Long-Term Impact of Liver Transplantation on Respiratory Function and Nutritional Status in Children and Adults With Cystic Fibrosis

Authors


Corresponding authors: J. K. Dowman, j.k.dowman@bham.ac.uk and PN Newsome, p.n.newsome@bham.ac.uk

Abstract

Early liver transplant (LT) has been advocated for patients with cystic fibrosis liver disease (CFLD) and evidence of deterioration in nutritional state and respiratory function to prevent further decline. However, the impact of single LT on long-term respiratory function and nutritional status has not been adequately addressed. We performed a retrospective analysis of the outcomes of 40 (21 adult/19 pediatric) patients with CFLD transplanted between 1987 and 2009 with median follow-up of 47.8 months (range 4–180). One and five-year actuarial survival rates were 85%/64% for adult and 90%/85% for pediatric LT cohorts, respectively. Lung function remained stable until 4 years (FEV1% predicted; pretransplant 48.4% vs. 45.9%, 4 years posttransplant) but declined by 5 years (42.4%). Up to 4 years posttransplant mean annual decline in FEV1% was lower (0.74%; p = 0.04) compared with the predicted 3% annual decline in CF patients with comorbidity including diabetes. Number of courses of intravenous antibiotics was reduced following LT, from 3.9/year pretransplant to 1.1/year, 5 years posttransplant. Body mass index was preserved posttransplant; 18.0 kg/m2 (range 15–24.3) pretransplant versus 19.6 kg/m2 (range 16.4–22.7) 5 years posttransplant. In conclusion, LT is an effective treatment for selected patients with cirrhosis due to CFLD, stabilizing aspects of long-term lung function and preserving nutritional status.

Abbreviations: 
CF

cystic fibrosis

CFLD

cystic fibrosis liver disease

cGFR

calculated glomerular filtration rate

CKD

chronic kidney disease

DIOS

distal intestinal obstruction syndrome

FEV1

forced expiratory volume in 1 second

FVC

forced vital capacity

IV

intravenous

LT

liver transplant

MELD

model for end-stage liver disease

PA

Pseudomonasaeruginosa

PHT

portal hypertension

PELD

pediatric end-stage liver disease

SDS

standard deviation score

Introduction

Cystic fibrosis (CF) is an autosomal recessive multisystem disorder, primarily affecting the lungs, but also pancreas, gastrointestinal tract and liver (1). With increased life expectancy, cystic fibrosis liver disease (CFLD) has emerged as a significant cause of morbidity and mortality (2–5). Approximately 3–5% of CF individuals develop portal hypertension due to severe cirrhosis, with a median age of diagnosis of 10–11 years, and 90% being diagnosed before 20 years of age (6). Liver transplantation is an established treatment for patients with CFLD (7–9), although there is uncertainty on the optimal indications and timing of transplant in this group. Some advocate preemptive LT when faced with declining lung function and muscle mass in the absence of overt liver synthetic failure, so as to arrest further decline (9). However, although favorable effects of liver transplantation on nutritional status have been reported (7,10), recent series have suggested that LT may in fact hasten the decline in lung function (7). With the exception of the series by Nash et al. (7), most other reports are relatively small in size and have limited follow-up. Thus, it has been difficult to determine the long-term outcome of liver transplantation in adult and pediatric patients with CF, as well as the impact on other organ systems.

The purpose of this study was to review the indications and outcomes of liver transplantation in adults and children with CFLD who underwent liver transplantation in the two tertiary referral liver transplant centers in Birmingham, UK. Our particular focus was the impact of liver transplantation on long-term respiratory function and nutritional status.

Subjects and Methods

A retrospective review was carried out of all patients who underwent liver transplantation for CFLD, either in isolation or in combination with other organs, at Birmingham Queen Elizabeth Hospital (QEH) Liver Unit and Birmingham Children's Hospital (BCH) between 1987 and 2009. Pediatric patients were classified as those who underwent liver transplantation up to the age of 16.5 years. Care of the majority of patients transplanted at BCH was transferred to the adult services at QEH at some stage after their 16th birthday. All of the single LTs were performed in Birmingham but three of the four triple transplant procedures were performed at another UK center.

