Potentially inappropriate prescribing for adults living with diabetes mellitus: a scoping review

Background People living with diabetes often experience multiple morbidity and polypharmacy, increasing their risk of potentially inappropriate prescribing. Inappropriate prescribing is associated with poorer health outcomes. Aim The aim of this scoping review was to explore and map studies conducted on potentially inappropriate prescribing among adults living with diabetes and to identify gaps regarding identification and assessment of potentially inappropriate prescribing in this group. Method Studies that reported any type of potentially inappropriate prescribing were included. Studies conducted on people aged < 18 years or with a diagnosis of gestational diabetes or prediabetes were excluded. No restrictions to language, study design, publication status, geographic area, or clinical setting were applied in selecting the studies. Articles were systematically searched from 11 databases. Results Of the 190 included studies, the majority (63.7%) were conducted in high-income countries. None of the studies used an explicit tool specifically designed to identify potentially inappropriate prescribing among people with diabetes. The most frequently studied potentially inappropriate prescribing in high-income countries was contraindication while in low- and middle-income countries prescribing omission was the most common. Software and websites were mostly used for identifying drug-drug interactions. The specific events and conditions that were considered as inappropriate were inconsistent across studies. Conclusion Contraindications, prescribing omissions and dosing problems were the most commonly studied types of potentially inappropriate prescribing. Prescribers should carefully consider the individual prescribing recommendations of medications. Future studies focusing on the development of explicit tools to identify potentially inappropriate prescribing for adults living with diabetes are needed. Supplementary information The online version contains supplementary material available at 10.1007/s11096-022-01414-7.


Introduction
Inappropriate prescribing is a potential threat to patient safety [1]. It may be manifested in the form of overprescribing, misprescribing or underprescribing [2,3]. Overprescribing is the prescription of a medication that has no clear indication or prescribing two or more agents for the same purpose while a single agent is sufficient. Misprescribing includes prescribing medications at the wrong dose, duration, or frequency, not choosing the first line option or prescribing medications with significant interactions. Underprescribing involves the omission of a beneficial medication [4].
Inappropriate prescribing may result in adverse drug reactions, hospitalization, worsening of the disease and death. [4] People with two or more potentially inappropriate prescribing (PIP) indicators have been reported to have twice the risk of emergency visits and adverse events [5]. Nearly one-third of adverse drug reactions in a Swedish older population were attributed to PIP [6]. In addition to increased risk of morbidity and mortality, PIP can also increase health care utilization. A study conducted in Northern Ireland in 2010 estimated that the annual gross cost of PIP was more than €6 million [7].
The two most common factors that contribute to PIP are polypharmacy (5 or more medications) and multiple morbidity (MM) [8,9]. One of the high-risk patient groups for polypharmacy and MM are people with diabetes mellitus (DM). The prevalence of DM is increasing globally [10]. The International Diabetes Federation (IDF) reported that 463 million people had diabetes in the year 2019 and this is expected to increase to 700 million by 2045 [11]. Diabetes costs the world more than a trillion dollars currently and is estimated to cost US$2.1 trillion by the year 2030 [12]. Globally, about 5 million deaths per year are due to DM [13]. Most people with DM have MM or complications that can include hypertension, dyslipidaemia, nephropathy, neuropathy, retinopathy, cardiovascular, cerebrovascular or peripheral vascular diseases and as a result receive multiple medications. Correspondingly, PIP among people with DM has been reported to be high in a range of international studies [14][15][16][17][18].
Studies have focused on one or more types of PIP (unnecessary drug therapy, omission, contraindication, inappropriate drug selection (IDS), dosing problems, and drug-drug interactions (DDI)) often utilizing different tools for the same type of PIPs. However, no published review article mapping studies on all types of PIP among people with DM was found in our preliminary search on the Cochrane Database of Systematic Reviews, PubMed, JBI Database of Systematic Reviews and Implementation Reports, PROS-PERO, Scopus, Informit, ProQuest, EBSCO and Google scholar databases.

