Phencyclidine (PCP)

David A. Gorelick and Robert L. Balster

Phencyclidine (commonly known as PCP or "angel dust") is a synthetic arylcyclohexylamine with a complex and unusual pharmacology that has captured the interest of researchers, clinicians, recreational drug users, and illegal drug dealers alike. Although human therapeutic use in the United States has ended and illicit use has waned over the past decade, scientific interest remains high, focused on the use of PCP as a probe for certain neurotransmitter systems (see Excitatory Amino Acid Neurotransmission) and as a model for certain psychiatric disorders (see Functional Brain-Imaging Studies in Schizophrenia. This chapter selectively reviews the basic neuropharmacology of PCP and its animal and human psychopharmacology.

PCP was developed in the 1950s as an injectable anesthetic. Animal testing revealed its unusual "cataleptoid anesthetic" properties; that is, in addition to producing anesthesia and analgesia, it made animals tranquil and serene—hence its trade names Sernyl and Sernylan (2). Clinical testing as an anesthetic revealed disturbing behavioral side effects during postoperative emergence, including dysphoria, confusion, delirium, and psychosis (8, 51, 66). These side effects attracted the interest of psychiatric researchers, who viewed PCP as producing a model psychosis useful in understanding schizophrenia (29). Subanesthetic doses administered to normal volunteers produced many schizophrenia-like symptoms, whereas schizophrenic patients suffered exacerbation of their symptoms. This research stopped when clinical trials were terminated in 1965, although PCP was later marketed for a time as a veterinary anesthetic (86).

Illicit recreational use of PCP first appeared during the mid-1960s on the West Coast in the form of oral capsules (2, 8, 51, 66). Its propensity for causing unpredictable dysphoric reactions gave PCP a bad street reputation, and use waned. It was classified as a Schedule III controlled substance in 1970. By the mid-1970s, illicit use had increased again, both because it could be easily and cheaply synthesized from legally available precursor chemicals and because users began smoking the drug, allowing individual titration of dosage to better control effects. In response, PCP and several analogues were reclassified into Schedule II, and controls were placed on the distribution of precursor chemicals. To prevent diversion of legal supplies, PCP was withdrawn from veterinary use in the United States in 1978. The close analogue ketamine, which has similar pharmacological properties, remains a legally marketed dissociative anesthetic not regulated under the Controlled Substances Act.

Illicit PCP use peaked in the United States in 1979, when prevalence of lifetime use was around 14% among high school seniors and young adults 18–25 years old, and recent (past 30 days) use prevalence was 2.4% for high school seniors (31). By 1992, lifetime and recent use prevalence had declined to 2.4% and 0.6%, respectively, among high school seniors and to 2.0% and 0.2% among young adults 19–28 years old. However, daily use has remained relatively stable among high school seniors over the past decade (0.1–0.3% prevalence), and PCP remains an important drug of abuse in some metropolitan areas and among certain sociodemographic groups, either taken alone or together with other illicit drugs such as marijuana ("primos," "wac," "zoom") or cocaine ("space base," "space cadet," "tragic magic"). Most PCP users also use or abuse other illicit drugs and alcohol, complicating efforts to identify effects of PCP from studies of clinical populations (13, 22, 23). Pharmacokinetic and pharmacodynamic interactions between smoked PCP and other concurrently smoked drugs (e.g., marijuana) may also complicate interpretation of PCP effects in drug users (41).

PCP and ketamine are arylcycloalkylamines, as are at least 20 identified PCP analogues and metabolites, some of which are themselves psychoactive and sold as "designer drugs" (1, 2). The structure–activity relationships among arylcycloalkylamines are well understood (38). As a lipophilic weak base, PCP readily crosses cell membranes and the blood–brain barrier, and its renal clearance is strongly dependent on urine pH (1, 7).

Compounds that structurally differ from the arylcycloalkylamines, but share significant aspects of PCP's neuropharmacological and behavioral effects, have proven to be useful tools for exploring the neuropsychopharmacology of PCP. For example, the 1,3-substituted dioxolanes, etoxadrol and dexoxadrol, which produce a profile of effects nearly identical to that of PCP (2, 3), have been very useful for studying stereoselectivity and for modeling the PCP pharmacophore (71). Dizocilpine (MK-801), a potent ligand for PCP sites of action in the brain, is used extensively as a selective probe for the PCP receptor (2, 30).

