Safety and immunogenicity of a glycoprotein D genital herpes vaccine in healthy girls 10-17 years of age: results from a randomised, controlled, double-blind trial.

OBJECTIVE: The investigational AS04-adjuvanted herpes simplex virus type 2 (HSV-2) glycoprotein D (gD2) subunit prophylactic vaccine ('HSV vaccine'; GlaxoSmithKline Vaccines) has been shown to be well tolerated in adults, but limited data exist for pre-teen and adolescent girls, a likely target population. The primary objective of this study was to compare the occurrence of serious adverse events (SAEs) over 12 months between HSV vaccine recipients and saline recipients (placebo control group) in pre-teen and adolescent girls. The immunogenicity of the HSV vaccine was also assessed. METHODS: Healthy girls aged 10-17 years, stratified by age (10-15 years; 16-17 years), were randomised 2:1:1 to receive the HSV vaccine, a hepatitis A vaccine (Havrix™; HAV control) or placebo (saline) according to a 0-, 1-, 6-month schedule. Participants and study personnel not involved in the preparation or administration of vaccines were blinded to treatment. Safety and immunogenicity analyses were performed overall and by age (10-15 years; 16-17 years) and HSV serostatus. RESULTS: No statistically significant difference in the percentage of subjects with SAEs was observed between the HSV and saline group, or between the HSV and pooled control (HAV and saline) groups. The HSV vaccine was well tolerated, although a higher incidence of solicited local symptoms was observed in the HSV group than in the control group. Neither age nor HSV serostatus at the time of study entry had an impact on the safety profile of this vaccine. The HSV vaccine was immunogenic regardless of pre-vaccination HSV serostatus. Higher anti-gD geometric mean concentrations were observed in HSV-1 seropositive participants than in HSV-1 seronegative participants. CONCLUSION: The HSV vaccine had an acceptable safety profile, and was well tolerated and immunogenic when administered to girls aged 10-17 years regardless of age or HSV pre-vaccination serostatus.

Pathogenicity and convalescent excretion of Campylobacter in rural Egyptian children.

Campylobacter infection in developing countries has not received much public health attention because of the observation that infections are not associated with disease beyond the first 6 months of life. A cohort of 397 Egyptian children aged less than 3 years, who were observed twice weekly during 1995--1998, experienced an incidence of 0.6 episodes of Campylobacter diarrhea per child-year. A total of 13% of the Campylobacter diarrheal episodes were characterized by severe dehydration. Age-specific incidence rates (episodes per year) were 0.9 in infants aged less than 6 months, 1.5 in those 6--12 months, and 0.4 and 0.2 in the second and third years of life, respectively. Convalescent excretion of Campylobacter after a diarrheal episode might be enhancing transmission and contributing to this high incidence. Observed risk factors for Campylobacter diarrhea were poor hygienic conditions and the presence of animals in the house. Regardless of the child's age, a first infection by Campylobacter was associated with diarrhea (odds ratio = 2.45; 95% confidence interval: 1.61, 3.71); however, subsequent infections were associated with diarrhea only in children aged less than 6 months. This observation that natural infection did not confer protection during the first 6 months of life poses a challenge to vaccine development.

Seroepidemiology of Helicobacter pylori infection in a population of Egyptian children.

BACKGROUND: To describe the seroepidemiology of Helicobacter pylori infection in a population of Egyptian children under 3 years. METHODS: A cohort of children under 36 months, residing in Abu Homos, Egypt, were visited at home twice weekly. Information regarding the child's breastfeeding status was obtained, and periodic anthropometric and household hygiene surveys were performed. In June 1997, a serosurvey was conducted on 187 study participants over 6 months old. The serosurvey was repeated in October 1997. All sera were tested for IgG antibodies to H. pylori. RESULTS: The June prevalence of H. pylori infection was 10%, and the incidence from June to October was 15%. Between June and October, 8 (42%) of 19 children that were positive for H. pylori infection seroreverted to negative. All seroreversions occurred in children 6-17 months. Other than age, no sociodemographic or environmental factor was significantly associated with incident H. pylori infection. There was no significant differences in the weight-for-age, weight-for-height, and height-for-age z-scores between children with and without prevalent H. pylori infection. CONCLUSIONS: Infection with H. pylori is common in Egyptian children under 3 years old and is not associated with malnutrition. No predictors for H. pylori infection were found. Our preliminary evidence for transient H. pylori infections in young children needs to be confirmed in a prospective cohort study, and predictors for persistent infection should be sought, since only these may be relevant to the known sequellae of infection.

