Since its discovery in 1984 Helicobacter pylori has been recognised as the major causal factor of the histological changes leading to severe gastroduodenal disease, including peptic ulcer disease, gastric adenocarcinoma and gastric mucosa-associated lymphoid tissue (MALT) lymphoma and was classified by the International Agency for Research on Cancer (IARC) as a group I carcinogen in 1994.1^–5 Essentially, all of those infected will have chronic gastritis, although most infections remain asymptomatic.6

TRANSMISSION

How H pylori is usually acquired and its routes of transmission are still unclear.2 Current knowledge implies various pathways of person-to-person transmission. Both faeco-oral as well as oro-oral transmission7 and, more recently, gastro-oral transmission (that is, by vomiting) early in life have been hypothesised, but proof is lacking.2 It is possible that all three routes, together with water and food contamination, may occur and may be of different relevance among various populations.8^–10 Direct human-to-human contacts have been suggested as the primary route in industrialised countries while the faeco-oral route, also through contaminated water, has been suggested in developing countries.4 11^–14 In epidemiological studies, the importance of low socioeconomic conditions, crowding and interfamilial spread on acquisition of H pylori infection have been documented.11 There is currently no evidence for zoonotic transmission, although H pylori is found in some non-human primates and occasionally in other animals.8 15

From the literature of the early 1990s emerged the possibility of occupational acquisition of H pylori; the subject was recently reviewed.10 According to this review, the risk seemed to be increased for gastroenterologists and some types of nurses. Variation in study design including study power, selection or response of control group or failure to match for, stratify or correct by logistic regression for potential confounders or baseline occupational risk factors could account for some of the observed differences.

In institutions for people with intellectual disability, the prevalence of H pylori infection is high among the residents.16 17 The close personal contact inherent in personal care may expose workers in these institutions to an increased risk of acquiring H pylori.

We conducted a cross-sectional study to determine whether healthcare workers in these institutions have an increased rate of H pylori infection compared to a non-exposed control population and to identify risk factors for infection.

SUBJECTS AND METHODS

Study population

Personnel at two institutions for children with intellectual disability having a documented high prevalence of H pylori infection in the residents (86% presenting antibodies against H pylori)17 were invited to take part in this study. In these institutions children with moderate to severe intellectual disability are supported in their daily living activities. Up to 90% live permanently in the institution. Approximately 66% have to be supported while eating (including gastric tube feeding), 75% are incontinent. Prevalence rate of H pylori antibodies was compared to a control group of non-exposed workers. The control group consisted of workers mostly in insurance and banking companies and they agreed to be tested for H pylori when they volunteered for a check-up, oriented to cardiovascular risk factors (serum cholesterol, glycaemia, hypertension, obesity, etc). Based on published data from Belgium,18 prevalence was assumed to be 30% in the non-exposed population. In order to be able to detect a 10% difference in prevalence, with a power of 80% and a level of significance of 0.05, a sample size of at least 370 in both groups was needed. The total number of subjects were 765 in the two institutions, and all workers were invited to take part in the study as it would have been difficult for psychological reasons to include only a sample. Workers in the control group were recruited as they presented themselves for the check-up. The ethics committee of the University Hospital of Ghent, Belgium approved this study. Written informed consent was obtained from each subject.

Data collection

In the study serology was used for the diagnosis of H pylori infection. Serum samples were tested for H pylori IgG antibodies using a second generation immunoglobulin G enzyme immunoassay (Cobas Core Anti-Helicobacter pylori EIA; Roche SA, Basel, Switzerland) according to the manufacturer’s instructions. This method has been shown to have a sensitivity of 93% and specificity of 95% in patients.19

Questionnaire

All participants received a questionnaire. Information was collected on:

• Demographic and socioeconomic factors such as age, sex, education of the father, education of the mother, number of household members and number of children sleeping in the same bedroom during childhood as indicators of crowding

• Medical history such as history and symptoms of upper gastrointestinal diseases and history of medication

• Lifestyle factors such as smoking and tropical journeys

• Employment data such as type of contact with residents (changing napkins with stools, washing and feeding residents, personal contact) and seniority in the institution.

In the control group the questions regarding contact with residents were omitted.

Statistical analysis

In the analysis we first compared study and control groups for a number of risk factors.

The proportion of H pylori positivity through occupation was estimated by calculating the attributable risk using the formula:

We used the χ² test or the independent-samples t test where appropriate to analyse differences between the study group and the control group. The Mann-Whitney test was used for non-normally distributed variables. A statistical significance level of 5% was used. Percentages were directly adjusted for age by using age groups 20–29, 30–39, 40–49 and ⩾50 years, using the study group as reference group. Variables associated with H pylori positivity by univariate analysis were entered into a binary logistic regression analysis (backward stepwise likelihood ratio).

