Gastric acid-suppressive therapy and community-acquired respiratory infections
Dr R. Laheij, University Medical Center St. Radboud, Department of Gastroenterology, PO Box 9101, 6500 HB Nijmegen, The Netherlands.
Background : Bacteria and viruses have been detected in the stomach of patients during acid-suppressive therapy.
Aim : To investigate whether subjects using acid-suppressive drugs more often develop community-acquired respiratory infections when compared to those who do not use acid-suppressive drugs.
Methods : 700 study subjects were recruited during a single week in December 2002. Information on the prevalence of clinical manifestations of infections and complications in the preceding month was assessed by questionnaire. Furthermore, subjects were asked to report antibiotic therapy and physician visits related to possible infection.
Results : Questionnaires were returned by 405 subjects (58%). Consumption of acid-suppressive drugs was reported by 91 individuals, of whom 79 used proton-pump inhibitors (20%) and 12 H2-receptor antagonists (3%). Overall, 101 (25%) responders reported clinical manifestations of respiratory infection in the preceding month. Subjects using acid-suppressive drugs were 2.34 times [95% confidence interval (CI) 1.4–4.1] more likely to have clinical manifestations of infection than individuals not using acid-suppressive drugs. Subjects using acid-suppressive drugs visited a physician 3.72 times more often (95% CI 2.1–6.8) for an infection and received antibiotic therapy 4.19 times more often (95% CI 2.2–8.1) in comparison to individuals not using acid-suppressive drugs.
Conclusions : Subjects using acid-suppressive drugs more often reported community-acquired respiratory infections in comparison to those who did not use acid-suppressive drugs.
Ingestion of pathogens can cause many different infections. However, most of these organisms rarely succeed in surviving the passage to the intestine because of the body's defence mechanisms. Gastric secretion of acid is an important defence mechanism of the stomach. Manipulation of acid secretion by acid-suppressive drugs alters the effectiveness of this barrier.1–3 Viruses have been detected in gastric mucosa biopsies of patients using acid-suppressive therapy.1 Hayase et al. reported the existence of type B influenza genes in the mucosa even in periods when influenza was not epidemic. Besides influenza virus, several species of bacteria that normally colonize the oral cavity and pharynx have been found in the stomach and duodenum.3–7
The presence of pathogens in the stomach has been associated with the occurrence of nosocomial respiratory infection, in particular ventilator-associated pneumonia.7–9 During prolonged mechanical ventilation the oropharynx, sinuses, dentition and stomach of patients using acid-suppressive therapy become colonized with pathogenic bacteria. Colonized secretions pool in the oropharynx and subglottic space, and gain access to the lower airways to multiply. The degree of bacterial overgrowth depends on the degree of reduction in gastric acid secretion and is therefore considerably higher in patients treated with a proton-pump inhibitor compared with a H2-receptor antagonist.4, 5
Limited information is available about the clinical consequences of pathogens in the stomach due to acid-suppressive therapy and community-acquired respiratory infection. Community-acquired respiratory infection is more common than nosocomial respiratory infection, and is also associated with significant morbidity and mortality. Identifying risk factors for the prevention of infection might be of interest, especially with the current outbreak of severe acute respiratory syndrome (SARS) and avian influenza in The Netherlands. The aim of this study was to investigate whether subjects using acid-suppressive drugs have a higher incidence of community-acquired respiratory infections in comparison to those who do not use acid-suppressive drugs. Because a specific diagnosis is seldom necessary or available to guide treatment, we focused on clinical manifestation instead of the type of pathogen involved.
This study was conducted at the department of Gastroenterology, UMCN St Radboud, Nijmegen, The Netherlands. The study subjects were recruited during 1 week in December 2002. Three hundred and fifty subjects who underwent endoscopy, mostly sigmoidoscopy and colonoscopy, and 350 nonhospitalized subjects living in the community were asked to participate. Subjects not undergoing endoscopy were selected by zip code residing in the same geographical area. Data collection was by questionnaire. The subjects were unaware of the hypothesis.
The questionnaire was sent to assess personal characteristics, anthropometric variables, and general lifestyle habits such as weight, length, smoking, alcohol or coffee consumption, travel to foreign countries and related medical history. Infections were identified by personal reporting of clinical manifestations of respiratory infection. Information on the prevalence of clinical manifestations in the preceding month included fever, common cold, influenza, laryngitis, otitis media, pneumonia and sinusitis. Subjects were asked to report whether a physician was visited, and antibiotics were prescribed per clinical manifestation. Furthermore, subjects were asked to report current medication use, doses and duration. Specific information about acid-suppressive drug use in the preceding month was assessed by mentioning all available proton-pump inhibitors (esomeprazole, omeprazole, lansoprazole, pantoprazole and rabeprazole) and H2 receptor antagonists (cimitidine, ranitidine, nizatidine and roxatidine).
