Can I just take one 800 mg tablet at night so I finally sleep?
Knock: eight hundred milligrams is not a lullaby
Highest-stakes line: 800 mg is a later PHN rung on a three-times-daily clock. It is not evidence that gabapentin is a sleep tablet.
The knock is a lilac blister stamped 800 mg and a message that says take it for sleep. That is the wrong room. Neurontin film-coated 800 mg tablets exist so a later daily total - 2400 mg as 800 three times a day, for example - can be built after 100, 300, and 400 mg capsules have already done the early nights. They are not a hypnotic start.
Postherpetic neuralgia is the pain indication the US label actually signed. Adjunctive partial-onset seizures is the other. Painful diabetic neuropathy has a famous trial and a thin official stamp. Insomnia, restless nights after a long shift, 'just something gentler than a z-drug' - those arrived by habit, not by a primary sleep endpoint.
I will walk this as a pastel card. Knock on the 800 mg misunderstanding. Chart Rowbotham and Rice. Ward-round Backonja and the off-label hallway. Sign out the opioid stack and the taper. If you need creatinine cut-steps or a full capsule-by-capsule climb, open the gabapentin profile.
Morphine plus gabapentin: combo ink and a 2019 warning
Gilron 2005 (NEJM) is the combination paper I still put on the peach wall. Single-centre, four-period crossover: morphine, gabapentin, both, and a lorazepam-flavoured active placebo. Target ceilings ran to morphine 120 mg alone, gabapentin 3200 mg alone, and a lower pair together (morphine 60 mg plus gabapentin 2400 mg). Mean pain was lowest on the combination (about 3.06) versus morphine 3.70, gabapentin 3.72, and placebo 4.49. Each drug needed a smaller mean daily dose in the combo weeks.
That is efficacy ink, not a safety holiday. The Neurontin label already notes that a 60 mg controlled-release morphine capsule two hours before 600 mg gabapentin raised gabapentin AUC by 44% in a small study. Morphine kinetics barely moved. You cannot treat the pair as two innocent neighbours.
December 2019, FDA: serious breathing problems with gabapentinoids when respiratory risk is present - opioids and other CNS depressants, COPD, older age. New label warnings. A push for more abuse-potential work, especially with opioids. Start low if you must co-prescribe. Watch sedation and respiratory rate. I say this in full so nobody reads Gilron as 'always add morphine.'
On a Dublin round I will still combine when neuropathic pain is brutal and both drugs are already in play at modest doses. I will not open a sleep complaint with 800 mg plus leftover oxycodone. That stack is how a secondary diary becomes an ICU transfer.
Off-label rooms that never got a chart
After the pain diaries were published, gabapentin spilled into hallways that never ran a Rowbotham. Perioperative anxiety. Alcohol-withdrawal adjuncts. Migraine prevention. Restless legs (Horizant, a different gabapentin-enacarbil stamp, is its own story). Hot flashes. 'Something for sleep' on a discharge bundle. The evidence in those rooms is mixed, small, or simply absent as a primary endpoint.
I do not scold every off-label mark. I scold the ones that skip the pain clock and skip the kidney line and skip the other sedatives on the kardex. A person who cannot sleep because PHN is burning at 2 a.m. is still in a pain room. A person who wants 800 mg because a neighbour uses it as a nightcap is not.
Misuse signals grew as opioids got harder to write. That does not make every PHN patient a diversion case. It does mean I no longer shrug when a young adult wants a high-dose film-coat with no rash history and no seizure history. The card should look like a trial population, or I ask why it does not.
If you came here from the parasite field study, you are in a different pastel room. Cross-link only if you need it: Stromectol field card is worms and platforms, not calcium-channel ligands.
Taper evidence, not a hard stop
The Neurontin circular is blunt on discontinuation. Abrupt stop can raise seizure frequency in people who take it for epilepsy. Withdrawal symptoms - anxiety, insomnia, nausea, pain, sweating - show up in case series even when the original indication was pain. The labeled minimum taper is about one week. Longer exposure and higher daily totals usually get a slower staircase on my ward.
