Introduction

Since the first recognition of bacteria meningitis (BM) in 1805₁, mortality from this infection has remained unacceptably high and developmental disabilities and neurologic sequelae following neonatal and infantile meningitis still occur in significant number of survivals. Specific aetiological diagnosis of BM in developing countries is often difficult. Although Gram staining of CSF sediment is a very useful cheap and fairly rapid method of identification of organism, the sensitivity in developing countries is only 25-40%₂ when compared to 80-85% in developed countries₃. Cultures of CSF are also infrequently performed in many health institutions in developing countries and sensitivity does not exceed 40% with results available only after 2-3days₂,₄. Other methods such as latex antigen tests (LAT)₅ and polymerase chain reaction (PCR)₆ that are highly sensitive and specific are expensive and not available for routine use in developing countries.

In Nigeria, epidemics of meningitis due to Neisseria meningitidis have been reported from the Northern region since the 1950s₇,₈,₉,₁₀,₁₁,₁₂, and in the South, epidemic and sporadic cases₁₃,₁₄,₁₅,₁₆,₁₇,₁₈ of BM due to Streptococcus pneumoniae, Haemophilus influenzae and enteric Gram negative bacilli have been reported by different tertiary health institutions. Several reviews₁₉,₂₀,₂₁ have suggested that the choice of antimicrobial agent for the initial empiric therapy of bacterial meningitis when there is no CSF Gram stain result or when CSF Gram stain is not diagnostic, should depend on the age, geographical location and immunological status of the patient.

In many tertiary health institutions in Nigeria, financial constraints have compelled many clinicians to employ empiric therapy for patients with meningitis without CSF analysis or culture. In this study, we assess the use of empiric therapy for suspected BM in our institution in order to determine its suitability in a resource poor country.

Materials And Method

Medical records of 172 patients clinically diagnosed and managed as BM in Ladoke Akintola University Teaching Hospital, Osogbo, between 2004 and 2005 were retrieved for analysis. Information retrieved included demographic data, height, weight, presenting complaints, results of physical examination, predisposing factors, complications, outcome and problems encountered during management. In those investigated, results of microbiological and other laboratory investigations were also retrieved.

Data entry and management were done with Microsoft Excel on IBM ThinkPad computer. All analyses and calculations were performed using GraphPad software (GraphPad Software Inc, San Diego, USA). Relationship between categorical variables was done using Chi square or Fisher's exact test and for continuous variables using Student's t-test or Mann Whitney test as appropriate and P < 0.05 was taken as significant value.

Result

Demographic characteristics

Of the 172 patients with suspected BM reviewed over the study period, 108 were males and 64 were females giving a male to female ratio of 1.7 to 1. The age of the patients ranged from 1 month to 80 years. The age group 16-50 years constituted the majority (36.2%) followed by age group above 50 years (18.6%), 6-15 years (17.4%), 1-5 years (12.8%), infants (8.1%) and neonates (8.1%) (Table 1).

Figure 1

Table 1: Age and sex distribution of patients with meningitis in LAUTECH Teaching Hospital (2004-2005)

Diagnosis

The common symptoms at presentation were fever in 75.6% of cases, impaired consciousness (50.0%), headache (32.6%), convulsion (27.9%), neck stiffness/pain (18.6%) and others (45.9%). Signs of meningeal irritation such as neck stiffness, Kernig's, Brudzinski and Babinski were found in only 62.8%, 43.0%, 33.7% and 6.2% respectively (Table 2).

Figure 2

Table 2: Clinical manifestations of meningitis in LAUTECH Teaching Hospital

Only 68 (42.0%) patients were investigated by CSF microscopy and culture. CSF cell morphology (protein > 45mg/dL, WBC > 50cells/ml and sugar < 2.2mmol/L) was suggestive of BM in 30 (41.1%) patients, 18 (26.5%) of these were positive for bacteria on Gram stain and 6 (8.8%) of these 18 had positive culture, Streptococcus pneumoniae in 5 and Escherichia coli in 1 (Table3).

Figure 3

Table 3: CSF parameters among patients with bacterial meningitis in LAUTECH Teaching Hospital (n = 68)

Outcome

Sixty two (38.3%) patients recovered completely before being discharged home, while 10 (8.1%) had varied neurological deficits (such as cranial nerve deficit, exaggerated deep tendon reflexes, quadriparesis and quadrilplegia) and 20 (10.5%) discharged self against medical advice (Table 4).

