Abstract
Abstract
Purpose
The Swedish Prescribed Drugs and Health Cohort (SPREDH) is a population-based cohort based on merged data from four nationwide health data registers in Sweden. SPREDH provides opportunities to investigate how the use of various medications influences cancer risk, cancer prognosis and many other outcomes. The cohort was recently updated to include a longer follow-up, more patients and additional drugs.
Participants
SPREDH currently includes 9 454 340 users of selected medications, who have been followed up for a total of 138 015 003 person-years from 1 July 2005 to 31 December 2024, that is, for up to 19½ years.
Findings to date
SPREDH includes data from the Swedish Prescribed Drug Register, Patient Register, Cancer Register and Cause of Death Register. Available data include participants’ characteristics, use of medication, healthcare utilisation, diagnoses (including detailed information on cancers), surgical procedures and dates and causes of death. The original version of SPREDH has been used for 10 original studies published in scientific journals, primarily in the fields of gastroenterology and oncology. The updated version of SPREDH includes 1 382 698 participants who have developed a cancer during the follow-up.
Future plans
The newly updated and extended version of SPREDH enables studies with a wide range of study designs and hypotheses, especially pharmacoepidemiological studies evaluating how the use of certain medications affects the risk and prognosis of cancer and other diseases. It also allows for comparative research across classes of medications, as well as investigations of drug utilisation, safety and effectiveness.
Keywords: EPIDEMIOLOGY, Gastrointestinal tumours, Epidemiology
STRENGTHS AND LIMITATIONS OF THIS STUDY.
The Swedish Prescribed Drugs and Health Cohort is a population-based and nationwide cohort using complete and high-quality data from national health data registers in Sweden.
The study includes as many as approximately 9.5 million participants with up to nearly 20 years of complete follow-up and has strong statistical power.
The extensive information is prospectively collected and includes participants’ characteristics, medications, diagnoses, surgical procedures and mortality, and there are also possibilities to link to additional registers and databases to obtain further detailed information.
There is a lack of information on the use of medications before the study period, that is, before the Prescribed Drug Register was established (1 July 2005).
Data on some potential confounders are missing, for example, direct data regarding body mass index, tobacco smoking, alcohol consumption and dietary factors.
Introduction
The nationwide population-based health data registers, the personal identity number system and the uniform healthcare system in Sweden offer outstanding opportunities for population-based medical research and have been described as a ‘goldmine’.1 2 Sweden and other Nordic countries have established nationwide registers of prescriptions dispensed at pharmacies, providing drug exposure data that can be linked to health outcomes recorded in other health data registers or studies. The Swedish Prescribed Drug Register, established on 1 July 2005, contains almost all prescriptions in the country, corresponding to over 100 million annually.3 4 When linked to other national health data registers, this register provides a valuable resource for investigating how prescribed drugs influence various health outcomes.
The Swedish Prescribed Drugs and Health Cohort (SPREDH) was initially created using data between 2005 and 2015 from the Swedish Prescribed Drug Register combined with three other national health data registers. The cohort profile of the first version of SPREDH described a population-based cohort of over 8 million individuals who had used any of the selected common medications.5
We have recently updated SPREDH by extending the number of participants, the medication groups and the length of follow-up until 31 December 2024, that is, 9 more years. We here present the second version of SPREDH, summarise the research output from the first version and outline some future research plans.
Cohort description
Participants
The updated SPREDH includes all individuals who purchased any of the medications presented in table 1 between 1 July 2005 and 31 December 2024 as recorded in the Swedish Prescribed Drug Register. The original SPREDH included medications that are known antagonists and modulators of sex hormones or affect sex hormone levels and fertility, as well as medications with similar medical indications but no known hormonal effects. The updated cohort additionally includes a wide range of commonly used medications: (1) drugs for treating cardiovascular diseases, including antihypertensives, beta receptor blockers, calcium channel blockers and agents acting on the renin–angiotensin system; (2) antibacterial medications for systemic use; and (3) nitroimidazole derivatives. These medication groups are the three most commonly prescribed drugs in Sweden in recent years.6 All participants in SPREDH have been followed up from the date of the first dispense of any of the listed medications until death (through linkage to the Causes of Death Register) or end of study period (31 December 2024), whichever occurred first.
Table 1. Medications included in the Swedish Prescribed Drugs and Health Cohort (SPREDH).
| Medication | Anatomical Therapeutic Chemical classification code |
|---|---|
| Sex hormones and modulators of the genital system | G03 |
| Drugs for peptic ulcer and gastro-oesophageal reflux disease | A02B |
| Metformin and other medications in the treatment of diabetes | A10 |
| Drugs used against benign prostatic hypertrophy | G04C |
| Endocrine therapy | L02 |
| Non-steroidal anti-inflammatory drugs and other anti-inflammatory agents, analgesics and platelet aggregation inhibitors | M01, N02, B01AC |
| Antihypertensives | C02* |
| Spironolactone and other diuretics | C03 |
| Beta-blocking agents | C07* |
| Calcium channel blockers | C08* |
| Agents acting on the renin–angiotensin system | C09* |
| Statins and other lipid-modifying agents | C10 |
| Antibacterials for systemic use | J01* |
| Nitroimidazole derivatives | P01AB* |
Added in the recent update of SPREDH.
Data sources
SPREDH includes data from four Swedish national health data registers, which are maintained by the governmental agency the Swedish National Board of Health and Welfare (‘Socialstyrelsen’) and funded by the Swedish government as part of the national health data infrastructure: (1) Prescribed Drug Register, (2) Patient Register, (3) Cancer Register and (4) Cause of Death Register. These registers have nationwide complete coverage and high data quality. The linkages of data between these registers for the individual participant are enabled through the use of the personal identity number assigned to each Swedish resident at birth or immigration.7 All these registers contain information on participants’ characteristics (including age, sex and place of residence) and the unique personal identity number. The collected data also include information available in one of the four registers, that is, medication use, diagnoses and surgical procedures, cancer diagnosis, and dates and causes of death. The four registers are briefly presented below:
The Swedish Prescribed Drug Register contains data on all medications prescribed and dispensed in outpatient care for each individual in the entire Swedish population. Data collection was automated through connection to the administrative system of the dispensing pharmacies at the time of purchase and reporting is mandatory for all pharmacies. The medications are classified according to the Anatomical Therapeutic Chemical classification system. The defined daily doses are used to quantify the drug utilisation. The information includes date of prescribing and dispensing, product details (substance, brand name, formulation and package size), amount and dosage. Drugs sold by prescription in outpatient care, which are completely recorded in the Swedish Prescribed Drug Register, account for at least 95% of the total drug utilisation by prescription in Sweden, while a small non-recorded proportion is used during hospital care.3 4
The Swedish Patient Register covers all inpatient care in Sweden from 1987 onwards and all specialised outpatient care since 2001. Information in the Patient Register includes dates of admission and discharge, diagnoses and surgical procedures. The diagnoses are coded according to the Swedish version of the International Classification of Disease (ICD) system. Inpatient data are nearly 100% complete. Specialised outpatient data are nearly 100% complete for public healthcare, but approximately 80% overall due to under-reporting from private healthcare providers.8
The Swedish Cancer Register started nationwide in 1958 and has approximately 96% overall completeness of recording of newly detected cancer diagnoses.9 It includes data on date of cancer diagnosis, basis of diagnosis, anatomic site, histological type and tumour stage. Sites of malignancies have been coded according to different versions of the ICD system over time in the Swedish Cancer Register. However, codes according to the 7th version of ICD (ICD-7) are available for the whole registration period from 1958 onwards for consistency.
The Swedish Cause of Death Register, initiated in 1952, covers all deaths of persons who were registered in Sweden the year they died, including those who died abroad. The register has 100% completeness for date of death.10 This register also includes data on the underlying cause (primary cause) and contributing causes (secondary causes) of death, and 96% of individuals who have died have a specific underlying cause of death recorded.10
The participants in SPREDH are included from when the Prescribed Drug Register started in 2005, but information for the participants is available from the earliest possible date in the three other registers, which have a longer history. Key variables obtained from each of the four registers are presented in figure 1. Register-based research is exempt from informed consent from individual participants in Sweden.11
Figure 1. Registers contributing data to the Swedish Prescribed Drugs and Health Cohort and the major variables included. All registers have data on the personal identity numbers and participants’ characteristics.
Patient and public involvement
We collaborate with a patient partnership group with a history of oesophageal or gastric cancer. They contribute to study planning and provide valuable feedback on research priorities and research questions to ensure they are of relevance to patients and the broader public.
Findings to date
The earlier cohort
To date, the first version of SPREDH has been used for 10 published original studies on various exposures and outcomes, predominantly in gastroenterology and oncology journals.12,21 These studies are summarised in table 2.