Astrovirus diarrhea in Egyptian children.

This study describes the epidemiology of astrovirus diarrhea among a population-based cohort of 397 children aged <3 years residing in rural Egypt from 1995 to 1998. The age-specific incidence rates of astrovirus diarrheal episodes per person-year were 0.38 for infants aged <6 months, 0.40 for those aged 6-11 months, 0.16 for those aged 12-23 months, and 0.05 for those aged 24-35 months. The overall incidence rate of astrovirus diarrhea was the same as that of rotavirus diarrhea, 0.19 episodes per person-year. Astrovirus infection was pathogenic and associated with severe dehydration in 17% of the cases. The most frequent serotype was HAstV-1, and, in order of decreasing frequency, HAstV-5, HAstV-8 and HAstV-3, HAstV-6, HAstV-4, and HAstV-2. In determining whether astrovirus diarrhea was associated with a reduced incidence of subsequent disease, there was evidence to suggest HAstV-1 homotypic immunity but not heterotypic immunity. Because we observed 38% of the incidence of astrovirus diarrhea to occur in infants aged <6 months, a candidate astrovirus vaccine would have to confer immunity very early in life.

Epidemiology of rotavirus diarrhea in Egyptian children and implications for disease control.

Reliable epidemiologic data are essential for formulating effective policy to control rotavirus disease through immunization. The objective of this study was to describe the epidemiology of rotavirus diarrhea in a population-based cohort of children under 3 years of age residing in Abu Homos, Egypt, in 1995-1996. Rotavirus diarrhea incidence rates (episodes per person-year) were 0.13 for infants aged <6 months, 0.61 for those aged 6-11 months, 0.17 for those aged 12-23 months, and 0.15 for those aged 24-35 months. Fifty-six percent of children with rotavirus diarrhea had clinical dehydration; 90% of rotavirus diarrheal episodes occurred between July and November. In infants under 1 year of age, receipt of breast milk was associated with a lower incidence of rotavirus diarrhea. No other sociodemographic or environmental factor was found to be significantly associated with rotavirus diarrhea. Of 46 rotavirus isolates with strains identified, 41 (89%) were G serotypes 1 and 2. Rotavirus diarrhea was a major cause of morbidity in this cohort. Promotion of breastfeeding may exert a protective effect in young infants in this setting, but improvements in water and sanitation are unlikely to be effective preventive measures. The use of effective immunization against rotavirus in early infancy should be considered a public health priority.

PIP: This study describes the epidemiology of rotavirus diarrhea in a population-based cohort of children under 3 years of age residing in Abu Homos, Egypt, during 1995-96. Samples consisted of a cohort of children under the age of 24 months assembled from two villages in the vicinity of Abu Homos. The age-specific incidence rates of rotavirus diarrheal episodes per person-year were 0.13 for infants aged 6 months, 0.61 for those aged 6-11 months, 0.17 for those aged 12-23 months, and 0.15 for those aged 24-35 months. No rotavirus diarrheal incidence occurred in infants under 20 weeks of age. The monthly incidence rates of rotavirus diarrhea demonstrate that 90% of the disease episodes occurred during the warmer months of July-November, with a peak incidence in August. In infants under 1 year of age, breast-feeding was associated with a lower incidence of rotavirus diarrhea. Promotion of breast-feeding may employ a protective effect in young infants in this setting, but improvements in water and sanitation are unlikely to be effective preventive measures.

Characterization of unusual G8 rotavirus strains isolated from Egyptian children.

