Wednesday, October 24, 2007

Cryptococcal Meningitis

A woman had been on the wards for over a week- she was known to have AIDS and cryptococcal meningitis, and had self-discontinued her outpatient treatment. She was anywhere from 30 to 33 weeks pregnant, based on a very vague last menstrual period. She had an 8-year-old child, and her husband had already died of AIDS.

Crytpococcal meningitis is largely an illness of HIV/AIDS patients. Patients usually present with a headache of mental status changes, and the diagnosis is made through a lumbar puncture (LP) and staining of the cerebrospinal fluid with India Ink. Personally, I have seen it once - in medical school (Downstate, where else?). The Internal Medicine service here in Uganda sees it frequently - at least 1-2 per week. But the Obs/Gyn department rarely ever sees it.

This patient had been diagnosed as an outpatient by LP, and had been treated with fluconazole. She had stopped taking it - it is not clear why, although probably because she couldn't afford it. By the time she was admitted, she had not taken it for 2 weeks. Therefore, her risk of resistance was high.

The true treatment for Crytpococcal meningitis is Amphotericin (or, as it is often called, Amphoterrible, because of its awful side effects. Doctors shudder when discussing it.). It is contraindicated in pregnancy (as far as I know), and anyway it's not an option here, because it is too expensive. Instead, the hospital here is trying a regimen of high-dose fluconazole. Whereas the dose this patient was taking was 400mg once a day, the internal medicine service is treating patients with up to 1200mg per day. We don't know the effects of this in pregnancy.

On Friday, I rounded with Dr. Wasswa. I have come to appreciate his insights, as he has seen a lot and is quite detail-oriented. We reviewed the patient’s progress.
On admission, the patent had undergone another LP, which revealed elevated intracranial pressure and again a positive India Ink stain and CSF culture for cryptococcus. She was now taking 400mg of fluconazole daily by mouth, which her mother would crush up and dissolve into liquid for her. The patient was doing very poorly. She was conscious, but just barely, and was worsening daily. Her mother was her attendant, and we could see how caring and attentive she was. The patient said very little, except at one point she shouted "I can't see!" in Runyankole. The mother told us that her vision and hearing were going.

Dr. Wasswa decided to call the medicine consult again. Perhaps the patient needed another LP, or a different medication, or steroids? Neither of us was very familiar with the management of this complicated illness. We discussed whether or not to deliver her. We weren't sure of the gestational age. Was the pregnancy making her condition worse? No. Would she be helped by delivery? No, not like a pre-eclamptic would. Would delivery make her worse? Possibly. Probably. But it would also allow us to give aggressive medication without worrying about the fetus.

We were not sure of her gestational age. The absolute minimal gestational age for survival here is 28 weeks, and even then they rarely survive. Her LMP was very uncertain; we were estimating 30-33 weeks, but what if it was less? What if she had irregular menses from chronic disease? In the US, we would immediately do a bedside ultrasound to get a sense of the fetal measurements and weight. Here, there is one ultrasound (in fact, it’s the imaging modality. There is no X-ray, CT scan or MRI) and it is a tiny screen (barely bigger than an iPod) and can’t take any measurements. There is a better ultrasound machine in the MSF (Medecins Sans Frontiers) research center, and I could do fetal measurements there, but the patient would have to be moved way over to the center, and she was not stable or cooperative enough.

What would be the best mode of delivery? Vaginal delivery would be best for her, but an induction would be likely to fail at her gestational age, and a cesarean after failed induction has increased risks of infection, blood loss and essentially every complication. Plus, her HIV was probably advanced, and her viral load high (although she was on anti-retrovirals, Cryptococcal meningitis usually only happens when the immune system is severely compromised. In Mbarara, we can only check CD4 count, not viral load, but we didn't even have that for her.) . So then her likelihood of transmission to the fetus would be high in a vaginal delivery. So perhaps we should just do a cesarean section. But in that case, we would be performing an elective abdominal surgery on a severely ill woman - it could easily kill her.

With all these uncertainties, we decided it would be safer not to deliver her yet. We would call the medicine service and see what their recommendations were. If they felt we should deliver her and somehow we could get her Amphotericin, then that might be an indication.

By the afternoon, no one had called medicine. I ran into Jane, a British internist working for a year as a consultant in Mbarara, and told her about the patient’s worsening status. Immediately concerned, she went straight to the ward with me to evaluate her.

With her hearing and vision loss, the patient probably had high intracranial pressure. She needed an emergent LP, and probably daily LPs until her pressures improved. She should be transferred to the cryptococcal service on the medicine ward, where the fluconazole is provided free of charge (the patient’s mother had been purchasing it herself, and it was difficult for them), and would get better monitoring. We spoke to Tony, a senior consultant and an expert in HIV neurology. He agreed with the plan. We discussed whether to increase the dose of fluconazole to 1200mg. Fluconazole is avoided in pregnancy, mostly because its effects are not well known, and 1200mg is a very high dose. On the other hand, this woman was going to die without it. Even with it, Tony estimated her chances of survival at about 50%. She was now in the third trimester, so anomalies are not a concern, and regardless: no mother, no baby. We agreed to increase the dose.

There were no beds on the medicine ward, and the patient would have had to sleep on the floor, so we left her on the antepartum ward, but Jane said they would round on her in the morning. The next morning, the patient had still not gotten the higher dose of fluconazole; Jane gave it herself.

Later that evening, she developed intractable seizures (here, they call them “fits,” or they say “the patient was fitting”) and went into cardiopulmonary arrest. Jane was there. They tried to rescusitate, but couldn’t, and postmortem cesareans are rare here. Mother and fetus died.

No comments: