HIV-associated tuberculous meningitis with hyponatraemia and a treatment-timing dilemma
A 32-year-old man, an agricultural labourer, presents with 3 weeks of worsening headache, low-grade fever and night sweats, and 3 days of confusion. He was diagnosed with HIV 6 weeks ago and has not started antiretroviral therapy; CD4 count 48 cells/µL. He is drowsy, GCS 13, with neck stiffness and a right sixth-nerve palsy. CSF: 190 white cells/µL (88% lymphocytes), protein 2.6 g/L, glucose 1.3 mmol/L (serum 6.2 mmol/L), opening pressure 28 cm H₂O. CSF cryptococcal antigen and India ink are negative; CSF Xpert MTB/RIF is positive and rifampicin-susceptible. MRI shows basal meningeal enhancement, communicating hydrocephalus and two ring-enhancing tuberculomas. Serum sodium is 119 mmol/L, urine osmolality 520 mOsm/kg, and he is clinically euvolaemic. He was started on isoniazid, rifampicin, pyrazinamide and ethambutol 2 days ago and is receiving 0.9% saline at 80 mL/h.
1. Would you give adjunctive corticosteroids? Justify the dose and duration, list the caveats in an HIV-positive patient, and explain what you would want excluded first.
Read the worked answer
Yes - give dexamethasone, but with clear eyes open about the evidence.
Reasoning:
- Adjunctive steroids reduce death and disability in tuberculous meningitis, particularly in patients with reduced consciousness (stage II–III disease); he is GCS 13 with hydrocephalus and tuberculomas, which is the group that benefits most.
- The benefit in HIV co-infection is less certain - subgroup analyses suggest attenuation - but the standard of care in most guidelines is still to give steroids to HIV-positive patients with TBM, especially with impaired consciousness.
Regimen: dexamethasone 0.3–0.4 mg/kg/day intravenously for one week, then taper over 4–8 weeks (e.g. 0.3, 0.2, 0.1 mg/kg each for a week). Do not stop abruptly.
What must be excluded first:
- Cryptococcal meningitis - steroids in cryptococcal disease increase mortality. He has a negative CSF cryptococcal antigen and India ink, which is reassuring; a serum CrAg adds little when CSF is negative but is cheap.
- Other immunosuppression-related infections that steroids might worsen.
Caveats in HIV: hyperglycaemia, GI bleeding (add a proton-pump inhibitor), avitaminosis, unmasking of CMV or Pneumocystis, and a possible increase in HIV viral replication (not a reason to withhold). Steroids also treat paradoxical IRIS, so anticipate that some worsening later may need more, not less, steroid. Review the sodium and potassium daily.
2. Explain the likely mechanism of the hyponatraemia and how you would correct it safely. What would you change about his current fluid prescription?
Read the worked answer
Mechanism. Sodium 119 mmol/L, euvolaemia, urine osmolality 520 mOsm/kg (inappropriately concentrated relative to a low serum osmolality) points to SIADH, common in TBM from meningeal inflammation, raised intracranial pressure and drug effects. The alternative is cerebral salt wasting, which is also common in TBM but produces hypovolaemia, high urine sodium and polyuria. The two are distinguished clinically and by a careful fluid challenge: in SIADH, isotonic saline tends to worsen hyponatraemia; in cerebral salt wasting, it corrects it. His euvolaemic status favours SIADH, but reassess volume status and urine output repeatedly.
What I would change: running 0.9% saline at 80 mL/h in probable SIADH may be counterproductive - restrict total intake to about 800–1000 mL/day, stop hypotonic fluids, and review all drugs. Check serum cortisol (TB adrenalitis) and TSH before accepting SIADH.
Correction:
- Correct no faster than 8–10 mmol/L in 24 hours to avoid osmotic demyelination; check sodium 4–6 hourly initially.
- If he seizes or becomes comatose, give a 100–150 mL bolus of 3% saline and aim for a 4–6 mmol/L rise, then stop.
- If cerebral salt wasting is confirmed, use isotonic saline and consider fludrocortisone.
- Look for a treatable cause: inadequate drainage of hydrocephalus can drive both SIADH and deterioration.
3. When would you start antiretroviral therapy, with which regimen, and how would you recognise and manage immune reconstitution inflammatory syndrome?
Read the worked answer
Timing. Start ART early but not simultaneously - after ATT has been established, generally within 2–8 weeks, commonly at 2–4 weeks. A randomised trial in TB meningitis found no mortality advantage for starting at 2 weeks versus 8 weeks, with more adverse events in the early arm, so do not rush; but do not drift beyond 8 weeks either, because AIDS-related mortality rises. This contrasts with cryptococcal meningitis, where ART is deliberately deferred for about 4–6 weeks.
Regimen. Tenofovir/lamivudine plus dolutegravir 50 mg twice daily, because rifampicin halves dolutegravir exposure; or efavirenz 600 mg with rifampicin (no dose adjustment). Avoid nevirapine and ritonavir-boosted protease inhibitors with rifampicin. Add cotrimoxazole prophylaxis while CD4 <200 cells/µL, checking for interactions and adherence support.
IRIS. Paradoxical IRIS typically appears 2–8 weeks after ART: recurrent headache, fever, worsening meningism, enlarging or new tuberculomas, and worsening hydrocephalus, with a rising CD4 and falling viral load. Management:
- Exclude treatment failure, drug resistance and a second infection (cryptococcus, bacterial meningitis) by repeating CSF and imaging.
- Continue ATT and ART; do not stop either.
- Give corticosteroids (e.g. prednisolone 1 mg/kg/day tapering) for significant paradoxical reactions.
- Escalate neurosurgery for hydrocephalus; thalidomide or infliximab for steroid-refractory disease.
- Educate the patient and record the plan: he will need prolonged follow-up and repeat imaging.