Nelson's syndrome is a potentially life-threatening condition that does not infrequently develop following total bilateral adrenalectomy (TBA) for the treatment of Cushing's disease. In this review article, we discuss some controversial aspects of Nelson's syndrome including diagnosis, predictive factors, aetiology, pathology and management based on data from the existing literature and the experience of our own tertiary centre.
Definitive diagnostic criteria for Nelson's syndrome are lacking. We argue in favour of a new set of criteria. We propose that Nelson's syndrome should be diagnosed in any patient with prior TBA for the treatment of Cushing's disease and with at least one of the following criteria: i) an expanding pituitary mass lesion compared with pre-TBA images; ii) an elevated 0800 h plasma level of ACTH (>500 ng/l) in addition to progressive elevations of ACTH (a rise of >30%) on at least three consecutive occasions. Regarding predictive factors for the development of Nelson's syndrome post TBA, current evidence favours the presence of residual pituitary tumour on magnetic resonance imaging (MRI) post transsphenoidal surgery (TSS); an aggressive subtype of corticotrophinoma (based on MRI growth rapidity and histology of TSS samples); lack of prophylactic neoadjuvant pituitary radiotherapy at the time of TBA and a rapid rise of ACTH levels in year 1 post TBA. Finally, more studies are needed to assess the efficacy of therapeutic strategies in Nelson's syndrome, including the alkylating agent, temozolomide, which holds promise as a novel and effective therapeutic agent in the treatment of associated aggressive corticotroph tumours. It is timely to review these controversies and to suggest guidelines for future audit.
In 1958, Don Nelson et al. described the first case of the eponymous syndrome in a 33-year-old woman who had undergone total bilateral adrenalectomy (TBA) 3 years previously for the treatment of refractory Cushing's disease. She presented with visual field defects, skin hyperpigmentation, raised plasma ACTH and a large sellar mass shown on skull X-ray (1). The sellar mass was a pituitary corticotrophinoma, and its surgical removal led to symptom resolution (1). In 1960, the first series of patients with Nelson's syndrome was published (2), to be followed by further case series (Table 1).
A summary of the main case series for Nelson's syndrome, demonstrating the heterogeneity of clinical and pathological features, and the diagnostic criteria used.
|Case study||Patients (n)||Follow-up (years)||Nelson's syndrome (%)||Tumour growth (%)||Plasma ACTH level (ng/l)||NF (%)||SH (%)||Diagnostic criteria for Nelson's syndrome|
|(4)||32 (children)||1–5.5||25||25||NR||6 (VFDs)||56||CD treated with TBA; hyperpigmentation; Sella turcica enlargement|
|(20)||50||1–22 (12)||28||24||450–8000 RI assay||12 (VFDs)||28||CD treated with TBA; hyperpigmentation; radiological evidence of pituitary tumour or VFDs|
|(13)||38||1–20 (10)||29||29||>240 RI assay||5 (VFDs)||84||CD treated with TBA; pituitary enlargement; hyperpigmentation; increased plasma ACTH|
|(38)||16||NR||25||25||NR||NR||63||CD treated with TBA; hyperpigmentation|
|(30)||30||2–14||47||47||>900†||3 (VFDs)||57||CD treated with TBA; pituitary enlargement; raised plasma ACTH; hyperpigmentation|
|(3)||44||Mean: 11||23||9||500–3000||NR||34||CD treated with TBA; radiological evidence of nonexpanding pituitary microadenoma; visual field defects; raised plasma ACTH; hyperpigmentation|
|(31)||39||1||28||13||220–3900 (1163)||NR||28||CD treated with TBA; enlargement of previous pituitary tumour or development of a new tumour; raised plasma ACTH; hyperpigmentation|
|(15)||53||11 (4.6)||32||32||133–3840 (1269)IR assay||2||NR||CD treated with TBA; enlargement of previous pituitary tumour or development of a new tumour; raised plasma ACTH|
VFD, visual field defect; NR, not recorded; CD, Cushing's disease; TBA, total bilateral adrenalectomy; NF, Neuro-ophthalmic feature; SH, skin hyperpigmented; RI, radioimmuno assay; IR, immuno radiometric assay. Figures in parentheses are mean values. †RI assay before 1991 and IR assay after 1991.
Nelson's syndrome does not occur infrequently and is a potentially life-threatening complication of TBA for the treatment of refractory Cushing's disease with an incidence of 8–43% in adults (3) and 25–66% in children (4, 5). Nelson's syndrome can develop up to 24 years post TBA, with mean development at 15 years post TBA (3). In the first published series of patients with Nelson's syndrome, mortality was 12% (6), although mortality rates have been much lower in the subsequent series (including absent mortality in two series (7, 8)), probably due to earlier diagnosis and improved management. In the pre-computed tomography (CT)/magnetic resonance imaging (MRI) era, Nelson's tumours were often detected late through clinical manifestations of invasion and compression of surrounding structures. MRI has enabled early detection of enlarging corticotrophinomas post TBA. The central pathological feature of Nelson's syndrome is an underlying ACTH-secreting pituitary adenoma (corticotrophinoma). There is evidence that positive immunostaining for ACTH within pituitary tumours is associated with a more aggressive behaviour of the tumour. In one study, it was shown that non-functioning pituitary adenomas with positive immunoreactivity for ACTH behaved more aggressively than immunonegative tumours (9). Furthermore, it has also been shown that ACTH-secreting pituitary macroadenomas are often more invasive than other pituitary adenomas (10).