We report the first detection of P[14], G8 rotaviruses isolated in Egypt from the stool of children participating in a 3 year study of rotavirus epidemiology. Two strains, EGY1850 and EGY2295, were characterized by a serotyping enzyme immunoassay (EIA), virus neutralization, and sequence analysis of the genes encoding VP7 and the VP8* portion of the VP4 gene. These two strains shared a high level of homology of their VP7s (87.8% nucleotide [nt], 97.2% amino acid [aa]) and VP4s (89.6% nt, 97.1% aa) and had the highest VP7 identity to serotype G8 (> 82% nt, > 92% aa) and VP4 identity to genotype P[14] (> or = 81% nt, > 91% aa) strains. Serological results with a VP7 G8-specific and VP4 P[14]-specific neutralizing monoclonal antibodies supported the genetic classification of EGY1850 and EGY2295 as P[14], G8. Genogroup analysis supports earlier findings that human G8 rotaviruses may be genetically related to bovine rotaviruses. These findings demonstrate that our understanding of the geographic distribution of rotavirus strains is incomplete, emphasize the need to monitor rotavirus serotypes, and extend the known distribution of serotype G8 and genotype P[14] strains in Africa.

Epidemiology of enterotoxigenic Escherichia coli diarrhea in a pediatric cohort in a periurban area of lower Egypt.

Enterotoxigenic Escherichia coli (ETEC) are diverse pathogens that express heat-labile (LT) and/or heat-stable (ST) enterotoxins, yet little is known about whether epidemiologic patterns of pediatric ETEC diarrhea vary by the expressed ETEC toxin phenotype. In total, 242 Egyptian children aged <3 years were prospectively followed in 1993-1995. ETEC episodes were detected during twice-weekly home visits, and asymptomatic ETEC excretion was identified from monthly cross-sectional surveys. ETEC episodes were 0.6 per child-year. ST-only ETEC was 2.6 times (P<.001) more common in warmer than cooler months, while LT-only ETEC showed no seasonal variation. Ownership of a household sanitary latrine, but not breast-feeding, was associated with a lower risk of both enterotoxin phenotypes. Coexpression of a colonization factor by LT- or ST-only ETEC strengthened the association with diarrhea. These findings indicate that the epidemiologic patterns of LT-only and ST-only ETEC are not identical and that disease interventions should include improved household sanitation.

Oral, inactivated, whole cell enterotoxigenic Escherichia coli plus cholera toxin B subunit vaccine: results of the initial evaluation in children. PRIDE Study Group.

Two randomized, double-blinded trials assessed the safety and immunogenicity of an oral, killed enterotoxigenic Escherichia coli (ETEC) plus cholera toxin B subunit vaccine in Egyptian children. Two doses of vaccine or E. coli K-12 were given 2 weeks apart to 105 6- to 12-year-olds and 97 2- to 5-year-olds. Safety was monitored for 3 days after each dose. Blood was collected before immunization and 7 days after each dose to measure immune responses. Few children reported postdosing symptoms, with no differences in the frequency of symptoms between treatment groups. Most vaccinees had an IgA antibody-secreting cell response against colonization factor antigen I (100%, 6-12 years; 95%, 2-5 years), coli surface antigen 2 (92%, 6-12 years; 83%, 2-5 years), and coli surface antigen 4 (93%, 6-12 years). Vaccination evoked a >/=4-fold rise in antitoxic IgA and IgG titers in 93% and 81% of children, respectively. In conclusion, the oral ETEC vaccine was safe and immunogenic in 2- to 12-year-old children, justifying further evaluation in infants.

Seasonality as a determinant of the efficacy of praziquantel in population-based chemotherapy: lessons from the practice.

During October-November 1991, 1356 male farmers, 18-40 years old from a village in Fayoum Governorate, Egypt, were examined for Schistosoma haematobium infection. The prevalence of infection was 22.2%. Infected farmers were treated immediately with praziquantel at the recommended dose of 40 mg/kg body weight in a single, oral dose. Twelve weeks after treatment 86/262 (32.8%) infected farmers were negative for S. haematobium eggs in urine. In another study conducted in a satellite village in the Nile Delta (Beheira Governorate) where S. mansoni infection is prevalent, all residents of both sexes between 5 and 50 years of age (n = 858) were examined for S. mansoni. The prevalence of S. mansoni was 69.0%. Infected subjects were treated with the same dose of praziquantel during January-February 94, with an overall resultant efficacy of 85.5% (471/551) and 97.2% (103/106) in 18-40 years old males 8-10 weeks post treatment. The high cure rate in the second study was probably because treatment took place 2 months after the end of the high transmission season.

The prevalence and intensity of Schistosoma mansoni infection among residents in two Egyptian villages, at two and twelve months post-treatment with 40 and 60 mg praziquantel per kg body weight at these two villages respectively.

Two different regiments of praziquantel, 40 mg/kg in a single dose and 60 mg/kg in two divided doses administered 6 hr apart for the treatment of S. mansoni infection, were compared at two villages of Beheira Governorate. All residents of both sexes between 5 and 50 yrs old from two rural communities (Farshout and Om El-Laban) participated in this study. The prevalence of S. mansoni infection was 58% at Farshout and 69% at Om El-Laban. Infected subjects received 60 mg/kg of praziquantel in two divided doses 6 hr apart at Farshout and 40 mg/kg in a single dose at Om El-Laban. Two months post-treatment, better cure rates (96% vs 85%, P < 0.001) were achieved with 60 mg/kg of praziquantel. Twelve months post-treatment the prevalence of infection (45% for both villages) increased, but remained lower than before treatment. This may indicate that transmission continued from cercariae transported by water from nearby hyperendemic areas. The prevalence and intensity of infection were higher in the younger (5-19 y) than in the older (20-50 y) age group both before treatment and one year post-treatment. The same applies to the incidence and reinfection of the treated subjects. We recommend that 60 mg/kg in two divided doses six hours apart be used for school children. For logistical reasons, the single dose, 40 mg/kg regimen, should be used for mass treatment. For better control of disease, mass chemotherapy with praziquantel as suggested above is essential for the whole endemic region rather than only for a focal area.

Praziquantel in the treatment of Schistosoma mansoni infection: comparison of 40 and 60 mg/kg bodyweight regimens.

Two different regimens of praziquantel, 40 mg/kg in a single dose and 60 mg/kg in two divided doses administered 6 hr apart, for the treatment of Schistosoma mansoni infection were compared in Abu Homos, Egypt. Participants in this study included 1,588 subjects of both sexes between five and 50 years old from two rural communities; 730 from Farshout and 858 from Om El-Laban. The prevalence of S. mansoni infection was 57.9% in Farshout and 69.0% in Om El-Laban. Infected subjects received 60 mg/kg of praziquantel in two divided doses 6 hr apart at Farshout and 40 mg/kg in a single dose at Om El-Laban. Eight to ten weeks after treatment, three stool samples were collected from each treated subject on three consecutive days and evaluated quantitatively by the modified Kato technique. The failure rate was significantly higher at Om El-Laban than at Farshout (14.5% versus 4.1% (odds ratio [OR] = 3.95, P < 0.0001). A significantly higher failure rate was consistent after adjustment for age and number of excreted S. mansoni eggs per gram of feces before treatment (OR = 3.75, P < 0.0001). In spite of differences among cure rates between the two regimens, we recommend 60 mg/kg on an individual basis but not for population treatment, since administration of a single dose of a drug is more practical than two doses given 6 hr apart.

Prevalence of anti-Rift-Valley-fever IgM antibody in abattoir workers in the Nile delta during the 1993 outbreak in Egypt.

In the early summer of 1993, an outbreak of Rift Valley fever (RVF) was reported among both humans and animals in Aswan Governorate, Upper Egypt. To determine whether RVF infection had spread to the Nile delta region of the country, we carried out a cross-sectional survey of 1181 occupationally exposed abattoir workers (97% male; age 10-72 years) in 15 governorates of Egypt in November 1993. The overall prevalence of anti-RVF virus IgM antibody was 2% (range: 0% (7 governorates) to 10%). The highest prevalences were in Ismailia (10%) and Sharqîya (8%) Governorates. None of the seropositive subjects reported having experienced an episode of fever in the 2 months prior to the study. The prevalence of antibody was significantly higher (P < 0.05) among workers employed in high-risk jobs such as cutting animals' throats (relative risk (RR = 2.24)) and handling animal parts (RR = 2.37). The findings suggest that abattoir workers represent a useful sentinel population for surveillance of RVF.

Evaluation of a twice-a-week application of 1% niclosamide lotion in preventing Schistosoma haematobium reinfection.