The similarities in the cellular and pharmacological actions of PCP and sigma-agonist benzomorphans led to great confusion in the literature when it was discovered, using radioligand binding, that N-allylnormetazocine (NANM) bound with high affinity to a site in nervous tissue that was clearly not the same as the historical PCP/sigma receptor thought responsible for their shared pharmacology (30). Nonetheless, this site quickly came to be called a sigma receptor (56). There are now hundreds of drugs identified with high affinity for this site (77), including (+)-3-PPP, di-tolyl guanidine, and certain potential novel antipsychotics. There is also evidence for heterogeneity of sigma sites (77). It is increasingly clear that these high-affinity sigma sites are not responsible for the behavioral/psychological effects of PCP, and may not mediate the psychotomimetic effects of any drug class (30, 47).

PCP administration in animals has effects on a number of neurotransmitter systems, including dopaminergic agonist effects, complex actions on both nicotinic and muscarinic cholinergic systems, N-methyl-D-aspartate (NMDA) antagonist effects, and poorly understood interactions with noradrenergic and serotonergic neurotransmission (30). Determining which of these neuropharmacological effects are responsible for which of the pharmacological, behavioral, and psychological effects of PCP has been the subject of extensive investigation.

PCP acts as an indirect-acting dopaminergic agonist in many model systems (15, 30). It remains unclear whether this is primarily an effect of PCP acting directly on sites within dopaminergic synapses or an indirect consequence of PCP activating presynaptic dopaminergic neurons, presumably through its primary actions at the NMDA receptor complex.

The best-characterized cellular action of PCP results from its interaction with a specific and unique binding site in neural tissues which satisfies many of the biochemical and pharmacological criteria for a physiologically relevant receptor (30, 87). Because of the affinity of benzomorphan sigma-agonist opioids for this site, this "PCP receptor" was known for a time as the PCP/sigma receptor. Actions at this site are an important neural basis for the behavioral effects of PCP-like drugs (2, 30). This conclusion is primarily based on the excellent concordance between binding site affinity and the potency and efficacy of arylcyclohexylamines and PCP-like drugs from other chemical classes for the production of PCP-like behavioral effects in animal studies.

The structure and physiological function of this PCP binding site have been clarified over the past decade. PCP has been identified as a noncompetitive NMDA antagonist probably acting as a channel blocker in the NMDA receptor complex. The previously identified PCP receptor is this site in the channel of the glutamate ionophore; there is an excellent correlation between the potencies of arylcyclohexylamines and other PCP-like drugs as NMDA antagonists and their activities at the PCP receptor site (35). In addition, there is a substantial overlap in the anatomical distribution of PCP and NMDA binding sites. The best evidence of a PCP-associated site on the NMDA receptor complex comes from expression cloning studies of NMDA receptor subunits that contain PCP-sensitive channels (46). By implication, this also suggests that NMDA antagonism probably plays an important role in producing the behavioral and psychological effects of PCP (82).

The explosion of knowledge in recent years regarding the neuropharmacology of glutamatergic receptors (see Excitatory Amino Acid Neurotransmission and Functional Brain-Imaging Studies in Schizophrenia ) and the structure, distribution, and function of the NMDA-receptor complex (35, 45) has helped clarify many of the pharmacological actions of PCP-like drugs. The key features of the NMDA receptor complex that explain PCP pharmacology include the following: Glutamate activates the NMDA site, leading to the opening of cationic channels resulting in Ca2+ and Na+ influx. Currents resulting from NMDA site activation are voltage-sensitive due to a block by Mg2+, which is overcome when the neuron is sufficiently depolarized. Thus, the use dependency of NMDA receptors provides a means by which they may participate in neuroplasticity and in the destabilization of membranes during convulsions. Excessive Ca2+ flux resulting from NMDA receptor activation during neural injury also appears to play an important role in excitotoxicity.

Another important feature of the NMDA receptor is the large number of allosteric modulatory sites on the receptor complex. In addition to the PCP site in the channel that blocks ion flux, there is a well-characterized glycine co-agonist site. The presence of glycine facilitates channel opening by agonists at the NMDA site, and competitive antagonists at the glycine site can function as noncompetitive NMDA antagonists. There also appears to be a polyamine regulatory site. The development of drugs with selective agonist or antagonist actions at each of these sites has provided a means of studying the pharmacology of the NMDA receptor.

During the 1980s, endogenous PCP-like activity was isolated from a variety of animal and human tissues, including cerebrospinal fluid (CSF), brain, and gastrointestinal (GI) tract (11). This activity was characterized as PCP-like because of its specific binding to PCP receptors and its ability to block NMDA-induced dopamine release in vitro. Initial attempts to isolate, purify, and identify the endogenous PCP-like ligand resulted in a small polypeptide. These results have never been convincingly replicated, so that the existence of an endogenous PCP-like compound remains uncertain.

