Answers

Short answers

1. In view of this patient’s background, low serum-ascites albumin gradient, low grade fever, negative cytology, and unhelpful cross sectional radiology the most likely diagnosis is tuberculous peritonitis. A normal chest radiograph does not exclude this diagnosis.

2. The diagnosis should be confirmed by laparoscopy and biopsy of omental tuberculous deposits in the peritoneal cavity.

3. Anti-tuberculous chemotherapy for six months minimum.

Long answers
1. Diagnosis of tuberculous peritonitis
One of the clues to the diagnosis in this case is the serum-ascites albumin gradient. This value is determined by subtracting the ascitic albumin from the serum albumin, using values both obtained on the same day. In this patient the value is 9 g/l. Serum-ascites albumin gradient values of >11 g/L are found in patients with cirrhosis and portal hypertension.¹ Similar values are seen in patients with ascites caused by heart failure. All the other causes of ascites given in the list below typically produce a serum-ascites albumin gradient of <11 g/l:

• Metastatic cancer
  
• Infective peritonitis:
  

  Tuberculous
  

  Pyogenic
  

  
  
• Nephrotic syndrome
  
• Pancreatic ascites
  

In view of this patient’s background, history, clinical findings, negative cytology, and unhelpful cross sectional radiology the most likely diagnosis is tuberculous peritonitis. A normal chest radiograph does not exclude the diagnosis as the radiograph is normal in more than 75% of cases of tuberculous peritonitis.²

2. Making the diagnosis of tuberculous peritonitis
Tuberculous peritonitis remains a major problem in the developing world, and because of its association with HIV co-infection, the disease is undergoing a resurgence in areas of the world that were thought to have low prevalence.³ In 2006, nearly 8500 cases of tuberculosis were reported in the United Kingdom. Of the 5000 cases in migrants, 2500 were extrapulmonary.⁴ Abdominal tuberculosis is commonly reported to be the sixth most common site of extrapulmonary tuberculosis. The disease is seen most commonly between the ages of 35-45 with an equal gender distribution. Even in areas where tuberculous peritonitis is endemic, diagnosis remains a challenge. The insidious onset, lack of specific clinical features, limited yield of most diagnostic tests, and difficulties with isolating mycobacteria from ascitic fluid mean that the most important diagnostic tool remains a high index of suspicion. In particular, the diagnosis should be entertained in those who have lived or travelled from an endemic area, those with HIV, and those with advanced chronic liver disease.

Mycobacterium tuberculosis enters the peritoneal cavity via a ruptured lymph node or haematogenous spread from a pulmonary focus.⁵Concomitant active pulmonary disease, however, is rare.⁶ Patients typically present with malaise, low grade fever (59%) and ascites (73%).⁷ The ascitic fluid is usually protein rich and contains numerous white cells (predominantly lymphocytes in 68%). Most patients have ascitic cell counts of between 500 and 1500 cells/ml.⁸ The fluid may be bloody. As mentioned above, the serum-ascites albumin gradient can be useful. Although its specificity is low, in a recent case series all 26 patients were found to have a low serum-ascites albumin gradient.⁹ Z-N staining of ascitic fluid is often negative (97% of cases with proven tuberculous peritonitis), and ascitic fluid culture can take up to six weeks to yield a result with a sensitivity of only 20%.⁶ The yield of culture positive results seems to be improved when 1000 ml of centrifuged ascites is cultured, rather than the more traditional 10-50 ml, with or without the BACTEC system for accelerating mycobacterial identification.¹⁰

In patients like the one in question where tuberculous peritonitis is suspected, it is inappropriate to await the culture results to establish the diagnosis. Such patients should have a laparoscopy, which is a safe diagnostic manoeuvre that will readily establish the diagnosis in most patients.¹¹ In 1992 Bhargava and colleagues¹² described the typical laparoscopic findings of tuberculous peritonitis, which are divided into three subtypes as follows:

• Thickened hyperaemic peritoneum with ascites and scattered (parietal peritoneal, omental, and bowel wall) whitish miliary nodules (66%)
  
• Thickened hyperaemic peritoneum with ascites and adhesions (21%)
  
• Fibroadhesive type with no ascites (13%).
  

A systematic review by Sanai and Bzeizi in 2005¹³ showed sensitivity rates of 93% and specificity of 98% (on cumulative data) when laparoscopic findings are combined with histological results. It is important that histological and microbiological samples are taken at laparoscopy to exclude peritoneal carcinomatosis, sarcoidosis, and starch peritonitis as all three may mimic tuberculous peritonitis, and to allow treatment to be guided by the antibacterial specificities of the organisms identified.

3. Treatment of tuberculous peritonitis
The National Institute for Health and Clinical Excellence (NICE)¹⁴ has published guidance on treatment of peritoneal tuberculosis, suggesting first line treatment should be six months of standard anti-tuberculous chemotherapy. This standard treatment consists of two months of rifampicin, isoniazid, pyrazinamide, and ethambutol followed by a further four months of rifampicin and isoniazid. Although it is unnecessary to carry out contact tracing in tuberculous peritonitis unless patients have coexisting pulmonary disease, it remains mandatory to report confirmed infection.

Cite this as: BMJ 2008;337:a1252