Surgical Endoscopy

, Volume 25, Issue 9, pp 2911–2918

The effects of cefazolin on cirrhotic patients with acute variceal hemorrhage after endoscopic interventions

Authors

  • Huang-Wei Xu
    • Division of Hepato-Gastroenterology, Department of Internal Medicine, Chang Gung Memorial Hospital-Kaohsiung Medical CenterChang Gung University College of Medicine
  • Jing-Houng Wang
    • Division of Hepato-Gastroenterology, Department of Internal Medicine, Chang Gung Memorial Hospital-Kaohsiung Medical CenterChang Gung University College of Medicine
  • Moan-Shane Tsai
    • Division of Infectious Diseases, Department of Internal Medicine, Chang Gung Memorial Hospital-Kaohsiung Medical CenterChang Gung University College of Medicine
  • Keng-Liang Wu
    • Division of Hepato-Gastroenterology, Department of Internal Medicine, Chang Gung Memorial Hospital-Kaohsiung Medical CenterChang Gung University College of Medicine
  • Shue-Shian Chiou
    • Division of Hepato-Gastroenterology, Department of Internal Medicine, Chang Gung Memorial Hospital-Kaohsiung Medical CenterChang Gung University College of Medicine
  • Chi-Sin Changchien
    • Division of Hepato-Gastroenterology, Department of Internal Medicine, Chang Gung Memorial Hospital-Kaohsiung Medical CenterChang Gung University College of Medicine
  • Tsung Hui Hu
    • Division of Hepato-Gastroenterology, Department of Internal Medicine, Chang Gung Memorial Hospital-Kaohsiung Medical CenterChang Gung University College of Medicine
  • Sheng-Nan Lu
    • Division of Hepato-Gastroenterology, Department of Internal Medicine, Chang Gung Memorial Hospital-Kaohsiung Medical CenterChang Gung University College of Medicine
    • Division of Hepato-Gastroenterology, Department of Internal Medicine, Chang Gung Memorial Hospital-Kaohsiung Medical CenterChang Gung University College of Medicine
Article

DOI: 10.1007/s00464-011-1642-0

Cite this article as:
Xu, H., Wang, J., Tsai, M. et al. Surg Endosc (2011) 25: 2911. doi:10.1007/s00464-011-1642-0

Abstract

Background

The American Association for the Study of Liver Diseases (AASLD) guidelines recommend that antibiotic prophylaxis should be instituted in any patient with cirrhosis and gastrointestinal hemorrhage, and that oral norfloxacin, intravenous ciprofloxacin, and ceftriaxone are preferable. However, the antimicrobial spectrum of the first generation of cephalosporins (cefazolin) covers a wide range of bacteria species, including community-acquired strains of Escherichia coli and Klebsiella pneumoniae, but their efficacy as prophylactic antibiotics in cirrhotic patients with acute hemorrhage was seldom warranted in the literature. This study aimed to explore the effects of cefazolin on the outcome of cirrhotic patients with acute variceal hemorrhage after endoscopic interventions.

Methods

A cross-sectional, retrospective chart review study was conducted on cirrhotic patients with acute variceal hemorrhage who underwent endoscopic procedures in a medical center. Cirrhotic patients who did not receive antibiotics were classified as group A (n = 63) while patients who received intravenous cefazolin 1 g q8 h for 2–7 days were classified as group B (n = 50). The end points were the prevention of infection, length of hospital stay, time of rebleeding, and death.

Results

A total of 113 patients were studied (male/female: 82/31; age: 56.8 ± 13.5 years). The incidence of infection (including proven infections) and bacteremia were significantly lower in group B patients (38.1% vs. 16.0%, P = 0.010; 17.5% vs. 4.0%, P = 0.026; 9.5% vs. 0%, P = 0.033, respectively). The no prophylactic antibiotics treatment was the independent risk factor. There was no significant difference between the two groups with respect to the source of bleeding, type of endoscopic intervention, length of hospital stay, and mortality. Actuarial probability of remaining free of early rebleeding (<7 days) was P = 0.105 by log-rank test for all cirrhosis patients and P = 0.085 for Child-Pugh class A patients.

Conclusions

The use of cefazolin in cirrhotic patients after endoscopic interventions for acute variceal hemorrhage reduced infections. A trend of actuarial probability of remaining free of early rebleeding (<7 days) was observed, especially in Child-Pugh class A patients. This study may be hampered by the small sample size and more large-scale studies are mandatory.

