Cushing's Disease in Older Patients

Presentation and Outcome

Nidan Qiao; Brooke Swearingen; Nicholas A. Tritos


Clin Endocrinol. 2018;89(4):444-453. 

In This Article


The typical CD patient is described as a young woman with central obesity, fatigue, muscle wasting, hirsutism and other typical physical findings (such as wide, violaceous striae and spontaneous ecchymoses). Cushing's disease at the extremes of the age spectrum is much less common, and its presentation in older patients is not well described. Although there is a substantial body of literature discussing symptoms and outcomes in paediatric patients with CD,[18,19] no studies have specifically addressed these issues in the older population. In this study, we report on clinical symptoms, biochemical features and surgical outcomes in patients with CD greater than 60 years old.

While in the literature CD is about three times more prevalent in women than in men,[14,20] our data show that the female:male ratio decreases in older patients from about 7.2:1 in the younger patient population to about 2.8:1 in the older population. This trend towards equalization in gender ratio was also seen in a large US-based study,[21] where the female/male ratio was 7.9 in patients between 35 and 44 years of age, 4.0 in patients between 45 and 54 years of age and 3.5 in patients between 55 and 64 years of age. Our study also found that weight gain as a presenting complaint tended to be more common in younger patients, and the BMI was significantly higher in the younger group. Conversely, muscle wasting was more common in older patients. Previous case reports in older patients with CD also suggested that their symptoms primarily included proximal myopathy, fractures, and psychiatric complaints. Some cases in the older age group were found incidentally, perhaps because these degenerative manifestations can be difficult to distinguish from those associated with the normal ageing process.[6–8] In a review of pituitary adenomas in the elderly patients,[22] Giuseppe et al concluded that the clinical picture of CD in older patients may be milder than in younger patients, but this difference in apparent severity may in fact represent a different response to the effects of chronic hypercortisolemia. Based on differences in presentation, we might speculate that the prevalence of catabolic manifestations of CD in the two groups is different. Older CD patients seemed to have a more catabolic phenotype at presentation in our study, although no difference in several anabolic hormone levels (thyroxine, testosterone and IGF-1) was found between the two age groups. It is also possible that there is age-related variability in tissue sensitivity to glucocorticoids, possibly due to differences in glucocorticoid receptor sensitivity and intracellular cortisol signalling.[23,24] Finally, it is possible that older patients may have a longer duration of hypercortisolism prior to clinical recognition.

In the published literature, weight gain and central obesity were usually combined for analysis and the reported proportion of overweight/obesity is 60%–100%.[25,26] The prevalence of central obesity alone was reported as 14%–42.9%.[26,27] Overall, our findings are in agreement with published data. If we combine the prevalence of weight gain and central obesity in our study, it is 80.0% in younger patients and 71.1% in older patients.

Our data showed that both younger and older patients with CD demonstrate an increased risk of multiple fractures, compared to those patients with NFPA. It is difficult to separate the effects of hypercortisolemia from age-related bone fragility. Previous reports are ambiguous about the predictive value of age on fractures in patients with CD. Zhanna et al[28] found that the age of patients who had fractures was higher than patients who did not, although the difference was not significant. Trementino reported that the age of fractured versus nonfractured CS patients was not significantly different, but suggested that the fracture risk increased with longer exposure to hypercortisolemia.[29] In our study, 17.8% of the older patients had history of multiple fractures comparing to 8.9% of the younger patients (P = NS). Although the result was not significant, the trend was clear. Tiganescu et al[30] mentioned that glucocorticoid excess and age may have similar result on skin changes; we did not identify a reported difference in skin fragility with age.

The incidence of venous thromboembolism is reported to increase markedly with ageing.[31] We observed the same increased risk in this study. The causes of this greater risk in older patients are thought to stem from multiple etiologies, including age-related changes in inflammatory pathways or comorbid conditions.[32] The comparative effectiveness of using prophylactic anticoagulation in older patients with CD should be investigated in future studies.