PCP and other arylcyclohexylamines (such as ketamine) produce a unique profile of psychopharmacological effects (2) that may be conceptualized as having dopaminergic stimulant, depressant-like, and other unique components. (See also Adaptive Processes Regulating Tolerance to Behavioral Effects of Drugs, Animal Models of Drug Addiction, and Animal Models of Psychiatric Disorders, for related discussions.)

PCP produces many amphetamine-like stimulant effects, particularly in rodent species, including sympathetic nervous system stimulation, increased locomotor activity, stereotyped behavior, ipsilateral rotation in substantia nigra-lesioned rats, suppression of plasma prolactin levels, increases in low rates of scheduled-controlled operant behavior, and enhancement of effects of amphetamine-like drugs (2). The locomotor effects, at least, are mediated by dopamine release in the nucleus accumbens (6, 68), but this may be less a result of any direct PCP action at the dopamine synapse than of PCP's NMDA antagonist effects, which block NMDA receptors on dopaminergic VTA neurons (78). PCP itself has little affinity for postsynaptic dopamine receptors, but does have some affinity for the dopamine transporter through a low-affinity, PCP2 binding site (30, 60). The discovery of a potent dopamine uptake inhibitor, BTCP, which is a structural analogue of PCP, lends support to the idea that PCP may exert some of its dopaminergic actions this way (44). On the other hand, PCP typically has a much higher in vitro affinity for the NMDA channel site (Ki = 50–100 nM) than it does for the dopamine transporter (Ki 0.1 mM) (30). It is even less potent as a dopamine-releasing agent. Furthermore, some drugs, such as dizocilpine, which produce PCP-like dopaminergic behavioral effects, have little, if any, in vitro affinity for the dopamine transporter (58), yet have high affinity for the PCP site on the NMDA receptor (30). This suggests that the dopaminergic effects of PCP-like drugs seen in vivo, and in many functional in vitro assays, may be secondary to NMDA antagonism (30). There are many studies showing functional interrelationships between glutamatergic and dopaminergic systems in brain (see Functional Brain-Imaging Studies in Schizophrenia). The cellular basis for the dopaminergic actions of PCP may be relevant to the possible relationship between PCP and schizophrenia, a question of considerable current research interest (6, 15, 18, 78).

PCP produces many effects similar to those produced by classical depressant drugs, such as the barbiturates, benzodiazepines, and ethanol (2). These include motor incoordination and muscle relaxation, anticonvulsant effects, and enhancement of the toxic, anesthetic, and behavioral effects of depressant drugs. Although PCP's anesthetic effects might be included here, it is clear that the dissociative anesthesia produced by PCP-like drugs differs qualitatively from anesthesia produced by depressant drugs and volatile anesthetics.

Another depressant-like effect of PCP that has attracted recent interest is its activity in various animal models predictive of anxiolytic activity. PCP and related drugs such as ketamine and dizocilpine generally increase punished behaviors in rats and pigeons, although the magnitude and range of effective doses is typically less than is seen with benzodiazepines under comparable conditions (81), and antipunishment effects are not seen in squirrel monkeys (39). Dizocilpine is also active in a variety of nonpunishment procedures, as are other types of NMDA antagonists (81), suggesting (a) the possibility that NMDA antagonists may be developed as anxiolytics and (b) a possible role for the NMDA receptor in anxiety disorders.

PCP and ketamine produce antinociceptive effects in a variety of animal tests through a nonopioid mechanism (17). The potency relationships among PCP-like drugs suggest that NMDA antagonism is the basis for their antinociceptive effects. This is consistent with the increasing evidence for a role for NMDA receptors in nociception. The poor separation of antinociceptive from other behavioral effects of PCP-like drugs suggests why they may not be clinically useful analgesics.

PCP and other PCP-like NMDA antagonists produce deficits in sensorimotor gating in several animal models, deficits similar to those produced by dopaminergic agonists that are considered models for the analogous deficits found in schizophrenic patients (2). Evidence from brain lesion, drug interaction, and structure–activity studies suggests that PCP's effects in the pre-pulse inhibition of the startle model are mediated through its NMDA-antagonist actions (33).

In animal studies, PCP produces a unique profile of discriminative stimulus effects different from its stimulant and depressant effects (2, 3). In animals trained to discriminate PCP or another PCP-like drug such as ketamine, drugs from other classes, including stimulants, depressants, and hallucinogens, fail to fully substitute (3). Despite numerous attempts, no drug has been found that consistently antagonizes PCP discrimination. This failure to obtain substitution or antagonism with selective agonists and antagonists for a large number of known receptor systems is an important basis for ruling out these systems as mediating PCP discrimination.

