Research

Su Wood, Duncan Petty, Liz Glidewell and DK Theo Raynor

Application of prescribing recommendations in older people with reduced function: a cross-sectional study in general practice

INTRODUCTION Abstract stay were associated with reduced kidney Prescribing when a patient’s kidney function function.7 Background is reduced should include consideration The National Institute for Health and Kidney function reduces with age, increasing of the risk from increased blood levels Care Excellence (NICE) guidance on chronic the risk of harm from increased blood levels of 8 many medicines. Although estimated glomerular when drugs excreted renally are eliminated kidney disease (CKD) recommends ‘review filtration rate (eGFR) is reported for prescribing more slowly.1 With ageing, a progressive of ’, and now the NICE acute decisions in those aged ≥65 years, creatinine loss of functional capabilities in most body kidney injury guidelines9 and alerts require clearance (Cockcroft–Gault) gives a more accurate organs, changed responses to receptor GPs to review medicines after an acute estimate of kidney function. stimulation, and decreased homeostatic kidney injury episode. These include drugs Aim mechanisms have implications for drug that are a direct risk to the kidney, such To explore the extent of prescribing outside handling. Of these changes, excretion is as non-steroidal anti-inflammatory drugs recommendations for people aged ≥65 years with reduced kidney function in primary care and the most significant and important age- (NSAIDs), angiotensin-converting enzyme to assess the impact of using eGFR instead of related pharmacokinetic change, and is inhibitors (ACEIs) and angiotensin II creatinine clearance to calculate kidney function. both predictable and measurable. Two- receptor blockers (ARBs); and drugs that Design and setting thirds of people aged 70–80 years have are more likely to cause adverse drug A cross-sectional survey of anonymised approximately half the kidney function of a reactions when blood levels are increased prescribing data in people aged ≥65 years from all ‘young adult’,2,3 with an average decline of as a result of slower elimination, such 80 general practices (70 900 patients) in a north of England former primary care trust. approximately 1 ml/minute annually after as metformin, gabapentin, and pregabalin. 30 years old, decreasing more rapidly after Reduced kidney function can also reduce Method 65 years of age.4 the effectiveness of some drugs, such as The prevalence of prescribing outside recommendations was analysed for eight Harm and hospitalisation are potential thiazides and nitrofurantoin. exemplar drugs. Data were collected for age, outcomes when recommendations for Estimated glomerular filtration rate sex, actual weight, serum creatinine, and altering prescribing are not applied. In other (eGFR) has been reported with pathology eGFR. Kidney function as creatinine clearance European countries, a large primary care results since the introduction of the CKD (Cockcroft–Gault) was calculated using actual body weight and estimated ideal body weight. study found drug use not recommended guidelines in 2003, and is calculated using when kidney function was reduced in people the ‘Modification of Diet in Renal Disease’ Results aged 65 years caused a 40% increase in equation,10 and more recently, the CKD Kidney function was too low for recommended ≥ 5 prescribing in 4–40% of people aged≥ 65 years, all-cause mortality; one-third of hospital Epidemiology Collaboration (CKD-EPI) and in 24–80% of people aged≥ 85 years despite admissions related to adverse drug formula.11 eGFR was developed to provide more than 90% of patients having recent reactions were judged to be caused by staging for CKD and does not have a recorded kidney function results. Using eGFR 6 overestimated kidney function for 3–28% of those renal impairment in those aged ≥65 years; weight component, therefore, making it aged ≥65 years, and for 13–58% of those aged and one-third of adverse drug reactions in straightforward for pathology to report. ≥85 years. Increased age predicted higher odds people aged ≥65 years during their hospital Drug-dosing studies, however, have used of having a kidney function estimate too low for recommended prescribing of the study drugs. Conclusion Prescribing recommendations when kidney S Wood, PhD, MPharm, PGDip, MRPharmS-IP, Address for correspondence function is reduced are not applied for many research fellow, Academic Unit of Primary Care, Su Wood, Academic Unit of Primary Care, Leeds people aged ≥65 years in primary care. Using Institute of Health Sciences, University of Leeds, Institute of Health Sciences, Worsley Building, eGFR considerably overestimates kidney Leeds. D Petty, PhD IP, FRPharmS, research University of Leeds, Leeds LS2 9UT, UK. function for prescribing and, therefore, creatinine practitioner in primary care, School of Pharmacy Email: [email protected] clearance (Cockcroft–Gault) should be assessed when prescribing for these people. Interventions and Medical Sciences, University of Bradford, Submitted: 11 July 2017; Editor’s response: 15 are needed to aid prescribers when kidney Bradford. L Glidewell, PhD, FHEA, lecturer in August 2017; final acceptance: 21 December 2017. function is reduced. primary care, Academic Unit of Primary Care, ©British Journal of General Practice Institute of Health Sciences, University of Leeds, This is the full-length article (published online Keywords Leeds. DKT Raynor, PhD, BPharm, professor 24 Apr 2018) of an abridged version published in Cockcroft–Gault; creatinine; drug prescriptions; of pharmacy practice, School of Healthcare, print. Cite this version as: Br J Gen Pract 2018; DOI: kidney function; older people; prescribing University of Leeds, Leeds. https://doi.org/10.3399/bjgp18X695993 recommendations; renal insufficiency.

