Telephone: +390432559801; FAX: +39043242097
Prevalence and risk factors for delayed adrenal insufficiency after liver transplantation
Article first published online: 25 JUN 2008
Copyright © 2008 American Association for the Study of Liver Diseases
Volume 14, Issue 7, pages 1014–1019, July 2008
How to Cite
Toniutto, P., Fabris, C., Fumolo, E., Bitetto, D., Fornasiere, E., Falleti, E., Rapetti, R., Minisini, R. and Pirisi, M. (2008), Prevalence and risk factors for delayed adrenal insufficiency after liver transplantation. Liver Transpl, 14: 1014–1019. doi: 10.1002/lt.21465
- Issue published online: 25 JUN 2008
- Article first published online: 25 JUN 2008
- Manuscript Accepted: 19 JAN 2008
- Manuscript Received: 3 SEP 2007
- Ricerca Sanitaria Finalizzata program (Regione Piemonte, Italy)
Liver transplantation (LT) recipients are at risk for early and delayed adrenal insufficiency for multiple reasons. Although early adrenal insufficiency is known to occur in a high proportion of recipients maintained on steroid-free immunosuppressive regimens, the prevalence and risk factors associated with delayed functional adrenal gland atrophy (FAGA) are unknown because routine evaluation for this condition is not standard practice among LT centers. We investigated a group of 87 patients (64 males) transplanted for end-stage liver disease related to different etiologies. All underwent a standard corticotropin stimulation test (CST) when, after gradual steroid tapering, they had been maintained for at least 1 week on oral prednisone at a daily dose of 5 mg. FAGA, defined by a serum cortisol concentration that, 60 minutes after corticotropin administration, did not double the baseline level and remained <20 μg/dL, was diagnosed in 23/87 patients (26.4%). Stepwise logistic regression analysis selected as significant predictors of FAGA the cumulative dosage of corticosteroids administered (P < 0.01), the increase in the body mass index after LT (P < 0.01), a low serum cholesterol concentration (P = 0.005), and a high adrenocorticotropin hormone (ACTH) serum level (P < 0.05) at the time CST was performed. In conclusion, FAGA is a common condition among LT recipients who are maintained on prolonged corticosteroid immunosuppressive treatment. Factors associated with FAGA include the cumulative steroid dose, weight changes after LT, and ACTH and cholesterol levels at the time of steroid withdrawal. Liver Transpl 14:1014–1019, 2008. © 2008 AASLD.
Relative adrenal insufficiency (RAI) is an uncommon clinical condition that can develop after conventional surgery, with an incidence of approximately 0.5%.1, 2 Previous studies have shown that RAI occurs more frequently in critically ill patients with severe infection, trauma, or aggressive surgery.3, 4 Moreover, RAI has been recently described in patients with end-stage liver disease, a setting for which the definition of hepatoadrenal syndrome has been proposed.5 Although the mechanisms responsible for RAI in patients with advanced liver diseases have not been fully elucidated, decreased availability of cholesterol6 and high levels of endotoxin and proinflammatory cytokines, such as tumor necrosis factor alpha,7, 8 may play a major role.
Factors able to induce an adrenal crisis, such as hypotension, infection, and severe blood loss, may all complicate liver transplantation (LT) in the early postoperative period. Glucocorticoids, included in the majority of immunosuppressive protocols adopted after LT, prevent LT recipients from developing early RAI after LT but greatly increase the risk of late functional adrenal gland atrophy (FAGA). Indeed, the prolonged use of corticosteroids is a major cause of adrenal suppression, and corticosteroid discontinuation can therefore trigger overt FAGA.9 LT recipients are therefore at risk for adrenal insufficiency in their first year after the transplant operation for multiple reasons. However, routine evaluation for this condition is not standard practice in LT centers; no information concerning the prevalence and risk factors associated with FAGA in liver transplant patients are available in the literature.
In this study, we investigated, in a group of patients transplanted for end-stage liver disease related to different etiologies and maintained on long-term immunosuppressive regimens based on calcineurin inhibitors, the prevalence and risk factors associated with FAGA at the time of corticosteroid withdrawal.
