- Keith R Neal (), lecturer in communicable disease controla,
- Helen M Scott, medical studenta,
- Richard C B Slack, consultant in communicable disease controlb,
- Richard F A Logan, reader in clinical epidemiologya
- a Department of Public Health Medicine, University of Nottingham, University Hospital, Queen's Medical Centre, Nottingham NG7 2UH
- b Nottingham Health Authority, Nottingham NG1 6GN
- Correspondence to: Dr Neal.
- Accepted 21 July 1995
Gastric acid protects against enteric infections,1 and patients who have had gastric surgery or are taking H2antagonists are more susceptible to salmonella infection.2 3 Antibiotic treatment also increases the risk of infection.3 It is not known whether these factors are also associated with campylobacter infection, for which statutory notifications now exceed those for salmonella.4 We conducted a case-control study to assess whether gastric antisecretory drugs, antibiotics, and abdominal surgery are associated with campylobacter infection.
Patients, methods, and results
Between January 1992 and August 1994, 243 notified cases of campylobacter infection, confirmed by faecal culture, were identified in people aged 45 and over in two of the local district councils within Nottingham Health Authority. Thirty two cases were excluded (non-resident (four), general practitioner declined (19), patient died and notes unobtainable (six), and notes unobtainable at general practice (three)), leaving 211 (123 women). The minimum age was 45 because people over this age have higher rates of prescribing by general practitioners. Controls were identified as the next two patients matched for sex and age within two years in the practice computerised records. No controls were excluded.
Data on previous surgical operations; prescriptions for H2 antagonists, proton pump inhibitors, antibiotics, hydroxocobalamin, and corticosteroids; and regular precriptions and other drugs used before infection were extracted from the general practice records. Data were analysed by conditional logistic regression using the EGRET package with the magnitude of associations measured by odds ratios. The study had 80% power to detect a 2.5-fold risk, given that 4% of the general population was exposed.
Omeprazole treatment in the month before infection was associated with a 10-fold increased risk of campylobacter infection (table 1). This was independently significant only for current use. The association with H2 antagonists was not significant after omeprazole use was controlled for. Antibiotic treatment in the two to 12 months before infection was associated with a relative risk of 2. No associations were seen with previous gastric or colonic surgery, pernicious anaemia, corticosteroids, use of other drugs, or the number of regular prescriptions. Analyses of subgroups by age (over 65, under 65) and sex showed the same associations.
Our results show that use of omeprazole predisposes to clinical campylobacter infection. The finding that current users but not former users of omeprazole were at increased risk suggests that the relation is causal. Omeprazole probably increases the risk or severity of infection by reducing the gastric killing of ingested organisms. Some cases of diarrhoea with omeprazole may be infective and should be investigated by faecal culture.
No relation was seen with H2 antagonists despite the power of the study to detect a relative risk of 2.5 for H2 antagonist use in the month before infection. H2 antagonists reduce gastric acidity less than proton pump inhibitors, which could leave sufficient acid to reduce the ingested dose of campylobacter organisms. The increased acid suppression with omeprazole allows more organisms to survive, so increasing the risk of clinical infection. Campylobacters are more acid sensitive than are salmonella.4 Severity of salmonella infection is related to size of infecting dose,5 and the same may apply to campylobacter. The greater acid sensitivity of campylobacter may also explain the lack of association seen with previous gastric surgery, which produces relatively modest reductions in acid secretion. Our results also suggest that antibiotics increase the risk of campylobacter infections, as with salmonella3; use of antibiotics may predispose to infection by altering bowel flora.
The absence of an association with hydroxocobalamin, which was used as a proxy for pernicious anaemia, and similar findings for salmonella,3 is surprising and suggests that other factors are important. Many of our patients had recently started taking omeprazole; temporary reduction in acid may be important, and different factors may operate in long term acid reduction, such as changes in bacterial flora, which protect against campylobacter and salmonella infection.
In conclusion, proton pump inhibitors lead to a significant increased risk of campylobacter infections in people aged 45 or over, an effect not seen with H2 antagonists or previous gastric surgery. This can be explained by differences in acid suppression and the pH sensitivity of campylobacter.
We thank Mrs G Campion and Mrs M Edmonds for help in collecting data.
Conflict of interest None.