Trends in Proton Pump Inhibitor Use in Sweden by Sex and Age: A Drug Utilisation Study

Proton pump inhibitors (PPIs) are globally among the most utilised and prescribed drugs [1]. They were first launched in Sweden in 1988 (FDA approved in 1989) as primary treatment for gastric acid-related disorders and peptic ulcer prophylaxis [2]. Sweden was the first country making (low-dose) PPIs available without prescription in 1999, and despite initial resistance, many countries followed including the UK and the USA [2‐4].

A recent systematic review estimated that approximately one-fourth of adults use PPIs globally, ranging between 4 and 33%, and reported to increase over time in several settings [5]. The variety in PPI utilisation definitions is high from incident users to long-term maintenance users, defined as at least 6, 12 or 36 months’ exposure [5‐11]. A meta-analysis assessed that 25–70% of prescribing is without an appropriate indication, with common inappropriate durations and excessive dosages and rebound acid hypersecretion complicating discontinuation [12‐16]. Many PPI users become maintenance users while the recommended treatment duration for the standard gastro-intestinal indications is 2–12 weeks [13, 17‐19]. Proton pump inhibitors may also be prescribed as gastro-protectant, to prevent peptic ulcers, when using non-steroidal anti-inflammatory drugs (NSAIDs), corticosteroids, anti-coagulants/antithrombotic treatment or selective serotonin reuptake inhibitors (SSRIs) [20, 21]. Their widespread use is nevertheless insufficiently supported by evidence, and it is recommended that Helicobacter pylori screening (and eradication) is performed before long-term NSAID treatment [21]. Gastroprotection for anti-thrombogenic treatment should be used only if additional risk factors for gastrointestinal bleeding are present [20]. In children, PPIs are often used off-label, and quality of the underlying evidence has been questioned for several age-groups and indications, particularly neonates [22, 23]. In infants, PPIs are often administered for gastro-oesophageal reflux, particularly in neonatal intensive care units (NICUs), despite uncertain clinical benefit and increased risk of infections and necrotising enterocolitis [24‐26]. One-year PPI-treatment is recommended after oesophageal atresia surgery [27], despite questionable efficacy and potential increased risks of eosinophilic oesophagitis and microbiome disturbances [28‐31].

Proton pump inhibitors have been saving lives and have increased the quality of life for many, at least in the short term, but there is room for optimising prescribing. Suboptimal prescribing contributes to increased healthcare costs [32, 33] and is not without risk. Proton pump inhibitors have being associated with microbiome disruptions [28, 34, 35], and various adverse events in both adults and children including, among others, increased risks of infections (gastro-intestinal), cancer, kidney failure, adverse outcomes during pregnancy and neurodevelopmental and neurological disorders [26, 28, 36‐47]. Looking at life-course epidemiology methodology, we suspect a sensitive period effect, in which PPI use may be more harmful during certain phases of life including infancy and in utero—compared to an accumulation model where higher cumulated doses would be most harmful (dose-response effect), irrespective of the life-period of exposure [48]. Treatment response to PPIs and risk of side effects are also reported to be sex dependent [49].

Therefore, the aim of this study was to explore population-level trends in prescribed PPIs in Sweden, in different age groups and by sex; and to assess if potential PPI indications and alternative therapies followed the same trends.

This descriptive nationwide drug utilisation study is based on the Swedish Prescribed Drug Registry, which was established in July 2005 [50]. This registry contains information on all prescribed and dispensed out-patient drugs, and is nationwide complete [50]. As the study is based on the publicly available Statistical Database, containing anonymised and aggregated statics, there was no need for an informed consent or ethical approval. All prescribed and dispensed drugs are categorised according to the anatomic therapeutic chemical (ATC) classification developed by the WHO. Data on PPI exposure and other relevant drugs were extracted per calendar year (2006–2023) and differences by age, sex and trends over time were explored.

The number of PPI users and prescriptions increased steadily over time, for children and adults, and both sexes (Fig. 1, Online Fig. A.1), from 69.8 to 98.8 users per 1000 inhabitants overall. In 2023, approximately 10.6% of adult men and 14.4% of adult women were using prescribed PPIs, and 1.0% of boys and 1.5% of girls (aged < 20 years) (Fig. 1a–b).

