Prescribing Gabapentin

Why Gabapentin Matters (and Why It's Everywhere)

Gabapentin is one of the most prescribed drugs in America, and almost none of that prescribing is for what it's actually approved to treat. Fewer than 1% of outpatient gabapentin prescriptions are written for an FDA-labeled indication. The rest is off-label: anxiety, insomnia, alcohol use disorder, generalized pain, restless legs, hot flashes, akathisia, bipolar "augmentation." It has become psychiatry's utility infielder, reached for when a clinician wants a non-addictive, hepatically clean, sleep-deepening, anxiety-softening agent and doesn't want to add another benzodiazepine or another antipsychotic.

Some of that reputation is earned. Gabapentin has real, if modest, roles in alcohol and cannabis use disorders, in a specific flavor of insomnia, and as an anxiety adjunct in patients with substance use histories. And it has two genuine structural advantages that keep it in the toolkit: it is not metabolized by the liver, it is excreted unchanged by the kidney, so it has essentially no CYP450 drug interactions, and by itself it is remarkably safe in overdose.

But a great deal of gabapentin prescribing outruns the evidence. It was the subject of one of the largest off-label-marketing settlements in pharmaceutical history (Pfizer/Warner-Lambert paid roughly $430 million in 2004 over the promotion of Neurontin for uses it was never shown to treat). It does not work for low back pain, sciatica, spinal stenosis, or migraine. It is not a mood stabilizer. And the field is steadily waking up to the fact that gabapentin has a real, under-recognized dependence and withdrawal syndrome, and a genuine misuse liability, especially in combination with opioids, where it contributes to fatal respiratory depression.

This guide is about prescribing it deliberately: knowing the handful of situations where it earns its place, dosing it around its one weird pharmacokinetic quirk, respecting its renal clearance, and never mistaking "well-tolerated and unscheduled" for "harmless."

The Mechanism, and the Misconception Worth Correcting

Start here, because the name misleads almost everyone.

Gabapentin does NOT act on the GABA-A receptor. Despite "GABA" sitting in its name and its structural resemblance to the neurotransmitter, gabapentin does not bind GABA-A, does not bind GABA-B in any clinically meaningful way, and is not a benzodiazepine-like or barbiturate-like drug. This is the single most common misconception about the medication, and it has clinical consequences: it is why gabapentin does not produce classic benzodiazepine tolerance, why its abuse liability (though real) is different in character, and why it is not a reliable substitute for a benzodiazepine in acute anxiety or in benzodiazepine withdrawal.

What it actually does: gabapentin binds the alpha-2-delta (α2δ) subunit of voltage-gated calcium channels in the CNS. By binding this auxiliary subunit, it reduces calcium influx at presynaptic terminals and dampens the release of excitatory neurotransmitters (glutamate, norepinephrine, substance P) in states of neuronal hyperexcitability. This is the same target pregabalin hits: the two drugs are a class, the gabapentinoids, defined by this α2δ mechanism, not by GABA.

The downstream effect is a calming of overactive circuits (which plausibly explains the benefit in neuropathic pain from aberrant firing, in anxiety, and in withdrawal states) without the GABA-A-mediated respiratory drive suppression seen with benzodiazepines when gabapentin is used alone. The respiratory danger appears when it is combined with true CNS depressants (see Toxicity).

Part 1: Indications: Who Is Gabapentin For?

FDA-Approved Uses

• Postherpetic neuralgia (the neuropathic pain that lingers after shingles)
• Partial (focal) seizures, as adjunctive therapy in epilepsy
• Restless legs syndrome, but specifically via the prodrug gabapentin enacarbil (Horizant); plain gabapentin is used off-label for RLS and works well

Note what is not on this list: no anxiety disorder, no insomnia, no bipolar disorder, no alcohol use disorder. Gabapentin has zero FDA approvals for any primary psychiatric condition. Every psychiatric use below is off-label.

The Off-Label Uses That Are Actually Worth It

Alcohol use disorder is the best psychiatric case for gabapentin, arguably its strongest off-label indication in our field. A meta-analysis of seven RCTs found a medium effect size (around 0.4) for reducing heavy-drinking days, and long-term use for relapse prevention is endorsed by APA's alcohol use disorder guidelines. Beyond the drinking outcomes, it does something benzodiazepines and naltrexone don't do as cleanly: it improves the sleep, anxiety, and craving of early and protracted abstinence, the symptom cluster that drives so many relapses. Typical relapse-prevention dosing is around 900 mg/day (range 300 to 3,600).

