Indian J Gastroenterol DOI 10.1007/s12664-014-0462-2
ORIGINAL ARTICLE
Readmission after living donor liver transplantation: Predictors, causes, and outcomes Raghavendra Nagaraja & Naimish Mehta & Vinay Kumaran & Vibha Varma & Sorabh Kapoor & Samiran Nundy
Received: 8 November 2013 / Accepted: 31 March 2014 # Indian Society of Gastroenterology 2014
Abstract Background Complications following liver transplantation requiring readmission may be serious and potentially life threatening. Most reports on readmission have been about after deceased donor liver transplantation (DDLT). We hypothesized that readmission after living donor liver transplantation (LDLT) is due to different reasons and analyzed our experience. Methods We retrospectively analyzed the records of 172 consecutive patients who underwent liver transplantation at our institute between January 2010 and June 2012. The primary outcome measure was readmission. We classified readmission into early (3 months). Results The study population was 140 after excluding pediatric patients (12), DDLT recipients (2), and those who died during the index admission (18). Their median age was 42 years, and there were 117 males and 23 females. Thirtyeight patients were readmitted (56 episodes) after LDLT. There were 35 early and 21 late readmission episodes. The most common cause for early readmissions was infection (46 %) and that for late readmissions was biliary stricture (62 %). On univariate analysis, pretransplant portal vein
The study was conducted at Sir Ganga Ram Hospital, New Delhi. R. Nagaraja (*) Department of Surgical Gastroenterology and Liver Transplantation, Manipal Hospital, 98, HAL Airport Road, Bengaluru 560 017, India e-mail: [email protected] N. Mehta : S. Nundy Department of Surgical Gastroenterology and Liver Transplantation, Sir Ganga Ram Hospital, Rajinder Nagar, New Delhi 110 060, India V. Kumaran : V. Varma : S. Kapoor Department of HPB and Liver Transplantation, Reliance Kokilaben Dhirubhai Ambani Hospital, Lokhandwala Complex Road, Andheri (West), Mumbai 400 061, India
thrombosis, more than one bile duct in the liver graft, revised arterial anastomosis or two arteries in the graft, and higher serum alkaline phosphatase levels at discharge were significantly associated with readmission. Readmission was also significantly associated with a higher overall mortality than non-readmission in which there was no mortality. Conclusion Pretransplant portal vein thrombosis, more than one bile duct in the liver graft, revision of the arterial anastomosis or two arteries in the graft, and higher serum alkaline phosphatase levels at discharge were significantly associated with readmission. Infective and biliary complications were the commonest causes of early and late readmission after LDLT. Keywords Biliary complication . Infection . Liver transplantation . Portal vein thrombosis . Readmission
Introduction Readmission after major abdominal surgery has been a major health-care subject of debate of late [1–5]. This is particularly so in Western countries where insurance companies play a major role in payment of health-care costs. However, probably the more important issue is the cause for readmission, medical and social problems faced by the patient and family members because of this, as well as the financial implications to individuals in countries where most patients pay out of their pocket for their health-care. Readmission rates after abdominal surgeries have also been suggested to be markers of healthcare quality in recent studies [5]. Liver transplantation, being a difficult abdominal operation, may have complications following the procedure that require readmission, which may be serious and potentially life threatening. The literature on readmission following liver transplantation is limited. We found only a few published articles on this subject, and these mainly involved deceased
Indian J Gastroenterol
donor liver transplantation (DDLT) [6, 7]. Living donor liver transplantation (LDLT) which is probably more complex technically is likely to be associated with different readmission rates and have different predictive factors for readmission. We therefore performed a study at our institution which has predominantly LDLT experience to examine the rates and risk factors of readmission after the procedure.
Methods We retrospectively analyzed, from a prospectively maintained database, the details of 172 consecutive patients who underwent liver transplantation at our institute between January 2010 and June 2012. We included all adult (>12 years) patients who underwent LDLT and who were discharged after the transplantation (n 140). Pediatric recipients (n 12), patients undergoing DDLT (n 2), and those who died during the index admission (n 18) after the transplant were excluded. The patients were followed postoperatively for at least 6 months (median 30 months; range 6–35 months). We studied their demographic factors and clinical presentation (preoperative variables), intraoperative variables, postoperative stay and complications, causes for readmission, and outcome after readmission. The primary outcome measure was readmission. For the purpose of analysis, we divided the patients into two groups and compared patients who were readmitted with those who were not. We also subclassified readmission into early (3 months). All categorical variables were expressed as number with percentages and continuous values as medians with range. The categorical variables were analyzed using Fisher’s exact test and the continuous variables, with the Mann-Whitney U test. The statistical analysis was performed using GraphPad Software, Inc (USA).
Results A total of 140 patients were included in the study. Their median age was 42 years (range 20–66 years), and there were 117 males and 23 females. Thirty-eight patients were readmitted (56 episodes) after LDLT. There were 35 (28 patients) early readmissions, 21 (15 patients) late readmissions, and 5 patients had both early and late readmissions. The demographic characteristics and presentation (Table 1) were comparable between the patients who were readmitted and those who were not except for portal vein thrombosis (PVT). PVT was significantly more frequent in patients who were readmitted (26 % vs. 3 %; p 0.0001). The donor characteristics in the two groups were similar (Table 2). In total, 13 patients had preoperative imaging-diagnosed portal vein thrombosis. Two of them were found to have a recanalized portal vein at the time of surgery. Seven patients
Table 1 Comparison of preoperative variables in the recipient groups Readmitted (38)
Non-readmitted p-value (102)
Age in years, median (range)
49.5 (24–66)
48 (20–66)
Male gender (%)
34 (89)
83 (81)
BMI median (range)
23.4 (18.5–29.4) 26 (16.9–31.2)
Pretransplant stay (days), median (range) CTP score, median (range)
2 (1–22)
2 (1–40)
11 (6–14)
0.34 0.31 0.61 0.9
11 (5–14)
0.8
MELD score, median (range) 22 (13–35)
21 (6–45)
0.6
Hepatocellular carcinoma, 8 (21) n (%) Portal vein thrombosis, n (%) 10 (26)
18 (18)
0.6
Chronic liver disease, n (%)
37 (97)
90 (88)
0.31
Comorbidity, n (%)
14 (37)
32 (31)
0.54
3 (3)
0.0001
Cause of liver disease Hepatitis B
6
Hepatitis C
14
28
7
31
Encephalopathy
26
61
Hepatorenal syndrome
10
24
9
17
5
25
15
42
Ethanolic
16
Decompensation
Ascites SBP Gastrointestinal bleed
BMI body mass index, CTP Child-Turcotte-Pugh score, MELD model for end-stage liver disease, SBP spontaneous bacterial peritonitis
had grade 1 portal vein thrombosis, Yerdel et al. [8], while the more severe grades of 2, 3, and 4 were seen in two patients each. In patients with grades 1 and 2 (n 9) PVT, portal vein thrombectomy was successful and provided adequate inflow. In patients with grade 3 PVT, thrombectomy was successful in one, and the other patient required a jump graft from SMV. Thrombectomy was not successful in patients with grade 4 thrombosis, cavoportal hemitransposition was done in one, and the other underwent renoportal transposition. Of the several intraoperative variables analyzed (Table 3), readmitted patients more frequently had >1 bile duct in the Table 2 Donor characteristics of readmitted vs. non-readmitted recipients Readmitted (38) Non-readmitted (102) p-value Age (years) 35.6 (19–54) Gender (male) 27 BMI 24.5 (17.8–32) LAI 8.7 (1–20) Comorbidity 9 Hospital stay (days) 8 (7–13)
33.4 (19–59) 55 24.2 (18.7–35.2) 10.2 (1–25) 20 8 (5–14)
BMI body mass index, LAI liver attenuation index
0.26 0.08 0.68 0.10 0.64 0.65
Indian J Gastroenterol Table 3 Intraoperative and postoperative variables in the recipient groups Right lobe Ascites (mL) Operative time (min) Warm ischemia time (min), median (range) Cold ischemia time (min), median (range) Graft weight, median (range) GRWR, median (range) >1 hepatic duct in graft, n (%) Roux-en-Y biliary anastomosis Double arterial anastomosis/arterial revision, n (%) Blood transfusion (units), median (range) Hospital stay (days), median (range) ICU stay (days), median (range) Complications (all), n (%) Severe complicationsa, n (%)
GRWR graft recipient weight ratio a
Clavien grades 3 and 4 [9]
Investigations at discharge Total leukocyte count (per cumm) Serum creatinine (mg/dL) Serum bilirubin (mg/dL) SGOT (IU/mL) SGPT (IU/mL) Alkaline phosphatase (IU/L) Tacrolimus level
