Antibiotic prophylaxis in percutaneous endoscopic gastrostomy (PEG): a prospective randomized clinical trial

BACKGROUND AND STUDY AIMS

The most frequent complication reported for percutaneous endoscopic gastrostomy (PEG) is peristomal wound infection which occurs in as many as 30 % of patients. In the studies published so far, the question of whether antibiotic prophylaxis reduces the incidence of peristomal wound infection has remained controversial. We therefore conducted a prospective, randomized trial to determine whether antibiotic prophylaxis can reduce the incidence of peristomal wound infection associated with PEG.

PATIENTS AND METHODS

During a 2-year-period a total of 347 patients (251 men, 96 women; mean age 60.2+/-5) were included in our study. In all cases, a 9-French gauge Freka gastrostomy tube was inserted using the "pull" technique. The patients were randomly assigned to three different treatment groups (group 1, antibiotic prophylaxis with 2 g cefotaxime; group 2, antibiotic prophylaxis with 4g piperacillin plus 0.5g tazobactam; group 3, no antibiotic prophylaxis). For a period of one week after gastrostomy, the peristomal area was evaluated using a wound score. According to the score, peristomal infections were classified as being of degree I-IV. Infections of up to degree III were regarded as minor, while an infection of degree IV was defined as a severe complication.

RESULTS

PEG tubes were successfully placed in 336/347 patients (97 %). Of the 336 patients, 29 were excluded because of incomplete follow up (dropout rate 9%). Cases included were: in group 1, n = 101; in group 2, n = 100, and in group 3, n = 106. There was no mortality related to PEG insertion. There were five instances of severe complications (1.6%). Patients who received antibiotic prophylaxis had mean daily combined wound scores that were significantly lower than those of patients without antibiotic prophylaxis (P < 0.01). The control group (no antibiotic prophylaxis) exhibited significantly more peristomal wound infections of degree III (n = 8) than antibiotic prophylaxis group 1 (n = 0, P < 0.001) or antibiotic prophylaxis group 2 (n = 1, P < 0.012). Three patients in group 3 (no antibiotic prophylaxis) and one patient in group 2 developed peritonitis.

CONCLUSIONS

Severe wound infections requiring medical or endoscopic intervention are very rare events after PEG insertion. Antibiotic prophylaxis significantly reduces the risk of peristomal wound infection associated with PEG insertion. Antibiotic prophylaxis, therefore, is to be recommended as a general measure in percutaneous endoscopic gastrostomy.

K-ras point mutations are rare events in premalignant forms of Barrett's oesophagus

OBJECTIVE

In Barrett's adenocarcinomas, in contrast to squamous oesophageal carcinomas, K-ras point mutations are thought to be a frequent event. The frequency of K-ras point mutations in premalignant forms of Barrett's oesophagus (metaplasia, dysplasia) leading to adenocarcinoma with increased risk is currently not known. To establish the frequency of K-ras mutations in premalignant forms of Barrett's oesophagus, we investigated oesophageal biopsy specimens with Barrett's metaplastic and dysplastic epithelium for point mutations in the K-ras gene/codons 12, 13.

DESIGN

A total of 412 biopsies from patients with Barrett's oesophagus were histologically classified into biopsies with metaplasia (n = 252), dysplasia (n = 105) and adenocarcinoma (n = 11), as well as biopsies distant from disease (normal, n = 37 and hyperplastic squamous epithelium, n = 7).

METHODS

DNA from biopsy specimens was amplified by polymerase chain reaction (PCR) with a modified primer for generating a restriction site in the case of wild type in codon 12. Wild-type or point mutations in the K-ras gene/codons 12, 13 were detected by restriction fragment length analysis of the PCR products.

RESULTS

Point mutations in K-ras/codon 12 were found in 9 biopsies (n = 1 in metaplasia, n = 4 in dysplasias, n = 4 in adenocarcinomas). All the other biopsies showed the wild type of K-ras/codon 12. No K-ras/codon 13 mutation (GGCgly-->GACasp) was observed.

CONCLUSION

Mutations in K-ras/codon 12 were rarely found in premalignant forms of Barrett's oesophagus. Whereas the screening for K-ras point mutations in metaplastic sites of Barrett's epithelium seems not to be of practical value, the screening for mutations in dysplastic lesions might be helpful to estimate the individual risk for progression of Barrett's epithelium to adenocarcinoma. A further evaluation in larger numbers of patients is needed.

Inhibition of histamine release in vitro by a blocking factor from human serum: comparison with the iron binding proteins transferrin and lactoferrin

Recently, we described the presence of a blocking factor (BF) in rat serum, which inhibited the histamine release from rat mast cells in vivo and in vitro. The blocking activity was demonstrated in human serum as well. Qualitative analysis of the purified preparations demonstrated a major component with an apparent molecular weight of 70,000 daltons. In human serum the blocking factor was identified as transferrin (TF) by serological and biochemical methods. BF (DEAE-peak 1) and the iron binding proteins transferrin and lactoferrin (LF) are shown to inhibit the histamine release in vitro. The dose response curves reveal that inhibition by these proteins is dependent on their degree of iron saturation. Furthermore, unlike lactoferrin, the effects of transferrin and BF (DEAE-peak 1) follow the same pattern. Their mechanism of action remains to be elucidated.

Purification of serum proteins with inhibitory activity on the histamine release in vitro and/or in vivo

Recently, we described a factor in rat and human serum with inhibitory activity on the 2-hour passive cutaneous anaphylactic (PCA) reaction in rat skin. The Ca-ionophore-induced histamine release from rat peritoneal mast cells was inhibited as well. In human serum, inhibitory activity for the in vitro histamine release from rat mast cells was shown to be present within two serum fractions. One molecule was purified by 50-70% ammonium sulfate precipitation and subsequent DEAE chromatography at pH 5.4. The main component showed complete identity with transferrin. Commercially available transferrin was active in vitro as well. The dose-response plot revealed two inhibition maxima at transferrin concentrations of 0.05-5 ng/ml and at more than 1 mg/ml, respectively. Subsequently it was shown that the degree of iron saturation was critical for the decrease in histamine release. The mediator release in vivo (PCA) could not be inhibited by the iron transport protein. Inhibitory activity in vivo was however mediated by the second serum fraction, which was characterized by a molecular weight of approximately 150,000 daltons and an IEP within the alkaline range. These data suggest that exogenous factors may be potent modulators of inflammatory reactions.