Download pdf - Effect of Lactobacillus GG supplementation on pulmonary exacerbations in patients with cystic fibrosis: A pilot study

ARTICLE IN PRESS

Available at www.sciencedirect.com

Clinical Nutrition (2007) 26, 322–328

0261-5614/$ - see frdoi:10.1016/j.clnu.

�Corresponding a

E-mail address:

http://intl.elsevierhealth.com/journals/clnu

ORIGINAL ARTICLE

Effect of Lactobacillus GG supplementation onpulmonary exacerbations in patients with cysticfibrosis: A pilot study

Eugenia Bruzzesea, Valeria Raiaa, Maria Immacolata Spagnuoloa,Monica Volpicellia, Giulio De Marcoa, Luigi Maiurib,c, Alfredo Guarinoa,�

aDepartment of Pediatrics, University of Naples ‘‘Federico II’’, Via S. Pansini 5, 80131 Naples, ItalybInstitute of Child Health, University College London, London, UKcEuropean Institute of Research in Cystic Fibrosis, Verona, Italy

Received 14 July 2006; accepted 19 January 2007

KEYWORDSCystic fibrosis;Probiotics;Pulmonary exacerba-tion

ont matter & 20072007.01.004

uthor. Tel.: +39 081

[email protected] (

SummaryBackground & aims: Probiotics reduce intestinal inflammation in children with cysticfibrosis (CF). We want to determine the effects of Lactobacillus GG (LGG) on pulmonaryexacerbations in CF.Methods: A prospective, randomized, placebo-controlled, cross-over study was per-formed. Nineteen children received LGG for 6 months and then shifted to oral rehydrationsolution (ORS) for 6 months. In parallel nineteen received ORS and then shifted to LGG.Main outcome parameters were: incidence of pulmonary exacerbations and of hospitaladmissions, forced expiratory volume (FEV1), and modifications of body weight.Results: Patients treated with LGG showed a reduction of pulmonary exacerbations(Median 1 vs. 2 , range 4 vs. 4, median difference 1, CI 95% 0.5–1.5; p ¼ 0.0035) and ofhospital admissions (Median 0 vs. 1, range 3 vs. 2, median difference 1, CI95% 1.0–1.5;p ¼ 0.001) compared to patients treated with ORS. LGG resulted in a greater increase inFEV1 (3.6%75.2 vs. 0.9%75; p ¼ 0.02) and body weight (1.5 kg71.8 vs. 0.7 kg71.8;p ¼ 0.02).Conclusions: LGG reduces pulmonary exacerbations and hospital admissions in patientswith CF. These suggest that probiotics may delay respiratory impairment and that arelationship exists between intestinal and pulmonary inflammation.& 2007 Elsevier Ltd and European Society for Clinical Nutrition and Metabolism. All rightsreserved.

Elsevier Ltd and European Society for Clinical Nutrition and Metabolism. All rights reserved.

7464232; fax: +39 081 5451278.

A. Guarino).

dx.doi.org/10.1016/j.clnu.2007.01.004

mailto:[email protected]

ARTICLE IN PRESS

Lactobacillus GG as adjunctive therapy in cystic fibrosis 323

Introduction

The major cause of morbidity and mortality in cystic fibrosis(CF) is progressive inflammatory lung disease which even-tually results in end-stage respiratory failure in 90% of CFpatients.1 During the disease, most children becomecolonized with Pseudomonas aeruginosa and undergo amore rapid clinical course with progressive impairment ofrespiratory function.2 Therefore, colonization with Pseudo-monas identifies a subset of patients at increased risk ofpulmonary exacerbations and is associated with a decreasein survival rate.3 In an attempt to reduce the rate andseverity of pulmonary exacerbations, children with CF areput on heavy load of antibiotics.

Intestinal inflammation is another typical feature of CF4

and is much more common than previously thought.Recently, we reported that fecal calprotectin concentrationand rectal nitric oxide production are increased in virtuallyall children with CF, suggesting that intestine is a targetorgan in CF and is constantly in an inflammatory state.5

Probiotics are live bacteria that are administered orallyand may decrease the severity and duration of childhoodgastroenteritis.6 However, there is evidence that probioticsprevent relapses of chronic inflammatory bowel diseaseswhen given in adjunct to standard therapy.7 Probiotics havealso been successfully used in the treatment8 and preven-tion9 of atopic disorders. Interestingly, probiotic supple-mentation reduced the incidence of respiratory infections inchildren attending day care center.10

