Archives ofDisease in Childhood 1992; 67: 940-943

Management and response to treatment ofHelicobacter pylon' gastritis

M J Mahony, J I Wyatt, J M Littlewood

AbstractGastritis associated with Helicobacter pylonwas present in gastric biopsies from 24/95(25%) children and adolescents undergoingendoscopy for recurrent abdominal pain andupper gastrointestinal symptoms. H pylonassociated gastritis occurred mainly in olderchildren (8-16 years) and was significantlyassociated with low socioeconomic class anda family history of peptic ulcer disease. Antralnodularity was a common endoscopic findingin H pylon positive children. Eighteen child-ren, all over 5 years of age, were treated withtripotassium dicitratobismuthate (De-Nol) fortwo months and ampicillin for two weeks. In12 children follow up gastric biopsies wereobtained six weeks after completion of treat-ment. In 9/12 (75%) children H pylori waseradicated, and gastritis improved.

Colonisation of the stomach with Helicobacter(formerly Campylobacter) pylorn is now firmlyestablished as an important cause of chronicgastritis and peptic ulcer disease in adults.' 2Several studies have also established a role forthis organism in children with abdominal painand upper gastrointestinal symptoms.?5 Weprospectively evaluated the prevalence ofH pylori gastritis in children undergoing endo-scopy for abdominal pain and upper gastro-intestinal symptoms, and report here the clinical,socioenconomic, endoscopic and histopatho-logical features, and the response to treatment.The systemic immune response to H pyloni wasalso studied in these children and has beenreported separately.6

St James'sUniversity Hospital,LeedsDepartment ofPaediatricsM J MahonyJ M LittlewoodDepartment of PathologyJ I WyattCorrespondence to:Dr M J Mahony,Paediatric Unit,Limerick Regional Hospital,Dooradoyle,Limerick,Republic of Ireland.Accepted I March 1992

MethodsA prospective study was performed betweenJuly 1987 and July 1991 on consecutive childrenpresenting with recurrent abdominal pain, pro-tracted vomiting, haematemesis, and melaena.Patients with recurrent abdominal pain andprotracted vomiting had symptoms of at leastthree months' duration. Ninety five children(51 females), age range 6 months to 16 years,median 10 years, were investigated by uppergastrointestinal endoscopy and mucosal biopsy.Endoscopy was performed using Olympus GIFP3, Olympus XP20, and Fujinon UGI PEpaediatric endoscopes by two paediatricianendoscopists (MJM, JML). Intravenous sedationusing diazepam 0-5-1 mg/kg, maximum 25 mg,was used in 93 children; two children hadendoscopy under general anaesthesia. At leasttwo antral biopsies were taken from each child,

and in 55 an additional antral biopsy specimenwas taken for gastric urease (CLO test, DeltaWest).

Biopsy specimens were fixed in neutralbuffered formalin and stained with haematoxylinand eosin for grading of gastritis and Giemsa fordetection of H pylorn.7 All specimens wereexamined by one histopathogist (JIW). Gastritiswas graded using the Sydney classification:briefly, this grades the degree of mononuclearcell infiltration, neutrophil infiltration, atrophy,and intestinal metaplasia in gastric biopsy speci-mens, each on a scale of 0-3. In addition theintensity ofH pylori colonisation is assessed alsoon a scale of 0o3.8Treatment with tripotassium dicitratobismu-

thate (De-Nol, Brocades) and ampicillin wasgiven to children over 5 years of age withH pylori associated gastritis. Children with'lymphocytic gastritis' were excluded. Eighteenchildren with H pylon associated gastritis,aged 5-15, median 12-5 years, were treatedwith tripotassium dicitratobismuthate for twomonths, and ampicillin for the first two weeksof treatment. Tripotassium dicitratobismuthatedosage was 240 mg twice daily in those over 10years, and 120 mg twice daily in those under 10years. The ampicillin dose was 500 mg fourtimes a day in those over 10, and 250 mg fourtimes a day in those under 10 years. Follow upantral biopsies were obtained six weeks aftercompletion of treatment in 12/18 children andwere assessed for eradication of H pylori andseverity of gastritis. For the purpose of compar-ing biopsy specimens taken before and aftertreatment the grades of mononuclear cell infil-tration (0-3) and neutrophil infiltration (0-3)were added to devise a gastritis score (range0-6).Bismuth concentrations were measured in

whole blood at completion of tripotassiumdicitratobismuthate treatment by hydridegeneration atomic absorption spectroscopy(Rooney Laboratories).9

Informed consent was obtained from thechildren's parents, and the study was approvedby the hospital ethics committee of Leeds EastHealth Authority.