Proton pump inhibitor use is clearly and consistently more common in women in all age-groups, except the group younger than 5 years which is dominated by boys (Fig. 1c–d). The use has increased in all paediatric and adult age-groups. The number of paediatric users and prescriptions doubled. In adults, the number of users increased to 41.4%, and the number of prescriptions to 65.9% (Fig. 1a–b). In adults, the steepest (and continuing) increase was seen among the two oldest age groups, 65–84 years, and > 85 years, responsible for the largest absolute number of prescriptions (Fig. 1c–d, Fig. A.1). In children, the increase seems to level off in the 15- to 19-year age group, with the number of users still increasing in the younger groups (Fig. 1c–d). Increasing age is correlated with a higher prevalence of PPI use, with 5.3/1000 girls and 6.0/1000 boys aged < 5 years (in 2023), compared to 389/1000 women and 370/1000 of men aged > 85 years (Fig. 2, Online Table A.3). The average annual number of prescriptions was highest for the oldest individuals (aged ≥ 85 years) with 9.5 (men) and 11.6 (women) prescriptions, based on 2023 data. In young children (aged < 5 years), an annual average of 2.5 prescriptions per user was recorded (Fig. 2).

Proton pump inhibitor use has increased drastically in Sweden since 2006, and is more common in women than men in all age groups apart from small children (aged < 5 years). Alternative therapies for reflux or ulcers were rarely prescribed in Sweden (~1% of all prescriptions in 2023). Omeprazole is clearly the most common PPI, representing almost 65% of PPI use in adults; while esomeprazole became more popular in children, representing half of the prescriptions by 2023. Over the study period (2006–2023), PPI use has doubled in children; and increased by approximately 50% in adults; with the use still increasing in children aged < 10 years, and adults aged ≥ 65 years. The average annual number of PPI prescriptions per user is high among elderly (aged 85+ years) with 9.5 and 11.6 dispensed prescriptions in men and women, respectively, suggesting mainly maintenance use.

As acid rebound is common after PPI treatment cessation, this may contribute to the global increase in maintenance users, contradictory to the treatment guidelines for most gastro-intestinal indications indicating short-term use [5, 12]. Yet, several guidelines and educational interventions are also available to help deprescribing, including reducing doses, stopping and going to on-demand use [15, 17, 55]. Interestingly, gastro-intestinal diagnoses have decreased in Sweden over time, and annual Patient Registry gastro-intestinal diagnoses only explain a minority of PPI use and do not explain the female predominance among PPI users. Potential ulcerogenic drugs were common, and NSAIDs, SSRIs and corticosteroids were used more frequently in women. Anti-thrombotic agents were mostly prescribed in (older) men. Yet, as both databases were not linked, we cannot assess if these prescriptions occurred among PPI users.

The main strength of this project is the large and valid Prescribed Drug Registry, which automatically collects data on all prescribed and dispensed outpatient drugs in Sweden [50]. This registry is entirely digitalised, and therefore we have an additional year of data compared to the Patient Registry, for which several checks and merges are needed before becoming available for analyses [52]. The Swedish and other Nordic Drug registries are unique resources for pharmaco-epidemiological research, as to date, very few countries have high-quality nationwide registries linking prescribed drug use to health data [56]. Even with publicly available data, we were able to assess differences by age, sex and calendar year, PPI subtypes and estimate the average number of prescriptions per user, and compare trends of different drug classes and diagnoses. The validity of annual reporting of the Patient Registry seems suboptimal for chronic diseases, as not everyone needs hospitalisation or outpatient care, and chronic diseases may not always be recorded for each visit. Yet, we did expect a higher coverage of gastro-intestinal diagnosis as these annual numbers only explain a fraction of PPI use. This may suggest widespread gastro-intestinal diagnostics and PPI initiation in primary care, PPI initiation without valid diagnoses, regular maintenance therapy without at least annual evaluation in specialist outpatient care, and/or a large proportion of PPI use for gastroprotection. Unfortunately, there is no nationwide primary-care patient registry available in Sweden.

Should we have ordered individual-level data from the National Board of Health and Welfare, which is crucial for association studies, we could have investigated the presence of these diagnoses over a longer period, e.g., 10 years. However, the exact indications are not recorded in the Drug Registry (as this information is not provided to the pharmacy and registry), which complicates linking Drug data to actual diagnoses. By using public data, we also lacked DDD information [57, 58], and could only calculate the average number of dispensed prescriptions per user, and not estimate the average number of days exposed as various package sizes are available in Sweden. With a maximal 3-month supply per prescription in Sweden [59], the 2.5 prescriptions on average in small children may still reflect up to 7.5 months of exposure. To our knowledge, there were no changes in package sizes over time, which may explain any differences in the number of prescriptions.