donor liver graft (68 % vs. 44 %; p 0.013) and underwent either double arterial anastomosis (n 4) or arterial revision (n 3) or single arterial reconstruction (n 1) with two arteries in the graft (21 % vs. 8 %; p 0.038). The other variables were not significantly different between the two groups. Eight right-lobe recipients underwent Roux-en-Y biliary anastomosis at the time of transplantation. Two patients had a preoperative diagnosis of overlap syndrome (autoimmune liver disease and primary sclerosing cholangitis) and underwent a planned Roux-en-Y anastomosis. One patient with extensive (grade 4) portomesenteric thrombosis with encased bile duct required enteric anastomosis. In the remaining five patients, Roux-en-Y anastomosis was constructed in addition to duct-to-duct anastomosis to accommodate the extra bile ducts in the liver graft. The postoperative intensive care unit (ICU) and hospital stay (Table 3), frequency of complications, and most of the investigations at discharge were similar in the two groups. Only serum alkaline phosphatase (ALP) at discharge was significantly higher in readmitted patients (199 vs. 118 IU/L; p 0.0012). The common specific complications were rejection (n 16), re-exploration for intraabdominal bleeding (n 11), and biliary complications (n 6). Respiratory complications including pneumonitis and collapse/effusion needing treatment were
Readmitted (38)
Non-readmitted (102)
p-value
35 1,750 676 48 (30–79) 78 (30–147)
99 2,000 679 46 (25–79) 79 (35–209)
0.34 0.51 0.92 0.67 0.79
687.5 (419–1150) 1 (0.76–1.52) 26 (68) 4 8 (21) 6 (0–13) 13 (9–102) 5 (2–23) 28 (74) 13 (34.2) 7,700 0.75 2.2 47.5 79 199 6.9
721 (454–1206) 1 (0.73–1.6) 45 (44) 4 8 (8) 6 (0–18) 12 (6–120) 5 (2–25) 71 (70) 27 (26.5) 7,600 0.7 1.6 42 72 118 6.4
0.44 0.56 0.013 0.21 0.038 0.86 0.15 0.92 0.68 0.4 0.079 0.27 0.21 0.84 0.93 0.0012 0.89
seen in 13 patients. Neurological complications including tremors, psychosis, foot drop, depression, seizures, proximal myopathy, cranial nerve palsies, and hemi/quadriparesis were encountered in 26 patients. Seven patients with intraabdominal collection (other than bile leak) required percutaneous drainage. Prolonged ventilation and ICU requirement were found in 11 patients. The other complications were ascitic fluid drainage from the wound, raised total leukocyte count needing continuation of antibiotics, and thrombocytopenia and dyselectrolytemia needing correction. The frequency of these complications, either all complications or severe ones (Dindo-Clavien grades 3 and 4 [9]), was not different in patients who were readmitted after transplantation, compared to those who were not readmitted (data not shown for individual complications due to lack of space). Most of these complications resolved before discharge without needing readmission. On analyzing their subsequent outcome, we found that readmitted patients had higher overall mortality than nonreadmitted ones (8 % vs. 0 %; p 0.01). All those who died fell into the early admission group, and their deaths were due to infection. The most common cause for early readmissions (Table 4) was infections (16, 46 %), and other common causes were acute rejection (6, 17 %) and bile leaks—biloma (4, 11 %). Rarer causes for early readmissions were renal dysfunction,
Indian J Gastroenterol Table 4 Causes of early and late readmission episodes (56) Early readmission Infection Intraabdominal collection CMV Mucormycosis Wound infection
35 16 6 5 2 2
Raised TLC Rejection Bile leak Renal dysfunction Others Dyselectrolytemia Anemia Complicated hernia Intracranial bleed Late readmission Biliary complication Chronic graft dysfunction/rejection Recurrence (HCV/HCC) Others Hernia SAIO Infection Gastrointestinal bleed
1 6 4 3 6 2 2 1 1 21 13 4 2 6 1 2 2 1
HCV hepatitis C virus, HCC hepatocellular carcinoma, SAIO subacute intestinal obstruction, TLC total leukocyte count, CMV cytomegalovirus a
More than one cause found in some, and few patients had more than one admission
obstructed hernia, dyselectrolytemia/anemia, and intracranial bleeding. The majority of late readmissions (Table 4) were due to biliary strictures (13, 62 %). The other causes for late readmissions were relatively less common and included chronic graft dysfunction, recurrence of liver disease, and incisional hernia. Rarer causes were respiratory infection, subacute obstruction, gastrointestinal bleeding, and myocardial infarction. Of the five patients who were readmitted both early and late, two had early readmissions for bile leak, none of whom have developed biliary stricture. They were admitted for incisional hernia and chest infection. In the same five patients, three had late readmission for biliary stricture, and their early readmissions were for acute rejection, cytomegalovirus (CMV) infection, and fungal infection of the wound.
