Review article: the long-term use of proton-pump inhibitors
Dr A. Raghunath, St Andrews Group Practice, Marmaduke Health Centre, Hessle Road, Hull, E. Yorkshire, HU3 3BH, UK.
More than 15 years after the launch of omeprazole in 1988, proton-pump inhibitors remain central to the management of acid-suppression disorders and are unchallenged with regard to their efficacy and popularity among doctors and patients. They are considered safe despite early concerns about the possibility of an association with cancer and gastric atrophy; current concerns about long-term proton-pump inhibitor therapy are centred mainly on a possible association with fundic gland polyps and between Helicobacter pylori and gastric atrophic changes. Long-term proton-pump inhibitor usage accounts for the majority of the total proton-pump inhibitor usage. Long-term usage is difficult to define and most patients take proton-pump inhibitors non-continuously. Data indicate that a substantial proportion of long-term users do not have a clear indication for their therapy and there is thus room for reduction or rationalization of treatment. Overall, on-demand therapy is more cost-effective than continuous therapy and should be considered wherever possible.
Since the introduction of omeprazole in 1988, sales of proton-pump inhibitors (PPIs) have increased greatly and constitute 55% of the $20 billion spend on the world gastrointestinal market. By 2000, worldwide sales of omeprazole totalled more than $6 billion with year-on increase of nearly 10%.1 The sales of PPIs (more than $10 billion) are second only to sales of the statins.2 At the same time, the cost of managing acid-related gastrointestinal disorders continues to escalate, driven mainly by the direct cost of the drugs.3, 4 Concerns have been expressed at this seemingly alarming increase in drug expenditure, particularly in the US, Europe and Australia for indications that are mainly of benign nature, mainly gastro-oesophageal reflux disease (GERD).5–8
Long-term PPI prescribing accounts for the majority of the total PPI drug bill8–10 and questions have been raised about the appropriateness of long-term, repeat prescribing in primary care in a significant number of patients.11 Qualitative studies have shown that primary care physicians view PPIs as being very effective and relatively safe and perceive withdrawal or reduction of long-term PPIs as difficult to achieve.12–14 Some primary care physicians may also feel pressure from patients to initiate and continue PPIs.15
What is long-term PPI prescribing?
There is no consensus or agreed definition on what constitutes long-term PPI prescribing. In studies on long-term acid suppression, varying methods have been used to identify patients on long-term PPIs. The definition of long-term PPI usage has varied: from one repeated prescription over 12 months to continuous therapy for periods ranging from 4 to >12 months.16–24
Despite recommendations to patients to take PPIs on a continuous basis, there is evidence that patients take them only intermittently or on-demand.19–21 Moreover, there is now a body of evidence advocating the use of on-demand or intermittent therapy on a long-term basis.25, 26 A recent trial concluded that on-demand therapy was significantly more cost-effective than continuous therapy.27
Rates of long-term PPI use, age and sex determinants
Rates of long-term PPI usage will necessarily vary depending on how ‘long term’ has been defined. Variation in studies is also related to the year of study, geographic location, size of study population and other factors (Table 1). The majority of studies on long-term acid suppression included H2RAs and these formed the bulk of long-term repeat prescribing in the earlier studies.
Table 1. Rates of long-term PPI use
|Ryder et al.18||UK||6 months or more||60 148||0.05|
|Chen et al.28; Roberts and Bateman29||UK||All PPI prescribing in a defined period||41 GP practices||0.4*|
|Rubin et al.20||UK||>12 months||24 400||0.44*|
|Goudie et al.16||UK||Repeat prescribing register||15 495||0.2|
|Boutet et al.17||Sicily, UK||Repeat prescribing register||42 GP practices||5%†|
|Hungin et al.21||UK||6 months or more||46 650||0.45|
|Prach et al.23||UK||More than 56 days/year||35 000||NA|
|Roberts and Bateman29; Vetvick and Straand95||Norway||Daily defined units (DDU)||17 105||0.97|
|Jacobson et al.30; Ahnfeldt-Mollerup et al.96||Denmark||Constant treatment||7160||0.87|
|Hurenkamp et al.19||Holland||More than 12 weeks||46 813||0.5|
|Tsai et al.27; Chen et al.28||Taiwan||Prescription items and DDU||NA||NA|
|Jacobson et al.30||USA||More than 90 days||168 727||1.6|
|Raghunath and Hungin22||UK||6 months or more||46 933||1.7|
|Lassen et al.8||Dennmark||>180 DDU/year||470 000||1.2|
|Majumdar et al.24||USA||Equal to or more than 1 year||216 720||1.5‡|
More recent studies from Europe and North America have ascertained higher rates of use of long-term PPIs in keeping with increased volume of PPI prescribing. In individual studies, up to a sixfold variation in repeat prescribing rates between practices have been described.17, 22 The reasons for these inter-practice variations are not entirely clear but may reflect the prescribing behaviour of physicians, variations in the availability of open-access endoscopy services, demographic and other characteristics of practices.14 Evaluation of crude prescribing rates, without reference to patient demography, may be unreliable as a guide to levels of usage.