Alcohol withdrawal. Gabapentin has been studied head-to-head against benzodiazepines and phenobarbital for withdrawal and performs reasonably in mild-to-moderate cases, improving cravings, anxiety, and sleep. A representative protocol: start 1,200 to 2,400 mg/day divided TID/QID, taper to around 600 mg/day over 4 to 7 days, then step down by around 300 mg every few days. The caveat is important: a few withdrawal seizures were observed during gabapentin tapers in trials (not statistically significant, but clinically notable). For that reason, reserve gabapentin monotherapy for mild-to-moderate withdrawal, or use it as an adjunct to a benzodiazepine regimen rather than a replacement in patients at real seizure or DT risk.

Cannabis use disorder. Gabapentin is essentially the only pharmacologic option with placebo-controlled trial data here: a single small trial showed reduced withdrawal symptoms, increased abstinence, and improved executive function. A studied regimen is 1,200 mg/day (e.g., 300 mg AM, 300 mg midday, 600 mg QHS). Evidence is thin (one trial), but in a space with no approved options, it is a defensible try.

Anxiety is a modest, second- or third-line use, and here the honesty matters. The gabapentin anxiety evidence base is small: two RCTs (n approximately 172) covering social phobia and panic. In social anxiety disorder, response was around 32% vs 14% placebo at high doses (averaging around 2,900 mg/day). In panic disorder, gabapentin did not separate from placebo overall (a signal only in a severe subgroup at high doses). For nonspecific/situational anxiety (perioperative, cancer-related), lower doses of 300 to 1,200 mg/day helped in larger trials. Bottom line: gabapentin's anxiety effect is real but weaker than SSRIs (response over 50%) and weaker than its cousin pregabalin (8 RCTs, n over 2,000). Where it earns a place is in the anxious patient with a substance use history or benzodiazepine risk, where you want anxiolysis without a controlled substance.

Insomnia, but only a specific kind. Gabapentin at 200 to 900 mg QHS does not help patients fall asleep faster. What it does is deepen sleep, increase slow-wave sleep, and reduce nocturnal awakenings. So it's the wrong drug for sleep-onset insomnia and the right one for the patient who wakes repeatedly or sleeps shallowly, particularly the patient with comorbid alcohol recovery or menopausal hot flashes. In most insomnia algorithms it sits below trazodone and mirtazapine; reach for it after those, not before.

Postmenopausal vasomotor symptoms (hot flashes). Several controlled trials support gabapentin for hot flashes, a useful two-for-one in the menopausal patient who also has insomnia or anxiety.

Antipsychotic-induced akathisia. Open-label data (300 to 3,600 mg/day) show promise for this miserable, adherence-wrecking side effect. Reasonable to try when propranolol and benztropine fall short.

Restless legs syndrome. Effective in a half-dozen small studies at 300 mg at bedtime (up to 2,400 mg studied). Prefer plain gabapentin over Horizant: it's cheaper, food-independent, taken at bedtime, and lacks Horizant's next-day driving-impairment warning and acetaldehyde-related toxicity concerns.

Where Gabapentin Does NOT Work: Stop Reaching for It Here

• Bipolar disorder / mania. Gabapentin failed large monotherapy trials and is not a mood stabilizer. The only defensible bipolar use is as an adjunct in a patient whose comorbid anxiety or substance use is the actual target, and even then the mood-stabilizing data (two small trials, n approximately 85 total) are unconvincing. For mood stabilization, use lithium, valproate, carbamazepine, lamotrigine, or an antipsychotic. In head-to-head NIMH data, lamotrigine beat gabapentin outright.
• Low back pain, sciatica, spinal stenosis, and migraine. A well-designed BMJ review found gabapentinoids ineffective for these, despite enormous real-world prescribing.
• Cocaine, methamphetamine, benzodiazepine, and opioid use disorders. No benefit, and in opioid use disorder it is actively harmful, associated with increased overdose and hospitalization.

Part 2: The One Pharmacokinetic Quirk You Must Understand

Gabapentin has an unusual absorption profile, and everything about how you dose it flows from this single fact.

