Screening for comorbid conditions in patients enrolled in the SODA registry: a 2-year observational analysis

In keeping with guidelines and consensus statements on the diagnosis and treatment of acromegaly-associated comorbidities [2, 6, 7], this report examined the frequency of monitoring for comorbidities and described the findings in a cohort of patients with acromegaly from the SODA registry. The frequency of optional testing and comorbidity screening recorded at enrollment, 1, or 2 years was low overall, potentially reflecting the ongoing challenges and geographic differences in managing this population of patients. Approximately 50% or fewer of patients had results from sleep studies, echocardiograms, gallbladder sonographies, or colonoscopies during study enrollment, and overall follow-up was infrequently reported during the 2-year observation period despite abnormal test results in some patients at enrollment. The reasons for the low frequency of testing for comorbidities are not exactly clear, but may involve differences in assessment practices for patients with a history of acromegaly vs patients at the time of initial diagnosis. Also, practitioners may not be performing certain follow-up assessments if a patient with a history of acromegaly does not also have a history of the comorbid condition. Furthermore, if a patient achieves and sustains target biochemical treatment goals, they may not be perceived of as being at an increased risk for some of the associated comorbid conditions, even though evidence has shown that some potentially severe comorbidities, including arthropathy and sleep apnea, may persist even after long-term biochemical control of acromegaly [19, 20]. Approximately two-thirds of the patients enrolled in the SODA study were recruited from sites in academic medical centers and one-third were from community private practices [17]. However, whether the number of recruited patients per study site was proportional to the volume of acromegalic patients cared for at a particular location is not known. Therefore, no conclusions can be made regarding the frequency of testing for comorbidities and the volume of acromegalic patients seen in a particular center. Limitations attributed to the observational study design may have also contributed to the low frequencies of testing reported. Specifically, patients may have had a particular study performed but the result was either unknown or not recorded in the SODA database by the investigator. In addition, the timeframe of the SODA study should be considered, as the study used a data cut-off of 2014, which was prior to the publication of the most recent acromegaly management guidelines. It should be noted, however, that the incidence of comorbidities was similar and in line with that seen in other observational studies of patients with acromegaly [21, 22].

Cardiovascular comorbidities, including hypertension (predominantly diastolic), left ventricular hypertrophy (LVH), diastolic and systolic dysfunction, arrhythmias, myocardial infarction, enlarged great vessel diameters, and valve diseases are highly prevalent in acromegaly and are considered a major cause of mortality [12, 14‐16, 23‐27]. Among them, hypertension [16] and cardiomyopathy [24] are considered the main cardiac risk factors that directly impact mortality. Three of 4 deaths in SODA resulted from cardiovascular complications, and all 4 patients had hypertension. Hypertension in SODA was documented in almost half (45.2%) of patients at enrollment, consistent with 47.5% in the US acromegaly registry reported by the Pituitary Center at Cedars-Sinai Medical Center (CSMC-PC) [28] but higher than rates documented in other studies, including registry studies (rates ranging from 22% to 41.3%) [9, 14, 21‐24, 29, 30]

The prevalence of different features of cardiomyopathy is 3.3–14.2 times higher in the acromegalic than non-acromegalic population, with the disease duration as its major determinant [24, 31]. Guidelines recommend routine echocardiogram and electrocardiogram at diagnosis and annually during follow-up, especially in patients who have evidence of ventricular hypertrophy by electrocardiography or who are symptomatic, particularly if older [2, 6]. In SODA, only 3 (1.2%) patients had documented cardiomyopathy at enrollment based on patient medical histories. Taking into account 4.3 years as the median time since acromegaly diagnosis, we can assume that cardiomyopathy was under-assessed in this cohort of patients. Indeed, just under half of the patients had an echocardiogram at enrollment, and very few patients had one at follow-up visits (<10%), of whom nearly half had abnormal findings. These data suggest that, in real-world practice, guideline recommendations are not always followed. However, according to the Endocrine Society 2014 Acromegaly Guidelines, the role of pretreatment echocardiogram has not been defined; rather, thorough cardiac evaluation may be indicated by suggestive clinical findings, particularly in perioperative patients [6]. Also, according to a cardiac MRI study, patients with active acromegaly might have a lower (5%) prevalence of LVH than previously reported using echocardiogram [32].