e378 British Journal of General Practice, May 2018 (www.medicines.org.uk/emc/) when kidney How this fits in function is reduced, for people aged ≥65 years across a healthcare population. Internationally, studies have found that The study also set out to determine the recommendations for prescribing in patients with reduced kidney function are often not impact of using eGFR instead of creatinine applied, with increased risk of patient harm clearance (Cockcroft–Gault) on drug and and hospital admissions. How medicines dosage decisions. affected by renal impairment are prescribed to those aged ≥65 years in primary care was METHOD previously unknown. This survey found many A cross-sectional survey was conducted people aged ≥65 years in UK general practice on prescribing data from all 80 general had a recorded estimate of kidney function but their prescribed drugs had not been practices in a north of England former adjusted. Data were available to calculate primary care trust (PCT). creatinine clearance (Cockcroft–Gault); Two drugs, or drug classes, were when creatinine clearance estimates were identified in each of four categories in applied even more patients were identified reduced kidney function: as at a risk. The findings from this study have informed an update to the British National • drugs that should be avoided; Formulary to recommend the Cockcroft– Gault formula as the preferred method for • drugs that should have a dose reduction; estimating renal function in older people. • drugs that are ineffective; and • drugs that require caution as they are known to frequently cause adverse drug the Cockcroft–Gault equation12 to estimate reactions in renal impairment. kidney function as ‘creatinine clearance’ (CrCl). Cockcroft–Gault includes a weight The eight choices of drugs were based component: ideal body weight to factor on findings in a previous case-note review muscle mass, not body fat, as creatinine of five practices in the PCT.24 Drugs were is produced from muscle turnover.13 The selected on the strength of the BNF/SPC equations give progressively different results (summaries of product characteristics) with increasing age, and studies focused on recommendations and the literature5,21–22,25–32 prescribing for older people have concluded impact on patients,28,33–40 and independent that Cockcroft–Gault should continue to expert advice from a renal physician, be used for prescribing decisions.14,15 geriatrician, hospital renal pharmacist, and Studies using gentamicin have shown that an antibiotics expert pharmacist (Table 1). blood levels are more closely estimated by creatinine clearance: while Cockcroft– Data collection Gault underestimates kidney function and Data collected were age, sex, actual weight effect on drug blood levels by 10% across and date last recorded, serum creatinine all older ages, eGFR overestimates kidney and date last recorded, eGFR and date last function increasingly as age increases recorded, or no eGFR ever recorded. These by 29% up to 69%,16 meaning the safer data were used to calculate creatinine estimate to use is Cockcroft–Gault.17,18 clearance using actual body weight Using eGFR for prescribing decisions, (CrCl-AW) and creatinine clearance using rather than Cockcroft–Gault, has been estimated ideal body weight (CrCl-IBW). shown to lead to more frequent major Searches were run for all patients bleeding events with glycoprotein-IIb-IIa- prescribed each drug by a PCT data analyst inhibitors,19 and increases the likelihood of at the end of October 2011 to extract adverse drug reactions,20 hospitalisation,6 data from general practice systems. All or ineffectiveness.21 practices used the TPP SystmOne clinical International studies have found information system. prescribing outside recommendations for older people with reduced kidney Data analysis function,5,22 but, while a similar problem Data quality and missing data analyses has been shown in a UK hospital,23 to were performed, and a descriptive analysis the authors’ knowledge, no studies have to give: been conducted in UK general practice. The aim of this study was to assess • the numbers of people aged ≥65 years 2 whether drugs are prescribed according with eGFR <60 ml/minute/1.73 m (NICE 8 to the British National Formulary (BNF)13 CKD level); and and the Electronic Medicines Compendium • the numbers of people aged ≥65 years (eMC), summary of product characteristics taking each of the study drugs.

British Journal of General Practice, May 2018 e379 Table 1. Choices of drugs and drug classes investigated in the study

Prescribing outside Category: recommendations, recommendation or dose suggested References where when kidney may need review, prescribing outside Comments function is BNF13/SPC in a case-note recommendations from independent reduced Drug recommendation review study,24 % was reported Evidence of potential impact for patients expertsa Avoid Alendronic acid SPC: not recommended 4.0 Breton et al 5 • Alendronic acid accumulates in the bones, Agreed to be BNF: avoid if CrCl Khanal et al 22 with a terminal half-life >10 years41 included <35 ml/min • MHRA review of risks: associated with an increased risk of atypical stress fractures of the proximal femoral shaft; low but increased risk of atrial fibrillation33 Metformin SPC: contraindication 2.0 Breton et al 5 • Increased risk of lactic acidosis especially Suggested by a <30 ml/min Khanal et al (2015)22 with dehydration34 renal physician and Schmidt-Mende a geriatrician as et al (2012)25 they see admissions to hospital caused by metformin when kidney function is low Reduce the SPC/BNF: CrCl 8.8 Breton et al 5 • Total drug (AUC) 2–3 times greater when Agreed to be dose <30 ml/min: dosages kidney function is low [personal included as a widely >10 mg/day should communication Merck Sharp & Dohme, 2011] -used drug with an be carefully considered (that is, greater than the effect of the alternative available and, if deemed necessary, drug interaction with amlodipine: AUC for (atorvastatin does implemented cautiously simvastatin increased by 1.58–1.77 fold35) not need dose • MHRA: increased plasma concentrations of alteration when simvastatin are associated with an increased kidney function is risk of myopathy and/or rhabdomyolysis35 low)42 Gabapentin Table of reduced doses 0.2 Breton et al 5 • High risk from the common adverse effects Added at the and pregabalin at specified levels of CrCl of somnolence, dizziness, ataxia, and fatigue43,44 suggestion of the • There has been a 46% rise in prescribing of renal pharmacist gabapentin and 53% rise in pregabalin because of frequent prescribing since 201136 side effects seen in the renal unit Ineffective Thiazides BNF: likely to be 17.0 Howard et al 26 • Unlikely to be effective below 30 ml/min45 Agreed to be ineffective at CrCl • Increased risk of adverse events included <30 ml/min such as electrolyte imbalance25 Nitrofurantoin BNF: likely to be 0.2 (on repeat) Farag et al 27 • MHRA: The antibacterial efficacy depends on Agreed to be ineffective at Geerts et al 28 the renal secretion of nitrofurantoin into the included; issue CrCl <45 ml/min Howard and Wood21 urinary tract. In patients with renal raised after SPC: contraindicated impairment, renal secretion of nitrofurantoin discussion with a <45 ml/min; may be is reduced. This may reduce the antibacterial pharmacist with used with caution efficacy, increase the risk of side effects (for international 30–44 ml/min — only example, nausea, vomiting, loss of expertise in prescribe to such appetite), and may result in treatment failures37 antibiotics patients to treat lower • The drug may not work increasing risk from urinary tract infection infection.46 An audit in a large GP practice with suspected or proven found older people with renal impairment multidrug-resistant were more likely to need further antibiotics21 pathogens when the • Raised blood levels increases the risk of benefits of nitrofurantoin pulmonary, hepatic, neurological, are considered to haematological, and gastrointestinal side outweigh the risks of effects during treatment; <50 ml/min side effects significantly increased the risk of pulmonary adverse events leading to hospitalisation (HR = 4.1, 95% CI = 1.39 to 13.09)28 ...continued

