Polypharmacy increases the risk of adverse effects and medicine interactions
The majority of older people have more than one medical condition, more than one prescriber and take more than one medicine.1 Polypharmacy
increases the risk of adverse effects and medicine interactions and the physiological changes that accompany ageing alter
the handling and response to medicines. These factors make prescribing decisions, about both starting and stopping medicines
in older people, difficult. The net result may be that the medicine regimen poses more risks than benefits.
The use of evidence based guidelines tends to increase the number of medicines required to meet specific targets, however
this may not always be appropriate in older people because they are often poorly represented in clinical trials.2,3 Guidelines
are population based and do not take into account individual variation, co-morbidities and patient preference. The clinical
status of older people and the approach to decisions about the use of medicines for this population vary widely. There
is often no clear consensus.
Individualised assessment that includes a review of the need for each drug can simplify treatment regimens and reduce
the potential for harm. Medicine withdrawal may be the best clinical decision and result in significant clinical
benefits including a reduction in falls. The key reasons for stopping medicines in older people include a decreased
risk of adverse effects, a reduction in the potential for medicine interactions and to simplify a prescription regimen.
How do you decide which medicines can be stopped?
Reviewing the treatment needs of an older patient who is taking multiple medicines can be a complex and time consuming
task. It can be difficult to know which medicine, if any, to stop. In some circumstances, the only way to know whether
or not to stop a medicine is to actually stop it and see what happens.4
One way of considering if any medicines are able to be stopped is to group them into two categories – those that improve
day-to-day quality of life and those that prevent future illness (see Box below). Thinking about medicines in this way
may provide a good starting point for talking with patients about the role and importance of each medicine. It may also
allow a discussion about any variance between the goals of the doctor and those of the patient – the question is “what
is the clinical and personal significance of this specific drug for this particular patient?”
An estimation of life expectancy may be difficult and needs to take into account not only the person’s age but
also how many medical problems they have, how severe they are and how well they function.5 Diagnosis of a life-limiting
illness may make the decision to stop long term medicines easier, but in reality, approaches range from stopping everything
(except palliative medicine) to continuing all medicines until the patient is physically unable to take them.6
Best practice tip: Consider medicine regimens to have an expiry or “best
before” date. Make sure that medicine reviews are carried out by this date unless there has been a change in the
patient’s clinical condition that prompts an earlier review. Use your PMS to make sure that reviews are done at
least annually for older people.
Factors to consider when deciding if a medicine can be stopped include:
- The wishes of the patient
- Clinical indication and benefit
- Duration of use
- The prescribing cascade
Medicines can be grouped as:
- Those that keep the patient well and improve day-to-day quality of life e.g. analgesics, thyroxine or anti-anginals.
In some cases, if these medicines are stopped, the patient may become ill or unable to function. However, some drugs
may be able to be stepped down, stopped or used on an as required basis (prn) e.g. a protein pump inhibitor (PPI).
- Those that are used for the prevention of illness in the future e.g. statins, aspirin, warfarin or bisphosphonates.
A decision about whether to stop medicines such as these should include consideration of the risks and benefits of treatment
for that particular patient, the length of time required for benefit and the life expectancy of the patient.
The wishes of the patient
The majority of people who take medicines would prefer not to, or at least to take only those that are really needed.
Patients often ask if any of their pills can be stopped. It is easy to say “no” and continue on with the consultation
but ideally this query should be viewed as a chance for a medicine review. Suggest a “brown bag review”, where
the patient brings in all their medicine, including over-the-counter and complementary/herbal medicines and also those
that they are not actually taking but are still in their cupboard. Consider involving the practice nurse and pharmacist
for a team approach.
See BPJ 11 (Feb, 2008) “Principles
of prescribing for elderly people” for more information on a “brown bag review”.
Check that there is still a valid clinical indication and ongoing clinical benefit for each medicine. It may be necessary
to review the patient notes to check what the original indication was. Consider if the benefit of the drug has already
been achieved or if the clinical condition of the patient has changed.