Discussion Most studies concerning readmission after major surgeries have been published in recent years and have been from the
countries where insurance companies play a major role in the funding of health-care [1–7]. The objectives of these studies have been either to describe the cost burden due to readmission or to define the predictors for readmission. Some of these studies have suggested that readmission rates are markers of health-care quality, both of the surgeon and hospital [5]. The reports from countries where insurance companies do not play a dominant role are few, and the reasons may be different, as the patient has to pay for his/her health-care in private hospitals. This may result in lower readmission rates as both the doctor and patient adopt a more conservative approach towards readmission. Pereira et al. [6] recently published their experience on readmission after DDLT. In their 8-year experience involving 766 patients, they reported a 30-day readmission rate of 45 % with about 8.5 % of patients being discharged to institutional care from the hospital. They found that patients who were sicker before transplantation were more likely to be readmitted. In particular, a 90-day pretransplant hospital admission, low serum albumin and higher creatinine levels, and preoperative PVT correlate with readmission rates. In addition to these, postoperative complications and lower education levels increased the risk of readmission. They also reported a significantly lower survival in these readmitted patients. In an earlier publication, Shankar et al. [7] studied predictive factors of readmission in 208 patients after liver transplantation (six LDLT). They reported a 90-day readmission rate of 30 %, with a shorter posttransplant ICU stay having a negative correlation, and hepatitis C viral (HCV) end-stage liver disease correlates positively with readmission. In our study, in addition to looking into the factors predicting readmission after LDLT, we also analyzed the cause and outcome of such readmissions. The readmission rate in the present study after LDLT was 27.1 % (38 patients, 56 episodes). Of these, 28 patients had early (
ORIGINAL ARTICLE
Readmission after living donor liver transplantation: Predictors, causes, and outcomes Raghavendra Nagaraja & Naimish Mehta & Vinay Kumaran & Vibha Varma & Sorabh Kapoor & Samiran Nundy
Received: 8 November 2013 / Accepted: 31 March 2014 # Indian Society of Gastroenterology 2014
Abstract Background Complications following liver transplantation requiring readmission may be serious and potentially life threatening. Most reports on readmission have been about after deceased donor liver transplantation (DDLT). We hypothesized that readmission after living donor liver transplantation (LDLT) is due to different reasons and analyzed our experience. Methods We retrospectively analyzed the records of 172 consecutive patients who underwent liver transplantation at our institute between January 2010 and June 2012. The primary outcome measure was readmission. We classified readmission into early (3 months). Results The study population was 140 after excluding pediatric patients (12), DDLT recipients (2), and those who died during the index admission (18). Their median age was 42 years, and there were 117 males and 23 females. Thirtyeight patients were readmitted (56 episodes) after LDLT. There were 35 early and 21 late readmission episodes. The most common cause for early readmissions was infection (46 %) and that for late readmissions was biliary stricture (62 %). On univariate analysis, pretransplant portal vein
The study was conducted at Sir Ganga Ram Hospital, New Delhi. R. Nagaraja (*) Department of Surgical Gastroenterology and Liver Transplantation, Manipal Hospital, 98, HAL Airport Road, Bengaluru 560 017, India e-mail: [email protected] N. Mehta : S. Nundy Department of Surgical Gastroenterology and Liver Transplantation, Sir Ganga Ram Hospital, Rajinder Nagar, New Delhi 110 060, India V. Kumaran : V. Varma : S. Kapoor Department of HPB and Liver Transplantation, Reliance Kokilaben Dhirubhai Ambani Hospital, Lokhandwala Complex Road, Andheri (West), Mumbai 400 061, India
thrombosis, more than one bile duct in the liver graft, revised arterial anastomosis or two arteries in the graft, and higher serum alkaline phosphatase levels at discharge were significantly associated with readmission. Readmission was also significantly associated with a higher overall mortality than non-readmission in which there was no mortality. Conclusion Pretransplant portal vein thrombosis, more than one bile duct in the liver graft, revision of the arterial anastomosis or two arteries in the graft, and higher serum alkaline phosphatase levels at discharge were significantly associated with readmission. Infective and biliary complications were the commonest causes of early and late readmission after LDLT. Keywords Biliary complication . Infection . Liver transplantation . Portal vein thrombosis . Readmission