As we did not have individual level data, we could not assess PPI utilisation in specific populations, e.g., pregnancy or individuals with peptic ulcers. We also had to define our paediatric cohort as everyone aged < 20 years, as age was only provided in 5-year groups, and we could not investigate infants separately. Furthermore, we lacked information on over-the-counter PPI use, available at a higher price, and smaller packages in pharmacies and larger grocery stores. It has been estimated that 84% of the Swedish drug utilization is prescribed in the outpatient setting (77% of the total expenditure), with the remaining part being in-hospital use and over-the-counter consumption, yet exact data of specific classes such as PPIs were not provided [50].

Our Swedish results are not necessarily generalisable to other countries, as drug availability and pricing/reimbursement, demographics and prescription practices and guidelines may vary.

We did find few comparable nationwide PPI utilisation studies. An Icelandic study (2003–2015) also described a “considerable” increase in overall outpatient PPI use among adults, especially among older adults, and warned for increasing treatment durations [8]. They also described a high concurrent use of NSAIDs and anti-thrombogenic drugs [8]. A Danish study (2002–2014) also reported on an increased use in adults, particularly of maintenance users and particularly among the oldest group, with only a moderate increase of ulcerogenic drugs [7]. The Danish proportion of prescriptions of omeprazole, pantoprazole and lansoprazole seemed rather equal, while esomeprazole was rarely used [7], while Sweden is still clearly dominated by omeprazole, followed by esomeprazole. The Danish and Icelandic PPI cohorts were also clearly dominated by women in each (adult) age category [7, 8]. Another Danish study on children (2000–2015) reported a markable PPI use increase (8-fold increase of days exposed to PPIs), although duration of use was usually rather short (< 1 month) [60]. A French study (2009–2019) on paediatric PPI use (5.3% of the paediatric population), reported an important increase over time, especially among infants, yet they also reported an effect of the introduced prescription guidelines [61]. An Irish study (2009–2018) reported an increase of antacids in infants, but mainly in H2-receptor antagonists [62].

The higher prevalence of PPI use in women is intriguing as it cannot be explained by the typical gastro-intestinal indications such as peptic ulcers and gastro-oesophageal reflux. We suspect NSAID, SSRI and corticosteroid use are likely culprits, as these were more common in women in almost all age groups. In our previous research, 1.5–3% of women used PPIs during pregnancy in Sweden [42‐44, 63], which could only explain part of the higher prevalence in women of reproductive age. In older age groups, PPIs are often prescribed in combination with antithrombotic drugs, but these are more commonly used by men.

We are particularly worried about the doubling of PPI use in children, as off-label and unlicensed use seem rampant [64, 65]. There are concerning reports of the effect of PPIs on the microbiome [28, 34, 35]. Especially the early life period (first 2–3 years) is of concern, as the microbiome is acquired and a healthy microbiome should be established with potential long-term health consequences [66‐68]. Proton pump inhibitors are not recommended for infants, and only approved for certain conditions in children aged ≥ 1 year [69]. Childhood PPI use is often off-label for chronic cough or suspected gastroesophageal reflux disease (GERD), which can be difficult to diagnose and may have little benefit [23, 70‐72].

It seems that evidence-based guidelines are needed for PPI prescription for different indications and age groups, and that PPI stewardship programmes are warranted to avoid unnecessary treatment and to decrease duration of treatment [15, 61, 73‐75]. While sex difference in efficacy and safety in drug use remain largely under-investigated in general, and also for PPIs, we urge for sex-stratified analyses in PPI research, from utilisation and healthcare access studies to association and mechanistic studies [49, 76‐81].

Proton pump inhibitor utilisation and potential underlying indications vary markedly between women and men, and by age, and the main gastro-intestinal indications do not explain the female predominance and increase in PPI use over time. The increasing use of PPIs in the youngest age groups, and the amount of unwarranted PPI use overall highlight the need for further investigation into the sex- and age-specific indications for the prescriptions and long-term efficacy and safety.