This work is licensed under a Abstract
Graphical Abstract
Abstract
Objective
Characterize disease-specific mortality rates in patients with acromegaly on pegvisomant and identify pertinent risk factors, including on-therapy insulin-like growth factor I (IGF-I) levels.
Design
Retrospective cohort analysis of ACROSTUDY, a global surveillance study of patients with acromegaly receiving pegvisomant.
Methods
Cumulative incidence function was used to estimate disease-specific mortality, and regression analyses to characterized risk factors. Disease-specific standardized mortality ratios (SMR) were calculated; Poisson regression models characterized the association between disease-specific SMR, IGF-I, and other risk factors.
Results
2077 patients were followed (median: 4.1 years). Mortality (HR, 95% CI) secondary to cardiovascular/cerebrovascular causes increased with higher on-treatment IGF-I (1.97 (1.45–2.67), P < 0.0001) and older age at enrollment (1.10 (1.07–1.13), P < 0.0001). Mortality secondary to malignant (1.57 (1.17–2.09), P = 0.0024) or respiratory (1.64 (1.23–2.19), P = 0.0008) causes increased with higher on-treatment IGF-I. Younger attained age (0.93 (0.91–0.96), P < 0.0001), younger age (<35 vs > 50 years) at diagnosis (3.64 (1.33–9.93), P = 0.0117), higher on-treatment IGF-I (1.69 (1.12–2.55), P = 0.0127), and pituitary radiotherapy (2.25 (1.09–4.63), P = 0.0280) were associated with higher SMR (95% CI) for cardiovascular/cerebrovascular causes. Younger attained age (0.93 (0.89–0.96), P < 0.0001), higher IGF-I at enrollment (>2 × vs <1 × upper limit of normal: 4.89 (1.09–21.8), P = 0.0378), and malignancy at enrollment (7.05 (2.36–21.03), P = 0.0005) were associated with higher SMR (95% CI) for malignant causes. Younger age (35–50 vs >50 years) at diagnosis (4.50 (1.08–18.83), P = 0.0394) and sleep apnea (4.98 (1.34–18.53), P = 0.0168) were associated with higher SMR ratios for respiratory causes.
Conclusions
Younger age, higher on-therapy IGF-I, and radiotherapy were associated with higher SMR for cardiovascular/cerebrovascular causes, highlighting the importance of achieving IGF-I normalization.
Keywords: acromegaly, mortality, pegvisomant, radiotherapy
Introduction
Acromegaly has been associated with increased all-cause mortality, which has been attributed to cardiovascular, cerebrovascular, respiratory, and possibly malignant disorders (1, 2, 3, 4). The primary treatment of acromegaly generally involves pituitary surgery (5). In addition, medical therapy and/or radiotherapy are recommended for patients whose disease is not controlled postoperatively (5). Excess all-cause mortality is mitigated in patients who achieve endocrine control of acromegaly (1, 6, 7). However, it is not clear whether excess disease-specific mortality is mitigated in patients with acromegaly whose disease is controlled (based on serum IGF-I normalization).
Pegvisomant is a genetically engineered growth hormone receptor antagonist that normalizes serum insulin-like growth factor I (IGF-I) levels in 75–89% of patients with acromegaly (8, 9, 10, 11). It has been previously reported that patients with acromegaly treated with pegvisomant, either alone or together with other agents (somatostatin receptor ligands or cabergoline), and who achieved IGF-I normalization had all-cause mortality rates that were indistinguishable from those in the general population (7).
We hypothesized that disease-specific mortality is associated with higher on-treatment serum IGF-I levels and is mitigated in patients with acromegaly whose disease is controlled, based on on-treatment serum IGF-I levels that did not exceed the upper limit of the corresponding normal range. To test this hypothesis, we analyzed data from ACROSTUDY, a large database of patients with acromegaly treated with pegvisomant, either alone or in combination with other therapies (10, 12, 13, 14).
Methods
ACROSTUDY is a global, multicenter, observational pharmacoepidemiologic surveillance study, which was established in 2004 and concluded in December 2017 (10, 12, 13, 14). The database includes information on 2,221 patients with acromegaly who were treated with pegvisomant, generally administered after surgery, either alone or in combination with other agents (such as somatostatin receptor ligands or cabergoline).