That insomnia in withdrawal is a cruel irony for anyone who was using the tablet 'for sleep.' The rebound night is not proof they had a sleep indication. It is proof the nervous system noticed a ligand leaving. I chart the taper in writing: which film-coat, which day, who to call if jerking or a breakthrough seizure appears.
I do not cut 2400 mg to zero on a Friday because a discharge pharmacist wants a short list. Three 800 mg tablets a day become two, then the 600s, then the 400s, over more than a week if the person has been on the drug for months. Renal impairment stretches the tail; you may need smaller steps, not a macho stop.
Evidence here is more label-plus-series than a giant RCT of taper speeds. I still treat the one-week floor as a floor, not a target for everyone. Sign the plan with the same clinician who owns the pain or seizure diagnosis.
Ward-round: Backonja's DPN numbers, still off the US stamp
Backonja 1998 (JAMA) is the diabetic-neuropathy trial residents quote as if it were a second official pain indication. It is not, on the US Neurontin stamp. It is still a real clock. Patients with painful diabetic neuropathy climbed week by week - 900, then 1800, then 2400, then 3600 mg a day if they could stand it - against placebo.
Mean daily pain moved from 6.4 to 3.9 on gabapentin and from 6.5 to 5.1 on placebo (P<0.001). Global 'moderate or much improved' sat near 60% versus 33%. Sleep interference improved as a companion diary, again secondary. Somnolence and dizziness led the adverse list. That is a neuropathic-pain signal, not a primary insomnia program.
I still offer this paper when a person with burning feet asks whether the molecule 'does anything.' I also say the honest next sentence: guidelines may list gabapentin for painful DPN; the US tablet stamp that matches this site's 800 mg lock is PHN plus seizures. Off-label is allowed. Pretending the stamp already includes DPN is sloppy.
And I do not open DPN care with an 800 mg tablet on night one. The trial climb started at hundreds, not at the top film-coat. Kidneys set the ceiling. That ceiling is on the profile, not here.
Chart: Rowbotham and Rice on the PHN clock
1993-94
US gabapentin approval as adjunctive therapy for partial seizures. Pain is not yet on the stamp.
1998
Rowbotham JAMA PHN: 229 randomised, up to 3600 mg/day. Pain 6.3 to 4.2 versus 6.5 to 6.0. Sleep is a secondary diary.
1998
Backonja JAMA DPN: forced climb to 3600 mg/day. Pain 6.4 to 3.9 versus 6.5 to 5.1. Still not a US labeled pain indication.
2001
Rice Pain PHN: 334 adults, 1800 or 2400 mg/day. About -34% pain change versus -16% placebo. 800 mg is a later packing chip, not the opener.
2005 / 2019
Gilron NEJM combination study, then FDA December 2019 respiratory warning when gabapentinoids stack with opioids or frail lungs.
Rowbotham's 1998 JAMA trial is the first PHN room I keep. Sixteen US centres, August 1996 to July 1997, 229 adults with pain at least three months after the zoster rash healed. Four-week titration toward 3600 mg a day or the most they could stand, then four weeks fixed. Average daily pain (0-10) fell from 6.3 to 4.2 on gabapentin and from 6.5 to 6.0 on placebo (P<0.001). Sleep interference moved with the pain score. It was not a sleep-disorder trial.
Eighty-nine of 113 on drug finished; 95 of 116 on placebo finished. Somnolence, dizziness, ataxia, and edema showed up more on gabapentin. Withdrawals were in the same neighbourhood (13.3% versus 9.5%). That pattern still shows on a ward: people feel fogged on the way up, then the diary, if it is going to move, has often already twitched by week one.
