Drugs to Avoid in 2017

OUTLOOK

2017 Update

A more recent version is available online at english.prescrire.org Use the “Search” function with keywords “drugs to avoid (Prescrire's annual review)”

Towards better patient care: drugs to avoid in 2017

ABSTRACT harm-benefit balance, and the trial design must be relevant. Tailored supportive care is the best ●● To help healthcare professionals and patients option when there are no available treatments choose high-quality treatments that minimise the capable of improving prognosis or quality of life, risk of adverse effects, in early 2017 we updated beyond their placebo effect. the list of drugs that Prescrire advises health pro- Rev Prescrire 2017; 37 (400): 137-148 fessionals and patients to avoid.

●● Prescrire’s assessments of the harm-benefit balance of new drugs and indications are based on a his is Prescrire’s fifth consecutive annual rigorous procedure that includes a systematic and review of “drugs to avoid” (1,2). The drugs reproducible literature search, identification of Tlisted here are clearly more dangerous than patient-relevant outcomes, prioritisation of the sup- beneficial and should therefore not be used. The porting data based on the strength of evidence, aim is to help health professionals choose safe, comparison with standard treatments, and an anal- effective treatments and thereby avoid harming ysis of both known and potential adverse effects. their patients.

●● This fifth annual review of drugs to avoid has been extended to cover all drugs examined by A reliable, rigorous and independent Prescr ire between 2010 and 2016 and authorised methodology in the European Union, whereas previous reviews only considered drugs marketed in France. We What data sources and methodology do we use to identified 91 drugs that are more harmful than assess the harm-benefit balance of a given drug? beneficial in all the indications for which they have The following review concerns drugs and indica- been authorised in France or in the Europe- tions on which we published detailed analyses in an Union. our French edition over a seven-year period, from 2010 to 2016. Some drugs and indications were ●● In most cases, when drug therapy is really examined for the first time, while others were necessary, other drugs with a better harm-benefit re-evaluated as new data on efficacy or adverse balance are available. effects became available. All our publications are intended to provide ●● Even in serious situations, when no effective health professionals (and thereby their patients) treatment exists, there is no justification for pre- with the clear, independent, reliable and up-to-date scribing a drug with no proven efficacy that pro- information they need, free from conflicts of inter- vokes severe adverse effects. It may be acceptable est and commercial pressures. to test these drugs in clinical trials, but patients Prescrire is structured in such a way as to guar- must be informed of the uncertainty over their antee the quality of the information provided to our

Prescrire International • April 2017 • Volume 26 N° 181 • Page 108-1 Downloaded from english.prescrire.org on 02/10/2021 Copyright(c)Prescrire. For personal use only. OUTLOOK

subscribers. The Editorial Staff comprise a broad example, it can be difficult to attribute a specific range of health professionals working in various outcome to a particular drug, as other factors must sectors and free from conflicts of interest. We also be taken into account, including the natural history call on an extensive network of external reviewers of the disease, the placebo effect, the effect of an- (specialists, methodologists, and practitioners rep- other treatment the patient may not have men- resentative of our readership), and each article tioned, or a change in lifestyle or diet. Similarly, a undergoes multiple quality controls and doctor who sees an improvement in certain patients cross-checking at each step of the editorial process may be unaware that many other patients have been (see About Prescrire > How we work at english. harmed by the same treatment (3). prescrire.org). Our editorial process is a collective The best way to overcome this subjective bias one, as symbolised by the “©Prescrire” signature. due to non-comparative evaluation of a few pa- Prescrire is also fiercely independent. Our work tients is to prioritise well-conducted clinical studies, is funded solely and entirely by our subscribers. No particularly double-blind randomised trials versus company, professional organisation, insurance sys- standard care (3,4). tem, government agency or health authority has any financial influence whatsoever over the content Serious conditions with no effective treat- of our publications. ment: patients should be informed of the consequences of interventions. When faced Comparison with standard treatments. The with a serious condition for which there is no effective harm-benefit balance of a given drug has to be con- treatment, some patients opt to forgo treatment while tinually re-evaluated as new data on efficacy or ad- others are willing to try any drug that might bring verse effects become available. Likewise, treatment them even temporary relief, despite a risk of serious options evolve as new drugs arrive on the market. adverse effects. Not all drugs are equal: some offer a therapeutic When the short-term prognosis is poor, some advantage, while others are more harmful than health professionals may propose “last-chance” beneficial and should not be used (3). treatments without fully informing the patient of All Prescrire’s assessments of new drugs and the harms, either intentionally or unwittingly. indications are based on a systematic and repro- Yet patients in this situation must not be treated ducible literature search. The resulting data are then as guinea pigs. It is very useful to enrol patients analysed collectively by our editorial staff, using an into clinical studies provided they are informed of established procedure: the harms and the uncertain nature of the possible –– Efficacydata are prioritised: most weight is given benefits, and that the results are published in order to studies providing robust supporting evidence, to advance medical knowledge. i.e. well-conducted, double-blind, randomised con- But patients must be made aware that they have trolled trials; the option of refusing to participate in clinical trials or –– The new drug is compared with a carefully cho- to receive last-chance treatments with an uncertain sen standard treatment, if one exists (not necessar- harm-benefit balance. They must also be reassured ily a drug); that, if they do refuse, they will not be abandoned but –– The accent is placed on those clinical endpoints will continue to receive the best available care. Even most relevant to the patients concerned. This means though they are not aimed at modifying the outcome that we often ignore surrogate endpoints such as of the underlying disease, supportive care and symp- simple laboratory markers that have not been shown tomatic treatment are key elements of patient care. to correlate with a favourable clinical outcome (4,5). By their very nature, clinical trials involve a high degree of uncertainty. In contrast, drugs used for Careful analysis of adverse effects. Adverse routine care must have an acceptable harm-benefit effects can be more difficult to analyse, as they are balance. Marketing authorisation should only be often less thoroughly documented than efficacy, granted on the basis of proven efficacy relative to and this discrepancy must be taken into account. standard care, and an acceptable adverse effect The adverse effect profile of each drug is as- profile: in general, little, if any, extra information on sessed by examining data from clinical trials and efficacy is collected once marketing authorisation animal pharmacotoxicology studies, and any has been granted (3). pharmacological affiliation. The fact that a new drug has been granted marketing authorisation does not signify that its 91 authorised drugs more dangerous harm-benefit balancehas been fully documented. than beneficial Indeed, rare but serious adverse effects may only emerge after several years of routine use (3). This review lists drugs that have an unfavourable harm-benefit balance in all their authorised indica- Empirical data and personal experience: risk tions, in other words drugs that should be removed of bias. Empirical assessment of a drug’s harm- from the market on account of their toxicity. Drugs benefit balance based on individual experience can with an unfavourable harm-benefit balance in help to guide further research but is subject to major certain situations but not in others have not been bias and represents only weak evidence (3,4). For included.