In the logistic model H pylori infection status was the dependent variable. Adjusted odds ratios (ORs) were calculated to identify variables that were independently associated with H pylori positivity. The reference category for OR estimates was no exposure. Data were analysed using the SPSS statistical software package for Macintosh (version 10.0).

RESULTS

A total of 671 of the 765 staff members in the institutions (response rate 87.7%) participated in the study; in the control group refusals were not noted. Complete data were available for 587 people in the study group and 390 in the control group and analysis was done on this group.

The main characteristics of both study and control groups are given in table 1.

The mean age was lower and the percentage of females was clearly higher in the study group, as was the percentage of smoking. Body mass index (BMI) and the percentage reporting tropical journeys were higher in the control group. The number of household members during childhood was slightly lower in the control group. Both groups were comparable for other potential risk factors for H pylori infection such as parent’s education and number of children sharing the same bedroom during childhood. The prevalence of upper gastrointestinal diseases symptoms and history of gastroscopy was not different between both groups (data not shown); only a history of gastric ulcer treatment was more frequently reported in controls (7.2% versus 2.0% in the exposed group) (p = 0.04).

The prevalence of H pylori antibodies in the study group was 40.6% compared to 29.2% in the control group. This difference was statistically significant (p<0.001). The seroprevalence of H pylori was higher in older age groups in both study and control group (see fig 1). In the exposed group the age-stratified seroprevalence increased significantly with length of employment in the institution in each age group, except in those above 50 years of age (data not shown). The age-standardised prevalence of H pylori antibodies in the control group was 28.3 (95% CI 23.8 to 32.8); the difference between study and control groups remained after standardisation for age and even increased slightly.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1 Prevalence of seropositivity for Helicobacter pylori by age group in exposed and non-exposed subjects.

Crude odds ratio (OR) was 1.68; adjusting for the confounding effect of age, gender, BMI, smoking, tropical journeys and number of household members during childhood resulted in an even higher, significantly increased adjusted OR of 1.98 (95% CI 1.42 to 2.69).

Based on these results, the attributable risk of being occupationally infected by H pylori was calculated to be 49.5%.

In order to explore the independent contribution of each significant occupational exposure variable on infection of H pylori, we used multiple logistic regression analysis to adjust the ORs for the confounding effect of age, BMI, years of employment, father’s education, mother’s education, number of children sharing the same bedroom, personal contact, faecal contact, washing and feeding of inhabitants (variables associated with H pylori positivity in univariate analysis) (table 2).

Only faecal contact with residents showed an increased adjusted OR for the risk for H pylori infection (p = 0.001): OR = 3.98 (95% CI 1.68 to 9.46) for regular faecal contact and OR = 2.11 (95% CI 0.88 to 5.02) for occasional faecal contact.

DISCUSSION

In a study in healthcare workers in two institutions for people with intellectual disability, with a documented high prevalence of H pylori infection (86%), the crude OR for infection was 1.68, but the adjusted occupational risk of acquiring H pylori infection was almost double (adjusted OR  = 1.98) that of a control group of non-exposed workers.

The attributable risk was 49.5%, meaning that almost half of the infections in this group of workers were caused by the occupational exposure and thus were potentially preventable.

As the controls were self-selected, this could have attracted people with symptoms and/or a history of upper gastrointestinal diseases with a higher prevalence of H pylori than the general population. However, results showed that only a history of gastric ulcer treatment was more frequently reported among controls. As it is not known what kind of treatment (eradication or symptomatic) they received, this difference could have affected the results in at least two (opposite) ways: firstly, the controls could have a lower prevalence of H pylori after having been treated (and H pylori being eradicated) more frequently; secondly, the controls could have a higher prevalence of H pylori if they had not been treated by eradication therapy for H pylori.

The main strength of this study is the precise (self) classification of employees as to the nature of their occupational contact with residents with a disability, which allowed the investigation to demonstrate the relative importance of faeco-oral transmission in this population. This, however, does not mean that other transmission routes do not occur, but we were not able to demonstrate their importance. One of the reasons for that could be because it is not easy to determine quantitatively the exposure to oral secretions and vomiting by residents in institutions.