Absorption is saturable and nonlinear. Gabapentin is absorbed in the small intestine by an active, saturable transporter. Once you saturate that transporter (around 900 to 1,200 mg per dose), additional milligrams are absorbed at a steeply diminishing rate. The numbers make it vivid:

Doubling the dose from 1,200 to 2,400 mg does not double the drug you actually get. This has two direct consequences:

1. Higher is not proportionally more. Don't assume a patient failing at 1,800 mg TID will get a linear boost at 3,600 mg. You get diminishing returns, and above a point, mostly side effects.
2. You must divide the total daily dose. Because per-dose absorption plateaus, splitting the daily total into smaller, more frequent doses gets more drug into the patient than the same total given once or twice.

The half-life is short: about 6 hours. Combined with saturable absorption, this is the reason gabapentin is dosed three times daily (TID) for anxiety and pain. A once- or twice-daily schedule produces troughs where the drug has largely washed out. (Sleep and RLS indications are the exception: a single bedtime dose is appropriate when you only need nighttime coverage.)

Part 3: How to Start and Dose

Formulation

Use generic immediate-release gabapentin: it's cheap and flexible. It comes in 100, 300, 400 mg capsules and 600, 800 mg tablets, plus a 250 mg/5 mL solution for fine titration. Extended-release and prodrug forms (Gralise, Horizant) exist for specific FDA indications but add cost with little psychiatric advantage.

Start Low, Titrate Over Days to Weeks

• Start 100 to 300 mg at bedtime. Starting low blunts the early dizziness and sedation that otherwise sour patients on the drug.
• Titrate up over several days to a couple of weeks as tolerated, adding a daytime dose and building toward a TID schedule for anxiety or pain.
• A reasonable escalation for anxiety: 300 mg QHS → 300 mg BID → 300 mg TID → titrate further by 300 mg increments to effect and tolerability.

Typical Dosing Ranges by Indication

The usual effective range across indications is 900 to 3,600 mg/day, and, because of saturable absorption, most patients don't gain much above around 1,800 to 2,400 mg/day. Reserve the top of the range for genuine partial responders, and remember it's divided TID.

No Serum Levels

Unlike lithium, gabapentin has a broad therapeutic index and requires no drug-level monitoring. You titrate to clinical effect and tolerability, not to a number.

Part 4: Monitoring

Gabapentin is refreshingly low-maintenance to monitor: the mirror image of lithium.

• No serum drug levels. No therapeutic window to track.
• No hepatic monitoring. It isn't metabolized by the liver, so no LFTs are needed on its account.
• Check renal function. Because gabapentin is cleared entirely by the kidneys, an estimate of renal function (creatinine clearance or eGFR) is the one lab that matters, at baseline and periodically, especially in the elderly and anyone with known or evolving kidney disease. Renal function drives the dose (see Special Populations).
• Clinical monitoring for the CNS effects that actually limit the drug: sedation, dizziness, imbalance, and fall risk, assessed by asking, not by labs.

Part 5: Side Effects and How to Manage Them

Gabapentin is generally well tolerated, and it lacks the metabolic, hepatic, and cardiac liabilities of many alternatives. Its side effects are overwhelmingly CNS and dose-dependent, and most improve with slower titration or a lower dose.

The Common Ones (CNS)

• Sedation and fatigue: the most common, and useful when you're dosing for sleep, a nuisance when you're not. Load the dose toward bedtime.
• Dizziness and imbalance / ataxia: dose-related; the main driver of falls, especially in the elderly. Notably, gabapentin causes less ataxia (and less weight gain) than pregabalin, a real tolerability edge over its cousin.
• Visual changes / nystagmus: usually mild and dose-related.
• Tremor: occasional.

Management: titrate more slowly, lower the dose, and concentrate more of the daily total at night. Counsel explicitly about driving and operating machinery until the patient knows how it affects them.

The Ones People Forget About

• Peripheral edema. Gabapentinoids cause dose-dependent peripheral edema that is not cardiac or renal in origin. It's easy to misattribute. If a patient develops ankle swelling after starting or increasing gabapentin, suspect the drug before launching a cardiac workup, though don't ignore genuine cardiac risk.
• Weight gain. Real, but less than with pregabalin. Worth mentioning to patients up front.