Prescrire International • April 2017 • Volume 26 N° 181 • Page 108-2 Downloaded from english.prescrire.org on 02/10/2021 Copyright(c)Prescrire. For personal use only. OUTLOOK

Notable changes in the 2017 update

nly one drug from the list of drugs that Prescrire ad Six of the new products examined by Prescrire in 2016 Ovises health professionals and patients to avoid was have an unfavourable harm-benefit balance in all their withdrawn from the market in 2016: the recombinant urate approved indications, and three of them are cancer drugs: oxidase pegloticase. Its European marketing authorisation, nintedanib for non-small cell lung cancer and for idiopath- for severe gout, was withdrawn at the request of the phar- ic pulmonary fibrosis,olaparib for ovarian cancer, pano- maceutical company concerned (Prescrire Int n° 180). binostat for multiple myeloma, mepolizumab for asthma, ciclosporin eye drops for dry eye disease, and idebenone Panitumumab, varenicline: Prescrire reviewing new for Leber’s hereditary optic neuropathy. data in 2017. The only drugs listed in our 2016 review of drugs to avoid that do not feature in this year’s review are Additions authorised at European level but not market- panitumumab for colorectal cancers and varenicline for ed in France. Prescrire analyses all drugs that receive smoking cessation. This is because we are currently authorisation through European or French marketing re-evaluating their harm-benefit balance in light of new authorisation procedures. In previous years, we only con- data published in 2016. sidered drugs marketed in France when compiling our list of drugs to avoid in order to provide better patient care. Additions: ambroxol, capsaicin, various antineoplastics. This year, for the benefit of readers who do not work or The adverse effects of the mucolytics ambroxol and brom- live in France, we have expanded our review to include all hexine are better known now that they have been in use the drugs examined by Prescrire between 2010 and 2016 for a long time. The hypersensitivity reactions and and having European marketing authorisation, regardless life-threatening cutaneous disorders they cause make their of their availability in France. harm-benefit balance unfavourable. Although these Ten drugs to avoid were added to the list as a result of adverse effects are rare, they are unacceptable for drugs this approach. All but the first are unavailable in France as that have no efficacy beyond a placebo effect and that are of early 2017: alemtuzumab for multiple sclerosis, alogliptin indicated for minor ailments such as cough or sore throat. (alone or combined with metformin), canagliflozin, dapagli- The data on dronedarone in atrial fibrillation andcapsa - flozin and pioglitazone for type 2 diabetes, the fixed-dose icin in neuropathic pain led us to add these drugs to the combinations bupropion + naltrexone for weight loss, list of ones to avoid. mannitol inhalation powder for cystic fibrosis,mifamur- We have also added the vasoconstrictor phenylephrine, tide for osteosarcoma, ranolazine for angina, and vernaka- authorised as a decongestant for nasal use, which we had lant for atrial fibrillation. erroneously omitted from previous lists. ©Prescrire

This fifth annual review of drugs to avoid has options are briefly mentioned, as are situations been extended to cover all the drugs examined by (serious or non-serious) in which there is no suit- Prescrire between 2010 and 2016 that are author able treatment. ised in the European Union. In previous reviews, The differences between this year’s and last we confined our assessment to drugs marketed in year’s lists are detailed in the inset above. France. As of early 2017, we have identified 91 drugs that are more dangerous than beneficial, 82 of which are marketed in France. Cardiology They are listed below, based first on the thera peutic area in which they are used and then in ••Aliskiren, an antihypertensive renin inhibitor, has alphabetical order of their international nonpropri- not been shown to prevent cardiovascular events. etary names (INNs). On the contrary, a trial in diabetic patients showed These 91 drugs comprise: that aliskiren was associated with an increase in –– Active substances with adverse effects that are cardiovascular events and renal failure (Prescrire disproportionate to the benefits they provide in a Int n° 106, 129, 166). It is better to choose one of the given situation; many established antihypertensive drugs such as –– Older drugs that have been superseded by new a thiazide diuretic or an angiotensin converting drugs with a better harm-benefit balance; enzyme (ACE) inhibitor. –– Recent drugs that have a less favourable harm- ••Bezafibrate, ciprofibrate and fenofibrate are benefit balance than existing options; cholesterol-lowering drugs with no proven efficacy –– Drugs that have no proven efficacy (beyond the in the prevention of cardiovascular events (beyond placebo effect) but that carry a risk of serious ad- the placebo effect), yet they all have numerous verse effects. adverse effects, including cutaneous, haemato The main reasons why these drugs are consid- logical and renal disorders (Prescrire Int n° 85, 117). ered to have an unfavourable harm-benefit balance When a fibrate is justified,gemfibrozil is the only are explained in each case. When available, better one that has been shown to prevent the cardio

Prescrire International • April 2017 • Volume 26 N° 181 • Page 108-3 Downloaded from english.prescrire.org on 02/10/2021 Copyright(c)Prescrire. For personal use only. OUTLOOK

vascular complications of hypercholesterolaemia, Dermatology - Allergy although renal function and serum creatine phos- phokinase levels must be closely monitored. ••Mequitazine, a sedating antihistamine with anti- ••Dronedarone, an antiarrhythmic chemically relat- muscarinic activity, authorised for allergies, has ed to amiodarone, is less effective than amiodarone only modest efficacy but carries a higher risk than at preventing atrial fibrillation recurrence, yet has other antihistamines of cardiac arrhythmias at least as many serious adverse effects, in particu through QT prolongation in patients who are slow lar hepatic, pulmonary and cardiac disorders CYP2D6 metabolisers (and CYP2D6 metaboliser (Prescrire Int n° 108, 120, 122; Rev Prescrire n° 339). status is rarely known) or when co-administered Amiodarone is a better option. with drugs that inhibit CYP2D6 (Rev Prescrire