Rice 2001 (Pain) then fixed the doses UK-style. Three hundred thirty-four adults, mean age 73, forced titration to 1800 or 2400 mg a day in three divided doses, seven weeks. Final pain change versus baseline: -34.5% at 1800, -34.4% at 2400, -15.7% on placebo. The gap versus placebo sat near 19% for both active arms. Sleep-interference diaries followed the same shape. More than 50% pain cut favoured drug. Dizziness and somnolence clustered in the climb, not as a sold bedtime effect.
Two PHN clocks. One US forced climb to 3600. One UK pair at 1800 and 2400 that looked interchangeable on the primary diary. Neither clock started at a lone 800 mg night dose. The 800 mg tablet is how you later pack a 2400 mg day, not how the trial opened.
Why 800 mg arrives late on the card
Where 800 mg sits on the climb
- Early chips: 100 / 300 / 400 mg capsules on the way to 900 mg by day three in the PHN stamp.
- Later film-coats: 600 and 800 mg tablets to pack 1800-3600 mg days.
- Rice's interchangeable PHN arms: 1800 and 2400 mg/day - 2400 is three 800s, not one bedtime swallow.
- Wrong knock: 800 mg 'for sleep' with no pain diary and no kidney line.
Neurontin capsules come as 100, 300, and 400 mg. Tablets come as 600 and 800 mg. The PHN label still opens at 300 mg on day one, 600 on day two, 900 on day three, then a climb toward 1800 mg a day as a usual effective total, divided. Some people later live at 2400 or 3600 if kidneys and fog allow. The 800 mg tablet is a packing tool for those later totals.
Three times 800 is 2400. That number sits next to Rice's higher fixed arm. It is not a starting ritual and it is not 'the sleep strength.' I have watched people jump to 800 mg at bedtime because the box looked decisive. They wake groggy, skip work, and declare the drug useless for pain they never titrated.
Renal clearance is the quiet governor. Gabapentin leaves unused in urine. Drop the creatinine clearance and the same 800 mg chip becomes a larger exposure. Elderly wards already carry the FDA's extra respiratory caution. Stacking a late-rung tablet onto a thin eGFR is how a 'small extra' becomes a fall.
So when this site locks 800 mg in the title, it is naming the later film-coat on an evidence card, not advertising a first-night hypnotic. Read it that way or you will mis-file the whole study.
Sign-out: pain trials first, 800 mg later, sleep last
Pain diaries first. Later 800 mg packing second. Sleep-as-indication last, which means almost never.
Sign-out in one peach strip. Rowbotham moved PHN from 6.3 to 4.2 against a flat placebo. Rice showed 1800 and 2400 mg/day looking like twins on the diary. Backonja moved DPN numbers without earning a US pain stamp. The 800 mg tablet packs later days. It does not open a sleep clinic.
Off-label hallways are crowded. Gilron's combination week is real and so is the 2019 breathing warning. Taper for at least a week, longer after a long stay at high totals. Do not hard-stop a seizure medicine on a whim.
Talk with the prescriber who can see your eGFR, your opioid list, and your actual diagnosis before anyone adds or pulls an 800 mg chip. This card teaches. It does not write your kardex.
Last Updated
From the postbag
Questions this study raised
Answered by Dr. Saoirse Byrne, MD · Internal medicine & clinical pharmacology
These knocks are about an 800 mg film-coat treated as a nightcap, about PHN numbers, and about stacks I will not shrug off. Names are editorial. Answers are teaching, not your taper sheet.
Not as a plan I would sign. The 800 mg film-coat is a later packing chip on a three-times-daily pain or seizure climb. Rowbotham and Rice never randomised a single bedtime 800 mg as insomnia therapy. Sleep diaries moved when pain moved. If nights are wrecked by PHN, the honest card is a slow climb toward a trial-like daily total, kidneys allowing - not a hypnotic shortcut. If there is no neuropathic diagnosis, I would not start gabapentin for sleep at all. Ask the clinician who knows your other sedatives before you invent a nightcap.
Rowbotham only dropped pain by about two points. Is that worth the fog?