Table 2. 10 published studies based on the first version of the Swedish Prescribed Drugs and Health Cohort.
| Year | Journal | Exposure | Main outcome | Major finding |
|---|---|---|---|---|
| 2019 | Br J Cancer | Metformin | Risk of gastric adenocarcinoma | No association |
| 2020 | Am J Gastroenterol | Metformin | Risk of oesophageal squamous cell carcinoma | Decreased risk (HR 0.68, 95% CI 0.54 to 0.85) |
| 2021 | Br J Cancer | Metformin | Disease-specific mortality in gastric adenocarcinoma | Decreased risk (HR 0.79, 95% CI 0.67 to 0.93) |
| 2022 | Cardiovasc Drugs Ther | Proton pump inhibitor plus clopidogrel | Risk of cardiovascular events | Increased risk of myocardial infarction (HR 1.23, 95% CI 1.15 to 1.32) |
| 2022 | Br J Cancer | Menopausal hormone therapy | Risk of oesophageal adenocarcinoma | Decreased risk (HR 0.78, 95% CI 0.63 to 0.97) |
| 2022 | Br J Cancer | 5α-reductase inhibitors | Risk of oesophageal and gastric cancer | Decreased risk of oesophageal squamous cell carcinoma (HR 0.49, 95% CI 0.37 to 0.65), not of oesophageal or gastric adenocarcinoma |
| 2022 | Eur J Gastroenterol Hepatol | Gastro-oesophageal reflux disease | Risk of atrial fibrillation | Increased risk in younger individuals within 1 year of gastro-oesophageal reflux diagnosis (HR 1.55, 95% CI 1.27 to 1.88) |
| 2022 | J Gastroenterol | Discontinuation of proton pump inhibitor | Risk of gastric and oesophageal adenocarcinoma | Decreased risk of gastric adenocarcinoma (IRR 0.81, 95% CI 0.67 to 0.98) and oesophageal adenocarcinoma (IRR 0.80, 95% CI 0.68 to 0.96) |
| 2023 | J Gastroenterol | Proton pump inhibitor | Risk of pneumonia | Increased risk (IRR 1.73, 95% CI 1.71 to 1.75) |
| 2023 | JNCI Cancer Spectr | Metformin | All-cause mortality in oesophageal cancer | Decreased risk (HR 0.86, 95% CI 0.75 to 1.00) |
IRR, incidence rate ratio.
The studied exposures include the medications metformin, menopausal hormone therapy, 5α-reductase inhibitors, proton pump inhibitors and clopidogrel and also gastro-oesophageal reflux disease. The outcomes were risk of upper gastrointestinal cancers, cardiovascular diseases and pneumonia, as well as mortality in patients who developed upper gastrointestinal cancer.
Among four studies that examined metformin, one study found that metformin users had a decreased risk of oesophageal squamous cell cancer,13 and another study found no reduced risk of gastric adenocarcinoma in metformin users.12 Two other studies found that metformin use was associated with improved survival in patients with a diagnosis of oesophageal and gastric cancer.14 21
The updated cohort
The updated cohort consists of 9 454 340 participants who have been followed up for a total of 138 015 003 person-years. The maximum length of follow-up is 19½ years. The cohort includes slightly more women (52.8%) than men (47.2%). The age at cohort entry ranged from 0 to 104 years. Participants were recruited primarily in 2005 (n=3 704 821; 39.2%) when the Prescribed Drug Register started, and in the 5 years from 2006 to 2010 (n=3 809 691; 40.3%), while the remaining 1 949 828 (20.6%) were recruited in 2016–2024 (table 3).
Table 3. Characteristics of participants in the current version of the Swedish Prescribed Drugs and Health Cohort.