PATIENTS AND METHODS
Data from 87 patients who underwent LT for end-stage liver disease at our institution from 1998 to 2005 were retrospectively analyzed. The selection process of the studied population is shown in Fig. 1. Patients were eligible if they survived at least 8 months after LT, had not been involved in clinical trials with experimental immunosuppressive regimens, and were able to attend the transplantation clinic for regular follow-up visits. Clinical and demographic characteristics of the studied population are presented in Table 1. None of the studied patients was on corticosteroids at the time of the transplant operation. All patients were maintained on an immunosuppressive regimen that was either cyclosporine-based or tacrolimus-based and was associated, in the first few months after LT, with corticosteroids. The cyclosporine dosage was calculated to obtain serum levels (measured 2 hours after the drug administration) ranging from 800 to 1200 μg/L in the first 6 weeks after transplant and from 600 to 800 μg/L thereafter. The tacrolimus dosage was calculated to obtain predose serum levels ranging from 10 to 15 μg/L in the first 6 weeks after transplant and from 5 to 10 μg/L thereafter. In all patients, the corticosteroid administration during the first week after LT was performed as follows: 1000 mg of methyl prednisolone intravenously during the transplant operation, 250 mg intravenously on the first and second day after LT, 125 mg intravenously on the third day after LT, and then 40 mg of oral prednisone each day up to the first week after LT. In the second postoperative week, corticosteroids began to be slowly tapered according to a schedule aiming to complete steroid withdrawal by the end of the sixth postoperative month. On clinical grounds, however, caring physicians were allowed to deviate from this scheme, and this resulted in shorter or longer corticosteroid courses: the 95% confidence intervals of the median, calculated according to the method of Campbell and Gardner,10 for the length of corticosteroid treatment after LT were therefore 100-125 days. In the presence of moderate to severe rejection episodes, 3 intravenous methyl prednisolone boluses of 1000 mg followed by 4 days of 40 mg of oral prednisone were added to the protocol corticosteroid dosages. Body weight was measured to the nearest 0.1 kg, and height was measured to the nearest 1 cm, with study participants wearing only underwear and no shoes in order to calculate the body mass index (BMI). BMI was determined both at the time of LT and at the time of the corticotropin stimulation test (CST), with the amount of ascites, which was estimated by abdominal ultrasound, being subtracted from the body weight.
|Recipient male gender||64 (73.6)|
|Recipient age in years||56 (21–67)|
|BMI in kilograms per meter squared||24.8 (17.0–34.6)|
|MELD score*||15 (9–30)|
|Etiology of liver disease|
|Donor male gender||58 (66.7)|
|Donor age in years||51 (14–77)|
|Tacrolimus-based immune-suppressive therapy||56 (64.4)|
|Length of corticosteroid therapy in days||117 (24–738)|
|Cumulative dose of corticosteroids in milligrams of prednisone||1480 (112–8532)|
|Patients with post-LT diabetes mellitus†||40 (45.9)|
|Patients who experienced rejection episodes‡||18 (20.7)|
According to a protocol enforced at our center, all patients had to undergo a standard CST (250 μg of corticotropin administered intravenously in 5 minutes)11 before complete steroid withdrawal when, after gradual tapering, a daily oral prednisone dose of 5 mg had been reached and maintained for at least 1 week. FAGA was diagnosed when the fasting serum cortisol concentration, 60 minutes after corticotropin administration, did not double the baseline level and remained <20 μg/dL.
Statistical analysis of data was performed with the BMDP Dynamic statistical software package, version 7.0 (Statistical Solutions, Cork, Ireland). Continuous variables were presented as medians (range), and categorical variables were presented as frequencies (%). Associations between categorical variables were performed by means of the chi-square test. Stepwise logistic regression analysis with a forward approach was employed to assess the variables or selected interactions between the variables independently associated with an abnormal CST. With this method, by means of an approximate asymptotic covariance estimate, at each step in the stepping process, one set of design variables was added to the model. The odds ratios and 95% confidence intervals of the selected variables were also calculated. The 95% confidence interval of the median of the length of corticosteroid treatment was calculated with MedCalc statistical software, version 18.104.22.168 (Mariakerke, Belgium).
The CST was abnormal in 23 (26.4%) and normal in 64 (73.6%) of the studied patients. In Table 2, their clinical and demographic characteristics, at the time of the transplant operation and during the follow-up after LT, are presented with respect to the CST results. A significant correlation was found between an abnormal CST and both the length and cumulative dosage of corticosteroid therapy administered.
|Clinical and Demographic Characteristics||Abnormal CST [n = 23 (26)]||Normal CST [n = 64 (74)]||P*|
|Recipient male gender||20 (87)||44 (69)||<0.10|
|Recipient age > 55 years||9 (39)||35 (55)||>0.10|
|BMI ≤ 25 kg/m2||17 (74)||34 (53)||<0.10|
|MELD score > 15||9 (39)||27 (42)||>0.10|
|Alcoholic etiology of liver disease||13 (57)||22 (34)||<0.10|
|Donor male gender||16 (70)||42 (66)||>0.10|
|Donor age > 50 years||12 (52)||33 (52)||>0.10|
|Patients with tacrolimus-based therapy||16 (70)||40 (63)||>0.10|
|Patients with length of corticosteroid therapy > 105 days||17 (74)||32 (50)||<0.05|
|Patients with cumulative dose of prednisone > 1350 mg||19 (83)||33 (52)||<0.01|
|Patients who experienced acute rejection episodes||6 (26)||12 (19)||>0.10|
Table 3 reports the associations between the clinical characteristics of the studied population during the perioperative period and the presence of an abnormal CST. Neither surgically related parameters nor the blood supply needed during the transplant operation was found to be associated with the occurrence of post-LT FAGA.
