Adrenal insufficiency is common amongst kidney transplant recipients receiving maintenance prednisolone and can be predicted using morning cortisol

ABSTRACT Background Long-term glucocorticoid therapy is a key component of immunosuppression for kidney transplant recipients (KTRs), leading to significant cumulative glucocorticoid exposure. The aims of this study are to investigate the prevalence of adrenal insufficiency (AI) in KTRs taking prednisolone and to develop a screening algorithm to identify patients at the highest risk of AI. Methods In this cross-sectional cohort study, 67 KTRs receiving prednisolone underwent a short synacthen test (SST) and measurement of cumulative glucocorticoid exposure. Results A total of 72% (n = 48) of participants failed the SST. Participants with AI had a higher daily prednisolone dose (4.9 versus 4.2 mg/day; P = .002) and greater cumulative glucocorticoid exposure (289 versus 111 mg/kg; P = .03) than those with intact adrenal function. Participants with AI had lower baseline cortisol than participants with intact adrenal function (143 versus 303 nmol/L; P < .001). Morning cortisol of >288 nmol/L predicted a normal SST with 100% specificity [95% confidence interval (CI) 92–100] and 70% sensitivity (95% CI 56–78%), therefore excluding AI. Conclusions Our results suggest KTRs are at a higher risk for AI than previously reported. A morning serum cortisol measurement is a useful screening tool in this cohort, reducing the need for stimulatory testing by 44%. KTRs with AI need education regarding glucocorticoid sick rules, similar to patients with other forms of AI.


INTRODUCTION
Adrenal insufficiency (AI) occurring after the suppression of the hypothalamic-pituitary-adrenal axis by exogenous glucocorticoids is termed tertiary adrenal insufficiency (TAI) [1]. TAI is associated with all forms of exogenous glucocorticoid medication and patients exposed to glucocorticoids for prolonged periods are at an increased risk of TAI [2]. Patients with undiagnosed TAI are at risk of developing a life-threatening adrenal crisis should they discontinue steroid therapy abruptly or if they encounter a physiological stressor such as an intercurrent illness or surgery without appropriate stress dose steroid dosing [3,4].
Glucocorticoid medications such as prednisolone are commonly used as a component of long-term maintenance immunosuppression after kidney transplantation, and high-dose glucocorticoids are the first-line treatment for acute rejection episodes [5,6]. Current transplant guidelines recommend daily doses of prednisolone of ≤5 mg [5,6]. As the median survival following a kidney transplant is excellent [7], many kidney transplant recipients (KTRs) receive a significant cumulative dose exposure of glucocorticoids over the lifetime of the graft and consequently are at risk of developing AI and associated adrenal crises. However, despite this, steroid safety advice and sick day rule education do not feature in the major clinical practice guidelines for the management of KTRs [5,6].
A small number of studies have reported a 44-80% prevalence of TAI in the KTRs population [8][9][10][11] (Table 1); meta-analysis data suggest that the absolute risk of TAI in this population is 56.2 [(95% confidence interval (CI) 42.9-68.6] [2]. However, most published studies have small participant numbers [8][9][10][11][12][13][14][15][16], were performed in an era of higher maintenance prednisolone doses (5-20 mg prednisolone/day) and vary in their methods of adrenal function assessment [10,[13][14][15]. More recent studies of patients treated with lower maintenance prednisolone (5-7.5 mg prednisolone/day) suggest the prevalence is 31-60% [14][15][16]. It is therefore Adrenal insufficiency among KTRs Naik et al. [9] K T R (n = 10) 10 mg 80 SST SST on participants ≤5mg, 4/5 fail, 1 participant symptomatic AI Canaflax et al. [10] K T R (n = 18) 0.1-1 mg/kg/day 44 250 μg SST 50% of SST failures died within 2 years, 12 months pred >25 g cumulative dose predicts SST failure Rodgers et al. [11] K T R (n = 21) 15  difficult to predict which KTR patients are likely to develop TAI on modern glucocorticoid regimens. While there is some evidence in the paediatric population that prescribed daily dose correlates with risk of TAI [15], the duration of glucocorticoid exposure has not consistently correlated with the risk of TAI [14,16] and there is a lack of published data detailing how cumulative glucocorticoid exposure correlates with the risk of TAI. TAI can be asymptomatic outside of the setting of an adrenal crisis [2] and therefore clinicians rely heavily on laboratory investigations for diagnosis. Stimulatory tests such as the short synacthen test (SST) and insulin tolerance test are the gold standard for the diagnosis of AI [17,18] and are safe and welltolerated investigations. The use of a morning serum cortisol value as a screening tool has been reported in other patient cohorts to reduce the need for stimulatory testing by 40% [19][20][21]. This would represent time-and resource-saving benefits for both patients and healthcare systems; however, it has not yet been validated in this patient cohort.
We hypothesize that TAI is highly prevalent in the KTR patient population and that it is underdiagnosed and therefore undertreated. We suggest that the risk of TAI correlates with both daily prednisolone dose and cumulative glucocorticoid exposure.
The primary aim of this study is to determine the prevalence of TAI in renal transplant recipients receiving maintenance immunosuppression ≤5 mg prednisolone/day. Secondary aims were to assess the relationship between TAI and cumulative dose exposure to prednisolone, to evaluate the use of a morning serum cortisol concentration as a screening tool for the diagnosis of TAI in this group and to assess knowledge of steroid sick day rules in the KTR population.