Long-term PPI usage rates increase with age29 and according to one study, a rate of >14% in the 65+-year age group was ascertained compared with under 2% in the 15–44-year age group.17 Another study showed that 65% of patients receiving long-term acid-suppressing treatment were aged ≥60 years and that 25% were >75 years.18 Lassen et al.,8 in a large population-based cohort study, showed that 83% of long-term PPI users were >50 years old. The mean ages and gender ratios in various studies are shown in Table 2.
Table 2. Long-term PPIs and demographic characteristics
|Boutet et al.17||NA||15–44 |
|Ryder et al.18||64 (4–93)||1.27:1||NA|
|Rubin et al.20||NA||1.32:1||NA|
|Hungin et al.21||60 (14–91)||0.88:1||NA|
|Vetvick and Straand95||63||1.22:1||NA|
|Hurenkamp et al.19||61||0.81:1||<0.05|
|Raghunath and Hungin22||65 (18–91)||0.78:1||<0.05|
|Jacobson et al.30||52 (21–90)||1.22:1||NA|
|Majumdar et al.24||55 ± 14||0.88:1||NA|
Who is prescribed long-term PPIs?
Table 3 shows the indications for long-term PPI prescribing from various studies. In long-term PPI users, the prevalence of GERD or related conditions such as stricture, Barrett's and hiatus hernia ranged from 29% to 59%, of peptic ulcer from 3% to 37%, non-ulcer dyspepsia from 6% to 21%, uninvestigated dyspepsia from 7% to 33% and of use of aspirin or non-steroidal antiinflammatory drugs (NSAIDS) (mostly uninvestigated) from 5% to 41%. In a study by Jacobson et al.30 only 27% of all patients on long-term PPIs had undergone endoscopy.
Table 3. Long-term PPI users and diagnostic categories
|Ryder et al.18*||25||42||8||7||8|
|Goudie et al.24*||29||27||23||21||NK|
|Boutet et al.17†||29||12||8||31||16|
|Hungin et al.21†||54||3||6||22||29|
|Vetvick and Straand95||48||37||12||7||8|
|Hurenkamp et al.19||37||29||14||20||5|
|Jacobson et al.30||NK||0.6||NK||21||NK|
|Raghunath and Hungin14||47||10||14||32||15|
|Lassen et al.30||28||23||NK||33||41|
|Majumdar et al.24||59||6||13||22||26|
Patients with peptic ulcer would benefit from Helicobacter pylori eradication therapy and the opportunity to reduce or stop use of long-term PPIs exists in some of them.31 Lassen et al.31 ascertained a reduction of 24% in mean daily use of antisecretory drugs following H. pylori eradication in patients with uncomplicated peptic ulcers who were on long-term acid suppression therapy. Eradication is also likely to offer a chance of cure in H. pylori-positive patients with non-ulcer dyspepsia on long-term PPIs.32
The 2004 NICE guidelines in England have advocated further rationalization of the use of endoscopy and suggest intermittent or on-demand PPI therapy to control dyspepsia symptoms.33
Compliance with continuous PPI therapy is poor and only a minority of patients request their prescriptions regularly. The chief factors determining whether or not patients take their PPIs is the presence or severity of symptoms and the desire to remain in personal control, together with a fear of side-effects and lack of knowledge about the drugs.34 A recent systematic review ascertained that between 26% and 71% of GERD patients can be adequately managed without the need for continuous therapy.35 A large family practice audit involving over 7000 patients on PPI therapy (mostly long-term) in 91 practices in UK ascertained it was possible in a significant number of patients to stop therapy or to change to a much less expensive regimen without adverse consequences.36 Chan et al.37 in a retrospective 2-year longitudinal study in America concluded that drug costs were more than fourfold higher (P < 0.001) when PPIs rather than H2RAs were prescribed initially although nondrug costs and health services use did not decrease.