Extraction of results
Data extraction was performed using Colandr (Science for  Nature and People Partnership, Conservation International,  and DataKind, CA, USA), an online application for conducting systematic synthesis of evidence. Data extracted from the studies included author, year, country, publication type (e.g. original article), study design, study population, setting, sample size, types of PIP investigated, prevalence of PIP, medications involved, examples of PIP events, and criteria used in the identification of PIP. Pretesting of the data extraction form was done on a random sample of 10 articles and modified to ensure that all the required information was databases were searched from their inception to December 2020: PubMed, EBSCO, Web of Science, ProQuest, Scopus, Cochrane, EMBASE, and Informit. In addition, some grey literature sources including Open Dissertation.org, Open Access Theses and Dissertation, and BIELEFELD Academic Search Engine (BASE) were searched for any unpublished work. In a third step, manual searches of reference lists of included studies were searched for additional articles. The final search results were exported to EndNote X8.2 (Clarivate Analytics, PA, USA) and duplicates were removed. A sample search result on PubMed database is shown in the electronic supplementary material (Supplementary Table 1).

Fig. 1
Preferred Reporting Items for Systematic reviews and Meta-Analyses extension for Scoping Reviews flow diagram 41.1% of the studies. Unnecessary drug therapy was the least frequently reported PIP ( Table 2).
The most frequently studied type of PIP in high-income countries (HICs) was contraindication, while in low-and middle-income countries (LMICs) prescribing omissions were more common. Prescribing omissions were more frequently studied in the outpatient setting, whereas contraindication dominated in inpatients and community dwelling patients. The use of software and websites as reference/criteria was more frequently reported in studies that identified DDIs, while clinical practice guidelines (CPG) and explicitly listed criteria were common in studies that identified prescribing omissions and contraindications, respectively (Fig. 2).

Specific PIP events and medications involved
The most common PIP events were contraindicationsprescribing metformin in the presence of contraindications (e.g. renal failure) and prescribing medications listed in the Beers Criteria or STOPP list for older adults. Not prescribing antiplatelet and lipid lowering agents for eligible individuals were the most commonly reported prescribing omissions. Prescribing incorrect insulin or metformin doses, not adapting the dose of a medication for renal insufficiency and not keeping to the maximum daily dose recommendations were some of the dosing problems reported in the reviewed studies. Examples of specific events in each class of PIP are shown in Table 3. All PIP events reported in the included studies and medications involved are summarized captured. One study team member (MB) extracted the data and this was manually checked and verified by others (GD, JS).

Data synthesis
Data was exported to Excel then to IBM SPSS version 25 for descriptive statistical analysis. Some of the variables (e.g. year of publication, study country) were categorized into groups. Countries were categorized as high-income and low-and middle-income based on 2020-2021 World Bank classification [21]. The mapping of the included studies was undertaken in terms of the type of PIPs studied against study area, setting, year of publication, and criteria used to assess PIP. All DM related PIPs reported from each study were listed and presented in tables. Prevalence of PIP was summarized using median and interquartile range (IQR).

Study selection
As shown in the PRISMA extension for scoping review (PRISMA-ScR) flow diagram (Fig. 1), the systematic search resulted in 21,172 published articles and 490 items of grey literature. After removing duplicates, screening the titles, abstracts and full text and adding articles from the reference lists of included studies a total of 190 articles were found eligible for this scoping review.

Study characteristics
Of the 190 included studies, 75.8% were published in the last 10 years. The majority (63.7%) of studies were conducted in high-income countries. Methodologically, 74.7% had a cross-sectional nature ( Table 1). Detailed characteristics of included studies are provided in the electronic supplementary material (see Supplementary Table 2).

Criteria for assessment of PIP
Appropriateness of medication therapy was assessed by using a variety of standard references and criteria across the studies. These criteria are summarized in Table 2.
Sulphonylureas should be used with caution because of their risk of serious hypoglycaemia [35]. If the use of sulphonylureas in older individuals is the only option, short acting sulphonylureas (e.g. glipizide) are preferred.
in the electronic supplementary material (see Supplementary Table 3).

Prevalence of potentially inappropriate prescribing
Prevalence of PIP was reported from the included studies in two different ways. Some of the studies reported prevalence for the number of adults with DM in the sample and others took the total number of drug related problems (DRPs). The reported prevalence for each type of PIP varies greatly across the studies ( Table 4).