Examples may include:
- A patient no longer taking a NSAID who now may not need a gastroprotective agent
- A patient with treated congestive heart failure who may be able to have their diuretic reduced or stopped
- A patient who has been successful with lifestyle modifications (such as a healthier diet, weight loss and smoking
cessation) who may no longer require an antihypertensive
- A patient diagnosed with a terminal illness and a prognosis that makes the benefit of preventative medicines such
as bisphosphonates and statins unlikely to be realised in that patient’s lifetime
Check that the medicine is appropriate for use in an older person (see below). Two large studies have shown that approximately
21% of medicines used in older people may be inappropriate.7,8 Some medicines are less safe for use in older
people and more likely to cause adverse effects or have an increased risk of medicine interactions.
Defining inappropriate medicine use
Several screening tools have been developed to help identify potentially inappropriate prescribing in older people
such as the Beers’ criteria and the more recent STOPP/START tools (Screening Tool of Older Person’s Prescriptions
and Screening Tool to Alert doctors to Right Treatment).10 Although widely used in research, these tools currently
have limited value in primary care because they are not easy to use, time consuming and the Beers’ criteria in
particular includes medicines that in some situations may be appropriate e.g. dipyridamole, oxybutinin and doxazosin.
These tools should not be used as a substitute for careful clinical judgement.11
Examples of medicines that may be inappropriate in older people include:
- Amitriptyline (particularly doses > 50mg), which is likely to cause sedation and also has strong anticholinergic actions
- nortriptyline is recommended as a safer alternative if a tricyclic is required.
- Benzodiazepines, which can cause excessive sedation and increase the risk of falls.
- Dextropropoxyphene, which can cause confusion and excessive sedation particularly in older people. Evidence shows
that it is no more effective for pain than regular paracetamol use.9 Due to this unfavourable risk-benefit
balance, dextropropoxyphene is being withdrawn from the New Zealand market on 1 August 2010.
See BPJ 26 (Mar, 2010) “Dextropropoxyphene
containing medicines to be withdrawn”.
Consider if there are any other solutions that may be effective e.g. diet modification instead of a statin for reducing
cholesterol, or non-pharmacological treatment instead of an antipsychotic for the behavioural and psychological symptoms
One way to “stop” medicines is not to start them in the first place
Prescribing for older people can be difficult. When considering any new medicine for an older person, check if it is
appropriate by considering the following questions: (Adapted from Holmes, 2006)2
- Is there an indication for the drug?
- Is the medicine effective for the condition?
- Are there clinically significant drug-drug interactions?
- Are there clinically significant drug-disease/condition interactions?
- Is there unnecessary duplication with other medicines?
- Is the likely duration of therapy known and acceptable to both doctor and patient?
- Will the patient take the medicine – what are the likely adverse effects, is the dose correct, are the directions
- Is this medicine the least expensive alternative compared with others of equal usefulness?
See BPJ 11 (Feb, 2008) “Dilemmas
in prescribing for older people”
Duration of use
Check how long the patient has been on the medicine. Some medicines are repeated for years as it can be quicker and
simpler to maintain the status quo.
Check if there was a clear understanding at the time of initiation about the expected duration of use, particularly
if the medicine was started in a secondary care setting. For example, a patient who was initiated on dipyridamole after
a TIA, who has had no further ischaemic events, may be able to stop this medicine after two years.
Check whether there is still an indication for the use of the medicine and if so, that its use is still consistent with
recommendations in current guidelines. Consider if there are more up-to-date drugs on the market that may be superior
Check if the patient is taking all of their prescribed medicines. If not, ask them why? Did they understand the aims
of treatment and did they experience any adverse effects? If the patient has remained well without the medicine, and is
unlikely to suffer harm if it is not taken, consider stopping it e.g. corticosteroid inhalers for COPD.
The prescribing cascade
When a patient presents with new symptoms, consider an adverse medicine reaction as a possible cause. The aim is to
avoid the prescribing cascade where additional medicines are initiated to treat adverse effects (both recognised and unrecognised)
of other medicines.