Data were collected from case notes and/or databases and included demographic details, indication for liver transplantation, transplant details, survival, respiratory microbiology and pulmonary function, nutritional status and posttransplant complications including prevalence of renal impairment, diabetic status and distal intestinal obstruction syndrome (DIOS). Data time points included time of assessment/pretransplant, at 3, 6, and 12 months posttransplant and yearly thereafter until death or 5 years posttransplant. Although the transplant assessment and prioritization on the waiting list of many patients predated the introduction of the Pediatric End-Stage Liver Disease and Model for End-Stage Liver Disease (MELD) scoring systems, these were calculated retrospectively where possible.

In adults, spirometry was assessed using forced expiratory volume (FEV1) and forced vital capacity (FVC) and expressed as percentage of the predicted value. In children and adolescents, lung volumes and standing heights are not linearly related and thus lung function values in our pediatric cohort could only be compared accurately by employing standard deviation or z-scores in which a z-score of 0 is equivalent to the normal population median or 50th percentile (11).

In the adult cohort, nutritional status was expressed as median body mass index, calculated as weight (kg)/height (m)2. Dry weights were estimated in patients with ascites (12). In the pediatric cohort, height and weight were recorded and expressed as a standard deviation score (SDS or z-score), in which an SDS of 0 is equivalent to the normal population median or 50th percentile. Renal function was expressed as calculated glomerular filtration rate (cGFR) in mL/min/1.73 m2 using the Cockcroft–Gault Formula in adults and Counahan–Barratt Formula (height in centimeters multiplied by a constant of 40 and divided by creatinine in μmol/L) in children. Chronic renal impairment was defined as cGFR <60 mL/min/1.73 m2 (chronic kidney disease stage ≤3).

Statistical analysis included descriptive statistics and Wilcoxon Signed Rank Test to analyze significance of changes. The changes in FEV1% predicted and FVC% predicted over time were analyzed using generalized estimating equations. The patient was set and the subject variable, with the time point at which each measurement was taken used as the within-subject variable. An unstructured correlation matrix was used, and the time of the measurement was included as the covariate in each model. For comparison of the annual decline in FEV1% predicted in our cohort compared with that reported in the literature for nontransplanted patients, the gradient of our cohort was compared to the reported gradient of −3, using a one-sample t-test.

Results

Indication for LT

As detailed in Figure 1, 36 adults and 27 children with CFLD were assessed for transplant, of whom 35 underwent sole liver transplantation and 5 underwent combined liver, double lung and heart (n = 4) or liver/double lung (n = 1) transplantation. Pretransplant patient demographics, including indication for transplantation and MELD scores are shown in Table 1. The majority of adult patients were transplanted in accordance with standard criteria of liver decompensation, with only a few being transplanted due to portal hypertensive bleeding. In the pediatric cohort, 11 patients underwent preemptive liver transplantation due to deteriorating lung function in the setting of advanced compensated cirrhosis, with the intention of arresting the decline in lung function. Of the patients assessed but not transplanted, 5 died before reaching transplantation, and the remainder were considered to have either too early- or late-stage disease to warrant surgery. The presence of the ET12 strain of Burkholderia cenocepacia is now considered an absolute contraindication to transplantation due to its association with poor outcome. However, transplantation is more frequently contraindicated due to a combination of relative factors, such as FEV1 <40–50% predicted and poor nutritional status.

Figure 1.

Assessment, listing and transplant details for patients transplanted for CFLD between 1987 and 2009. (A) Adult and (B) Pediatric cases. IFaLD = intestinal failure-associated liver disease.

Table 1.  Demographics and indication for transplant in patients transplanted for CFLD (1987–2009)
 Adult liver or liver/lung transplant (n = 17)1Pediatric liver transplant (n = 19)
  1. 1Data on transplant indication unavailable for 1 adult patient.

  2. 2Decompensated liver disease: hepatic synthetic dysfunction included hypoalbuminemia, coagulopathy, ascites, jaundice, poor/deteriorating nutritional status, encephalopathy.