A number of comorbidities were more often seen in older CD patients, including hypertension, diabetes mellitus, history of cardiovascular disease and history of deep vein thrombosis. The higher burden of comorbidities may be attributed to both increased cortisol levels and increased age. However, we saw increases in these comorbidities in comparison with both control groups, suggesting that the effects may be cumulative and likely attributed to hypercortisolism rather than ageing alone. Previous studies have reported that the mortality risk in CD is increased compared to the general population; diabetes, hypertension and uncontrolled hypercortisolism have been reported to predict risk of death.[33] There may be a disproportionately increased mortality risk in CD patients with older age. However, our data did not allow us to estimate the long-term mortality risk in our population.

The risk of operative complications was similar in the two groups, and no age-related difference in adverse events was observed. The prevalence of SIADH was low in our cohort compared to the literature, but it may well have been underestimated in this retrospective study. Surgical complications were also uncommon compared to the literature.[34] The 30 day postoperative mortality rate was low both in young (0 death) and older patients (1 deaths); there was one perioperative death 1 week after surgery in an older patient in a rehab hospital, possibly due to pulmonary embolus; this is similar to our previous report.[35]

Our data showed that the surgical remission rates were similar in both younger and older populations and that surgery remains safe and effective in older age group. Postoperative remission rates vary widely in the literature, depending on the adopted remission criteria, ranging from 65% to 93%.[15] In our study, no significant difference was observed in initial remission rate, remission rate after early second operation and overall remission between the two age groups. In a recent meta-analysis, remission rate was around 78% in CD patients.[15] Early reoperation remains a successful treatment strategy in older patients, with reasonably high remission rates. Nadia et al and Shirvani et al reported that an approximately 70% remission rate can be achieved in early reoperation;[36,37] our success rate after early reoperation was 9/16 (56%) overall.

We found a significant difference in recurrence risk between the two groups, with a relatively low recurrence rate in older patients. Older patients had a nonsignificantly shorter mean follow-up, but the 5-year recurrence-free survival was 77.7% in younger as opposed to 92.7% in older patients (P = 0.049). The significance of this observation is unclear. Recurrence in younger patients was comparable to a pooled analysis of previous studies (average recurrence rate was 14%, 95%CI: 10%–18%).[38] There were no consistent results regarding the predictive value of age on recurrence in previous studies. Jessica et al concluded that no demographic variables predicted risk of recurrence,[5] although Bansa et al reported that recurrence was more common in a younger population as opposed to the group that experienced sustained remission.[39] One possible explanation for the difference in recurrence rate was that the Ki-67 index tended to be lower in older patients, although the difference was not significant. There was also an increase in the proportion of macroadenomas in the younger population, and this has been associated in the literature with a higher ki-67 index and possibly a greater risk of recurrence.[40–42] In contrast, other studies have not supported the value of pathological indices (increased mitotic index and Ki-67) in predicting disease recurrence.[43]

All patients with adequate available follow-up had eventual control of their hypercortisolemia, most commonly with a combination of medical and radiation therapy. This suggest that, while surgery remains first-line treatment, adjunctive treatments are important in long-term disease control. Our patient population with persistent disease was too small to address the relative benefits of second-line treatments in the older population.

There are several limitations in this study, primarily due to its retrospective nature. First, ideally, symptoms would have been quantified using a questionnaire in a prospective manner. These data were not available in our retrospective study. However, all of the evaluations we reviewed were conducted by a small group of neuroendocrinologists in our unit who are skilled in the evaluation of the patients with Cushing's disease and were consistently abstracted from the electronic medical records. Many of the symptoms of Cushing's syndrome are difficult to quantify in a retrospective study. Nonetheless, even if our retrospective review of symptoms and signs is not specifically quantitative, we would argue that this would be more likely to introduce random as opposed to systematic error. This might dilute differences between age groups, but would be unlikely to magnify them. Second, we have insufficient data on disease duration, magnitude of weight gain, bone density data as well as prevalence of comorbidities after disease remission in our records to make a meaningful comparison. Future studies may focus on more quantitative assessments, including variations in body fat composition, bone densitometry, and muscle strength in older CD patients.[44] Third, we currently save samples from the IPSS for prolactin assay if the ACTH central:peripheral ratios are equivocal. This was not our practice during the period of chart review. Fourth, it is conceivable that the higher prevalence of hirsutism observed in younger patients might be attributed to a lower attention to this clinical problem in older patients. Nonetheless, our data suggest that CD in older patients may present a relatively distinct clinical entity.