Data extracted from the ACROSTUDY database include age at diagnosis of acromegaly, age at entry into ACROSTUDY, attained age (as previously defined) (7), sex, serum IGF-I levels (measured locally at each participating study site and expressed as ‘fold levels’ above the upper limit of the respective normal range (ULN)), presence of pituitary hormone deficiencies, comorbidities at study entry (diabetes mellitus, hypertension, dyslipidemia, cardiovascular/cerebrovascular disease, sleep apnea, malignancy), pegvisomant dose, additional therapies for acromegaly (surgery, medical therapies besides pegvisomant, radiation therapy), date of birth, dates of first visit and last visit in ACROSTUDY, and date of death (if applicable). Briefly, attained age is the age each patient reached from their first visit date to their last date in ACROSTUDY, taking into consideration their date of birth and the number of patient-years accrued in ACROSTUDY.
Data on causes of death were also extracted from the ACROSTUDY database and were adjudicated by a subgroup of investigators (NAT, LC, KKM) before data analysis. Data on 36 patients whose cause of death was initially unclear were reviewed, including all available narrative information from case report forms, to adjudicate causes of death. Any disagreements during adjudication were resolved, and consensus was achieved at an online meeting. Causes of death of interest were categorized as cardiovascular/cerebrovascular, respiratory (including sleep apnea), and malignant disorders. Sensitivity analyses were conducted in which eight patients whose cause of death remained unknown after adjudication were assumed to have a cardiovascular or cerebrovascular cause of death.
All patients with available IGF-I data at study entry and during follow-up were included in the analyses. The continuous IGF-I/ULN data were classified time-dependently during ACROSTUDY in six IGF-I/ULN classes (<0.7, 0.7 to <1.0, 1.0 to <1.25, 1.25–2.0, 2.0–2.57, and ≥2.57), where class values were represented by their respective mean values (0.52, 0.84, 1.11, 1.53, 2.24, and 3.59). In tables, these were reduced to three classes (given as <1 IGF-I/ULN, 1 to <2 IGF-I/ULN, and ≥2 IGF-I/ULN). The time-dependent IGF-I/ULN classification took into account blood sample dates, wherein each patient could contribute patient-years and be classified in different IGF-I/ULN classes during follow-up, depending on measurement values at different sample dates during the follow-up period. As an example, a patient who was alive after 3 years of follow-up and had IGF-I/ULN values measured at baseline and years 1, 2, 3 (IGF-I/ULN: 2.26, 1.09, 0.70, 0.89, respectively) would have a time-dependent covariate (classified in three classes) that equals: (≥2, 1 to <2, <1, <1) during the respective time period. Data were analyzed with IGF-I/ULN as a continuous variable as well as log-transformed.
To account for competing risks of disease-specific mortality attributed to cardiovascular/cerebrovascular, respiratory, and malignant disorders, we calculated the cumulative incidence function (CIF) separately for each event of interest. The CIF represents the probability that a specific event of interest occurs by time (t) without the subject experiencing a competing risk event (i.e., death due to another cause) by this time. Gray’s test was used to compare ≥2 CIFs. This test is analogous to the log-rank test comparing Kaplan–Meier curves and uses a modified chi-squared test statistic. When possible (based on dataset size), cause-specific Cox proportional hazards analysis was used to estimate cause-specific hazard rates for each event type, adjusting for age and sex.
Disease-specific mortality for the study population was also compared with that in the general (reference) population using the standard mortality ratio (SMR), defined as the ratio between the observed and expected number of cause-specific deaths encountered during the observation period (measured in patient-years from ACROSTUDY entry until the last recorded visit or date of death). The expected number of cause-specific deaths was computed based on the 2016 update of estimates for cause-specific mortality (World Health Organization Global Burden of Disease (WHO GBD)) and the number of patient-years in the ACROSTUDY cohort (15). Stratification was performed by attained age, sex, and the patients’ respective country, following the classification used in the WHO GBD project. When possible (based on dataset size), multiplicative multiple Poisson regression analyses were conducted to characterize the association between on-therapy serum IGF-I levels and SMR, while adjusting for potential confounding factors.
