Chapter 11
Anxiety and Depression: Drugs

R. Hamish McAllister-Williams and Sarah Yates

Institute of Neuroscience, Newcastle University, Newcastle upon Tyne, UK

When should the use of medication be considered to treat anxiety and depression?

Depression and anxiety are commonly comorbid and this can sometimes complicate thoughts around treatment. Fortunately the first-line pharmacological treatment of choice for both depression and most anxiety disorders are antidepressants, particularly selective serotonin reuptake inhibitors (SSRIs). There can be a perception that these ‘psychological’ disorders should be managed with ‘psychological’ treatments and that physical interventions are simply artificial crutches that can end up preventing the individual dealing with their ‘difficulties’. However, there is clear evidence of biological abnormalities associated with depression and anxiety, and strong evidence for the use of antidepressants to treat these disorders. Nevertheless, it is important not to immediately jump to medication for any patient presenting with symptoms of depression and anxiety. The NICE guidelines for both depression and anxiety advocate a stepped approach to management, with antidepressants indicated at step 3, or when there is a previous history of depression. For patients with predominantly anxiety symptoms it is usually appropriate to at least try psychological and behavioural techniques (see Chapter 10) before resorting to medication. For patients presenting predominantly with depression the evidence suggests that if they have moderate to severe illnesses they will do better with medication than with psychological interventions and best of all with a combination of the two.

Do antidepressants actually work?

In recent years there have been a number of high-profile publications and news reports questioning whether antidepressants have any clinically significant benefit in the treatment of patients with depression and anxiety. These have been criticised on many levels. A central issue is that in randomised controlled trials (RCTs) of antidepressants a large placebo effect is seen. As a result the difference in effect of the active drug often seems small in comparison. Conversely the effect of psychological interventions often appears large in RCTs. However, this is because such therapies are most usually compared against waiting list controls or ‘treatment as usual’. In direct head-to-head comparisons in well-controlled studies, psychological interventions such as CBT are generally of similar efficacy to antidepressants in mild to moderate depression in working-age adults.

How do antidepressants work?

The notion that depression is due simply to low levels of monoamines such as serotonin (5-HT) or noradrenaline, and that antidepressants work by increasing the brain levels of these neurotransmitters, is not tenable. The neurobiology underlying depression and the mechanism of antidepressants is more complex.

There is strong empirical evidence that stress can precipitate episodes of depression. However, not all individuals when stressed become depressed. The risk of developing a depressive episode appears to relate to the relative balance between resilience and vulnerability within an individual. One mechanism of resilience involves 5-HT neurotransmission within the brain. This system appears able to help prevent a negative cognitive bias. This type of negative cognitive bias is what is challenged in cognitive-behavioural therapy (CBT), since such bias can drive low mood and low mood drives negative cognitions causing a potential vicious circle. The 5-HT mechanism of resilience can be made vulnerable to dysfunction by a variety of factors including genetic and environmental influences. The latter include both early life adversity and current stress.

It is known that a range of different antidepressant drugs and electroconvulsive therapy (ECT) all have effects on 5-HT neurotransmission via different receptors or neuronal functions. The net effect appears to be common between treatments, that is, to enhance transmission in the specific 5-HT pathway that is involved with resilience and prevention of negative cognitive bias. This may help to explain the synergy seen in RCTs between antidepressants and CBT. The latter is challenging the negative cognitive bias in depression while the former is helping to support a more positive bias.

How should antidepressants be used to treat depression and anxiety?

We have put together an algorithm that can be used when treating patients with depression and/or anxiety with antidepressants (Figure 11.1). It should be stressed that this is just one of many possible algorithms that could be produced and there is no hard and fast way that things must be done.

c11-fig-0001 — **Figure 11.1** Algorithm for treating patients with depression and/or anxiety with antidepressants.

The starting point in treatment, as ever in medicine, is making a diagnosis. As described above, the decision to offer medication should be made for patients with moderate to severe illness. A patient’s knowledge about antidepressants should be explored and any concerns addressed. It is important to review the patient and monitor the effects of treatment. In general, we are poor at measuring and assessing depression. One reason for this is a lack of consensus as to the best (and most practical) way of doing this. One possibility is to use the PHQ-9 (see Appendix 2): a score in excess of 10 would certainly be supportive of using medication. The decision to treat should, however, be made in a more clinically rounded way than simply on the basis of a score from a scale.