Probiotics may protect from gastrointestinal diseases byacting on intestinal permeability, as witnessed by theircapacity to improve intestinal barrier function.11 However,it is becoming increasingly clear that several effectsinduced by probiotics are related to modifications ofimmune response,12 which may explain the clinicaleffects of probiotics observed in non-intestinal and insystemic diseases. Several probiotic strains have beeninvestigated in a number of clinical conditions. The bestcharacterized probiotic is Lactobacillus rhamnosus strain GG(LGG), for which several clinical effects have beenidentified. Recent data showed that LGG affects theexpression of genes involved in immune response andinflammation.13

Children with CF may be ideal candidates for probioticsupplementation. Indeed the intestinal microflora of thesechildren is often abnormal due to massive exposure toantibiotics, and in addition their intestinal permeability isincreased suggesting disruption of intestinal barrier func-tion.14 Recently we have reported that LGG reduces nitricoxide and fecal calprotectin, two markers of intestinalinflammation, in CF children.5 Furthermore cystic fibrosis isa prototype of chronic inflammatory disease and is asso-ciated to a dysregulation of inflammatory response andinnate immune mediators.15

The increased inflammation in the intestine inchildren who also have chronic lung inflammation,together with the anti-inflammatory effects observed inthe distal intestine,5 raised the hypothesis that probioticadministration may reduce the incidence and/or severity ofpulmonary exacerbations in children with CF. We havetested this hypothesis through a randomized cross-overstudy.

Patients and methods

Children with CF, diagnosed by an elevated sweat Cl�

concentration, with moderate/severe disease and chroni-cally infected with Pseudomonas, were enrolled in thisprospective, randomized, placebo-controlled, single blind,cross-over trial.

All patients had pancreatic insufficiency as judged bysteatorrhea and clinical symptoms. The severity of pulmon-ary disease was evaluated according to the criteria of the CFFoundation16 Patients were defined as colonized with P.aeruginosa when this was found in their sputum for at least 6months in at least three consecutive cultures. All CFpatients underwent courses of suppressive antimicrobialtherapy, usually with oral or intravenous administration oftwo different antibiotics for 14–21 d, each 3 months,according to the accepted clinical practice at our CFreference center and based on specific guidelines.17 Theyalso underwent preventive administration of inhaled anti-biotics. All were enrolled within one month from the lastsuppressive therapy and were divided in two groups.Those in group A were started on LGG (6� 109 CFU/d)dissolved in oral rehydration solution (ORS) for 6 months(period 1) and then shifted to ORS per day for further 6months (period 2). Conversely, children in group B receivedORS for 6 months (period 1) and were then shifted to LGG(period 2). LGG or ORS were assumed, blindly, in a singledaily dose by the patients. The LGG dosage used for theenrolled patients is directly available in the market. Awashout period of 4 weeks was introduced between LGG andORS administration.

The randomization was constructed using the randomnumbers table with a block design for groups of ten patients.The patients were enrolled over a period of 6 months. Themethods to prevent selection biases were not be undertakenbecause very stringent criteria of inclusion were chosen tolimit the baseline group differences.

The outcome measures of efficacy were recorded by oneof the investigator who was unaware about the administra-tion of LGG/ORS. They reflected the incidence and severityof pulmonary exacerbations, respiratory function, generalinflammation and general clinical conditions and included:

(a)
Number of episodes of pulmonary exacerbations. Thesewere identified on the basis of an increase in pulmonarysymptoms and airway secretions, as reported by the CFFoundation Criteria.16 The duration in days of eachepisode of pulmonary exacerbation corresponded to theduration of antibiotic therapy prescribed for thatspecific episode. Days of prophylactic antibiotic admin-istration were not taken into account in the evaluationof the duration of respiratory infections. However, nocase of pulmonary exacerbation occurred during pro-phylactic treatment.
(b)
Number and duration of hospital admissions required forpulmonary exacerbations.
(c)
Route of administration of antibiotics (oral vs. intrave-nous administration). This reflected the severity ofrespiratory episodes. Hospital admission and discharge,as well as the choice and the duration of antibiotictreatment, were established by the physicians in charge

ARTICLE IN PRESS

T

SMHMM

E. Bruzzese et al.324

of each patient, who were blind to the administration ofLGG/ORS.