Sex (M/F)13/411/8
Age (years), median (range)23.2 (16.7–34.9)11.8 (9.5–16.5)
MELD score, median11 
Indication for liver transplantation:  
1. Decompensated liver disease213 (76%)8 (42%)
2. Severe progressive portal hypertension (PHT) ± intractable variceal bleeding3 (18%)0
3. Progressive compensated cirrhosis/PHT, with early liver transplantation when lung function is preserved011 (58%)
Waiting time (weeks), median (range)  
Liver/lung68 
Liver15 (4–112)17 (0.5–33.5)

Survival after LT for CFLD

Four adult patients underwent triple transplantation, of whom three died with a median survival of 13 days (0–66 days) (Figure 2). The remaining patient required an additional renal transplant at 18 years but was still alive at 20 years posttransplant. The single liver/double lung recipient remains well at 3 years. Median follow-up time for the 17 adult patients undergoing sole liver (n = 16), and liver/lung (n = 1) transplantation was 47.8 (range 4–180) months. Eleven patients remain alive, with 1- and 5-year actuarial survival rates of 85% and 64% (Figure 3), which are comparable with the rates for adult liver transplantation in our center during the same period (82% and 73%). Early deaths at 15 and 33 days were the result of Burkholderia cenocepacia-related pneumonia and sepsis/DIOS, respectively. The four late deaths occurred at 2, 3.7, 6.3 and 8.8 years and were from CF-related respiratory complications.

Figure 2.

Survival following liver transplantation for CFLD between 1987 and 2009.

Figure 3.

Lung function in adults and children transplanted for CFLD. (A) and (B) Pretransplant lung function in adult single liver recipients was relatively poor with a median predicted FEV1 of 48.4% and predicted FVC of 69.3%. Lung function remained stable until 4 years posttransplant but started to reduce by 5 years. (C) and (D) Pediatric lung function trends showed improved median FEV1/FVC z-scores during the first year posttransplant (p = 0.5/0.6), stabilization in the second year, but with further deterioration by 5 years posttransplant (p = 0.01/0.03) compared with baseline. Some initial stabilization, particularly FVC, may reflect resolution of effects of severe portal hypertension (pulmonary edema, intrapulmonary shunting, organomegaly, ascites). In these box and whisker plots, boxes illustrate the 25th and 75th percentiles and whiskers refer to one standard deviation above and below the mean of the data. Boxes without whiskers indicate that these data points lie within the box.

12/19 pediatric recipients remain alive with overall survival 63% (follow-up range 29–162 months), and 1- and 5-year actuarial survival of 90% and 85%. Two early deaths at 22 days and 9 months resulted from sepsis and multiorgan failure following hepatic artery and hepatic vein thromboses, respectively. Apart from one death due to a cerebrovascular event at 15.7 years (6.25 years after retransplantation), the four remaining late deaths at 3.3, 6.6, 6.9 and 6.9 years were all due to respiratory complications of CF.

Lung function after liver transplantation

Pretransplant lung function in the adult cohort was relatively poor, with a median FEV1 of 48.4% predicted and FVC of 71.4% predicted. FEV1 remained unchanged until 4 years (FEV1 pretransplant 48.4% vs. 45.9% predicted at 4 years posttransplant; p = 0.499, 95% CI =−2.88, 1.40) but started to decline by 5 years (42.4%) (Figure 3(A)). The mean annual decline in FEV1% predicted of −0.74% up to 4 years posttransplant was significantly improved (p = 0.042) compared with the reported decline for a CF patient without liver transplantation of 3% (13–15) (Figure 4). FVC% predicted posttransplant was seen to decline at an annual rate of 2.93%, although the 4-year value (64%) was not significantly different to the pretransplant value (p = 0.083 [95% CI =−6.25, 0.39]). An FEV1 of below 50% predicted did not result in a worse outcome compared to those patients with an FEV1 of >50% predicted (one death in each subgroup) although this was based on small numbers. Lung transplantation following liver transplantation has not yet been performed in any of our patients, although one died while undergoing assessment for lung transplantation and another is currently on an active lung transplant waiting list.

