📝 Tegretol, Epitec, and Stevens-Johnson Syndrome — A Watchful-Eye Briefing

A note before the medicine

Before any of the clinical content: nothing in what was described sounds like Stevens-Johnson syndrome was actually happening. Flu-like symptoms come up constantly in healthy children and adults; they overlap with the very early hours of an SJS prodrome but they overlap with viral illness, post-vaccination malaise, sinusitis, and a dozen other ordinary things much more often. The fact that the question was asked is the safety system working — the watchful eye that catches the rare cases is the same watchful eye that asks "should I be worried?" about the much more common ones.

Update at the time of writing: Tegretol has already been stopped, and there's a follow-up appointment with the doctor in two days. That's exactly the right call. Removing the suspect drug is the single most important intervention in any possible drug reaction — it costs almost nothing if the symptoms turn out to be unrelated, and it dramatically reduces the risk if they were related. The two-day window before seeing the doctor is also useful diagnostically: if the symptoms continue to settle off the drug, that's reassuring and corroborative; if anything new appears (mouth ulcers, eye involvement, blistering, a spreading rash), there's a clear escalation path to an emergency room rather than waiting for the appointment.

What this briefing is for: making sure that the next appointment is a confident, informed conversation rather than a guessing game — and that you have the framework to answer "should I be worried?" quickly the next time it comes up.

1. What Stevens-Johnson syndrome actually is

SJS is a rare, severe, immune-mediated reaction — usually to a medication — in which the immune system mistakes the body's own skin and mucous membrane cells for something that needs destroying. CD8+ cytotoxic T-cells, helped by a protein called granulysin, attack keratinocytes (skin cells), causing them to die in sheets. The result resembles a severe internal burn.

It sits on a spectrum with toxic epidermal necrolysis (TEN), which is the same disease at higher severity:

Condition	Body surface area affected	Mortality
SJS	Less than 10%	~10%
SJS/TEN overlap	10–30%	~25%
TEN	More than 30%	~30–50%

It's rare — about 2–7 cases per million people per year. SJS is roughly 3× more common than TEN.

Medications cause more than 80% of cases. The biggest offenders, in rough order of risk:

1. Allopurinol (gout)
2. Sulfa antibiotics (cotrimoxazole / Bactrim)
3. Carbamazepine (Tegretol)
4. Lamotrigine (Epitec / Lamictin)
5. Phenytoin
6. Oxcarbazepine
7. Phenobarbital
8. Nevirapine (HIV drug)
9. NSAIDs (especially the oxicam class)

Two of the top four are in this household. That's why this question is worth asking carefully — but it's also why these specific drugs are titrated and monitored the way they are, in everyone, all the time.

2. The progression — what it looks like vs. what it doesn't

What separates an SJS prodrome from a viral illness isn't the early hours — it's what happens over the next 1–4 days.

The prodrome (1–3 days before the rash):

• Fever, often above 39 °C
• Sore throat, cough, runny nose
• Burning, sore, or red eyes — this is more telling than the others
• Muscle aches, fatigue
• Skin pain — and this is the key signal — pain that is disproportionate to anything you can see

This stage genuinely does look like a viral illness, which is why early misdiagnosis happens. The thing that pivots it is the next stage.

The rash stage (days 3–4):

• Red, purple, or dusky-grey flat spots that start on the face, upper chest, palms and soles
• "Atypical target" lesions — flatter, less concentric than the classic erythema multiforme target
• The rash spreads centrifugally
• Mucosal involvement is the hallmark — mouth ulcers, hemorrhagic crusting on the lips, eye redness with discharge, painful urination, painful swallowing. Mucosal involvement appears in more than 90% of SJS cases.

The blistering stage:

• Flaccid blisters form on the red patches, then rupture
• Sheets of skin peel away — looks like a burn
• Nikolsky's sign positive — gentle lateral pressure makes the top layer of skin slide off

A clean rule of thumb for parents: fever and sore throat alone is virtually never SJS. Fever + sore throat + mouth ulcers + sore eyes + skin pain + a spreading rash on the face/chest is what flips the picture. The combination, not any single sign.

3. Why Tegretol and Epitec specifically

Both of these drugs sit in a family called aromatic anticonvulsants — drugs whose chemical structure includes an aromatic ring. The body metabolises them through cytochrome P-450 enzymes into intermediate compounds called arene oxides, and in some people those intermediates trigger a T-cell-mediated immune response.