The treatment options for Cushing's corticotrophinomas that are refractory to transsphenoidal surgery (TSS) include i) repeat TSS; ii) drugs that block cortisol production; iii) pituitary irradiation and iv) TBA (11). Although generally out of favour (12), laparoscopic TBA can be useful when a corticotrophinoma is undetectable, surgically unresectable or there is recurrence following TSS. In these cases, TBA offers 85–100% success rate in controlling hypercortisolaemia (3, 11) and is particularly useful in patients with life-threatening manifestations of hypercortisolism (13). Although TBA is usually a definitive form of treatment for refractory Cushing's disease with immediate control of hypercortisolaemia, a limitation is the potential for the subsequent development of Nelson's syndrome (14).
The pathophysiology of Nelson's syndrome and the factors leading to its onset and progression are poorly understood. Its effective management is a challenge. In this review, we address some key unanswered questions relating to Nelson's syndrome:
How should Nelson's syndrome be diagnosed?
What factors predict the onset of Nelson's syndrome and influence subsequent progression?
What is the aetiology of Nelson's syndrome?
What are the pathological features of corticotrophinomas in Nelson's syndrome?
How should Nelson's syndrome be managed?
How should Nelson's syndrome be diagnosed?
Clinical, biochemical and radiological features
Historically, Nelson's syndrome usually presented with late clinical manifestations of an enlarging pituitary corticotrophinoma including frank visual field defects (15, 16, 17) and cranial nerve palsies (8), due to tumour compressive effects and invasion into surrounding structures. Such late features are less frequent at presentation of Nelson's syndrome in modern times, but hyperpigmentation of skin and mucous membranes (particularly on extensor surfaces, flexures, over scars and on the areolae (Fig. 1)) remains an important clinical manifestation. Metastases from a corticotroph carcinoma are unusual (16, 18, 19). Other clinical features include headaches, pituitary apoplexy (17, 20), diabetes insipidus (21), panhypopituitarism, testicular pain and oligospermia (22). Nelson's syndrome occasionally presents with paraovarian tumours (23) or paratesticular tumours (20) from hyperstimulation of adrenal rest cells within gonadal tissues (24). This atypical complication of adrenal rest cell hyperplasia/stimulation can sometimes masquerade as a relapse of Cushing's disease and be mistaken for this (22).
The availability of sensitive ACTH assays and high-resolution MRI and CT imaging modalities has enabled earlier diagnosis of Nelson's syndrome through identification of an enlarging pituitary corticotrophinoma (Figs 2 and 3) associated with elevated and rising plasma ACTH levels. Despite this, hyperpigmentation still occurs in up to 42% (25) and visual field defects in 10–57% of patients with Nelson's syndrome (26). The heterogeneity of the clinical presentation of Nelson's syndrome is demonstrated in Table 1.
The most consistent and reliable biochemical feature of Nelson's syndrome is a marked elevation of plasma ACTH, which continues to rise after adrenalectomy. In a comparison of nine patients with Nelson's syndrome and nine patients with Cushing's disease, basal secretion of ACTH was up to six fold higher and volume of pulsatile ACTH secretion was nine fold larger in Nelson's syndrome (27). Surprisingly, ACTH secretion followed a more orderly sequence in Nelson's syndrome compared with Cushing's disease and diurnal variation was maintained (27). The timing of collection of the blood sample for ACTH measurement in relation to the last steroid dose is important: the blood sample should be taken at 0800 h, 20 h following the last dose of glucocorticoid and before the morning dose of glucocorticoid (28). Plasma ACTH is sometimes also measured 2 h following the morning dose of glucocorticoid.
The radiological features of Nelson's syndrome are those of an enlarging pituitary mass. It is imperative that comparisons of pituitary appearances are made with those from a post-TSS scan (26). MRI provides the best imaging modality, with detection of tumours as small as 3 mm. High resolution (3 mm cuts) should be used so that early tumour progression can be identified (26). The onset of Nelson's syndrome has been described as late as 24 years after TBA (15). Therefore, lifelong pituitary imaging should be performed on all patients following TBA treatment of refractory Cushing's disease, and any increase in tumour size in this scenario is significant. The first post-TBA pituitary scan should be performed at 3 months (26). Our Oxford-based practice is to repeat MRI scans at 6-month intervals for 2 years and thereafter yearly, based on the observations that 20% of Nelson's tumours develop in the first year and 35% develop in the first 2 years after TBA (15). This protocol will provide data for future audit.
The diagnosis of Nelson's syndrome is controversial, and a consensus opinion is lacking. This hampers direct comparison and merging (to improve power) of data between case series. Ultimately, multiple sets of diagnostic criteria also adversely affect patient care. Existing diagnostic criteria for Nelson's syndrome focus on an enlarging pituitary tumour post TBA, rising plasma ACTH levels and presence of hyperpigmentation. While we accept that development of hyperpigmentation is a useful clinical sign of Nelson's syndrome, its inclusion as a diagnostic criterion poses problems. First, there is a lack of evidence relating to a close correlation between the ACTH levels and intensity of skin pigmentation, and a suggestion that in some patients, hyperpigmentation may only develop in the context of extremely high ACTH levels (29), raising issues regarding sensitivity. Conversely, the objectivity of measurements derived from pituitary MRI and ACTH assays lend themselves well for inclusion as diagnostic criteria, particularly as Nelson's syndrome should usually be diagnosed in the context of an enlarging pituitary corticotrophinoma post TBA surgery (15). It is clearly important to demonstrate pituitary tumour expansion rather than just the presence of a remnant (26).