Figure 4.

Change in FEV1% and FVC% predicted in adults for first 4 years posttransplant. Changes in (A) FEV1% and (B) FVC% predicted in adult single liver recipients for up to 4 years posttransplant and comparison with the 3–5% annual decline reported in nontransplanted CF adults. The annual decline in FEV1% of 0.74% predicted in our cohort is significantly improved (p = 0.042) compared with an estimated 3% annual decline.

In the pediatric cohort, median (min, max) pretransplant percentage of predicted FEV1 and FVC were 78% (51%, 110%) and 89.5% (56%, 115%), respectively, which was higher than in the adult cohort. Median FEV1 and FVC z-scores pretransplant were −1.36 and −0.84, respectively. At 6 months posttransplant, this had improved to −1.07 and −0.87 (p = 0.79 and 0.33, respectively) although this did not achieve statistical significance (Figure 3B). By 24 months median FEV1 z-scores (−1.43) had deteriorated, while median FVC z-scores (−0.57) continued to improve slowly suggesting a degree of stabilization. At 5 years posttransplant, both FEV1 and FVC z-scores had deteriorated significantly from baseline (p = 0.01/0.03, respectively). Four children died from respiratory failure at 3.3, 6.6, 6.9 and 6.9 years posttransplant.

Bacterial colonization and antibiotic requirement after liver transplantation

Data on pretransplant chronic infection obtained from sputum culture were available in 19/21 (90%) of the adult cohort and in all children. A total of 17/19 adults were infected with Pseudomonas aeruginosa (PA). Two had multiresistant strains of PA but this did not portend a poorer outcome. One early death at 15 days posttransplant was caused by ‘cepacia syndrome’ in a patient with pretransplant infection with the ET12 strain of Burkholderia cenocepacia (formerly genomovar III). Requirement for intravenous antibiotic courses to treat pulmonary exacerbations reduced from an average of 3.9/year pretransplant, to 1.5/year at 1 year and 1.1/year at 5 years posttransplant. A total of 18/19 children had preoperative chronic PA infection, of whom only one had a multiresistant strain. This patient died at 22 days from an unrelated cause (HAT) although sepsis was a contributory factor. No Burkholderia species were isolated from any pediatric patients pretransplant and there were no early postoperative deaths directly from respiratory causes.

Nutritional status after liver transplantation

Median body mass index of adult patients was maintained from pretransplant values of 18.0 kg/m2 (range 15–24.3, adjusted for ascites), to 18.7 kg/m2 at 1 year (range 14.3–22.2) and 19.6 kg/m2 at 5 years posttransplant (range 16.4–22.7) (Figure 5A). In the pediatric cohort, weight and height SDS were significantly below normal at all time points (Figures 5B and C). During the first 6 months posttransplant, median weight and height z-scores dropped significantly from −0.85 and −1.18 to −1.15 and −1.63, respectively (p = 0.003 and 0.002). Both then stabilized until 12–24 months posttransplant, followed by further deterioration up to 5 years posttransplant. A total of 8/19 patients received supplemental enteral nutrition via nasogastric tube (n = 5) or gastrostomy (n = 3).

Figure 5.

Nutritional trends in adults and children transplanted for CFLD. (A) Adult single liver recipients showed stabilization in body mass index at 1 and 5 years post liver transplantation. Low body mass index at liver transplantation did not influence overall survival. (B) and (C) In the pediatric cohort, weight and height SDS were significantly below normal at all time points, with a clear fall in median SDS for both weight and height at 6 months, with no evidence of recovery from this up to 60 months. A total of 8/19 received nutritional support via nasogastric tube (5) or gastrostomy (3) In these box and whisker plots, boxes illustrate the 25th and 75th percentiles and whiskers refer to one standard deviation above and below the mean of the data. Boxes without whiskers indicate that these data points lie within the box.