e380 British Journal of General Practice, May 2018 Table 1 continued. Choices of drugs and drug classes investigated in the study

Caution as NSAIDs BNF: caution in reduced 1.7 Evans et al 29 • Sodium and water retention may occur Agreed to be adverse drug kidney function (study Guthrie et al 30 and renal function may deteriorate, possibly included: NSAIDs reactions are parameter set at Howard et al 26 leading to renal failure; deterioration in renal affect all stages of likely <30 ml/min) Ingrasciotta et al 31 function has also been reported after kidney function and topical use13 are frequently a • Chronic use of NSAIDs is a risk factor cause of hospital for progression of CKD8 admission • Frequently cited as causing admissions38 ACEIs and BNF: caution in reduced 26.0 Breton et al 5 • Ramipril SPC: CrCl <60 ml/min maximum Although ACEIs and ARBs kidney function (study Handler et al 32 dose is 5 mg47 ARBs are used in parameter set at • Hyperkalaemia and other side effects of renal disease, they <30 ml/min) Khanal et al 22 ACEIs are more common, and the dose can also be Schmidt-Mende may need to be reduced13 nephrotoxic so the et al 25 • Can cause impairment of kidney function consultant experts which may progress and become severe suggested inclusion (at particular risk are older people)13 • STOP-ACEi study is investigating whether the risk to the kidney may outweigh any beneficial effect if <30 ml/min39 aIndependent experts were a renal physician with a national role, geriatrician, hospital renal pharmacist, and an antibiotics expert pharmacist. ACEIs= angiotensin-converting enzyme inhibitors. ARBs = angiotensin II receptor blockers. AUC = area under the curve. BNF = British National Formulary. CrCl = creatinine clearance. HR = hazard ratio. MHRA = Medicines and Healthcare products Regulatory Agency. NSAIDs= non-steroidal anti-inflammatory drugs. SPC = summary of product characteristics.

For each of the eight drugs the number RESULTS of patients in age bands ≥65, 65–74, 75–84, Thirteen per cent of the PCT population and ≥85 years were assessed; for seven were aged ≥65 years (70 900/549 533), and drugs on the repeat list, and 1.8% (9723) were aged ≥85 years; 26% of for nitrofurantoin, at least one prescription those aged ≥65 years, and 50% of those in the previous 12 months, described aged ≥85 years, had a documented eGFR as ‘patient drug events’. For each drug ≥60 ml/minute/1.73 m2. the following data were collected and/or A recent kidney function test in the calculated: previous 15 months was found in the patient record for 83% of patients prescribed • the number of ’patient drug events’ NSAIDs and for 97% of patients prescribed where a serum creatinine level had been metformin. Only 0.2–9.7% of the patient recorded in the previous 15 months, and drug events had no eGFR, and 0–5.7% did the estimated kidney function was too not include the parameters to be able to low for the recommended use, calculated calculate CrCl-IBW. using eGFR, CrCl-AW and CrCl-IBW; Figure 1 shows the numbers of patients • the number of ‘patient drug events’ prescribed each of the drugs studied; an that would be missed if eGFR, or actual ACEI or ARB was found to be prescribed for weight in Cockcroft–Gault, were used; 40% of the 70 900 patients ≥65 years. • the number of ‘patient drug events’ for Prevalence of low kidney function each drug where there was no kidney The number of patient drug events where function estimate on the record; a recent kidney function was available but • Spearman correlation between age and was too low for the recommended drug use level of kidney function (using SPSS ranged from 39.6% (95% confidence interval version 21); and [CI] = 35.8 to 43.4%) for nitrofurantoin, to • Logistic regression analysis to explore 3.5% (95% CI = 0 to 11.2.0%) for NSAIDs the effect of people aged ≥65 years on in those aged ≥65 years, and 79.5% the likelihood of having a kidney function (95% CI = 75.1 to 83.9%) (nitrofurantoin), to too low for appropriate prescribing of the 29.2% (95% CI = 0.5 to 57.9%) (NSAIDs) for drug (using SPSS version 21). the patients aged ≥85 years (Table 2). Figure 2 shows an example line plot Data oversight was provided by the charting the range of kidney function supervisory research team. (CrCl-IBW in the previous 15 months) for

British Journal of General Practice, May 2018 e381 Figure 1. Number of people aged≥ 65 years prescribed each of the study drugs. ACEI = angiotensin-converting Alendronic acid (n = 4230) 6% enzyme inhibitor. ARB = angiotensin II receptor blocker. NSAID = non-steroidal anti-inflammatory drug. Metformin (n = 6431) 9.1%

