Trigeminal Neuralgia: A Hyponatremia Reading Against Two Years of Pain Control
Carbamazepine controls her facial pain this well precisely because it works for the large majority of trigeminal neuralgia patients who can tolerate it. A confusional fall and a sodium of 124 test whether tolerating it any longer is actually the safer choice.
Y.W. describes the pain that brought her to neurology two years ago the way most trigeminal neuralgia patients eventually learn to: not as a headache, but as electricity, a jolt through the right side of her face triggered by something as small as brushing her teeth or a gust of wind, severe enough in the weeks before diagnosis that she’d stopped chewing on that side entirely and lost eleven pounds without trying. She is 68, retired from three decades as a public-accounting firm’s office manager, and of Han Chinese descent — a detail that mattered directly at the start of her treatment, since carbamazepine carries a real, FDA-labeled risk of Stevens-Johnson syndrome and toxic epidermal necrolysis strongly associated with the HLA-B*1502 allele, carried at meaningfully higher frequency in patients of Han Chinese and broader Southeast Asian ancestry. She was tested before her first dose, came back negative, and started carbamazepine with that specific risk genuinely screened rather than assumed away.
Two years later, her facial pain is essentially gone — full response, the outcome carbamazepine achieves in the large majority of trigeminal neuralgia patients who can tolerate it, and the reason it remains first-line despite everything else about it. What brought her back this month wasn’t pain; it was a fall at home, mild confusion her daughter noticed on a phone call, and a basic metabolic panel drawn at urgent care that came back with a sodium of 124, attributed to carbamazepine’s well-documented syndrome of inappropriate antidiuretic hormone secretion — a real class effect, not an idiosyncratic reaction unique to her.
Berghuis et al. (Epilepsia 2017), following 1,421 patients on one drug or the other, set the threshold for severe hyponatremia at a sodium of 128 or below. Hers is 124. She is not approaching that line, she is four points past it — inside the band where the comparative figures for the two drugs stop resembling each other. Seven percent of carbamazepine patients reached it; 22% of oxcarbazepine patients did. Dong et al. (Neurology 2005) found the same direction independently, 12.4% against 2.8%.
The instinct when a drug produces a tolerability problem is to slide sideways within its class toward the agent with the friendlier reputation, and oxcarbazepine has earned that reputation on cognitive and dermatologic grounds. On this one axis it is the worse drug, and the switch most likely to be proposed first is the switch most likely to put her back in an emergency department with a lower sodium than the one she came in with.
A fall, a sodium of 124, and two years of near-complete pain control
I'd switch her to oxcarbazepine. Same mechanism, same reason it works for her pain, and it's generally considered better tolerated overall — a reasonable lateral move within a class that's already proven itself for her.
I need to correct the premise before we go further — oxcarbazepine actually carries a higher documented incidence of clinically significant hyponatremia than carbamazepine, not a lower one. Berghuis found hyponatremia in 26% of carbamazepine patients against 46% on oxcarbazepine, and severe hyponatremia in 7% against 22%; Dong reported the same direction a decade earlier. Switching her for this specific complication doesn't solve the problem, it roughly triples her odds of ending up back here.
If we're moving off carbamazepine for the SIADH specifically, lamotrigine is the real second-line option that doesn't share that mechanism. It comes with a real, if modest, efficacy gap relative to either of the first two drugs, but it actually addresses what sent her here.
You're right that oxcarbazepine was the wrong instinct as stated — I'm not going to defend that. But I'd ask whether we need to switch class at all yet. Her pain control is essentially complete, and that's not guaranteed to survive a switch to anything, including lamotrigine.
Dose-reduce carbamazepine, start fluid restriction, and recheck her sodium in a week with an explicit threshold for what “not working” looks like. If conservative management can't keep her stable, lamotrigine is there, correctly reasoned this time. But I'd rather try the option that doesn't risk her pain control first.
One caveat on my own plan, since the next decision runs off that sodium: doing both at once means a normal result won't tell us which one produced it, and we'll be guessing later about whether her pain-controlling dose can safely come back up. I'd accept that here — at 124 with a fall already behind her I'm not willing to run them sequentially just to get a cleaner answer — but we should write down that the recheck answers whether she is safe, not what fixed her.
Agreed: carbamazepine dose reduced, fluid restriction initiated, sodium rechecked in one week with an explicit threshold — if it doesn't normalize, move to lamotrigine rather than oxcarbazepine.
The Pain Medicine Specialist is more willing to accept some real efficacy loss with lamotrigine for the sake of safety, given how severe her pretreatment pain was.
The Clinical Pharmacologist wanted to explore adding low-dose baclofen to carbamazepine first, rather than switching entirely, given a real alternative combination exists. Not agreed.