The cross-sectional design might be considered the main weakness of this study, as it is not designed to test causality. On the other hand, cross-sectional studies are considered more valuable for studying diseases with a long duration and short latency period after exposure and not causing healthy worker bias when early symptoms of the disease cause workers to terminate employment.20 These three conditions are fulfilled for H pylori: seroconversion is mostly lifelong (if not treated), the latency period between infection and seroconversion is short and, as also shown in our study, most infected people do not complain of signs and symptoms of the infection.2 6

Recent studies have also shown the importance of transmission from other infected members of the household.21 In our study we did not test members of the workers’ household but household transmission should be comparable in both exposed and non-exposed subjects.

Comparison with other studies

Although studies of H pylori infection in institutionalised individuals with intellectual disability have consistently reported much higher rates of infection (60–90%) compared to matched groups in the general population (30–40%),16 17 22^–25 relatively few studies have been performed to study the occupational risk in healthcare workers caring for them.

In a Dutch study, prevalence of H pylori infection among the employees of an institution for people with intellectual disability was increased in those employees who had had intensive contact with the residents (31.6% for direct intensive physical contact and 14% for restricted physical contact or without direct contact), those who had worked longer than 5 years in the institute and those who had upper abdominal complaints.23 Although the occupational risk was comparable to our study, the prevalence of H pylori was higher in our study. This may be due to a different transmission risk between Belgium and The Netherlands in childhood, which has also been documented for other infections in childhood—for example, hepatitis A infection.26

Another study27 compared the seroprevalence of H pylori in personnel performing gastrointestinal endoscopy and healthcare workers caring for developmentally disabled individuals. Seroprevalence of H pylori was 24% in endoscopy personnel and 47% among the healthcare workers, comparable to our results.

Because these institutions are often regarded as a home for their residents (in our study, 90% lived there permanently), it is reasonable to assume that, unlike hospitals, infection control procedures based on standard precautions are not strictly enforced. This probably leads to less hygienic working conditions than in a hospital, with a higher transmission risk of infectious agents.28^–30

Studies on other healthcare workers have been reviewed.10 Almost all cross-sectional studies and all published cohort studies showed an increased risk for healthcare workers. Since this review, a cross-sectional study of gastroscopy nurses has shown no excess risk.31

Preventive measures

There are very few data available concerning protective measures which are effective in preventing occupational H pylori transmission, particularly in institutions for people with intellectual disability.28

Primary prevention

The prevalence of H pylori has consistently been shown to be higher in institutionalised people with intellectual disability, so deinstitutionalisation of people with intellectual disability could be the single most important primary prevention measure. It might be argued that deinstitutionalisation could shift the risk of infection from the workers to patients’ relatives. However, this risk in a family setting would be much smaller than that of workers in an institution, as it has been shown that the prevalence of H pylori in non-institutionalised people with intellectual disability is about half that of institutionalised people.16 Screening and subsequent treatment of infected people with intellectual disability could be an alternative, but the diagnostic tests are less reliable, H pylori eradication rate is lower, with treatment side-effects and, particularly, recurrences rates much higher than in the general population.16

Other preventive measures including implementation of standard precautions are recommended, particularly the consistent use of protective gloves and handwashing whenever faecal contact can occur. Gloves should be changed after contact with each patient. Hands should be washed immediately after gloves are removed. Workers should be instructed accordingly and an appropriate infrastructure, including widely available sinks, should be in place. There must be supervision and observation of compliance with existing guidelines and recommendations, and periodic review of both practice and guidance.

Data on other healthcare workers, particularly nurses in hospitals at risk for H pylori infection, have shown that these preventive measures are effective in preventing infection by H pylori.32

Secondary prevention

Secondary prevention is here defined as screening of asymptomatic infected workers and subsequent treatment33; up to now, this has only been recommended in high-risk groups.34 35

As H pylori is considered by IARC to be a group I carcinogen and thus gastric cancer could be defined, based on the results of our study, as a work-related cancer, screening of asymptomatic workers could be considered.

However, in the absence of controlled studies showing the benefits of screening for H pylori, screening in healthcare workers working with people with intellectual disability cannot be recommended, but this could change in the future if evidence of its effectiveness were available.

Implications for occupational health policy

Based on our study and other scientific evidence that is currently available we would certainly recommend that the problem of occupational risk of H pylori posed by taking care of institutionalised people with intellectual disability be recognised by all concerned (occupational physicians, employees, employers, public health authorities). We would further stress the importance of implementing universal precautions, particularly the consistent use of gloves and handwashing while handling faecal material.

Unanswered questions and further research

We have shown an increased prevalence of H pylori infection among healthcare workers, through faeco-oral transmission, but questions on other ways of transmission remain unanswered. Future research should include cohort studies and an even more detailed assessment of exposure to all possible transmission routes of H pylori (faeco-oral as well as oro-oral and gastro-oral).

To assess the value of screening healthcare workers for H pylori is an important research question.34