Participant recruitment
Participants were recruited consecutively from the National Kidney Transplant outpatient department at Beaumont Hospital, Dublin, Ireland. Eligible patients were KTRs ≥18 years of age, ≥1 year post-transplant, established on a stable maintenance dose of oral prednisolone for ≥6 months, with stable 238 graft function [estimated glomerular filtration rate (eGFR) >30 mL/min/1.73 m 2 ]. Exclusion criteria included known hypothalamic-pituitary or adrenal pathology, known adrenal insufficiency, critically ill patients, pregnancy or oral oestrogen medication within 6 weeks of study enrolment. Participants with impaired graft function (eGFR <30 mL/min/1.73 m 2 ) were excluded from the study due to changes in endogenous cortisol metabolism and binding globulin affinity and changes in prednisolone pharmacokinetics in the latter stages of chronic kidney disease [22][23][24].

Study visit
Patients attended for a single visit, during which a detailed medical history was obtained and electronic and paper healthcare records were reviewed. A detailed cumulative glucocorticoid exposure history was calculated, which included pretransplant, perioperative and post-transplant exposure to glucocorticoids and exposure to glucocorticoid medication for reasons unrelated to transplantation (such as the treatment of underlying inflammatory conditions), using patient interview and hospital chart review. The results were finalized after review of the National Kidney Disease Clinical Patient Management System, a national kidney disease electronic patient management system that is used in the care of patients with chronic kidney disease. Knowledge of steroid sick day rules advice, medical alert card and emergency hydrocortisone administration was assessed by participant interview.
Participants were requested to withhold their morning medications, as per local practice, to obtain trough drug levels. Fasting laboratory investigations were performed, including full blood count; renal, liver and bone profiles; total serum bicarbonate, magnesium, serum albumin, tacrolimus or sirolimus levels as appropriate and plasma adrenocorticotropin (ACTH), using in-hospital assays (see below for details). Cortisol binding globulin concentration was measured using a competitive radioimmunoassay (DIAsource Immunoassays, Louvain-la-Neuve, Belgium) [intra-assay coefficient of variation (CV) 8.6% at 23 μg/mL and 3.9% at 83 μg/mL].
Blood samples were taken prior to (0 minutes) and 30 minutes after 250 μg intravenous bolus of synthetic ACTH was administered (Synacthen, Alfa Sigma, Rome, Italy). Serum cortisol concentration was measured using the Elecsys II Cortisol immunoassay (Roche Diagnostics, Mannheim, Germany), with a CV of 1.8-2.1% and cross-reactivity with prednisolone of 7.9%. All laboratory samples were obtained before 10 a.m. A peak cortisol response to SST >430 nmol/L was regarded as a normal response [25,26]. Participants who did not achieve a peak cortisol response of >430 nmol/L were educated regarding sick day steroid rules, in accordance with Society for Endocrinology recommendations [4]. Baseline ACTH concentration was analysed using the Elecsys ACTH immunoassay (Roche Diagnostics).
The study was reviewed and approved by the Beaumont Hospital Research Ethics Committee (reference 19/72). Written informed consent was obtained from each participant prior to enrolment in the study and the study was conducted in accordance with the ethical standards laid down in the 1964 Declaration of Helsinki.