Returned questionnaires were checked for accuracy and completeness. The data were entered automatically into a database by the use of a scanner supplemented by sensitive checking of the data. Statistical analysis was undertaken with SAS version 8.0. For categorical variables, differences were assessed by a χ2-test or Fisher's exact test; for continuous variables t-tests were employed as appropriate. We calculated an overall respiratory infection rate by combining the reported incidence of common cold, influenza, laryngitis, otitis media, pneumonia and sinusitis per person. Multiple regression analyses with entry of potentially confounding variables were used in the multivariate analysis. The confounding variables were age, obesity, smoking, asthma/chronic obstructive pulmonary disease (COPD), alcohol consumption and traveling. To study the difference in outcome between proton-pump inhibitors and H2 receptor antagonists we performed a multivariate regression analysis including confounding variables. An unadjusted logistic regression analysis was performed in order to identify differences in occurrence of the individual respiratory infections between patients with and without acid-suppressive therapy. We did not perform a multivariate regression analysis regarding the individual respiratory infections because the number of infections was insufficient. Missing values were excluded from the analyses.
Overall 405 out of the 700 questionnaires (58%) that were sent were returned; 226 of the respondents were male (57%) and the mean age was 55 years. Consumption of acid-suppressive drugs was reported by 91 individuals, of whom 79 subjects used proton-pump inhibitors (20%) and 12 subjects H2-receptor antagonists (3%). Subjects using acid-suppressive drugs were older (P = 0.03), more obese (P = 0.05), consumed less alcohol (P = 0.001) and visited more foreign countries (P = 0.05) than those who did not use acid-suppressive drugs (Table 1).
Table 1. Patient characteristics
|Males, no. (%)||177 (58)||49 (57)|
|Age, no. (%)|
| Younger than 45 years||100 (32)||16 (18)|
| Between 45 and 60 years||97 (31)||30 (33)|
| Above 60 years||117 (37)||44 (49)|
|Asthma/COPD||26 (8)||13 (14)|
|Obesity, no. (%)||86 (28)||35 (39)|
|Current smoking, no. (%)||190 (61)||58 (64)|
|Current alcohol drinking, no. (%)||195 (62)||38 (42)|
|Foreign country visit, no. (%)||3 (1)||4 (4)|
Overall 49 subjects reported fever in the preceding month, 16 out of 91 subjects (18%) using acid-suppressive therapy and 33 out of 314 subjects (11%) not using acid-suppressive therapy (P = 0.06). A total of 101 (25%) responders reported at least one respiratory infection in the preceding month; 26 reported a combination of infections. Fifty-six (14%) subjects reported clinical manifestations of common cold and 71 (18%) other respiratory infections. Eighteen out of the 56 subjects who reported common cold and 25 out of the 71 reporting other respiratory infections used acid-suppressive drugs. Subjects using acid-suppressive drugs were 2.06 times [95% confidence interval (CI) 1.2–3.4] more likely to have clinical manifestations of respiratory infection in the preceding month than individuals not using acid-suppressive drugs. This finding did not change [odds ratio (OR) 2.34, 95% CI 1.4–4.1] after adjustment for age, obesity, smoking, asthma/COPD, alcohol consumption and traveling.
Although all individual infections were more often reported by acid-suppressive drug users, only sinusitis (4% vs. 10%) and pneumonia (0% vs. 6%) were statistically significantly more reported (Table 2). Patients using proton-pump inhibitors (OR 2.21, 95% CI 1.3–3.9) and not H2 receptor antagonists (OR 3.07, 95% CI 0.9–10) more often reported respiratory infection. The age of the patients was also significantly associated with the occurrence of respiratory infection (OR 0.57, 95% CI 0.4–0.8). Furthermore, subjects using acid-suppressive drugs visited a physician 3.72 times more often (95% CI 2.1–6.8) for an infection and received antibiotic therapy 4.19 times more often (95% CI 2.2–8.1) in comparison to individuals not using acid-suppressive drugs. Adjustment for the confounding variables did not change the observed association between acid-suppressive drug use and both consulting a physician (OR 3.64, 95% CI 2.0–6.8) and antibiotic drugs prescribed (OR 3.92, 95% CI 2.0–7.7).