Renal function after liver transplantation

Renal function declined posttransplant in adult patients with a prevalence of renal impairment of 13% at 2 years, rising to 23% at 5 years posttransplant (Figure 6A). However, this compares favorably with the decline in renal function observed in our general LT population, where prevalence of cGFR <60 mL/min at 5 years was 57% (16). Two adults required renal replacement therapy, including the single triple transplant survivor who underwent a live donor renal transplant at 18 years posttransplant, and one single liver recipient who required dialysis after 43.5 months.

Figure 6.

(A) Prevalence of chronic renal failure (defined by cGFR <60 mL/min for >3 months or need for renal replacement therapy) posttransplant in adult cohort and (B) Median cGFR trends in children transplanted for CFLD. Adult single liver recipients showed increased prevalence of renal impairment but this was lower than observed in the non-CFLD transplant population in our center. Two patients required renal replacement therapy. In the pediatric cohort, a significant fall in median cGFR was observed in the first 6 months posttransplant (p < 0.05), with stabilization thereafter. This trend was similar to that observed in the non-CFLD transplant population. Two patients required renal replacement therapy. In these box and whisker plots, boxes illustrate the 25th and 75th percentiles and whiskers refer to one standard deviation above and below the mean of the data. Boxes without whiskers indicate that these data points lie within the box.

Renal function was normal at time of transplant for all pediatric patients, with median cGFR 120.9 mL/min/1.72 m2. There was a significant fall in median cGFR in the first 6 months posttransplant to 64 mL/min/1.72 m2, p ≤ 0.0001, followed by stabilization (Figure 6B). By 2 years posttransplant, a significant and sustained recovery of median cGFR to within the normal range (77 mL/min/1.72 m2) was observed (p = 0.02). Only 1 patient progressed to chronic renal failure during the first 5 years posttransplant and 2 patients required renal replacement therapy at 7 and 9.5 years, respectively. The latter patient had been regrafted 2 years previously for de novo autoimmune cirrhosis.

Pancreatic insufficiency after LT

Data on pretransplant pancreatic insufficiency were available for 20/21 (95%) of the adult cohort and all pediatric patients. Exocrine pancreatic insufficiency was universal at the time of assessment and all patients were prescribed pancreatic enzyme supplements. Of the 20/21 adult patients for whom data were available, 13 (65%) had a diagnosis of CF-related glucose intolerance and were insulin-dependent at the time of LT assessment. An additional two (10%) patients developed de novo insulin dependency posttransplant at 2 and 3 years, respectively, (75% total).

A total of 8/19 (42%) pediatric recipients had a pretransplant diagnosis of glucose intolerance and were insulin-dependent. An additional 7/19 (37%) patients developed glucose intolerance and insulin dependency posttransplant. Five of these seven patients developed de novo insulin dependency in the early postoperative period of whom three were able to discontinue insulin at 9, 12 and 16 months posttransplant, with one recommencing insulin therapy at 36 months posttransplant. The additional 2 patients developed de novo insulin dependency at 4 and 8 years posttransplant, respectively. At 5 years posttransplant, 12/19 (63%) of patients in the pediatric cohort were insulin-dependent.

Immunosuppression

The immunosuppression regimes utilized in these patient cohorts were similar to the general LT population. All adult patients were commenced on calcineurin inhibitor-based triple therapy with steroids and azathioprine. Of those receiving a single organ transplant, 4 patients received ciclosporin, 1 of whom was switched to tacrolimus on day 8, and the remainder (n = 8 for whom data were available) received tacrolimus-based triple therapy. The 4 patients transplanted since 2006 received mycophenolate mofetil (MMF) in place of azathioprine, consistent with our practice in general LT recipients at high risk of renal dysfunction. One triple organ recipient died perioperatively but the remaining three were commenced on ciclosporin-based triple therapy, consistent with accepted practice at that time. In the pediatric cohort, most patients transplanted in the earlier period received ciclosporin-based triple therapy (n = 9) with steroids and azathioprine, after which time tacrolimus replaced ciclosporin as the standard calcineurin inhibitor. However, in contrast to the adult cohort, most pediatric patients transplanted since 2000 received only dual therapy with tacrolimus and steroids (n = 8), although 1 patient later commenced MMF and another also received induction dacluzimab therapy.

