

when TSB was declining. Newborns did not receive pho-

totherapy until the end of the study. Informed consent was

obtained from the parents, and the study was approved byLocal Institutional Ethics Committee. For TSB measure-

ments, a blood capillary sample was drawn in all neonates,

centrifuged for 5 min and assessed by colorimetric method

(Elmed Ginevri, Roma, Italy).

Faecal a1AT quantitation

a1AT is a serum antiprotease, not catalyzed by digestive

enzymes, although it is presently unknown whether a1AT

undergoes significant digestion by microbial proteolytic en-zymes in the neonatal gut. Quantification of stool a1AT is

an established, noninvasive method both in clinical research

and practice to study protein leakage from the vascular com-

partment into the intestinal lumen (5). Stool samples from

both groups were collected and stored at 80C until theday of analysis. Stool a1AT was measured by radial im-

munodiffusion; LC Partigen plates and human standard

serum proteins LC-V (Dade Behring Spa, Milano, Italy) were

used. Faecal samples were weighted, diluted 1:4 with saline

C2008 The Author(s)/Journal Compilation C2008 Foundation Acta Pdiatrica/Acta Pdiatrica 200897, pp. 751753 751

8/13/2019 Bilirrubina 31874289[1]

2/4

Bilirubin, gut permeability and milk intolerance Raimondi et al.

solution 0.9% and homogenized over 2 min, then centrifuged

at 4.500 g over 20 min. Standard serum proteins LC-V wereused to obtain a calibration slope nondiluted (solution I),

diluted 1:2 (solution II) and 1:4 (solution III) with saline

solution 0.9%. Each LC Partigen plate was incubated with

20 L of solution I, II, III and patients samples for 72h. Dur-

ing that interval, specific antibodies in the agarose gel linkeda1AT or standard serum proteins LC-V and formed immuno-

complexes; the diameter of the ring caused by precipitationof immunocomplexes was measured with a ruler by two in-

dependent blinded observers using a calibrated lens. We re-

lated the diameter of the rings of the patients samples to

the calibration slope to calculate the concentration of a1ATin milligrams per gram of stool and then we referred each

result to 1.5 g of stool (wet weight).

Studying the prevalence of CMPI in formerly

hyperbilirubinemic infants

Prevalence of cows milk protein intolerance (CMPI) in the

first year of life was prospectively investigated in a cohort of

consecutive neonates with moderate hyperbilirubinemia, asdefined above, and controls matched for gender, birthweight,

family income, maternal age, maternal smoking during preg-

nancy and parental history of atopic diseases.

All infants were born at the University Federico II Mater-

nity Division between January 2005 and May 2006. Parental

informed consent was obtained. Preterm newborns and/orwith pathologic jaundice were excluded.

At the age of 12 months, parents were administered a ques-

tionnaire focused on the clinical manifestations of CMPI. All

children had attended the clinics of primary care paediatri-

cians, and medical records of all food intolerance episodeswere carefully reviewed. According to the literature (6), only

medical diagnoses based on resolution of symptoms after

elimination of cows milk proteins (CMPs) from the diet

were considered.

Statistics

Data, expressed as mean standard deviation, were

analysed with the help of a statistical software package

(SPSS/PC+, SPSS, Inc., Chicago, IL, USA). Faecal a1AT

content in hyperbilirubinemic neonates and matched con-trols were investigated with the t-test for paired values. Pear-

sons coefficient was used to study the relation between TSB

and faecal a1AT in all patients. Wilcoxon rank-sum test was

used for those patients who had two stool samples taken for

a1AT determination. Finally, chi-square tables were built toanalyse the incidence of CMPI in a cohort of hyperbiliru-

binemic infants and matched controls. The sample size had

been previously calculated with the help of a commercial

software package (EpiInfo, EpiSoftware, Ravenna, Italy) set-

ting the confidence interval at 95%, the power of the study

at 80%; the prevalence of CMPI in the infantile populationunexposed to jaundice was estimated at 3%. Three hundred

fifteen infants were requested in each group for an estimated

odds ratio equal to 3.