Sleep apnea, primarily obstructive, is a frequent comorbidity in acromegaly due to soft tissue thickening and edema of the tongue, pharynx, and upper airways, with a reported prevalence of 25%–60% [5] and ~69% [6]. Although biochemical control of acromegaly usually improves sleep apnea, it has been shown to persist in ~40% of biochemically controlled patients [6, 33]. Guidelines recommend an annual Epworth Sleepiness Scale or sleep study during follow-up of acromegaly [2‐4, 6]. Less than one-third of patients had completed a sleep study at SODA enrollment, and very few (<3%) had entered a sleep study at follow-up visits. At enrollment, the majority of tested patients had sleep apnea (79.2%), consistent with other reports [33‐35], although the CSMC-PC and ACROSTUDY reported significantly lower (22.5% and 17%, respectively) prevalence of sleep apnea [22, 28]. However, these numbers are not strictly comparable, since symptomatic patients were more likely to get sleep studies, and thus the true prevalence may not be reflected. Also, the SODA database did not distinguish between obstructive or central sleep apnea. Nonetheless, given that sleep apnea is independently associated with hypertension and cardiovascular disease [33, 34] and proposed to account for up to 25% of the excess mortality in untreated acromegaly [33], there may be potential for improved outcomes in acromegaly patients if sleep apnea screening is increased.

The frequent prevalence of DM in SODA patients (25.3%), along with older-age, higher-BMI, and higher coexistence of hypertension and statin use in DM vs non-DM patients noted in the previous 2-year SODA analysis [18], were findings similar to those reported in several other studies [6, 9, 22] and registries [21, 28, 29, 36]. Hyperglycemia at diagnosis and ongoing SRL treatment are independent predictive factors of persistent or new glucose abnormalities during follow-up [37]. In addition, patients with comorbid DM are reported to have a lower survival rate than patients without DM [38]. Therefore, all patients with acromegaly should be tested for glucose intolerance and DM, appropriately treated, and followed up as indicated for these conditions. Guidelines recommend an oral glucose tolerance test at diagnosis, fasting blood glucose every 6 months (particularly in uncontrolled disease and during SRL therapy), and HbA1c every 6 months if diabetes is present [2‐4, 6]. In SODA, less than half of patients had HbA1c results reported after 1 and 2 years of LAN therapy.

Lowering GH levels improves glycemic control and increases insulin sensitivity in acromegaly; however, SRL therapy may exert variable effects on glucose metabolism, with worsening due to inhibition of insulin secretion and improvement due to improved insulin sensitivity with acromegaly control [39, 40]. Although the number of patients with reported HbA1c levels in this SODA analysis were small, similar proportions of patients were reported in each HbA1c level at M12 and M24 within each DM and non-DM subgroup. These findings support other reports of the relatively minor impact of SRLs on glucose homeostasis [39], in particular LAN [41, 42], which is less detrimental to diabetes than the next-generation SRL pasireotide [43].

Reported biochemical control rates were similar among patients at M12 and M24, independent of glucose homeostasis levels. No significant differences between DM and non-DM groups were observed in patients who achieved both ≤2.5 µg/L and <1.0 µg/L GH control after 1 and 2 years of treatment. Conversely, the data revealed higher rates of IGF-1 <ULN in non-DM vs DM patients after 2 years of treatment. The similar mean LAN 28-day dose equivalent and LAN EDI use, and higher use of the 120-mg LAN dose among DM vs non-DM patients, suggests that inadequate LAN dosing would not account for the higher IGF-1 levels in the DM group at M24. One likely explanation is the hyperinsulinism in the DM group, which enhanced synthesis of IGF-1 through upregulation of hepatic GH receptors [17, 44, 45]. The disconnect between the similar GH levels in DM vs non-DM patients and the difference in IGF-1 may also be explained by the effect of obesity [45] and hyperinsulinism both increasing hepatic GH sensitivity [46] and increasing free fatty acids, which suppress GH release [47]. These findings are in agreement with another recent analysis of SODA data, which indicated that more obese vs non-obese patients achieved GH <1.0 µg/L after both 1 and 2 years of LAN treatment [18]. Notably, this trend of GH control was opposite to IGF-1 control. These data are in line with those of Matta et al [48], who report metabolic syndrome markers such as higher fasting blood glucose and systolic blood pressure, in treated patients with acromegaly with a high IGF-1 and GH <1 µg/L.