In the adult cohort, the standard immunosuppression protocol indicates that steroids are discontinued by 3 months, although we cannot exclude some patients receiving future courses of steroids as treatment for their lung disease. However, in the pediatric cohort, steroids were continued (n = 7), or restarted at between 1 and 5 years posttransplant after being discontinued at 3 months (n = 6), in 13/17 patients who survived beyond 1 year. Steroids were continued/restarted either due to the immunosuppression protocol at the time, as treatment for lung disease, or in some cases due to de novo autoimmune hepatitis in the allograft.

DIOS after LT

Five adult patients had a history of DIOS pretransplant. Of these, one died of “cepacia syndrome” on day 15 posttransplant without developing DIOS and one died at 33 days posttransplant from DIOS, with associated sepsis and multiorgan failure. This patient had a history of recurrent episodes of DIOS pretransplant and was therefore at high risk of developing bowel obstruction postoperatively. This prompted introduction of a new protocol at our unit, whereby high-risk patients have a temporary ileostomy sited at the time of transplant, which is reversed after 3 months, in addition to standard prophylactic measures such as polyethylene glycol lavage solution, gastrograffin and elemental feeding. Since the introduction of this protocol, 3 patients with a previous history of DIOS have been transplanted, with ileostomy reversal between 3 and 5 months. Although some ileostomy-associated complications have been observed, mainly related to fluid balance and electrolyte disturbances, none of these patients have developed postoperative DIOS. Prior to the new protocol, one other patient with no prior history of DIOS developed this complication postoperatively, but made an uneventful recovery after receiving gastrograffin on day 9 posttransplant.

A total of 2/19 (10.5%) pediatric patients had a history of meconium ileus/DIOS pretransplant but did not develop DIOS posttransplant. A total of 4/19 (21%) patients without a prior history of this condition experienced DIOS posttransplant. Only one of these cases occurred early (on day 4) and was managed successfully with conservative treatment. Three other patients had later symptoms of bowel obstruction at 8 months, 8 years and 9.5 years posttransplant, none of whom required surgery.

Discussion

This is the largest single-center series of the long-term outcomes of liver transplantation for adult and pediatric patients with CFLD. Overall survival rates after sole liver transplantation in adults and children were 85% and 90% at 1 year and 65% and 85% at 5 years, respectively. These survival rates are similar to those undergoing liver transplantation for other indications at our center. Most deaths in both adult and pediatric cohorts occurred late and were due to respiratory complications of CF. The patient who received combined liver/lung transplantation remains well at 3 years posttransplant with no significant postoperative complications. As observed elsewhere (7), triple organ transplantation was associated with unacceptably high mortality rates and this program was subsequently discontinued.

In the adult cohort a markedly increased male: female sex ratio was observed, which is consistent with previous reports (7,17,18). The mechanism by which male sex increases risk of CFLD remains to be established, although it has been suggested that endocrine factors may exert a protective effect on female CF patients (19). A difference in the indications for LT was observed between adult and pediatric recipients. While most (76%) adults were transplanted for decompensated liver disease, with or without intractable variceal bleeding, this was lower in children (42%). Early or preemptive liver transplantation was performed in both groups, but was commoner in children (58%). In this latter group, selection and timing for transplant was based on the impact liver transplantation may also have on lung function and nutritional status to avoid the need for combined lung/LT.