There is a paucity of community-based studies on the symptoms and quality of life in patients on long-term continuous PPI therapy. It was ascertained that, in more than 50% of patients on continuous PPI therapy, significant symptoms of GERD and dyspepsia were still present.30 Another study obtained similar results and found that a simple patient-directed intervention was able to reduce long-term PPI use without affecting the severity of dyspepsia or quality of life.38 Patients on long-term PPIs and without H. pylori infection appear to have worse reflux symptoms and quality of life compared with those who are infected.39
Long-term continuous PPI therapy, however, appears to have a firm place in the management of Barrett's oesophagus based on current evidence. Regression of columnar and replacement with squamous epithelium has been observed in several studies.40–43 However, regression of dysplasia has not as yet been clinically observed to any significant extent44 and as yet it is unclear if long-term continuous PPI therapy prevents oesophago-gastric adenocarcinoma arising from Barrett's oesophagus. Long-term epidemiological studies are required to provide further answers in this area.
Are long-term PPIs safe?
In a large mortality study of 18 000 patients treated with omeprazole (25% on long-term therapy), the authors concluded that increases in mortality associated with treatment were due to pre-existing illness, including pre-existing severe oesophageal disease. There was no evidence of an increased risk of oesophageal adenocarcinoma in those without oesophageal mucosal damage recorded at registration.45 This is reassuring and, in particular, there was no increase in gastric cancer although the length of follow-up was insufficient to allay all fears, as the delay between vagotomy and the observed increase in gastric cancer can be more than 10 years. It was therefore recommended that surveillance would need to continue for sometime more yet.46
Long-term PPIs and gastric polyps
Over the last 18 years, over 720 million prescriptions have been issued for PPIs worldwide largely for long-term prescribing.47 Concerns were initially raised from preclinical trials with omeprazole of the development of enterochromaffin-like (ECL) cell carcinoid tumours in the oxyntic mucosa of rats secondary to hypergastrinaemia.48 However, further experiments by Carlsson et al.49 discounted the direct effect of omeprazole for such changes. Large randomized clinical trials and long-standing experience have confirmed the high efficacy and safety profile of long-term PPI therapy.50
Sustained increase in gastrin levels as a result of long-term acid suppression by PPIs have been shown in several studies.51–54 Other studies have shown that gastrin levels in most patients remain in the normal range while on long-term acid suppression treatment55–57 and more importantly, no significant changes in ECL cell numerical densities or gastric endocrine cell histology were observed. The findings of the study with the greatest experience with omeprazole of 11 years in 230 patients with a mean follow-up of 6.5 years showed only minor changes in serum gastrin and gastric histology.58 However, it is true that some stimulation of ECL cells do occur following long-term PPI therapy as indicated by increased levels of chromogranin A (an endocrine cell product)59 and it has been suggested that blood levels of this chemical could be used to survey patients on long-term PPIs.60
Long-term PPI therapy may be associated with benign fundic polyps,61–63 but these are considered rare, innocuous and reversible on stopping therapy.58, 64 Case reports have shown that typical fundic polyps with cystic dilatations can occur on long-term omeprazole therapy in the absence of H. pylori infection and normal serum gastrin levels.63
In a retrospective study of 231 complicated GERD patients on long-term treatment with PPIs, gastric polyps developed in 17 (10 males and seven females; 7.3%). The mean interval of PPI use after which an endoscopy revealed gastric polyps was 32.5 months. Endoscopy established that typical polyps were generally small (<1 cm), sessile, multiple, and whitish pink with a mottled partially translucent surface. The polyps were most often present in the proximal/mid-gastric body. None of the polyps contained dysplasia or carcinoma.62
A recent case–control study concluded that long-term PPI therapy may not be associated with the development of fundic polyps. In this retrospective 12-month study, the frequency of fundic polyps in 2251 patients without H. pylori infection receiving PPI therapy (duration of treatment at least 4 weeks) was compared with a control group of 28 096 patients who did not have H. pylori infection and were not receiving PPI therapy. Polyps were identified with an identical frequency in both groups (5.0% in the control and 5.2% in the PPI group). No significant differences were present between the groups with respect to the presence of gastritis or age or sex.65
A recent retrospective study on children on long-term omeprazole therapy found that seven of the 31 children developed gastric polyps and or nodules over a mean follow-up period of 10–48 months. There were no dysplastic changes in the gastric mucosa or polyps in any of the children. There were no significant differences between the seven children with and the 24 without polyps/nodules with respect to age, gastrin concentrations, or dose and duration of omeprazole therapy.66 Fundic polyps can also disappear after cessation of long-term PPI therapy.67
Thus, it seems highly likely that gastric polyps associated with the use of long-term PPI therapy are benign and unlikely to have any potentially serious clinical consequences. A prospective study is required however, to establish their incidence, natural history and clinical significance.