In terms of which antidepressant to use, there are a large number to choose from, but SSRIs are recommended first line. In our algorithm we suggest either citalopram or sertraline because these drugs have lower risks of drug interactions than fluoxetine or paroxetine and because citalopram is cheaper than escitalopram (Boxes 11.1 and 11.2).

Concerns have been expressed that SSRIs might increase the risk of suicide, particularly in adolescents and young adults. This could relate to the fact that lack of energy and poor motivation (which might lessen the chance of completed suicide) may respond faster than low mood following treatment. This is probably independent of class of antidepressant. If a patient has significant suicidal ideation it is important to review the patient after no more than a week.

For all patients it is worthwhile reviewing how the patient is tolerating the antidepressant at 2 weeks. The major issues to look for are whether they are experiencing problems with nausea, increased anxiety and sexual dysfunction. The first two tend to occur early in treatment and reduce rapidly over time. The latter, however, tends to be persistent. In the algorithm, there is a recommendation to increase the dose of sertraline from 50 mg/day to 100 mg/day after 2 weeks. This is on the basis that lower doses are better tolerated when first started. Although this principle has not been applied to citalopram, it can be considered in patients who have difficulty tolerating medication and in patients with high levels of anxiety. In these individuals increased anxiety in the first week or so of treatment can be problematic and therefore commencing the antidepressant at half dose and then increasing after 1–2 weeks can be helpful.

The main time point for review is then after 4–6 weeks. The evidence strongly demonstrates that if a patient has shown absolutely no response after 4 weeks, or minimal improvement after 6 weeks of treatment, they are extremely unlikely to respond to this antidepressant.

If the patient has shown a full response, the next step is to review risks of relapse (see Box 11.3). If there are no risks of relapse a patient should be treated for at least 6–12 months following full remission of symptoms.

If a patient has shown partial improvement it is particularly worthwhile looking to see whether there are ongoing sleep problems and addressing these either with psychological interventions specifically targeting insomnia or possibly using hypnotics such as zopiclone (7.5 mg) or trazodone (50–100 mg). Note that low-dose tricyclic antidepressants (TCAs) should not be used due to interactions between SSRIs and TCAs (see below). Alternatively or additionally consideration can be given for increasing the dose of the antidepressant (see Figure 11.1). It is then important to review again after a further 4–6 weeks to assess the effectiveness of this increased dose.

If a patient shows no response then the first question to ask is whether the patient is adhering to medication. If they are, the steps described above (for those patients with partial remission) can be considered. Alternatively it may be decided to switch antidepressant. This may also be considered for patients with partial adherence, for whom the antidepressant of choice is fluoxetine (Box 11.4)due to its long half-life (the active metabolite of fluoxetine has a half-life of 1 week). Alternatively if a patient is not tolerating a drug due to agitation, insomnia, nausea or sexual dysfunction, or you are looking for an alternative to an SSRI, mirtazapine can make a good choice (Box 11.5).

Occasionally patients find it extremely difficult to tolerate SSRIs because they can lead to ‘emotional blunting’. This refers to not only a reduction in the ability to feel sad emotions but also happy emotions. In such a case the use of a drug that selectively targets noradrenaline rather than 5-HT can be helpful, such as reboxetine (Box 11.6) or lofepramine (Box 11.7).

If the issue leading to switching relates simply to non-response, then a number of options are possible, including trying a second SSRI or switching to a noradrenaline reuptake inhibitor such as lofepramine or reboxetine, a serotonin and noradrenaline reuptake inhibitor (SNRI) such as venlafaxine or duloxetine (Box 11.8), or a drug with a different mechanism of action such as mirtazapine.

Regarding how to switch antidepressants, we generally recommend abrupt switching from one to the other. This is possible when switching to or from an SSRI, a SNRI (e.g., venlafaxine and duloxetine), mirtazapine or reboxetine (i.e., most of the medications listed in our algorithm and most commonly used in general practice). The only exception to this is if a patient is on high doses of antidepressants, in which case it may be advisable to reduce the dose for a week (or 4–5 weeks in the case of fluoxetine) before making the abrupt switch. Otherwise, with the exception of monoamine oxidase inhibitors, the only switch likely to be encountered that needs more care is from an SSRI to a TCA. In such a situation taper and stop the SSRI and wait 4–7 days (or 4–5 weeks if switching from fluoxetine) and then introduce the TCA. Further information can be found in the Maudsley Prescribing Guidelines.

