There are three concerns regarding the use of antidepressants to treat suicidal patients. First, it is important to verify the diagnosis of depression. As we have emphasized repeatedly, the presence of suicidality alone is not sufficient for the diagnosis of depressive disorder, and there is no firm evidence that antidepressant medication is helpful with suicidality per se. Do not diagnose depressive disorder without using adequate criteria (American Psychiatric Association 2013); do not assume that suicidal thoughts or actions justify this diagnosis. When the diagnosis is incorrect, the medical treatment has little chance of working, and your patient will expect a positive change that will not occur. This failed expectation runs the risk of increasing the patient’s suicidality.

Second, if antidepressants are indicated, make sure the number of pills in the bottle as prescribed makes up less than a lethal dose. For SSRIs, lethal dosing is not much of a problem—most individuals would need to take several months’ worth to get into serious trouble. For TCAs, staying below a lethal dose generally requires prescribing for 1–2 weeks at a time, keeping the total amount available less than 1,500–2,000 mg. Work with pharmacies to promote this plan. For example, you can write four 1-week prescriptions, dated to cover a month, rather than writing one prescription for the entire period. If this method proves difficult, family members or friends can be recruited to help keep your patient supplied with a reasonable amount of medication.

The problem with such a management technique is that although it promotes safety, it can also emphasize passivity and dependence. It is important to work to gradually increase a sense of competency and security for your patient in the self-managing of medication. For example, you should rehearse ways in which your patient might take the initiative to discuss with the pharmacist ways of obtaining smaller prescriptions more frequently. Of course, patients will always be able to hoard medication, increasing the risk of fatal overdose. Dealing with the total available dose as a treatment issue may make hoarding less likely.

Another technique involves packaging medication in individual wrappings the same way that over-the-counter medications are frequently dispensed. Unfortunately, this method is not readily available in the United States. Many patients who overdose do so impulsively. They may be angry or upset, and often they are consuming alcohol. There is very little planning or lead time between when the person decides to overdose and when the person ingests the medication. With a bottle of medication, most of the pills are usually consumed and the person assumes that the act of taking the pills will be lethal. The situation would be quite different with individually packaged pills. Unwrapping each pill might interfere with the impulsivity of the moment and make the situation safer.

A third potentially troublesome aspect of prescribing antidepressants, particularly those in the SSRI class of medications, is the possibility that these medications may have an iatrogenic effect on suicidality in certain populations. Currently, no medications, including antidepressant medications of different classes (Braun et al. 2016), have been proven to eliminate suicidal ideation, impulses, or behaviors. Rather, in the early 1990s, scattered anecdotal reports linked the use of an SSRI for depression to increased suicidality. The pharmaceutical industry examined this connection in a number of data reanalyses and reported no evidence of any such connection for any SSRI that was studied. In 2003, however, researchers secured these original data sets under the Freedom of Information Act and reanalyzed the data (Healy 2003). The conclusion of the reanalysis was that there is a statistically significant association between the use of an SSRI and increased suicidality, particularly among adolescents.

There was a confound in these reanalyses of SSRIs. The original studies from which the data were taken excluded patients with any form of current or recent suicidal behavior as well as patients with potentially suicide-enhancing factors, such as drug or alcohol use and physical illness. Thus, the base rates of suicidal behavior were low. To the best of our knowledge, no studies of antidepressant medication have included actively suicidal, depressed individuals and used suicidal behaviors as outcome measures. Keep abreast of developments in this area, and do not assume that what is true for one antidepressant is true for another. If more studies and reanalyses are forthcoming, you should remember that antidepressants are a diverse group of medications, both kinetically and dynamically. Almost all of these medications, some more than others, have effects on multiple neurotransmitters. For example, an increase in suicidality might be associated with an increase in agitation, which might be a product of the neurological side-effect akathisia. Akathisia is generally believed to be a dopamine-modulated problem. Accordingly, this side effect is likely to be seen in the profile of some antidepressants but not others.

Nevertheless, there is cause for concern about the relationship between antidepressants and suicide. The phenomenon of “paradoxical suicidality” has been documented as a rare occurrence in mood disorders studies, such as the Treatment for Adolescents with Depression Study (TADS) with children and adolescents (March et al. 2009). Although controversial, this phenomenon is speculated to be a consequence of some improvement in somatic aspects of depression (e.g., low energy, amotivation) without commensurate improvement in negative cognitions (e.g., hopelessness, distorted belief that death is preferable to living).

Another extrapyramidal symptom, akinesia, is difficulty initiating movement. Akinesia can be quite uncomfortable and is related to increasing suicidality. As a chronic side effect, akinesia gives many patients a blunted and unresponsive appearance. Facial muscles do not work well, arms do not swing normally during walking, and the patient looks stilted and odd. The overall result can be medication-enhanced difficulties with communication, which then results in social isolation. If not diagnosed and adequately addressed, both akathisia and akinesia can be instrumental in producing or enhancing suicidality.