Previous reports on the effect of liver transplantation on lung function in patients with CFLD are conflicting. Although several studies have suggested a short-term improvement following liver transplantation (9,18,20,21), a later study demonstrated a more rapid decline in lung function than would have been expected without transplantation (7). Studies reporting improvements in lung function were largely in pediatric populations, who may have a more reversible type of lung disease than is observed in adults (7). Importantly, these studies have also been based on small numbers with very short follow-up times. We report that lung function in adults as determined by FEV1 declines at a rate of 0.74%/year over the first 4 years. This is a marked improvement compared with the predicted 3%/year decline in FEV1 for adult CF patients without liver transplantation (13–15). Although studies report annual reductions of FEV1% as low as 1%, these studies do not take into account the comorbidity found in patients with concomitant liver disease, which is likely to be associated with a worse rate of decline. Our patients had an average of 3.9 chest infections/year pretransplant, which is reported in other CF nonliver disease cohorts to be associated with a significantly increased annual rate of decline in FEV1% (13). In pediatric patients, decline in FEV1 is dependent on baseline lung function in addition to growth and nutritional status (22), but has been reported to vary from 1.1% to 2.4%/year (23). Although FEV1 is generally regarded as the better marker for describing lung function, it is interesting that our cohort have a posttransplant annual decline in FVC% of 2.93%. The reason for the disparity is not clear and cautions against the use of just FEV1 when assessing such patients. However, the marked reduction in requirement for intravenous antibiotics for chest exacerbations seen after transplant in our cohort is an important observation that has not previously been reported, and suggests a form of stabilization.

Some of the initial improvement in lung function may reflect resolution of effects of severe portal hypertension such as pulmonary edema, intrapulmonary shunting, ascites and organomegaly (20), in addition to reduced infection resulting from improved immune, particularly neutrophil, function (24).

Of note, the median FEV1 at listing in our adult cohort was 48.4% predicted, thus including several patients with an FEV1 less than 50% predicted, who would not have been offered isolated liver transplantation in some other centers (7). Poorer lung function at baseline did not seem to adversely impact survival and these patients achieved similar stabilization of lung function compared to those adults with higher FEV1 at baseline. These results suggest that isolated liver transplantation may be feasible in patients with moderately reduced lung function (i.e. FEV1 40–50% predicted).

Chronic infection with PA was observed in the majority of patients in both cohorts, often in combination with other organisms. The presence of multiresistant PA did not portend a poorer outcome than antibiotic sensitive PA. Infection with the ET12 strain of Burkholderia cenocepacia led to ‘cepacia syndrome’ and early posttransplant death in 1 patient and, being associated with a poorer outcome following lung transplantation (25–27), is now considered a contraindication for single or multiple organ transplantation at our center.

Previous studies have reported mixed findings with regard to nutritional status post-LT (7,10), although these studies had limited follow-up. Body mass index in our adult population was maintained, with a small but nonsignificant improvement after 5 years. This may be at least partly attributable to the observed stabilization of lung function, as there is a well-established relationship between lung function and nutritional status in CF (28). The more advanced preoperative lung disease in our cohort compared with other groups (7,10) may also have reduced the degree of nutritional improvement. Poor nutritional status has previously been associated with worse outcome after liver (29) and lung transplantation (30), and it has been suggested that patients with nutritional deterioration should undergo early and elective liver transplantation (10). However, in our study lower pretransplant body mass index did not affect 1- or 5-year survival. In keeping with previous reports, we found no improvement in overall nutritional status posttransplant in the pediatric cohort, despite a large number of children receiving nutritional support. This contrasts with the nutritional improvement seen in children transplanted for other chronic liver diseases and may reflect the unchanged impact of CF on overall nutritional status, including pancreatic insufficiency, suppurative lung disease and delayed puberty.

Although renal function in our adult cohort declined after liver transplantation, the prevalence of chronic renal impairment at both 2 and 5 years was in keeping with our general LT population. These results also compare favorably with previous reports, where up to 40% developed significant renal impairment (7). In the pediatric cohort, a similar trend was observed as in the general pediatric LT population, with a significant deterioration of renal function in the first 6 months, followed by stabilization up to and improvement after 12 months posttransplant (31,32). These findings suggest that adults and children undergoing liver transplantation for CFLD do not have a higher risk of developing renal impairment compared to the general liver transplantation population, despite the high prevalence of diabetes and increased pre- and posttransplant exposure to potentially nephrotoxic antibiotics such as aminoglycosides.