|Clinical Characteristics||Abnormal CST [n = 23 (26)]||Normal CST [n = 64 (74)]||P†|
|Graft cold ischemia time > 420 minutes||11 (48)||39 (61)||>0.10|
|Operative time > 360 minutes||17 (74)||46 (72)||>0.10|
|>5 blood units transfused||11 (48)||36 (56)||>0.10|
|Presence in the ICU > 7 days||2 (9)||14 (22)||>0.10|
|Sepsis in the ICU*||0 (0)||4 (6)||>0.10|
The associations between clinical and laboratory parameters evaluated at the moment of CST performance and the presence of an abnormal CST are reported in Table 4. FAGA was associated with serum concentrations of total cholesterol and adrenocorticotropin hormone (ACTH) and with body mass changes after LT.
|Clinical Laboratory Parameters at the Time of CST||Abnormal CST [n = 23 (26)]||Normal CST [n = 64 (74)]||P†|
|Post-LT BMI increase > 1 kg/m2 (n = 24)*||11 (48)||13 (20)||<0.02|
|Serum cholesterol concentration ≤ 125 mg/dL (n = 9)||5 (22)||4 (6)||<0.05|
|Serum ACTH concentration > 26 pg/mL (n = 32)||13 (57)||19 (30)||<0.05|
|Fasting serum glucose level > 120 mg/dL (n = 14)||5 (22)||9 (14)||>0.10|
|MBP > 100 mm Hg (n = 42)||12 (52)||30 (47)||>0.10|
|Plasma potassium concentration > 4.8 mEq/L (n = 13)||2 (9)||11 (17)||>0.10|
Stepwise logistic regression analysis with a forward approach was performed to identify design variables independently associated with an abnormal CST. Predictor variables included all those reported in Tables 2 to 4. Furthermore, the following interactions were considered: the Model for End-Stage Liver Disease score with each of the parameters reported in Table 3 and all the possible associations between the variables reported in the same table. At step 1, the analysis selected as a significant predictor of FAGA development the cumulative dosage of corticosteroids administered (improvement χ square: 7.350, P < 0.01); at step 2, it selected the serum cholesterol concentration at the moment of CST (improvement χ square: 7.860, P = 0.005); at step 3, it selected the ACTH serum level at the moment of CST (improvement χ square: 5.268, P < 0.05); and at step 4, it selected the increase of the BMI value after LT (improvement χ square: 6.936, P < 0.01). The pertinent statistical values, obtained from the logistic regression analysis, of the independent predictor variables are reported in Table 5.
|Dose of prednisone > 1350 mg||2.071||0.834||2.48||7.93||1.51–41.7|
|Post-LT BMI increase > 1 kg/m2*||1.621||0.639||2.53||5.06||1.42–18.0|
|Serum ACTH > 26 pg/mL||1.574||0.624||2.52||4.82||1.40–16.7|
|Serum cholesterol ≤ 125 mg/dL||−2.857||1.02||−2.79||0.06||0.01–0.44|
Prolonged treatment with corticosteroids followed by their withdrawal can induce 3 possible effects on the adrenal glands.12 In some cases, the patient retains a normal function of the hypothalamic pituitary adrenal (HPA) axis, and FAGA does not develop; on the contrary, other patients may develop a central adrenal insufficiency. In this case, the suppression of the corticotropin-releasing hormone production by the hypothalamus and/or of the ACTH by the pituitary gland during corticosteroid treatment may outlast the discontinuation of steroid therapy. Finally, the most severely affected glucocorticoid-treated patients have complete HPA suppression characterized by adrenal gland atrophy. In this condition, the adrenal glands cannot generate cortisol promptly even if directly stimulated by exogenous ACTH. To assess FAGA and to discriminate this condition from central adrenal insufficiency, several diagnostic tests may be adopted, that is, insulin hypoglycemia, a metyrapone test, a corticotropin-releasing hormone test, and low-dose (1 μg) or high-dose (250 μg) CST.9 In this study, high-dose CST was employed because the results given by this test offer the best diagnostic discrimination of patients affected by FAGA.13
In the present series, 23 of 87 (26.4%) of the liver-transplanted patients have been found to present an abnormal CST and then diagnosed to have FAGA. To the best of our knowledge, this is the first study to report the prevalence of this clinical condition in liver-transplanted patients. A very high prevalence (92%) of RAI has recently been reported in patients who underwent LT and were maintained on steroid-free immunosuppressive regimens.6 However, 2 major comments should be outlined to differentiate the prevalence of RAI in liver-transplanted patients and the prevalence of FAGA that we found in our work. Although RAI develops early after LT as the result of the stress derived from the transplant operation, FAGA is the result of steroid withdrawal. In effect, the prevalence of adrenal insufficiency was lower (61%) in LT patients studied later after the transplant operation.6 Furthermore, a low-dose CST instead of a high-dose CST was previously adopted to diagnose RAI in these patients; this probably led to overestimation of the prevalence of RAI.