Simvastatin >10mg (n = 21 733) 30.7%

Gabapentin/pregabalin (n = 1448) 2.0%

Nitrofurantoin (n = 3489) 4.9%

Thiazide (n = 12 098) 17.1%

NSAID (n = 3363) 5%

ACEI/ARB (n = 28 254) 39.9% 0% 20% 40% 60% 80% 100%

nitrofurantoin at each age level for those suggested alteration or stopping of the aged ≥65 years. All cases below the line study drugs would be missed for 1–10% of at CrCl-IBW 45 ml/minute had a kidney patients aged ≥65 years. function likely to be too low for the drug to be effective, and more likely to cause Effect of increased age adverse drug reactions.28 Kidney function reduced with age for all groups studied using Spearman correlation, Different ways to estimate kidney for example, for nitrofurantoin (r = –0.608, function n = 3185, P<0.001). Logistic regression Using eGFR to estimate the level of kidney analysis found that, compared with age function suggested that a much lower band 65–74 years, patient drug events for number of alterations were required for those aged 75–84 years had higher odds of drug or dose choice than would be the case being prescribed drugs where the kidney using CrCl-IBW (Figure 3). Use of eGFR, function was too low, and even higher rather than CrCl-IBW, would mean that odds for those ≥85 years (Table 3). For the suggested alteration or stopping of the example, for nitrofurantoin, using CrCl-IBW study drugs would be missed for 3–28% of the patient drug events for people aged those aged ≥65 years, and for 13–58% for 75–84 years had 5.64 (95% CI = 4.58 to 6.95) those aged ≥85 years. times greater odds of being prescribed Using actual weight in the Cockcroft– drugs where the kidney function was too Gault equation also found fewer patients low for recommended use, and 29.23 who might need prescribing alteration, but (95% CI = 22.75 to 37.56) times greater odds to a lesser extent than using eGFR: the for those aged ≥85 years.

160.0

140.0

120.0

100.0

80.0

60.0

40.0

20.0 Kidney function CrCl-IBW, ml/min 0.0 Figure 2. Range of kidney function using creatinine clearance (Cockcroft–Gault) calculated using 65 65 67 68 69 70 71 72 73 74 75 76 77 78 78 79 80 81 82 83 85 86 87 89 90 93 Age, years estimated ideal body weight (CrCl-IBW) in the previous 15 months, for nitrofurantoin prescribed at each age CrCl-IBW 45 ml/min level ≥65 years.

e382 British Journal of General Practice, May 2018 DISCUSSION Summary Prescribing of drugs outside recommendations for use in reduced kidney function was widespread for the eight representative drugs in the study

population. Using eGFR, rather than CrCl- IBW, considerably underestimated 85 years the potential risk to patients, particularly for

Aged ≥ those aged ≥85 years.

recommended % 95% CI Prescribing recommendations in reduced < kidney function were not applied for a large

function Kidney anti-inflammatory drug.non-steroidal = number of people aged ≥65 years, in this first UK general practice study. Assessment of Cockcroft-Gault creatinine clearance to estimate kidney function level when making prescribing decisions for people aged ≥65 years will find more at risk of higher drug blood levels than eGFR, and reduce the risk of harm.

Strengths and limitations The study population was from a large Aged 75–84 years

recommended % 95% CI Total PCT with a broad range of prescribers. A < deprivation score higher than the England function Kidney average, a lower life expectancy, and a slightly younger population,40 suggests that the study findings might be an

900 (10.2%) patient drug events). underestimation of the problem in the UK. Only a small number of data were excluded, where no kidney function data were available, giving confidence in the findings being representative. Anonymised ‘patient drug events’ were collected, but it is likely that many patients were prescribed more than one drug. Aged 65–74 years using calculated estimated ideal body (Cockcroft–Gault) weight. clearance creatinine NSAID

=

recommended % 95% CI Total Inferences cannot be made on the group years years with a kidney function estimate (CrCl-CG IBW) in the 15 previous months below < of drugs as a whole in this study, or the function Kidney 65

≥ prevalence of a patient having more than Total one drug affected by level of kidney function. A case-note review in five GP practices found that 25% of patients aged ≥65 years were prescribed an average of two drugs when their kidney function was too low for recommended use, and 70 different drugs were involved.24 years 65 years angiotensin II receptor blocker. CrCl-IBW angiotensin blocker. II receptor = Comparison with existing literature

Aged ≥ International prevalence studies in primary and secondary care have reported large recommended % 95% CI < numbers of drug use and dosing outside 5,22,25–32

function Kidney recommendations. A French 3400 6276 804 267 23.6 28.1 19.2 to 3185 4.3 7.3 1.2 to 1120 1262 2483 39.6 3491 49 43.4 35.8 to 86 30 1206 4.4 3.5 5.1 3.6 to 11.2 0 to 141 0.9 0.91 0.81 to 1532 11.7 1577 13.4 10.0 to 2319 342 1242 8 124 22.3 24.6 20.1 to 535 5.3 0.5 748 0.57 0.44 to 5.8 4.9 to 43.1 47.0 39.1 to 413 762 466 737 55.2 113 36 61.8 to 48.6 586 24.2 4.7 79.5 32.1 to 16.4 5.9 3.5 to 83.9 75.1 to 144 42 29.2 57.9 0.5 to Total