(d)
Forced expiratory volume in 1 s (FEV1) measured at thebeginning and end of either ORS or LGG period. FEV1 wasselected as this is the variable that best reflects thestatus of lung function throughout the course of CF lungdisease19 and is the best clinical predictor of a pooroutcome.20 For each test, at least 3 expiratorymaneuvers were performed and the best result wasselected.21 FEV1 values were expressed as % modifica-tions of the value measured prior to start ORS or LGG;
(e)
Body weight and body mass index (BMI) at enrolment andafter LGG and ORS. Both parameters were expressed asmodifications of basal values and were collected at theonset of each treatment period.
(f)
Serum immunoglobulin concentrations (IgA, IgG andIgM), which reflect the inflammatory state in childrenwith CF18 Ig were determined by standard method.
All parameters were determined at the onset and at theend of ORS and LGG and were recorded in children who werein stable clinical conditions. Clinical stability was defined asthe absence of substantial, acute onset clinical problemsthat could affect the outcome measures included in theclinical protocol, at the time of enrolment, shift oftreatment and/or exit from treatment. Outcome measureswere evaluated within two weeks after the end of eachobservation period. Compliance with probiotic assumptionwas checked with the mothers of children enrolled duringroutine clinical controls.

The Ethics Committee of our institution approved thestudy. Informed consent was obtained from parents of allchildren enrolled. It needs to be considered that probioticsare classified as ‘‘food additives’’ and not drugs by theItalian Health Ministry, and this was clearly stated when theconsent was requested.

Statistical analysis

Simple size was calculated assuming a mean incidence ofpulmonary exacerbations in the control group of 4 episodes/child/12 months, and a reduction of about 30% in theincidence during the treatment, with a significance levelalpha of 0.025, and a power of 0.8. With this assumption andby considering the cross-over study design, number ofchildren needed was 30 for each group. However, due tothe very selective inclusion criteria, we were able to enrolonly 19 patients for each group.

able 1 Clinical baseline features of the enrolled populations

Grou

ex (M/F) 6/13ean age (year) 12.8omozygosis for Delta F508 7/19ean time of Pseudomonas colonization (years) 5.5 (ean value of FEV1 (%) 63.57

�Group A was started on LGG and shifted to ORS, whereas group B

Data were initially analyzed by pooling the data of allchildren receiving probiotics or ORS (i.e. A1+B2 vs. A2+B1).In the analysis of pooling data we use the Unpaired T-test toevaluate the difference in FEV1 and body weight and theMann–Whitney test to compare the number of pulmonaryexacerbation and the number of hospital admissions.

Then the differences of probiotic or ORS administrationwere evaluated in parallel groups (A1 vs. B1 and B2 vs. A2).The Mann–Whitney U test was used to compare the numberof pulmonary exacerbation and the number of hospitaladmission, while the unpaired T-test was used to evaluatethe differences in FEV1 values and body weight. Finally, theanalysis was performed in longitudinal groups, i.e. for thesame population before and after shifting from LGG to ORS(A1 vs. A2) or vice versa (B1 vs. B2) and the Wilcoxon signedrank test was used to compare the number of pulmonaryexacerbation and the number of hospital admission whilethe paired T-test was used to compare the FEV1 values andthe body weight modifications.

For the non parametric procedures the results areexpressed as median and range, median difference and CI95% of the median difference , for the parametric tests theresults are expressed as mean7DS, mean difference and CI95% of the mean difference. po0.05 was consideredstatistically significant.

Results

Thirty-eight of the 43 patients enrolled, completed thestudy (16 males, mean age 13.274.2 y, range 5–23). Sixteenhad the Delta F508/Delta F508 mutation; 13 had Delta F508/other; 7 had other/other and in two the mutation wasundetermined. One of the 5 patients who did not completethe study was unable to receive probiotics because ofvomiting, and whereas the other 4 were clinically unstableat the end of probiotic or ORS administration. The totalnumber available for evaluation was 19 in each group. Thegeneral features of children in the two groups are listed inTable 1.

Figure 1 schematically depicts the cross-over study.Evaluation of outcome parameters showed the followingdata.

(a)

.�

p A

(ran

rang19

sta

Incidence and severity of pulmonary exacerbations:Overall, 116 episodes of pulmonary exacerbations wererecorded in the 12 months of observation in the 38children enrolled. As shown in Fig. 2, the incidence ofpulmonary exacerbations was significantly higher in

(n ¼ 19) Group B (n ¼ 19) p

10/9 NSge 5–18) 13.7 (6–18) NS

10/19 NSe 1–16) 6.1 (range 1–14) NS.5 45.3713.3 o0.05

rted on ORS and shifted to LGG.