Statistical analysis was performed usingFisher's exact probability test, and Wilcoxon'ssigned pair test.

ResultsGastritis was present in antral biopsies from25/95 children (26%). In 24 children (25%), agerange 9 months to 15 years, median 12-5 years,chronic gastritis was associated with H pylori

940

on January 30, 2020 by guest. Protected by copyright.

http://adc.bmj.com

/A

rch Dis C

hild: first published as 10.1136/adc.67.7.940 on 1 July 1992. Dow

nloaded from

Management and response to treatment ofHelicobacter pylori gastritis

40- L H pylori

negative

H pylori301- 1 positive

c

._'a

0 20-0

0

z

10

0-4 48 8-12 12-16Age (years)

Figure I Age distribution ofthe study population.

colonisation (23 histology positive; one CLOpositive, histology negative). Gastric histologywas normal in the other 70 children and H pyloriwas not detected on Giemsa staining or CLOtest.The CLO test was performed on 55 antral

biopsies and was positive in 13. All 13 biopsieshad gastritis on histology and 12/13 werepositive for H pylorn on Giemsa staining. Theother 42 biopsies showed normal gastric histo-logy and all were negative for H pylori onhistology.

CLINICAL FEATURESThe age range of the children studied and thosewith H pyloni associated gastritis is shown infigure 1. H pylon gastritis was present mainly inolder children and adolescents aged 8-16 years,but was present in five young children aged9 months, 15 months, 2 years, and two aged5 years. The sex distribution ofH pylori gastritiswas 10 boys and 14 girls. A family history ofpeptic ulcer disease in first degree relatives waspresent in 6/24 (25%) H pylori positive childrenwith gastritis compared with 2/70 (3%) of thosewho wereH pyloin negative, with normal gastrichistology (p<0 003).

Although several children had multiplesymptoms, the major symptom leading to endo-scopy was recurrent abdominal pain. This wasfurther subdivided into epigastric pain and

Social class distribution of H pylori positive and H pylori negative children

Non-manual classes Manual classes

1 2 3NM 3M 4 5

H pylori positive 0 0 3 6 7 8

3 (12 5%) 21 (87 51%)*

H pylori negative 3 18 6 26 8 10

27 (38%) 44 (62%)

Study population 3 18 9 32 15 18(3 1%) (19%) (9.5%) (33 7"%) (15-81%) (18-9%))

30 (31-6%) 65 (68 41%)

West Yorkshire 5% 23% 12% 36% 18% 6%(1981 census)

400/o 60%)

*Significant difference between positive and negative groups (p<O002).

central periumbilical pain. Epigastric pain wascommoner in children with H pylon gastritis,occurring in 13/24 (54%), compared with 28/70(40%) of those with normal histology, althoughthis difference was not statistically significant(p=0086). Central periumbilical abdominalpain was more common in H pylori negativechildren with normal gastric histology, 27/70(39%), compared with 5/24 (20%) of those withH pylori positive gastritis, but this differencedid not reach statistical significance (p=0069).Recurrent vomiting was the second most fre-quent symptom but was poorly discriminatorybetween the two groups, being present in 4/24(17)% and 12/70 (17%) patients with andwithout H pylon gastritis respectively. Otherpresenting symptoms in H pylon positivechildren were haematemesis (one child), andmelaena (one child). One child with H pylongastritis had refractory iron deficiency anaemia.

SOCIAL CLASSThe social class of the children studied wasassigned according to the occupation of the headof the household using the registrar general'sclassification, and is shown in the table. Forpurposes of comparison, non-manual classes 1,2, and3NMwere grouped together andcomparedwith manual classes 3M, 4, and 5. The socialclass distribution of the study population wasbroadly similar to that of the population of WestYorkshire,'° but there was a marked social classdifference between H pylon positive and nega-tive children. Children of parents with man-ual occupations were overrepresented in theH pylon positive gastritis group 21/24 (87 5%)compared with 44/71 (62%) in the H pylonnegative group (p<002), and no H pylon'positive child came from social class 1 or 2.

