Antibiotics in Necrotising Pancreatitis – Historical Developments
The idea of preventing pancreatic sepsis through application of systemic antibiotics is rather old. In the 1970s, a number of uncontrolled clinical series failed to show any benefit when antibiotics were given to patients suffering from acute pancreatitis.^13–15 These studies suffered from some methodological drawbacks as the enrolled patients had mild pancreatitis with only little risk for developing bacterial infections, and the antibiotics used in these studies were not capable of penetrating the human pancreas.

The detection of bacterial infection of pancreatic necrosis as a major determinant for the prognosis of the patients in the mid 1980s re-attracted attention to the problem of antibiotic prophylaxis.7 As a consequence, considerable efforts were made during the early 1990s to identify antibiotics that were – on a theoretical basis – best suited as prophylactic antibiotic drugs in patients with necrotising pancreatitis. For this purpose, the antibiotics should be active against the bacterial spectrum of pancreatic infection that comprises Gram-negative bacteria, anaerobes and Gram-positive bacteria.^{7,11,12,16,17} Taking into account both the selective uptake of antibiotic drugs into the human pancreas and the bacterial spectrum, pharmacokinetic studies have shown that imipenem, quinolones or thirdgeneration cephalosporins should be the prophylactic drugs of choice.^18–20

These pharmacokinetic data have been supported by experimental studies that showed that systemic ciprofloxacin and systemic imipenem were capable of reducing both the incidence of septic complications and mortality in a rat model of severe acute pancreatitis if these drugs were given over a period of seven days, compared with an untreated control group.²¹ The first results of clinical studies that met the requirements necessary for reliable conclusions were seen in 1993 (see Table 1). These studies, which included patients with pancreatic necrosis and antibiotic prophylaxis, were performed with antibiotics of adequate pharmakokinetic profiles. The first results were very promising; the first clinical study published by Pederzoli²² showed a significant advantage for prophylaxis with imipenem in reducing pancreatic and non-pancreatic septic complications compared with a non-antibiotic control group. Even more promising were the results of the Sainio study that demonstrated a significant reduction in mortality by cefuroxime prophylaxis.²³ Most interestingly, this effect was not attributable to prevention of pancreatic sepsis, as the incidence of infected necrosis did not differ between both treatment groups.

In 1996 and 1997, two studies were published where either ofloxacin in combination with metronidazole or ceftazidime in combination with amikacin and metronidazole were investigated.^24,25 The fact that none of these series was able to show a beneficial effect of antibiotic prophylaxis could be attributed to the small number of patients recruited.

In the late 1990s, efforts were made to compare different antibiotic regimes in their capability to prevent pancreatic infection and improve the patient’s prognosis.

Bassi and co-workers have compared imipenem (3x500mg/day) and pefloxacin (2x400mg/day) in a controlled clinical trial in patients with necrosis of more than 50% of the gland’s parenchyma.²⁶ Interestingly, imipenem was superior to pefloxacin in reducing the rate of infected pancreatic necrosis (10% versus 34%), but this did not result in a significant reduction of mortality (10% versus 24%). It is remarkable that in the pefloxacin group, the incidence of infected necrosis was comparable with the rate of infection seen in former series with untreated control groups.^22,23

In another series, two different regimes of imipenem application were compared.²⁷ Patients either received imipenem prophylaxis immediately after hospital admission or had delayed treatment with imipenem when they did not respond to conservative treatment. It is interesting to note that the clinical condition of nine patients receiving delayed treatment improved after initiation of imipenem therapy; they recovered without surgical intervention. It remains unclear whether the response of these patients is due to the antibacterial effect of imipenem or its capability to interfere in the inflammatory mediator cascade responsible for the systemic complications of the disease.

As carbapenems are currently regarded as the antibiotics of choice in necrotising pancreatitis, it is not surprising that different members of this group of antibiotics have been compared with regard to their effects in severe acute pancreatitis. Nor is it surprising, given that these drugs are among the most otent antibiotics, that a recent comparison of meropenem (3x500mg/day) and imipenem (4x500mg/day) failed to show superiority of one of the regimes over the other.²⁸ It is, however, remarkable to see that the incidence of pancreatic infection (11.4% versus 13.6%) and mortality (13.6% versus 11.4%) was low in both treatment groups. As the patient population of this study overwhelmingly consisted of patients with extended pancreatic necrosis (the amount of necrotic areas exceeding 50% of the gland’s parenchyma), thus representing the group with the poorest prognosis, the authors interpret their results as proof for the beneficial potential of carbapenem prophylaxis.