The Reassuring Part

• No black-box warning.
• No hepatotoxicity.
• No QT prolongation or significant cardiac toxicity.
• No metabolic-syndrome burden like the antipsychotics.

The trade-off for all that tolerability is not a scary side effect: it's the withdrawal and misuse profile covered in Parts 7 and 8, which the drug's easygoing reputation tends to hide.

Part 6: Overdose and Toxicity

Gabapentin Alone Is Remarkably Safe in Overdose

Taken by itself, gabapentin has a wide margin. Large ingestions typically produce sedation, dizziness, slurred speech, ataxia, and sometimes lethargy: unpleasant and impairing, but rarely life-threatening, and usually managed with supportive care. This is one of gabapentin's genuine advantages over benzodiazepines and, especially, over the drugs it sometimes substitutes for.

The Danger Is Combination, and It's an FDA Warning

The critical exception: gabapentin combined with opioids or other CNS depressants can cause serious, sometimes fatal respiratory depression. In 2019 the FDA issued a warning about this specific risk after reports of respiratory depression and deaths when gabapentinoids were taken with opioids, benzodiazepines, or in patients with underlying respiratory compromise. Both gabapentin and pregabalin suppress respiratory drive, and that suppression is dangerously additive with opioids, which also potentiate the opioid "high," a key reason for the misuse pattern.

Highest-risk populations for respiratory depression:

• Patients on concurrent opioids (prescribed or illicit)
• Patients on benzodiazepines or other sedatives
• Patients with COPD or other respiratory disease
• Elderly patients
• Patients with renal impairment (accumulation)

Part 7: Drug Interactions: Where Gabapentin Genuinely Shines

This is gabapentin's marquee advantage, and it's worth stating plainly.

Essentially No Pharmacokinetic Interactions

• Gabapentin is not metabolized by the liver. It is excreted unchanged by the kidneys.
• Therefore it has no CYP450-mediated drug interactions: it neither induces nor inhibits hepatic enzymes, and its levels aren't altered by drugs that do.
• This makes it exceptionally easy to combine with complex regimens: patients on multiple CYP-metabolized psychotropics, patients on antiretrovirals, transplant patients on tacrolimus, patients with polypharmacy where every added drug threatens a new interaction. Gabapentin slots in cleanly.

For a clinician juggling a medically complicated patient, this is a real and underappreciated selling point, one that lithium, valproate, carbamazepine, and most antidepressants can't match.

The One Interaction That Matters Is Pharmacodynamic, Not Pharmacokinetic

• Additive CNS and respiratory depression with opioids, benzodiazepines, alcohol, and other sedatives (see Toxicity). This is not a metabolic interaction; it's overlapping depressant physiology, and it's the dangerous one.

A Minor Practical Note

• Antacids (aluminum/magnesium) can modestly reduce gabapentin absorption; separate dosing by around 2 hours if a patient uses them heavily. Rarely clinically decisive.

Part 8: Discontinuation, Dependence, and Misuse: The Under-Recognized Part

The friendliest thing about gabapentin (no scheduling in most states, no black box, well tolerated) is exactly what causes clinicians to underestimate its dependence and misuse potential. Correct that.

There Is a Real Withdrawal Syndrome: Taper, Don't Stop Abruptly

Because gabapentin does not produce classic benzodiazepine-style tolerance, it was long assumed to be free of withdrawal. It is not. Abrupt discontinuation, especially after longer, higher-dose use, can produce a withdrawal syndrome:

• Anxiety, restlessness, agitation
• Insomnia
• Sweating, tremor
• Nausea
• A lowered seizure threshold: the reason this matters most in patients with epilepsy or heavy alcohol history

How to stop: taper gradually. A practical approach is to decrease by around 300 mg every few days, slowing further near the end and individualizing to the patient's dose, duration, and seizure risk. In someone on high chronic doses or with a seizure history, go slower. Never stop a long-term, high-dose patient cold.