Statistical analysis
Data are expressed as number (%) for categorical variables and as median [interquartile range (IQR)] for non-parametric continuous variables. The t-test was used to compare single variables and the Mann-Whitney test was used to determine the statistical significance of the analysis of multiple comparisons of non-parametric variables. Fisher's exact test was used to analyse the association between categorical variables. A receiver operating characteristics (ROC) analysis was performed to assess basal cortisol as a predictor of SST response, with an area under the curve (AUC) of 1.0 representing perfect discrimination. A P-value <.05 was considered statistically significant. Statistical analysis was performed using the GraphPad Prism software package (GraphPad Software, San Diego, CA, USA).

Cohort description
A total of 67 participants were recruited between December 2019 and November 2020, 64% of whom were male ( Table 2). The median age of participants was 52 years (IQR 39-64) and the median duration since transplant was 79 months (IQR 46-146). The median duration of dialysis prior to transplantation was 24 months (IQR 11-42). The most common transplant type was a deceased donor (72%) ( Table 2). Twenty-one percent of participants had more than one previous renal transplant. Most participants were also taking a calcineurin inhibitor (81%) and mycophenolate mofetil (82%) as immunosuppressive therapy, with a median tacrolimus level at the time of study visit of 6.6 ng/mL (IQR 5.6-7.8). The primary renal diagnosis in the study cohort was reflective of that of the Irish national KTR population ( Figure 1).

Prevalence of AI
A total of 48 participants (72%) failed an SST, indicating AI. Only 19 participants had a normal response, indicating preserved adrenal function. Participants with AI had a lower eGFR than those with intact adrenal function [63 mL/min/ 1.73 m 2 (IQR 52-8) versus 81 (60-98); P = .03], but there was no significant difference in age, body mass index, duration since transplantation, history of a previous transplant, use of calcineurin inhibitors or mycophenolate mofetil, daily tacrolimus dose (mg/kg), pretransplant glucocorticoid exposure or exposure to other glucocorticoid preparations between the two groups.

Glucocorticoid exposure
Cumulative glucocorticoid exposure was associated with a risk of AI. Participants with AI had a higher median daily dose of prednisolone [4.9 mg/day (IQR 4.7-5) versus 4.2 (3.6-4.8); P = .002], a greater absolute cumulative glucocorticoid Adrenal insufficiency among KTRs
Similarly, basal serum cortisol >288 nmol/L predicts a normal SST response with 100% specificity (95% CI 92-100) 240  and 70% sensitivity (95% CI 56-78) and therefore participants with basal serum cortisol concentration >288 nmol/L may not require SST testing. However, participants with a basal serum cortisol concentration that falls between these threshold values (130-288 nmol/L) would require an SST to exclude the presence of AI. Implementing these threshold values could reduce the number of participants requiring SST by 47%.

Steroid sick day rules
Only nine patients (13%) were aware of the sick day rules and no patient carried a medical alert bracelet or steroid card.