Introduction Readmission after major abdominal surgery has been a major health-care subject of debate of late [1–5]. This is particularly so in Western countries where insurance companies play a major role in payment of health-care costs. However, probably the more important issue is the cause for readmission, medical and social problems faced by the patient and family members because of this, as well as the financial implications to individuals in countries where most patients pay out of their pocket for their health-care. Readmission rates after abdominal surgeries have also been suggested to be markers of healthcare quality in recent studies [5]. Liver transplantation, being a difficult abdominal operation, may have complications following the procedure that require readmission, which may be serious and potentially life threatening. The literature on readmission following liver transplantation is limited. We found only a few published articles on this subject, and these mainly involved deceased
Indian J Gastroenterol
donor liver transplantation (DDLT) [6, 7]. Living donor liver transplantation (LDLT) which is probably more complex technically is likely to be associated with different readmission rates and have different predictive factors for readmission. We therefore performed a study at our institution which has predominantly LDLT experience to examine the rates and risk factors of readmission after the procedure.
Methods We retrospectively analyzed, from a prospectively maintained database, the details of 172 consecutive patients who underwent liver transplantation at our institute between January 2010 and June 2012. We included all adult (>12 years) patients who underwent LDLT and who were discharged after the transplantation (n 140). Pediatric recipients (n 12), patients undergoing DDLT (n 2), and those who died during the index admission (n 18) after the transplant were excluded. The patients were followed postoperatively for at least 6 months (median 30 months; range 6–35 months). We studied their demographic factors and clinical presentation (preoperative variables), intraoperative variables, postoperative stay and complications, causes for readmission, and outcome after readmission. The primary outcome measure was readmission. For the purpose of analysis, we divided the patients into two groups and compared patients who were readmitted with those who were not. We also subclassified readmission into early (3 months). All categorical variables were expressed as number with percentages and continuous values as medians with range. The categorical variables were analyzed using Fisher’s exact test and the continuous variables, with the Mann-Whitney U test. The statistical analysis was performed using GraphPad Software, Inc (USA).
Results A total of 140 patients were included in the study. Their median age was 42 years (range 20–66 years), and there were 117 males and 23 females. Thirty-eight patients were readmitted (56 episodes) after LDLT. There were 35 (28 patients) early readmissions, 21 (15 patients) late readmissions, and 5 patients had both early and late readmissions. The demographic characteristics and presentation (Table 1) were comparable between the patients who were readmitted and those who were not except for portal vein thrombosis (PVT). PVT was significantly more frequent in patients who were readmitted (26 % vs. 3 %; p 0.0001). The donor characteristics in the two groups were similar (Table 2). In total, 13 patients had preoperative imaging-diagnosed portal vein thrombosis. Two of them were found to have a recanalized portal vein at the time of surgery. Seven patients
Table 1 Comparison of preoperative variables in the recipient groups Readmitted (38)
Non-readmitted p-value (102)
Age in years, median (range)
49.5 (24–66)
48 (20–66)
Male gender (%)
34 (89)
83 (81)
BMI median (range)
23.4 (18.5–29.4) 26 (16.9–31.2)
Pretransplant stay (days), median (range) CTP score, median (range)
2 (1–22)
2 (1–40)
11 (6–14)
0.34 0.31 0.61 0.9
11 (5–14)
0.8
MELD score, median (range) 22 (13–35)
21 (6–45)
0.6
Hepatocellular carcinoma, 8 (21) n (%) Portal vein thrombosis, n (%) 10 (26)
18 (18)
0.6
Chronic liver disease, n (%)
37 (97)