Keywords

CefazolinCirrhosisVariceal hemorrhageBacterial infectionsRebleeding

Bacterial infections are common in patients with cirrhosis who are hospitalized because of gastrointestinal hemorrhage [1]. Many clinical trials proved that prophylactic antibiotics with oral norfloxacin, intravenous ciprofloxacin, or ceftriaxone in cirrhotic patients with gastrointestinal (GI) bleeding can reduce infections and mortality [25]. Previous studies showed that enteric aerobic Gram-negative bacteria (GNB) are the most common causative organisms in patients with acute hemorrhage [2, 3, 68]. However, there are some epidemiological changes in the etiology of infections because Gram-positive cocci (GPC) have markedly increased in cirrhotic patients, probably due to more invasive procedures in treating liver diseases in recent years [5]. Quinolones are the most commonly used prophylactic antibiotics in cirrhotic patients with GI hemorrhage because of the antibacterial activity against GNB bacteria. The broad-spectrum antibiotic intravenous ceftriaxone was superior to norfloxacin in areas with high quinolone resistance, particularly in those patients who cannot receive oral administration of any medications due to acute GI bleeding [9]. The American Association for the Study of Liver Diseases (AASLD) guidelines recommended that antibiotic prophylaxis should be instituted in any patient with cirrhosis and GI hemorrhage, and oral norfloxacin (400 mg twice daily), intravenous ciprofloxacin, and intravenous ceftriaxone (1 g/day) may be preferable [10].

The antimicrobial spectrum of cefazolin covers a wide range of bacteria species, including community-acquired strains of Escherichia coli and Klebsiella pneumoniae [11]. Theoretically, it is reasonable to assume that cefazolin should be efficacious as a prophylactic antibiotic in cirrhotic patients with acute hemorrhage but was seldom warranted in the literature. We therefore conducted the current retrospective outcome study to explore the effects of the intravenous first-generation antibiotic cefazolin on cirrhotic patients with acute variceal hemorrhage after endoscopic interventions.

Patients and methods

Patients

A cross-sectional, retrospective chart review study was conducted on cirrhotic patients with acute variceal hemorrhage who underwent endoscopic procedures in Chang Gung Memorial Hospital-Kaohsiung Medical Center between July 2009 and April 2010. A total of 113 patients who met the criteria were included in the study (male/female: 82/31; age, 57.45 ± 13.25 years). Patients with existing signs of infection (body temperature [BT] >38°C, white blood cell [WBC] count >11,000/μl) or who received oral or parenteral antibiotics in the week prior to the procedure were excluded. The prescription of an antimicrobial agent was at the discretion of the attending physician. The diagnosis of cirrhosis was made according to the clinical, laboratory, abdominal ultrasonographic, or histological findings [12]. The severity of cirrhosis was classified according to Pugh’s modification of Child’s classification [13].

The studied patients were divided into two groups. Patients who did not receive antibiotic after the endoscopic procedure were classified as group A (n = 63), while those who received intravenous cefazolin 1 g every 8 h for 2–7 days were classified as group B (n = 50). The choice of antibiotics, dose, and duration of therapy was determined by the clinicians. Two sets of blood culture were obtained before administering the antibiotics. When cirrhotic patients with acute gastrointestinal hemorrhage arrived at the emergency room (ER), vital signs were checked and laboratory tests, including white blood cells, hemoglobin, platelet count, prothrombin time, albumin, and bilirubin, were obtained. Terlipressin (Glypressin®) or octreotide (Sandostatin®) was administered for 3 days. A nasogastric tube, two intravenous catheters, or central venous catheters were placed according to clinical indication. Gastric varices (GV) or esophageal varices (EV) bleeding was diagnosed by (1) clinical signs of hematemesis, coffee ground vomitus, hematochezia, or melena; (2) endoscopic signs of active bleeding, adherent blood clots, white nipple signs, or erosions on varices; and/or (3) large varices with a red-color sign without other bleeding sources. The endoscopic procedure was arranged within 24 h of the patient’s arrival at the ER. Endoscopic variceal ligation (EVL) or endoscopic variceal injection sclerotherapy (EIS) was performed by an experienced endoscopist.

Definitions

Rebleeding was defined as a new onset of hematemesis, coffee-ground vomitus, hematochezia, or melena, with an increasing pulse rate of over 110 beats/min and decreasing blood pressure below 90 mmHg after a 24 h period of stable vital signs and hematocrit following endoscopic treatment. Early rebleeding was defined as rebleeding within 7 days of enrollment after initial control of bleeding.