- A patient taking 50 mg amitriptyline for pain presents with incontinence and/or constipation. Amitriptyline has anticholinergic
actions which can cause urinary retention leading to overflow incontinence. If this is not recognised, oxybutynin may
be prescribed, which aggravates the incontinence because it also has anticholinergic actions. In addition, the patient
then becomes constipated and a laxative is prescribed. One medicine has led to the use of three others. Stopping the
amitriptyline and finding an alternative medicine for the pain may be the best action
- A patient taking a calcium channel blocker presents with ankle swelling. Avoid prescribing a diuretic as they are
not effective in this situation.
- For any patients taking NSAIDs who develop dyspepsia, hypertension or heart failure consider the NSAID as a potential
How to stop medicines
Take a stepwise approach to stopping medicines
A four step process can be used when stopping medicines:13
- Recognise the need to stop
- Reduce or stop one medicine at a time
- Consider if the medicine can be stopped abruptly or should be tapered
- Check for benefit or harm after each medicine has been stopped
Recognise the need to stop a medicine
When the patient presents for a renewal of medicine ask if they have any new symptoms (including adverse effects)
or any concerns about their medicine. Has the clinical condition of the patient changed? Consider the preferences of
Are there drugs that can be stopped? If more than one medicine can be stopped, which one should be stopped first? This
relies on clinical judgement and consideration of factors such as medicines most likely to cause adverse effects or without
Reduce or stop one medicine at a time
Try to reduce or stop only one medicine at one time. If problems develop it is then easier to know what the likely
cause may be.
Taper medicines when appropriate
To reduce the likelihood of an adverse withdrawal event, many medicines should be tapered. It can be difficult
to determine which can just be stopped and which should be tapered. Therefore if in doubt taper, as it is safer. For
many medicines the first step in tapering is to halve the dose. The dose should then be tapered in a stepwise manner
to establish if the patient’s symptoms, conditions or risks can be managed with a lower dose or whether the medicine
can be stopped completely.13
Generally there will be plenty of time to taper a medicine. If the medicine is being discontinued because toxicity is
a concern, then a more ambitious taper can be undertaken or the medicine stopped abruptly. Once tapering has begun, ask
the patient to note any symptoms that may suggest a more gradual withdrawal is required e.g. dyspepsia with reduced dose
If intolerable symptoms occur following a decrease in the dose or after the medicine has been stopped, then it may be
necessary to restart the previously prescribed dose and then try tapering again, but at a more gradual rate.
Check for benefit or harm after each medicine has been stopped
Ask the patient if any changes have occurred after a medicine has been stopped. Beneficial effects should be
noted to reinforce that the decision to reduce or stop the medicine was correct. There is also evidence that the beneficial
effects of some medicines may persist even after the medicine is stopped e.g. treatment with a bisphosphonate for five
years, gives an ongoing reduction in risk of fracture for a further five years.14
If symptoms of the initial condition return and are troublesome, despite gradual tapering, then it may be that the medicine
cannot be stopped completely. The patient may however be able to be managed on a reduced dose e.g. 10 mg PPI rather than
|A general guide to tapering medicine:
|Halve the dose. At the next scheduled visit review progress, then either:
- Maintain (at half dose)
- Continue to taper (e.g. quarter dose)
- View the discontinuation process as a trial
- Stop one medicine at a time so that any withdrawal event(s) can be easily attributed to the medicine that is being
- Time taken to taper may vary from days to weeks to months
Specific guidance on stopping medicines
| Oral Corticosteroids
Antidepressants should be tapered rather than stopped abruptly, to reduce the risk of developing a discontinuation syndrome
and to allow time to assess the possible re-emergence of depressive symptoms (Table 1). Antidepressant discontinuation
syndrome is more likely with a longer duration of treatment and a shorter half-life of the treatment drug.15
Antidepressants should normally be withdrawn over at least a four week period. Patients may experience withdrawal symptoms
but usually these are mild and self-limiting. If these symptoms are not tolerated, it may be necessary to resume the previous
dose and then reduce the antidepressant more slowly.16
Antidepressant Discontinuation Syndrome16,22
|Antidepressant discontinuation syndrome can occur with rapid discontinuation of any antidepressant.