The clinical reality:

• Carbamazepine (Tegretol): ~1–6 cases of SJS/TEN per 10,000 new users in non-Asian populations. In Asian populations carrying HLA-B*15:02, that risk can be 10× higher — which is why Taiwan, Singapore, Hong Kong, and the FDA recommend pre-prescription HLA testing for patients of Asian ancestry.
• Lamotrigine (Epitec): ~44 cases per 100,000 new users in published cohorts. The risk is concentrated in the first 8 weeks and is sharply higher with rapid titration or with concurrent valproate.

Both drugs are still widely prescribed, including to children, adolescents, and adults, because for most people the benefits massively outweigh the risk. The risk is not random — it's predictable in shape (first 8 weeks, dose-dependent, ancestry-modulated) — and the prescribing protocols are designed around that predictability.

4. Cross-reactivity between Tegretol and Epitec

This is the subtle bit that doesn't always come up in a routine appointment.

Even though carbamazepine and lamotrigine are chemically different, they share enough structural features (and metabolic pathways) that T-cells primed by one of them can cross-recognise the other. Cellular studies have shown carbamazepine-induced T-cells reacting to lamotrigine.

What this means practically:

• If someone has had any severe cutaneous adverse reaction (SCAR) to one drug in the aromatic anticonvulsant family, they should avoid all of them lifelong. That includes carbamazepine, oxcarbazepine, phenytoin, phenobarbital, primidone, and lamotrigine. Cross-reactivity rates across the class are reported as high as 75%.
• Switching from Tegretol to Epitec after a rash from Tegretol is one of the riskier moves in this drug class. It's not absolutely contraindicated in every guideline, but most modern dermatology and epilepsy specialists would jump straight to a non-aromatic alternative.
• A documented SCAR to one of these drugs is something the family should keep in writing. A medical alert card or note in the wallet/phone listing the trigger drug AND the whole aromatic anticonvulsant class is the standard way to prevent accidental re-exposure in an emergency.

If neither drug has caused a reaction so far, the cross-reactivity issue is theoretical and the family is well into the safer post-8-week window — this is a "useful to know" item, not an active concern.

5. The South African / Namibian prescribing context

A few things worth understanding about how this is handled locally:

HLA-B*15:02 screening is not routine in SA, Namibia, or Botswana. The reasoning isn't sloppy — the specific allele that drives carbamazepine-SJS is largely an Asian allele. Most Black South Africans don't carry it. But the picture is more nuanced for some ancestries:

Ancestry	HLA-B*15:02 prevalence	Screening worth requesting?
Indigenous Black African	Very low	Probably not
White European	Very low	No
Indian / South Asian	~2–4%	Yes — worth requesting
Cape Malay	~2–6%	Yes — worth requesting
Chinese / Vietnamese	~10–15%	Yes — strongly recommended
Mixed-ancestry with any Asian or Cape Malay heritage	Variable	Worth asking

The test costs roughly R1,500–R3,000 in SA and is available through Lancet, Ampath, and PathCare. In Namibia, private referrals to SA labs are the typical route.

The SA system relies on clinical vigilance during the first 8 weeks rather than upfront genetic testing. The titration protocols, the "stop if any rash" instruction, and the parent-as-watchful-eye safety net are how the system catches the rare reactions. That system is what was working in this case.

The 2025 SA epilepsy guidelines update is also worth knowing about. As of June 2025, lamotrigine is now first-line for epilepsy in SA across all ages — the change was made on efficacy grounds, not safety grounds, but it does mean that if Tegretol ever needs to be reconsidered, the prescriber's default alternative for epilepsy isn't another aromatic AED.

6. The lamotrigine titration — why slow really matters

This is the single biggest safety lever for Epitec, and the protocol is specific.

Standard titration (no carbamazepine, no valproate):

• Weeks 1–2: 25 mg once daily
• Weeks 3–4: 50 mg once daily
• Week 5: 100 mg daily
• Week 6: 200 mg daily (typical maintenance)

With valproate on board (slower, because valproate raises lamotrigine levels):

• Weeks 1–2: 25 mg every other day
• Weeks 3–4: 25 mg daily
• Week 5: 50 mg daily, escalating slowly thereafter

With carbamazepine on board (faster, because carbamazepine lowers lamotrigine levels):

• Can reach maintenance more quickly, but the SJS risk profile changes. Specialists balance the trade-off case by case.

If the prescriber's titration matches one of these patterns, that's reassurance. If the dose is being escalated faster than this without a clear reason — that's a fair question to ask.

The Fujii 2020 study compared slow titration (n=195) to standard titration (n=84) head-to-head and found rash rates of 7.7% vs 27.4%. Slow titration is not a polite suggestion — it's a 3.5× difference in skin-reaction risk.