Based on the available literature (outlined in Table 1), we propose a new set of diagnostic criteria for the diagnosis of Nelson's syndrome for discussion (Table 2). We hope that these new diagnostic criteria will supersede the existing multiple sets (Table 1), for application in both clinical and research spheres. A single set of diagnostic criteria will assist in comparison of data between future case series and will optimise our clinical approach to patients with Nelson's syndrome. We believe it to be imperative that every patient with Cushing's disease has close follow-up surveillance with MRI of the pituitary post TBA surgery. This practice should be widely and routinely adopted. Pituitary remnant tumour re-growth forms an important element of Nelson's syndrome, and the early detection of tumour re-growth (prior to development into a macroadenoma) through MRI surveillance is essential so that appropriate therapeutic manipulation can be initiated.
Suggested criteria for diagnosis of Nelson's syndrome.
|For a diagnosis of Nelson's syndrome to be made, a patient must have had prior treatment with total bilateral adrenalectomy (TBA) for Cushing's disease (regardless of prior transsphenoidal surgery) in addition to at least one of the following two criteria (based on Refs (13) and (15)):|
|An expanding pituitary mass lesion post TBA surgery (shown on MRI or CT scan) compared with MRI of the pituitary prior to TBA surgery|
|An elevated level of ACTH to >500 ng/l from a single plasma sample taken at 0800 h prior to steroid administration and post TBA surgery, in addition to progressive elevations of ACTH levels from plasma samples taken on at least three consecutive occasions at different time-points post TBA surgery (a rise of ACTH by >30% of the initial post-TBA sample)|
The cut-off for plasma ACTH to define Nelson's syndrome is controversial, with a level >200 ng/l (NR 0–80 ng/l) used by some authors (26). In one study, ACTH levels >700 ng/l post TBA for refractory Cushing's disease only occurred in those patients who developed Nelson's syndrome (15). In three of the largest series on Nelson's syndrome, levels of ACTH in those patients who developed Nelson's syndrome were >500 ng/l (3), >450 ng/l (20) and >900 ng/l (30), and often >1000 ng/l (Table 1). In two other series, mean ACTH level in those patients who developed Nelson's syndrome was >1100 ng/l (15, 31). These data form the rationale for our proposed cut-off for ACTH level at 500 ng/l. We also include patients with three consecutive elevations of plasma ACTH levels post TBA surgery (a rise of ACTH >30% of the initial post-TBA sample).
Our proposed new diagnostic criteria for Nelson's syndrome are set out in Table 2. Crucially, a prerequisite requirement for a diagnosis of Nelson's syndrome is prior TBA surgery in a patient with pre-existing Cushing's disease. We propose that in this scenario, a diagnosis of Nelson's syndrome can be made if at least one of the following two criteria is satisfied: i) an expanding pituitary mass lesion post TBA surgery (shown on MRI or CT scan) compared with MRI of the pituitary prior to TBA surgery; ii) an elevated level of ACTH >500 ng/l from a single plasma sample taken at 0800 h prior to steroid administration and post TBA surgery, in addition to progressive elevations of ACTH levels from plasma samples taken on at least three consecutive occasions at different time-points post TBA surgery (a rise of ACTH by >30% of the initial post-TBA sample). The second biochemical criterion takes account of some uncertainty regarding a fixed cut-off level of ACTH based on current evidence, and incorporates what we believe to be important in Nelson's syndrome: progressive elevations of plasma ACTH. Our suggestion of a 30% rise in ACTH is arbitrary and is designed to avoid a diagnosis of Nelson's syndrome being made on the basis of relatively small and clinically insignificant progressive rises in ACTH that may occur through pure chance or variability within the assay used. We suggest that pre-steroid plasma levels of ACTH are checked immediately post TBA surgery, then at 3-months post TBA, then at 6-month intervals for the first 2 years, and yearly thereafter. The adequacy of the cut-off values for ACTH proposed here should be a focus for future audit, and the criteria should then be modified accordingly.
What factors predict the onset of Nelson's syndrome and influence subsequent progression?
One key unanswered question relates to the identification of those post-TBA patients who are most at risk of subsequent development of Nelson's syndrome. A clearer understanding of predictive factors would enable a more targeted approach for close surveillance directed towards those patients most at risk of early corticotrophinoma development, and would inform decisions regarding TBA surgery. Potential predictive factors for the onset and progression of Nelson's syndrome are discussed below.
ACTH levels in the first post-operative year
A rapid rise in plasma ACTH levels in the year following TBA may be associated with a greater risk of developing Nelson's syndrome (8, 30, 32), and is probably the best validated predictive factor currently available (probably due to its association with corticotrophinoma progression (15)). However, up to 42% of post-TBA patients have persistently elevated (>200 ng/l) plasma ACTH levels in long-term follow-up studies (8), but only a proportion of these go on to develop Nelson's syndrome calling into question the utility of post-TBA ACTH levels as a predictive factor. It has been suggested that an increase in ACTH >100 ng/l in the first post-operative year may be a useful predictor of Nelson's syndrome (15), although this needs further study. Furthermore, there is no evidence that pre-TBA (baseline) plasma ACTH levels or ACTH levels following metyrapone suppression are predictive of Nelson's syndrome development post TBA, which is also an area for future research.
Radiological evidence of residual pituitary tumour prior to TBA
The presence of residual pituitary tumour following TSS for Cushing's disease (pre-TBA) has been shown to be predictive for the subsequent development of Nelson's syndrome post TBA in most studies (7, 30, 33, 34). In one series, 30% of patients with residual tumour at TSS subsequently developed Nelson's syndrome compared with 26% in those patients without residual tumour post TSS (31). In a further study on 30 patients who had undergone TBA for refractory Cushing's disease, 30% of those who developed Nelson's syndrome had radiological evidence of residual pituitary tumour at TBA (group 1) compared with 17% of those who did not subsequently develop Nelson's syndrome (group 2). On surgical exploration, the proportion with macroscopically visible residual tumour was shown to be 55 and 33% in groups 1 and 2 respectively, illustrating the point that the presence of residual pituitary tumour in these subjects is nearly twice as high following surgical exploration as that suggested on imaging (7).
