Current issues of ACP Journal Club are published in Annals of Internal Medicine


Norfloxacin reduced bacterial infections and total antibiotic costs in patients with cirrhosis who developed gastrointestinal hemorrhage

ACP J Club. 1993 Mar-April;118:44. doi:10.7326/ACPJC-1993-118-2-044

Source Citation

Soriano G, Guarner C, Tomás A, et al. Norfloxacin prevents bacterial infection in cirrhotics with gastrointestinal hemorrhage. Gastroenterology. 1992 Oct;103:1267-72.



To investigate whether selective intestinal decontamination with norfloxacin can prevent bacterial infections in patients with cirrhosis who develop gastrointestinal hemorrhage.


Randomized controlled trial with follow-up to hospital discharge.


Gastrointestinal hemorrhage unit of a university teaching hospital in Spain.


All patients had cirrhosis and were hospitalized with gastrointestinal hemorrhage. Exclusion criteria were signs of infection, treatment with antibiotics in the 2 weeks before admission, or transfer from another hospital. 128 patients were randomized and 119 were analyzed (mean age 60 y, 61 men). 8 patients died or had surgery within 24 hours of admission and were excluded. Follow-up was 99%.


60 patients received norfloxacin, 400 mg twice daily orally or through a nasogastric tube for 7 days. Norfloxacin was started immediately after initial endoscopy and restarted after recurrence of hemorrhage. 59 control patients received identical care without norfloxacin. If infection developed, patients were treated with antibiotics.

Main outcome measures

Physical examinations and blood counts were done daily to detect bacterial infections and drug side effects. At admission, and if there was clinically suspected infection, patients had chest radiographs and urine and blood cultures and analyses. Spontaneous bacterial peritonitis was defined as ≥ 250 neutrophils/µL; bacteremia and urinary infections were culture defined; and respiratory infections were diagnosed by physical, analytic, and radiographic data. Chart audit provided mortality data.

Main results

The groups did not differ for mortality although fewer deaths occurred in the norfloxacin group (4 vs 7 deaths). Fewer patients taking norfloxacin developed infections (P = 0.001); bacteremia, spontaneous bacterial peritonitis, or both (P < 0.05); and urinary infections (P = 0.001) (Table). The groups did not differ for respiratory infections (7% vs 7%, P = 0.96). The total cost of antibiotics was less for those who took norfloxacin ($80 vs $210 per patient). Norfloxacin had no side effects.


Selective intestinal decontamination with norfloxacin in patients with cirrhosis who developed gastrointestinal hemorrhage reduced the incidence of bacterial infections and total antibiotic costs.

Source of funding: Madaus Cerafarm Laboratories (norfloxacin).

For article reprint: Dr. C. Guarner, Liver Unit, Department of Gastroenterology, Hospital de la Santa Creu i Sant Pau, Avinguda St. Antoni M. Claret 167, 08025 Barcelona, Spain. FAX 34-3-456-0190.

*Numbers calculated from data in article.

Table. Norfloxacin vs no norfloxacin for bacterial infections in patients with cirrhosis who developed gastrointestinal hemorrhage*

Outcomes Norfloxacin No norfloxacin RRR (95% CI) NNT (CI)
Infections 10% 37% 73% (41 to 88) 4 (2 to 8)
Bacteremia or spontaneous bacterial peritonitis or both 3% 17% 80% (25 to 95) 8 (4 to 32)
Urinary tract infections 0 19% 100% (67 to 100) 6 (3 to 9)

*Abbreviations defined in Glossary; RRR, NNT, and CI calculated from data in article.


In this exciting report, Soriano and colleagues show a reduced incidence of infection in patients with cirrhosis and hemorrhage by the prophylactic use of twice-daily oral or intravenous norfloxacin, a fluoroquinolone that kills most colonic aerobic bacterial groups while preserving the anaerobic flora. The norfloxacin-treated group had a reduced incidence of infection during 2 weeks of hospitalization because of a decreased incidence of both bacteremia and urinary tract infection (spontaneous bacterial peritonitis alone was not affected).

Two concerns preclude routine incorporation of norfloxacin prophylaxis into my practice. First, bacterial resistance to fluoroquinolones is increasing, and mutations conferring resistance can occur during a brief course of therapy (1). It would be important, therefore, to show significantly reduced morbidity or mortality before using treatment that could potentially change hospital flora. No difference in mortality, however, was noted between treatment and control groups. Although some cost savings were found in the norfloxacin group, this may reflect hospital flora, treatment regimens, and local drug prices.

Second, the organisms found less frequently in the treated group were aerobic gram-negative bacilli and enterococci. Although 60% to 70% of patients in both groups had urinary catheters, the authors do not report whether these infections occurred disproportionately in the catheterized patients. If so, the effect of norfloxacin on infection risk in the noncatheterized patients with cirrhosis might diminish, but the benefit would be clearer for patients who cannot avoid catheterization.

The authors deserve high praise for completing this difficult study. I eagerly await further delineation of the role of antimicrobial prophylaxis in liver disease.

Joseph M. Alcorn, MD
University of New MexicoAlbuquerque, New Mexico, USA


1. Segreti J, Gootz TD, Goodman LJ, et al. High-level quinolone resistance in clinical isolates of Campylobacter jejuni. J Infect Dis. 1992;165:667-70.