ENDOSCOPYEndoscopic abnormalities were present in 20/24children withH pylon gastritis. The commonestabnormality was antral nodularity (fig 2) foundin 13/24; five children had antral erythema andgastric erosions were present in two. Duodenalulcers were diagnosed in two children (aged 15and 13 years): one was associated with gastricerosions and coexistent H pylon gastritis onantral histology, the other was an isolatedfinding and was not associated with H pylongastritis. The principal endoscopic finding inthe H pylon negative, normal gastric histologygroup was macroscopic oesophagitis, present in14/70, and microscopic oesophagitis was foundin a further eight on oesophageal biopsy.

HISTOLOGYOf the 25 children with chronic gastritis, 22 hadtypical helicobacter associated chronic gastritison histology. Three children showed the patternof 'lymphocytic gastritis'."1: one of these hadhelicobacter on histological staining, and onewas positive on CLO test, while all three hadpositive H pylon serology.6 Thus 3/25 (12%) ofchildren with chronic gastritis showed thisspecial type of gastritis, a frequency higher than

"I".

941

7//7/-

on January 30, 2020 by guest. Protected by copyright.

http://adc.bmj.com

/A

rch Dis C

hild: first published as 10.1136/adc.67.7.940 on 1 July 1992. Dow

nloaded from

Mahony, Wyatt, Littlewood

0)

00

C.

._0

H pylorieradicated

Before Aftertreatment treatment

Figure 2 Antral nodularity seen at endoscopy in a childwith H pylori gastritis.

that seen in adults, where it is present in 2% ofpatients with chronic gastritis." None of thebiopsy specimens showed reactive gastritis or anyof the other special forms of gastritis recognisedby the Sydney system.The 22 children with helicobacter associated

chronic gastritis (excluding cases of lymphocyticgastritis) were graded according to the Sydneysystem. All showed some increase in chronicinflammatory cells, and lymphoid follicles werepresent in biopsies from 12/22 (55%) children(fig 3). The presence of 'activity' of chronicgastritis is assessed by the degree of neutrophilinfiltration of the epithelium. This was presentin specimens from 10/22 (45%) children only,although it is characteristically present in adultswith helicobacter gastritis. 12None of the biopsy specimens showed archi-

tectural irregularity to suggest there had beenany atrophy of antral glands, and none of themshowed intestinal metaplasia.

RESULTS OF TREATMENTIn the 12 children in whom antral biopsyspecimens before and after treatment were

studied, H pylori eradication was achieved in

Figure 3 Lymphoidfollicle in the antral mucosa ofa childwith H pylori gastritis.

H pyloripersistent

Before Aftertreatment treatment

Figure 4 Severity ofgastritis in antral biopsy specimensbefore and after treatment, showing significant reduction(p<0 025) in gastritis score where H pylori was eradicated.

9/12 (75%). Gastritis scores improved in theH pylon eradication group from mean (SD) of2-55 (1-13) before treatment to 1 (0-5) aftertreatment (p<0025) (fig 4). Symptoms resolvedin seven children but were unchanged in twodespite eradication ofH pylori and improvementin gastritis score. In the three children in whomH pylori was not eradicated, gastritis score wasnot significantly altered, and all remainedsymptomatic. Six other children had identicaltreatment and all experienced an improvementin symptoms, but follow up endoscopy wasdeclined by their parents. The 15 year oldadolescent with duodenal ulcer and coexistentH pylori gastritis was treated with tripotassiumdicitratobismuthate and ampicillin at dosesmentioned above, and ranitidine 150 mg twicedaily. Follow up endoscopy 11 months afterpresentation showed healing of the ulcer butmucosal biopsies were not obtained.No child developed side effects from treat-

ment. Bismuth concentrations ranged from 3 to29 ng/ml, mean 14-8 ng/ml. Bismuth toxity isassociated with concentrations above 50 ng/ml. 13