Surveillance imaging and detection of residual pituitary tumour (and therefore the predictive value of radiology) are limited by the resolution of MRI. Our protocol scan uses T1W spin-echo sequences performed at 1.5 T in coronal and sagittal planes as 3 mm thick contiguous slices (with or without gadolinium enhancement). An alternative approach is to use a T1W gradient-echo technique with a 3D Fourier synthesis. Comparison with previous scans is mandatory. It is possible that future development of computer-assisted 3D MRI volume measurement technology will enable the detection of pituitary tumour development post TBA surgery to be more objective, improving its predictive value for Nelson's tumour development.
Administration of neoadjuvant radiotherapy post TBA surgery
This may delay and possibly protect against development of Nelson's syndrome (7, 13, 31). A study on 39 patients followed up over a period of 15 years post TBA surgery demonstrated that for patients who received neoadjuvant radiotherapy, none developed Nelson's syndrome compared with 50% of those who did not receive neoadjuvant radiotherapy (31). In a further study, 25% of those patients receiving prophylactic pituitary radiotherapy developed Nelson's syndrome compared with 50% of those who did not (7). However, there is a controversy regarding the efficacy of neoadjuvant radiotherapy post TBA surgery, with some series suggesting that lack of prophylactic pituitary radiotherapy is not a risk factor for the subsequent development of Nelson's syndrome (8, 16, 33, 35). There is no current consensus on the use of neoadjuvant pituitary radiotherapy post TBA in patients with Cushing's disease. Although neoadjuvant pituitary radiotherapy may reduce the occurrence of Nelson's syndrome in those patients with residual pituitary tumour (7), the potential benefits of this procedure need to be weighed carefully with a high probability of the development of adverse sequelae following the administration of pituitary radiotherapy. A possible alternative to the use of pituitary radiotherapy in those Cushing's disease patients with pituitary remnant tissue at the time of TBA is the use of radiosurgery, which should also be a focus for future research. Our practice in Oxford is to administer neoadjuvant pituitary radiotherapy to those patients with pituitary remnant tissue who undergo TBA surgery, but not to administer it in those patients who do not have post-operative residual pituitary tissue. Again, this appraisal will be audited in the future.
Residual adrenal remnant after TBA
There is very little evidence to support the idea that the presence of a small adrenal remnant post TBA provides any protection from the subsequent development of Nelson's syndrome (35). Adrenal remnant tissue also leaves the potential for recurrence of hypercortisolaemia and its sequelae if ACTH levels remain high. This was demonstrated in a large study of Cushing's disease patients treated with TBA in which 12 (27%) had adrenal remnants. Of these 12 patients, 2 developed early recurrence of Cushing's disease from hyperfunctioning adrenal remnant tissue (3).
Duration of Cushing's disease prior to TBA
Duration of Cushing's disease prior to TBA has been associated with the risk of future development of Nelson's syndrome (15). In one study of 43 patients, followed up for a median of 10 years, the patients who went on to develop pituitary enlargement and hyperpigmentation had symptoms of Cushing's disease for twice as long (5 years) prior to TBA compared with those who developed hyperpigmentation alone or neither hyperpigmentation nor pituitary enlargement (13). However, one study on just seven patients unsuccessfully treated with pituitary surgery for Cushing's disease who subsequently underwent TBA showed no correlation between Cushing's disease duration and likelihood of Nelson's syndrome development (36). Duration of Cushing's disease prior to TBA would seem plausible as a predictive factor for Nelson's syndrome, although clearly more data are required.
A younger age at TBA seems to be predictive for the subsequent development of Nelson's syndrome (16, 37, 38) with children being at particularly high risk (4, 5). One possible explanation for this is that children may harbour more aggressive subtypes of corticotrophinomas (39). However, one study showed that although 50% of patients developed Nelson's syndrome within 3 years of TBA, age did not seem to be predictive (15). Once again, further studies are required for validation.
High urinary cortisol
For some tumours (although not microadenomas), pre-TBA urinary cortisol serves as a useful marker of corticotrophinoma tumour size and functionality, and may be a useful predictor of the subsequent Nelson's syndrome development (3, 33, 40). However, the evidence is not conclusive, and a correlation between urinary cortisol levels and the risk of subsequent development of Nelson's syndrome has not been demonstrated in some studies (13, 30, 32). Pre-operative urinary cortisol levels cannot therefore be advocated as a reliable predictive factor for the subsequent development of Nelson's syndrome.
Insufficient exogenous steroid replacement therapy post TBA surgery
Standard steroid replacement doses are recommended for patients following TBA surgery for recurrent Cushing's disease. Suboptimal or absent steroid replacement therapy post TBA may increase the risk of developing Nelson's syndrome (41, 42) or even result in transformation of a pre-existing adenoma into a more aggressive tumour subtype (3, 26). However, most studies have not shown this (3, 13, 32).
Lack of cortisol suppression on high-dose dexamethasone prior to TBA surgery
We hypothesise that the 20% of patients who lack suppression of cortisol on high-dose dexamethasone before initial hypophysectomy may predict the subsequent development of Nelson's syndrome. In one study on 48 patients who underwent TBA surgery for Cushing's disease, pre-TBA cortisol suppressibility following the administration of dexamethasone (8 and 16 mg) was not predictive for the subsequent development of Nelson's syndrome (37). However, this was a relatively small study with only eight patients who developed Nelson's syndrome (37). This hypothesis has not been assessed in a sufficiently large cohort and should be a focus for future research.