ARTICLE IN PRESS

A1 B2

B1 A2

LGG LGG

ORSORS

Period 1 (six months) (4 weeks)

Period 2 (six months)

Figure 1 Cross-over design of study. Group A was started onLGG for 6 months (period 1) and then shifted to ORS (period 2);Group B was started ORS for 6 months (period 1) and then wasshifted to LGG (period 2). A washout period of 4 weeks wasintroduced between LGG and ORS.

0

1

2

3

4

5

Num

ber

of e

piso

des

LGGORS

A1 A2

All children Period 1 Period 2

B2B1A2+B1 (ORS)

A1+B2 (LGG)

p=0.003

p=0.03

p=0.02

p=0.05 p=0.02

Figure 2 Episodes of pulmonary exacerbations. Pulmonaryepisodes were significantly decreased in patients receiving LGGbased on pooled dates from patients before and after shifting(p ¼ 0.003). Data from parallel groups showed a lower numberof episodes in children treated with LGG although thedifference was significant only in period 1 (p ¼ 0.02). Long-itudinal data showed that the number of pulmonary exacerba-tions significantly decreased in group A and group B during LGGadministration (p ¼ 0.035 and 0.02, respectively).

0

1

2

3

4

5

A2

Num

ber

of h

ospi

talis

atio

ns

LGGORS

p=0.001


p=0.03

p=0.01p=0.1

p=0.01

A1+B2 (LGG)

A2+B1(ORS)

A1 B1 B2

Figure 3 Number of hospital admissions. The number ofhospital admissions was significantly decreased during LGGadministration as judged by pooled data (p ¼ 0.001). Data fromparallel groups showed a lower number of hospital admission inchildren during LGG significant only in period 1 (p ¼ 0.01).When data were analyzed in longitudinal groups a significantlower number of hospital admission in children during LGGadministration irrespectively of either when started on LGG orORS was observed (p ¼ 0.03 and 0.01, respectively).


children on ORS (71 episodes) than in those on LGGadministration (45 episodes). Data from parallel groups(A1 vs. B1 and B2 vs. A2) showed a decreased rate ofpulmonary exacerbations in children receiving LGGcompared to those receiving ORS in both periodsalthough the difference between parallel groups wassignificant only in period 2 (Median 2 vs. 1, range 4 vs. 4,median difference 1, CI 95% 0.1–2; p ¼ 0.02) (Fig. 2).Longitudinal data, as shown in Fig. 2, showed that thedifference in the number of pulmonary exacerbationsbefore and after cross-over, was significantly reducedwhen children assumed LGG in both group A (Median 1vs. 2, range 3 vs. 4, median difference 1, CI 95% 0.1–2;p ¼ 0.035) and group B (Median 2 vs. 1, range 4 vs. 4,median difference 1, CI 95% 0–2; p ¼ p ¼ 0.02).The mean duration of pulmonary exacerbations was notsignificantly reduced in children receiving LGG or ORS as

judged by the duration of antibiotic therapy. This wastrue for both parallel and longitudinal data.

(b)
Hospital admissions and length of hospital stay: Thenumber of hospital admissions because of pulmonaryexacerbations was significantly reduced, during LGGadministration (16 vs. 32 admissions during ORS). Datafrom parallel groups showed a decreased rate of hospitaladmissions in children receiving LGG versus those on ORSin both period 1 and 2 (A1 vs. B1 and A2 vs. B2), althoughthe difference was significant only in period 1 (Median 0vs. 1, range 2 vs. 2 median difference 1 CI 95% 0.1–1;p ¼ 0.01) (Fig. 3). When data were analyzed in long-itudinal groups, the difference was significant for bothgroup A and B (A1 vs. A2 median 0 vs. 1, range 2 vs. 2,median difference 0.1; CI 95% 0.1–0.5 p ¼ 0.03; B1 vs.B2 median 1 vs. 0, range 2 vs. 3, median difference 1, CI95% 0.1–1.5, p ¼ 0.01).The mean duration of hospital stay per pulmonaryepisode did not differ significantly in children receivingLGG or ORS.
(c)
Route of administration of antibiotics: Overall, therewas a trend toward fewer exacerbations requiringparenteral antibiotics in children receiving LGG,although the difference was not significant (16/45episodes were treated with parenteral antibiotics duringLGG vs. 35/71 during ORS). The ratio between parent-eral and oral antibiotics was 0.55 during LGG and 0.97during ORS.
(d)
Pulmonary function: Coupled FEV1 values at the begin-ning and the end of each treatment ere available for 29out of the 38 patients enrolled. No significant modifica-tions were observed after 6 months of ORS(51.8%716.4–52.7%716.1) in pooled data. Incontrast, the mean FEV1 value was significantly in-creased compared to baseline values after LGG