A randomized double-blind trial was conducted to assess the efficacy of a twice-a-week application of 1% niclosamide lotion for prevention of Schistosoma haematobium reinfection. Six hundred farmers in Fayoum, Egypt, 18-40 years of age, were treated to cure their S. haematobium infection, then randomly assigned to self-apply niclosamide or placebo lotion to their limbs, neck, and torso. Subjects were exposed to schistosomal-infested water during routine irrigation activities from April to October 1992. Three hundred fifty subjects met the inclusion criteria and completed the trial, 169 (48.3%) in the niclosamide group and 181 (51.7%) in the placebo group. The subjects assigned to the niclosamide-treated group were comparable with those in the placebo group in age (27.2 versus 27.8 years), total water contact (101.9 versus 109.0 hr), lotion application compliance (93.5% versus 90.6%), and avoidance of whole body water contact (94.7% versus 96.7%). The reinfection rate with S. haematobium was 30.8% in the niclosamide-treated group and 28.2% in the placebo group. Niclosamide lotion applied to the limbs and trunk twice a week failed to prevent S. haematobium reinfection.

Chlamydia trachomatis associated with chronic dysuria among patients with Schistosoma haematobium.

Eighty-seven Egyptians with Schistosoma haematobium infection (current or past) and chronic dysuria seen at the Theodor Bilharz Research Institute outpatient urology clinic in Imbaba, Cairo, Egypt were investigated for Chlamydia trachomatis infection. Direct fluorescent antibody tests (DFAT) showed 30 urethral swabs (35%) and 40 urine sediments (46%) were positive for Chlamydia. Indirect fluorescent antibody testing showed 29 (34%) serum samples and 25 (29%) urethral smears were positive for Chlamydia. There was a highly significant association between patients with active Schistosoma infection and those with urine and urethral specimens positive for Chlamydia by DFAT (P < 0.01).

Field trial of 1% niclosamide as a topical antipenetrant to Schistosoma mansoni cercariae.

A randomized, double-blind, placebo-controlled field trial of a topical antipenetrant lotion, 1% niclosamide, applied daily to the upper and lower limbs of farmers occupationally exposed to Schistosoma mansoni cercarial-infested water, was conducted in the Nile Delta to assess its safety and efficacy in preventing reinfection. Farmers aged 18-40 years were treated to cure their S. mansoni infections three months prior to the onset of the trial. Subjects were randomly assigned to receive niclosamide or placebo lotion that was self-applied daily for five months. A total of 186 subjects met the inclusion criteria and completed the trial. The exposure to schistosomal-infested water occurred during routine irrigation activities from June to November 1991. Stool specimens were evaluated monthly during and for two months following the lotion application period. The subjects applying the niclosamide lotion were comparable to those applying placebo lotion in age (mean 30 years for both), total water contact (184.5 hr versus 173.8 hr), reported lotion application compliance (88% versus 92%), and reported water contact involving skin exposure other than upper and lower limbs (23% versus 27%). The schistosomal reinfection rate was lower in the niclosamide group (53.3%) compared with the placebo lotion group (71.3%), (P < 0.02). Increased protection might be obtained with total body application for shorter, less intense, water contact exposures.

Serosurvey of prevalence of human immunodeficiency virus amongst high risk groups in Port Sudan, Sudan.

During October, 1987, 593 sera were collected from risk groups in Sudan's only major deepwater port, Port Sudan. The risk groups included prostitutes, lorry drivers and prisoners. A large proportion of the study participants practised high risk behaviour which included sexual promiscuity, medical treatment by injection, scarification and tattooing. Despite high risk behaviour and evidence of a high prevalence of hepatitis B infection, a virus transmitted in a manner similar to HIV, no study participants were positive for HIV infection. This data suggests that the prevalence of HIV infection amongst high risk groups in Port Sudan is very low. These findings confirm a lack of clinical cases of AIDS in hospitalized patients in Port Sudan and the small number of reported cases in other areas of northern Sudan.

PIP: The high prevalence of hepatitis B markers in the Sudan (up to 80% of those surveyed) suggests the potential for a rapid spread of human immunodeficiency virus (HIV) since both viruses are transmitted in similar ways. Although clinical cases of acquired immunodeficiency syndrome (AIDS) have not been reported from Port Sudan, southern Sudan borders on several countries with a high prevalence of HIV infection. Sudan's National AIDS Committee plans a series of surveys to determine the prevalence of HIV infection in high risk groups and the general population in several geographic regions. The 1st such survey was conducted in Port Sudan in 1987 among 593 high-risk individuals (203 prostitutes, 103 lorry drivers, 118 prisoners, and 169 in mixed occupations). The study population included 330 males and 263 females. About half of the participants were married and in the 21-30-year age group. Over 75% had been exposed to hepatitis B and 76% had been treated for malaria, largely through injection. Overall, the incidence of non-sex-related risk factors for HIV infection among Port Sudan subjects was: injection, 48%; scarification, 40%; and tatoos, 38%. 32% reported a prior history of a sexually transmitted disease. 71% of the males had used prostitutes. Surprisingly, no study participants were positive for HIV infection. This finding presumably reflects Port Sudan's geographic isolation from other Central and East African countries with large numbers of HIV-positive individuals. On the other hand, Port Sudan is the site of importation of all goods by sea into the country and many people from other African and Arab countries are associated with the seaport. Thus, once the HIV virus is introduced by infected persons from other areas, the risk factors suggest the potential for rapid transmission.