RESULTS

Faecal a1AT determination

Twenty casecontrol pairs were enrolled in the study. Stool

a1AT quantification was successfully performed in all stud-

ied newborns. No statistically significant difference was

found between hyperbilirubinemic and nonhyperbilirubine-

mic group for the following: mean gestational age (38.08 0.9 weeks vs. 38.78 1.64), birthweight (3057.50 243.69 g

vs. 3107.78 391.46 g), Apgar score (8.7 0.7 vs. 8.9 0.3)

and sex. Two neonates from each group were on formula and

breast milk feeding; all other neonates in both groups were

on breast milk. At the time when stool samples were col-lected, the mean serum bilirubin was statistically different

between hyperbilirubinemic infants and controls (14.34 1.95 mg/dL vs. 4.58 1.99; p < 0.01). The mean stool a1AT

was also statistically different between hyperbilirubinemic

and nonhyperbilirubinemic group (0.68 0.28 mg/g vs.

0.25 0.11 mg/g; p < 0.01). When plotting the two setsof samples from hyperbilirubinemic babies, a significant de-

crease of a1AT was associated with a declining TSB while

controls showed no variation (Fig. 1). Finally, Figure 2 de-

picts a direct correlation between TSB and a1AT (r = 0.853;

p < 0.01), when all samples were pooled together.

Studying CMPI prevalence in formerly

hyperbilirubinemic infants

Seventeen out of 353 neonates in the hyperbilirubinemic

group had signs and symptoms related to CMPI, but because

three of them were not adequately challenged, only 14 CMPI

diagnoses were considered versus 4 out of 339 in the con-

trol group ( 2 = 4.018, p = 0.045). Basic clinical features ofthese diagnoses included atopic dermatitis (3 control infants,

11 hyperbilirubinemic infants), blood in the stools (3 cases

in the jaundiced group), isolated vomiting (1 control baby)and poor growth (3 hyperbilirubinemic infants). Four babies

in the jaundiced group had positive prick tests to CMP.

0,0

0,2

0,4

0,6

0,8

Second

sampleFirst

sample

Hyperbilirubinemic neonates

Controls

a1AT

(mg/grfeces)

Figure 1 Time-course of a1At stool excretion in 10 hyperbilirubinemic-control

matches. A first stool sample was taken at day of life 69 and a second sample

at day of life 1012 for determination while TSB was declining; a1At = alpha

1 antitrypsin; TSB = total serum bilirubin.

752 C2008 The Author(s)/Journal Compilation C2008 Foundation Acta Pdiatrica/Acta Pdiatrica 200897, pp. 751753

8/13/2019 Bilirrubina 31874289[1]

3/4

Raimondi et al. Bilirubin, gut permeability and milk intolerance

0 5 10 15 200,0

0,2

0,4

0,6

0,8

1,0

1,2

1,4

1,6

1,8

fecala1AT(mg/g)

Figure 2 Correlation plot between faecal a1AT obtained at day of life 69 and

TSB in the hyperbilirubinemia and control groups (r = 0.85; p < 0.01).

DISCUSSION

The present work confirms that hyperbilirubinemic babies

have a higher intestinal permeability than control babies.Compared to the urinary disaccharides recovery, a vali-

dated method to investigate mucosal to serosal permeabilityto small probes (7), a1AT studies the blood to lumen (i.e.

serosal to mucosal) percolation of a larger marker. Similarly

in an in vitro system, we were able to show that UCB pro-

motes the passage of a 40 kDa marker, the size of many food

allergens, from the lumen to the serosa (3) and vice versa(F Raimondiunpublished results). The link between hyper-

bilirubinemia and intestinal protein loss is further strength-

ened by the linear relation depicted in Figures 1 and 2. An

increased intestinal permeability has been shown in sev-

eral immune-mediated conditions. Indeed, large epidemi-

ologic studies have linked neonatal jaundice and/or pho-totherapy to infantile bronchial asthma or diabetes (8,9).