On an 11-point diary, 6.3 to 4.2 versus 6.5 to 6.0 is a real group effect and still leaves many people in moderate pain. That is neuropathic pain, not a miracle. Rice's responders (more than 50% cut) are the ones who feel the drug was 'worth it.' Fog, dizziness, and edema are how others leave the climb. I do not promise a zero. I promise a diary and a stop rule if function gets worse than the pain was. Your own clinician has to weigh the two-point mean against your gait and your job.
I have burning diabetic feet. Why isn't that on the US stamp if Backonja was positive?
Because a positive JAMA trial is not the same act as an FDA pain indication. Backonja's climb to 3600 mg moved mean pain from 6.4 to 3.9 against 6.5 to 5.1, and more people felt moderately or much improved. Guidelines often still list gabapentin for painful DPN. The Neurontin US pain stamp that matches this card is postherpetic neuralgia. Off-label DPN use is a conversation, not a lie about the label. I would still start low and watch kidneys, not open at 800 mg.
Rice said 1800 and 2400 looked the same. Why would anyone go to three 800s?
Because some people on a desk still titrate toward Rowbotham's higher tolerated totals, and because 2400 is a clean three-times-800 pack once you already know the drug helps. Rice's primary diary did not reward the extra 600 mg in that seven-week UK design. More milligrams can mean more somnolence without more relief. I treat 1800 as a usual PHN destination and 2400 as a later experiment, not a badge. The 800 mg tablet is convenient packaging, not a target you owe the blister.
Gilron said the combo beat either drug. Should my oxycodone go up with Neurontin?
Gilron used morphine, not oxycodone, in a crossover with lower mean doses of each drug during combination weeks, and mean pain near 3.06 versus the mid-3s on monotherapy. That is not an order to raise your opioid. The label also records a 44% gabapentin AUC rise after 60 mg CR morphine before 600 mg gabapentin. FDA's December 2019 warning is about breathing when gabapentinoids meet opioids, other sedatives, COPD, or older lungs. If both drugs are already on your list, the usual move is caution and monitoring, not an automatic uptitration. Do not add 800 mg on top of leftover oxycodone to chase sleep.
How fast can I stop? I only used it for two restless months.
The circular's floor is a taper over at least one week, and that floor exists because abrupt stops have raised seizure risk in epilepsy and produced withdrawal - including insomnia - in people who were not taking it for seizures. Two months is not 'nothing,' especially if you reached an 800 mg three-times-daily pack. I would still step down through smaller chips rather than jump from 2400 to zero on a Sunday. If you ever had seizures, do not self-taper. Call the prescriber who owns that diagnosis.
My discharge list has gabapentin 'for sleep' and no pain diagnosis. What now?
Bring the list back to the team that wrote it, or to your GP, and ask which trial they thought they were copying. There isn't a primary insomnia program behind that line. You may be carrying fog, a fall risk, and - if an opioid travelled with you - a respiratory stack for a problem gabapentin was not shown to treat. A taper plan belongs to a named clinician. Do not stop hard on the drive home, and do not refill forever because the discharge bundle felt official.
Does improved sleep on the PHN diary mean I can skip the pain diagnosis?
No. Those diaries were secondary. People slept a little better when the burning eased. That is not a license to write 800 mg on an insomnia template. If your nights are wrecked and there is no PHN, no DPN trial-like picture, and no seizure history, look at sleep hygiene, other medicines, mood, and alcohol before you borrow a neuropathic drug. I would rather fix the cause of the night than borrow Rice's somnolence as a feature.
Where is the actual titration rail if this page will not retell it?
On the gabapentin profile - creatinine steps, capsule order, holds. This study is the evidence envelope: Rowbotham, Rice, Backonja, the late 800 mg film-coat, sprawl, Gilron, the 2019 warning, and a one-week-or-longer taper. For a completely different Oracel field card see Stromectol two-tablet bands versus viral claims. Neither page replaces a prescriber who has met you.
Read these as teaching on a general point, not as a care plan for whoever asked. Anything specific to you belongs with a prescriber who can see your notes, bloods and full medicine list at once.