ARTICLE IN PRESS

FEV

1 in

crea

se (

%)

LGG

ORS

A1+B2 (LGG)

A2+B1 (ORS)

A1 B2B1 A2

10,00

8,00

6,00

4,00

2,00

0,00

p=0.02p= 0.1 p= 0.06

p=0.004

p=0.2


Figure 4 FEV1 modifications. Mean FEV1 was significantlydecreased in patients receiving LGG based on pooled datesfrom patients before and after shifting (p ¼ 0.02). Data fromparallel groups showed an higher FEV1 in children treated withLGG. Longitudinal data showed that the mean FEV1 significantlyincreased in group B during LGG administration (p ¼ 0.004).Conversely, mean FEV1 decreased, although not significantly, ingroup A during ORS administration.

0,00

1,00

2,00

3,00

4,00

5,00

body

wei

ght i

ncre

ase

(kg)

LGG

ORS

Period 1 Period 2All children

p = 0.02 p = 0.03

p = 0.01

p = 0.2

p = 0.2

A1+B2(LGG)

A2+B1(ORS)

A1 B2B1 A2

Figure 5 Body weight changes. Body weight increasedsignificantly in children after LGG administration as judged bypooled data (p ¼ 0.02). Data from parallel groups showed amore pronounced increase in body weight in children duringLGG than during ORS statistically significant only in period 1(p ¼ 0.03). Longitudinal data showed that the increase in bodyweight was significant in group A (p ¼ 0.01), but not in group B.


(53.8%719.2–57.4%718.7; po0.008). Individual FEV1values increased in 20 out of the 29 children that wereexamined during LGG administration. Data from parallelgroups showed an increase in FEV1 in children treatedwith LGG compared to children treated with ORS in bothperiod 1 and period 2, although the differences were notsignificant (Fig. 4). When we separately analyzed theresults for longitudinal groups, the mean delta value ofFEV1 was significantly increased, during LGG adminis-tration, in the group of patients started on ORS and thanshifted to LGG (0.4774.41 vs. 4.474.84, mean differ-ence 3.93, CI 1.13–6.64; p ¼ 0.004), whereas thedifference was not significant for children started onLGG and then shifted on ORS (Fig. 4).

(e)
Body weight and BMI modifications: The pattern of bodyweight modifications was significantly different in LGGand ORS, with a greater weight gain after LGG than afterORS. When data were analyzed for parallel groups therewas an increase in weight gain in children treated withLGG compared to ORS in both period 1 and 2 (A1 vs. B1and A2 vs. B2) (Fig. 5) although the difference wassignificant only in period 1 (1.98 kg71.93 vs.0.78 kg71.97; mean difference 1.2, CI 95% 0.08–2.48;p ¼ 0.03). The analysis of modifications of body weightfor longitudinal groups showed a significant increase inbody weight only in group A (1.98 kg71.93 vs.0.6171.65, mean difference 1.37, CI 95% 0.19–2.56;p ¼ 0.01) , but not in group B (Fig. 5).No significant differences in BMI were observed inparallel and longitudinal groups.
(f)
Immunoglobulin concentrations: Coupled Ig concentra-tions were available for 27 of the 38 children. The meanserum IgG concentration increased significantly duringORS administration (from 1512+438 to 1633+424mg/dl;p ¼ 0.01), whereas it remained stable during LGGadministration (from 1587+427 to 1508+384mg/dl).
There were no differences in serum concentrations ofIgA and IgM during LGG compared to ORS.Side-effects of LGG administration. There were noevident side effects (beside vomiting in one child) thatled to LGG withdrawal early after enrollment. Overallacceptance of the probiotic was good, with thecomplaint being that another ‘‘drug’’ had to be takenfor a long time.

Discussion

This study shows that LGG exerts a protective effect againstpulmonary exacerbations in CF children colonized withPseudomonas, as judged by all the parameters included inthe analysis of pooled data. There were no changes inpatients’ treatment that could explain the observed reduc-tion in pulmonary exacerbations. Moreover, the samechildren served as both patients and controls therebyovercoming inter-patient variability.