All statistical analyses were performed using SAS (version 9.4, SAS Institute, USA) and the SAS procedures PROC LIFETEST, PROC PHREG, and PROC GENMOD. Data are presented as median (10th percentile, 90th percentile) or percentages (as appropriate). Two-tailed P values below 0.05 were considered statistically significant.
Results
Demographic and clinical characteristics of the study population are shown in Table 1. After excluding 144 patients with missing IGF-I data (at study entry or follow-up), 2,077 patients with acromegaly were followed for a median (range) of 4.1 years (<1, 13.1), contributing 8,957 patient-years in ACROSTUDY. Pegvisomant therapy was initiated at a median dose of 10 mg daily and was advanced according to local medical practice. Of note, pegvisomant was started before entry into ACROSTUDY in 95% of patients (median interval: 0.9 years before enrollment). Nearly half of the patients had elevated (above normal) IGF-I levels at entry into ACROSTUDY. After excluding two deaths due to missing follow-up IGF-I data, there were 85 deaths in the analysis (37 women and 48 men), accounting for 4.1% of the study population. Causes of death included cardiovascular/cerebrovascular diseases (n = 30), malignancy (n = 14), respiratory disorders (n = 9), unknown/sudden death (n = 20), and other (n = 12). Etiologies for ‘other’ causes of death included septic shock, meningitis, acute abdomen, gastrointestinal bleeding, hepatic encephalopathy, and accident (resulting in shock).
Table 1.
Demographic and clinical characteristics of study participants.
| Variable | n = 2,077 |
|---|---|
| Age at diagnosis (acromegaly), n (%) | |
| <35 years | 683 (32.9) |
| 35–50 years | 801 (38.6) |
| >50 years | 587 (28.3) |
| Missing value | 6 (0.3) |
| Age at ACROSTUDY entry, mean (range), years | 51.6 (32.0–69.6) |
| Sex, females/males, n (%) | 1,009/1,068 (48.6/51.4) |
| IGF-I at study entry (fold elevation above ULN), n (%) | |
| ≤1 × ULN | 1,052 (50.7) |
| >1–2 × ULN | 767 (36.9) |
| >2 × ULN | 258 (12.4) |
| Comorbidities at ACROSTUDY entry, n (%) | |
| Hypopituitarism | 526 (25.3) |
| Hypertension | 820 (39.5) |
| Diabetes mellitus | 600 (28.9) |
| Cardiovascular disease | 332 (15.9) |
| Cerebrovascular disease | 36 (1.7) |
| Malignant disease | 69 (3.3) |
| Therapy | |
| Pituitary surgery, n (%) | 2,005 (96.5) |
| Pegvisomant dose at study entry, mean (range), mg/daily | 10 (5–20) |
| Medical therapy before study entry, n (%) | 1,723 (82.9) |
| Combination medical therapy during study, n (%) | 894 (43.0) |
| Radiotherapy, n (%) | 648 (31.2) |
Abbreviations: IGF-I, insulin-like growth factor I; ULN, upper limit of the normal (reference) range.
The cumulative incidence of disease-specific mortality secondary to cardiovascular/cerebrovascular (Fig. 1A), malignant (Fig. 1B), and respiratory (Fig. 1C) disorders increased with older age at ACROSTUDY entry. The mortality risk secondary to cardiovascular/cerebrovascular causes increased with higher time-dependent serum IGF-I levels (relative risk (RR) per fold-level: 1.74 (95% CI: 1.28–2.36), P = 0.0004). Similarly, the mortality risk (95% CI) secondary to malignant (1.57 (1.17–2.09), P = 0.0024) or respiratory (1.64 (1.23–2.19), P = 0.0008) causes increased with higher time-dependent serum IGF-I levels.
Figure 1.
Cumulative incidence of cause-specific mortality for cardiovascular/cerebrovascular causes (panel A), malignant causes (panel B), and respiratory causes (panel C) stratified by patient age at time of entry into ACROSTUDY. Disease-specific mortality rates increased with higher age at study entry.