••Ivabradine, an inhibitor of the cardiac If current, n° 337). A “non-sedating” antihistamine without can cause visual disturbances, cardiovascular dis- antimuscarinic activity, such as cetirizine or lorata- orders (including myocardial infarction), potential- dine, is a better option in this situation. ly severe bradycardia and other cardiac arrhyth- ••Omalizumab in chronic spontaneous urticaria mias. It has no advantages in either angina or heart (also see the Pulmonology - ENT section on page 7) failure (Prescrire Int n° 88, 110, 118, 155, 165). Estab- (Prescrire Int n° 161). lished treatments shown to be effective in angina ••Injectable promethazine, an antihistamine used to include beta-blockers and the calcium channel treat severe urticaria, can cause thrombosis, skin blockers amlodipine and verapamil. There are also necrosis and gangrene following extravasation or better options for heart failure: one is to refrain inadvertent injection into an artery (Rev Prescrire from adding another drug to an optimised treat- n° 327). Injectable dexchlorpheniramine, which does ment regimen; another is to use a beta-blocker with not appear to carry these risks, is a better option. a proven impact on mortality. •• Topical tacrolimus, an immunosuppressant used ••Nicorandil, a vasodilator with solely symptomatic in atopic eczema, can cause skin cancer and efficacy as a preventive treatment in effort angina, lymphoma, yet its efficacy is barely different from can cause severe mucocutaneous ulceration that of topical corticosteroids (a)(Prescrire Int (Prescrire Int n° 81, 95, 110, 132). A nitrate is a better n° 101, 110, 131; Rev Prescrire n° 367). Judicious use option to prevent angina attacks. of a topical corticosteroid to treat flare-ups is a ••Olmesartan, an angiotensin II receptor blocker better option in this situation. (ARB or sartan) that is no more effective than other ARBs against the complications of hypertension, can cause sprue-like enteropathy leading to chron- Diabetes - Nutrition ic diarrhoea (potentially severe) and weight loss, and, possibly, an increased risk of cardiovascular mortality (Prescrire Int n° 148, 171). It is better to Diabetes. Various glucose-lowering drugs have choose another of the many ARBs available, such an unfavourable harm-benefit balance. They reduce as losartan or valsartan, which do not appear to blood glucose slightly but have no proven efficacy have these adverse effects against the complications of diabetes (cardiovas- ••Ranolazine, an antianginal with a poorly under- cular events, renal failure, neurological disorders, stood mechanism, provokes adverse effects that etc.) yet many adverse effects. Far more reasonable are disproportionate to its minimal efficacy in re- choices are to use a proven treatment such as met- ducing the frequency of angina attacks, including: formin, or a sulfonylurea such as glibenclamide or gastrointestinal and neuropsychiatric disorders, an insulin if metformin is insufficiently effective palpitations, bradycardia, hypotension, QT prolon- and, in some cases, setting a higher HbA1c target. gation and peripheral oedema (Prescrire Int n° 102; ••The gliptins (dipeptidyl peptidase 4 (DPP-4) inhibi- Rev Prescrire n° 350, 386 suppl. 2-3-7). tors) alogliptin, linagliptin, saxagliptin, sitagliptin ••Trimetazidine, a drug with uncertain properties, is and vildagliptin, used alone or in combination with used in angina despite its only modest symptom- metformin, have an unfavourable adverse effect atic effects (shown mainly in stress tests), yet it can profile that includes serious hypersensitivity reac- cause parkinsonism, hallucinations and thrombo- tions such as anaphylaxis and Stevens-Johnson cytopenia (Prescrire Int n° 84, 100, 106). It is better syndrome, infections (urinary tract and upper respira to choose better-known treatments for angina, such tory tract infections), pancreatitis, bullous pemphi- as beta-blockers or the calcium-channel blockers goid and intestinal obstruction (Prescrire Int n° 121, amlodipine and verapamil. 135, 138, 158, 167; Rev Prescrire n° 365, 366, 379). ••Vernakalant, an injectable antiarrhythmic used in ••Canagliflozin and dapagliflozin can provoke hypo atrial fibrillation, has not been shown to reduce tension, urinary tract and genital infections, renal mortality or the incidence of thromboembolic or failure, ketoacidosis, haematocrit elevation (a risk cardiovascular events. Its adverse effects include various arrhythmias (Prescrire Int n° 127). It is better to use amiodarone for pharmacological cardio- a- Oral or injectable tacrolimus is a standard immunosup- version. pressant for transplant recipients, and in this situation its harm-benefit balance is clearly favourable (Rev Prescrire n° 386 suppl. 10-1).

Prescrire International • April 2017 • Volume 26 N° 181 • Page 108-4 Downloaded from english.prescrire.org on 02/10/2021 Copyright(c)Prescrire. For personal use only. OUTLOOK

