Early events in spontaneous bacterial peritonitis

The database regarding spontaneous bacterial peritonitis (SBP) has increased dramatically in the past 33 years since this phrase was first coined by Harold Conn.¹ In the remote past this easily treatable cause of deterioration of patients with advanced cirrhosis was underdiagnosed and undertreated. This undoubtedly led to many unnecessary deaths, which were probably viewed as mysterious at the time.

Now we know that SBP is quite common, with a prevalence of >20% on admission to the hospital, prior to the era of prevention.² We know who is at high risk—patients with cirrhosis and (a) prior SBP, (b) low protein ascites, or (c) gastrointestinal bleeding.³ We also know how to diagnose, treat, and even prevent this potentially fatal infection.^3,4

Even the “spontaneous” nature of this infection has been largely resolved in recent years. We now know that the gut is the source of most of the bacteria that eventually cause SBP.⁵ As cirrhosis develops in animals, gram negative bacteria increase in numbers in the gut.⁵ We know that the gut of animals and patients with advanced cirrhosis is more permeable to bacteria than the normal gut and more permeable than the gut in less advanced cirrhosis.^6,7 Once bacteria reach a critical concentration in the gut lumen, they “spill over”, and escape the gut, “translocating” to mesenteric lymph nodes. Then they can enter lymph, blood, and eventually ascitic fluid.⁶ If the ability of the ascitic fluid to assist macrophages and neutrophils in killing the errant bacteria is deficient, uncontrolled growth occurs.⁸ This is SBP. In general, the animal or patient dies if they develop this infection and it is not promptly diagnosed and treated.

Thus SBP is the result of failure of the gut to contain bacteria and failure of the immune system to kill the virulent bacteria once they have escaped the gut. Patients and animals have duplicative mechanisms of protection from bacteria. This makes great sense teleologically. Opsonins assist motile and fixed “professional” killers of bacteria, the neutrophils and Kupffer cells, respectively. Innate defenders against bacterial invasion include macrophages, dendritic cells, and natural killer cells. These cells synthesise proinflammatory cytokines and effector molecules which assist in killing bacteria.

Unfortunately, patients with advanced cirrhosis have been reported to have defects and dysfunction in many of these systems of protection.^8–11 It is then no surprise that these patients are vulnerable to infection by their own gut flora. To make matters worse, some of the effector molecules and cytokines that help kill the bacteria have undesired side effects. Nitric oxide (NO) is one of these effector molecules. Tumour necrosis factor (TNF) is one of the relevant cytokines. NO is probably the long sought after agent responsible for vasodilation that characterises advanced liver disease.¹² Bacterial infection leads to further elevations in these molecules.^13,14 NO and TNF are important mediators of the further vasodilation and renal failure that too often accompany SBP.^14–16

The good news here is that selective intestinal decontamination with poorly absorbed antibiotics reduces gut bacterial counts, reduces translocation rates, can prevent SBP in high risk subgroups, and can improve the hyperdynamic circulatory state of these patients.^3,17,18 Selective intestinal decontamination can even improve survival of rats with cirrhosis and ascites.¹⁹

This brings us to the most recent contribution to this line of investigation published by Francés and colleagues²⁰ in this issue of Gut[see page 860]. These investigators have previously shown that some patients with cirrhosis have bacterial DNA in their serum and ascitic fluid, and that the DNA is always present simultaneously in both body fluids.²¹ This provides molecular evidence of bacterial translocation. They have also shown that patients who subsequently develop SBP have a higher baseline ascitic fluid TNF level than patients who do not develop SBP.²² It is probable that in patients and in rats with cirrhosis that SBP is preceded by episodes of colonisation of blood and ascitic fluid with viable bacteria or translocated pieces of bacteria (for example, DNA).^23,24 Bacterial DNA can bind to the toll-like receptor 9 of cells of the innate immune system and activate them.²⁵ Host immune defences are able to kill the bacteria in these episodes of colonisation that do not progress to SBP. SBP occurs (a) when the organism is more virulent than the bacteria that were killed by host immune defences at the stage of colonisation, (b) when immune defences weaken, or (c) a combination of these events. It is the peritoneal macrophage that is the first line of defence against bacterial colonisation of ascitic fluid.²⁶ SBP occurs when macrophages fail to kill the bacteria and the second line of defence is called in, the neutrophils.

Based on the information presented above, one would predict that the presence of whole bacteria or DNA in serum and ascitic fluid would have consequences (for example, stimulation of immune defences, effector molecules, and cytokines). This could in turn impact on haemodynamics, renal function, and survival. These effector molecules and cytokines are two edged swords. They can protect from bacterial infection but they can also initiate a sequence of events than can lead to the patient’s death.

The elegant study in this issue furthers this line of investigation and provides the scientific rationale for new clinical studies, including a randomised controlled trial.²⁰ The authors harvested peritoneal macrophages from patients with cirrhosis and ascites. Approximately one third of their patients had detectable bacterial DNA. They divided the patients into two groups: those with and without bacterial DNA in serum and ascitic fluid. They measured macrophage production of NO metabolites and cytokines, including TNF, and compared the results between the two groups. The authors convincingly demonstrate that peritoneal macrophages from patients with cirrhosis and bacterial DNA in serum and ascitic fluid are markedly activated, as evidenced by increased NO synthesising ability and enhanced cytokine production.²⁰ This study provides further insight into very early events in the pathogenesis of SBP. Pieces of bacteria commonly escape the gut and end up in blood and ascitic fluid. In the process, a complex sequence of events occurs. The immune system is stimulated to contain the bacterial colonisation and protect the host from fatal infection. However, as a consequence, the effector molecules and cytokines are increased, setting the stage for worsening of the haemodynamic status, development of functional renal failure, and the possibility of death.

Now that this new subset of patients with molecular evidence of translocation has been identified, it is perhaps time to perform a prospective study following those patients who are positive for bacterial DNA and determining if their risks of SBP, hepatorenal syndrome, and death are higher than those of DNA negative controls. If excessive morbidity and/or mortality are documented, the next step would be to conduct a randomised controlled trial of selective intestinal decontamination versus placebo in the DNA positive group and determine if hepatorenal syndrome and death can be prevented.