Electronic Supplementary Material Table S1: Summary of studies not included in the systematic review

Study / Study Design / Participants/Settinga / Reason for exclusion
Ailabouni et al
2016
New Zealand / Qualitative (face-to-face semi-structured interviews) / General practitioners’ prescribing for ≥10 patients in residential care (N=10) / The study included a hypothetical patient profile and no medicines for actual patients were discontinued.
Coleman et al
2016
USA / Retrospective matched-cohort / Patients with NVAF newly initiated on rivaroxaban, dabigatran, or warfarin with a baseline CHA2DS2-VASc score ≥2, ≥2NVAF diagnosis codes, ≥6 months continuous medical and pharmacy benefit enrolment before oral anticoagulant initiation. Participants stratified in 3 categories, N=32634: Rivaroxaban (mean age 71.3 ± 11.1), Dabigatran (mean age70.9 ± 10.8), Warfarin (mean age 71.5 ± 11.3) / The study did not look at discontinuation associated with old age or in individuals with LLE and was focused on treatment persistence and discontinuation.
Garcia Rodriguez et al
2011
UK / Nested case-control / Individuals with first prescription of aspirin for secondary prevention of cardiovascular outcomes (aged between 50 to 84), N=39513 / Discontinuation associated with old age or in individuals with LLE was not an investigation in this study.
Greenspan et al
2008
USA / RCT, double-blinded, placebo-controlled, partial crossover / Individuals receiving androgen-derived therapy for non-metastatic prostate cancer (average age was mean of 71.4 ± 8.6), N=112 / The study did not investigate discontinuation as an outcome and did not concentrate on LLE related discontinuation.
Martin-Merino et al
2014
UK / Retrospective database / Patients prescribed low-dose ASA for secondary prevention of CVD and discontinued therapy for at least 90 days (mean age 68.5), N=11565 / The study investigated re-initiation rather than discontinuation as a primary outcome and did not deliberate on individuals with LLE.
McNeil et al
2016
USA / Observational / Adults with LLE (≥1 month), statin therapy for ≥3 months, recent deterioration in functional status (mean age 74.3 ± 11.5), N=244 / The primary focus of this study was to examine the burden of polypharmacy at end of life and did not investigate the discontinuation of medicines.
MekontoDessap et al
2011
France / Retrospective cohort / Patients with severe sepsis and septic shock, admitted in intensive care unit, in whom statin therapy had been discontinued or continued. Patients were stratified in 2 groups, N=76; discontinuation: (age 71.1, 61.7 – 78.7) and continuation (age 72.9, 62.3 – 79.6) / This study investigated the effect of continuation/discontinuation of statins based on its pleiotropic activity resulting in a potential benefit during sepsis rather than its role in preventing an event of a CVD.
NÍ ChrÓInÍN D et al
2015
Australia & New Zealand / Qualitative (anonymised electronic survey) / Members of Australian and New Zealand Society for Geriatric Medicine and New South Wales Geriatric Medicine Advanced Trainee (Specialist Registrar) network, N=134 / The study included case vignettes and no medicines for actual patients were discontinued.
Orkaby et al
2016
USA / Retrospective cohort / Veterans receiving warfarin for NVAF for at least 6 months, newly diagnosed with dementia, not enrolled in Medicare Advantage (mean age 79.5 ± 6.0), N=2572 / This study did not focus on the concept of LLE but medication use in the presence of multiple morbidity.
Saez et al
2016
UK / Retrospective cohort / Individuals with first prescription of aspirin for secondary prevention of cardiovascular outcomes who discontinued aspirin for at least 90 days(aged between 50 to 84), N=10453 / This study examined the outcomes associated with discontinuation and reinitiation of aspirin and did not focus on individuals with LLE.
Turner et al
2014
Australia / Cross-sectional / Individuals who attended the medical oncology unit at the Royal Adelaide Hospital (mean age 76.7), N=385 / The study does not examine discontinuation of preventive medicines in LLE.
Ostergaard et al
2014
Denmark / Retrospective database / Incident ischemic stroke patients in a Danish stroke registry, (median age 68.5), N=4670 / This study did not investigate discontinuation due to LLE.
Weimar et al
2013 / Multicenter, double-blind, randomized, placebo-controlled trial / Patients with recent non-cardioembolic ischemic stroke stratified in 4 groups: ASA+ERDP, N=7212 (mean age 65.5 ± 8.4), discontinued ASA + ERDP, N=2843 (mean age 67.4 ± 8.7), clopidogrel, N=7864 (mean age 65.8 ± 8.4), discontinued clopidogrel, N=2176 (mean age 67.5 ± 8.9) / In this study authors investigated the combination aspirin-dipyridamole compare to clopidogrel in preventing recurrent stroke and did not consider discontinuation based on LLE.
Xie et al
2016
USA / Retrospective database cohort / Patients with medical claims associated with a VTE diagnosis and who initiated warfarin within 10 days of the VTE diagnosis (mean age 60.2), N=21163 / In this study LLE in older individuals was not investigated.
Wargo et al
2013
USA / Retrospective analysis of clinical trial cohort / Patients with stage III or IV breast cancer, colorectal cancer, lung cancer and prostate cancer (average age 60.5 ± 10.9), N=484 / This study did not meet the inclusion criteria for mean age to be ≥ 65 years.
Hughes et al
2004
USA / Retrospective database cohort / Stroke survivors admitted to nursing homes for at least 6 months (aged ≥ 65 years), N=16579 / This study examined the discontinuation of anticoagulants and identified factors related to advanced age and deteriorating comorbid conditions however did not investigate the context of LLE.
Todd et al
2014
England / Prospective cohort study / Patients attending a specialist palliative care centre (mean age 70), N=132 / This study discusses the discontinuation of inappropriate medicines, including preventive medicines, and provides recommendations but does not investigate actual patients discontinuing medicines.
Legault et al
2011
Canada / Prospective clinical audit / Consecutive patients admitted to palliative care unit, N=127 / The mean age of participants was not available. The author was contacted with no response.
Mullvain et al
2015
USA / Retrospective / Patients with brain metastases receiving whole brain radiation therapy, N=206 / The mean age of the total study population was not available. The author was contacted with no response.

aAge reported in years

N/A – not applicable; NVAF – non-valvular atrial fibrillation; LLE – limited life expectancy; RCT – Randomised Controlled Trial; ASA – acetyl salicylic acid; ERDP – extended-release dipyridamole; VTE – venous thromboembolism