Prescribing with comorbid illnesses

Cardiovascular comorbidity

In a patient with cardiac morbidity, prescribing can be complicated. The most important point is to avoid drugs that can lead to arrhythmias, particularly TCAs; these are also contraindicated in the immediate aftermath of a myocardial infarction (MI). There is evidence that SSRIs, in particular sertraline, can be effective in treating depression post-MI as well as reducing cardiac morbidity in such patients. This is for a number of reasons, including an inhibition of platelet aggregation.

Venlafaxine and duloxetine have a propensity to increase blood pressure. They can be used in patients with hypertension as long as this is well controlled. These drugs are not pro-arrhythmic and are safe in patients with cardiac disease.

There have been recent concerns raised regarding the risks of citalopram and escitalopram causing a dose-dependent increase in the correct QT (QTc) interval in the electrocardiograph (ECG). This potentially increases the risk of arrhythmias. As a result they are not recommended in patients with long QTc or at risk of arrhythmias. Recommended maximum doses are 40 mg and 20 mg for citalopram and escitalopram, respectively (20 mg and 10 mg in the elderly). However, it should be noted that in a large epidemiology study of several hundreds of thousands of patients, increasing doses of citalopram are actually associated with a decreased risk of ventricular arrhythmia and all-cause mortality. It is important to effectively treat depression for the cardiac health of a patient. If in any doubt do an ECG to check the actual QTc for the individual patient.

Gastrointestinal complications

Selective serotonin reuptake inhibitors inhibit platelet aggregation and increase bleeding time. This is important in individuals who have a risk of GI bleeds, for example those on high-dose non-steroidal anti-inflammatory drugs (NSAIDs), and the elderly. In these patients, consider prescribing a gastroprotective drug. Venlafaxine and duloxetine have the same risks. If there is a major concern of bleeding, mirtazapine or reboxetine would be safer options or consider gastroprotection.

Pain syndromes

Patients with depression are three times as likely to be suffering from painful syndromes as patients without depression. Conversely patients experiencing pain are more likely to be depressed than the general population. The presence of pain also makes it less likely a patient will respond to antidepressants. As a result higher doses and the use of SNRIs such as venlafaxine and duloxetine are often warranted.

Another issue that commonly occurs in patients presenting with both depression and significant pain (of any origin) is that they are started on an SSRI for depression and then have a small dose of a TCA such as amitriptyline added at night for pain. SSRIs, particularly fluoxetine and paroxetine, inhibit the metabolism of TCAs increasing their plasma levels. This can mean that patients in this situation end up with toxic levels of TCAs despite them being prescribed at low dose. In these situations it is better to either treat the patient with a dose of amitriptyline that would be therapeutic for depression (usually 100–150 mg or more) and stopping the SSRI, or switching both SSRI and TCA to venlafaxine or duloxetine, which also have efficacy in treating pain.

Comorbid substance misuse

The management of depression in patients who are also misusing substances including alcohol or other illicit drugs is complex. In general it is best to be confident that the patient does have a depressive illness in addition to the substance misuse, for example, through observation that symptoms are present even when the patient is abstinent from alcohol or illicit drugs. In such circumstances antidepressants can be helpful, but in general it is difficult to achieve a good response unless the substance misuse is also addressed.

Case study: Francis’s story

Francis appears to have an underlying social phobia. This has led to him using alcohol to improve his confidence in social situations. However, his alcohol misuse has progressed so that he appears to be dependent on alcohol and experiences withdrawal symptoms. This has resulted in his mood becoming low. Regarding medication in this circumstance it would be important for Francis initially to receive help regarding his alcohol dependence.

Recommendations suggest that antidepressants may improve mood but not necessarily decrease alcohol use in those who are depressed with harmful or dependent alcohol use. Generally mood will only improve in those with a significant depressive disorder, and use of antidepressants should be restricted to this population and then with caution and monitored.

