- Top of page
The use of proton pump inhibitors (PPIs) is widespread and increasing. In the Netherlands, the number of PPI prescriptions increased by 71% between 1999 and 2003, from 2.3 to 4.0 million on a population of 16 million inhabitants, comprising 8% of the total pharmaceutical costs in our country.1 Worldwide, the annual sales of PPIs have surpassed $10 billion.
Proton pump inhibitors are indicated for the treatment of gastro-oesophageal reflux disease (GERD), peptic ulcer and, in combination with two suitable antibiotics, for the eradication of Helicobacter pylori infection. In addition, PPIs are used for gastroprotection in patients using non-steroidal anti-inflammatory drugs (NSAIDs) or aspirin.2
For most indications other than gastroprotection and severe reflux disease, PPIs should only be used for 4–8 weeks,3–6 but it is known that use of PPIs is generally much longer.7 Some investigators have claimed that PPI withdrawal may lead to rebound acid hypersecretion, which has been used as an explanation why PPIs are used for extended periods.8 Although PPIs are well tolerated, overuse should be avoided because PPIs increase the risk of pneumonia9 and Clostridium difficile-associated disease10 and give rise to increased healthcare costs.
Proton pump inhibitors are usually prescribed as a continuous once or twice daily regimen; however, there are indications that patients tend to use PPIs as needed, titrating their dosing regimen to the experienced symptoms.11, 12 Although on-demand or intermittent use may be cost-effective in certain patients,13–15 it may reduce effectiveness especially in patients requiring chronic treatment for erosive oesophagitis.16 Little is known about the extent of PPI underuse in indications requiring continuous therapy.
We conducted a retrospective cohort study in the Integrated Primary Care Information (IPCI) database to describe persistence and adherence with PPIs under daily circumstances. As persistence and adherence may be different for different indications, we described usage patterns in patients using PPIs for GERD, non-reflux dyspepsia or eradication of H. pylori and peptic ulcers separately.
- Top of page
The source population comprised 386 002 patients with a mean follow-up of 3.4 ± 2.4 years. Half of the population (50%) was male and the mean age (±s.d.) was 35.3 (±21.9) years.
The prevalence of PPI use (all indications) per calendar year increased by more than 130% from 2.5/100 person-years (95% CI: 2.4–2.7) in 1996 to 5.8/100 person-years (95% CI: 5.6–5.9) in 2003.
During the study period, 17 813 persons started PPIs, 16 311 (92%) of them were prescribed fixed dosing regimens. In this cohort of 16 311 new PPI users, the most frequent indications were GERD (27%), non-reflux dyspepsia (25%) and H. pylori-associated indications (15%). In 21% of patients, PPIs were given for the prevention or treatment of NSAID- or aspirin-related gastrointestinal complications. About 6% of patients used PPIs for other reasons, whereas for 7% no indication was recorded. The final study cohort comprised all patients who received PPIs with fixed dosing regimens for the treatment of GERD, non-reflux dyspepsia or H. pylori-associated indications (n = 10 833). In the GERD group, 353 patients had endoscopically confirmed oesophagitis LA grade A/B and 51 oesophagitis LA grade C/D. Eighty-one patients had BE. The mean age at time of the first PPI prescription was 50.3 ± 17.1 years and 45% for the study cohort was male.
Half of the study cohort received only one PPI prescription with the percentage of persons with only one prescription being higher for dyspepsia and H. pylori indications than for GERD (Table 1).
Table 1. Persistence to and adherence with proton pump inhibitor treatment by indication
| ||Overall (n = 10 833)||Total GERD (n = 4330)||Simple reflux (n = 3845)||Oesophagitis LA grade A/B (n = 353)|
|Number of prescriptions, n (%)|
| 1||5853 (54.0)||2041 (47.1)||1939 (50.4)||76 (21.5)|
| 2||2015 (18.6)||789 (18.2)||719 (18.7)||59 (16.7)|
| 3 or more||2965 (27.4)||1500 (34.7)||1187 (30.9)||218 (61.8)|
|Persistence, % (95% CI)|
| 6 months||30.3 (29.3–31.2)||38.7 (37.1–40.3)||34.6 (32.9–36.3)||65 (60–70)|
| 1 year||22.3 (21.4–23.2)||30.6 (29.0–32.2)||26.7 (25.1–27.3)||54 (48–60)|
| 2 years||17.4 (16.5–18.3)||24.4 (22.8–26.0)||20.9 (19.3–22.5)||42 (36–48)|
|PDC (mean±s.d.)||0.76 ± 0.24||0.77 ± 0.22||0.77 ± 0.23||0.80 ± 0.19|
|Adherence overall*, n (%)|
| <20%||75 (1.5)||23 (1.0)||23 (1.2)||0 (0)|
| 20–80%||2244 (45.1)||1004 (43.9)||857 (45.0)||111 (40.1)|
| >80%||2661 (53.4)||1262 (55.1)||1026 (53.8)||166 (59.9)|
| ||Oesophagitis LA grade C/D (n = 51)||Barrett's oesophagus (n = 81)||Non-reflux dyspepsia (n = 4111)||Helicobacter pylori-associated indications (n = 2392)|
|Number of prescriptions, n (%)|
| 1||8 (15.7)||18 (22.2)||2398 (58.3)||1414 (59.1)|
| 2||5 (9.8)||6 (7.4)||805 (19.6)||421 (17.6)|
| 3 or more||38 (74.5)||57 (70.4)||908 (22.1)||557 (23.3)|
|Persistence, % (95% CI)|
| 6 months||77 (65–89)||79 (70–88)||25.2 (23.7–26.1)||23.0 (21.1–24.9)|
| 1 year||73 (60–86)||72 (61–83)||16.8 (15.4–18.2)||15.5 (13.8–17.2)|
| 2 years||65 (50–80)||69 (58–80)||12.9 (11.6–14.3)||11.3 (9.7–12.9)|
|PDC (mean± s.d.)||0.81 ± 0.16||0.87 ± 0.14||0.76 ± 0.25||0.74 ± 0.25|
|Adherence overall*, n (%)|
| <20%||0 (0)||0 (0)||28 (1.6)||24 (2.5)|
| 20–80%||20 (46.5)||16 (25.4)||773 (45.1)||467 (47.8)|
| >80%||23 (53.5)||47 (74.6)||912 (53.2)||487 (49.8)|
Overall, 22% of patients continued PPI treatment for at least 1 year, 17% for at least 2 years. Persistence was significantly higher in patients using PPIs for GERD than in patients using PPIs for dyspepsia or H. pylori-associated indications (Figure 1). Among the patients with GERD, persistence was 39% (37–40%) after 6 months, 31% (29–32%) at 1 year and 24% (23–26%) at 2 years (Table 1). Within the GERD group, patients with endoscopically confirmed oesophagitis and BE were most persistent. Within these patient groups, persistence with PPIs was higher in patients with oesophagitis LA grade C/D and BE than in patients with oesophagitis LA grade A/B (Table 1).
Among patients with non-reflux dyspepsia and H. pylori-associated indications, which are indications that usually require only short-term treatment, around 16% was still using PPIs 1 year after start and around 12% was still using PPIs 2 years after start (Table 1).
The mean percentage of days covered during the treatment duration (PDC) in patients with at least two prescriptions was 76% and was similar for all indications. The distribution of PDC levels in low, moderate and high adherence did not differ between patients treated for GERD, non-reflux dyspepsia or H. pylori. Half of the patients (53%) had an adherence level of at least 80%, but at least 45% of patients used PPIs intermittently or on-demand (Table 1). Among GERD patients, adherence did not differ between patients with an endoscopy-proved diagnosis of oesophagitis or patients without such a diagnosis, but adherence was higher among patients with BE (Table 1).
Adherence decreased rapidly in the first 6 months after treatment start, but increased after that period (Figure 2), especially the percentage of patients with a low adherence level reduced. Figure 2 shows that for each moment in time the percentage of patients with intermittent PPI use (PDC ≤80%) was quite substantial.
More comorbidity, increasing age, more co-medication, a specialist visit and alcohol abuse were associated with a reduction in the odds for discontinuation, whereas a dosing frequency of PPIs of more than once daily increased the probability of discontinuation (Table 2). Predictors for discontinuation of PPIs were quite similar across indications.