Identified gaps regarding PIP for adults with DM
The following gaps were identified in the study of PIP for adults with DM.
• PIP is less studied in low-and middle-income countries (LMICs) where the risk of PIP could be high due to less efficient systems and resource scarcities. • PIP is less studied in nursing home and community dwelling adults living with DM. • The specific events and conditions that were considered as inappropriate prescribing were inconsistent across included studies. • There are no explicit tools/criteria solely designed to identify PIP among adults with DM.

Discussion
The current scoping review included 190 studies reporting on PIP in adults living with DM worldwide. To the authors' knowledge, this review is the first to comprehensively map, identify and assess PIP in this group. The extent and degree of identification of PIP varied with the type of tool/criteria used. Even though explicit criteria are less costly to use than implicit criteria, require less clinical judgement, and can ensure a higher degree of objectivity, only a quarter of the included studies used explicit criteria. Moreover, almost all the explicit tools used were not specifically designed to detect PIP among people with DM. The majority of explicit tools were designed to identify omissions and contraindications in older populations (e.g. Beers criteria, STOPP and START criteria).

Management of hyperglycaemia
Metformin is the preferred first line pharmacologic agent for people with T2DM, unless contraindicated, because metformin is safe, effective, inexpensive and has beneficial effects  [47][48][49]. Mixed results were reported from the included studies regarding the GFR limit used to consider metformin use as inappropriate in an adult with decreased renal function. Huang et al. considered GFR < 60 [50], Lamine et al. and Diab took < 45 [51], (Diab MI, unpublished), and other studies considered a GFR of < 30 mL/min/1.73m 2 [52,53] as the contraindicated level of kidney function for metformin. This difference arose because authors used different criteria as a reference to decide on metformin contraindications.
Beta-blocker use in a diabetic patient with frequent hypoglycaemic episodes is not recommended because of the high Glyburide, a longer acting sulphonylurea, should not be used in older adults because of the increased risk of hypoglycaemia in this patient group [36]. In line with this, many studies that investigated inappropriate prescribing reported that glyburide or chlorpropamide use in an older adult with DM was contraindicated [24, 26, 27, 30-32, 34, 37-43].
Thiazolidinediones (TZDs) used in the presence of contraindication was reported in nine studies, of which six stated that it was given to adults with heart failure. According to Medscape (WebMD, NY, USA), heart failure is a 'black box warning' for thiazolidinediones because they can cause or exacerbate congestive heart failure in some patients. That is why these medications are not recommended for a patient with symptomatic heart failure and their initiation in a patient with established NYHA class III or IV heart failure is contraindicated. [44].
Metabolism and clearance of many of the hypoglycaemic agents is via the kidneys [45]. As a result, prescribing of antidiabetic medications for people with renal impairment is challenging. One of the hypoglycaemic agents that needs special emphasis in the presence of renal impairment is metformin. It is excreted unchanged through the kidneys

Management of hypertension for adults with DM
Treatment of hypertension to a blood pressure target of < 140/90 mmHg reduces microvascular and cardiovascular events in people with DM [55]. Some of the included studies [14,56], (Soorapan S, unpublished) reported that adults with diabetes and hypertension were not prescribed antihypertensive therapy. The American Diabetes Association (ADA), 2019 guideline recommends that diabetic patients with confirmed office-based BP ≥ 140/90 should be