Figure 4

Table 4: Outcome of treatment of meningitis in LAUTECH Teaching Hospital

A total of 80 patients died giving a mortality rate of 45.5%. Mortality was significantly associated with severe dehydration (P=0.0319), those who did not have CSF microscopy or culture performed (P<0.0001) and those who did not receive antibiotic therapy (P=0.0008). (Table 4). The antibiotic combination that was significantly associated with favourable outcome was penicillin and chloramphenicol (P=0.021). Ampicillin/gentamicin (P=0.54), ampicillin/chloramphenicol/cephalosporins (P=0.41) and ampicillin/gentamicin/cephalosporin (P=0.46) combinations did not significantly influenced outcome. The presence of clinical symptoms such as convulsion, impaired consciousness and jaundice also did not significantly influenced outcome (P > 0.05). The greatest problem encountered during the course of patients care in this series was financial constraint as many patients were unable to pay for investigations and medications.

Figure 5

Table 5: Prognostic factors of meningitis in LAUTECH Teaching Hospital

Discussion

The result of this study showed that mortality from BM in Osogbo, Southwestern Nigeria is unacceptably high. The 45.6% mortality rate in this series is far higher than reports of previous studies in other parts of the country₁₃,₁₄,₁₅,₁₆,₁₇,₁₈ and elsewhere.₂₂ This underscores the fact that BM still remains a fatal disease with high mortality, and morbidity in those who recovers.₂₂,₂₃ The outcome of this infectious neurological disease has been shown to largely depend on age, type of organisms, time of initiation and appropriateness of antibiotic therapy, other intercurrent illnesses and immunological status of patient affected.₁₉,₂₀,₂₁ The high mortality in this report might not be unconnected with the fact that more than half (60%) of the patients were treated empirically without any recourse to CSF investigation. The likelihood of using inappropriate antimicrobial agents is high with empirical therapy.

Mortality from BM in our study was significantly higher in cases with dehydration and in cases where basic CSF investigations and antibiotic therapy could not be administered due to financial constraints. Especially in those who did not receive antibiotics, mortality was 100% indicating the grave consequence of BM in those who are not financially empowered in the society and also lend credence to the fact that BM is invariably fatal. Mortality was also high (60%) in cases of BM due to S. pneumoniae in this study and agrees with similar studies₁₃,₁₈in Southern Nigeria. Although BM due to Gram negative enteric bacilli, which is usually associated with high mortality, tends to be more prevalent in neonates₂₃, an inference cannot be made about this assertion in our study because of the small number of neonates involved.

Apart from the combination of crystalline penicillin and chloramphenicol therapy, which gave a statistically significant favorable outcome, other combination regimen such as ampicillin/gentamicin, ampicillin/chloramphenicol/cephalosporin and ampicillin/gentamicin/cephalosporin gave comparable but poorer outcome. Our observation of a favorable outcome with penicillin/chloramphenicol combination corroborates the findings of Ozumba₁₈ who used the same combination in Enugu, a centre in the same meningitis belt as our centre. However, some recent studies have shown an increase in the incidence and dissemination of strains of S. pneumoniae resistant to penicillin and cephalosporins₂₄, and strains of H. influenzae resistant to ampicillin and chloramphenicol₂₅, and the possibility of some of these resistant organisms being involved in BM in our series is very high. This may partly explain the high mortality recorded.

Aside from fever which was present in about 75% of patients, symptoms such as convulsion, altered consciousness, headache and neck pain/stiffness were found in less than 50% of our patients, and these symptoms were notably absent in the neonates. The classical signs of meningitis such as Kernig's and Brudzinski signs were elicited in less than 40% of patients, although when present were associated with poor prognosis. The implication is that clinical diagnosis alone is inadequate as a diagnostic modality for BM, especially in our environment where antibiotic abuse/misuse tends to modify the classical symptoms of BM and therefore creates diagnostic difficulties to unsuspecting physicians. Also, the presence of convulsion and impaired consciousness, which were associated with high mortality in some series₂₅, did not significantly influenced outcome in this study. Their presence should therefore not be seen as a discouraging sign but rather as a challenge to proper investigation and management.

The CSF parameters (sugar, protein and white cell counts) were suggestive of BM in 41% of those investigated by CSF analysis while the 26.5% sensitivity of CSF Gram stain in the study is comparable to what is normally obtained in developing countries₄ but culture is correspondingly less sensitive. Factors that may be responsible for this include frequent inappropriate use of antibiotics before presentation, delay in processing of CSF specimens and lack of laboratory facilities to isolate fastidious organisms. For mortality from BM to appreciably decrease, these factors must be addressed.

Conclusion

Our study shows that empiric antibiotic therapy of BM without CSF investigations is associated with extremely high case fatality. We suggest that in a resource poor country like Nigeria with widespread abject poverty, the initial management of BM especially the cost of performance of basic CSF investigation and antibiotic therapy should be borne by the government through the hospital management while the choice of empiric therapy should be guided by research findings.

Correspondence to

Dr. Taiwo Samuel S Department of Medical Microbiology, Ladoke Akintola University Teaching Hospital, PMB 5000, Osogbo, Nigeria E-mail: samtaiwo2003@yahoo.com Tel: 234 0 80 3343 6344