The diagnosis of spontaneous bacterial peritonitis was based on ≥250 neutrophils/μl in ascitic fluid [14]. Bacteremia was defined as the presence of viable bacteria in the blood.

Patients without any identified infection source but with BT ≥38°C, leukocytosis >11000/μl with neutrophilia, hypothermia ≤36°C, alteration in mental status, and a greater than expected hemodynamic instability and oliguria were considered as having possible infection and received on-demand antibiotics. Patients with infections were the total of patients with proven infections and possible infections and clinicians’ determination as an infectious episode. Urinary tract infection was diagnosed when there were a positive urine culture of ≥105 colonies/ml with urine neutrophil count of >10 neutrophils/μl and associated clinical pictures. The diagnosis of respiratory infections was made when there were positive clinical symptoms and signs and positive chest X-ray findings. Infections relevant to the procedure were those infectious episodes that occurred from inclusion in the trial to 10 days or those that occurred during the first hospitalization.

The end points were prevention of infection (patients with BT >38°C or WBC >11,000/μl, and by the clinician’s determination as an infectious episode), length of hospital stay (from the arrival at ER to discharge day), time of rebleeding, death (died during hospitalization).

Statistical analysis

All results were expressed as means ± standard deviations for continuous data and as frequencies or percentages for categorical data. Distributions of continuous variables were analyzed by the χ2 test, Fisher’s exact test, or independent-sample t test, depending upon the type of data analyzed for the two groups where appropriate. Kaplan–Meier analysis with the log-rank test was used to compare differences of rebleeding between the two groups. Stepwise logistic regression analysis was used to identify the independent variables. Statistical significance was taken as P < 0.05. All statistical analyses were performed using SPSS ver. 17 (SPSS, Inc., Chicago, IL, USA).

Results

In our 113 patients, there were 21 patients diagnosed by a liver biopsy; others were diagnosed by abdominal ultrasonography, and clinical and laboratory data. The etiologies of cirrhosis among the 113 patients in the current study were alcoholism (n = 28, 24.8%), hepatitis B (n = 53, 46.9%), hepatitis C (n = 37, 32.7%), and cryptogenic (n = 12, 10.6%) (Table 1). There were no significant differences in any of the clinical and laboratory data between the two groups except infections. Overall, there were 38.1% of infections in group A and 16.0% in group B (P = 0.010) (Table 2). There were 11 patients (17.5%) in group A and 2 patients (4.1%) in group B with proven infections (P = 0.026). Thirteen patients (20.6%) in group A and 6 patients (12.0%) in group B had possible infections (P = 0.223). Among them, three patients (4.8%) in group A suffered from pneumonia, and three patients (4.8%) in group A and one patient (2.0%) in group B had urinary tract infection (UTI). Spontaneous bacterial peritonitis (SBP) occurred in four patients (6.3%) in group A and one (2%) in group B (Table 2, Fig. 1). Of those patients who developed bacterial infections, six sets of bacterial cultures from blood samples yielded positive results. All six belonged to group A and were Escherichia coli (3), Klebsiella pneumoniae (2), and Haemophilus influenzae (1) (Table 2). The incidence of bacteremia was also significantly lower in group B patients (P = 0.033).
Table 1

Clinical characteristics of the patients

 

Total (N = 113) (%)

Age (years)

56.8 ± 13.5

Sex (M/F)

82 (72.6)/31 (27.4)

Etiology

 Alcoholic

28 (24.8)

 HBV

53 (46.9)

 HCV

37 (32.7)

Vitals at the ER

 BT (°C)

36.6 ± 0.5

 HR (beats/min)

90.4 ± 16.8

 SBP (mmHg)

123.2 ± 29.8

Laboratory data

 WBC (x109/L)

6.0 ± 2.1

 Hb (g/dl)

9.3 ± 2.0

 PLT (109/L)

108.8 ± 74.9

 PT (s)

12.8 ± 2.3

 Albumin (g/L)

3.1 ± 0.6

 Bilirubin (mg/dL)

1.9 ± 1.7

Child-Pugh class A/B/C

61 (53.1)/48 (42.5)/4 (3.5)

Bleeding source:EV/GV

88 (77.9)/25 (22.1)

Treatment: EVL/EIS

47 (41.6)/22 (19.5)

Patients with infections, n(%)

32 (28.3)

 Proved infections

13 (11.5)

 Possible infections

19 (16.8)

Type of proved infection:

 Bacteremia

6 (5.3)

 Pneumonia

3 (2.7)

 Urinary tract infection

4 (3.5)

 Spontaneous bacterial peritonitis

5 (4.4)

Organisms, n (%)

 Escherichia coli

3 (2.7)

 Klebsiella pneumoniae

2 (1.8)

 Haemophilus influenzae

1 (0.9)

Hospital days

10.4 ± 9.9

Death

5 (4.4)

HBV hepatitis B virus, HCV hepatitis C virus, ER emergency room, BT body temperature, HR heart rate, SBP systolic blood pressure, WBC white blood cells, Hb hemoglobin, PLT platelet count, PT prothrombin time, EV esophageal varices, GV gastric varices, EVL endoscopic variceal ligation, EIS endoscopic injection sclerotherapy

Table 2

Clinical characteristics of the patients in the no antibiotic prophylaxis group and the cefazolin group

 

Group A (N = 63) (%)

Group B (N = 50) (%)

P value

Age (years)

55.8 ± 13.4

58.1 ± 13.7

0.368

Sex (M/F)

46 (73.0)/17 (27.0)

36 (72.0)/14 (28.0)

0.904

Etiology

 Alcoholic

14 (22.2)

14 (28.0)

0.480

 HbsAg(+)

31 (49.2)

22 (44.0)

0.582

 Anti-HCV(+)

21 (33.3)

16 (32.0)

0.881

Vitals at the ER

 BT (°C)

36.6 ± 0.5

36.6 ± 0.5

0.977

 HR (beats/min)

89.8 ± 16.9

91.1 ± 16.7

0.673

 SBP (mmHg)

118.3 ± 30.4

129.3 ± 28.2

0.052

Laboratory data

 WBC (x109/L)

6.2 ± 2.1

5.9 ± 2.1

0.448

 Hb (g/dl)

9.1 ± 1.9

9.7 ± 2.0

0.127

 PLT (109/L)

115.5 ± 78.3

100.5 ± 70.1

0.294

 PT (s)

12.8 ± 2.7

12.8 ± 1.7

0.910

 Albumin (g/L)

3.1 ± 0.6

3.1 ± 0.5

0.934

 Bilirubin (mg/dl)

2.0 ± 2.0

1.8 ± 1.1

0.587

Child-Pugh class A/B/C

31 (49.2)/29 (46.0)/3 (4.8)

30 (60.0)/19 (38.0)/1 (2.0)

0.487

Bleeding source: EV/GV

51 (81.0)/12 (19.0)

37 (74.0)/13 (26.0)

0.377

Treatment: EVL/ EIS

26 (42.9)/13 (20.6)

20 (40.0)/9 (18.0)

0.831

Patients with infections, n (%)

24 (38.1)

8 (16.0)

0.010*

 Proved infections

11 (17.5)

2 (4.0)

0.026**

 Possible infections

13 (20.6)

6 (12.0)

0.223

Type of proved infection

 Bacteremia

6 (9.5)

0 (0.0)

0.033***

 Pneumonia

3 (4.8)

0 (0.0)

0.253

 Urinary tract infection

3 (4.8)

1 (2.0)

0.628

 Spontaneous bacterial peritonitis

4 (6.3)

1 (2.0)

0.381

Organisms, n (%)

 Escherichia coli

3 (4.8)

0 (0.0)

0.253

 Klebsiella pneumoniae

2 (3.2)

0 (0.0)

0.502

 Haemophilus influenzae

1 (1.6)

0 (0.0)

0.1

Hospital days

9.2 ± 7.9

11.8 ± 12.0

0.171

Death

3 (4.8)

2 (4.0)

1

HBV hepatitis B virus, HCV hepatitis C virus, ER emergency room, BT body temperature, HR heart rate, SBP systolic blood pressure, WBC white blood cells, Hb hemoglobin, PLT platelet count, PT prothrombin time, EV esophageal varices, GV gastric varices, EVL endoscopic variceal ligation, EIS endoscopic injection sclerotherapy

P = 0.010, ** P = 0.026, *** P = 0.033

https://static-content.springer.com/image/art%3A10.1007%2Fs00464-011-1642-0/MediaObjects/464_2011_1642_Fig1_HTML.gif
Fig. 1

Types of bacterial infections in both groups of patients, including bacteremia, pneumonia, urinary tract infection, and spontaneous bacterial peritonitis. Gray column, no antibiotics group (group A); black column, cefazolin group (group B). UTI urinary tract infection, SBP spontaneous bacterial peritonitis

As shown in Table 2, there was no significant difference between the two groups with respect to bleeding source (EV or GV), type of endoscopic therapy (EVL or EIS), length of hospital stay, and mortality. The mean (±SD) hospital stay was 9.2 (±7.9) days in group A versus 11.8 (±12.0) days in group B. Age, gender, etiology of cirrhosis, presence of esophageal or gastric varices, laboratory data (white blood cells, hemoglobin, platelet count, prothrombin time, albumin, bilirubin), Child-Pugh classification, source of bleeding, and prophylactic treatment using cefazolin were all selected as independent variables in a multivariate stepwise logistic regression analysis, with the acquisition of infection as the only independent variable. Antibiotic prophylaxis with cefazolin was estimated to be an independent predictor of acquiring bacterial infection in cirrhotic patients with acute variceal hemorrhage after endoscopic interventions (odds ratio = 0.298; 95% confidence interval CI = 0.112-0.796; P = 0.016).

Hemostatic outcomes

The hemostatic outcome data is summarized in Table 3. The rebleeding rate was actually higher in group A (44.4%) than in group B (34.0%) but did not reach a significant difference (P = 0.206). For all the cirrhotic patients, the actuarial probability of remaining free of rebleeding was not significant different between the two groups (P = 0.550 by log-rank test) (Fig. 2). The actuarial probability of remaining free of early rebleeding (<7 days) was higher in group A than in group B, but again did not reach statistical significance although a trend was observed (P = 0.105 by log-rank test) (Fig. 3). However, if we considered only Child-Pugh class A cirrhotic patients, the overall actuarial probability of remaining free of rebleeding between the two groups was P = 0.229 by log-rank test (Fig. 4), and the trend of significance became more obvious in those with early rebleeding (<7 days) (P = 0.085 by log-rank test) (Fig. 5).
Table 3

Hemostatic outcome in patients with variceal bleeding following endoscopic treatment

 

Group A (N = 60) (%)

Group B (N = 49) (%)

P value

No. of rebleeding

28 (44.4%)

17 (34.0%)

0.206

Time of rebleeding

 Early (≤6 weeks)

15

11

0.802

  <7 days

6

1

0.131

  7–14 days

3

3

1

  15–42 days

6

7

0.459

 Late (>6 weeks)

13

6

0.223

https://static-content.springer.com/image/art%3A10.1007%2Fs00464-011-1642-0/MediaObjects/464_2011_1642_Fig2_HTML.gif
Fig. 2

Actuarial probability of remaining free of rebleeding in all cirrhotic patients for the no antibiotic prophylaxis group (group A) and the cefazolin group (group B) (P = 0.550 by log-rank test)

https://static-content.springer.com/image/art%3A10.1007%2Fs00464-011-1642-0/MediaObjects/464_2011_1642_Fig3_HTML.gif
Fig. 3

Actuarial probability of remaining free of early rebleeding (<7 days) in all cirrhotic patients for the no antibiotic prophylaxis group (group A) and the cefazolin group (group B) (P = 0.105 by log-rank test)

https://static-content.springer.com/image/art%3A10.1007%2Fs00464-011-1642-0/MediaObjects/464_2011_1642_Fig4_HTML.gif
Fig. 4

Actuarial probability of remaining free of rebleeding in the Child-Pugh class A cirrhotic patients for the no antibiotic prophylaxis group (group A) and the cefazolin group (group B) (P = 0.229 by log-rank test)

https://static-content.springer.com/image/art%3A10.1007%2Fs00464-011-1642-0/MediaObjects/464_2011_1642_Fig5_HTML.gif
Fig. 5

Actuarial probability of remaining free of early rebleeding (<7 days) in the Child-Pugh class A cirrhotic patients for the no antibiotic prophylaxis group (group A) and the cefazolin group (group B) (P = 0.085 by log-rank test)

Discussion

Antibiotic prophylaxis for cirrhotic patients with gastrointestinal hemorrhage is efficacious in reducing the number of deaths and bacterial infections and should be practiced [10]. Quinolones or third-generation cephalosporins were tested in trials with a median duration of treatment of 7 days [1, 10]. Obviously, all the previous trails used strong antibiotics before or after the endoscopic procedure [1]. The reports on the effect of first-generation cephalosporins on cirrhotic patients with acute gastrointestinal hemorrhage are scarce. Our results showed that the use of cefazolin as a prophylactic antibiotic in cirrhotic patients after endoscopic interventions for acute variceal hemorrhage was efficacious in reducing infections, but did not shorten the hospital stay, shorten rebleeding, or decrease mortality. However, a trend of actuarial probability of remaining free of early rebleeding (<7 days) was observed in patients with cefazolin prophylaxis.

Bacterial infections occur more frequently in cirrhotic patients with upper gastrointestinal hemorrhage than in those admitted to hospital with other forms of decompensation such as encephalopathy and may be associated with mortality [25, 15]. Early publications revealed that the most frequent bacterial infections seen in cirrhotic patients are UTI (12–29%), most frequently caused by GNB (E. coli or Klebsiella); SBP (7–23%), caused by GNB and aerobic GPC, usually as a single organism; respiratory tract infections (6–10%), predominantly caused by GNB and Staphylococcus species; and primary bacteremia (4–11%) [1]. However, there has been a change in the epidemiology of bacterial infections in cirrhotic patients in the past decade which is probably related to the changes in the practice of physicians, particularly those who use invasive procedures such as EVL, EIS, arterial embolization, or percutaneous ablation of hepatocellular carcinoma [5]. On the other hand, long-term norfloxacin has been prescribed to prevent infections caused by GNB but not those caused by GPB. Norfloxacin prevents infections caused by GNB but not those caused by GPC. All these features may explain the epidemiological changes of acute bacterial infections in cirrhotic patients [16]. The result of current study was compatible with the recent reports on epidemiological changes. The incidence of infection in patients with acute variceal hemorrhage who did not receive antibiotic prophylaxis after endoscopic procedures was 38.1%. Among them, SBP (6.3%), pneumonia (4.8%), and UTI (4.8%) were the major causes of infection. Patients with bacteremia (9.5%) all belonged to this group. The probable mechanism for infections in cirrhotic patients may be bacterial translocation caused by predisposition to intestinal bacterial overgrowth, intestinal dysmotility, and increased intestinal permeability [17]. In addition, bacterial byproducts such as endotoxin can cause an increase in portal pressure, impairment of liver function, and worsening of hemostasis. It had been reported that infections were present in up to 20–50% of the cirrhotic patients while hospitalized [1, 18]. The role of intestinal bacteria as a major source of infection in cirrhotic patients may explain why prophylactic antibiotics, both orally and intravenously, are so effective in reducing infections during variceal bleeding. There are only a few types of intestinal bacteria, such as Escherichia coli, Klebsiella pneumoniae, Enterobacteriaceae, Pseudomonas aeruginosa, Enterococci, and some Streptococci, that are able to translocate into monolymphocytic neutrophils [17]. Cirrhotic patients may also be infected by other spontaneous infections caused by intestinal bacteria as a consequence of bacterial translocation besides SBP. It was reported that selective intestinal decontamination with antibiotics reduces bacterial translocation [19, 20] and decreases the risk of infections, especially those due to GNB [16]. Thus, it is a matter of consequence that cycling events that involve damage to the intestinal barrier leading to bacterial translocation, endotoxemia, impairment of liver function, and increase in portal pressure possibly cause further damage to the gut. Infection is associated with failure to control variceal hemorrhage and early variceal rebleeding in these patients [21]. The possible pathophysiological pathway behind this is that infections cause impairment of liver function and coagulation which lead to variceal hemorrhage. Three phases of events occur during the process: endotoxemia, bacterial translocation, and bacterial infections. First, the bleeding event was triggered by the priming of monocytes and an increase of nitric oxide (NO) and tumor-necrotic factor-α (TNF-α) [22]. Then, a further increase of NO and TNF-α and systemic vasodilatation reduce response to vasoconstrictor in the translocation phase. Eventually, the result is bacterial infections that cause variceal hemorrhage and probable mortality [17].

There was evidence of the benefit of using prophylactic antibiotics for cirrhotic patients who underwent endoscopic intervention for bleeding varices, but most studies achieved this goal by using strong potent antibiotics. They used a quinolone in doses ranging from 400 mg of ofloxacin, 400–800 mg of norfloxacin, and 400–1000 mg of ciprofloxacin. Other regimens included a quinolone given with amoxicillin-clavulanate, nonabsorbable antibiotics, and imipenem-cilastin [2, 3, 68, 23, 24]. There were other studies that used intravenous ceftriaxone, a third-generation cephalosporin, which may be superior to quinolone in the prophylaxis of bacterial infections in patients with advanced cirrhosis and hemorrhage in a high-quinolone-resistance area [9]. In terms of clinical efficacy, ceftriaxone proved to be a better option than quinolone, particularly in patients with advanced cirrhosis and UGI hemorrhage in a high-quinolone-resistance area, because ceftriaxone offers activity against both Gram-positive and Gram-negative organisms. Although cefazolin provides a narrower spectrum of activity against both Gram-negative and Gram-positive bacteria than ceftriaxone, most community-acquired bacteria strains were susceptible to cefazolin. Thus, it is reasonable to assume, and it was demonstrated in this study, that a first-generation cephalosporin agent offered comparable clinical efficacy in preventing bacterial infection in cirrhotic patients with hemorrhage. As mentioned earlier, the first-generation cephalosporin cefazolin is effective for both Gram-positive and Gram-negative microorganisms, and the current study has proven that it reduced infections in cirrhotic patients with acute variceal hemorrhage. What about the effect on hemostatic outcome? It is optimistic that if such a goal can be attained by prescribing first-generation antibiotics, antibiotic-resistant strains could be avoided. Besides, medical costs may be reduced. Unfortunately, the current study was not powered to support the point that prescribing the first-line cephalosporin as prophylaxis antibiotics prevented rebleeding. However, more rebleeding occurred in group A than in group B, but the cumulative rate did not reach a significant difference (44.4 vs. 34.0%, P = 0.206). A trend of actuarial probability of remaining free of early rebleeding (<7 days) was also observed in Child-Pugh class A patients with cefazolin prophylaxis (P = 0.085 by log-rank test). There may be an underestimation due to the relatively small sample size in this study. Therefore, more studies with a larger sample size are urged to further clarify this issue. The possible mechanism behind this could be explained by the fact that cefazolin could not reverse the systemic vasodilatation and prevent early variceal rebleeding [20]. Two studies had shown that prophylactic antibiotics could prevent rebleeding in cirrhotic patients with acute hemorrhage. Hou et al. [25] showed that the risk of recurrent bleeding may be reduced by giving prophylactic antibiotics (intravenous ofloxacin 200 mg q12 h for 2 days followed by oral ofloxacin 200 mg q12 h for 5 days) in cirrhotic patients with esophageal varices bleeding. Jun et al. [26] showed that antibiotic prophylaxis using third-generation cephalosporins (cefotaxime) prevented both bacterial infection and early rebleeding in patients at the first acute gastroesophageal variceal hemorrhage.

The cost of the drugs is also an important consideration when prescribing antibiotics, besides the efficacy and drug resistance. In Taiwan, the National Health Insurance (NHI) drug price of a single dose of ceftriaxone (Rocephin, 500 mg) is 4.9 US dollars, ciprofloxacin (200 mg) is 10 US dollars, and cefazolin (Stazolin, 1000 mg) is 0.69 US dollars [Bureau of National Health Insurance Taiwan (http://www.nhi.gov.tw)]. Therefore, the use of cefazolin instead of third-generation cephalosporins as prophylactic antibiotics may be an economical choice in terms of preventing infections after endoscopic interventions for acute variceal hemorrhage in cirrhotic patients.

This study has limitations. First, it is a retrospective chart review study and the sample size is relatively small so bias may exist. Second, blood culture was not performed for all the enrolled patients.

In conclusion, the use of cefazolin in cirrhotic patients after endoscopic interventions for acute variceal hemorrhage reduced infections and bacteremia. A trend of actuarial probability of remaining free of early rebleeding (<7 days) was observed in patients with cefazolin prophylaxis rebleeding (<7 days), but did not shorten hospital stay or decrease mortality. This issue may be hampered by the small sample size and therefore more large-scale studies are mandatory.

Disclosures

Drs. Huang-Wei Xu, Jing-Houng Wang, Moan-Shane Tsai, Keng-Liang Wu, Hue-Shian Chiou, Chi-Sin Changchien, Tsung Hui Hu, Sheng-Nan Lu, and Seng-Kee Chuah have no conflicts of interest or financial ties to disclose.

Copyright information

© Springer Science+Business Media, LLC 2011