Pain and medicines over the months ahead

For her partner, 10 September 2026. The longer view of the drugs: where the dose sits, how it changes, what else exists, how it stays safe over months, and how it comes down afterwards. Every figure is sourced in research thread 20 (56 papers and guidelines, PubMed IDs checked). No trial population is skull base chordoma; where I am reasoning from general cancer-pain evidence it says so. The treating team's judgement comes first.

1. Where 20 mg twice a day sits

The NICE, European and ESMO cancer-pain guidelines all start oral morphine at 20 to 30 mg a day. At 40 mg a day of slow-release morphine plus 30 to 50 mg of Oramorph, she is on 70 to 90 mg a day: one step above a starting dose, low to moderate. Patients in the big cancer-pain trials average 160 to 270 mg a day. There is no maximum dose; side effects and diminishing returns set the limit. About 19 in 20 cancer patients get their pain to mild or none on oral opioids (Wiffen 2017, low-quality evidence).

2. How the dose moves

The rule. Count the breakthrough doses. Rescue should be about a sixth of the daily regular dose (Davies 2009). If she needs three or more rescues on most days, yesterday's total (regular plus rescue) becomes the new regular dose, split in two, and the rescue dose is recalculated. Her arithmetic on 9 September (40 plus 30 to 50 mg) points to 30 to 45 mg twice a day with a rescue of 10 to 15 mg. Steps are 30 to 50 percent at a time, no more often than once a day for the slow-release form.

When more is not the answer. Two rises in a row with no better relief, or pain that spreads and worsens after a rise, is a known trap called opioid-induced hyperalgesia, where the drug itself makes the nervous system more sensitive (Higgins 2019). At that point the move is to switch opioid or add a nerve-pain drug, not to go up again.

3. Switching opioids

No trial compares switching with staying; case series say most people do better after a switch (Quigley 2004, Cochrane). The reasons to switch are side effects that outgrow the benefit (sedation, fog, itch, nausea) or rises that stop working. The options, with the usual conversion from oral morphine: oxycodone tablets (about two thirds the milligrams), hydromorphone (a fifth), a fentanyl patch (for stable pain only), and methadone (specialist-only; it has some action on nerve pain but a long unpredictable half-life). Whatever the switch, the calculated new dose is cut by a quarter to a half and built back up, because tolerance does not carry over fully (European guideline, Caraceni 2012).

4. The nerve-pain drugs

Ranked by how many people need to take them for one to get good relief (Finnerup 2015, the international neuropathic pain guideline): amitriptyline or nortriptyline 3.6, duloxetine 6.4, gabapentin 7.2, pregabalin 7.7.

• Gabapentin. Licensed to 3,600 mg a day, but the gut absorbs less as the dose rises (about 60 percent at 900 mg, about a third at 3,600 mg), so the benefit flattens beyond 1,800 to 2,400 mg. The trial in cancer patients on opioids went to 1,800 mg and cut average pain by about one point in ten (Caraceni 2004). A fair trial is two to four weeks at 900 to 1,800 mg a day; if nothing by then, it is the wrong drug, not too little of it.
• Pregabalin. The swap if gabapentin fails: absorbed evenly at any dose, twice a day, 150 to 600 mg. In one trial of cancer nerve pain it beat both gabapentin and amitriptyline and reduced morphine use (Mishra 2012, single centre).
• Amitriptyline or nortriptyline at night, 10 to 75 mg. Old, cheap, best numbers, weakest modern trials. Nortriptyline is the less sedating one. Not with a heart rhythm problem.
• Duloxetine 60 mg. Good trials in diabetic and chemotherapy nerve pain, none in tumour pain.
• Carbamazepine or oxcarbazepine. The first-line drugs for classical trigeminal neuralgia, the second-long electric shocks. Her pain is continuous with flares, which is a different pattern, and in the largest series about one in ten trigeminal neuralgias were caused by a tumour; those patients responded to drugs at first, and removing the tumour was what fixed it (Cheng 1993). Carbamazepine also causes dizziness, double vision and unsteadiness, which would muddle her eye observations. If true shocks appear, oxcarbazepine is the cleaner choice.
• Dexamethasone. The steroid that shrinks swelling around a compressed nerve; guidelines support a short course. The evidence is modest: Cochrane found low-quality benefit at one week, and the best trial found no extra pain relief but better appetite and energy (Paulsen 2014). It is held back before biopsy because steroids can shrink a lymphoma and spoil the diagnosis; once tissue is taken, it is the first extra thing to ask the team about.