factor for thromboembolism), and possibly bladder, rectal preparations), used carefully and patiently, breast and prostate cancer ( Prescrire Int n° 147, 157, are safer than prucalopride. 169, 175). ••Pioglitazone has a burdensome adverse effect profile, including heart failure, bone fractures and Gynaecology - Endocrinology bladder cancer (Prescrire Int n° 129, 160). ••Tibolone, a synthetic steroid hormone used for Weight loss. As of early 2017, no drugs are capable postmenopausal hormone replacement therapy, of inducing lasting weight loss without risk. It is has androgenic, oestrogenic and progestogenic better to focus on dietary changes and physical ac- properties and carries a risk of cardiovascular dis- tivity, providing psychological support if necessary. orders, breast cancer and ovarian cancer (Prescrire ••The weight loss drug bupropion + naltrexone Int n° 83, 11, 137). When hormone therapy is chosen combines a drug chemically related to amphet- despite the inherent risks, the most reasonable amines (bupropion) with an opioid receptor antag- option is an oestrogen-progestogen combination, onist (also see the Psychiatry - Addiction section on used at the lowest possible dose and for the short- page 7) (Prescrire Int n° 164). est possible period. ••Orlistat has only a modest and transient effect on weight loss: patients lost about 3.5 kg more than with placebo over 12 to 24 months, with no evi- Infectious diseases dence of long-term efficacy. Gastrointestinal disorders are very common, while other adverse effects ••Moxifloxacin is no more effective than other include liver damage, hyperoxaluria, and bone fluoroquin olone antibiotics but can cause toxic fractures in adolescents. Orlistat alters the gastro- epidermal necrolysis and fulminant hepatitis and intestinal absorption of many nutrients (fat-soluble has also been linked to an increased risk of cardiac vitamins A, D, E and K), leading to a risk of deficien- disorders (Prescrire Int n° 62, 103; Rev Prescrire cy, and also reduces the efficacy of some drugs n° 371). Another fluoroquinolone such ascipro (thyroid hormones, some antiepileptics). Oral con- floxacin or ofloxacin is a better option. traceptive efficacy is reduced whenorlistat pro- ••Telithromycin has no advantages over other mac- vokes severe diarrhoea (Prescrire Int n° 57, 71, 107, rolide antibiotics but carries an increased risk of QT 110; Rev Prescrire n° 374). interval prolongation, hepatitis, visual disturbances and syncope (Prescrire Int n° n° 84, 88, 94, 106, 154). Another macrolide such as spiramycin or azithro- Gastroenterology mycin is a better option.

••Domperidone and droperidol, two neuroleptics, can provoke ventricular arrhythmias and sudden Neurology death. This is an unacceptable risk given the symptoms treated and these drugs’ weak efficacy against nausea and vomiting, and in the case of domperi- Alzheimer’s disease. The drugs available in done, gastroesophageal reflux Prescrire( Int n° 129, early 2017 for Alzheimer’s disease have only mini- 144, 175, 176, 179). Other drugs such as antacids mal and transient efficacy. They are also difficult to and omeprazole have a much better harm-benefit use because of their disproportionate adverse ef- balance in gastroesophageal reflux disease. In the fects and many interactions with other drugs. None rare situations in which treatment with an anti of the available drugs has been shown to slow emetic neuroleptic appears justified, it is better to progression toward dependence, yet all carry a risk choose metoclopramide, which also provokes seri- of life-threatening adverse effects and severe drug ous cardiac events but has proven efficacy against interactions (Prescrire Int n° 128, Rev Prescrire nausea and vomiting. It should be used at the n° 363, 364). It is better to focus on reorganising the lowest possible dose, taking drug interactions into patient’s daily life, keeping him or her active, and account and monitoring the patient frequently. providing support and help for caregivers and ••Prucalopride, a drug chemically related to neuro- family members. leptics, is authorised for chronic constipation but ••Donepezil, galantamine and rivastigmine, three shows only modest efficacy, in about one in six cholinesterase inhibitors, can cause gastrointestinal patients. Its adverse effect profile is poorly docu- disorders (including severe vomiting), neuro mented, particularly with respect to cardiovascular psychiatric disorders, cardiac disorders (including disorders (palpitations, ischaemic cardiovascular bradycardia, collapse and syncope), and cardiac events, possible QT prolongation), depression and conduction disorders (Prescrire Int n° 162, 166; Rev suicidal ideation and teratogenicity (Prescrire Int Prescrire n° 337, 340, 344, 349, 398). n° 116, 137, 175). There is no justification for expos- ••Memantine, an NMDA glutamate receptor antag- ing patients with simple constipation to such risks. onist, can cause neuropsychiatric disorders (such as If dietary measures are ineffective, then bulk- hallucinations, confusion, dizziness and headache) forming laxatives, osmotic laxatives or, very occa- that can lead to violent behaviour, as well as seizures sionally, other laxatives (lubricants, stimulants, or and heart failure (Rev Prescrire n° 359, 362, 374).

Prescrire International • April 2017 • Volume 26 N° 181 • Page 108-5 Downloaded from english.prescrire.org on 02/10/2021 Copyright(c)Prescrire. For personal use only. OUTLOOK

Multiple sclerosis. The standard “disease- ••Nintedanib, a tyrosine kinase inhibitor authorised modifying” treatment for multiple sclerosis is inter- as add-on therapy in combination with docetaxel feron beta, despite its limitations and many adverse for certain types of non-small cell lung cancer, has effects. The harm-benefit balance of the other not been shown to prolong survival but can pro- disease-modifying treatments is no better and some- voke many serious adverse effects related to its times clearly unfavourable. This applies in particular inhibitory effects on angiogenesis, including: ven to three immunosuppressants that have dispropor- ous thromboembolism, bleeding, hypertension, tionate adverse effects and should be avoided. gastrointestinal perforations and impaired wound ••Alemtuzumab, an antilymphocyte monoclonal healing (Rev Prescrire n° 389). antibody, has no proven efficacy and can provoke ••Olaparib has not been shown to prolong survival many serious and sometimes fatal adverse effects, when used as maintenance treatment for advanced in particular: infusion-related reactions (including ovarian cancer in women in remission. It has seri- atrial fibrillation and hypotension), infections, fre- ous adverse effects: haemopoietic disorders, quent autoimmune disorders (including autoim- myelodysplastic syndrome, acute myeloid leukae- mune thyroid disease, immune thrombocytopenic mia (Prescrire Int n° 178). purpura, cytopenia and renal disease) (Prescrire Int ••Panobinostat has not been shown to prolong n° 158; Rev Prescrire n° 384). survival in refractory or relapsed multiple mye ••Natalizumab, another monoclonal antibody, can loma, and provokes many, often serious, adverse lead to sometimes fatal opportunistic infections, effects that affect many vital functions, hastening including progressive multifocal leukoencephalop- the death of many patients (Prescrire Int n° 176). athy, potentially serious hypersensitivity reactions, ••Trabectedin showed no tangible efficacy in com- and liver damage (Rev Prescrire n° 330, 398; parative trials in ovarian cancer or soft-tissue sar- Prescrire Int n° 122, 158, 182). comas but has very frequent and severe gastro ••Teriflunomide has serious and potentially fatal intestinal, haematological, hepatic and muscular adverse effects, including liver damage, leukopenia adverse effects (Prescrire Int n° 102, 120; Rev and infections. There is also a risk of peripheral Prescrire n° 360). It is unreasonable to add trabec- neuropathy (Prescrire Int n° 158). tedin to platinum-based chemotherapy for ovarian cancer. When chemotherapy is ineffective in pa- Miscellaneous. A number of drugs used in mi- tients with soft-tissue sarcomas, it is best to focus graine and Parkinson’s disease should also be on appropriate supportive care. avoided. ••Vandetanib has not been shown to prolong sur- ••Flunarizine and oxetorone, two neuroleptics used vival in patients with metastatic or inoperable med- to prevent migraine attacks, have at best only mod- ullary thyroid cancer. Too many patients were lost est efficacy flunarizine( prevents about one attack to follow-up in placebo-controlled trials to show an every two months) but can cause extrapyramidal increase in progression-free survival. Serious ad- disorders, cardiac disorders and weight gain verse effects (diarrhoea, pneumonia, hypertension) (Prescrire Int n° 137). It is better to choose another occur in about one-third of patients. There is also a drug such as propranolol. risk of interstitial lung disease, torsades de pointes ••Tolcapone, an antiparkinsonian COMT inhibitor, and sudden death (Prescrire Int n° 131). can cause life-threatening liver damage (Prescrire ••Vinflunine has uncertain efficacy in advanced and Int n° 82, Rev Prescrire n° 330). When other treat- metastatic bladder cancer. A clinical trial provided ment options have been exhausted, entacapone is weak evidence that vinflunine increases median a better option. survival by two months at best compared with symptomatic treatment. There is a high risk of haematological adverse effects (including aplastic anaemia), Oncology - Haematology and a risk of serious infections and cardiovascular disorders (torsades de pointes, myocardial infarction, ischaemic heart disease), sometimes resulting in Antineoplastics. Various antineoplastic drugs death (Prescrire Int n° 112; Rev Prescrire n° 360). have a clearly unfavourable harm-benefit balance. They are often authorised for situations in which Cancer or cancer therapy complications. other treatments are ineffective. When exposure to Some drugs are authorised for the treatment of highly toxic drugs is not justified by proven benefits, cancer complications, such as the peritoneal com- it is better to focus on tailored symptomatic treat- plication malignant ascites, or for complications of ment and on improving the patient’s quality of life. cancer therapy. ••Mifamurtide is authorised as add-on therapy in ••Catumaxomab, for malignant ascites, provokes combination with other chemotherapy for osteo serious and possibly fatal adverse effects in more sarcoma but has not been shown to prolong sur than three-quarters of patients (Prescrire Int n° 109). vival and can provoke serious hypersensitivity re- Paracentesis is a better option, repeated as neces- actions, pleural and pericardial effusions, sary to control symptoms. neurological adverse effects and hearing loss ••Defibrotide, an antithrombotic authorised for se- (Prescrire Int n° 115; Rev Prescrire n° 341). It is bet- vere hepatic veno-occlusive disease following ter to propose chemotherapy without mifamurtide. haemopoietic stem cell transplantation, had no

Prescrire International • April 2017 • Volume 26 N° 181 • Page 108-6 Downloaded from english.prescrire.org on 02/10/2021 Copyright(c)Prescrire. For personal use only. OUTLOOK

more impact on mortality or complete disease re- cardia, arrhythmias, QT prolongation) due to their mission than symptomatic treatment in an unblind- noradrenergic activity. In addition, venlafaxine ed trial, but provokes sometimes fatal haemor overdoses are associated with a high risk of cardiac rhages (Prescrire Int n° 164). It is better to focus on arrest (Rev Prescrire n° 338, 343, 386; Prescrire Int preventive measures and symptomatic treatments. n° 170). ••Tianeptine, a drug with no proven efficacy, can cause hepatitis, life-threatening skin reactions (in- Ophthalmology cluding bullous rash) and addiction (Prescrire Int n° 127, 132). ••Ciclosporin eye drops, authorised for the treatment of dry eye disease with severe keratitis, fre- Other psychotropic drugs. Some other psycho quently provoke eye pain and irritation, have im- tropic drugs have unacceptable adverse effects: munosuppressive effects and may cause ocular or ••Dapoxetine, a “selective” SRI, is used for prema- periocular cancer, yet had no proven efficacy when ture ejaculation with sexual dissatisfaction. Its ad- compared with the same eye drops without ciclo- verse effects are disproportionate to its very mod- sporin (b)(Prescrire Int n° 181). It is better to use est efficacy and include aggressive outbursts, artificial tears for example for symptomatic relief, serotonin syndrome, and syncope (Prescrire Int several types of which are available. n° 105; Rev Prescrire n° 355). A psychological and ••Idebenone was no more effective than placebo in behavioural approach is a better option in this situ a trial in Leber’s hereditary optic neuropathy, and ation. carries a risk of adverse effects including hepatic ••Etifoxine, a drug poorly evaluated in anxiety, can disorders (Prescrire Int n° 179). As of 2017, there are cause hepatitis and severe hypersensitivity reac- no drugs with a favourable harm-benefit balance tions (including DRESS syndrome, Stevens- for patients with this rare disease. Johnson syndrome and toxic epidermal necrolysis) (Prescrire Int n° 136; Rev Prescrire n° 376). When an anxiolytic drug is justified, it is better to choose a Psychiatry - Addiction benzodiazepine, for the shortest possible period.