Long-term PPIs and atrophic gastritis
Concerns were first raised in 1996 by Kuipers et al.68 of increased occurrence of atrophic gastritis, itself a precursor of gastric cancer, in H. pylori-infected patients on long-term omeprazole therapy in their highly debated and somewhat controversial cohort study spanning two European countries. They studied patients from two separate cohorts who were being treated for reflux oesophagitis. Seventy-two patients were treated with fundoplication and 105 with omeprazole (20–40 mg once daily). In both cohorts, the patients were followed for an average of 5 years (range, 3–8). After fundoplication, the patients did not receive acid-suppressive therapy. The presence of H. pylori was assessed at the first visit by histological evaluation in the fundoplication group, and by histological and serologic evaluation in the omeprazole group. The patients were not treated for H. pylori infection. Before treatment and during follow-up, the patients underwent repeated gastroscopy, with biopsy sampling for histological evaluation. Among the patients treated with fundoplication, atrophic gastritis did not develop in any of the 31 who were infected with H. pylori at baseline or the 41 who were not infected; one patient infected with H. pylori had atrophic gastritis before treatment that persisted after treatment. Among the patients treated with omeprazole, none of whom had atrophic gastritis at baseline, atrophic gastritis developed in 18 of the 59 (30%) infected with H. pylori (P < 0.001) and two of the 46 (4%) who were not infected (P = 0.62). The authors concluded that patients with reflux oesophagitis and H. pylori infection treated with omeprazole were at increased risk of atrophic gastritis. Similar results have been obtained from some other studies.69, 70 In contrast, Lundell71 reported no evidence of accelerated development of atrophic gastritis in patients on long-term omeprazole, although the results have been challenged by some authors72, 73 who have ascertained that, H. pylori-infected patients did develop accelerated moderate to severe atrophy.
Other studies have found no evidence of acceleration of corpus atrophy in H. pylori-infected individuals on long-term PPI therapy for reflux disease.74–76 A recent randomized study by Kuipers et al.77 examined the progression of atrophy in H. pylori-infected subjects with reflux oesophagitis on PPI therapy. They ascertained that despite being on PPIs for ≥3 years, no progression to atrophy was observed in any subject.77 These results appear to support the decision taken by the FDA Gastrointestinal Drug Advisory Committee not to consider long-term PPI therapy as a gastric cancer risk.78 Thus, despite arguments in favour of long-term PPI therapy not accelerating or promoting gastric atrophy in the presence of H. pylori infection, several experts have advocated a test-and-treat strategy prior to instituting long-term PPI therapy or in patients already on such therapy.79 The rationale for this is evidence that there is a changing pattern of gastritis from antrum to corpus in H. pylori-infected patients given long-term PPI therapy,80 that corpus gastritis per se is associated with an increased risk for gastric cancer81 and that eradication of H. pylori under these circumstances can lead to regression and resolution of corpus predominant gastritis.77, 82
The changing patterns of H. pylori gastritis in long-standing acid suppression was documented in a recently conducted prospective, double-blind trial.83 The authors ascertained the effect on gastric histology of 12-month maintenance treatment with omeprazole in H. pylori-positive GERD patients randomly assigned to either an eradication or omeprazole-alone regime. A control group of 20 H. pylori-negative GERD patients also received omeprazole throughout the study period. Biopsies taken at baseline and at 12 months were graded ‘blind’ by a single observer according to the updated Sydney System. The 41 H. pylori-positive subjects with grade B or C oesophagitis were randomly assigned (20 to omeprazole alone, 21 to eradication) and 33 subjects completed the 12-month study. There was a significant decline in antral chronic inflammation in initially positive patients between baseline and end in both the eradication group (P = 0.035) and the omeprazole-alone group (P = 0.008). However, corpus chronic