| Participants, n (%) | Person-years | |
|---|---|---|
| Total | 9 454 340 (100.0) | 138 015 003 |
| Age at cohort entry, years | ||
| 0–19 | 1 534 685 (16.2) | 25 193 313 |
| 20–44 | 3 509 671 (37.1) | 52 640 325 |
| 45–64 | 2 648 736 (28.0) | 41 555 729 |
| 65+ | 1 761 248 (18.6) | 18 625 635 |
| Sex | ||
| Male | 4 460 178 (47.2) | 62 312 816 |
| Female | 4 994 162 (52.8) | 75 702 186 |
| Year of cohort entry | ||
| 2005 | 3 704 821 (39.2) | 58 466 432 |
| 2006–2010 | 3 809 691 (40.3) | 63 339 640 |
| 2011–2015 | 988 753 (10.5) | 11 434 792 |
| 2016–2020 | 610 699 (6.5) | 4 066 933 |
| 2021–2024 | 340 376 (3.6) | 707 204 |
Among all participants in SPREDH, 1 382 698 developed any cancer during the follow-up, which is almost two times the number of the original cohort (n=787 776). The most commonly diagnosed cancer types are those of digestive organs (16.0%), non-melanoma skin tumours (14.1%), prostate (13.3%), uterus (11.1%), breast cancer (10.5%) and melanoma (8.3%). Particularly, SPREDH includes a large number of patients with cancers in sex hormone dependent organs, including 144 635 patients with breast cancer (143 713 female patients and 922 male patients), 154 015 with uterine cancer, 14 933 with ovarian cancer, 183 616 with prostate cancer and 5240 with testicular cancer (table 4). There are more male cases of oesophageal and gastric cancers than female cases, particularly for oesophageal adenocarcinoma, where 81.6% are men (table 5).
Table 4. Malignancies occurring during follow-up in the Swedish Prescribed Drugs and Health Cohort.
| Tumour site | ICD-7 codes | Number (%) |
|---|---|---|
| All | 140–209 | 1 382 698 (100) |
| Lip, oral cavity and pharynx | 140–148 | 21 085 (1.5) |
| Digestive organs | 150–157 | 221 327 (16.0) |
| Oesophagus | 150 | 9112 (0.7) |
| Stomach | 151 | 16 708 (1.2) |
| Colon, rectum and anus | 153–154 | 144 242 (10.4) |
| Liver and biliary tract | 155 | 20 427 (1.5) |
| Pancreas | 157 | 23 163 (1.7) |
| Lung (including trachea and bronchus) | 162 | 75 565 (5.5) |
| Breast | 170 | 144 635 (10.5) |
| Uterus | 171, 172, 174 | 154 015 (11.1) |
| Ovary | 175.0 | 14 933 (1.1) |
| Prostate | 177 | 183 616 (13.3) |
| Testis | 178 | 5240 (0.4) |
| Bladder | 181.0 | 49 543 (3.6) |
| Kidney | 180 | 24 433 (1.8) |
| Melanoma | 190 | 114 201 (8.3) |
| Other skin tumours | 191 | 194 637 (14.1) |
| Thyroid | 194 | 9758 (0.7) |
| Hodgkin lymphoma | 201 | 3472 (0.3) |
| non-Hodgkin's lymphoma | 200, 202 | 36 238 (2.6) |
| Leukaemia | 204–207 | 31 577 (2.3) |
ICD-7, 7th version of the International Classification of Diseases.
Table 5. Oesophageal and gastric cancer occurring during follow-up in the Swedish Prescribed Drugs and Health Cohort.
| Cancer type | Total, n, (%) | Men, n (%) | Women, n (%) |
|---|---|---|---|
| Oesophageal adenocarcinoma | 4714 (100) | 3845 (81.6) | 869 (18.4) |
| Oesophageal squamous cell carcinoma | 3280 (100) | 1927 (58.7) | 1353 (41.3) |
| Gastric adenocarcinoma | 12 988 (100) | 8152 (62.8) | 4836 (37.2) |
Future plans
With more participants, extended follow-up and broader medication coverage, there are opportunities for a wide range of research examining numerous hypotheses and using various study designs. There are especially good possibilities for large pharmacoepidemiological studies evaluating if and how certain medication use influences the risk of cancer and other chronic diseases, and also comparative research across drug classes and specific medications, and assessment of drug utilisation, safety and effectiveness. In addition, linkage with the Swedish Cancer Register and the Cause of Death Register provides excellent opportunities for studies on cancer outcomes.
Strengths and limitations
SPREDH is a population-based and nationwide Swedish cohort using complete and high-quality data from national health data registers. With 9.5 million participants, nearly 20 years of follow-up and 138 million person-years under study, the cohort has strong statistical power for many potential studies, not the least for pharmacoepidemiological studies assessing cancer risk and prognosis. There are also possibilities to link SPREDH to additional registers and databases to obtain further detailed information using the unique personal identity number.