Factors responsible for HPA axis suppression and induction of FAGA are in part clearly defined. First of all, the type of steroids used is of primary importance because agents with more potent anti-inflammatory activity are responsible for a more pronounced ACTH suppression.14 In this respect, we adopted for all patients a single schedule of steroid administration, with methyl prednisolone administered intravenously early after LT and followed by oral prednisone. Among corticosteroids used in clinical practice, prednisone is characterized by intermediate anti-inflammatory potency and the ability to induce ACTH suppression.14 Because the type of corticosteroid treatment was identical in all patients, this variable has not been further considered. On the contrary, 2 major determinants of FAGA are represented by the duration and cumulative dose of glucocorticoid assumption.15 In effect, only these 2 variables among the other selected basal clinical and demographic characteristics of the patients were found to be significantly correlated with a pathological CST. FAGA was significantly more common among patients with a length of corticosteroid therapy > 105 days and receiving a total dose of prednisone > 1350 mg. It can be argued that another important factor for determining FAGA in our patients could be related to the timing adopted in the steroid-withdrawal scheme. Even if no clear indications about this issue are present in the literature, many experts suggest that the corticosteroid withdrawal should be performed slowly.9 In the present series, corticosteroid tapering to the dosage of 5 mg was completed by all patients within 6 months from the transplant operation, a period long enough to exclude the possibility that FAGA developed because of excessively rapid steroid tapering. It has been reported that the adrenal response is suppressed for up to 14 days even after short-term glucocorticoid treatment.16 Thus, it could be reasonable to suggest maintaining for a longer time a dosage of 5 mg of prednisone before a CST is performed or, as an alternative, tapering the prednisone dose to 2.5 mg daily before complete withdrawal.
Interestingly, no associations were observed between the presence of FAGA and either the severity of pretransplant liver disease, that is, the Model for End-Stage Liver Disease score, or the occurrence of several known clinical determinants of surgical stress during LT, such as blood loss, surgical time, and presence of sepsis. On the contrary, these factors have been found to be associated with the development of RAI in liver-transplanted patients adopting a steroid-free immunosuppressive regimen.6 Our interpretation is that the corticosteroid treatment adopted in our protocols protected the patients from the early occurrence of RAI after LT; however, while protecting our patients from the development of RAI, we may render them more prone to develop FAGA.
Can we identify at the moment of corticosteroid withdrawal which patients will develop FAGA and therefore reserve only for these subjects the performance of CST to avoid adrenal crisis? Surely the information obtainable about the duration and total dose of corticosteroids administered may be valuable.15 However, the total corticosteroid dose administered cannot always be easily obtained. Among the parameters associated with prolonged corticosteroid administration, only a significant rise (>1 kg/m2) of BMI was associated with an abnormal CST value, whereas no association was found with fasting blood glucose, potassium serum levels, and mean blood pressure. It may be argued that in the context of LT, the assumption of corticosteroids may magnify their anabolic effect with a tendency to develop a major body mass increase, whereas other metabolic and circulatory side effects may be blunted by the concomitant use of calcineurin inhibitors. The usefulness of measuring ACTH in these patients is indubitable, and high ACTH levels could prompt clinicians to obtain a CST.11 However, that a parameter routinely evaluated such as serum cholesterol is also associated with an abnormal CST value is perhaps more interesting from both clinical and pathophysiological standpoints. Low levels of high-density lipoprotein cholesterol are significantly related to RAI in patients with liver diseases, whereas low-density lipoprotein cholesterol levels are also decreased in these patients in comparison with those without RAI, although without reaching statistical significance.6 Cholesterol is the principal precursor of steroid genesis. At rest and during stress, about 80% of circulating cortisol is derived from plasma cholesterol, the remaining 20% being synthesized in situ from acetate and other precursors.17
In conclusion, FAGA is a common condition among LT recipients who are maintained on prolonged corticosteroid immunosuppressive treatment. Because the risks and costs of the high-dose CST are limited, this test might be performed systematically when steroid withdrawal is completed. Factors associated with FAGA include the cumulative steroid dose, weight changes after LT, and ACTH and cholesterol levels at the time of steroid withdrawal.