10 805 797 26 109 7.4 2521 9.1 5.7 to 9.7 multicentre 10.3 9.0 to 5371 12 663 62 288 prospective 1.2 1.20 to 1.11 2.3 2.31 2.24 to 4189 10 280 primary 303 1062 care 10.3 7.2 10.5 10.2 to 7.6 6.9 to study 3166 1245 1171 432 37.0 38.5 35.5 to 34.7 38.3 to 31.1 found 13.3% of patients aged ≥65 years with a kidney function too low for recommended use, 52.5% in those with eGFR 30–59, and 96% with eGFR <30 ml/minute/1.73 m2.5 10 mg 19 434 1465 7.5 8.5 6.6 to 9894 A retrospective 167 1.7 1.72 1.65 to 7494 case 643 record 8.6 8.8 8.4 to analysis 2046 of 655 32.0 34.1 29.9 to patients aged >70 years in a UK hospital angiotensin-converting enzyme inhibitor. ARB enzyme angiotensin-converting inhibitor. = found 13% had ‘potentially inappropriate prescribing’.23 Table 2. Number of Table drugs for aged people prescribed Drug acid Alendronic Metformin > Simvastatin Gabapentin or pregabalin Nitrofurantoin 1209 Thiazide NSAID 132 ACEI/ARB 10.9 25.7 0 to 598 43 7.2 9.4 5.0 to 450 56 12.4 17.3 7.6 to 161 33 20.5 40.7 0.3 to that for recommended that use for recommended ACEI Patients with no eGFR or on kidney the function record, data older than 15 for months, the excluded were final analysis (9911/96 A recent kidney function level was

British Journal of General Practice, May 2018 e383 Where kidney function is reduced: eGFR CrCl Cockcroft-Gault

Alendronic acid Alendronic acid Avoid Metformin Metformin

Simvastatin >10mg Simvastatin >10mg Reduce the dose Gabapentin/Pregabalin Gabapentin/Pregabalin

Nitrofurantoin Nitrofurantoin Ineffective Thiazide Thiazide

NSAIDs NSAIDs ‘Caution’ ACEIs/ARBs ACEIs/ARBs

0% 20% 40% 60% 80% 100% 0% 20% 40% 60% 80% 100% 65–74 years 75–84 years ≥85 years

Figure 3. Percentage of patients found with a kidney recorded in the patient record for most was used,50 to a Spanish hospital study of function too low for recommended use of the drug (90%) people aged ≥65 years prescribed nephrotoxic drugs, which found 65% dosing using estimated glomerular filtration rate compared the included drugs in the current study. The outside recommendations.51 A US primary with creatinine clearance (Cockcroft–Gault) in the UK Quality Outcomes Framework48 targets care study found that 40% of patients should three older age bands. ACEIs = angiotensin-converting may have helped because kidney function have been recommended different doses, enzyme inhibitor. ARBs = angiotensin II receptor blockers. CrCI = creatinine clearance. eGFR = estimated testing has been incentivised, for example, and 22% of those reviewed would have had glomerular filtration rate. NSAIDs = non-steroidal anti- in diabetes. However, renal impairment different recommendations based on the 52 inflammatory drugs. did not seem to lead to altered prescribing equation chosen. The current study found in all four categories of drugs studied, that using eGFR would mean that many which may impact on patient safety from people needing a review of prescribing increased blood drug levels, sensitivity, would be missed. It also showed that or ineffectiveness. For all drugs studied, using actual weight in the Cockcroft–Gault the dose could have been reduced or an equation would miss drugs that might need alternative treatment chosen (such as review; an ideal body weight, or actual body atorvastatin, which is not affected by level weight if lower, should therefore be used in of kidney function) instead of simvastatin, the Cockcroft–Gault calculation to give an or another antihypertensive or antidiabetic indication of muscle mass.13,14 medication. Increased age predicted higher odds Implications for research and practice of having a kidney function estimate too Treatment of multimorbidity has resulted low for recommended prescribing. This in increasing numbers of people aged analysis did not address the possibility of ≥65 years being prescribed more long- confounders, but there is broad evidence term medicines, the mean being seven for of progressive loss of kidney function with people aged ≥80 years,53 and 60% of the ageing.2,3 Those aged ≥65 years are more billion items dispensed in the community likely to have a lower kidney function, and in England are for people aged ≥65 years.54 so a lower reserve to react to assaults As many people aged ≥65 years are such as dehydration or nephrotoxicity of prescribed multiple medications, there drugs. Renal impairment, polypharmacy, is the likelihood that they will be taking and identified drugs such as NSAIDs are more than one drug eliminated via the included as ‘deficits’ that can occur with kidneys, increasing the complexity of the ageing and combine to increase ‘frailty’ impact. Applying BNF recommendations which, in turn, increases the risk of adverse for prescribing (or avoidance of prescribing) outcomes.49 in renal impairment could reduce the risk of The equation used to estimate kidney adverse drug reactions, particularly for the function for prescribing decisions made oldest and most frail patients. a substantial difference to whether In 2017 the BNF reviewed the prescribing might be reviewed because of evidence from this study and updated its renal impairment in people aged ≥65 years. recommendation in the ‘Prescribing in Many other studies report variance Renal Impairment’ section to state that in estimating of kidney function when ‘... the Cockcroft and Gault formula is the different equations are used, ranging from preferred method for estimating renal an Italian hospital study that found 9.8% function in elderly patients aged 75 years dosing outside recommendations if eGFR and over’.13,55 Electronic patient record

e384 British Journal of General Practice, May 2018 Table 3. The odds of having a kidney function too low for the drug with increased age compared with those aged 65–74 years, using eGFR, creatinine clearance (Cockcroft–Gault) calculated using actual body weight, and using ideal body weight