Limitations include potential misclassification of exposures due to unavailable medication data before 2005, lack of lifestyle data (eg, direct information regarding body weight, tobacco smoking, alcohol consumption and dietary habits) and missing data on over-the-counter or hospital-only drug use. These limitations are common in register-based research, but can often be addressed through careful study design.22 23 Exposure-response analyses and time-latency analyses may be helpful, and it is often possible to use proxies for missing variables, for example, obesity diagnosis or type 2 diabetes as a proxy for obesity, tobacco-related diagnoses as a proxy for tobacco smoking, and alcohol-related diagnoses as a proxy for alcohol overconsumption. In one of our previous studies, we also used a rule-out approach to testify that the observed reduced risk of oesophageal squamous cell carcinoma among metformin users was unlikely to be explained by unmeasured or residual confounding.13 In addition, because a prescription is valid for a maximum of 12 months in Sweden, it is possible to identify ‘incident’ (new) users of medications by excluding those with a first known prescription in the first year of the cohort establishment, that is, before July 2006. Sensitivity analyses conditional on various aspects of the study design can also help validate the robustness of the findings.
Through the Swedish personal identity number, integration with other data, including genetic data, is possible for selected subgroups with available biobank materials. However, such linkages require specific ethical approvals and are not part of the current SPREDH infrastructure.
Collaboration
All data from SPREDH presented in this article are stored on a safe server at Karolinska Institutet, Sweden, and handled confidentially according to current laws and regulations. Only the research team has access to the data. Researchers interested in collaboration for research studies using SPREDH are welcome to contact the corresponding author, S-HX (shaohua.xie@ki.se).
Footnotes
Funding: Supported by funding from the Swedish Cancer Society (grant numbers 222038 and 243414). The grant provider had no role in the design, conduct, analysis or reporting of the cohort.
Prepublication history for this paper is available online. To view these files, please visit the journal online (https://doi.org/10.1136/bmjopen-2025-104455).
Data availability free text: Data sharing is restricted, and approvals are required from the relevant authorities. Researchers interested in collaboration are encouraged to contact the corresponding author, S-HX (shaohua.xie@ki.se).
Patient consent for publication: Not applicable.
Provenance and peer review: Not commissioned; externally peer reviewed.
Patient and public involvement: Patients and/or the public were involved in the design, or conduct, or reporting, or dissemination plans of this research. Refer to the Cohort description section for further details.
Ethics approval: The first version of SPREDH was approved in 2016 by the Regional Ethical Review Board in Stockholm, Sweden (reference number 2016/982-31/4). Subsequent approvals included cancer prognosis and mortality as studied outcomes (reference number 2018/271-32) and inclusion of additional medications (reference number 2024-01293-02). The governmental agency managing the registers, the Swedish National Board of Health and Welfare (Socialstyrelsen), approved the data disclosure.
Data availability statement
Data are available upon reasonable request.
References
- 1.Calltorp J, Adami HO, Aström H, et al. Country profile: Sweden. Lancet. 1996;347:587–94. doi: 10.1016/s0140-6736(96)91278-7. [DOI] [PubMed] [Google Scholar]
- 2.Thygesen LC, Ersbøll AK. When the entire population is the sample: strengths and limitations in register-based epidemiology. Eur J Epidemiol. 2014;29:551–8. doi: 10.1007/s10654-013-9873-0. [DOI] [PubMed] [Google Scholar]
- 3.Wettermark B, Hammar N, Fored CM, et al. The new Swedish Prescribed Drug Register--opportunities for pharmacoepidemiological research and experience from the first six months. Pharmacoepidemiol Drug Saf. 2007;16:726–35. doi: 10.1002/pds.1294. [DOI] [PubMed] [Google Scholar]
- 4.Wallerstedt SM, Wettermark B, Hoffmann M. The First Decade with the Swedish Prescribed Drug Register - A Systematic Review of the Output in the Scientific Literature. Basic Clin Pharmacol Toxicol. 2016;119:464–9. doi: 10.1111/bcpt.12613. [DOI] [PubMed] [Google Scholar]
- 5.Xie SH, Santoni G, Mattsson F, et al. Cohort profile: the Swedish Prescribed Drugs and Health Cohort (SPREDH) BMJ Open. 2019;9:e023155. doi: 10.1136/bmjopen-2018-023155. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.(Socialstyrelsen) TNBoHaW Statistics on medicines (Statistik om läkemedel) [17-Apr-2025]. https://www.socialstyrelsen.se/statistik-och-data/statistik/alla-statistikamnen/lakemedel/ Available. Accessed.