HIV infection surveillance in Mogadishu, Somalia.

A group of 89 prostitutes and 45 patients attending sexually transmitted disease clinics in Mogadishu, Somalia were examined for evidence of HIV infection. Both groups reported more than 1 sexual partner routinely and had sexual contacts with prostitutes. There was a significant amount of sexually transmitted diseases (STDs) in these two groups, with 11.2% and 6.7% respectively being culture positive for N. gonorrhoea. Among the prostitutes, 28.1% were positive for antibodies to T. pallidum while only 4.4% of the STD patients were positive. One isolate of N. gonorrhoea was resistant to penicillin. All study participants were negative for antibodies to HIV suggesting an extremely low prevalence of HIV in high risk behaviour groups in the capital city of Somalia.

PIP: 89 prostitutes and 45 men attending the sexually transmitted disease (STD) clinic in Mogadishu, Somalia, were tested for HIV (human immunodeficiency virus) with the Abbott ELISA (enzyme-linked immuno-sorbent assay) test, cultured for gonorrhea, and screened for syphilis. There were no sera positive for HIV. 11% of the prostitutes and 7% of the men had positive gonorrhea cultures; 28% of the prostitutes and 4% of the men were positive for syphilis; 1 of the men had penicillin-resistant N. gonorrhoea with a beta-lactamase test. An epidemiological questionnaire was administered to the subjects. Most were aged 20-29; 67% were married; 80% of the men and 22.5% of the women were soldiers. 40% of the men reported use of prostitutes. Stated numbers of sexual contacts were 1.87/week for the prostitutes, and 1.51/week for the men. Data were also reported on occupations, recent injections, immunizations, intravenous drug use, surgery, blood transfusions and scarification.

Risk factors of gastric precancerous lesions in a high-risk Colombian population. I. Salt.

A case-control study for stomach cancer was conducted in a high-risk population in Nariño, Colombia to determine the risk of gastric precancerous lesions associated with salt intake measured by sodium-to-creatinine ratio of a single urine sample. Gastric biopsies and urine samples were collected from 263 individuals. Urinary sodium-to-creatinine ratios were studied in relation to histological data from the biopsies. Significantly high odds ratios for precancerous lesions (chronic atrophic gastritis, intestinal metaplasia, and dysplasia) were associated with higher sodium-to-creatinine ratios. Adjusted odds ratios (OR) of 2.50 for chronic atrophic gastritis and 7.24 for dysplasia were found. The association with intestinal metaplasia was weaker and not significant (OR = 1.57). Furthermore, an excess risk associated with adding salt to food at the table was found among patients with precancerous lesions (OR = 1.80). These findings support the two-step involvement of salt in the process of gastric precancerous lesions.

Risk factors of gastric precancerous lesions in a high-risk Colombian population. II. Nitrate and nitrite.

Gastric nitrite content was studied in relation to precancerous lesions of the stomach in a case-control study conducted in a high-risk Colombian population. The proportion of detectable nitrite in gastric juice and the mean pH were significantly higher among those with precancerous lesions (chronic atrophic gastritis, intestinal metaplasia, and dysplasia) than among the controls (normal and superficial gastritis); the proportion and mean pH increased with the progression of histological changes from normal to dysplasia. Nitrite was not detectable in gastric juice with a pH less than 5.0. A positive association was found between the proportion of detectable nitrite and the risk of gastric precancerous lesions. Odds ratios of 4.39 for intestinal metaplasia and 24.72 for dysplasia remained significant after controlling for confounders. This finding suggests that nitrite may be a precursor of a mutagen that targets gastric epithelial cells.