The "stop at any rash" rule applies for the first 8 weeks of lamotrigine. Any rash, even one that looks benign, is a stop-and-call moment. Most of those rashes turn out to be benign and the drug can be cautiously restarted later if the prescriber agrees. But during the first 8 weeks, the cost of pausing and being wrong is small; the cost of continuing through a real reaction is enormous. The asymmetry justifies the rule.

7. Alternatives if a switch ever becomes necessary

This is the section that exists so the family knows there are good options if needed — not because a switch is currently on the table.

For epilepsy:

Drug	Family	SJS risk	SA availability
Levetiracetam (Keppra, Levetin)	Non-aromatic, broad-spectrum	Very low	On EML, public-sector primary care level
Sodium valproate (Epilim)	Non-aromatic	Very low	On EML, widely available — but avoid in females of childbearing potential due to teratogenicity
Topiramate (Topamax)	Non-aromatic	Low	On EML for treatment-resistant epilepsy
Gabapentin / Pregabalin (Lyrica)	Non-aromatic	Very low	Available privately, mostly used for focal epilepsy and neuropathic pain
Lacosamide (Vimpat)	Non-aromatic	Very low	Private sector, expensive

The standard "we need to leave the aromatic class" answer for epilepsy in SA is levetiracetam. It's effective, on the EML, well-tolerated, and has one of the lowest SCAR signals of any AED. Main side effect is irritability/mood change in some patients.

For bipolar disorder:

Drug	Phase	SJS risk	Notes
Lithium	All phases (gold standard)	None	Requires blood-level monitoring; uniquely reduces suicide risk
Sodium valproate (Epilim)	Mania, maintenance	Very low	Avoid in females of childbearing potential
Quetiapine (Seroquel)	All phases	None	Atypical antipsychotic, generic available cheaply
Lurasidone (Latuda)	Bipolar depression	None	Best efficacy/tolerability balance for the depressive phase
Olanzapine, aripiprazole, cariprazine	Mania, maintenance	None	Atypical antipsychotics, varying SA availability

The standard "we need to leave the aromatic class" answer for bipolar in SA is lithium and/or quetiapine, depending on which phase needs covering. Lurasidone is the strongest add-on for the depressive pole if budget allows.

8. The four questions to ask the prescriber

When you're sitting in front of the doctor next, these are the questions that will both reassure you and surface anything that might warrant attention:

1. "What does the titration plan look like over the next 8 weeks, and what would warrant slowing it down or stopping?" Listen for whether the timing matches the standard or adjusted protocols above. Listen for whether there's a clear "stop drug, call us" instruction tied to specific signs.

2. "Given our family's ancestry, would HLA-B*15:02 testing be appropriate?" This is most relevant if anyone in the family has Indian, Chinese, Cape Malay, Filipino, or Thai heritage. The test is available in SA privately. The doctor may or may not have raised it depending on what they know about your background.

3. "What specific signs would justify a same-day call rather than waiting for the next scheduled appointment?" A good answer will mention any rash, any mouth/eye/genital ulceration, any blistering, any disproportionate skin pain, and any combination of fever + skin or mucosal signs in the first 8 weeks. If the doctor's answer is vaguer than that, you can fairly ask for specifics.

4. "If a reaction ever happened, what would the family carry as a written record of which drugs to avoid?" This is a forward-looking question that frames you as the engaged parent the doctor wants you to be. The right answer is a list including the trigger drug AND the whole aromatic anticonvulsant family — and ideally a medical alert tag or wallet card for the affected person.

9. The honest summary

These are powerful drugs that work well for most people who take them. They also carry one specific, well-characterised, predictable, monitorable risk that medical practice has built guardrails around: titration protocols, the 8-week vigilance window, the "stop at any rash" rule, regional HLA screening where it makes sense, and parents who notice when something is off.

The episode that triggered this briefing is the system working. A flu-like illness in someone on these medications got noticed, got questioned, didn't progress, and turned into an opportunity to lock in the framework for next time. That's the right loop.

If a future episode looks different — mucosal involvement, disproportionate skin pain, blistering, a rash that doesn't fit a viral pattern, especially in those first 8 weeks — the framework above tells you exactly what to do: stop the drug, call the prescriber, and head to a hospital with a dermatology or burn unit if the pattern matches. Speed matters in SJS not because every hour counts, but because every day of continued exposure to the trigger drug raises the risk.

You're not the parent who missed something. You're the parent who asked. Keep doing that.