PCP- or ketamine-like discriminative stimulus effects are consistently produced by compounds with PCP-site-selective NMDA antagonist effects (2, 3, 82, 83). The fact that all potent and selective NMDA channel blockers produce PCP or ketamine-like discriminative stimulus effects with a potency consistent with their affinity for the PCP site in the channel is strong support of the hypothesis that this blockade is an important neural mechanism for PCP discrimination (3). PCP-site NMDA channel blockers that have been tested in humans (NANM, dexoxadrol, dextrorphan, dizocilpine) tend to produce dysphoric psychological effects (3, 47, 73), and a number of arylcyclohexylamine analogues of PCP (TCP and PCE) have appeared as "designer drugs" of abuse. Taken together, these data support the hypothesis that the ability of drugs to produce PCP-like discriminative stimulus effects in animals is predictive of their potential to produce PCP-like subjective effects and abuse liability in humans and that this effect is the result of NMDA channel blockade.

PCP has reinforcing effects in all animal species in which it has been studied (2), a characteristic that distinguishes its behavioral pharmacology from that of classical hallucinogens. It appears that all NMDA channel blockers studied, including other arylcyclohexylamines, (+)-NANM, dizocilpine, dexoxadrol, and etoxadrol, have reinforcing effects (2, 82, 83). As with a number of other classes of drug reinforcers, the self-administration of PCP-like drugs by animals can be modified by changing food deprivation conditions, response requirements per injection, and the availability of alternative reinforcers (2). The sensitivity of PCP self-administration to these and other variables suggests that animal studies can be used to explore the complex interactions between pharmacology and context that underlie individual differences in vulnerability to PCP abuse (9).

Tolerance to the behavioral effects of PCP in animals has been shown in a number of studies (2). Generally, only two- to fourfold shifts in dose–effect curves are seen when moderate, behaviorally active doses are given repeatedly. This magnitude of tolerance is less than is generally seen with other classes of drugs of abuse, such as the opioids and sympathomimetic stimulants. Most tolerance is pharmacodynamic, rather than biodispositional, and learning plays an important role in the development and persistence of tolerance (2, 80)) and its situational specificity (67).

PCP can produce physical dependence in animal studies (2)—for example, in rhesus monkeys who self-administered very large daily doses. The excitatory discontinuation syndrome included tremors, oculomotor hyperactivity, bruxism, fearfulness, vocalizations, diarrhea, and, in some animals, emesis and convulsions. The time course of the syndrome corresponded to clearance of drug from the body. At lower, less behaviorally toxic doses, an easily observable withdrawal syndrome is not typically produced, but subtler changes in learned behavior following cessation of repeated PCP administration can be reliably demonstrated (2, 10, 80). There is some evidence for cross-dependence between PCP and other PCP-like NMDA antagonists, such as ketamine and (+)-NANM. The neural basis of dependence on PCP is not known. Studies of possible changes in PCP/NMDA receptor regulation with repeated dosing have yielded conflicting results (42, 79).

Physical dependence on PCP may occur in some human chronic daily users, but is not common in psychologically dependent users presenting for drug abuse treatment, and has never been systematically studied in humans (2, 22, 23). Burn patients do show tolerance to the analgesic effects of ketamine, and up to fourfold tolerance to PCP's behavioral/psychological effects has been reported anecdotally in some abusers (1, 8). No obvious PCP withdrawal syndrome has been reported in human users, although one group has described a possible discontinuation syndrome occurring within 1 day of drug cessation among outpatients who were daily users for at least 3 months (8). The syndrome included depression, anxiety, irritability, anergia, hypersomnia, increased appetite, poor memory, confused thoughts, and increased craving for PCP. Given the nonspecific nature of these symptoms, it remains unclear to what extent they represent a true drug-withdrawal syndrome.

PCP, or other PCP-like NMDA antagonists such as dizocilpine, given repeatedly in combination with other drugs of abuse, can block or reduce development of tolerance and dependence [e.g., to opioids (74)] and sensitization [e.g., to effects of amphetamines, cocaine, and nicotine (32, 63)]. Competitive as well as noncompetitive NMDA antagonists are able to block tolerance development (72). Although the neural basis by which NMDA antagonists modify tolerance and dependence development is not known, it is possible that their interference with NMDA-receptor-mediated neuroadaptive processes is involved.

The functionality of NMDA receptor activation is activity-dependent, as evidenced by voltage-dependent Mg2+ blockade of the associated ion channel. This suggests their possible important role in integration of cell firing and neuronal plasticity. Consistent with this role is evidence that NMDA receptors are involved in long-term potentiation (LTP) in the hippocampus, a form of activity-dependent synaptic plasticity (40). Inhibition of LTP by PCP and other NMDA antagonists could account for their consistent disruption of learning and memory in a wide variety of animal models (82, 83).

PCP is well-absorbed and readily penetrates the central nervous system after intravenous, inhalational (smoked), intranasal, oral, and percutaneous administration (7, 53), and it can be passively absorbed from the environment (61). The rapidity of onset of action varies with the route of administration: Within 1 min with intravenous and inhalation, several minutes with intranasal, and 20–40 min after oral (7). The time to peak effect also varies: 10 min with intravenous, 5–30 min with inhalation, and 90 min with oral administration (7).

PCP intoxication may last 4–8 hr after recreational doses, with some users reporting subjective effects for 24–48 hr (51, 66). This prolonged duration of action contrasts with PCP's rapid clearance from blood (distribution [alpha] half-life: 1–4 hr) by uptake into brain and other fatty tissues, and may be related to tissue stores and a gastroenteric circulation facilitated by pH-dependent GI absorption (7). The plasma [beta] elimination half-life is 7–50 hr (mean: 17.6 hr) in normal volunteers and somewhat longer (11–89 hr) in overdose patients (7). Chronic users often show an asymptotic [gamma] elimination phase lasting several weeks, which probably represents prolonged, slow release of PCP from saturated fatty tissue stores. (PCP and metabolites have been shown to persist for weeks in rat brain and other fatty tissue.) This may account for positive urine toxicology results occurring several weeks after the last PCP use (22, 64).

PCP is metabolized mainly by cytochrome P-450dependent mixed function oxidases in liver (and placenta), resulting in about 75% of an administered dose becoming polar (chiefly ring hydroxylated) metabolites that are eliminated in the urine (7, 41). None of these polar metabolites are known to contribute to the behavioral effects of PCP administration, but more than a third of minor metabolites are still not chemically identified.

PCP's pharmacokinetic parameters and metabolite patterns appear to be similar over a wide range of doses regardless of route of administration or chronicity of use (7). Inhalation introduces a new set of compounds by pyrolysis of PCP to 1-phenyl-1-cyclohexene (PC) and piperidine, which then undergo their own oxidative metabolism (41). In mice, PC and its metabolites are 100 times less behaviorally active than PCP, but their human pharmacology is unknown. PCP metabolism shows substantial individual variation. One factor may be cigarette smoking, which is associated with increased in vitro PCP hydroxylation by human liver (54). Pregnant and postpartum women and neonates tend to have different metabolite patterns from other subjects studied, suggesting that there may be hormonal influences on PCP metabolism (7). Genetics may also be a factor, because rate of PCP metabolism in mouse liver is controlled by a single gene on the X chromosome or chromosome 17 (26).

As described above for animals, PCP produces a variety of behavioral and psychological effects in human recreational users, effects which may be influenced by dose, rate, route of administration, and prior experience with the drug (51, 66) (see section entitled "Tolerance and Dependence," above). These include hypesthesia and analgesia, sedative or depressant-like effects (feelings of calmness, depression, psychic numbing, anergia, impaired concentration, ataxia, and analgesia), stimulant-like effects (feelings of euphoria, power, invulnerability, anxiety, insomnia, and anorexia), and hallucinogenic or psychotomimetic effects (distortions of time perception and body image, synesthesias, illusions and hallucinations, depersonalization, derealization, thought disorganization, paranoid ideation, and bizarre behavior).

There is little direct human data indicating which neuropharmacological mechanisms mediate these psychopharmacological effects of PCP. In experimental studies with normal volunteers, these effects occur at doses up to 5 mg orally and 1–2 mg i.v. or by inhalation, producing serum concentrations up to 100 ng/ml (0.4 mM) (29). At these concentrations, PCP has significant interactions only with its NMDA receptor and dopamine transporter binding sites. Such pharmacokinetic correlations, as well as limited human structure–activity correlation studies with PCP analogues, suggest that analgesic and psychotomimetic effects are mediated by the PCP binding site in the NMDA receptor complex. PCP NMDA receptor sites are widely distributed in human brain, with highest densities in cerebellum, hippocampus, and temporal cortex (79). The contribution of PCP's dopaminergic actions to its human stimulant-like and reinforcing effects remains unknown.