Endocrine pancreatic insufficiency is a common complication of CF, with reported prevalence of CF-related diabetes (CFRD) increasing from 9% below the age of 10 years to 43% above the age of 30 years, with annual age-dependent incidence rates ranging from 4% to 9% (33). The rates of 65% and 42% of adult and pediatric patients in our cohort with a pretransplant diagnosis of CFRD were therefore higher than might be expected. New-onset diabetes is a frequent posttransplant complication, related to a number of factors including the use of steroids and other diabetogenic immunosuppressive agents (34). In our cohort, a further 2 (10%) adult and 7 (37%) pediatric patients developed new-onset diabetes posttransplantation. In comparison, a recent study of 96 adults undergoing liver transplantation at our center reported rates of pretransplant diabetes of 73% in patients transplanted for nonalcoholic steatohepatitis (NASH), a condition in which diabetes is highly prevalent, and 29% in patients transplanted for other chronic liver diseases, increasing to 81% and 43%, respectively at 2 years posttransplant (35). A larger study from Israel, including 252 adult LT recipients, reported a pretransplant diabetes prevalence of 14.4%, rising to 39.6% posttransplant, with a mean follow-up of 6.2 years (36). A recent analysis using the Organ Procurement and Transplant Network/United Network for Organ Sharing database reported an adult posttransplant incidence of new-onset diabetes of 26.4% over a median follow-up period of 685 days (34). The significant proportion of pediatric patients continuing or resuming steroids at 1 year posttransplant may have contributed to the high incidence of posttransplant diabetes observed in this group.

DIOS, formerly known as meconium ileus equivalent, is reported to occur in 10–20% of patients with CF (37,38), being more common with age (39) and associated with pancreatic insufficiency (37,40). Risk factors include a previous history of DIOS or abdominal surgery, and acute potentially life-threatening intestinal obstruction can occur in patients with CF undergoing liver transplantation. Interestingly, DIOS occurred more frequently in our pediatric cohort, although occurred as a late presentation in 75% of cases, with none of these children having had pretransplant DIOS. Following the early posttransplant death of an adult patient from DIOS, our unit introduced a new posttransplant DIOS management protocol, which includes placement of a temporary ileostomy at time of transplant in high-risk individuals. Results for 3 patients transplanted according to this protocol have been positive with no further episodes of DIOS observed.

Our study has several strengths. It is the largest single-center series with long-term follow-up data in both adult and pediatric cohorts to be reported to date. Importantly, this includes detailed data on pulmonary function, renal function and nutritional status (adjusting for ascites and peripheral edema), which are not recorded in registry studies (41). As a consequence, we are able to provide novel data on the impact of bacterial colonization and the need for posttransplant intravenous antibiotics. Furthermore, our observation and change in practice of cases of perioperative DIOS provides important clinical information to clinicians managing CF patients receiving liver transplantation.

Our study is weakened by the fact it is retrospective, although this is the case for all previous publications of such cohorts. As it is a single-center study, we have been able to collect most of the requisite information to overcome many of the detractions of a retrospective series. However, for some of the later time points, the number of patients with available data is reduced due either to them not reaching those time points yet or due to transfer of their care to other regional CF units where not all the data were available. Nevertheless, there are still sufficient data points to address the clinically important question of the impact of single LT on lung function and nutritional status.

In conclusion, we suggest that liver transplantation is an effective treatment for patients with decompensated cirrhosis due to CFLD, even in the presence of moderately impaired lung function. Our study provides novel data on long-term lung function after LT for CFLD; it appears to slow down the deterioration of some aspects of lung function (FEV1%) over the first 4 years, as well as reducing the frequency of respiratory infections requiring intravenous antibiotics. Our findings have implications for decision making in relation to the optimal timing of liver transplantation in patients with incipient end-stage CFLD and declining lung function. These data support the view that there should be a lower threshold for considering liver transplantation in patients with CFLD when faced with declining lung function and nutritional status.

Acknowledgments

JK Dowman is a Wellcome Trust Clinical Research Fellow and is receiving funding from the Wellcome Trust for liver research unrelated to this work. This manuscript has not been influenced by this association.

Disclosure

The authors of this manuscript have no conflicts of interest to disclose as described by the American Journal of Transplantation.

Ancillary