Pregnancy

It is important always to consider the possibility of pregnancy in any woman of child-bearing age when prescribing antidepressants. There are a number of concerns about the use of medication during pregnancy, which means the threshold for prescribing should be higher than in other clinical situations. Alternative options such as CBT or even ECT should be considered. However, if a patient is suffering from significant depression or anxiety it is important for this to be effectively treated. There is evidence that antenatal depression and anxiety has detrimental effects on the unborn child (see Chapter 5).

It is important to consider the patient’s past history of response to treatment. If they have never been on an antidepressant, some drugs have more evidence to help guide decisions than others. This is not just related to the amount of time the drug has been in clinical practice. There are far more data relating to the use of SSRIs in pregnancy than there are relating to the use of TCAs. If a patient has previously responded well to a particular drug, particularly if this was only after failure to respond to others, there needs to be a good reason to switch to an alternative. There is no point in exposing a foetus to any risk if the drug doesn’t work for that mother. Specialist advice might need to be sought.

Prescribing to patients of different ages

The management of depression in adolescents is also complex. The evidence is that antidepressants may work but they may not be as effective as in adults. There is also the concern that there is a stronger association between the use of antidepressants and increased risk of suicide than in older individuals. The evidence base as to which antidepressant to use is also extremely limited. In general it is probably wise to seek specialist advice before using antidepressants in adolescents (see Chapter 2).

The same principles should be applied in the elderly as in younger working-age adults. However, there are a few extra things to consider. The first is that response to treatment can be somewhat slower in the elderly, and rather than making a decision to change an antidepressant at 6 weeks, it may be wise to wait and review after 8–10 weeks. A second is that the elderly or physically frail individuals may be more susceptible to side-effects. As a result it can be helpful to start with lower doses and increase these more gradually than one would do in more physically healthy individuals. It is important to judge the dose used against the response and tolerability of the individual patient. There is no inherent reason why an elderly patient cannot receive the same dose as a younger individual if they are physically fit. Finally, another consideration in the elderly is that these individuals may be on other medication and care needs to be taken with regards to drug interactions. For this reason citalopram and sertraline are better options than fluoxetine and paroxetine.

Next step treatments

Having followed the algorithm shown in Figure 11.1, the majority of patients will show a response but a significant minority will not. There are a number of options than can be employed in such situations. The GP could consider switching the antidepressant for a second time. There are a number of antidepressants that have been shown to have a small but statistically significantly greater efficacy than others. These include venlafaxine, escitalopram, mirtazapine, amitriptyline and clomipramine. There are also theoretical reasons to think that duloxetine might also fall into this category. Note that this list includes two TCAs: amitriptyline and clomipramine. In general it is not recommended to use TCAs in primary care for the management of depression. This is due to their risks in overdose and adverse effect profile. The safest and best tolerated TCA is lofepramine, and this is a potential alternative to reboxetine in the algorithm. Otherwise amitriptyline and clomipramine are worthwhile options to consider for more treatment-refractory patients. Clomipramine is particularly useful for refractory anxiety disorders, especially obsessive-compulsive disorder or comorbid obsessional symptoms. In specialist mental health care, beyond switching antidepressants the next major alternative is augmenting or combining drugs. This will only be outlined here.

There is somewhat conflicting evidence to support various antidepressant combinations. The most commonly employed is the combination of mirtazapine with either an SSRI or an SNRI. These combinations tend to be well tolerated since mirtazapine actually reduces many of the side-effects seen with SSRIs or SNRIs. Traditionally one of the most common augmentation strategies is the addition of lithium, but this should only be initiated by a specialist. Lithium can be added to any of the regularly used antidepressants. An increasingly favoured augmentation strategy is the addition of an atypical antipsychotic to an SSRI to an SSNRI. In particular quetiapine now has a licence for this use in patients with a suboptimal response to antidepressant monotherapy. Again, such prescribing would be recommended by a specialist rather than being initiated by a GP. The GP, however, should monitor the patient for the metabolic complications of such atypical antipsychotics. While not licensed in Europe, there is also evidence to support the use of aripiprazole in such situations. Other augmentation strategies can be employed but they tend to be more used in highly specialised affective disorders centres. These include the use of triiodothyronine, l-tryptophan, modafinil and pramipexole. Electroconvulsive therapy (ECT) can be a highly effective treatment not only for patients with severe depression with psychomotor retardation and/or psychosis but also for patients with severe depression that has not responded well to medication. Further information regarding prescribing can be found in the British Association of Psychopharmacology guidelines.