Consistent dose–effect or concentration–effect correlations have never been established in clinical studies (29, 51, 66). Serum concentrations in several case series of intoxicated patients have ranged from 0 to 800 ng/ml, whereas subjects arrested for public intoxication or driving under the influence have had concentrations from 7 to 240 ng/ml. The reasons for this wide variation in tolerated plasma concentrations are not well understood, but probably include (a) poor correspondence between PCP concentrations in body fluids and binding at the sites of action in the brain and (b) individual variation in sensitivity to phencyclidine. No relevant data on human brain or CSF PCP concentrations are available in the literature.

Acute PCP intoxication can produce three stages of behavioral/psychological toxicity: (i) behavioral toxicity with only mild neurologic and physiologic abnormalities, (ii) stupor or light coma with responsiveness to pain, and (iii) deep coma with unresponsiveness to pain (1, 51, 66, 86). Manifestations of behavioral toxicity can be loosely grouped into several clinical patterns that resemble psychiatric syndromes, and thus should be considered in the differential diagnosis of many newly presenting psychiatric patients. The commonest pattern seen in health care settings is that of delirium (i.e., disorientation, confusion, impaired recent memory, labile and inappropriate affect, and impaired judgment), without other evidence of psychosis. Delirium lasting up to several days is quite common in patients recovering from PCP-induced coma, and may occur transiently as the final phase of any episode of PCP intoxication. Other common patterns include (a) psychosis, with hallucinations and (usually paranoid) delusions occurring in an alert and oriented patient, and (b) catatonia, with negativism, mutism, blank staring, and catalepsy. The rarest patterns in health care settings (but much commoner in users not seeking or needing acute treatment) are (a) euphoria occurring in oriented patients without psychosis and (b) lethargy or sedation, also occurring without psychosis.

PCP intoxication may be accompanied by agitated or bizarre behavior regardless of the primary symptom pattern (1, 51, 66). PCP has gained a reputation as causing violent behavior (5, 13). Coupled with the drug's alleged analgesic and strength-enhancing effects, this has given PCP users a reputation for being especially dangerous to interact with. The bases for this reputation have never been confirmed by direct scientific evidence (5). All published case series lack one or more methodologic factors needed to support firm conclusions, including objective confirmation of PCP use around the time of the violent behavior, exclusion of concurrent use of other drugs, distinction between intended violence and agitation or psychosis, and knowledge of any prior history of violent behavior. Large-sample, population-based studies do not suggest a special propensity for violence or criminal behavior among PCP users. For example, lifetime and prior-month PCP users are overrepresented four- to eightfold compared to the general population among state prison inmates, but this is no greater than the four-to twelvefold overrepresentation of users of other illicit drugs such as cocaine and heroin (37). Also among state prison inmates, 5.6% of lifetime PCP users reported being under the influence of the drug while committing their crime, compared with 29% of lifetime cocaine users and 23% of lifetime opiate users (37). Data such as these are consistent with the association between illicit drug use and antisocial, including violent, behavior, but do not suggest any special risk for such behaviors in PCP users. Animal studies using experimental models of aggression also do not support any special propensity for PCP to produce aggression (2).

PCP users may be more at risk for trauma (including accidents) in general. Substantially more PCP users who present to hospital emergency departments do so because of trauma (7.7% in 1991) than do users of opiates (3.4%) or cocaine (5.3%) (48), and more PCP users whose deaths are investigated by medical examiners die because of a contributing external physical event (36.4% in 1991) than do users of opiates (8.1%) or cocaine (25.2%) (49). These findings may reflect the ability of PCP to impair both psychomotor function and judgment.

PCP intoxication produces a variety of neurological and cardiovascular effects mostly related to its sympathomimetic actions and disruption of cerebellar function (1, 51, 66). Common neurological effects include increased muscle tone, tremor, brisk deep tendon reflexes, nystagmus (especially vertical), and ataxia. Pupil size is variable: It is often normal or enlarged, and it is usually reduced only during actual coma. Common cardiovascular effects, probably due to both sympathomimetic action and decreased baroreceptor activity, include moderate elevations in heart rate (typically 20–30 beats/min) and blood pressure (typically 10–20 mm Hg, with systolic greater than diastolic), resulting in increased cardiac output. Noncardiovascular sympathomimetic effects include diaphoresis, lacrimation, and increased bronchial and salivary secretions.

At doses intentionally taken by PCP users, the drug does not significantly influence respiration, metabolic rate, or GI motility (1, 51, 66). At higher doses, PCP causes a variety of serious medical complications, including coma, seizures, hyperthermia, intracranial hemorrhage, apnea, and acute rhabdomyolysis (often resulting in myoglobinuria and acute renal failure). Direct depression of myocardial contractility and decreased peripheral vascular resistance may cause hypotension and circulatory collapse. Other medical complications, such as diarrhea, abdominal cramps, and hematemesis, may be due to by-products of synthesis or pyrolysis and other contaminants in street samples, rather than PCP itself (62).

In the absence of clinically usable antagonists, treatment of acute PCP intoxication—that is, reduction of environmental stimulation and control of remaining behavioral toxicity with sedatives (for anxiety, agitation) or high-potency neuroleptics (for psychosis), as appropriate (1, 51, 66)—remains symptomatic. A variety of other medications, such as calcium channel blockers, have seen scattered use (55). There are no comparative controlled clinical trials to favor the use of any particular medication over others. In severe intoxication with serious physiologic abnormalities, enhancement of PCP renal clearance by aggressive acidification of the urine can be useful (16, 36), although care must be taken to avoid producing metabolic acidosis (with consequent risk of renal compromise) and to exclude the presence of other drugs whose renal clearance might be retarded by urine acidification.

PCP use clearly can produce psychological dependence, as manifested in users who enter drug abuse treatment reporting that use is rewarding for them and who have great difficulty in stopping use despite knowledge of adverse consequences (2, 22, 23). The specific subjective effect of PCP that reinforces drug-taking may vary with the individual. Three patterns of acute intoxication responses have been described as desirable by PCP abusers in treatment: euphoria/stimulation, depression/sedation, and hallucinogenic effects (including religious experiences) (13, 22, 23). Two psychological effects commonly reported by PCP abusers as motivating their continued drug use are: (i) feelings of power, strength, and invulnerability; and (ii) psychic numbing used as self-medication for dysphoria (22, 23). In the absence of relevant epidemiologic data, it is unknown what proportion of PCP users become dependent, or what frequency or duration of PCP use is associated with dependence. Among high school seniors in 1992, 25% of lifetime PCP users used the drug within a 30-day period, compared with 22%, 21%, and 25% of lifetime LSD, cocaine, and heroin users, respectively, using their drug within a 30-day period (31). By contrast, 17% of current (within past 30 days) PCP users were daily users, compared with 5%, 8%, and 17%, respectively, of current LSD, cocaine, and heroin users. Almost all PCP abusers in drug abuse treatment have been using the drug for several years, although only about one-third have been using it daily (22, 23).

There is little available data on the treatment of PCP abuse/dependence and few controlled clinical trials (13). Most published studies describe psychosocial treatment methods such as outpatient group therapy and long-term residential treatment, chiefly in adolescents and young adults. Clinical experience and the available outcome data suggest that currently used treatment methods tend to have low long-term success rates (13, 22, 23). Even less data are available on pharmacological treatment of PCP abuse/dependence. Desipramine (150 mg daily) and buspirone (10 mg t.i.d.) have been used in conjunction with counseling in small, double-blind, placebocontrolled outpatient trials, with significant improvement in psychological symptoms (especially depression) but no effect on PCP use (all groups had high retention and abstinence rates) (19, 20).

A variety of persisting behavioral/psychological changes have been associated with PCP use, including psychosis, depression, anxiety, personality change, and neuropsychological impairment (1, 2, 12, 13, 24). Persistent depressed mood, sometimes severe enough to require psychiatric treatment, can follow any episode of PCP intoxication. Persisting psychosis may follow acute PCP-induced psychosis, especially in patients with preexisting schizophrenia (24). One recent test–retest study found some neuropsychological impairment persisting over several weeks of drug abstinence (12). Most studies have used retrospective data collection, lacked appropriate comparison groups, or had other methodological flaws making it impossible to distinguish actual PCP effects from preexisting conditions or from the consequences of other factors frequently associated with PCP use, such as other substance use, head injury, and a drug-abusing lifestyle. When an appropriate comparison group is used, there is often no significant difference between the PCP users and non-PCP-using drug users.

PCP causes a variety of degenerative changes in rat and human neurons and astrocytes in vitro, including vacuolization, inhibition of microtubule function, suppression of axon outgrowth, and cell death (43, 50). These effects are time- and concentration-dependent, tending to occur only after several days of exposure to PCP concentrations (100–500 mM) substantially higher than the plasma concentrations found in human PCP users, but probably closer to concentrations achieved in the fetuses of PCP-abusing mothers. These neuropathic effects may be mediated by blockade of inactivating potassium channels rather than by other actions of PCP because they occur only at concentrations far above those required for action at PCP's NMDA and dopaminergic receptor sites and are produced by other potassium channel blockers but not by other NMDA or sigma receptor ligands.

