How Can We Help?
Meningitis
Meningitis.
The clinical performance of various aspects of the history and examination in suspected meningitis has been only poorly studied, and test performance is likely to be related to disease prevalence. For instance, where cryptococcal meningitis is common, lack of neck stiffness is a unhelpful for excluding meningitis, as it is only present in about 20% of immunosuppressed patients.
One review[3] found that only 50% of patients with meningitis had a headache, and only 30% had nausea or vomiting. However 99% had at least one of fever, neck stiffness or altered mental state, and all had jolt accentuation of headache. Rather depressingly, in the light of how industriously these are taught, Kernig’s sign had only been evaluated once since the original report, and Brudzinski’s not at all.
Interpretation of CSF results
A CSF result can only be described as normal if there is a normal protein and glucose with 5 or few cells (all lymphocytes) and a negative Indian Ink stain. In immunosuppressed patients the changes may be minor, but there will usually be at least one abnormal index. Gram stain and results of bacterial culture are important.
Perspective – dip sticks in the CSF
A study[4] on 234 children in Kuwait looked at the diagnostic value of popping a urine dipsticks into the CSF. Sensitivity and specificity for the diagnosis of meningitis were 97% and 100% respectively, and kappa for agreement with the laboratory findings was excellent at 0.97 (CI 0.94 – 0.99.) Cut-points used were: Bacterial if protein >0.5g/l and glucose <2.8 mmol/l, or if protein >0.5g/l, glucose >2.8 mmol/l, but cells > 500/microlitre. Viral if glucose >2.8 mmol/l, any protein, and cells 10-70/microlitre.
Perspective – Clinical contraindications to lumbar puncture
There are several clinical situations where doing a CT before lumbar puncture is clearly appropriate, however in other situations there may be unjustified caution which could have clinical consequences related to delaying lumbar puncture while trying to get the patient to a CT scanner. There is no single factor that adequately predicts when it is safe to do an LP straight away, but the list below has been proposed – the absence of any of the features (i.e. none of them present) had a negative predictive value of 97%[5] for the presence of CT abnormalities making LP likely to be dangerous.
If all of the items below are absent, an LP without CT is probably safe:
- Any focal neurology, including difficulty speaking or visual problems
- Age > 60 yeaars
- HIV positive
- History of previous CNS problems
- Fitsin the previous week
- Decreased consciousness
- Difficulty answering two questions one after another
- Difficulty in following two commands one after another
- (Papilloedema)
The difficulty with a study of this nature is extrapolating to areas with different disease prevalence. I n this study only 4% of patients had a focal lesion with mass effect.
Diagnostic performance of CSF examination in diagnosing bacterial meningitis
Most studies on this topic look at the differentiation of bacterial from viral meningitis[6]. A CSF to blood glucose ratio of less than 0.4 had an LR+ of 18, more than 500 WCCs/microlitre 15, and a CSF lactate of more than 3.5 mmol/l had a LR+ of 21 (95% CI 14 to 32.)
Having done the CT, when do the findings make lumbar puncture unwise[7]?
- If there is midline shift
- If there is a posterior fossa mass or bleed or obliteration of the 4th ventricle
- If there is obliteration of the basilar cisterns (the area around the optic chiasm, in the region of the pituitary, is called the suprachiasmatic cistern, and the area anterior to the midbrain is called the basilar cistern)
- Loss of cerebral sulci due to diffusely raised pressure is probably less of a risk, but should be considered a relative contra-indication in most circumstances.
Treatment of bacterial meningitis.
Bacterial meningitis is ideally managed in a setting of reliable culture and sensitivity results. Partial treatment prior to presentation and suboptimal microbiology services may necessitate a more empiric approach.
Initial management of patients with suspected meningitis should be with ceftriaxone 2 g 12 hourly IVI for ten days (half dose if GFR < 10 ml/min).
In adults, the organism is usually a pneumococcus or meningococcus. The prevalence of resistant pneumococci varies, and reflects levels of access to health care as well as the prescribing habits of local clinicians. Clinically important resistance is high level, and is less common than low level resistance.
Corticosteroids are not recommended in bacterial meningitis.
Tuberculous meningitis can present initially with a CSF neutrophil predominance.
A delay in initiating antibiotics is associated with worse outcome. If there is any suspicion that a patient has bacterial meningitis, either do a lumbar puncture and then start the antibiotic, or if an LP is not thought to be safe, do a blood culture, start the antibiotic, and only then organise the CT scan referral.
