Bacterial infection monitoring in the early period after liver transplantation

Liver transplantation (LT) has become an established treatment for properly selected patients with end stage liver disease, primary and some secondary hepatic malignancies and some rare metabolic liver diseases. Increased use of potent immunosuppressive agents has dramatically reduced the incidence of rejection in the transplant population, while increasing patient susceptibility to opportunistic infections and cancer [1]. Infection and rejection remain the major causes of morbidity after LT, together accounting for up to 85% of deaths [2].

Many studies have reported that infection is one of the main factors influencing morbidity and mortality in posttransplant patients. Approximately 80% of solid-organ transplant recipients suffer at least one significant episode of infection during the first year following transplantation [3]. Most of the infections have a bacterial etiology and they occur early after transplantation [4, 5]. Infections in this patient population are notoriously difficult to diagnose because their usual signs and symptoms, such as fever and leukocytosis, may be masked or absent. Mild infections are dangerous in immunosuppressed patients [6].

An investigation of the methods useful for early detection of hidden infections is needed. Kim et al. [7] reported that periodic microbiologic surveillance is useful for predicting posttransplantation pneumonia and intra-abdominal infection.

This study comprises an analysis of bacterial infections that occur during the early period after LT in adults. Laboratory culture tests were performed every day and the results were analyzed to determine if an analysis can improve the detection of early infections.

This study included 129 adults who underwent LT from January 2013 to December 2013 at a single center institution. These patients had no symptoms or signs of infection before transplantation. Blood, urine, and sputum cultures were collected from patients the day before transplantation, and there was no evidence of infection. Bacterial cultures were performed every day from the day of transplantation to 1 week after LT. Samples of blood, urine, sputum, and drained ascites were collected for culture and identification of the isolated microorganisms. Cultures and microorganism identification were conducted in accordance with standard microbiological procedures [8].

In this study, cultures harboring bacteria were identified as positive cultures, while cultures that did not contain bacteria were identified as negative cultures. In addition, after the operation, infection was defined as the occurrence of fever above 38℃, use of antibiotics, positive culture and a sudden increase in leukocyte count or C-reactive protein level to 2–3 times the upper limit of normal.

Among the 129 LT recipients, 99 (76.7%) were men and 30 (23.3%) were women. Eighty-eight patients (68.2%) received living donor LT (LDLT), and 41 patients (31.8%) received deceased donor LT (DDLT). The model for end-stage liver disease (MELD) score of 87 patients (67.4%) was less than 20 points, and the MELD score of 42 patients (32.6%) was more than 20 points. Forty-seven patients (36.4%) were classified as Child-Pugh Class A, 31 (24%) were B, and 51 (39.5%) were C. Thirty-five patients (27.1%) had positive cultures, and 94 patients (72.9%) had negative cultures (Table 1). The following factors were significantly different between the positive and negative culture groups: LDLT vs. DDLT (odds ratio [OR], 3.269; 95% confidence interval [CI], 1.45–7.37; P = 0.003), MELD score (OR, 4.364; 95% CI, 1.92–9.93; P < 0.001), and Child-Pugh classification (OR, 2.71; 95% CI, 1.22–6.00; P = 0.007).

Table 1
Demographic and clinical characteristics of liver transplantation recipients

• Click for larger image

Of 129 patients, 30 patients had at least one positive preoperative blood, urine, and sputum culture, and 26 patients had antibiotics before the transplantation. Fourteen patients received LDLT and 16 patients received DDLT. The most common cause was spontaneous bacterial peritonitis in 11 patients followed by urinary tract infection in 3 patients, bacteremia in 2 patients, pneumonia in 2 patients and other infections in 8 patients. Fourteen had a negative culture for 1 week after LT, and 16 patients had a positive culture.

Table 2 shows the types of bacteria detected in positive cultures. More than 1 type of bacteria were detected in 1 patient.