In some studies, colon polyps and/or cancer have been reported to occur more frequently in acromegaly than in the general population, although data from other studies have not supported this association [49, 50]. There is general agreement, however, that early colonoscopy screening and regular surveillance is justified [38, 49, 51‐54]. The Screening Guidelines for Colorectal Cancer and Polyps in Acromegaly [52] recommend regular colonoscopic screening, starting at 40 years of age, with frequency of repeat colonoscopy depending on findings at original screening and acromegaly activity. Almost half of SODA patients had undergone a colonoscopy by enrollment, many within 3 years, and 35.3% had polyps. These findings are higher than in other studies [9, 10, 55] and in the Belgian (27.2%) [21] and CSMC-PC (20.0%, polyps or colon cancer) [28] registries. Follow-up colonoscopies were recorded in fewer numbers of SODA patients.

Although the impact of acromegaly and its control on neoplasia risk and mortality are controversial [6, 8, 27], the presence of cancer and the last IGF-1 level are considered significant mortality predictors in acromegaly [9, 12, 38]. Malignancies of various systems were reported in 10.8% of SODA patients at enrollment, which is consistent with 10.5% in the Belgian registry [21]. Thyroid cancer is a commonly reported cancer in acromegaly[6, 9], followed by breast, lung, ovarian, and lymphoma [9]. Consistent with these reports, thyroid, skin, and breast were the most common cancer types in the SODA cohort, followed by brain, blood (lymphoma), and prostate.

Hypopituitarism in acromegaly may develop due to tumor compression or as a result of surgical or radiation treatment. Acromegaly guidelines recommend assessing for hypopituitarism and adequate replacement of adrenal, gonadal, and thyroid insufficiency; patients who receive radiotherapy need lifelong monitoring of pituitary function [2, 4, 6]. In SODA, 44.0% of enrolled patients presented with pituitary hormone deficiencies, which is higher than the rates reported in several European registries [21, 30, 56]. Similar to the German registry [56], pituitary insufficiency was more commonly reported in males compared to females. The prevalences of TSH and gonadotropin deficiencies were higher in SODA than in the German registry[56], CSMC-PC registry [28], and ACROSTUDY [22]. Interestingly, as in the German registry [56], gonadotropin deficiency in SODA was more commonly reported in males (49.1%) vs females (7.2%). The cause for this apparent gender difference is unknown but it is important to consider in registry studies that it may be due to ascertainment bias or potentially less rigorous documentation of gonadal status in women. This gender difference does not appear to be related to tumor size, surgery, or radiation exposure as there were no gender differences noted for these variables in SODA [18]. ACTH deficiency in SODA (11.6%) was similar to the German registry (11.8%) [56], but lower than in the CSMC-PC registry (14.9%) [28] and ACROSTUDY (15%) [22]. ACTH deficiency can lead to adrenal crisis during acute illness and may cause adverse metabolic effects due to chronic supra-physiological glucocorticoid replacement, which are associated with increased mortality in patients with acromegaly [2, 27, 56, 57]. None of the 4 patients who died in SODA had ACTH deficiency.

Gallbladder stones were present in 17.8% of SODA patients with a gallbladder sonography performed at enrollment, which falls between the rates reported in the ACROSTUDY (7%) [22] and Belgian registry (23.4%) [21]. Of note, gallbladder sonography was performed in only 18.7% of SODA patients. The Endocrine Society guideline does not suggest routine abdominal ultrasound to monitor for gallstone disease in a patient receiving an SRL [6]; however, ultrasound should be performed if a patient has signs and symptoms of gallstone disease. In SODA, the frequency of gallbladder ultrasound was relatively low throughout the observation period, which could be in part because the study protocol excluded patients with symptomatic, untreated biliary lithiasis. Moreover, a retrospective case cohort study of 31 consecutive newly diagnosed patients with acromegaly revealed a significantly increased prevalence of gallbladder polyps compared with a control group [58].

Given that SODA is an observational study, heterogeneity of data collection across centers is an inherent limitation, in addition to the fact that patients in SODA were enrolled at different stages of disease and treatment. Follow-up in observational studies is generally not as active or as standardized as in randomized trials; therefore, ascertainment of outcomes may be incomplete or inaccurate. The optional studies recorded in the database may not represent a completely accurate account of all the studies actually performed due to the possibility of under-reporting. Although registries are typically more generalizable to ‘real world’ practice because of their observational design, entry into a registry may not be as strictly monitored compared with randomized trials. This may weaken the generalizability of findings obtained from analysis of registry data. The subgroup comparisons were also limited by small sample sizes in some instances.

In conclusion, this analysis of acromegaly patients from the observational SODA registry found less frequent real-life monitoring of comorbid conditions than recommended by recent treatment guidelines. Additional prospective analysis will be needed to further identify and address potential barriers to managing acromegaly-associated comorbidities and the impact of screening on survival.