90 (88)
0.31
Comorbidity, n (%)
14 (37)
32 (31)
0.54
3 (3)
0.0001
Cause of liver disease Hepatitis B
6
Hepatitis C
14
28
7
31
Encephalopathy
26
61
Hepatorenal syndrome
10
24
9
17
5
25
15
42
Ethanolic
16
Decompensation
Ascites SBP Gastrointestinal bleed
BMI body mass index, CTP Child-Turcotte-Pugh score, MELD model for end-stage liver disease, SBP spontaneous bacterial peritonitis
had grade 1 portal vein thrombosis, Yerdel et al. [8], while the more severe grades of 2, 3, and 4 were seen in two patients each. In patients with grades 1 and 2 (n 9) PVT, portal vein thrombectomy was successful and provided adequate inflow. In patients with grade 3 PVT, thrombectomy was successful in one, and the other patient required a jump graft from SMV. Thrombectomy was not successful in patients with grade 4 thrombosis, cavoportal hemitransposition was done in one, and the other underwent renoportal transposition. Of the several intraoperative variables analyzed (Table 3), readmitted patients more frequently had >1 bile duct in the Table 2 Donor characteristics of readmitted vs. non-readmitted recipients Readmitted (38) Non-readmitted (102) p-value Age (years) 35.6 (19–54) Gender (male) 27 BMI 24.5 (17.8–32) LAI 8.7 (1–20) Comorbidity 9 Hospital stay (days) 8 (7–13)
33.4 (19–59) 55 24.2 (18.7–35.2) 10.2 (1–25) 20 8 (5–14)
BMI body mass index, LAI liver attenuation index
0.26 0.08 0.68 0.10 0.64 0.65
Indian J Gastroenterol Table 3 Intraoperative and postoperative variables in the recipient groups Right lobe Ascites (mL) Operative time (min) Warm ischemia time (min), median (range) Cold ischemia time (min), median (range) Graft weight, median (range) GRWR, median (range) >1 hepatic duct in graft, n (%) Roux-en-Y biliary anastomosis Double arterial anastomosis/arterial revision, n (%) Blood transfusion (units), median (range) Hospital stay (days), median (range) ICU stay (days), median (range) Complications (all), n (%) Severe complicationsa, n (%)
GRWR graft recipient weight ratio a
Clavien grades 3 and 4 [9]
Investigations at discharge Total leukocyte count (per cumm) Serum creatinine (mg/dL) Serum bilirubin (mg/dL) SGOT (IU/mL) SGPT (IU/mL) Alkaline phosphatase (IU/L) Tacrolimus level
donor liver graft (68 % vs. 44 %; p 0.013) and underwent either double arterial anastomosis (n 4) or arterial revision (n 3) or single arterial reconstruction (n 1) with two arteries in the graft (21 % vs. 8 %; p 0.038). The other variables were not significantly different between the two groups. Eight right-lobe recipients underwent Roux-en-Y biliary anastomosis at the time of transplantation. Two patients had a preoperative diagnosis of overlap syndrome (autoimmune liver disease and primary sclerosing cholangitis) and underwent a planned Roux-en-Y anastomosis. One patient with extensive (grade 4) portomesenteric thrombosis with encased bile duct required enteric anastomosis. In the remaining five patients, Roux-en-Y anastomosis was constructed in addition to duct-to-duct anastomosis to accommodate the extra bile ducts in the liver graft. The postoperative intensive care unit (ICU) and hospital stay (Table 3), frequency of complications, and most of the investigations at discharge were similar in the two groups. Only serum alkaline phosphatase (ALP) at discharge was significantly higher in readmitted patients (199 vs. 118 IU/L; p 0.0012). The common specific complications were rejection (n 16), re-exploration for intraabdominal bleeding (n 11), and biliary complications (n 6). Respiratory complications including pneumonitis and collapse/effusion needing treatment were
Readmitted (38)
Non-readmitted (102)
p-value
35 1,750 676 48 (30–79) 78 (30–147)
99 2,000 679 46 (25–79) 79 (35–209)
0.34 0.51 0.92 0.67 0.79
687.5 (419–1150) 1 (0.76–1.52) 26 (68) 4 8 (21) 6 (0–13) 13 (9–102) 5 (2–23) 28 (74) 13 (34.2) 7,700 0.75 2.2 47.5 79 199 6.9
721 (454–1206) 1 (0.73–1.6) 45 (44) 4 8 (8) 6 (0–18) 12 (6–120) 5 (2–25) 71 (70) 27 (26.5) 7,600 0.7 1.6 42 72 118 6.4
0.44 0.56 0.013 0.21 0.038 0.86 0.15 0.92 0.68 0.4 0.079 0.27 0.21 0.84 0.93 0.0012 0.89