Table 2. Prevalence of clinical features of respiratory infections by acid-suppressive drug use
|Overall||68 (22)||33 (36)||2.06 (1.2–3.4)|
|Common cold||38 (12)||18 (20)||1.26 (0.8–2.1)|
|Influenza||29 (9)||12 (13)||1.49 (0.7–3.1)|
|Laryngitis||13 (4)||7 (8)||1.93 (0.8–5.0)|
|Otitis media||6 (2)||3 (3)||1.75 (0.4–7.1)|
|Pneumonia||1 (0)||5 (6)||18.2 (2–158)|
|Sinusitis||12 (4)||9 (10)||2.72 (1.1–6.7)|
Acid-suppressive therapy is associated with changes of the immune system leading to bacterial overgrowth in the stomach and influenza viruses in the gastric mucosa.1, 3 In the present study we investigated whether patients using acid-suppressive drugs have a higher incidence of community-acquired respiratory infections in comparison to those who do not use acid-suppressive drugs. Our results suggest that patients using acid-suppressive drugs have a 100% higher overall risk of becoming infected in comparison to those not on acid-suppressive therapy: in general, these infections do lead to more physician visits and antibiotic therapy. Because all measured clinical manifestations of respiratory infections were more frequently reported by subjects using acid-suppressive drugs this seems a real effect and not simple coincidence (unlike the finding that there is a statistical difference in infections between patients using proton-pump inhibitors and not H2-receptor antagonists). This is probably an erroneous conclusion resulting from insufficient sample size. Only a few patients used H2-receptor antagonists. Other studies showed that bacterial overgrowth does depend on the degree of reduction in gastric acid secretion and is considerably higher in patients treated with a proton-pump inhibitor compared to those treated with an H2-receptor antagonist.4, 5 However, viruses rather then bacteria are the causal pathogens of most respiratory infections.
Many organisms live in the gastrointestinal tract, and a wide range of organisms are capable of infecting the gastrointestinal tract.10 The stomach contents of healthy subjects harbour only transient organisms, although it may be colonized by acid-tolerant lactobacilli, streptococci and Helicobacter pylori. The upper intestine is only lightly colonized, but populations increase markedly in the ileum with streptococci, lactobacilli, enterobacteriaceae and bacteroides. Gastric and duodenal overgrowth frequently occurs when using acid-suppressive drugs.4–7 Approximately 50% of patients either using proton-pump inhibitors or H2-receptor antagonists have bacterial overgrowth. Bacteria identified in these studies belong to species colonizing the oral cavity and pharynx. Faecal types of bacteria were also found. Most of these bacteria are causative agents of infections throughout the body.
Furthermore, the results of previous studies showed that there is a relation between gastroduodenal dysfunction and bacterial colonization of the lungs.8, 11 In 11 out of 19 patients bacteria were isolated from stomach contents that were identical to those isolated form lower respiratory tract isolates.11 Patients with gastroduodenal dysfunction are often using acid-suppressive drugs to deal with dyspeptic symptoms such as regurgitation. A previous study investigating the effect of H2 receptor antagonist use on gastric flora and sputum showed that in 7 out of 12 patients gram-negative bacteria that may cause pneumonia were observed in sputum culture as well as in gastric juice.12 This was also reported in another study of Simms et al., who examined the role of gastric colonization in the development of pneumonia in critically ill trauma patients.9 Acid-suppressive therapy increases the number of bacteria in the upper gastrointestinal tract that may be pathogens for pneumonia.
Finally, proton-pump inhibitors diminish the defences of the body once pathogens successfully penetrate the body.13–15 The results from several in vitro studies have shown that proton-pump inhibitors impair neutrophil function and natural killer cell activity by decreasing extracellular and intracellular oxygen intermediate production. Both neutrophils and natural killer cells are the first line of cellular defence after bacterial invasion. An impairment of cellular defence might increase susceptibility to bacterial infection. H2 receptor antagonist use does not seem to impair neutrophil chemotaxis or phagocytosis.16
Infectious diseases are the most common afflictions of humans and are a major source of morbidity and mortality. Although infectious diseases affect subjects of all ages, several groups who are at increased risk for complications have been identified. The burden of respiratory infection is more pronounced in subjects with asthma and obstructive pulmonary disease, and immunocompromised and elderly patients. To lessen the impact of infections, precautionary measures are being recommended as a way of preventing infections and their complications. Patients using acid-suppressive drugs may be classified as a high-risk group for infectious diseases, in particular with extensive and severe outbreaks of epidemic diseases.
A limitation of this study is that not all approached subjects returned the questionnaire. The responders may have had more infections than those who did not respond to our request. Overall many responders reported having an infection in the winter season, when the questionnaire was sent. This is, however, not exceptional for this season of the year.17 Furthermore, the questionnaire we used was not validated. We asked subjects to report infection and medication use in the preceding month. This period might be too long for collecting this kind of information. Moreover, symptoms of infection are atypical and range from asymptomatic to severe illness, which might lead to misclassification. Taking these aspects into consideration, more studies are needed to confirm our findings that subjects using acid-suppressive drugs more often develop a community-acquired respiratory infection which leads to more physician visits and more prescriptions for antibiotics in comparison to those not using acid-suppressive drugs.