inflammation increased in the omeprazole-alone group (P = 0.0156) but decreased in the eradication group. The change towards corpus predominance between baseline and end for the omeprazole-alone group was highly significant (P = 0.0078). Furthermore, five of 11 in the omeprazole-alone group developed mild corpus atrophy, compared with none of eight who had undergone H. pylori eradication. The change in frequency of corpus atrophy between the two groups was also significant (P = 0.02). The authors concluded that in H. pylori-positive subjects with GERD, long-term acid suppression lead to a shift from antral- to corpus-predominant gastritis and an increase in corpus atrophy that could be prevented by prior eradication. It was recommended that H. pylori infection should be eradicated prior to long-term acid suppression with PPIs.
Rebound acid hypersecretion following stopping of long-term PPIs is only seen in H. pylori-negative patients; its clinical consequences are uncertain although it has been suggested that this may be responsible for rebound reflux symptoms commonly seen in clinical practice.84, 85
Long-term PPIs and enteric infections
Several studies have reported colonization and bacterial overgrowth in the stomach, jejunum and small intestine in patients on long-term PPIs occurring as a result of profound acid suppression.86–88 Despite concerns that these findings may lead to increased infections and malabsorption, such clinical consequences have not been realized. Recent studies have, however, indicated an increased incidence of community-acquired respiratory infections and pneumonia in long-term PPI users compared with non-users.89, 90 These findings need to be confirmed in further large prospective studies. In addition, the occurrence of non-H. pylori bacterial growth in the stomach, in the presence of acid inhibition in H. pylori-positive patients, leads to increased cytokine levels and a higher risk of atrophic gastritis. Why this gastric co-infection leads to atrophic gastritis is not fully elucidated but the effect of the bacteria appears to be synergistic.91
Long-term PPIs and vitamin B12 deficiency
A recent case–control study reported a significantly increased risk of vitamin B12 deficiency (odds ratio 4.45; 95% confidence interval 1.47–13.34) in chronic PPI users of >12 months compared with non-users or past and short-term users.92 In a large population-based study, a low vitamin B12 was detected in 2.5% of males aged 51–65 years, and in 75% of these individuals, atrophic gastritis with ongoing H. pylori infection was present. In this study, although a low vitamin B12 was associated with a high homocysteine level, regardless of the presence or absence of atrophic gastritis, those with atrophic gastritis and a low vitamin B12 had a significantly higher homocysteine level, and a high homocysteine level with milder levels of vitamin B12 deficiency compared with those without gastric atrophy. This suggests that the consequences of a low vitamin B12 appears earlier in those with gastric atrophy. A high homocysteine level has been associated with vascular and neurodegenerative disorders.93 Further prospective studies are required to confirm this finding.
Long-term PPIs in childhood
Although PPIs are widely used in childhood, few studies have focused on their safety. A recent study in young rats comparing PPIs with placebo revealed that PPIs led to suppressed body weight gain and reduced bone mineral density. Studies in children assessing these are needed.94
Proton-pump inhibitors are highly effective drugs that have revolutionized the management of acid–pepsin disorders over the last two decades. Although the use of long-term PPIs appear to have a high margin of safety, further long-term studies and monitoring is required to ensure that is indeed the case, particularly in the areas of upper gastrointestinal cancers and infections. Further research is required in relation to H. pylori, long-term PPIs and upper gastrointestinal systems but evidence to date suggests that the risk vs. benefit is in favour of testing and eradication in those requiring long-term treatment. The cost effectiveness of long-term PPI use has been challenged by several experts and as such family doctors are being encouraged through guidelines to consider on-demand, low-dose maintenance, or intermittent therapy in most patients.