Symptoms are variable.
Typical symptoms include – Flu-like symptoms, Insomnia, Nausea,
disturbances and Hyperarousal (anxiety/agitation) (FINISH).
|Symptoms are likely to appear within one week of rapid dose reduction or abrupt discontinuation of an antidepressant.
Symptoms are often mild and short lived and resolve without treatment in about ten days. For patients with more severe
symptoms the pre-reduction dose may need to be restarted which usually results in resolution of symptoms within 24 hours.
Subsequent tapering then needs to be at a slower rate.
|Table 1: A guide to discontinuing antidepressants
|General tapering guide
|An antidepressant should not be stopped abruptly if it has been taken for six weeks or more
The dose should be reduced gradually over at least four weeks, or longer if withdrawal symptoms emerge17
|Wide range of symptoms including anxiety, gastrointestinal disturbance, headache,
insomnia, irritability, malaise, myalgia, recurrence of depression
|SSRIs and venlafaxine
Taper slowly over several weeks or months e.g. reduce by 25% every four to six weeks for drugs with a
Fluoxetine at low doses may not need to be tapered, as it has a long half-life.15
|Mild self limiting symptoms (above) may occur within a few days. There may be a delay before symptoms
present for patients on higher doses of fluoxetine because of the longer half-life.
Discontinuation syndrome appears to occur more frequently with paroxetine and venlafaxine. This may partly
be due to the shorter half-life of these drugs.
Tricyclic and related antidepressants (e.g. mianserin) should be withdrawn slowly18
e.g. reduce by 25% every four weeks
|In addition to antidepressant discontinuation syndrome, rapid withdrawal may produce symptoms associated
with cholinergic rebound (e.g. agitation, headache, sweating, gastrointestinal symptoms), parkinsonism and problems
with balance.15 This is more likely with the more potently anticholinergic TCAs such as amitriptyline.
|Neuropsychiatric symptoms may be more prominent and include severe anxiety, agitation, altered
sleep, hallucinations, delirium and paranoid psychosis16
Regular and prolonged use of hypnotics should be avoided because of the risk of tolerance to effects, dependence and
an increased risk of adverse events.19,20
Patients who have taken benzodiazepines on a long term basis should be withdrawn gradually over a number of months (e.g.
six months). The longer a patient has been taking a benzodiazepine, the more likely they are to develop dependence and
There are a wide range of withdrawal symptoms (Table 2) and some may be similar to those for which the benzodiazepine
was originally prescribed. Some patients may experience withdrawal symptoms such as rebound insomnia and anxiety after
only two to four weeks of treatment. Withdrawal symptoms can continue for weeks or months after stopping a benzodiazepine.18 An
awareness of this may help prevent additional medicines being prescribed for these symptoms.
Abrupt withdrawal may result in confusion, toxic psychosis, seizures or a condition termed benzodiazepine withdrawal
syndrome which is similar to delirium tremens.18 Typical symptoms of this include insomnia, loss of appetite,
weight loss, sweating, perspiration, tinnitus and disturbances of perception. Benzodiazepine withdrawal syndrome can occur
within one day of stopping a short-acting benzodiazepine or up to three weeks after stopping a long-acting benzodiazepine.18
Successful discontinuation may result in improvements in cognitive and psychomotor function, particularly in older people.
Patients may be more alert and have increased working memory, reaction times and balance.12 Alternative strategies
for insomnia (e.g. sleep compression, See BPJ 14) may be required as improvements
may not occur for some months after the benzodiazepine has been stopped.21
|Table 2: A guide to discontinuing benzodiazepines
|Slowly taper the dose in steps of approximately one-eighth of the daily dose every two weeks18
If withdrawal symptoms occur, maintain at the current dose until symptoms settle and then continue to
taper, usually at a slower rate
|Wide range of symptoms including anxiety, mood changes, insomnia, palpitations, tremor,
headache, gastrointestinal disturbance, muscle stiffness and spasms
Benzodiazepine withdrawal syndrome
|Alternative withdrawal method18
- Transfer patient to an equivalent daily dose of diazepam, preferably taken at night
- Reduce the dose of diazepam every two to three weeks by 2 or 2.5 mg. If withdrawal symptoms occur, maintain this
dose until there is improvement.