Table 2
Bacterial and fungal strains in the first week after liver transplantation

• Click for larger image

Fig. 1 shows the number of positive culture samples. The results of blood, urine, sputum, and ascites culture performed for 1 week after LT are presented. Sputum culture was more likely to be positive than any other kind of culture.

Fig. 1
Number of culture positive samples. POD, postoperative day.

• Click for larger image

Fig. 2 shows the incidence (panel A) and proportion (panel B) of infections occurring up to 12 weeks after LT. In cultures performed within 1 week of LT, there was a statistically significant difference in the incidence rate between the positive culture group and the negative culture group at 1–2 weeks, but the occurrence of infection decreased within 4 weeks and the incidence increased again from 4 weeks to 12 weeks. Since patients are normally discharged after 3 weeks, it is not clear whether infections occurring after 4 weeks are linked with a positive culture obtained within the first week after LT. This is evident in Table 3. The incidence of infection at 4–12 weeks was higher in the positive culture group than the negative culture group (n = 12 [28.6%] vs. n = 10 [12.8%]), and most infections in both groups were associated with surgical complications (83% vs. 73%). Also, cholangitis caused by biliary stricture occurred more frequently in the negative culture group.

Fig. 2
(A) Infection events occurring between culture positive and culture negative. (B) Proportion of infection event after liver transplantation.

• Click for larger image

Table 3
Causes of infection at 4–12 weeks after liver transplantation

• Click for larger image

In this study, statistically significant differences between the positive culture group and the negative culture group were noted in LDLT and DDLT, MELD score, and Child-Pugh classification. However, the number of patients with positive and negative cultures within the first week after transplantation did not affect the number of infections within 1 month. Even though bacteria were most commonly seen in sputum samples in this study, surgical complications (biliary stricture or bile leakage) were more common sources of infection than was pneumonia after 4 weeks.

LT recipients are generally given immunosuppressants. Also, pretransplant conditions are variable and some factors cannot be controlled. Uncontrolled pretransplant variables such as underlying liver disease, lymphocyte mismatch, and a history of surgery may be associated with posttransplant conditions such as the presence of infectious and noninfectious diseases. These pre- and posttransplant conditions affect posttransplant management and length of stay in intensive care unit and hospital. A pretransplant MELD score of more than 25 was linked with patient and graft survival [10]. A high MELD score was indicative of a poor outcome. Our data revealed that 6 of the spontaneous bacterial peritonitis and 2 of the bacteremia were detected preoperative culture in DDLT patient and that patients with positive cultures tended to have higher MELD scores than those with negative cultures. This, suggests that patients who receive DDLT and have a high MELD score are more susceptible to infection during the preoperative period and within the first week after LT.

Posttransplant infections occurring after 1 month of surgery are primarily attributable to surgical and technical complexity, wound infection, urinary tract infection, catheter-related infection, and pneumonia [11]. Opportunistic infections are generally uncommon during the early period after transplantation, since the full effect of immunosuppression is not reached surgery-related infection is most common in the early period after LT. Our data revealed that both the positive and negative culture groups after LT experienced a gradual decrease in the number of infection events. Therefore, it is important to consider the number of cultures immediately after LT in patients without any evidence of pretransplantation infection. However, in patients with a high MELD score, who receive DDLT, and who develop infectious episodes in the pretransplant period, it is dangerous to reduce the number of serial culture tests for infection monitoring.

A limitation of our study is a retrospective study that included small number of LT patients. Another limitation of our study is that we could not identify the operational events and techniques that may have influenced surgical complications because of the retrospective nature of the study design. We also could not evaluate the biliary complexes in living donors compared to those in deceased donors. The biliary system could be a critical factor in biliary complications other than infection. A future large prospective study could provide more informative data.

In conclusion, we suggest reducing the number of culture tests performed in patients immediately after LT because full immune suppression is not reached and bacterial culture results at 1 week after LT are unlikely to be linked to infection. However, in patients with high MELD scores, who receive DDLT, or who have evidence of preoperative infection, the option of reducing the number of culture test should be seriously reconsidered.