In proportional hazards regression, the mortality risk (95% CI) secondary to cardiovascular/cerebrovascular causes increased with higher time-dependent serum IGF-I levels (1.97 (1.45–2.67)) per IGF-I mean value category (or 97% increase per fold-level of IGF-I/upper limit of normal (ULN)), P < 0.0001) and older age (in years) at ACROSTUDY entry (1.10 (95% CI: 1.07–1.13), P < 0.0001), whereas the risk was lower in women vs men (0.55 (95% CI: 0.31–0.97), P = 0.0395).
Disease-specific mortality rates in ACROSTUDY and the general population in relation to attained age are shown in Fig. 2. SMR ratios secondary to cardiovascular/cerebrovascular causes increased with higher time-dependent serum IGF-I levels and decreased with older attained age (Table 2). Time-dependent IGF-I levels above the normal range were associated with elevated SMR ratios for cardiovascular/cerebrovascular causes.
Figure 2.
Cardiovascular/cerebrovascular mortality rates (panel A), mortality rates secondary to malignant causes (panel B), and respiratory causes (panel C) are shown in relation to attained age in patients with acromegaly participating in ACROSTUDY (blue line) and individuals from the general population in ACROSTUDY countries (red line).
Table 2.
Standardized mortality ratios for cardiovascular/cerebrovascular mortality, stratified by serum IGF-I, attained age, and sex.
| IGF-I (time-dependent) | <1 × ULN | 1–2 × ULN | >2 × ULN | Trend: SMR per IGF-I/ULN fold-level (CI)* | P value for trend | |
|---|---|---|---|---|---|---|
| Crude SMR (95% CI) | Estimate | 0.72 | 1.25 | 2.74 | 1.69 | 0.0127 |
| Lower | 0.4 | 0.62 | 0.74 | 1.12 | ||
| Upper | 1.19 | 2.24 | 7.01 | 2.55 | ||
| Crude SMR (95% CI) | Estimate | 1 (ref) | 1.58 | 2.84 | 1.56 | 0.037 |
| Relative to <1 × ULN | Lower | 0.71 | 0.82 | 1.03 | ||
| Upper | 3.51 | 9.81 | 2.38 | |||
| Poisson regression: attained age- and sex- adjusted SMR | 40 years† | 5.59 | 8.95 | 17.23 | – | – |
| 2.34 | 3.69 | 5.49 | – | – | ||
| 13.32 | 21.7 | 54.06 | – | – | ||
| 50 years† | 2.77 | 4.44 | 8.54 | – | – | |
| 1.43 | 2.2 | 3.06 | – | – | ||
| 5.37 | 8.96 | 23.89 | – | – | ||
| 55 years† | 1.95 | 3.12 | 6.02 | – | – | |
| 1.09 | 1.65 | 2.22 | – | – | ||
| 3.49 | 5.91 | 16.28 | – | – | ||
| 65 years† | 0.97 | 1.55 | 2.98 | – | – | |
| 0.58 | 0.86 | 1.11 | – | – | ||
| 1.6 | 2.8 | 8 | – | – | ||
| 70 years† | 0.68 | 1.09 | 2.1 | – | – | |
| 0.4 | 0.59 | 0.76 | – | – | ||
| 1.15 | 2.02 | 5.78 | – | – | ||
| Relative to <1 × ULN† | 1 (ref) | 1.6 | 3.09 | – | – | |
| 0.73 | 1.02 | – | – | |||
| 3.49 | 9.36 | – | – |
Trend estimates are based on six IGF-I/ULN classes (<0.7, 0.7 to <1.0, 1.0 to <1.25, 1.25–2.0, 2.0–2.57, and ≥2.57), where class values are represented by the respective mean values (0.52, 0.84, 1.11, 1.53, 2.24, and 3.59).
Data are shown as SMR estimates, with lower and upper 95% CI.
Abbreviations: IGF-I, insulin-like growth factor I; ref, referent; SMR, standardized mortality ratio; ULN, upper limit of normal range.