- Continue to reduce the dose, if necessary by smaller amounts. It is better to reduce too slowly rather than too
- Stop diazepam completely. The withdrawal period may vary from about four weeks to more than one year.
|Approximate equivalent doses for diazepam 5 mg:
- ≡ lorazepam 0.5–1 mg
- ≡ nitrazepam 2.5–5 mg
- ≡ oxazepam 15 mg
- ≡ temazepam 10 mg
- ≡ triazolam 0.25 mg
- ≡ zopiclone 7.5 mg
Beta-blockers are the cardiovascular medicine most often associated with adverse withdrawal events. Abrupt withdrawal
may cause rebound hypertension, tachycardia, arrhythmia or angina. These events may be physiological withdrawal reactions
or an exacerbation of the underlying condition.13 Gradual dose reduction is required (Table 3).
|Table 3: A guide to discontinuing antihypertensives18
|General tapering guide
|Most antihypertensives should be tapered. Taper dose at approximately monthly intervals,
over three to six months.
||Wide range depending on the specific medicine and the condition being treated. May include ankle
oedema, weight gain, headache, tachycardia, increased blood pressure, worsening heart failure or angina, myocardial
Gradual dose reduction necessary
|Sudden withdrawal may cause or exacerbate angina
|Calcium channel blockers
Consider gradual reduction
|Sudden withdrawal may exacerbate angina
It may not be practical to cut tablets so either stop or consider alternate day dosing initially then
twice weekly dosing
|Possible exacerbation of the underlying condition
|Angiotensin-converting enzyme inhibitors
Consider gradual reduction
| Possible exacerbation of the underlying condition
The decision to stop a statin is based on an assessment of individual benefits and risks. For example, stopping may
be justified in a person at relatively low risk of a cardiovascular event, who is also poorly compliant or experiencing
troublesome adverse effects. In most cases statins can be stopped without the need for tapering.
Statins should not be stopped in patients admitted with (or with a history of ) cardiovascular events including acute
coronary syndrome, myocardial infarction and stroke.
In older people taking warfarin, low initial and maintenance dosages are recommended (e.g. dose adjusted to maintain
the INR at the lower end of the range of 2–3).23 The optimum duration of warfarin therapy is determined by
the condition being treated and its severity.23
Some clinicians tail off long-term treatment over several weeks but the need for this is unclear. It is possible to
stop abruptly rather than taper (Table 4).24
|Table 4: A guide to discontinuing warfarin
Taper over several weeks
|A rebound hypercoagulable state with a risk of thrombosis, has been reported in some patients but
this can occur even if the dose is tapered and may reflect the initial pro-thrombotic state for which treatment was
Consider stopping NSAID therapy when the risks associated with treatment outweigh the benefit. Risks associated with
NSAIDs usually relate to declining renal function in the older age group and adverse gastrointestinal effects. NSAIDs
may also reduce the effectiveness of antihypertensive therapy. Some patients may tolerate abrupt discontinuation but tapering
the dose allows for other analgesics to be introduced or increased (Table 5).
|Table 5: A guide to discontinuing NSAIDs
|Consider prn use or regular use at a lower dose
Can be stopped abruptly
Halve the dose for two to four weeks then stop
Review the need for gastric protection therapy i.e. PPI or H2RA
|Recurrence of pain, arthritis or gout symptoms
Many people remain on acid suppressants despite there being no ongoing clinical indication e.g. NSAID stopped or H.
pylori successfully treated. It is often possible to maintain symptom control on a lower dose or on an as needed
basis rather than on long term high dose maintenance therapy.
Tapering the dose of an acid suppressant (both PPIs and H2RAs) is recommended because of the risk of rebound
hypersecretion of gastric acid (Table 6).25 If rebound hyperacidity is mistaken for a return of the underlying
condition then acid suppressants may be restarted unnecessarily. Following discontinuance of omeprazole therapy, gastric
acid secretion returns to baseline over a three to five day period.23
|Table 6: A guide to discontinuing acid suppressants
|General tapering guide
|Halve the dose for four to eight weeks then stop (or step down to a less potent agent)
Consider providing an antacid for dyspepsia symptoms
|Recurrence of oesophagitis and indigestion symptoms
|Proton pump inhibitors (PPI)
Consider alternate day dosing. Capsules cannot be halved.
Consider stepping down to an H2RA if a more gradual taper is required
|Stopping PPIs suddenly can cause rebound hypersecretion of acid
|Histamine receptor antagonists (H2RA)
|Rebound dyspepsia has also been described after stopping H2RA therapy abruptly
The beneficial effects (e.g. prevention of bone loss, improved bone mineral density [BMD], and reduced risk of both
vertebral and non-vertebral fractures) are seen within one year of starting treatment. Beneficial effects on BMD persist
after stopping the drug.26 The half-life of bisphosphonates appears to be very long (probably up to several
years) because of skeletal storage.27 This may explain continued effectiveness after discontinuation.
Alendronate can be stopped abruptly without the need for tapering.
For correspondence regarding stopping alendronate therapy, see "Correspondence: Guidance for stopping alendronate", BPJ 46 (September, 2012).
Administration of oral corticosteroids for more than three weeks, or shorter courses at high doses (≥40mg prednisone),
can lead to suppression of the hypothalamic-pituitary-adrenal (HPA) axis.
Tapering may not be required for some patients who have received low to moderate doses (e.g. <40mg prednisone) of
corticosteroids for less than three weeks, however this will depend on the patient and the disease being treated.
Withdrawal should not be abrupt for the majority of patients who have been taking systemic corticosteroids for more
than three weeks. These patients should generally have their corticosteroid slowly tapered to allow the HPA axis to recover
(over weeks or months).23,26 The rate of dose reduction will vary depending on the original dose and duration
of treatment, the disease and the response of the patient.26 If the disease flares up during withdrawal, the
dose may need to be increased and the subsequent withdrawal to be more gradual. An increase in dose may also be required
during periods of stress e.g. infection, trauma or surgery. There are several methods for tapering oral corticosteroids
|Table 7: A guide to discontinuing oral corticosteroids
|For patients who have been on corticosteroid treatment for three weeks or longer reduce the dose,
e.g. at a rate of 2.5–5 mg every one to three days. Once the dose has reached 5–10 mg daily, reduce the dose more slowly,
e.g. by 1 mg each week.
Reduce more slowly initially if it is likely that the disease will relapse e.g. 2.5–5 mg every one to
Patients on longer term treatment may require withdrawal at a more gradual rate over many months (such
as a reduction of 1 mg every three to four weeks)
|Withdrawal effects include: anorexia, hypotension, nausea, weakness, fever, myalgia,
arthralgia, weight loss
Increase dose if disease relapses then taper more gradually
Increase dose during periods of stress e.g. infection, trauma or surgery
A degree of inhibition of the HPA axis may persist for six to twelve months (or longer) after prolonged
high dose treatment is withdrawn; steroid therapy may need to be re-instituted during periods of stress
The majority of patients respond initially to levodopa and its use improves the quality of life. After two years or
more, benefit is reduced as the disease progresses and late complications emerge.26 The long term use of levodopa
is limited by motor complications and drug induced dyskinesias.
If antiparkinsonian drug therapy is reduced abruptly, or discontinued, a symptom complex resembling neuroleptic malignant
syndrome can occur (Table 8).23
|Table 8: A guide to discontinuing antiparkinsonian medicines
|Antiparkinsonian medicines should not be stopped abruptly as there is a small risk
of neuroleptic malignant syndrome18
Reduce the dose gradually over about four weeks
Sinemet CR tablets are scored and may be administered as half tablets
|Hypotension, psychosis, pulmonary embolism, rigidity, tremor
A symptom complex (resembling the neuroleptic malignant syndrome) may occur. Symptoms include muscular
rigidity, elevated body temperature, mental changes, diaphoresis, tachycardia, and tachypnea. There may be an increase
in serum creatine kinase concentration.23