Abstract

Introduction
Pleural effusion is a common clinical presentation with a broad differential diagnosis. The aetiology of pleural effusions varies geographically, but most investigation guidelines are based upon factors pertinent to the industrialized societies where malignancies predominate. 1–3 However, in developing nations infections – especially TB and parapneumonic effusions (PPE) are more prevalent. 4–7 The purpose of this study was to determine the characteristics and causes of pleural effusions in adults in Kumasi, and to develop a diagnostic algorithm that precludes the need for investigations which may not be available.
Methods
This prospective study was done at the Komfo Anokye Teaching Hospital (KATH) Kumasi, Ghana. Adults (>15 years) admitted to the hospital with pleural effusion were considered eligible for the study. The investigations included chest X-rays, full blood count, HIV serology, sputum microscopy and culture (including TB), pleural aspiration for protein, lactate dehydrogenase (LDH), microscopy, differential cell count and culture (including TB) and cytology. Effusions were classified radiologically as large unilateral (L > 50% of hemithorax) small unilateral (S < 50% of hemithorax) and bilateral (B). They were also classified as exudates (pleural fluid protein >30 g/dL) or transudates. Pleural biopsy is not available at KATH; the diagnosis of TB was based upon clinical assessment and responses to trials of standard anti-TB chemotherapy.
Results
We assessed120 consecutive adults with pleural effusion. Nineteen were excluded due to: no evidence of informed consent (8), the patient being too unwell (9), pleural fluid had not obtained (2). Forty-seven men and 54 women were included in the analysis (Table 1). The men were younger than the women and more likely to have smoked, but none had been exposed to asbestos. HIV serology was positive in 38.6% with a slight female predominance, and there was a high prevalence of anaemia. The common symptoms were:
cough (90.1%), usually productive (82.4%); dyspnoea (89%); fever (72.3%); chest pain (70%); weight loss (60%). Patient characteristics Chi squared with Yates correction
a
P < 0.005
b
P < 0.001
cMale haemoglobin <12 g/dL and female haemoglobin <11 g/dL)
The characteristics of TB and non-TB effusions
Chi squared with Yates correction
a P < 0.001
b P < 0.005
c P < 0.05
Discussion
Studies of pleural effusion in adults in sub-Saharan Africa have established that the differential diagnosis is between TB and other causes. In previous studies TB was the cause of pleural effusion in 66–95%, with positive HIV serology in 42–72%. 4–7 Our findings were similar: we attributed 63.5% of pleural exudates to TB of which 55.5% were HIV positive. However, in this study the diagnosis of TB was based almost entirely on clinical judgement and the response to anti-tuberculous chemotherapy. Pleural biopsy is not available at KATH but pleural fluid is routinely aspirated for culture and biochemistry. However, of the investigations performed only the effusion size (on chest X-ray), HIV serology and pleural protein had any discriminating value. Most patients with transudates could have been identified clinically and would not have required pleural aspiration. Similarly, most PPE are associated with a clinical evidence of pneumonia; pleural aspiration is necessary only if effusions are large or if empyema is suspected. For patients with malignant effusions there is often other clinical evidence of malignancy.
We propose a simple algorithm that would ensure that patients are treated appropriately while minimizing the need for laboratory investigations (Figure 1). The most important laboratory test is HIV serology. HIV-positive patients with large effusions should be given a trial of anti-TB therapy without further investigation. Other HIV-positive patients should also be treated for TB if an alternative cause for the effusion is not apparent after minimal investigation. Only HIV-negative patients (with a pleural exudate) would require further diagnostic investigations. Had we followed this algorithm less than half of the patients would have required diagnostic pleural aspiration and few would have required further investigation. However, this algorithm requires further evaluation.
An algorithm for pleural effusions. Numbers in parentheses: number of cases with TB/total for each group. CCF, congestive cardiac failure; CRF, chronic renal failure
Conclusions and recommendations
Pleural effusions in Ghanaian adults are usually due to TB and are associated with HIV infection. Minimal laboratory investigations are required to establish a diagnosis.