5. The specialist tier

Complex cancer pain belongs with the hospital pain team and the supportive (palliative) care team together, and the referral is a symptom-control tool, not a statement about outlook: early involvement improves quality of life and mood in trials (Temel 2010; Cochrane 2017). What they can add:

• A sphenopalatine ganglion block, local anaesthetic through the nose to the nerve cluster behind it. Case series in head and neck cancer report relief lasting days to months; low risk, repeatable; the first block to ask about.
• Gasserian (trigeminal) ganglion procedures. Long established for neuralgia, small recent cancer series, very low-quality evidence overall (Fujimura 2026). The needle passes into Meckel's cave, next to where her tumour already sits against the carotid, so only an interventional consultant with her images can say whether it is safe.
• Ketamine. The one large trial found no benefit over placebo when added to opioids, with side effects in one in six (Hardy 2012). A last resort.
• Intravenous lidocaine. Relief for the hours after the infusion, nothing lasting (Zhu 2019). A bridge for a crisis, not a plan.

Ask for the pain team and supportive care referrals together now, through Mr Syrris's team or the GP, not after a crisis.

6. Staying safe over months

• Bowels. Regular senna plus a macrogol sachet from day one, increased with every opioid step; if laxatives fail, naloxegol tablets block the opioid in the gut without touching the pain relief (NICE TA345).
• Nausea and sedation fade within one to two weeks of each change; the week after each rise is the risky window for thinking and balance.
• Gabapentin plus morphine together raise the risk of dangerously slow breathing; in a large study co-prescription raised opioid-related deaths by about half (Gomes 2017; MHRA warning 2017). At her doses the absolute risk is small; the rules are no alcohol, no extra sedatives, slow gabapentin steps, and someone awake in the house on the night after each change.
• Driving. UK law sets a blood morphine limit, with a defence for prescribed use taken as directed and not impaired. That is moot for now: the DVLA bars driving with uncontrolled double vision. Later, a stable dose does not stop her driving, but never in the days after a dose change.
• Dependence is not addiction. Months on morphine produce tolerance and physical dependence, which are expected and reversible. Addiction (compulsive use despite harm) is uncommon in cancer pain without a prior substance problem, never zero, somewhat higher in younger adults (Ballantyne 2003). The protections: one prescriber, slow-release as the backbone, every rescue dose counted in the diary, and a written taper plan before surgery.

7. Coming down afterwards

After surgery and proton therapy the pain should fall and the dose follows it. After months of use the guidelines reduce opioids by 5 to 20 percent every two to four weeks, slower near the end (VA/DoD 2022; NICE NG215). Withdrawal (sweats, cramps, diarrhoea, anxiety, restless legs) means the step was too big: back one step and hold. Pain and function usually improve, not worsen, with a supported reduction (Frank 2017). Gabapentin comes down over at least a week, in 300 mg steps every three to seven days after months of use; stopping it dead can cause agitation and, rarely, seizures.

Honest caveats: nerve pain that has been present for a year can outlast the tumour if the nerve is damaged; compression pain usually eases as the mass comes off the nerve, and nobody can promise which hers is. After endoscopic surgery the main complaints are nasal crusting and sinus discomfort for three to six months; in 100 chordoma patients treated with protons, severe pain during treatment affected 3 and lasting headache 2 (Chhabra 2023). No chordoma study reports pain scores before and after treatment; that gap is real.

The asks, in order

1. Fold the breakthrough count into the regular dose with Dr Liu this week, using the rule in section 2.
2. A written gabapentin schedule with a stop point: 900 to 1,800 mg a day for two to four weeks, then keep or swap to pregabalin.
3. Referral to the Southampton pain team and supportive care team, asked for now.
4. Dexamethasone on the list of questions for the team the day the biopsy is taken.
5. A taper plan written before surgery, so the way down is agreed before the way up is finished.

Deferred option, added 14 September 2026: psilocybin, after treatment and not before

Kept here so it is not lost. The full analysis is in em-psilocybin-pain.md and research thread 23.

Psilocybin is not an analgesic and there is no trial anywhere showing it lowers pain intensity in any patient group. The clinical signal is confined to migraine and cluster headache, and the newest of those trials found an ordinary antihistamine did as well.

The reason it is recorded here rather than dismissed is the phenotype argument. Every condition the field studies (fibromyalgia, chronic low back pain, migraine, cluster, phantom limb, drug-refractory neuropathic pain) is one where the peripheral generator is absent, minimal or static and the pathology sits in central processing. That is what 5-HT2A driven network reorganisation is aimed at.

the patient does not have that phenotype now. She may have it later. If she has surgery and proton therapy and is left with persistent neuropathic facial pain from a static injury to the trigeminal system, with no growing lesion behind it, her pain converts from structural to nociplastic, which is exactly the population the hypothesis addresses. Persistent facial neuropathic pain after skull base treatment is common and is currently treated badly.

So: revisit twelve to eighteen months after definitive treatment, if pain persists, and only then. Conditions for revisiting: histology known and disease controlled; the seizure question resolved with an EEG; and by then, ideally, a trial with an active placebo and a pain-intensity primary endpoint that has reported. Nothing meeting that last description exists in September 2026.

Also relevant to the taper in the section above: the psilocybin literature on opioid reduction works by restoring motivation and reducing craving. The patient has no deficit in either, so it offers nothing to her taper. What blocks her taper is unrelieved sensory pain, which is what the four asks above address.