Table 2. Predictors of discontinuation of and adherence with proton pump inhibitor treatment by indication
| N (mean ± s.d.)||RRadj¶ for discontinuation (95% CI) N = 10 833||RRadj†† for adherence >80% (95% CI) N = 4980|| N (mean ± s.d.)||RRadj¶ for discontinuation (95% CI) N = 4330|| RRadj†† for adherence >80% (95% CI) N = 2289|
| Non-reflux dyspepsia||4111||1.4 (1.3–1.4)||0.93 (0.81–1.1)||–*||–*||–*|
| Helicobacter pylori-associated indications||2392||1.4 (1.3–1.5)||0.85 (0.71–1.0)||–*||–*||–*|
|Age (year)||50.3 ± 17.1||0.99 (0.99–0.99)||1.0 (1.01–1.01)||51.8 ± 16.3||0.99 (0.99–0.99)||1.0 (1.0–1.01)|
|Co-medication (number of different drugs)||1.4 ± 1.6||0.97 (0.95–0.98)||1.1 (1.0–1.1)||1.3 ± 1.6||0.95 (0.92–0.98)||1.0 (0.98–1.1)|
| 1–5||4746||0.80 (0.76–0.84)||–†||1899||0.80 (0.74–0.87)||–†|
| ≥6||972||0.76 (0.69–0.84)||–†||402||0.79 (0.67–0.93)||–†|
|Specialist visit (at least once)||1384||0.61 (0.57–0.66)||1.4 (1.2–1.6)||1818||0.51 (0.45–0.57)||1.7 (1.3–2.1)|
|Frequency of daily PPI dosing >1||2074||1.2 (1.1–1.3)||0.71 (0.60–0.83)||486||–†||0.65 (1.3–2.1)|
|Alcohol abuse||103||0.79 (0.62–1.0)||–†||34||0.60 (0.37–0.98)||–†|
|Depression||817||–†||0.79 (0.64–0.97)||304||1.2 (1.0–1.3)||–†|
|Number of PPI prescriptions§|
| 3||894||–*||0.46 (0.39–0.54)||364||–*||0.37 (0.29–0.48)|
| 4–6||922||–*||0.35 (0.29–0.41)||461||–*||0.34 (0.27–0.43)|
| ≥7||1149||–*||0.58 (0.49–0.67)||675||–*||0.67 (0.55–0.86)|
| ||Non-reflux dyspepsia||H. pylori-associated indications|
|N (mean ± s.d.)||RRadj¶ for discontinuation (95% CI) N = 4111||RRadj†† for adherence>80% (95% CI) N = 1713|| N (mean ± s.d.)||RRadj¶ for discontinuation (95% CI) N = 2392||RRadj** for adherence >80% (95% CI) N = 978|
| Non-reflux dyspepsia||–*||–*||–*||–*||–*||–*|
| H. pylori-associated indications||–*||–*||–*||–*||–*||–*|
|Age (year)||48.6 ± 17.7||0.99 (0.99–0.99)||1.0 (1.01–1.02)||50.4 ± 17.4||0.99 (0.99–0.99)||1.0 (1.01–1.02)|
|Co-medication (number of different drugs)||1.3 ± 1.6||0.94 (0.92–0.97)||–†||1.6 ± 1.7||1.0 (0.97–1.0)||–†|
| 1–5||1681||0.77 (0.71–0.84)||–†||1166||0.85 (0.77–0.94)||–†|
| ≥6||359||0.78 (0.66–0.92)||–†||211||0.74 (0.60–0.92)||–†|
|Specialist visit (at least once)||263||0.61 (0.52–0.72)||–†||520||0.74 (0.66–0.83)||–†|
|Frequency of daily PPI dosing >1||600||–†||–†||988||1.4 (1.3–1.6)||0.79 (0.58–1.1)|
|Number of PPI prescriptions§|
| 3||317||–*||0.54 (0.42–0.71)||213||–*||0.53 (0.38–0.75)|
| 4–6||295||–*||0.35 (0.27–0.47)||166||–*||0.38 (0.26–0.55)|
| ≥7||296||–*||0.51 (0.39–0.68)||178||–*||0.45 (0.31–0.66)|
In patients with GERD the odds for high adherence (PDC >80%) increased with increasing age and a specialist visit, whereas the odds of a high adherence level decreased with an increasing number of PPI prescriptions and a PPI dosing frequency of more than once daily. In patients with non-reflux dyspepsia, the probability of a high adherence level increased with increasing age and decreased with more PPI prescriptions. In patients treated with PPIs for H. pylori the odds for high adherence increased with increasing age and decreased with more PPI prescriptions and a dosing regimen of more than once daily (Table 2).
- Top of page
This study investigated the pattern of PPI use in patients using these drugs for GERD, non-reflux dyspepsia and H. pylori-associated indications. More than 75% of patients stopped PPI treatment within 1 year. The fact that 25% of patients without a proper indication for PPI maintenance therapy, such as patients with non-reflux dyspepsia and H. pylori-associated indications,3, 4 still used PPIs after 6 months points to overuse. This finding is consistent with the literature that is available.7 The highest persistence was observed in patients with oesophagitis grade C/D and BE, for which maintenance PPI treatment is indeed recommended.4 On the other hand, PPIs might be underused by this group of patients in terms of adherence.
Low or moderate adherence occurred in almost half of all patients and reflects on-demand or intermittent use rather than continuous use. Surprisingly this percentage did not differ between indication groups, except for patients with BE who were most adherent. Adherence decreased particularly if patients were prescribed a multiple dosing regimen.
The clinical consequences of low adherence are probably limited if PPIs are solely prescribed for symptom control. Adjustment of intake to symptoms may then lead to decreased drug intake and cost savings. Studies examining on-demand/intermittent PPI treatment in patients with non-erosive reflux disease (NERD) have shown adequate symptom control in the majority of patients,14, 15 and a similar rate of symptom relief and overall quality of life score in patients on continuous or on-demand treatment,13 indicating that on-demand/intermittent PPI treatment may reduce the cost of therapy in patients with NERD.
In other patient groups, the consequences of intermittent use of PPIs may be less innocent. Studies have shown that patients with severe oesophagitis relapse frequently25 and benefit from continuous PPI maintenance therapy.26 Intermittent PPI therapy may not be sufficient to prevent relapses.16 In patients with BE, PPIs have been claimed to partly prevent or slow down the progression from intestinal metaplasia to dysplasia and eventually adenocarcinoma.27, 28
To our knowledge, only one (other) study previously investigated the rate of compliance with PPIs. This study evaluated compliance by questionnaire in patients using PPIs for more than 1 year and 71% of these patients reported taking their PPIs continuously on a daily basis.12 Our study is the first to examine adherence levels in a large group of PPI users by means of prescription data and by indication.
It can be debated if persistent use of PPIs in patients with an indication for which only short-term PPI use is recommended PPI use should be referred to as overuse, because these patients were probably gaining some symptomatic benefit. The management of functional dyspepsia can be difficult and continued therapy may be desirable in some patients. On the other hand, patients can continue treatment driven by habit rather than by symptoms. Our data did not allow studying this issue in further detail.
In this study we were not able to make a distinction between patterns of intermittent use such as ‘on-demand users’, who take their medication only when symptoms occur, and patients who take a full prescription on a continuous basis after a symptom relapse, and then have a gap until the next symptom episode. It was not possible to make a distinction between GP-directed or patient-directed intermittent use of PPIs. Therefore, one cannot conclude that patients with a low adherence level were not following GP instructions on intake.
Our GP database study has some limitations. First, it contains all PPI prescriptions issued by GPs, but may miss those prescribed by specialists. This may have led to a small underestimation of PPI persistence in our population. PPI treatment is, however, often initiated by a GP, at least in the Netherlands, and in case a medical specialist initiates the treatment, a follow-up prescription will be prescribed by the GP. Only 12% of patients consulted a specialist and persistence was higher in those who did. Secondly, as our data are based on prescriptions and not pharmacy delivery, an overestimation of PPI use may result from patients not collecting their prescription. We do, however, not believe that this has led to a large overestimation of PPI use, as most PPIs are initially prescribed for symptom control. Moreover, for the adherence analyses we included only patients with at least one refill, which makes it unlikely that they never used the medication. Finally, we do not have information on the patients’ additional over-the-counter use of antacids and H2-receptor antagonists. We did, however, analyse the available information on prescribed antacids and H2-receptor antagonists in the PPI users and found no significant association with either persistence or adherence.
Strengths of this study are the large sample size and the fact that the IPCI database contains a random sample of the general population, resulting in a random sample of PPI users within the Dutch population. We used complete electronic patient records and did not have to rely on self-reporting of patients, avoiding recall bias. Furthermore, the patients in this study were not aware of the fact that their PPI use was monitored, thereby avoiding the risk of changing patient's behaviour caused by the study itself.
In conclusion, the prevalence of PPI use in the Western population is high and further increasing. More than 75% of patients that start PPI treatment stop within 1 year. Overuse in terms of extended use is seen primarily in patients with non-reflux dyspepsia and those using PPIs for H. pylori-associated indications. About half of all patients use PPIs on a non-continuous basis, indicating intermittent/on-demand use. The consequences of intermittent use are probably minor in patients treated only for symptom control and therefore may lead to decreased drug use and cost savings. Intermittent/on-demand use may, however, be inadequate for patients requiring continuous maintenance use.