In summary, the following four factors emerge as potentially useful predictors of Nelson's syndrome post TBA: i) presence of residual pituitary tumour on MRI post TSS; ii) an aggressive subtype of corticotrophinoma defined on the basis of growth rapidity (MRI comparisons) and local tissue invasion (histological observations on samples taken during prior-TSS), discussed further in section entitled ‘What are the pathological features of corticotrophinomas in Nelson's syndrome?’; iii) lack of prophylactic neoadjuvant pituitary radiotherapy administered concurrently with TBA; and iv) rapid rise of ACTH levels in year 1 post TBA. As the reliability of these predictive factors is not well established, we currently advocate close long-term follow-up in all Cushing's disease patients post TBA.
What is the aetiology of Nelson's syndrome?
The aetiology of Nelson's syndrome is incompletely understood. One hypothesis is that a release of the suppressive negative-feedback effects of elevated endogenous cortisol levels on corticotroph cells following TBA, with concurrent restoration of hypothalamic CRH production (43) is sufficient to drive progressive corticotroph neoplasia in some patients. This is analogous to the rare development of pituitary neoplasia in occasional patients with Addison's disease (44). Evidence is lacking for an alternative hypothesis that de novo generation of a new corticotrophinoma (clonally different from the previous Cushing's corticotrophinoma) occurs following TBA.
Observations in rat models are consistent with the ‘released negative-feedback’ hypothesis for development of Nelson's syndrome in humans: adrenalectomised rats develop raised CRH levels (45), raised proopiomelanocortin (POMC) gene products and corticotroph cell hyperplasia (46); chronic CRH infusion results in corticotroph cell hyperplasia (47); administration of dexamethasone results in apoptosis of cells in the anterior portion of the pituitary (48); and, in adrenalectomised mice, increased hypothalamic-derived arginine vasopressin has been demonstrated (49) which in turn induces proliferation of murine corticotroph cell lines in vitro (50). In humans with Nelson's syndrome, stimulation of ACTH release has been demonstrated in response to vasopressin (51) and GNRH (52). However, there is less evidence in support of the ‘released negative-feedback’ hypothesis in humans. Although cortisol has been shown to lower the proliferation rate of human pituitary adenoma cells in vitro (53), secretion of POMC-derived peptides from patients with Cushing's disease is not attenuated by glucocorticoids (54). Furthermore, corticotrophs from Nelson's syndrome tumours in humans exhibit an attenuated negative-feedback response to glucocorticoids in vivo (55). The number of membrane-bound CRH receptors is increased in human pituitary corticotrophinomas (56), and the corticotrophs of Nelson's adenomas have been shown to be variably sensitive to CRH administration in both in vivo (57) and in vitro studies (58). Therefore, human evidence for the ‘released negative-feedback’ hypothesis for development of Nelson's syndrome seems to be lacking. Furthermore, there are inconsistencies in that not all post-TBA patients develop Nelson's syndrome after long-term follow-up, and most patients who develop Nelson's syndrome also have adequate exogenous steroid replacement therapy.
An important clue for the aetiology of Nelson's syndrome is the observation that this condition occurs in those patients with Cushing's disease and aggressive forms of corticotrophinomas (15). This subgroup of patients also tend to be those who relapse following TSS and who subsequently require TBA. Although the majority (52–85%) of patients treated with TSS for Cushing's disease go into remission (59), recurrence rates within a year post TSS are between 7 and 12% (60). Nelson's syndrome can develop up to 24 years following TBA (30, 33, 38, 61). It has also been demonstrated that patients with invasive corticotrophinomas (shown on TSS-derived samples) are at greater risk of subsequent (and earlier) development of Nelson's syndrome (62) compared with less aggressive forms (3, 31). Although corticotrophinoma aggressiveness is an important concept when considering Nelson's syndrome aetiology, other factors may also be implicated. Further research is needed in this area including cellular-based in vitro studies on the characteristics of corticotrophinomas (such as CRH/glucocorticoid receptor number and function (61)) that progress to Nelson's syndrome.
What are the pathological features of corticotrophinomas in Nelson's syndrome?
The corticotrophinomas of Nelson's syndrome and Cushing's disease originate from the same corticotroph cells: POMC mRNA is identical to that expressed in normal corticotroph cells and is processed in the usual way (63). There are histological and molecular similarities between these two tumour types (61, 64). The corticotroph cells in Nelson's syndrome tumours are monoclonal in origin (65), and in both Nelson's and Cushing's corticotrophinomas, expression of functional CRH and vasopressin V3 receptors persists (66), with some evidence that receptor numbers are increased (56). Two isoforms of glucocorticoid receptor are expressed in corticotroph cells, and these are retained in Nelson's syndrome (67). Loss of heterozygosity at the glucocorticoid receptor locus has been demonstrated in some Nelson's syndrome tumours (68).
Patients with Cushing's disease and aggressive subtypes of corticotrophinoma are more likely to develop Nelson's syndrome post TBA (30). It is possible that histopathological and immunocytochemical appearances (percentage of mitoses and Ki67-immunopositive nuclei) of corticotrophinoma specimens derived during TSS in patients with Cushing's disease would be a useful predictive marker for the subsequent development of Nelson's syndrome. However, there are insufficient data concerning the biological nature of pre-TBA corticotrophinomas in patients who develop Nelson's syndrome compared with those who do not develop Nelson's syndrome. It is also unclear whether corticotrophinomas removed at the time of onset of Nelson's syndrome are inherently more aggressive than sporadic corticotrophinomas. It is likely that the stage of tumour progression at the time of surgery for Nelson's tumours is different today than it was in the pre-MRI era, giving a historical bias towards more aggressive tumours. Compared with Cushing's disease, Nelson's tumours have a tendency towards being macroadenomas and more invasive (69). Nelson's tumours also have significantly higher proliferation (1.1 vs 0.5%) and lower p27 labelling indices (13 vs 28%) than Cushing's adenomas (69). Although there was no difference in p53 expression, comparison of p53 expression between invasive and non-invasive Nelson's tumours showed a positive correlation with invasiveness. In 12 patients with Cushing's disease, mitotic figures or Ki-67-expressing nuclei in corticotroph adenomas at baseline did not predict adenoma progression post TBA (15).