Second-generation antipsychotics are considered by many to be a significant pharmacological advance. They may offer better treatment for some of the severe symptoms of schizophrenia, particularly negative symptoms and some aspects of cognitive impairment. Their different side-effect profile (compared with that of first-generation medications) may also offer advantages. Second-generation antipsychotics are less likely to produce akathisia and akinesia, lessening at least the increased risk of suicide those effects might produce. Be aware, however, that the metabolic side effects of some newer agents may make matters worse for your suicidal patient. We are particularly concerned with metabolic syndrome, which heralds the onset of type 2 diabetes (diabetes has long been associated with increased risk of depression), and with excessive weight gain, which could adversely affect self-esteem.

Among the antipsychotics, clozapine, generally viewed in the United States as a third-line intervention in schizophrenia, was shown in a large double-blind study (Meltzer et al. 2003) to have significantly better effects than olanzapine on reducing both suicidal ideation and suicide attempts in patients with schizophrenia. The authors of this study are to be commended. This study is a well-designed, prospective medication trial that included suicidal individuals and examined suicidal behaviors.

The judicious use of multiple medications certainly has a place in modern psychiatry, but dangers do arise. We note these dangers here because it is not unusual to find a suicidal person taking several different medications, especially if he or she has made several suicide attempts and if he or she has several different psychiatric diagnoses.

First, some patients are exposed to polymedication regimens because health care information is not being shared among providers and distribution points. Several physicians may have prescribed medication without knowing about the others’ involvement. One of us (J. A. C.) looked at the medications being prescribed by other clinics for approximately 600 psychiatric patients in a large county hospital and found that 22 of these patients were being given psychoactive medications (either antianxiety or antidepressant agents) by clinic physicians. In most cases, these medications were not recorded in the psychiatry chart. The information became available only when the hospital opened up its integrated pharmacy database. If you have access to one of these databases, use it to check your patients.

A related issue is the continuation of medications after the prescribing physician quits the case. Some physicians have an unfortunate tendency to add new medications but not subtract old ones, and some pharmacies continue refills indefinitely. One of us (J. A. C.) treated a distressing case of tardive dyskinesia that emerged after a 48-year-old woman had been treated with thioridazine for 15 years. This medication, at 50 mg per night, had originally been prescribed for insomnia. The treating physician had died, and the local drugstore continued to refill the prescription for years. The woman saw other physicians during this time for the treatment of depression. None of them were aware of the ongoing antipsychotic treatment. When you become involved in the care of a patient, make sure that polypharmacy has been arrived at by a rational process.

A second danger with polymedication regimens is that they have a multiplicative impact on side effects. An informal rule is that side-effect potential squares with each additional medication. Two pills have 4 times the side-effect potential of one pill, three pills have 9 times the potential, four pills have 16 times the potential, and five pills (this is the point where you really need to start thinking things over) have 25 times the side-effect potential of one pill. In addition to side effects increasing rapidly in number, individual side effects can also be made worse. For example, a patient taking several medications with anticholinergic properties can experience significant constipation, a condition sometimes not reported and often not asked about.

Third, medications can interact with the pharmacokinetic properties of one another in a variety of ways, creating swings in blood levels that can lead to both adverse events and ineffective medication levels. You must understand the pharmacokinetics and pharmacodynamics of the medications you prescribe so as to avoid or treat these interactive problems. A systematic review of nonpsychotropic agents revealed that the effects of medications for physical disorders such as heart disease on suicidality are unknown (Gorton et al. 2016), although medication interactions are common and well documented (Wynn et al. 2008). All your knowledge ceases to be of much use when a patient is taking three or more kinds of pills at once. At this point, pharmacotherapy can become so complex that no one knows what is going on or what might happen.

The following are a few rules that will help to keep you out of trouble with polypharmacy regimens:

1. Have a good reason for adding a medication and document it, including measures for evaluating the effect and target symptoms to be treated.
   
2. Use medication in a sufficient dose and for a sufficient length of time to determine whether it is working. Do not add another medication until you are sure the patient has taken the first one as prescribed, at a therapeutic level, and for long enough to verify whether it does or does not produce a benefit.
   
3. Use response criteria to gauge effectiveness. Stop the medication if it is not working. If possible, discontinue the medication in a tapering manner to avoid adverse rebound or withdrawal phenomena.
   
4. Keep your patient as active as possible in sizing up the effect of the new medication (e.g., include your patient’s assessment in the clinical trial process).
   
5. Change only one medication at a time when possible. It is difficult enough to gauge the effect of the addition or subtraction of one pharmacological agent, let alone two or three.
   
6. Be cautious about adding medicines simply because a suicidal crisis is present. The recurrence of suicidality does not necessarily mean that treatment is failing. Suicidality should be viewed as a clinical “sign” of distress and active, if compromised, problem solving by the patient. Suicidality should be addressed constructively, as reflected in the philosophy of therapy illustrated throughout this book.