The outcome parameters from pooled data of childrenreceiving LGG or ORS supported the efficacy of probioticsand comparative analysis of data before and after cross-overprovided further information. Children receiving LGG vs.those receiving ORS in parallel or vs. themselves aftershifting (in longitudinal analysis) consistently had a betteroutcome. The patients in group B, who first received ORSonly, actually started the second period with the samefeatures found at the baseline as judged by longitudinalanalysis and they were found to have a decreased incidenceof pulmonary exacerbation and of the others main outcomesat the end of the second period. These data add to theefficacy of probiotic administration. The parallel databefore and after crossing over showed a reduced numberof pulmonary exacerbations during LGG and a decrease inhospital admissions.

However the difference tended to be less evident for dataobtained from parallel groups of children after crossing over.

ARTICLE IN PRESS


The pattern of secondary outcomes supported a protectiveeffect of probiotics on pulmonary exacerbations.

We observed an increase in predicted FEV1 during LGGthat suggested that the reduction of exacerbation isreflected by change in pulmonary function in the patientsenrolled in whom data were available. Although the twopopulation were different in term of FEV1 this parameterimproved in the two group as judged by longitudinalanalysis. This suggests that the beneficial effects by theprobiotic are not related to the baseline features. Howeverwe feel that the functional data are not as strong as theincidence data but they add to the former and are well inline with immunologic (IgG ) results.

Given the close correlation between pulmonary functionand nutritional status in patients with CF18 the observedgreater increase of body weight in children receiving LGGmay well be a direct consequence of the decreased rate ofpulmonary exacerbations.

The hypothesis that probiotics exert an anti-infectiouseffect in sites other than intestine has been suggested bypreliminary data10,20 and begs the problem of the mechan-ism of the effect. A direct effect against Pseudomonasmay be considered. LGG administration reduced Pseudomo-nas bacteremia in irradiated mice and Lactobacillus caseiincreased the lung clearance of Pseudomonas in youngmice.21 In addition, Lactobacillus plantarum inhibitedthe pathogenic activity of Pseudomonas.22 However, in noneof our children Pseudomonas was eradicated (data notshown).

Several lines of evidence suggest that probiotics affectthe host immune/inflammatory response, as a consequenceof modifications of intestinal microflora.23,24 A recent studysuggested that the addition of Lactobacillus casei accel-erates the recovery of the innate immune response andimproves the specific immune mechanisms against arespiratory infection in malnourished mice induce byStreptococcus pneumoniae.25

Intestinal epithelial cells and the Gut-Associated-Lym-phoid-Tissue (GALT) are considered key players in driving theappropriate host’s adaptation to environmental chal-lenges.23 Several reports indicate that disruption of host/microbiota balance, by means of alteration of microbiotacomposition, leads to the disruption of mucosal hypo-responsiveness, potentially interfering with the maturationof dendritic cells that promote antigen-specific regulatory T-cell response.26 Lactobacilli have a number of effects on theintestinal immune response. They can downregulate den-dritic cell maturation resulting in the in vitro induction ofregulatory T-cells27, control the proinflammatory/antiapop-totic IkB/NFkB pathway, inhibit the constitutive geneexpression and production of IL-8 by epithelial cells,increase mucin expression and strengthen the intestinalepithelial barrier in vivo.27,28 In particular, the disruption ofthe intestinal epithelial barrier is central to the pathogen-esis of several inflammatory diseases. Interestingly, anincrease in intestinal permeability has been reported inatopic dermatitis and IDDM27,28 as well as in CF.29 Thesefindings suggest that probiotics may contribute in severalways to the first line host defence to environmentalchallenges.

CF is the prototype of chronic inflammatory diseasescharacterized by a dysregulation of the first-line host

defence to environmental challenges. In CF airways an‘intrinsic’ aberrant inflammatory response, apparentlyrelated to the activation of NF-kappaB or to a deficit ofthe lipoxin pathways, leads to an overt mucosal inflamma-tion upon bacterial challenge. In CF airways there is a lowtreshoold response to environmental triggers as shown by exvivo culture experiments in which a very early and dramaticinflammatory response is induced in CF nasal polyp mucosaeexposed to P. aeruginosa LPS.30 A similar behavior char-acterizes the CF small intestine where the role of dietaryantigens in sustaining mucosal inflammation has beendemonstrated.4 The control of such inappropriate inflam-matory response to environmental triggers, might be themain target of therapeutic strategies. Probiotics may bewell considered within this view.