Type of PIP Specific PIP events (examples) Unnecessary drug therapy
Prescribing -• antiplatelet (e.g. aspirin) or statin for illegible individuals • intensified antidiabetic medication for a patient with limited life expectancy or already at goal HbA1c • dual antihypertensive agents for a stage I hypertensive patient • loop diuretics in the absence of clinical signs of HF • aspirin to a patient with cardiovascular and/or coronary risk < 1.0 event/100 patients/year Abbreviations: ACEI = angiotensin converting enzyme inhibitor; AHA = American Heart Association; AKI = acute kidney injury; ARBs = angiotensin II receptor blockers; BG = blood glucose; CAD = coronary arterial disease; CCB = calcium channel blocker; CHD = coronary heart disease; CKD = chronic kidney disease; CrCl = creatinine clearance; CVD = cardiovascular disease; Cyp2C9 = cytochrome 2C9; DKA = diabetic ketoacidosis; DM = diabetes mellitus; DPP = dipeptidyl peptidase; eGFR = estimated glomerular filtration rate; ESRD = end stage renal disease; GFR = glomerular filtration rate; GLP = glucagon-like peptide; HbA1c = haemoglobin A1c; HCT = hydrochlorothiazide; HF = heart failure; HTN = hypertension; IHD = ischaemic heart disease; KCl = potassium chloride; MI = myocardial infarction; NSAIDs = non-steroidal anti-inflammatory drugs; PVD = peripheral vascular disease; SCr = serum creatinine; SGLT = sodium-glucose co-transporter; SR = sustained release; SSRI = selective serotonin reuptake inhibitor; T1DM = type 1 diabetes mellitus; T2DM = type 2 diabetes mellitus; TG = triglyceride; TIA = transient ischaemic attack; TZD = thiazolidinedione Only a few items relating to diabetes are usually included in these criteria. We strongly recommend the development of an up-to-date explicit tool that can be used to specifically address PIP for adults with diabetes. This review included a large number of studies conducted on PIP among people with DM without restrictions to the study area, year and language of publication. Not conducting a critical appraisal of included studies is a limitation of this review and may have resulted in the inclusion of lowquality studies.

Conclusion
PIP is common among people with DM. Contraindications, prescribing omissions and dosing problems were the most commonly reported PIPs. The specific events and conditions that were considered as inappropriate were inconsistent across included studies. PIP was less studied in low-and middle-income countries. There are no explicit criteria specifically designed to measure PIP for adults living with DM. Future studies focusing on the development of explicit tools to identify PIP for adults with DM are needed.
Funding No specific funding was received.

Open Access funding enabled and organized by CAUL and its Member Institutions
Conflicts of interest The authors have no conflicts of interest to declare.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons. org/licenses/by/4.0/. is the leading cause of end stage renal disease (ESRD) [59]. ACEIs and ARBs are the preferred agents for treatment of high blood pressure in people with diabetes and urine albumin to creatinine ratio (UACR) ≥ 30 or eGFR < 60 mL/ min/1.73 m 2 , as these protect against kidney disease progression [49].

Prevention and management of atherosclerotic cardiovascular disease (ASCVD)
People with diabetes have a higher risk of ASCVD as compared to non-diabetic individuals [60]. Lipid lowering therapy is recommended in people with diabetes and a prior CVD or high risk for CVD [22,23]. A meta-analysis supports the use of a high-or moderate intensity statin as a first line therapy for LDL cholesterol lowering and cardio-protection [61]. Statins have been demonstrated to reduce the risk of CV events and death in people with DM when given as primary or secondary prevention [61,62]. Consequently, not prescribing a statin therapy to an eligible adult with DM was considered as inappropriate in 28 included studies [2, 24-34, 58, 63-74], (Diab MI, unpublished; Aketchi IE, unpublished; Langenhoven W, unpublished).
People with DM and an established ASCVD should be given an antiplatelet agent as a secondary prevention strategy unless there is a clear contraindication. Most of the diabetes guidelines recommend low dose aspirin (75-350 mg) for this group of people [22,23,49]. Omission of antiplatelet therapy (e.g. aspirin) for adults with diabetes and CVD is reported by some of the studies in this review [29,73,[75][76][77][78][79], (Soorapan S, unpublished; Aketchi IE, unpublished). Use of aspirin as a primary prevention may lead to more benefit than harm in people with high risk of CVD. The ADA guideline recommends the use of aspirin as a primary prevention for people with diabetes and age ≥ 50 with at least one additional major risk factor (family history of premature ASCVD, hypertension, dyslipidaemia, smoking, or chronic kidney disease/albuminuria). Similarly, many of the reviewed studies considered omission of antiplatelet therapy (e.g. aspirin) in adults with DM who are at risk of cardiovascular disease as PIP [24-27, 29-34, 56-58, 63, 75-85], (Soorapan S, unpublished; Langenhoven W, unpublished).
Inappropriate prescribing, especially in adults with diabetes, can result in increased morbidity and mortality and increased burden on the healthcare system and costs. Therefore, strategies to prevent and resolve PIP should be sought. Using an explicit tool to identify PIP which can be used as a physicians' desk reference is a less costly and easy to use strategy to prevent PIP. Tools exist to assess PIP in older people and new tools and criteria frequently emerge. However, they do not specifically target people with diabetes.