Univariable analyses of risk factors in relation to SMR ratios for cardiovascular/cerebrovascular causes, based on Poisson regression models, are shown in Table 3. Younger attained age (SMR ratio: 0.93 (95% CI: 0.91–0.96), P < 0.0001), younger age at diagnosis of acromegaly (RR: <35 vs >50 years, 3.64 (95% CI: 1.33–9.93), P = 0.0117), higher on-treatment (time-dependent) IGF-I levels (1.69 (95% CI: 1.12–2.55), P = 0.0127), and previous pituitary radiotherapy (2.25 (95% CI: 1.09–4.63), P = 0.0280) were associated with higher SMR ratios. On the other hand, sex, history of hypertension, diabetes mellitus, dyslipidemia, smoking, hypopituitarism, cardiovascular or cerebrovascular disease at study entry, sleep apnea, or combination medical therapy were not associated with SMR ratios for cardiovascular/cerebrovascular causes.
Table 3.
Cause-specific SMR ratios for cardiovascular/cerebrovascular disease in patients with acromegaly based on a Poisson regression model (univariable approach).
| Parameter (comorbidities were recorded at ACROSTUDY entry) | SMR ratio (95% CI) | P value |
|---|---|---|
| Age at diagnosis (<35 vs >50 years) | 3.64 (1.33–9.93) | 0.0117 |
| Attained age (RR per yr) | 0.93 (0.91–0.96) | <0.0001 |
| Sex (females vs males) | 0.84 (0.41–1.71) | 0.6236 |
| Time-dependent serum IGF-I (RR per fold IGF-I/ULN >1) | 1.69 (1.12–2.55) | 0.0127 |
| Pituitary radiotherapy (Y vs N) | 2.25 (1.09–4.63) | 0.0280 |
| Hypopituitarism (Y vs N) | 1.40 (0.64–3.07) | 0.3937 |
| Hypertension (N vs Y) | 1.10 (0.51–2.35) | 0.8060 |
| Diabetes mellitus (N vs Y) | 1.38 (0.66–2.91) | 0.3918 |
| Cardiovascular disease (Y vs N) | 1.75 (0.85–3.59) | 0.1306 |
| Cerebrovascular disease (Y vs N) | 2.17 (0.66–7.16) | 0.2021 |
| Dyslipidemia (Y vs N) | 0.22 (0.03–1.63) | 0.1391 |
| Smoking (Y vs N) | 0.88 (0.27–2.90) | 0.8386 |
| Sleep apnea (Y vs N) | 1.71 (0.78–3.72) | 0.1801 |
| Combination medical therapy (N vs Y) | 0.64 (0.31–1.31) | 0.2256 |
Abbreviations: IGF-I, insulin-like growth factor I; RR, relative risk; SMR, standardized mortality ratio; ULN, upper limit of the normal (reference) range.
Multivariable analyses of risk factors in relation to SMR ratios for cardiovascular/cerebrovascular causes, based on several Poisson regression models, are shown in Table 4. Younger attained age, higher time-dependent serum IGF-I levels, and previous pituitary radiotherapy generally remained robust as factors associated with higher SMR ratios pertaining to cardiovascular/cerebrovascular causes.
Table 4.
Cause-specific SMR ratios for cardiovascular/cerebrovascular disease in patients with acromegaly based on several Poisson regression models (multivariable approach).
| Parameter | SMR ratio (95% CI) | P value |
|---|---|---|
| Model 1 | ||
| Attained age (RR per yr) | 0.94 (0.91–0.96) | <0.0001 |
| Sex (females vs males) | 1.29 (0.60–2.77) | 0.5073 |
| Time-dependent serum IGF-I (RR per fold IGF-I/ULN >1) | 1.25 (0.98–1.61) | 0.0735 |
| Pituitary radiotherapy (Y vs N) | 1.96 (0.94–4.05) | 0.0707 |
| Model 2 | ||
| Sex (females vs males) | 0.84 (0.41–1.73) | 0.6377 |
| Time-dependent serum IGF-I (RR per fold IGF-I/ULN >1) | 1.69 (1.13–2.54) | 0.0113 |
| Pituitary radiotherapy (Y vs N) | 2.24 (1.08–4.60) | 0.0292 |
| Model 3 | ||
| Sex (females vs males) | 0.84 (0.41–1.72) | 0.6377 |
| Time-dependent serum IGF-I (RR per fold IGF-I/ULN >1) | 1.69 (1.13–2.55) | 0.0113 |
| Model 4 | ||
| Time-dependent serum IGF-I (RR per fold IGF-I/ULN >1) | 1.31 (1.02–1.68) | 0.0339 |
| Pituitary radiotherapy (Y vs N) | 2.28 (1.11–4.70) | 0.0251 |
Abbreviations: IGF-I, insulin-like growth factor I; RR, relative risk; SMR, standardized mortality ratio; ULN, upper limit of the normal (reference) range.