The heavy use of antibiotics and the perturbation ofmucosal microenvironment due to CFTR dysfunction maylead to a disruption of intestinal microecology. Weperformed a qualitative control of intestinal microflora ina small number of children receiving LGG, using thermalgradient electrophoresis, and found substantial differencescompared to baseline pattern (data not shown). In addition,serum IgG concentrations were decreased in CF children onLGG. Since the increase in immunoglobulin concentrationsreflects systemic inflammation in CF,19 it is conceivable thatLGG administration exerts a general anti-inflammatoryeffect in this patients. We have previously reported thatLGG downregulates intestinal inflammation in CF children.5

Since the gastrointestinal tract may in fact be exposed toany antigen that the respiratory tract is also exposed too,the intestinal immune system might act as a sensor even forinhaled antigens with generation of T regulatory cells thatmay colonize sites of inflammation outside the gut.25

Targeting the small intestine to achieve antiinflamatoryeffects might therefore represent a new therapeutic for CFpatients.

In conclusion, this study shows that long-term adminis-tration of L. casei strain GG to children with CF colonizedwith P. aeruginosa significantly decreases the incidence ofpulmonary exacerbations and increases body weight, inassociation with a decrease of IgG concentrations. Whateverthe mechanism, this work provides a novel strategy withwhich to prevent pulmonary exacerbations in patients withCF.

Acknowledgments

The authors thankfully acknowledge Jean Gilder for editor-ial assistance. The study was supported in part by a grantfrom Fondazione per la ricerca sulla fibrosi cistica. Thespecific contribution of each author was the following:

Eugenia Bruzzese contributed to the study design, writingof manuscript and data analysis.

Valeria Raia is chair of the reference center for cysticfibrosis and enrolled the patients.

Maria Immacolata Spagnuolo and Monica Volpicelli MDperformed all clinical investigation.

Giulio De Marco contributed to data analysis.Luigi Maiuri, contributed to critical review the manu-

script.

ARTICLE IN PRESS


Alfredo Guarino contributed to the study design, writingof manuscript, data analysis and overall coordinatingactivities.

No conflict of interest exists for any author.

References

1. Ramsey BW. Management of pulmonary disease in patients withcystic fibrosis. N Engl J Med 1996;335:179–88.

2. Kerem E, Corey M, Gold R, Levison H. Pulmonary function andclinical course in patients with cystic fibrosis after pulmonarycolonization with Pseudomonas aeruginosa. J Pediatr 1990;116:714–9.

3. Wilmott RW, Tyson SL, Matthew DJ. Cystic fibrosis survival rates.The Influences of allergy and Pseudomonas aeruginosa. Am J DisChild 1985;139:669–71.

4. Raia V, Maiuri L, de Ritis G, et al. Evidence of chronicinflammation in morphologically normal small intestine ofcystic fibrosis patients. Pediatr Res 2000;47:344–50.

5. Bruzzese E, Raia V, Gaudiello G, et al. Intestinal inflammationis a frequent feature of cystic fibrosis and is reduced byprobiotic administration. Aliment Pharmacol Ther 2004;20:813–9.

6. Allen SJ, Okoko B, Martinez E, Gregorio G, Dans LF. Probioticsfor treating infectious diarrhoea. Cochrane Database Syst Rev2004(2):CD003048.

7. Rembacken BJ, Snelling AM, Hawkey PM, Chalmers DM,Axon ATR. Non-pathogenic E. coli vs. mesalazine for thetreatment of ulcerative colitis: a randomised trial. Lancet 1999;354:635–9.

8. Isolauri E, Arvola T, Sutas Y, Moilanen E, Salminen S. Probioticsin the management of atopic eczema. Clin Exp Allergy2000;30:1604–10.

9. Kalliomaki M, Salminen S, Poussa T, Arvilommi H, Isolauri E.Probiotics and prevention of atopic disease: 4-years follow-upof a randomised placebo-controlled trial. Lancet 2003;361:1869–71.

10. Hatakka K, Savilahti E, Ponka A, et al. Effect of long termconsumption of probiotic milk on infection in children attendingday care centres: double blind, randomised trial. Br Med J2001;322:1327.