DiscussionThis study confirms the findings of otherworkers that H pylori gastritis is a significantcause of abdominal pain in the paediatric agegroup.?5 The clinical profile of paediatricH pylori gastritis that emerges from this study isof epigastric pain in the older child and adole-scent (8-16 years) from a family of low socio-ecomonic status and a family history of pepticulcer disease. However, we have found thatH pylon gastritis can occur in young children,and 20% of our cases were in children aged5 years or younger. The bias towards lowsocioeconomic groups in H pylori positivechildren suggests that social class is a factor inH pylori acquisition in childhood. Childrenfrom higher social class groups appear to be lesssusceptible to H pylori infection, and similarresults have been found in adults in the UnitedKingdom. 14 Although the reservoir of infectionand modes of spread ofH pyloni are not known,the social patterns are suggestive of faecal-oraltransmission, and social class differences mayreflect differences in sanitation and hygienicpractices. 15

942

""I

on January 30, 2020 by guest. Protected by copyright.

http://adc.bmj.com

/A

rch Dis C

hild: first published as 10.1136/adc.67.7.940 on 1 July 1992. Dow

nloaded from

Management and response to treatment ofHelicobacter pylori gastritis 943

Positive endoscopic findings are common inchildren with H pylon associated gastritis,especially antral nodularity. Lymphoid follicleswere present in biopsies from 55% children; thepresence of follicles was not correlated with theactivity of the gastritis, nor the density ofbacterial colonisation. As the antral mucosa inchildren is thinner than that in adults, it is likelythat the presence of large lymphoid folliclesaccounted for the nodularity apparent on endo-scopy.

Lymphocytic gastritis is an unusual patternof inflammation in adults, but was seen rathermore frequently in our paediatric biopsyspecimens (3/25). As in adults, these childrenshowed serological evidence ofHpylori infection,although the bacteria were apparent on histologyin only one of three children. The relationshipbetween lymphocytic gastritis and H pyloninfection is still uncertain.The precise indications for treatment and the

optimal treatment regimen for 'children withH pyloni gastritis remain to be established. 6 Wechoose treatment with tripotassium dicitrato-bismuthate and ampicillin as this combinationgave optimal results in a large adult study. 17 Weachieved a H pylon eradication rate of 75% andthis is similar to other paediatric studies using acombination of ampicillin and bismuth pre-parations.5 18 In adult studies better eradicationrates have been achieved with dual therapyusing a bismuth preparation and metronidazole,but there are problems with metronidazoleresistance. 9 Blood bismuth concentrations weremeasured because of concern about potentialtoxicity of bismuth preparations in children,and all were in the non-toxic range. Tripotassiumdicitratobismuthate was well tolerated and nochild developed side effects.The discovery of H pylorn in 1983, and the

subsequent demonstration of its role in gastro-duodenal pathology, has been one the mostexciting developments in gastroenterology inthe past decade. Chronic abdominal pain inchildren has always been an important part ofthe paediatrician's workload.20 This study andothers suggest a role for H pylori in childrenwith chronic abdominal pain and upper gas-trointestinal symptoms.5 Although relativelyuncommon in children, a strong associationexists between H pylon gastritis and duodenalulceration. Furthermore, eradication ofH pylorihas been shown to improve ulcer healing andto reduce ulcer relapse rate in children.2' Thenatural history ofH pylon gastritis is still beingdelineated, but present evidence suggests a slowprogression to atrophic gastritis and gastricatrophy over a period of 20-40 years. Atrophicgastritis and gastric atrophy are known to beprecursor lesions for the development of gastriccarcinoma,22 and this potential carcinogenicityof H pylori makes it a significant pathogen inchildren.Treatment with bismuth preparations and

antibiotics can eradicate H pylori and improve

gastritis in about 75% of patients, but long termfollow up and control studies are lacking.Recent reviews have sounded a note of cautionabout the indiscriminate treatment of H pylornespecially the risk of antibiotic resistance, bothin children'6 and in adults.'9 Serological detec-tion of H pylori IgG antibodies is valuable inscreening children with chronic abdominal painfor H pylorn gastritis,6 and perhaps endoscopyshould be reserved for detection of associatedduodenal ulcer disease. Treatment to eradicateH pylori would then be justified in view of therole of H pylon in duodenal ulcer healing andrelapse.2' Further research remains to be donebefore we fully understand the precise role ofH pylori in children with chronic gastrointestinalsymptoms, the indications for treatment, andthe optimal agents.