DISCUSSION
This is the largest cross-sectional study to report the risk of AI in KTRs receiving low-dose glucocorticoid-based immunosuppression. It is one of the few adult studies to examine the effect of cumulative glucocorticoid exposure on the risk of AI [10,11] and the only such study to include pretransplant glucocorticoid exposure. Our data would suggest that AI is common among KTRs receiving glucocorticoidbased immunosuppression, with a prevalence of 72%.
Patients receiving glucocorticoid doses ≥5 mg prednisolone or equivalent for at least 1 month are recognized to be at a high risk of AI [27] and are advised to carry steroid alert cards and follow steroid sick day rules [3,27]. Our data suggest that patients with prolonged exposure to even lower doses are at risk of AI. AI was seen in participants receiving doses as low as 2.5 mg of prednisolone/day and on alternateday dosing regimens. AI was seen as early as 14 months following transplantation and normal synacthen responses were seen in participants up to 25 years following transplantation. We would therefore caution against discounting the possibility of AI based on dose or duration of exposure alone.
The prevalence of AI in our cohort is higher than the previously reported rates in KTR cohorts receiving similar daily doses of prednisolone [12,14,16]  Hydrocortisone 100 mg at the start of the procedure and resume enteral double dose prednisolone for 24 hours after the procedure before returning to preprocedure maintenance dose