Smoking cessation. One drug authorised as a Antidepressants. Several drugs authorised for smoking cessation aid is no more effective than depression carry a greater risk of severe adverse nicotine and has more adverse effects. When a drug effects than other antidepressants without offering is needed to help with smoking cessation, nicotine greater efficacy. In general, antidepressants have is a better choice. only modest efficacy and often take some time to ••Bupropion, an amphetamine, can cause neuro- work. It is better to choose one of the longer estab- psychiatric disorders (including aggressiveness, lished antidepressants with an adequately docu- depression and suicidal ideation), potentially se- mented adverse effect profile. vere allergic reactions (including angioedema and ••Agomelatine has no proven efficacy beyond the Stevens-Johnson syndrome), addiction, and con- placebo effect, but can cause hepatitis and pancrea genital heart defects in children exposed to the drug titis, suicide and aggression, as well as serious skin in utero (Prescrire Int n° 131; Rev Prescrire n° 377). disorders including Stevens-Johnson syndrome (Prescrire Int n° 136, 137; Rev Prescrire n° 397). ••Duloxetine, a serotonin and norepinephrine re- Pulmonology - ENT uptake inhibitor, not only has the adverse effects of the so-called “selective” serotonin reuptake inhibi- ••Decongestants for oral or nasal use (ephedrine, tors (SSRIs) but also carries a risk of cardiac disorders naphazoline, oxymetazoline, phenylephrine, (hypertension, tachycardia, arrhythmias, etc.) due to pseudoephedrine and tuaminoheptane) are sym- its noradrenergic activity. Duloxetine can also cause pathomimetic vasoconstrictors. They can cause hepatitis and severe cutaneous hypersensitivity re- serious and even life-threatening cardiovascular actions such as Stevens-Johnson syndrome disorders, including hypertensive crisis, stroke and (Prescrire Int n° 85, 100, 111, 142; Rev Prescrire n° 384). arrhythmias, as well as ischaemic colitis. These ••Citalopram and escitalopram are SSRI antidepres- adverse effects are unacceptable for drugs indi sants that expose patients to a higher incidence of cated for minor, rapidly self-resolving ailments such QT prolongation and torsades de pointes than as the common cold (Prescrire Int n° 136, 172, 178; other SSRIs and worse outcomes in the event of 183; Rev Prescrire n° 312, 342, 345, 348, 361). overdose (Rev Prescrire n° 369, 396; Prescrire Int n° 170, 174). ••Milnacipran and venlafaxine, two non-tricyclic, non-SSRI, non-monoamine oxidase inhibitor (MAOI) antidepressants, have both serotonergic b- Oral or injectable ciclosporin is a standard immunosup- and noradrenergic activity. Not only do they have pressant for transplant recipients, and in this situation its the adverse effects of SSRI antidepressants, they harm-benefit balance is clearly favourable (Rev Prescrire also cause cardiac disorders (hypertension, tachy- n° 386 suppl. 10-1).

Prescrire International • April 2017 • Volume 26 N° 181 • Page 108-7 Downloaded from english.prescrire.org on 02/10/2021 Copyright(c)Prescrire. For personal use only. OUTLOOK

••Ambroxol and bromhexine are mucolytics with no ••Cox-2 inhibitors (coxibs) such as celecoxib, proven efficacy beyond a placebo effect, yet they etoricoxib and parecoxib have been linked to an carry a risk of anaphylactic reactions and severe, excess of cardiovascular events (including myocar- sometimes fatal cutaneous reactions such as ery- dial infarction and thrombosis) and skin reactions thema multiforme, Stevens-Johnson syndrome and by comparison with other equally effective NSAIDs toxic epidermal necrolysis (Rev Prescrire n° 400). (Prescrire Int n° 167; Rev Prescrire n° 344, 361, 374). These adverse effects are unacceptable for drugs •• Oral aceclofenac and oral diclofenac cause more used to relieve sore throat or cough. cardiovascular adverse effects (including myocar- ••Pholcodine, an opioid used as an antitussive, can dial infarction and heart failure) and more cardio- cause sensitisation to neuromuscular blocking vascular deaths than other equally effective NSAIDs agents used in general anaesthesia (Rev Prescrire (Prescrire Int n° 167; Rev Prescrire n° 362, 374). n° 349, 400). This serious adverse effect is not ••Ketoprofen gel causes more photosensitivity re- known to occur with other opioids. Cough is a actions (eczema, bullous rash) than other equally minor ailment that does not warrant taking such effective topical NSAIDs (Prescrire Int n° 109, 137). risks. When drug therapy is required for cough, it ••Piroxicam, when used systemically, carries an is better to choose dextromethorphan, despite its increased risk of gastrointestinal and cutaneous limitations (Rev Prescrire n° 358). disorders (including toxic epidermal necrolysis) but ••Tixocortol (sometimes combined with chlorhexi- is no more effective than other NSAIDs (Rev dine), a corticosteroid authorised for sore throat, P rescrire n° 321). can cause allergic reactions such as facial mucocutaneous oedema, glossitis or angioedema (Rev Osteoporosis. Several drugs authorised for osteo Prescrire n° 320). When a drug is needed to relieve porosis are best avoided because their efficacy is at sore throat, paracetamol is a better option, when best modest and they have potentially serious ad- taken at the appropriate dosage. verse effects. When non-drug measures plus calcium ••Omalizumab, an anti-IgE monoclonal antibody and vitamin D supplementation prove inadequate, approved for severe persistent asthma and chronic alendronic acid or an alternative, raloxifene, have a spontaneous urticaria, and the anti-interleukin-5 better harm-benefit balance than other options, monoclonal antibody mepolizumab, approved for despite the significant limitations of both drugs. severe asthma, provoke disproportionate adverse ••Denosumab 60 mg in osteoporosis has very mod- effects: infections, hypersensitivity reactions and est efficacy in the prevention of osteoporotic frac- cardiac disorders in the case of omalizumab tures and no efficacy for “bone loss” during pros- (Prescrire Int n° 115, 161; 179). Corticosteroid ther- tate cancer, but carries a disproportionate risk of apy at the lowest effective dose is a better option adverse effects, including back pain, musculoskel- in both of these situations. etal pain, and serious infections (including endo- ••Mannitol inhalation powder, authorised as a carditis) due to the immunosuppressive effects of mucolytic for patients with cystic fibrosis despite this monoclonal antibody (Prescrire Int n° 117, 130, the lack of convincing evidence of efficacy, can 168). There is no known satisfactory drug treatment cause bronchospasm and haemoptysis (Prescrire for “bone loss” (c). Int n° 148). It is best to choose other mucolytics such ••Strontium ranelate has only modest efficacy in the as dornase alfa in the absence of a better alterna- prevention of recurrent vertebral fractures. Yet its tive. adverse effects include neuropsychiatric disorders, ••Nintedanib, a tyrosine kinase inhibitor, has not cardiovascular disorders (including venous throm- been shown to prolong survival, prevent the pro- bosis and pulmonary embolism, myocardial infarc- gression of fibrosis or relieve symptoms in patients tion and cardiovascular death), and hypersensitivity with idiopathic pulmonary fibrosis, whereas it reactions including toxic epidermal necrolysis and causes hepatic disorders and many serious adverse DRESS (drug reaction with eosinophilia and system- effects due to its inhibitory effect on angiogenesis, ic symptoms) (Prescrire Int n° 117, 125, 139, 142, 156). including: venous thromboembolism, bleeding, hypertension, gastrointestinal perforations and im- Osteoarthritis. Drugs authorised for their sup- paired wound healing (Prescrire Int n° 173). It is posed effect on the process that results in osteo better to focus on symptomatic treatment. arthritis should be avoided because they have significant adverse effects but no proven efficacy beyond the placebo effect. There are no drugs with Rheumatology - Pain efficacy against joint degeneration and a favourable harm-benefit balance.