Further evidence for the aggressiveness of Nelson's adenomas comes from the analysis of pituitary carcinomas (with craniospinal or systemic dissemination) which are rare (0.1–0.2% of pituitary tumours (70)) and are usually functional. In one series, four out of seven corticotroph carcinomas arose in the context of Nelson's syndrome with an average interval of 15.3 years (18). Landman et al. showed that of the 33 ACTH-producing pituitary carcinomas, 15 (45%) occurred in the context of Nelson's syndrome (71). Morphologically, Nelson's tumour cells may have unique ultrastructural features such as inconspicuous type 1 cytokeratin filaments (unlike in Cushing's disease corticotrophinomas) (65), although this is not our experience (Fig. 4). Nelson's corticotrophinomas may have larger cells with significant pleomorphism (3, 72) (Fig. 4).
How should Nelson's syndrome be managed?
Pituitary surgery should be the first-line treatment option for Nelson's syndrome, particularly if there is compression of the optic apparatus (25, 42, 73) with success rates ranging from 10 to 70% (10, 25, 55, 74). Although the usual approach is transsphenoidal, a transcranial procedure may be adopted if there is extrasellar extension (33% of cases (10)). Successful surgery (with long-term remission and fewer complications) is more likely with small tumours confined to the sellar region (74). Overall mortality following surgical management of Nelson's syndrome is 5% (25, 73), and morbidity rates are high: up to 69% develop panhypopituitarism (26); 5% acquire a cranial nerve palsy; 15% have a cerebrospinal fluid (CSF) leak; and 8% develop meningitis (25, 73).
Progression of Nelson's syndrome tumours may occur, despite surgical intervention in some patients (25) and adjuvant radiotherapy may be required in 20–30% of such patients (25, 74, 75, 76). In a follow-up study of 15 patients who had been treated with adjuvant pituitary radiotherapy for surgically refractory Nelson's syndrome, 93% had amelioration of skin pigmentation (reflecting reduced levels of POMC products (77)). The available evidence base for administering adjuvant pituitary radiotherapy for all Cushing's disease patients following TBA surgery is limited. However, this approach would seem reasonable for those patients with any remnant corticotroph tumour tissue. Further administration of radiotherapy needs to be balanced with increased complications resulting from multiple courses of radiotherapy.
The limited data on the use of gamma knife surgery (GKS) in Nelson's syndrome are conflicting (78, 79, 80, 81). Although one study of Nelson's syndrome patients showed no tumour re-growth at 7 years post GKS therapy (82), a further study showed remission rates post GKS to be only 14% (83). GKS seems to be most effective in Nelson's syndrome when it is administered soon after TBA (82) and when the anatomical target is clear and discrete (less likely with prior surgery, which may render the tumour border indistinct (78, 81)). Potential adverse effects of GKS include panhypopituitarism (up to at least 7 years post surgery (84)) and cranial nerve palsies. GKS is therefore not suitable for tumours adjacent to the optic apparatus and those invading the cavernous sinus. The precise role of GKS in the treatment of Nelson's syndrome remains to be established.
Selective somatostatin analogues
Selective somatostatin analogues (SSAs) potentiate specific G-protein-coupled membrane-bound somatostatin receptors (SSTRs 1–5) (85). SSAs may decrease plasma ACTH levels (86) and reduce tumour volume (87) in some patients with Nelson's syndrome. Pasireotide (SOM 230), which has affinity to SSTRs 1, 2, 3 and 5, has been shown to inhibit ACTH secretion in vitro from human corticotroph cells taken from patients with Cushing's disease (88). Pasireotide has also been shown to inhibit ACTH release in vivo (89) and reduce proliferative rate of human corticotroph cells (85). Unfortunately, there is, as yet, no good human evidence currently to support the use of SSAs in patients with Nelson's syndrome. Clearly, more data are required to inform appropriate placement of SSAs in the treatment of Nelson's syndrome.
Peroxisome proliferator-activated receptor γ agonists
Peroxisome proliferator-activated receptor γ (PPARγ or PPARG) receptors are expressed in normal corticotrophs and are highly concentrated in corticotroph adenoma cells (90). It has been shown that murine corticotrophs that are cultured in vitro with the PPARγ ligand rosiglitazone transcribe fourfold less POMC mRNA (90). In a mouse model of Cushing's disease, rosiglitazone at a dose of 150 mg/kg per day caused cell cycle arrest, apoptosis and reduced ACTH secretion from corticotroph cells (90). However, human studies have provided no good evidence that rosiglitazone is beneficial in Nelson's syndrome. Although rosiglitazone was shown to reduce urinary free cortisol in 50% patients with hypercortisolaemia in one study (91), further studies on seven patients with Nelson's syndrome who received rosiglitazone 8 mg od (92) and on six patients with Nelson's syndrome who received rosiglitazone 12 mg od (93) demonstrated no lowering of ACTH levels. PPARG agonists remain unlicensed for the treatment of Nelson's syndrome, and the possible cardiovascular risks associated with rosiglitazone therapy (94) may impede further studies in this area.