Labour and vaginal delivery
Hydrocortisone 100 mg i.v. at onset of labour followed by hydrocortisone 50 mg i.v. every 6 hours AI in our cohort was also higher than that reported in patients receiving similar daily doses of glucocorticoids for other diagnoses [2,28]. There are several possible explanations for this, including drug-drug interactions between prednisolone and other immunosuppressants and medication adherence in this transplant population. The interaction between glucocorticoid medications and other immunosuppressants used in KTRs is complex. Both glucocorticoids and calcineurin inhibitors are substrates for cytochrome P450 3A4 (CYP3A4) enzymes and P-glycoprotein (P-gp) [29,30]. Glucocorticoids reduce the bioavailability of tacrolimus through enzymatic induction of CYP3A4 and changes the expression of P-gp [31][32][33]. In vitro studies have suggested that tacrolimus may potentiate the effects of glucocorticoids by enhancing nuclear translocation of the glucocorticoid receptor [34]. Pharmacokinetic studies of prednisolone in KTRs have identified high prednisolone tissue exposure, thought to be due to alterations in corticosteroid metabolism via the 11β-hydroxysteroid dehydrogenase enzymes [35]. Therefore, impaired prereceptor metabolism and regulation of glucocorticoid in target tissues may increase prednisolone exposure, thereby adding to the risk of adrenal insufficiency in KTRs.
It is notable that KTRs with lower eGFR were more likely to have AI in our cohort. Studies examining the effects of im-paired renal function on the pharmacokinetics of prednisolone have primarily been performed in patients with end-stage renal failure, where uraemia, hypoalbuminaemia and reduced protein binding are hypothesized to increase free prednisolone AUCs [29]. Our cohort displayed a relative preservation of renal function. It is unclear from the literature if mildmoderate renal impairment has an effect on prednisolone pharmacokinetics.
Another possible explanation for the observed high prevalence of AI in the KTR population is adherence to medication in this patient group. The Irish kidney transplant programme operates under the public health system, which offers affordable healthcare to KTRs [36]. It should be noted that the Irish kidney transplant programme has almost 100% recipient follow-up, with failure to attend quarterly transplant clinic appointments a rarity [37,38].
Any individual with AI is at risk of an adrenal crisis, which can be life-threatening, irrespective of the aetiology. The reported incidence of adrenal crises in patients with known AI is 5.2-8.3 crises/100 patient-years [39,40], even among patients who have received sick day rules education. Despite this, there are few reported cases of adrenal crises in KTRs in the literature [41,42]. It is important to consider that adrenal crisis may not be suspected due to a lack of awareness of the risk of AI in this group. Clinicians should have a high  clinical suspicion of adrenal crisis in this patient group and treat suspected adrenal crisis without delaying confirmatory testing. KTRs with confirmed AI should be given stress-dose hydrocortisone for an acute illness, diagnostic or therapeutic procedure in the same manner as patients with known AI (Figure 4) [3,4,27].
As the majority of KTRs remain on glucocorticoid medication lifelong as part of an immunosuppressive regimen [5], additional glucocorticoid replacement is rarely necessary on a day-to-day basis [2]. However, KTRs with AI receiving prednisolone doses lower than adrenal replacement doses (<3-5 mg/day) [17] may require regular additional glucocorticoid replacement, the need for which must be weighed against the risk of iatrogenic complications of glucocorticoid excess.
Our results also highlight the need for increased focus on patient and healthcare provider education regarding both the risk of AI in KTRs and the prevention of adrenal crisis [43]. Sick day rules education, medic alert bracelets or steroid cards are recommended for all patients with AI [4,17,27,44]. We would suggest that sick day rules education should form part of routine follow-up care in this patient group. Currently neither steroid sick day rules education nor screening for AI are included in the Kidney Disease: Improving Global Outcomes (KDIGO) guidelines for the care of KTRs [6]. A copy of the sick day rules guidance given to the patients enrolled in this study is available as Supplementary data.
Our data would suggest that there is a significant proportion of KTRs with undiagnosed AI. The gold standard for diagnosis of AI is dynamic stimulatory testing [17], which has resource implications when considering screening a large patient population. We suggest that a morning cortisol concentration, measured on a second-generation immunoassay, is a useful alternative to predict the risk of AI in KTRs dividing patients into three groups based on their risk of AI. Patients with a morning cortisol concentration <130 nmol/L should be considered to have AI and should follow sick day rules. KTRs with a morning cortisol concentration of 130-288 nmol/L should be considered a high risk of AI but require an SST to confirm the presence or absence of AI. As glucocorticoid therapy is likely to continue in this group, a safe alternative to testing would be to employ sick day rules and perform SST if glucocorticoid therapy is to be stopped or reduced. A morning cortisol Adrenal insufficiency among KTRs concentration >288 nmol/L virtually excludes AI in this population and therefore SST is not necessary. Repeat interval screening should be considered in this group, as ongoing exposure to glucocorticoids conveys an ongoing risk of AI.
Interval screening with a basal serum cortisol concentration is appropriate for identifying patients (maintained on immunosuppressive glucocorticoid therapy) who are at risk of AI. However, in the event of discontinuation or reduction of glucocorticoid dose below AI replacement dosing (<3-5 mg prednisolone/day), an SST is required to ensure adequate adrenal function. Should these patients have AI they are at risk of adrenal crisis in the absence of adequate glucocorticoid replacement and should be commenced on physiological replacement glucocorticoids such as hydrocortisone 10 mg in the morning and 5 mg in the afternoon. In this case, we would suggest referral to an endocrinologist to ensure that adequate glucocorticoid replacement is prescribed ( Figure 5).
In conclusion, AI is common in KTRs receiving low-dose prednisolone as immunosuppressive therapy and is largely undiagnosed. A greater awareness of this risk of AI among both KTRs and clinicians caring for this patient group is key in reducing the potential risk of an adrenal crisis and we would advocate that steroid sick day rules education should be considered as part of routine clinical care for all KTRs receiving glucocorticoid-based immunosuppression. A single morning serum cortisol concentration is a useful screening tool in this patient cohort and may help to identify patients at the highest risk of AI. Finally, we would strongly encourage clinicians to perform an assessment of adrenal function prior to stopping glucocorticoid therapy after a prolonged period of exposure or prior to reducing daily glucocorticoid dosing below an adrenal replacement dose.

SUPPLEMENTARY DATA
Supplementary data are available at ndt online.

C ONFLICT OF INTEREST STATEMENT
None declared. The results presented in this article have not been published previously in whole or part, except in abstract format.

DATA AVAIL ABILIT Y STATEMENT
The data underlying this article cannot be shared publicly, as participants of this study did not agree for their data to be shared publicly, so supporting data are not available.