seen in 13 patients. Neurological complications including tremors, psychosis, foot drop, depression, seizures, proximal myopathy, cranial nerve palsies, and hemi/quadriparesis were encountered in 26 patients. Seven patients with intraabdominal collection (other than bile leak) required percutaneous drainage. Prolonged ventilation and ICU requirement were found in 11 patients. The other complications were ascitic fluid drainage from the wound, raised total leukocyte count needing continuation of antibiotics, and thrombocytopenia and dyselectrolytemia needing correction. The frequency of these complications, either all complications or severe ones (Dindo-Clavien grades 3 and 4 [9]), was not different in patients who were readmitted after transplantation, compared to those who were not readmitted (data not shown for individual complications due to lack of space). Most of these complications resolved before discharge without needing readmission. On analyzing their subsequent outcome, we found that readmitted patients had higher overall mortality than nonreadmitted ones (8 % vs. 0 %; p 0.01). All those who died fell into the early admission group, and their deaths were due to infection. The most common cause for early readmissions (Table 4) was infections (16, 46 %), and other common causes were acute rejection (6, 17 %) and bile leaks—biloma (4, 11 %). Rarer causes for early readmissions were renal dysfunction,
Indian J Gastroenterol Table 4 Causes of early and late readmission episodes (56) Early readmission Infection Intraabdominal collection CMV Mucormycosis Wound infection
35 16 6 5 2 2
Raised TLC Rejection Bile leak Renal dysfunction Others Dyselectrolytemia Anemia Complicated hernia Intracranial bleed Late readmission Biliary complication Chronic graft dysfunction/rejection Recurrence (HCV/HCC) Others Hernia SAIO Infection Gastrointestinal bleed
1 6 4 3 6 2 2 1 1 21 13 4 2 6 1 2 2 1
HCV hepatitis C virus, HCC hepatocellular carcinoma, SAIO subacute intestinal obstruction, TLC total leukocyte count, CMV cytomegalovirus a
More than one cause found in some, and few patients had more than one admission
obstructed hernia, dyselectrolytemia/anemia, and intracranial bleeding. The majority of late readmissions (Table 4) were due to biliary strictures (13, 62 %). The other causes for late readmissions were relatively less common and included chronic graft dysfunction, recurrence of liver disease, and incisional hernia. Rarer causes were respiratory infection, subacute obstruction, gastrointestinal bleeding, and myocardial infarction. Of the five patients who were readmitted both early and late, two had early readmissions for bile leak, none of whom have developed biliary stricture. They were admitted for incisional hernia and chest infection. In the same five patients, three had late readmission for biliary stricture, and their early readmissions were for acute rejection, cytomegalovirus (CMV) infection, and fungal infection of the wound.
Discussion Most studies concerning readmission after major surgeries have been published in recent years and have been from the
countries where insurance companies play a major role in the funding of health-care [1–7]. The objectives of these studies have been either to describe the cost burden due to readmission or to define the predictors for readmission. Some of these studies have suggested that readmission rates are markers of health-care quality, both of the surgeon and hospital [5]. The reports from countries where insurance companies do not play a dominant role are few, and the reasons may be different, as the patient has to pay for his/her health-care in private hospitals. This may result in lower readmission rates as both the doctor and patient adopt a more conservative approach towards readmission. Pereira et al. [6] recently published their experience on readmission after DDLT. In their 8-year experience involving 766 patients, they reported a 30-day readmission rate of 45 % with about 8.5 % of patients being discharged to institutional care from the hospital. They found that patients who were sicker before transplantation were more likely to be readmitted. In particular, a 90-day pretransplant hospital admission, low serum albumin and higher creatinine levels, and preoperative PVT correlate with readmission rates. In addition to these, postoperative complications and lower education levels increased the risk of readmission. They also reported a significantly lower survival in these readmitted patients. In an earlier publication, Shankar et al. [7] studied predictive factors of readmission in 208 patients after liver transplantation (six LDLT). They reported a 90-day readmission rate of 30 %, with a shorter posttransplant ICU stay having a negative correlation, and hepatitis C viral (HCV) end-stage liver disease correlates positively with readmission. In our study, in addition to looking into the factors predicting readmission after LDLT, we also analyzed the cause and outcome of such readmissions. The readmission rate in the present study after LDLT was 27.1 % (38 patients, 56 episodes). Of these, 28 patients had early (