Univariable analyses of SMR ratios (95% CI) pertaining to malignant causes showed that younger attained age (0.93 (0.89–0.96), P < 0.0001), higher IGF-I levels at ACROSTUDY entry (RR: >2 × ULN vs < 1 × ULN, 4.89 (1.09–21.8), P = 0.0378), and history of malignancy at ACROSTUDY entry (7.05 (2.36–21.03), P = 0.0005), were associated with higher SMR ratios; sex, time-dependent IGF-I levels, history of hypertension, diabetes mellitus, hypopituitarism, and pituitary radiotherapy were not associated with SMR ratios (data not shown).
Univariable analyses of SMR ratios for respiratory causes showed that younger age at diagnosis of acromegaly (RR: 35–50 vs >50 years, 4.50 (1.08–18.83), P = 0.0394) and history of sleep apnea at ACROSTUDY entry (4.98 (1.34–18.53), P = 0.0168) were associated with higher SMR ratios; there was a trend towards an association between higher IGF-I levels at ACROSTUDY entry (RR: >2 × ULN vs <1 × ULN, 3.91 (95% CI: 0.93–16.36), P = 0.0619) with higher SMR ratios, albeit of borderline statistical significance. In contrast, sex, time-dependent IGF-I levels, history of hypertension, diabetes mellitus, hypopituitarism, radiation therapy, and combination medical therapy were not associated with SMR ratios for respiratory causes of death (data not shown).
A sensitivity analysis was subsequently conducted, in which eight patients with unknown/sudden causes of death were assumed to have a cardiovascular/cerebrovascular etiology. On univariable analysis (95% CI) Poisson regression model, younger age at diagnosis of acromegaly (RR: 35–50 vs >50 years, 2.65 (1.00–6.99), P = 0.0494), younger attained age (0.94 (0.92–0.97), P < 0.0001), higher on-treatment (time-dependent) IGF-I levels (1.75 (1.22–2.50), P = 0.0023), history of radiation therapy (2.14 (1.12–4.07), P = 0.0209), or cerebrovascular disease (2.96 (1.16–7.59), P = 0.0236) at ACROSTUDY entry were associated with higher SMR ratios for cardiovascular/cerebrovascular causes. In a multivariable analysis (95% CI) Poisson regression model, younger attained age (0.94 (0.92–0.97), P < 0.0001), higher on-treatment (time-dependent) IGF-I levels (1.30 (1.05–1.62), P = 0.0185), history of radiation therapy (1.98 (1.04–3.80), P = 0.0390), or cerebrovascular disease (4.21 (1.61–11.00), P = 0.0034) at ACROSTUDY entry were associated with higher SMR ratios for cardiovascular/cerebrovascular causes of death.
Discussion
Cardiovascular/cerebrovascular, neoplastic, and respiratory diseases are considered leading causes of death in patients with acromegaly (1, 6, 16, 17, 18). In some recent studies, malignant diseases have surpassed vascular diseases as the major cause of mortality in acromegaly (1, 19). Our findings are broadly consistent with these observations and affirm the significance of both cardiovascular/cerebrovascular and neoplastic diseases as substantial contributors to mortality risk in acromegaly.
The main goals of the present study were to examine risk factors for disease-specific mortality in patients with acromegaly on pegvisomant therapy in relation to on-treatment (time-dependent) IGF-I levels, age, sex, comorbidities, and other therapies. Our analyses showed that higher on-treatment serum IGF-I levels are associated with higher mortality risk secondary to cardiovascular/cerebrovascular, neoplastic, and respiratory diseases. Importantly, higher serum IGF-I levels (above the normal range) are also associated with higher SMR ratios for cardiovascular/cerebrovascular disease. These findings are novel and affirm the importance of achieving IGF-I normalization on therapy to mitigate cardiovascular/cerebrovascular mortality. Our present findings extend previous observations linking on-treatment serum IGF-I levels and all-cause mortality in acromegaly (7). Furthermore, observations from other studies have affirmed the association between on-treatment serum IGF-I levels and all-cause mortality (20). A recent study reported that mean on-treatment serum growth hormone levels are also associated with all-cause mortality (21).