Inhibition of reuptake of gamma amino butyric acid (GABA) (through the use of sodium valproate) within the hypothalamus results in decreased CRH release (85). It has therefore been postulated that GABA reuptake inhibition may be a useful therapeutic target in Nelson's syndrome (85). However, there is no evidence that sodium valproate is efficacious (86), with no lowering of plasma ACTH levels (85, 86) or reduced corticotrophinoma growth (85) when used in patients with Nelson's syndrome.
Pituitary adenomas express dopamine receptors ubiquitously, providing the rationale for the use of dopamine agonist (DA) therapies in the treatment of non-prolactinoma pituitary tumours. Bromocriptine decreases ACTH levels in Nelson's syndrome (85), and cabergoline can induce remission of Nelson's syndrome (95, 96) and tumour resolution (97). Further studies are required before widespread use of DAs can be advocated in patients with Nelson's syndrome. Further cellular studies are required to ascertain the identities of the different subclasses of dopamine receptors present on the surface of corticotrophinoma cells in patients with Nelson's syndrome, to enable development of novel therapies that target these receptors (98).
Temozolomide is an alkylating agent. There is some evidence from case reports that temozolomide may be effective in the management of some patients with aggressive pituitary tumours. One such report suggested that the immunoexpression of O(6)-methylguanine DNA methyltransferase (MGMT), a DNA repair protein, could predict the response to temozolomide therapy in patients with aggressive pituitary tumours (99). In patients with re-growth of non-functioning pituitary adenomas, low MGMT expression was shown to correlate with tumour responsiveness to temozolomide (100). A recent case report of an aggressive Nelson's tumour, which failed to respond to both surgical and radiotherapy (including gamma knife treatment), revealed an excellent response to temozolomide with a significant reduction in plasma ACTH and regression of the underlying tumour after just four cycles of treatment (101). There is clearly a need for larger studies on the use of temozolomide in patients with Nelson's syndrome, although we believe that it represents a promising treatment for this condition.
Surgery (with consideration of adjuvant pituitary radiotherapy) remains the treatment of choice for patients with Nelson's syndrome. Medical therapies (including SSAs and temozolomide) and their placements remain uncertain due to lack of current evidence, and this should be a focus for future research. A summary of a suggested screening and management protocol for patients with Nelson's syndrome is shown in Fig. 5.
Suggested future developments
There are a number of areas that could be addressed to resolve some existing controversies. Given that progression of corticotrophinomas post TBA is a central feature of Nelson's syndrome, there is a need for more accurate measurement of corticotrophinoma volume for the use in temporal comparisons. Visual comparison of tumours between scans remains a subjective process. A more objective technique would be to use 3D computer-aided measurements of pituitary tumour volumes for comparison. Future research possibly in a multicentre trial is also needed to clarify the predictive factors for development of Nelson's syndrome post TBA, and to establish and verify their role in clinical decision making. In particular, more research at the cellular level is needed to identify histopathological and immunocytochemical features of corticotrophinomas that are most susceptible to progression following TBA. One such example is the quantification of CRH-receptor and glucocorticoid-receptor densities within Nelson's tumours, and to explore associations of receptor densities with clinical and pathological features (such as tumour growth). Finally, development of novel therapeutic agents based on cellular characteristics (including specific corticotroph receptor subtypes) should be a priority.
Nelson's syndrome is one of the most challenging of all endocrine conditions. Although mortality rates have declined, this remains a condition associated with significant morbidity. The frequent aggressiveness of the underlying corticotrophinoma justifiably necessitates close screening. Its rarity and complexity, coupled with existing controversies regarding diagnosis and management, make Nelson's syndrome a difficult condition on which to build a robust and reliable evidence base. The diagnostic criteria outlined in this review should facilitate the merging of datasets and should enable group collaboration to generate clearer data to address the unanswered questions. Challenges for the future include resolution of existing controversies regarding aetiology, prediction of onset, diagnosis and effective management, which will ultimately improve the lives of patients who suffer from Nelson's syndrome.
Declaration of interest
The authors declare that there is no conflict of interest that could be perceived as prejudicing the impartiality of the review reported.
This review did not receive any specific grant from any funding agency in the public, commercial or not-for-profit sector.
We acknowledge all the patients, relatives, nurses and physicians who contributed to the ascertainment of the various clinical samples reported on in this review.
JenkinsPJTrainerPJPlowmanPNShandWSGrossmanABWassJABesserGM. The long-term outcome after adrenalectomy and prophylactic pituitary radiotherapy in adrenocorticotropin-dependent Cushing's syndrome. Journal of Clinical Endocrinology and Metabolism199580165–171doi:10.1210/jc.80.1.165.
BradleyKJWassJATurnerHE. Non-functioning pituitary adenomas with positive immunoreactivity for ACTH behave more aggressively than ACTH immunonegative tumours but do not recur more frequently. Clinical Endocrinology20035859–64doi:10.1046/j.1365-2265.2003.01674.x.
De TommasiCVanceMLOkonkwoDODialloALawsER. Surgical management of adrenocorticotropic hormone-secreting macroadenomas: outcome and challenges in patients with Cushing's disease or Nelson's syndrome. Journal of Neurosurgery2005103825–830doi:10.3171/jns.2005.103.5.0825.
PorterfieldJRThompsonGBYoungWFChowJTFryrearRSvan HeerdenJAFarleyDRAtkinsonJLDMeyerFBAbboudCFNippoldtTBNattNEricksonDVellaACarpenterPCRichardsMCarneyJALarsonDSchleckCChurchwardMGrantCS. Surgery for Cushing's syndrome: an historical review and recent ten-year experience. World Journal of Surgery200832659–677doi:10.1007/s00268-007-9387-6.