11. Isolauri E, Majamaa H, Arvola T, Rantala I, Virtanen E,Arvilommi H. Lactobacillus casei strain GG reverses increasedintestinal permeability induced by cow milk in suckling rats.Gastroenterology 1993;105:1643–50.

12. Perdigon G, Fuller R, Raya R. Lactic acid bacteria and theireffect on the immune system. Curr Issues Interest Microbiol2001;1:27–42.

13. Di Caro S, Tao H, Grillo A, et al. Effects of Lactobacillus GG ongenes expression pattern in small bowel mucosa. Dig Liver Dis2005;37:320–9.

14. Van Elburg RM, Uil JJ, Van Aalderen WMC, Mulder CJJ, HeymansHSA. Intestinal permeability in exocrine pancreatic insuffi-ciency due to cystic fibrosis or chronic pancreatitis. Pediatr Res1996;39:985–91.

15. Claeys S, Van Hoecke H, Holtappels G, et al. Nasal polyps inpatients with and without cystic fibrosis: a differentiation by

innate markers and inflammatory mediators. Clin Exp Allergy2005;35:467–72.

16. Cystic Fibrosis Foundation. Mycrobiology and infectious diseasein Cystic Fibrosis. Consensus Conferences: Concept in care.Bethesda, 1994, Vol V, sect 1.

17. Hoiby N. Prospects for the prevention and control of pseudo-monal infection in children with cystic fibrosis. Pediatr Drugs2000;2:451–63.

18. Corey M, McLaughlin FJ, Williams M, Levison H. A comparison ofsurvival, growth, and pulmonary function in patients with cysticfibrosis in Boston and Toronto. J Clin Epidemiol 1988;41:583–91.

19. De Vizia B, Raia V, Spano C, Pavlidis C, Coruzzo A, Alessio M.Effect of an 8-month treatment with omega-3 fatty acids(eicosapentaenoic and docosahexaenoic) in patients with cysticfibrosis. J Parenter Enteral Nutr 2003;27:52–7.

20. Elmer GW, Surawicz CM, Mc Farland LV. Biotherapeutic agents. Aneglected modality for the treatment and prevention ofselected intestinal and vaginal infections. J Am Med Assoc 1996;275:870–6.

21. Alvarez S, Herrero C, Bru E, Perdigon G. Effect of Lactobacilluscasei and yogurt administration on prevention of Pseudomonasaeruginosa infection in young mice. J Food Prot 2001;11:1768–74.

22. Valdez JC, Peral MC, Rachid M, Santana M, Perdigon G.Interference of Lactobacillus plantarum with Pseudomonasaeruginosa in vitro and in infected burns: the potential use ofprobiotics in wound treatment. Clin Microbiol Infect2005;11:472–9.

23. MacDonald TT, Monteleone G. Immunity, Inflammation andallergy in the gut. Science 2005;307:1920–5.

24. Noverr MC, Huffnagle GB. Does the microbiota regulate immuneresponses outside the gut? Trends Microbiol 2004;12:562–8.

25. Villena J, Racedo S, Aguero G, Bru E, Medina M, Alvarez S.Lactobacillus casei improves resistance to pneumococcalrespiratory infection in malnourished mice. J Nutr2005;135:1462–9.

26. Christensen HR, Frokiaer H, Pestka JJ. Lactobacilli differen-tially modulate expression of cytokines and maturationsurface markers in murine dendritic cells. J Immunol 2002;168:171–8.

27. Kalliomaki M, Walker AD. Physiologic and pathologic interac-tions of bacteria with gastrointestinal epithelium. In: Gastro-enterology clinics of North America: probiotics, prebiotics andcommensal bacteria: perspectives and clinical applications ingastroenterology, vol. 34. Philadelphia: Elsevier; 2005. p.383–99.

28. Isolauri E, Salminen S. Probiotics, gut inflammation and barrierfunction. In: Gastroenterology clinics of North America:probiotics, prebiotics and commensal bacteria: perspectivesand clinical applications in gastroenterology, vol. 34. Philadel-phia: Elsevier; 2005. p. 437–50.

29. Troncone R, Santamaria F, Ercolini P, Raia V, Panza G, de Ritis G.Increased serum antibody levels to dietary antigens in cysticfibrosis. Acta Pediatr 1994;83:440–1.

30. Raia V, Maiuri L, Ciacci C, et al. Inhibition of p38 mitogenactivated protein kinase controls airway inflammation in cysticfibrosis. Thorax 2005;60:773–80.