The present study also identified an excess disease-specific mortality risk among younger patients (over that in the general population); however, older individuals had higher absolute mortality rates (as anticipated). A similar pattern was previously reported in regard to all-cause mortality in patients with acromegaly (7). It is possible that younger individuals may have fewer competing health risks and may be more vulnerable to the deleterious consequences of acromegaly. It is also conceivable that younger patients may have more severe disease; these two possibilities are not mutually exclusive. Our data do not allow us to distinguish between these possible explanations (please see Supplementary Data (see section on Supplementary materials given at the end of the article)).
Radiation therapy to the sella was identified as a risk factor for excess cardiovascular/cerebrovascular mortality in the present study. This observation is consistent with previous findings reporting an association between pituitary radiotherapy and either all-cause mortality (in acromegaly) or increased risk of stroke (in patients with pituitary adenomas), presumably reflecting detrimental long-term effects of radiation exposure on brain vasculature (22, 23, 24). It is also conceivable that patients who underwent radiation therapy might have more severe disease and a higher prevalence of hypopituitarism, which might have potentially contributed to their higher cardiovascular risk.
Strengths of the present study include the large size of the study population and follow-up interval (median: 4.1 year; maximal: 13.1 years). However, 10% of patients had a short period of follow-up (minimum: <1 year). The number of deaths in each category is relatively limited. In addition, the present study design and analysis were retrospective, and pegvisomant therapy was titrated based on local clinical practice. It is therefore possible that the dose of this medication was not sufficiently uptitrated in some patients. On the other hand, the dataset reflects real-world practices in the management of patients with acromegaly. The cause of death was sudden and/or unknown for some patients in the present study; however, sensitivity analyses demonstrated the robustness of our observations. It is conceivable that medical therapies for hypertension, dyslipidemia, and diabetes mellitus might influence disease-specific mortality risk in this population; however, this information was not available in ACROSTUDY. The present study was not specifically designed to examine the presence of any association between subnormal serum IGF-I levels and disease-specific mortality, an area where more research is needed.
In conclusion, higher on-treatment serum IGF-I levels are associated with higher mortality secondary to cardiovascular/cerebrovascular, neoplastic, and respiratory diseases, and higher serum IGF-I levels (above the normal range), younger age, and pituitary radiotherapy are associated with higher excess SMR for cardiovascular/cerebrovascular disease. Younger patients have a higher excess mortality risk secondary to cardiovascular/cerebrovascular and malignant diseases. In the aggregate, the findings of the present study indicate a relevant role of serum IGF-I during therapy in acromegaly and suggest that IGF-I normalization is important for mitigating excess disease-specific mortality risk.
Supplementary materials
Declaration of interest
NAT and CJ have no declarations of interest. GV has received investigator-initiated research grants from Takeda, and consulting and/or lecture honoraria from Lundbeck, Recordati, Ascendis, Novo Nordisk, Pharmanovia, Pfizer, HRA Pharma, and Merck. BMKB has served as the PI of research grants to Massachusetts General Hospital from Crinetics, and as an occasional consultant to Amolyt, Camurus, Chiesi, Crinetics, Pfizer, and Recordati. KKM has received study medication and investigator-initiated research grants from Amgen, and has equity in Bristol-Myers Squibb, General Electric, Boston Scientific, and Becton Dickinson. MOC is employed by Pfizer Inc. DLT, MPW, and LC are employed by and own stocks in Pfizer Inc.
Funding
ACROSTUDY was sponsored by Pfizer. Statistical support and analysis for this manuscript were funded by Pfizer.
Author contribution statement
NAT helped in conceptualization, data curation, formal analysis, investigation, methodology, project administration, supervision, validation, and visualization. MOC was responsible for data curation, formal analysis. GV and CJ helped in investigation. DLT and MPW contributed to methodology. BMKB helped in conceptualization, formal analysis, investigation, and methodology. LC was responsible for project administration and validation. KKM helped in conceptualization, data curation, investigation, methodology, supervision, validation, and visualization. All authors contributed to the drafting of this article, reviewed the article, and approved the final version for submission.