Certain nonsteroidal anti-inflammatory drugs. Although nonsteroidal anti-inflammatory drugs (NSAIDs) share a similar adverse effect profile, c- A 120-mg strength denosumab product is authorised for some expose patients to less risk. When paracetamol use in patients with bone metastases from solid tumours. proves inadequate, ibuprofen and naproxen, used In this situation, denosumab offers no tangible advantages, at the lowest effective dose and for the shortest but its harms do not clearly outweigh its benefits (Prescrire possible period, are the least risky options. Int n° 130).

Prescrire International • April 2017 • Volume 26 N° 181 • Page 108-8 Downloaded from english.prescrire.org on 02/10/2021 Copyright(c)Prescrire. For personal use only. OUTLOOK

Drugs that should not be used: swifter action required to protect patients

his fifth edition of our annual review “Towards better marketing authorisation, granted in 2007, remains valid (1). Tpatient care: drugs to avoid” provides an opportunity The same is true of asenapine, a neuroleptic authorised in to examine the decisions taken in France by regulators and 2010 (Rev Prescrire n° 388) (1). pharmaceutical companies to protect patients from these drugs. Delisting: a slow process, sometimes challenged, sometimes partial. If a drug’s marketing authorisation is upheld, Various measures available. Regulators can decide to with- particularly at European level, a stopgap measure is to draw or suspend a drug’s marketing authorisation, remove reduce the number of patients exposed to it through de it from the list of drugs that qualify for reimbursement by listing, i.e. removal from the list of products that qualify the national health insurance system, or reduce the percent- for reimbursement by the French health insurance system. age of its cost that is reimbursed. Pharmaceutical com A number of trimetazidine-containing products, including panies can decide to stop marketing a product. copies, are still available in early 2017 despite being de Various combinations of these measures have been listed in 2012, suggesting that significant quantities are applied between 2013 and 2016, each to only a handful of still sold (Rev Prescrire n° 342). Strontium ranelate remains drugs, some of which are discussed below. available in early 2017 despite being delisted in 2015 (Rev Prescrire n° 377). Marketing authorisations suspended or withdrawn for Some delisting decisions have been challenged in court about 10 drugs, and some half-measures. Prescrire’s five by the pharmaceutical companies, as was the case for annual reviews during this period have identified about a ketoprofen gel (Prescrire Int n° 109, 112; Rev Prescrire hundred drugs to avoid, but only about ten have been n° 317), diacerein, glucosamine and olmesartan (Rev withdrawn from the market through suspension or with- Prescrire n° 395, 380). The French health minister has asked drawal of the marketing authorisations for products con- for a “treatment protocol” to be drawn up before consid- taining them. The French Health Products Agency (ANSM) ering delisting the 4 drugs for Alzheimer’s disease (Rev has taken such action far more frequently than the Euro- Prescrire n° 398) (3). pean Medicines Agency. Sometimes a drug is delisted for certain authorised indi- A number of long-marketed drugs had their marketing cations, while other authorised indications are reimbursed authorisations suspended in 2013: products containing at a reduced rate. For example, topical tacrolimus was meprobamate (Prescrire Int n° 148) and 5 ergot derivatives granted European marketing authorisation in 2002, then (Rev Prescrire n° 364). in 2014 its reimbursement by the French health insurance The marketing authorisation for indoramin was with- system was revoked for children and reduced for adults drawn in 2013, after 28 years on the market (Rev Prescrire (Rev Prescrire n° 245, 367). n° 356) (1). The marketing authorisation for floctafenine Sometimes reimbursement is reduced to 15% of the was revoked in 2015, after 40 years on the market (Rev product’s cost, which was the case in 2016 for agomelatine, Prescrire n° 321, 384) (1,2). authorised since 2009 (Rev Prescrire n° 397). Marketing authorisation for domperidone 20 mg was Mifamurtide and vernakalant obtained European mar- withdrawn in 2014, after years of procedures (Prescrire Int keting authorisations years ago but are not on sale in n° 175). But the 10-mg strength, authorised in France since France, perhaps due to the unfavourable opinion issued 1980, remains on the market (1). by the committee responsible for recommending whether new drugs should be funded by the national health insur- A few market withdrawals by pharmaceutical com ance system (see inset p. 108-3). panies. A theodrenaline + cafedrine combination and nimesulide were withdrawn from the market in 2013 In summary: do not wait for companies or regulators (Prescrire Int n° 147, Rev Prescrire n° 364). In 2014, a to act. The actions taken from 2013 to 2016 by regulators quinine-containing suppository for cramps was also with- and pharmaceutical companies to rid the market of drugs drawn (Rev Prescrire n° 377). Their French marketing that are more dangerous than beneficial have been slow authorisations therefore became null and void (1). and piecemeal, especially at European level. Quinine Vitamine C Grand° (Rev Prescrire n° 400) has It is in patients’ and health professionals’ interests to not been marketed since 2014, but its French marketing take matters into their own hands by avoiding these drugs authorisation, granted in 1997, remains valid, and other now. oral quinine-containing products for cramps are still avail- ©Prescrire able. In 2016, the European marketing authorisation for 1- ANSM “Répertoire des spécialités pharmaceutiques”. http://agence-prd. pegloticase, granted in 2013, was withdrawn when the ansm.sante.fr/php/ecodex/ accessed 6 January 2017. company stopped marketing it (Prescrire Int n° 180). 2- HAS - Commission de la transparence “Avis-Idarac” 15 November 2006. 3- Agence France Presse “Médicaments anti-Alzheimer: Touraine écarte le Iron dextran ceased to be marketed in France in 2015 “déremboursement”” 26 October 2016. www.leparisien.fr accessed (Rev Prescrire n° 349; Prescrire Int n° 151). Its European 6 January 2017: 3 pages.