There is currently no information directly linking the PCP-induced neuropathologic effects observed in animal and in vitro studies with the behavioral/psychological effects observed in human PCP users. Children born to PCP-using mothers exhibit hypertonicity, coarse tremor, irritability, poor visual tracking, and impaired attention as neonates and may continue to have abnormal EEG patterns, nystagmus, and impaired interactive behavior and organizational responses to environmental stimuli through several years of age (21, 57, 70, 75, 76). Prospective longitudinal studies are needed to determine the persistence and consequences of these abnormalities, along with appropriate comparison groups to distinguish specific effects of PCP exposure in utero from general effects of maternal drug use and a drug-using lifestyle.

Brain imaging studies in small numbers of adult chronic PCP users suggest that they may have decreased right cerebral cortical blood flow and frontal glucose metabolism, abnormalities similar to those found in schizophrenic patients (25, 84). The affinity and density of brain PCP and sigma receptor sites has been reported to be normal in adult chronic PCP users who died traumatically (79).

Maternal PCP use during pregnancy is associated with intrauterine growth retardation, precipitate labor, and fetal distress (reflected in meconium staining), but not with prematurity or any specific congenital malformation or teratogenic risk (21). Because these pregnancy-related problems also occur frequently with maternal use of other illicit drugs, it is unclear to what extent they represent a specific pharmacologic action of PCP.

Adult chronic PCP users do not generally show evidence of clinically significant cardiovascular, hematologic, renal, or hepatic toxicity (1). Although PCP in vitro has been shown to block epinephrine-induced activation and aggregation of platelets and to depress stimulated antibody and interleukin production (possibly mediated by sigma receptors on leukocytes) (28, 52), it has not been associated clinically with coagulation or immune dysfunction.

Although PCP itself will not see therapeutic use because of its adverse psychoactive effects, its use as an experimental tool for increasing understanding of the neuropsychopharmacology of NMDA receptors is likely to bear therapeutic fruit in the future. The development of potent and specific ligands for the PCP receptor and other sites associated with the NMDA–receptor complex is likely to result in improved anticonvulsants, neuroprotective agents for the treatment of stroke and other ischemic and anoxic brain injuries, and novel antipsychotic medications free of the side effects associated with the currently available antidopaminergic neuroleptics (14, 29, 30, 69).

1. Separating PCP-like effects from therapeutic effects of NMDA antagonists. To maximize therapeutic utility, an NMDA antagonist should not produce significant PCP-like behavioral/psychological side effects or have PCP-like abuse liability. All NMDA antagonists tested to date that act at the PCP site in the channel are very likely to produce PCP-like side effects in humans. However, two lines of evidence suggest that separation of therapeutic from adverse PCP-like effects may be possible. First, different classes of site-selective NMDA antagonists produce different profiles of neurochemical and behavioral effects in animals (18, 82, 83); for example, competitive NMDA antagonists may have a better "therapeutic index" than noncompetitive antagonists. Second, early results with systemically active drugs that act as glycine-site and polyamine-site NMDA antagonists show that they may produce behavioral effects that are even more dissimilar from those of PCP than are obtained with competitive antagonists (4, 65).

2. Subtypes of NMDA/PCP receptors. The NMDA receptor is composed of a number of different subunits. Different heteromeric NMDA channels, created by oocyte expression, show differential sensitivities to PCP, ketamine, dizocilpine, and NANM (85). Biochemical and neuroimaging studies have shown differences among PCP and NMDA receptors in different brain regions (27, 58) and between agonist- and antagonist-preferring forms (45). These early findings suggest that identification of subtypes of PCP receptor sites may allow development of PCP-site antagonists with more selective pharmacological properties.

3. Pharmacodynamics of PCP intoxication. The most recent human experimental studies of PCP intoxication occurred over a decade ago, without use of brain imaging, electroencephalographic brain mapping, or computerized neuropsychological testing. Future studies using these modern techniques, correlated with pharmacokinetic measurements, should help clarify the neuropsychopharmacological mechanisms of acute PCP intoxication, as well as provide experimental models for studying the neuropsychopharmacology of psychosis and for developing potential treatments for acute PCP intoxication. Such studies are already underway using the PCP analogue ketamine, thereby avoiding some of the potential problems posed by using PCP itself (34). Comparative controlled clinical trials are needed to evaluate treatments for PCP intoxication. Prospective, longitudinal studies with appropriate comparison groups, such as non-PCP-abusing drug users, are needed to identify the incidence, nature, and course of long-term consequences of PCP use.

4. Treatment of PCP abuse/dependence. Both pilot studies and controlled clinical trials are needed to develop and evaluate (a) improved treatments for PCP abuse/dependence, including both psychosocial methods such as cognitive therapy and behavior modification, and (b) new pharmacotherapies based on knowledge of the neuropharmacology of PCP.

published 2000