Although the mechanismremains unknown, a possible inter-

ference with neonatal gut immune response or gut function

at large has been suggested (10). Given the retrospective na-

ture of the study, it was impossible to separate the effects ofbilirubin from those of phototherapy. We have conducted a

prospective cohort study showing that formerly jaundiced

infants who did not reach the threshold of phototherapy

have a higher risk of developing CMPI, a common, fre-

quently immune-mediated disease. It has been reported thatsensitization to food protein may occur during lactation and

is not limited to formula-fed infants (11). Bilirubin may act as

a chaperone facilitating the sensitization. We cannot provethat a bilirubin-mediated increase in intestinal permeability

has a definite play in this phenomenon because we could notmeasure the intestinal permeability of such a large cohort.

However, the hypothesis remains attractive as jaundiced ba-

bies have an increased intestinal permeability and the latter

is related in a dose-dependent fashion to the severity of food

intolerance manifestations (12).

CONCLUSION

In conclusion, hyperbilirubinemia is associated to an in-

creased intestinal permeability. This, in turn, may be lead-

ing to an enhanced allergen sensitization that would explain

the higher number of CMPI diagnosed in formerly jaundicedbabies. Our work and recent data from other groups demon-

strate that neonatal jaundice has a longer-lasting effect on

individual health than previously expected. Future research

will tell if this influence is going to affect clinical decisionmaking.

References

1. Cereijido M, Contreras RG, Flores-Bentez D, Flores-Maldonado C, Larre I, Ruiz A, et al. New diseases derived orassociated with the tight junction.Arch Med Res2007; 38:46578.

2. Heyman M. Gut barrier dysfunction in food allergy.Eur JGastroenterol Hepatol2005; 17: 127985.

3. Raimondi F, Crivaro V, Capasso L, Maiuri L, Santoro P, TucciM, et al. Unconjugated bilirubin modulates the intestinalepithelial barrier function in a human-derived in vitro model.

Pediatr Res 2006; 60: 303.4. Indrio F, Raimondi F, Laforgia N, Riezzo G, Polimeno L,

Francavilla R. Effect of hyperbilirubinemia on intestinalpermeability on healthy newborns.Acta Paediatr2007; 96:735.

5. Florent C, LHirondel C, Desmazures C, Aymes C, Bernier JJ.Intestinal clearance of alpha 1-antitrypsin. A sensitive methodfor the detection of protein losing enteropathy.Gastroenterology1981; 81: 77780.

6. Novembre E, Vierucci A. Milk allergy/intolerance and atopicdermatitis in infancy and childhood. Allergy2001; 56 (Suppl.67): 1058.

7. DeMeo MT, Mutlu EA, Keshavarzian A, Tobin MC. Intestinalpermeation and gastrointestinal disease. J Clin Gastroenterol2002; 34: 38596

8. Dahlquist G, Patterson C, Soltesz G. Perinatal risk factors forchildhood type 1 diabetes in Europe.Diabetes Care1999; 22:1698702.

9. Aspberg S, Dahlquist G, Kahan T, Kallen B. Is neonatalphototerapy associated with an increased risk for hospitalizedchildhood bronchial asthma? Pediatr Allergy Immunol 2007;18: 3139.

10. Dahlquist G, Kallen B. Indications that phototherapy is a riskfactor for insulin dependent diabetes. Diabetes Care2003; 26:2478.

11. Heine RG, Hill DJ, Hosking CS. Primary prevention of atopicdermatitis in breast-fed infants: what is the evidence?J Pediatr2004; 144: 5647.

12. Ventura MT, Polimeno L, Amoruso AC, Gatti F, Annoscia E,Marinaro M, et al. Intestinal permeability in patientswith adverse reactions to food. Dig Liver Dis 2006; 38:7326.

C2008 The Author(s)/Journal Compilation C2008 Foundation Acta Pdiatrica/Acta Pdiatrica 200897, pp. 751753 753

8/13/2019 Bilirrubina 31874289[1]

4/4