Kidney function 65–74 years 75–84 years ≥85 years Drug calculation Odds ratio Odds ratio 95% CI Odds ratio 95% CI eGFR 1 1.96 1.08 to 3.56 5.49 3.07 to 9.81 Alendronic acid CrCl-AW 1 5.11 3.59 to 7.26 24.12 16.93 to 34.36 CrCl-IBW 1 6.28 4.61 to 8.57 26.95 19.56 to 37.13 eGFR 1 2.26 0.80 to 6.35 6.15 1.87 to 6.35 Metformin CrCl-AW 1 5.90 3.20 to 10.88 48.64 26.70 to 88.61 CrCl-IBW 1 6.52 4.36 to 9.75 36.93 24.33 to 58.05 eGFR 1 2.45 1.89 to 3.17 4.98 3.71 to 6.70 Simvastatin >10 mg CrCl-AW 1 5.03 4.06 to 6.23 30.64 24.73 to 37.95 CrCl-IBW 1 5.47 4.60 to 6.50 27.43 22.93 to 32.80 eGFR 1 2.73 1.27 to 5.89 4.28 1.78 to 10.26 Gabapentin/pregabalin CrCl-AW 1 2.97 1.55 to 5.72 9.96 5.25 to 18.88 CrCl-IBW 1 1.83 1.21 to 2.79 3.33 2.03 to 5.45 eGFR 1 3.03 4.20 to 4.18 5.24 3.78 to 7.27 Nitrofurantoin CrCl-AW 1 6.79 5.16 to 8.92 41.33 30.87 to 55.36 CrCl-IBW 1 5.64 4.58 to 6.95 29.23 22.75 to 37.56 eGFR 1 2.19 1.42 to 3.37 5.29 3.32 to 8.43 Thiazides CrCl-AW 1 5.60 3.96 to 7.93 49.67 35.52 to 69.44 CrCl-IBW 1 6.68 5.07 to 8.80 45.50 34.52 to 59.98 eGFR 1 4.20 1.05 to 16.84 3.68 0.38 to 35.57 NSAIDs CrCl-AW 1 3.50 1.52 to 8.03 43.93 20.31 to 95.03 CrCl-IBW 1 9.73 4.50 to 21.03 80.76 36.94 to 176.55 eGFR 1 1.81 1.49 to 2.21 3.84 3.09 to 4.78 ACEIs/ARBs CrCl-AW 1 4.73 3.98 to 5.61 29.01 24.27 to 34.38 CrCl-IBW 1 4.95 4.33 to 5.65 25.22 21.99 to 28.93 ACEIs = angiotensin-converting enzyme inhibitors. ARBs = angiotensin II receptor blockers. CrCl-AW = creatinine clearance (Cockcroft-Gault) calculated using actual body weight. CrCl-IBW = creatinine clearance (Cockcroft-Gault) calculated using ideal body weight. eGFR = estimated glomerular filtration rate. NSAIDs = non-steroidal anti-inflammatory drugs.

Funding systems, such as SystmOne and EMIS, have recommendations when prescribing, and to This research was carried out as part of a a ‘renal calculator’ to enable prescribers to inform an intervention to aid prescribing in self-funded PhD at the University of Leeds. calculate, and code, creatinine clearance reduced kidney function.57 Cockcroft–Gault; smartphone apps are also Guidelines should consider Ethical approval available and easy to use. Pharmacists recommendations for drug use and level Not required for this study. have been shown to significantly reduce of kidney function.58 CKD and acute kidney Provenance prescribing outside recommendations in injury guidelines could include guidance 56 Freely submitted; externally peer reviewed. reduced kidney function; they can audit, on prescribing (and avoiding prescribing) to highlight for review, and do medication reduce risk to the kidney and reduce adverse Competing interests reviews for high-risk populations. drug reactions from renally excreted drugs. The authors have declared no competing The data required to calculate creatinine The NHS England Think Kidneys resources interests. clearance and identify affected drugs are already available in the prescribing and include an acute kidney injury medicines Acknowledgements consultation systems; it would be possible optimisation toolkit (www.thinkkidneys. The authors thank the PCT research nhs.uk) aimed at prescribers in hospitals, coordinator Paul Carder, data analyst to develop patient- and drug-specific which could be usefully extended to primary Simon Falkner, and the statistics advice decision support to suggest safer doses care. Non-prescription medicines could be received from Professor Robert West. or alternative drugs, both at initiation and at medication review. Further research important for future research, for example Discuss this article has been undertaken to understand why NSAIDs bought from pharmacies or retail Contribute and read comments about this GPs often do not apply renal impairment outlets. article: bjgp.org/letters

British Journal of General Practice, May 2018 e385 with renally excreted drugs in primary care?Int J Pharm Pract 2011; 19(Suppl REFERENCES 2): 38.