Prescrire International • April 2017 • Volume 26 N° 181 • Page 108-9 Downloaded from english.prescrire.org on 02/10/2021 Copyright(c)Prescrire. For personal use only. OUTLOOK

••Diacerein causes gastrointestinal disorders (in- ous application, expose patients to the adverse cluding gastrointestinal bleeding and melanosis effects of corticosteroids and to salicylate hyper- coli), angioedema and hepatitis (Rev Prescrire sensitivity reactions. Other drugs such as oral n° 282, 321; Prescrire Int n° 159). paracetamol (at the appropriate dosage) and topical ••Glucosamine causes allergic reactions (angioede- ibuprofen have a better harm-benefit balance in ma, acute interstitial nephritis) and hepatitis patients with painful sprains or tendinopathy, in (Prescrire Int n° 84, 137; Rev Prescrire n° 380). conjunction with non-drug measures (rest, ice, splints). Miscellaneous. A number of other drugs used for specific types of pain or in rheumatology are best avoided. Putting patients first ••Capsaicin, a red chilli pepper extract authorised in patch form for neuropathic pain, is barely more Our analyses show that the harm-benefit balance effective than placebo but can provoke irritation, of the drugs listed here is unfavourable in all their severe pain and burns (Prescrire Int n° 108, 180). authorised indications. Yet some have been mar- Capsaicin remains an unreasonable choice even keted for many years and are commonly used. How when systemic pain medications or local ones such can one justify exposing patients to drugs that have as lidocaine medicated plasters fail to provide ad- more adverse effects than other members of the equate relief. same pharmacological class or other similarly ef- ••Muscle relaxants with no proven efficacy beyond fective drugs? And what justification is there for the placebo effect: methocarbamol has many ad- exposing patients to drugs with severe adverse verse effects, including gastrointestinal and cuta- effects but no proven impact (beyond the placebo neous disorders (angioedema), while thiocolchico- effect) on patient-relevant clinical outcomes? side, which is related to colchicine, causes It is necessary but not sufficient for healthcare diarrhoea, stomach pain, photodermatosis and professionals to remove these drugs from their list possibly convulsions; it is also genotoxic and tera- of useful treatments: regulators and health author- togenic (Prescrire Int n° 168, Rev Prescrire n° 282, ities must also take concrete steps to protect pa- 321, 313, 367, 400). There is no justification for ex- tients and promote the use of treatments that have posing patients to these adverse effects for such an acceptable harm-benefit balance. little efficacy. An effective analgesic such aspara The drugs listed above are more dangerous than cetamol is a better option, when taken at the ap- beneficial. There is no valid reason for them to re- propriate dosage. tain their marketing authorisations or continue to ••Quinine for cramps can have life-threatening ad- be marketed. verse effects including anaphylactic reactions, ©Prescrire haematological disorders (including thrombocyto- ▶▶ Translated from Rev Prescrire February 2017 penia, haemolytic anaemia, agranulocytosis, and Volume 37 N° 400 • Pages 137-148 pancytopenia) and cardiac arrhythmias. These adverse effects are disproportionate in view of its poor efficacy (d)(Rev Prescrire n° 337, 344). There are no drugs with a favourable harm-benefit bal- d- Quinine is a recommended treatment for malaria (Rev Prescrire n° 360). ance for patients with cramps. Regular stretching can be beneficial Rev( Prescrire n° 362). ••Colchimax° (colchicine + opium powder + tiemon ium) should be avoided in gout attacks because the action of opium powder and tiemonium can mask Selected references from Prescrire’s literature search the onset of diarrhoea, which is an early sign of 1- Prescrire Editorial Staff “Towards better patient care: drugs to avoid potentially fatal colchicine overdose (Prescrire Int in 2016” Prescrire Int 2016; 25 (170): 105-111. n° 147). A nonsteroidal anti-inflammatory drug or 2- Prescrire Editorial Staff “Towards better patient care: drugs to avoid” Prescrire Int 2013; 22 (137): 108-111. colchicine alone are better options for gout attacks. 3- Prescrire Rédaction “Des médicaments à écarter pour mieux soi- The dexamethasone + salicylamide + hydroxy gner: pourquoi?” Rev Prescrire 2013; 33 (360): 792-795. ethyl salicylate combination (Rev Prescrire n° 345) 4- Prescrire Editorial Staff “Determining the harm-benefit balance of an intervention: for each patient” Prescrire Int 2014; 23 (154): 274-277. and the prednisolone + dipropylene glycol salicy- 5- Prescrire Editorial Staff “Treatment goals: discuss them with the late combination (Rev Prescrire n° 338), for cutane- patient” Prescrire Int 2012; 21 (132): 276-278.

Prescrire International • April 2017 • Volume 26 N° 181 • Page 108-10 Downloaded from english.prescrire.org on 02/10/2021 Copyright(c)Prescrire. For personal use only.