- 7.Ludvigsson JF, Otterblad-Olausson P, Pettersson BU, et al. The Swedish personal identity number: possibilities and pitfalls in healthcare and medical research. Eur J Epidemiol. 2009;24:659–67. doi: 10.1007/s10654-009-9350-y. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Ludvigsson JF, Andersson E, Ekbom A, et al. External review and validation of the Swedish national inpatient register. BMC Public Health. 2011;11:450. doi: 10.1186/1471-2458-11-450. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Barlow L, Westergren K, Holmberg L, et al. The completeness of the Swedish Cancer Register: a sample survey for year 1998. Acta Oncol. 2009;48:27–33. doi: 10.1080/02841860802247664. [DOI] [PubMed] [Google Scholar]
- 10.Brooke HL, Talbäck M, Hörnblad J, et al. The Swedish cause of death register. Eur J Epidemiol. 2017;32:765–73. doi: 10.1007/s10654-017-0316-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Ludvigsson JF, Håberg SE, Knudsen GP, et al. Ethical aspects of registry-based research in the Nordic countries. Clin Epidemiol. 2015;7:491–508. doi: 10.2147/CLEP.S90589. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Zheng J, Xie SH, Santoni G, et al. Metformin use and risk of gastric adenocarcinoma in a Swedish population-based cohort study. Br J Cancer. 2019;121:877–82. doi: 10.1038/s41416-019-0598-z. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Wang QL, Santoni G, Ness-Jensen E, et al. Association Between Metformin Use and Risk of Esophageal Squamous Cell Carcinoma in a Population-Based Cohort Study. Am J Gastroenterol. 2020;115:73–8. doi: 10.14309/ajg.0000000000000478. [DOI] [PubMed] [Google Scholar]
- 14.Zheng J, Santoni G, Xie SH, et al. Improved prognosis in gastric adenocarcinoma among metformin users in a population-based study. Br J Cancer. 2021;125:277–83. doi: 10.1038/s41416-021-01408-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Holmberg D, Mattsson F, Xie S, et al. Risk of gastric and oesophageal adenocarcinoma following discontinuation of long-term proton-pump inhibitor therapy. J Gastroenterol. 2022;57:942–51. doi: 10.1007/s00535-022-01930-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Maret-Ouda J, Santoni G, Xie S, et al. Objectively confirmed gastroesophageal reflux disease and risk of atrial fibrillation: a population-based cohort study in Sweden. Eur J Gastroenterol Hepatol. 2022;34:1116–20. doi: 10.1097/MEG.0000000000002419. [DOI] [PubMed] [Google Scholar]
- 17.Maret-Ouda J, Santoni G, Xie S, et al. Proton Pump Inhibitor and Clopidogrel Use After Percutaneous Coronary Intervention and Risk of Major Cardiovascular Events. Cardiovasc Drugs Ther. 2022;36:1121–8. doi: 10.1007/s10557-021-07219-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Rabbani S, Santoni G, Lagergren J, et al. Use of anti-androgenic 5α-reductase inhibitors and risk of oesophageal and gastric cancer by histological type and anatomical sub-site. Br J Cancer. 2022;127:892–7. doi: 10.1038/s41416-022-01872-w. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Xie SH, Santoni G, Lagergren J. Menopausal hormone therapy and risk of oesophageal adenocarcinoma in a population-based cohort study. Br J Cancer. 2022;126:129–33. doi: 10.1038/s41416-021-01575-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Maret-Ouda J, Panula J, Santoni G, et al. Proton pump inhibitor use and risk of pneumonia: a self-controlled case series study. J Gastroenterol. 2023;58:734–40. doi: 10.1007/s00535-023-02007-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Wang QL, Santoni G, Lagergren J. Diabetes, metformin use, and survival in esophageal cancer: a population-based cohort study. JNCI Cancer Spectr . 2023;7:pkad043. doi: 10.1093/jncics/pkad043. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Olsen J. Register-based research: some methodological considerations. Scand J Public Health. 2011;39:225–9. doi: 10.1177/1403494811402719. [DOI] [PubMed] [Google Scholar]
- 23.Nørgaard M, Ehrenstein V, Vandenbroucke JP. Confounding in observational studies based on large health care databases: problems and potential solutions - a primer for the clinician. Clin Epidemiol. 2017;9:185–93. doi: 10.2147/CLEP.S129879. [DOI] [PMC free article] [PubMed] [Google Scholar]