1 Marshall BJ, Warren JR. Unidentified curved bacilli in thestomach of patients with gastritis and peptic ulceration.Lancet 1983;i:1273-5.

2 Dixon MF. Campylobacter pylori and chronic gastritis. In:Rathbone BJ, Heatley RV, eds. Campylobacter pylori andgastroduodenal disease. Oxford: Blackwell Scientific Publi-cations, 1989:106-16.

3 Drumm B, Sherman P, Cutz E, Karmali M. Association ofCampylobacter pylori on the gastric mucosa with antralgastritis in children. N EnglJ' Med 1987;316:1557-61.

4 Glassman MS, Schwarz SM, Medow MS, et al. Campylobacterpylori-related gastrointestinal disease in children: incidenceand clinical findings. Dig Dis Sci 1989;34:1501-4.

5 De Giacomo C, Fiocca R, Villani L, et al. Helicobacter pyloriinfection and chronic gastritis: clinical, serological, andhistological correlations in children treated with amoxicillinand colloidal bismuth subcitrate. J Pediatr GastroenterolNutr 1990;11:310-6.

6 Crabtree JE, Mahony MJ, Taylor JD, Heatley RV, LittlewoodJM, Tomkins DS. Immune responses to Helicobacterpylori in children with recurrent abdominal pain. J ClinPathol 1991;44:768-71.

7 Gray SF, Wyatt JI, Rathbone BJ. Simplified techniques foridentifying Campylobacter pyloridis. J Clin Pathol 1986;39:1279-80.

8 Price AB. The Sydney system: histological division. journalof Gastroenterology and Hepatology 1991;6:209-22.

9 Rooney RC. Determination of bismuth in blood and urine.Analyst 1976;101:749-52.

10 Office of Population Censuses and Surveys. West Yorkshirecounty census report. London: HMSO, 1981.

11 Dixon MF, Wyatt JI, Burke DA, Rathbone BJ. Lymphocyticgastritis-relationship to Campylobacter pylori infection.J7Pathol 1988;154:125-32.

12 Bayerdorffer E, Oertel H, Lehn N, et al. Topographic as-sociation between active gastritis and Campylobacter pyloricolonisation. J Clin Pathol 1989;42:834-9.

13 Bader JP. The safety profile of De-Nol. Digestion 1987;37suppl 2:53-9.

14 Sitas F, Forman D, Yarnell JWG, et al. Helicobacter pyloriinfection rates in relation to age and social class in a popula-tion of Welsh men. Gut 1991;32:25-8.

15 Graham DY. Helicobacter pylori: its epidemiology and itsrole in duodenal ulcer disease. J'ournal of Gastroenterologyand Hepatology 1991;6:105-13.

16 Drumm B. Helicobacter pylori. Arch Dis Child 1990;65:1278-82.

17 Rauws EAJ, Langenburg W, Houthoff HJ, Zanen HC,Tytgat GNJ. Campylobacter pyloridis-associated chronicactive antral gastritis: a prospective study of its prevalenceand the effects of antibacterial and antiulcer treatment.Gastroenterology 1988;94:33-40.

18 Drumm B, Sherman P, Chiasson D, Karmali M, Cutz E.Treatment of Campylobacter pylori-associated antral gas-tritis in children with bismuth subsalicylate and ampicillin..7 Plediatr 1988;113:908-12.

19 Axon ATR. Helicobacter pylori therapy: effect on pepticulcer disease. Journa of Gastroenterology and Hepatology1991;6:131-7.

20 Apley J. 7he child with abdominal pains. 2nd Ed. Oxford:Blackwell Scientific Publications. 1975.

21 Yeung CK, Fu KH, Yuen KY, et al. Helicobacter pylori andassociated duodenal ulcer. Arch Dis Child 1990;65:1212-6.

22 Correa P. Is gastric carcinoma an infectious disease? N Engl7Med 1991;325:1170-1.

on January 30, 2020 by guest. Protected by copyright.

http://adc.bmj.com

/A

rch Dis C

hild: first published as 10.1136/adc.67.7.940 on 1 July 1992. Dow

nloaded from