Clinical improvement may not always be dramatically rapid, but should be consistently better from day to day. A fluctuating course suggests intracranial pus, medication not being given regularly, or possibly resistance, although in the latter case it would be more likely that there would be progressive deterioration.
Perspectives – prognosis in bacterial meningitis
Meningococcal meningitis treated promptly with adequate doses of penicillin is a relatively benign illness, with reported mortality about 15%. When treated late (even a delay of only hours) prognosis is even worse. Pneumococcal meningitis has a fearsome reputation – In some studies in HIV positive persons, in-hospital mortality as high as 65% has been recorded[8].
Perspective – treatment of bacterial meningitis
A systematic review[9] comparing third generation cephalosporins with ‘conventional’ antibiotics (ampicillin-chloramphenicol) was unable to demonstrate a significant difference, although the authors make the cogent point that most of the studies date from the eighties or early nineties. At that time, penicillin resistance was lower. (ARR for death -1%, 95% CI –4% to +3%, ARR for treatment failure was also non-significant at –2%, 95% CI –5% to +2%)
The lack of a clear RCT evidence base for current recommendations on the use of third generation cephalosporins is of some concern but is unlikely to be remedied at this stage, particularly in the light of rising in vitro pneumococcal penicillin resistance.
Perspective – corticosteroids in acute bacterial meningitis
An early study on children in Malawi failed to demonstrate benefit[10], but was followed by work from the Netherlands[11] suggesting a mortality advantage in an environment where, due to good antibiotic policies, 100% of pneumococcal isolates were sensitive to penicillin.
Further work in the developing world has cast some doubt on this, with two studies failing to demonstrate benefit in somewhat different environments. In the Vietnamese study[12], there was no difference in mortality or rate of severe disability, although in the pre-specified subgroup with microbiologically proven bacterial meningitis there was a survival advantage. (This sub-analysis should be treated with reserve as there was heterogeneity in the results.)
The Malawian study[13] failed to demonstrate either a survival or a disability advantage in a patient group with a 90% HIV prevalence rate.
Mortality | Steroids | Placebo | ARR | NNT |
Netherlands | 0.07 | 0.15 | 0.08 (0.005 – 0.149) | 14 (7-185) |
Vietnam | 0.10 | 0.12 | 0.02 (-0.028 to +0.073) | 45 (-14 to inf to 36) |
Malawi | 0.52 | 0.55 | -0.04 (-0.11 to + 0.04) | Inf (-26 to inf to 9) |
Note the high mortality in both arms of the Malawi study. An individual patient meta-analysis[14] pooled the results from these three studies with two further studies in children (one from Malawi and one from South America) providing high quality double-blinded RCT data on 2029 patients. It confirmed the lack of any significant benefit in terms of either mortality or neurological disability, and that this result was consistent across sub-groups (i.e. was not influenced by variables such as HIV status or duration of symptoms before starting treatment.)
‘Door to antibiotic’ time.
A retrospective folder review of 123 patients from Canada[15] showed three factors strongly associated with mortality in acute bacterial meningitis: being afebrile on presentation (OR 39.4) severe mental obtundation (OR 12.6) and a door to antibiotic time of more than 6 hours (OR 8.4). The mortality rate of a door to antibiotic time of less than six hours was 5%, and it rose sharply to 45% at 6 to 8 hours, and to more than 75% after a delay of more than eight hours.
Tuberculous meningitis
This diagnosis can be missed easily, particularly in immunocompromised patients, and delays in starting treatment clearly influence prognosis. Suspect the condition in any patient with chronic worsening headache and be very careful with CSF interpretation. Early presentations may be with a polymorph predominance, which only becomes lymphocytic some days later. Suspect the diagnosis in any patient with low CSF glucose and even slightly raised protein, and do not be put off by rather unimpressive cell counts. TB PCR on CSF is worth doing, even though sensitivity is not that brilliant although the current next generation kit (GeneXpert Ultra) is reported to have sensitivity as high as 70% in HIV positive patients.1
Treatment is for nine months, and corticosteroids are recommended.
Bahr NC et al. Diagnostic accuracy of Xpert MTB/RIF Ultra for tuberculous meningitis in HIV-infected adults: a prospective cohort study. Lancet 2018. https://doi.org/10.1016/S1473-3099(17)30474-7 ↩