Given the fact that guidelines advise against polypharmacy, Christoph Correll (Feinstein Institute for Medical Research, Manhasset, New York, USA) had the arguably more difficult task – that of defending it in debate.

His case was based in part on arguing that trials which contributed to the meta-analyses concluding polypharmacy had no added benefit do not represent real life. Patients willing to enter trials, and fulfilling the criteria that make them eligible, are not the patients we see 90% of the time, he suggested. And the fact that 20 -25% of patients in everyday clinical practice receive more than one antipsychotic gives polypharmacy face validity.

Not so, came the counter-argument. The fact that a treatment persists is evidence that people believe in it, but the history of medicine shows that such beliefs are often unjustified.

Evidence suggests that polypharmacy brings with it an increased risk of medication interactions, side effects and poor adherence

It is true that meta-analyses have found that the only evidence of benefit for combination therapy derives from open label trials. Once these are excluded, there is no superiority over monotherapy, Dr Correll admitted. On the other hand, open-label trials reflect real practice. And trial results based on group means can dilute benefit in specific types of patient.

Patients on combinations are in greatest need

Those treated with a combination are a particularly difficult subgroup – often people with earlier onset disease, poorer functioning, more negative symptoms and treatment resistance. And polypharmacy tended to be prescribed by physicians who were senior and had known patients for more than two years, suggesting they had a good grasp of their circumstances.

In addition, Christoph Correll cited a randomized study showing that polypharmacy patients who were randomized to monotherapy had a higher need for medication change than those who continued on combination therapy.

The more drugs we give, the less the likelihood of compliance

In reply, Stefan Leucht (Technical University, Munich, Germany) had an array of arguments for thinking polypharmacy was perhaps a bad idea – in both principle and practice. For the following reasons, it should be used only in exceptional circumstances:

• Trials suggest combination therapy is associated with an increased burden of side effects, increased risk of drug-drug interactions, and a greater likelihood of non-compliance
• Drugs tend to be added at a time when patients are doing least well, which means that any apparent benefit can be due to the naturally fluctuating course of disease – so it is difficult to disentangle cause and effect
• If combination therapy is of benefit, it is not possible to decide which drug is responsible. The consequence is that patients continue to receive multiple agents because discontinue decisions are difficult to make
• Since antipsychotics are active at dopamine receptor targets in the brain, they essentially have the same mechanism of action. Once the majority of receptors are blocked, adding another drug may only result in a greater risk of adverse effects. Schizophrenia is not the same as hypertension, where entirely different classes of agent can each contribute efficacy.

Not all antipsychotics have the same profile: they are created differently

Christoph Correll responded by saying that if the last point were true, there would be no reason to switch monotherapies for schizophrenia. Indeed, treatments differ in their receptor profile and binding affinity. And Stefan Leucht conceded that, for example, there might in theory be a case for combining a full dopamine antagonist with a partial agonist. 

Need to exclude “pseudo resistance”

The topic was debated with eloquence, the wisdom of experience, and in a good-natured way. From the many contributions by the audience, it was an issue that excited interest. And the outcome was a sensible compromise.

In a series of informal votes, the meeting agreed that polypharmacy was something to be considered only after a range of other options had been tried and failed. These options included the exclusion of “pseudo-resistance” by trial of monotherapy with a long-acting injectable, ECT, CBT, and cognitive remediation. But combination therapy should not be excluded altogether.

On the other hand – Stefan Leucht cautioned -- if a patient does get better with polypharmacy, we do not know which drug caused the improvement. So the natural tendency is to continue with them all. If polypharmacy is adopted, its effects should be monitored carefully and the combination stopped if there is no benefit.