The Misuse Liability Is Real and Growing

By itself gabapentin is not strongly rewarding, which is why it isn't a federally scheduled substance. But recognition of its misuse has grown sharply:

• It potentiates opioids, enhancing the high and, dangerously, the respiratory depression. This is the core misuse pattern and the driver of gabapentin-involved overdose deaths.
• Dose escalation: some patients drift into a "more is better" pattern, taking supratherapeutic doses.
• Diversion in corrections and addiction settings is well documented.
• Scheduling is spreading. Gabapentin is not federally scheduled, but a growing number of states have made it a Schedule V controlled substance or added it to prescription-drug-monitoring programs precisely because of misuse concerns. (Pregabalin, by contrast, has been Schedule V federally since approval.) Know your state's status: it changes what you must document and report.

Part 9: Special Populations

Renal Impairment: The Mandatory Adjustment

This is the population where gabapentin dosing is not optional to get right. Because the drug is cleared entirely renally, impaired kidneys mean accumulation, oversedation, ataxia, and fall risk. Adjust to creatinine clearance:

Patients on hemodialysis require dosing around the dialysis schedule (gabapentin is dialyzable, so a supplemental post-dialysis dose is typically needed). Check renal function before you dose, not after a fall.

Elderly

The elderly get a double hit: age-related renal decline (so they need the renal adjustments above, whether or not anyone has calculated a clearance) plus heightened sensitivity to sedation, dizziness, and imbalance, which translates directly into fall and fracture risk. Start low, titrate slowly, keep the total dose modest, check renal function, and reassess gait and falls at each visit.

Hepatic Impairment

No dose adjustment needed: there's no hepatic metabolism to impair. Another quiet advantage in the medically complex or cirrhotic patient (where valproate, carbamazepine, and many psychotropics are problematic).

Pregnancy and Lactation

Data are more limited than for older agents, and the source material here is thin, so counsel from general principles rather than false precision. Gabapentin crosses the placenta and is excreted into breast milk. It is not among the highest-teratogenic-risk anticonvulsants (it is not valproate), but a reflexive assumption of safety isn't warranted either. Decisions should be individualized with the patient and their obstetrician, weighing the often modest psychiatric indication against uncertainty, and remembering that for many of gabapentin's off-label uses (mild insomnia, situational anxiety), there are non-drug or better-characterized alternatives in pregnancy.

Children and Adolescents

Used in pediatric epilepsy under specialist care. For psychiatric use, note the negative pediatric bipolar trial (ages 7 to 18), another reminder that gabapentin is not a mood stabilizer at any age.

Part 10: Gabapentin vs. the Alternatives

vs. Pregabalin (Lyrica): The Intra-Class Decision

They share the α2δ mechanism, but they diverge in ways that matter:

The practical read: for anxiety, pregabalin is the better-evidenced gabapentinoid and reasonable to try first. For tolerability (less weight gain, less ataxia) and cost, gabapentin wins. For the saturable-absorption reason, gabapentin is fussier to dose (TID, plateau above around 900 to 1,200 mg/dose); pregabalin's linear kinetics make it more predictable.

vs. Benzodiazepines (for Anxiety)

• Benzodiazepines: faster onset, stronger acute anxiolysis, better anxiety evidence, but tolerance, dependence, cognitive effects, and high abuse liability.
• Gabapentin: slower, more modest anxiolysis, but no classic tolerance and lower abuse liability. This is precisely why it's attractive in the anxious patient with a substance use disorder or benzodiazepine-risk profile: you get some anxiolysis without handing an addiction-prone patient a benzodiazepine. Just don't expect benzodiazepine-level relief, and don't use it for benzodiazepine withdrawal (different mechanism, unreliable there).

vs. SSRIs (for Anxiety)

SSRIs remain first-line for anxiety disorders, with response rates over 50% vs gabapentin's around 32% in social phobia. Gabapentin is an adjunct or an alternative for patients who can't tolerate or won't take an SSRI, or where a substance use history makes the gabapentinoid attractive.

Why Is Gabapentin So Overused, Then?

Because it's cheap, unscheduled (mostly), interaction-free, non-hepatotoxic, and feels safe, so it gets reached for reflexively for pain, sleep, and anxiety far beyond where the data support it. The corrective isn't to abandon it; it's to use it where it actually works (alcohol and cannabis use disorders, a specific insomnia phenotype, substance-comorbid anxiety, akathisia, RLS) and to stop using it where trials say it fails (back pain, sciatica, migraine, mania, and, dangerously, opioid use disorder).

The Bedside Cheat Sheet