1. Ashley C, Dunleavy A, eds. The renal drug handbook. 4th edn. Oxford: Radcliffe 25. Schmidt-Mende K, Wettermark B, Andersen M, Hasselström J. Potentially Medical Press, 2014. inappropriate drugs in elderly hypertensive patients with impaired renal function. Pharmacoepidemiol Drug Saf 2012; 21: 30. 2. Beers MH, Berkow R, eds. The Merck manual of geriatrics. 3rd edn. West Point, PA: Merck & Co. Inc., 2000. 26. Howard RL, Avery AJ, Howard PD, Partridge M. Investigation into the reasons for preventable drug related admissions to a medical admissions unit: 3. Michel JP, Beattie L, Martin FC, Walston JD.Oxford textbook of geriatric observational study. Qual Saf Health Care 2003; 12(4): 280–285. medicine. 3rd edn. Oxford University Press: Oxford, 2017. 27. Farag A, Garg AX, Li L, Jain AK. Dosing errors in prescribed antibiotics for older 4. Lindeman RD, Tobin J, Shock NW. Longitudinal studies on the rate of decline in persons with CKD: a retrospective time series analysis.Am J Kidney Dis 2014; renal function with age. J Am Geriatr Soc 1985; 33(4): 278–285. 63(3): 422–428. 5. Breton G, Froissart M, Janus N, et al. Inappropriate drug use and mortality 28. Geerts AF, Eppenga WL, Heerdink R, et al. Ineffectiveness and adverse events in community-dwelling elderly with impaired kidney function: the Three-City of nitrofurantoin in women with urinary tract infection and renal impairment in population-based study. Nephrol Dial Transplant 2011; 26(9): 2852–2859. primary care. Eur J Clin Pharmacol 2013; 69(9): 1701–1707. 6. Helldén A, Bergman U, von Euler M,et al. Adverse drug reactions and impaired 29. Evans JM, McGregor E, McMahon AD, et al. Non-steroidal anti-inflammatory renal function in elderly patients admitted to the emergency department: a drugs and hospitalization for acute renal failure.QJM 1995; 88(8): 551–557. retrospective study. Drugs Aging 2009; 26(7): 595–606. 30. Guthrie B, McCowan C, Davey P, et al. High risk prescribing in primary care 7. Corsonello A, Pedone C, Corica F, et al. Concealed renal insufficiency and patients particularly vulnerable to adverse drug events: cross sectional adverse drug reactions in elderly hospitalized patients.Arch Intern Med 2005; population database analysis in Scottish general practice.BMJ 2011; 342: 165(7): 790–795. d3514. 8. National Institute for Health and Care Excellence. in 31. Ingrasciotta Y, Sultana J, Giorgianni F, et al. The burden of nephrotoxic adults: assessment and management. Clinical guideline [CG182]. 2014. https:// drug prescriptions in patients with chronic kidney disease: a retrospective www.nice.org.uk/guidance/cg182 (accessed 14 Mar 2018). population-based study in Southern Italy. PloS One 2014; 9(2): e89072. 9. National Institute for Health and Care Excellence.Acute kidney injury: 32. Handler SM, Cheung PW, Culley CM, et al. Determining the incidence of drug- prevention, detection and management. Clinical guideline [CG169]. 2013. https:// www.nice.org.uk/Guidance/CG169 (accessed 14 Mar 2018). associated acute kidney injury in nursing home residents.J Am Med Dir Assoc 2014; 15(10): 719–724. 10. Levey AS, Bosch JP, Lewis JB, et al. A more accurate method to estimate glomerular filtration rate from serum creatinine: a new prediction equation. 33. Medicines and Healthcare products Regulatory Agency. : Modification of Diet in Renal Disease Study Group.Ann Intern Med 1999; 130(6): use and safety. 2014. https://www.gov.uk/government/publications/ 461–470. bisphosphonates-use-and-safety/bisphosphonates-use-and-safety (accessed 13 Mar 2018). 11. Levey AS, Stevens LA, Schmid CH, et al. A new equation to estimate glomerular filtration rate. Ann Intern Med 2009; 150(9): 604–612. 34. DeFronzo R, Fleming GA, Chen K, Bicsak TA. Metformin-associated lactic acidosis: current perspectives on causes and risk.Metabolism 2016; 65(2): 12. Cockcroft DW, Gault MH. Prediction of creatinine clearance from serum 20–29. creatinine. Nephron 1976; 16(1): 31–41. 35. Medicines and Healthcare products Regulatory Agency. Simvastatin: updated 13. BMJ Publishing Group, Royal Pharmaceutical Society. British national formulary advice on drug interactions. 2012. https://www.gov.uk/drug-safety-update/ 74. London: BMJ Publishing Group, RPS Publishing, 2017. simvastatin-updated-advice-on-drug-interactions (accessed 13 Mar 2018). 14. Helou R. Should we continue to use the Cockcroft-Gault formula?Nephron Clin 36. Public Health England, NHS England. Advice for prescribers on the risk of Pract 2010; 116(3): c172–c185. the misuse of pregabalin and gabapentin. PHE publications gateway number: 15. Spruill WJ, Wade WE, Cobb HH 3rd. Comparison of estimated glomerular 2014586. 2014. https://www.gov.uk/government/uploads/system/uploads/ filtration rate with estimated creatinine clearance in the dosing of drugs attachment_data/file/385791/PHE-NHS_England_pregabalin_and_gabapentin_ requiring adjustments in elderly patients with declining renal function.Am J advice_Dec_2014.pdf (accessed 13 Mar 2018). Geriatr Pharmacother 2008; 6(3): 153–160. 37. Medicines and Healthcare products Regulatory Agency. Nitrofurantoin now 16. Roberts GW, Ibsen PM, Schiøler CT. Modified diet in renal disease method contraindicated in most patients with an estimated glomerular filtration overestimates renal function in selected elderly patients.Age Ageing 2009; rate (eGFR) of less than 45ml/min/1.73m2. 2014. https://www.gov.uk/drug- 38(6): 698–703. safety-update/nitrofurantoin-now-contraindicated-in-most-patients-with-an- 17. Péquignot R, Belmin J, Chauvelier S, et al. Renal function in older hospital estimated-glomerular-filtration-rate-egfr-of-less-than-45-ml-min-1-73m2 patients is more accurately estimated using the Cockcroft-Gault formula than (accessed 13 Mar 2018). the modification diet in renal disease formula.J Am Geriatr Soc 2009; 57(9): 38. Wiffen P, Gill M, Edwards J, Moore A. Adverse drug reactions in hospital 1638–1643. patients: a systematic review of the prospective and retrospective 18. Dowling TC, Wang ES, Ferrucci L, Sorkin JD. Glomerular filtration rate studies. Bandolier Extra 2002. http://www.bandolier.org.uk/Extraforbando/ equations overestimate creatinine clearance in older individuals enrolled ADRPM.pdf (accessed 10 Apr 2018). in the Baltimore Longitudinal Study on Aging: impact on renal drug 39. Bhandari S, Ives N, Brettell EA, et al. Multicentre randomized controlled trial dosing. Pharmacotherapy 2013; 33(9): 912–921. of angiotensin-converting enzyme inhibitor/angiotensin receptor blocker 19. Melloni C, Peterson ED, Chen AY, et al. Cockcroft-Gault versus modification withdrawal in advanced renal disease: the STOP-ACEi trial.Nephrol Dial of diet in renal disease: importance of glomerular filtration rate formula for Transplant 2016; 31(2): 255–261. classification of chronic kidney disease in patients with non–ST-segment 40. Department of Health. Health Profile 2012: Bradford. http://webarchive. elevation acute coronary syndromes. J Am Coll Cardiol 2008; 51(10): 991–996. nationalarchives.gov.uk/20170106193052/http://www.apho.org.uk/resource/ 20. Corsonello A, Pedone C, Lattanzio F, et al. Association between glomerular item.aspx?RID=117066 (accessed 20 Mar 2018). filtration rate and adverse drug reactions in elderly hospitalized patients: the 41. The electronic Medicines Compendium (eMC). Fosamax once weekly 70mg role of the estimating equation. Drugs Aging 2011; 28(5): 379–390. tablets. Merck Sharp & Dohme Limited. 2016. https://www.medicines.org.uk/ 21. Howard P, Wood S. Outcomes with the use of nitrofurantoin in renal emc/product/1281 (accessed 27 Mar 2018). impairment in primary care ­— a pilot study. Eur J Hosp Pharm Sci 2013; 42. The electronic Medicines Compendium (eMC). Lipitor 80 mg film-coated 20(Suppl 1): A114. tablets. Pfizer Limited. 2017. https://www.medicines.org.uk/emc/product/5239/ 22. Khanal A, Peterson GM, Castelino RL, Jose MD. Potentially inappropriate smpc (accessed 27 Mar 2018). prescribing of renally cleared drugs in elderly patients in community and aged 43. The electronic Medicines Compendium (eMC). Neurontin 300 mg hard care settings. Drugs Aging 2015; 32(5): 391–400. capsules. Pfizer Limited. 2017. https://www.medicines.org.uk/emc/ 23. Jones SA, Bhandari S. The prevalence of potentially inappropriate medication product/3195/smpc (accessed 27 Mar 2018). prescribing in elderly patients with chronic kidney disease.Postgrad Med J 44. The electronic Medicines Compendium (eMC). Lyrica capsules. Pfizer Limited. 2013; 89(1051): 247–250. 2017. https://www.medicines.org.uk/emc/product/5539/smpc (accessed 27 Mar 24. Wood S, Petty D, Glidewell L, Raynor T. Are we over-dosing our elderly patients 2018).