AssiéGBahurelHCosteJSilveraSKujasMDuguéMKarrayFDoussetBBertheratJLegmannPBertagnaX. Corticotroph tumor progression after adrenalectomy in Cushing's disease: a reappraisal of Nelson's syndrome. Journal of Clinical Endocrinology and Metabolism200792172–179doi:10.1210/jc.2006-1328.
van AkenMOPereiraAMvan den BergGRomijnJAVeldhuisJDRoelfsemaF. Profound amplification of secretory-burst mass and anomalous regularity of ACTH secretory process in patients with Nelson's syndrome compared with Cushing's disease. Clinical Endocrinology200460765–772doi:10.1111/j.1365-2265.2004.02052.x.
KimuraNIshikawaTSasakiYSasanoNOnoderaKShimizuYKimuraISteinerDFNaguraH. Expression of prohormone convertase, PC2, in adrenocorticotropin-producing thymic carcinoid with elevated plasma corticotropin-releasing hormone. Journal of Clinical Endocrinology and Metabolism199681390–395doi:10.1210/jc.81.1.390.
KeminkLPietersGHermusASmalsAKloppenborgP. Patient's age is a simple predictive factor for the development of Nelson's syndrome after total adrenalectomy for Cushing's disease. Journal of Clinical Endocrinology and Metabolism199479887–889doi:10.1210/jc.79.3.887.
BoscaroMLudlamWHAtkinsonBGlusmanJEPetersennSReinckeMSnyderPTabarinABillerBMKFindlingJMelmedSDarbyCHHuKWangYFredaPUGrossmanABFrohmanLABertheratJ. Treatment of pituitary-dependent Cushing's disease with the multireceptor ligand somatostatin analog pasireotide (SOM230): a multicenter, phase II trial. Journal of Clinical Endocrinology and Metabolism200994115–122doi:10.1210/jc.2008-1008.
van WijkPAvan NeckJWRijnberkACroughsRJMolJA. Proliferation of the murine corticotropic tumour cell line AtT20 is affected by hypophysiotrophic hormones, growth factors and glucocorticoids. Molecular and Cellular Endocrinology199511113–19doi:10.1016/0303-7207(95)03541-E.
ResetićJReinerZLüdeckeDRiznar-ResetićVSeksoM. The effects of cortisol, 11-epicortisol, and lysine vasopressin on DNA and RNA synthesis in isolated human adrenocorticotropic hormone-secreting pituitary tumor cells. Steroids19905598–100doi:10.1016/0039-128X(90)90002-S.
OldfieldEHSchulteHMChrousosGPGoldPWBenkerGPetersonRECutlerGBJrLoriauxDL. Corticotropin-releasing hormone (CRH) stimulation in Nelson's syndrome: response of adrenocorticotropin secretion to pulse injection and continuous infusion of CRH. Journal of Clinical Endocrinology and Metabolism1986621020–1026doi:10.1210/jcem-62-5-1020.
ShibasakiTNakaharaMShizumeKKiyosawaYSudaTDemuraHKuwayamaAKageyamaNBenoitRLingN. Pituitary adenomas that caused Cushing's disease or Nelson's syndrome are not responsive to ovine corticotropin-releasing factor in vitro. Journal of Clinical Endocrinology and Metabolism198356414–416doi:10.1210/jcem-56-2-414.
HuizengaNATMde LangePKoperJWClaytonRNFarrellWEvan der LelyAJBrinkmannAOde JongFHLambertsSW. Human adrenocorticotropin-secreting pituitary adenomas show frequent loss of heterozygosity at the glucocorticoid receptor gene locus. Journal of Clinical Endocrinology and Metabolism199883917–921doi:10.1210/jc.83.3.917.
BatistaDLZhangXGejmanRAnsellPJZhouYJohnsonSASwearingenBHedley-WhyteETStratakisCAKlibanskiA. The effects of SOM230 on cell proliferation and adrenocorticotropin secretion in human corticotroph pituitary adenomas. Journal of Clinical Endocrinology and Metabolism2006914482–4488doi:10.1210/jc.2006-1245.
HoflandLJvan der HoekJFeeldersRvan AkenMOvan KoetsveldPMWaaijersMSprij-MooijDBrunsCWeckbeckerGde HerderWWBeckersALambertsSW. The multi-ligand somatostatin analogue SOM230 inhibits ACTH secretion by cultured human corticotroph adenomas via somatostatin receptor type 5. European Journal of Endocrinology2005152645–654doi:10.1530/eje.1.01876.
A Pilot Study of Prolonged High Dose Rosiglitazone Therapy (12 mg/day) in Nelson's Syndrome [Internet]. 2004 Nov 1; [cited 2009 April 28]. Available from: http://www.endocrine-abstracts.org/ea/0008/ea0008OC24.htm.
McCormackAIMcDonaldKLGillAJClarkSJBurtMGCampbellKABraundWJLittleNSCookRJGrossmanABRobinsonBGClifton-BlighRJ. Low O6-methylguanine-DNA methyltransferase (MGMT) expression and response to temozolomide in aggressive pituitary tumours. Clinical Endocrinology200971226–233doi:10.1111/j.1365-2265.2008.03487.x.
WidhalmGWolfsbergerSPreusserMWoehrerAKotterMRCzechTMarosiCKnospE. O(6)-methylguanine DNA methyltransferase immunoexpression in nonfunctioning pituitary adenomas: are progressive tumors potential candidates for temozolomide treatment?Cancer20091151070–1080doi:10.1002/cncr.24053.
(T M Barber and E Adams contributed equally to this work)