e386 British Journal of General Practice, May 2018 45. The electronic Medicines Compendium (eMC). Aprinox/bendroflumethiazide 52. Lessard BA, Zaiken K. Comparison of equations for dosing of medications 2.5mg tablets. Sovereign Medical. 2013. https://www.medicines.org.uk/emc/ requiring renal adjustment. J Am Pharm Assoc 2013; 53(1): 54–57. product/2240 (accessed 27 Mar 2018). 53. Petty DR, Zermansky AG, Alldred DP. The scale of repeat prescribing–time for 46. The electronic Medicines Compendium (eMC). Macrobid capsules 100mg B.P. an update. BMC Health Serv Res 2014; 14(1): 76. Concordia International — formerly AMCo. 2014. https://www.medicines.org.uk/ 54. NHS Digital. Focus on the health and care of older people — June 2014. http:// emc/product/429 (accessed 27 Mar 2018). digital.nhs.uk/catalogue/PUB14369 (accessed 14 Mar 2018). 47. The electronic Medicines Compendium (eMC). Tritace 10mg tablets. SANOFI 2018. https://www.medicines.org.uk/emc/product/2757/smpc (accessed 27 Mar 55. National Institute for Health and Care Excellence.Prescribing in renal 2018). impairment. Issues encountered in renal impairment. 2013. https://bnf.nice.org. uk/guidance/prescribing-in-renal-impairment.html (accessed 14 Mar 2018). 48. National Institute for Health and Care Excellence.NICE Quality and Outcomes Framework indicators. 2018. https://www.nice.org.uk/standards-and-indicators/ 56. Bhardwaja B, Carroll NM, Raebel MA, et al. Improving prescribing safety in qofindicators (accessed 14 Mar 2018). patients with renal insufficiency in the ambulatory setting: the Drug Renal Alert Pharmacy (DRAP) program. Pharmacotherapy 2011; 31(4): 346–356. 49. Rockwood K, Song X, MacKnight C, et al. A global clinical measure of fitness and frailty in elderly people. CMAJ 2005; 173(5): 489–495. 57. Wood S. Are recommendations for prescribing applied for older people with 50. Moranville MP, Jennings HR. Implications of using modification of diet in renal reduced kidney function in primary care? A mixed methods study to explore disease versus Cockcroft-Gault equations for renal dosing adjustments.Am J and improve implementation. PhD thesis. University of Leeds, 2016. http:// Health Syst Pharm 2009; 66(2): 154–161. etheses.whiterose.ac.uk/16504/ (accessed 22 Mar 2018). 51. Cabello-Muriel A, Urbieta-Sanz E, Iniesta-Navalón C, et al. Influence of 58. Dumbreck S, Flynn A, Nairn M, et al. Drug-disease and drug-drug interactions: equation used to estimate the renal function in dosage potentially nephrotoxic systematic examination of recommendations in 12 UK national clinical drug. Eur J Hosp Pharm Sci 2015; 22(1): 23–27. guidelines. BMJ 2015; 350: h949.

British Journal of General Practice, May 2018 e387