Ethics statement
Patients were enrolled in ACROSTUDY after providing informed consent at their study center and were treated by their physicians according to the local standard of care. ACROSTUDY was approved by institutional review boards (IRBs) at each participating medical center (IRBs listed in Supplementary Data). The study was conducted according to the principles of the Declaration of Helsinki (25).
Acknowledgments
The authors thank the patients and their families/caregivers, investigators, sub-investigators, research nurses, study coordinators, and operations staff who contributed to this study. Medical writing and editorial support were provided by Chu Kong Liew, PhD, CMPP, of Engage Scientific Solutions, and were funded by Pfizer.
References
- 1.Bolfi F, Neves AF, Boguszewski CL, et al. Mortality in acromegaly decreased in the last decade: a systematic review and meta-analysis. Eur J Endocrinol 2018. 179 59–71. ( 10.1530/EJE-18-0255) [DOI] [PubMed] [Google Scholar]
- 2.Colao A, Grasso LFS, Giustina A, et al. Acromegaly. Nat Rev Dis Primers 2019. 5 20. ( 10.1038/s41572-019-0071-6) [DOI] [PubMed] [Google Scholar]
- 3.Orme SM, McNally RJ, Cartwright RA, et al. Mortality and cancer incidence in acromegaly: a retrospective cohort study. United Kingdom acromegaly study group. J Clin Endocrinol Metab 1998. 83 2730–2734. ( 10.1210/jcem.83.8.5007) [DOI] [PubMed] [Google Scholar]
- 4.Dal J, Leisner MZ, Hermansen K, et al. Cancer incidence in patients with acromegaly: a cohort study and meta-analysis of the literature. J Clin Endocrinol Metab 2018. 103 2182–2188. ( 10.1210/jc.2017-02457) [DOI] [PubMed] [Google Scholar]
- 5.Katznelson L, Laws ER Jr, Melmed S, et al. Acromegaly: an endocrine society clinical practice guideline. J Clin Endocrinol Metab 2014. 99 3933–3951. ( 10.1210/jc.2014-2700) [DOI] [PubMed] [Google Scholar]
- 6.Esposito D, Ragnarsson O, Granfeldt D, et al. Decreasing mortality and changes in treatment patterns in patients with acromegaly from a nationwide study. Eur J Endocrinol 2018. 178 459–469. ( 10.1530/EJE-18-0015) [DOI] [PubMed] [Google Scholar]
- 7.Tritos NA, Mattsson AF, Vila G, et al. All-cause mortality in patients with acromegaly treated with pegvisomant: an ACROSTUDY analysis. Eur J Endocrinol 2020. 182 285–292. ( 10.1530/EJE-19-0794) [DOI] [PubMed] [Google Scholar]
- 8.Kopchick JJ. Discovery and mechanism of action of pegvisomant. Eur J Endocrinol 2003. 148 (Supplement 2) S21–S25. ( 10.1530/eje.0.148s021) [DOI] [PubMed] [Google Scholar]
- 9.Maamra M, Kopchick JJ, Strasburger CJ, et al. Pegvisomant, a growth hormone-specific antagonist, undergoes cellular internalization. J Clin Endocrinol Metab 2004. 89 4532–4537. ( 10.1210/jc.2003-031781) [DOI] [PubMed] [Google Scholar]
- 10.Fleseriu M, Fuhrer-Sakel D, van der Lely AJ, et al. More than a decade of real-world experience of pegvisomant for acromegaly: ACROSTUDY. Eur J Endocrinol 2021. 185 525–538. ( 10.1530/EJE-21-0239) [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Trainer PJ, Drake WM, Katznelson L, et al. Treatment of acromegaly with the growth hormone-receptor antagonist pegvisomant. N Engl J Med 2000. 342 1171–1177. ( 10.1056/NEJM200004203421604) [DOI] [PubMed] [Google Scholar]
- 12.Buchfelder M, Schlaffer S, Droste M, et al. The German ACROSTUDY: past and present. Eur J Endocrinol 2009. 161 (Supplement 1) S3–S10. ( 10.1530/EJE-09-0350) [DOI] [PubMed] [Google Scholar]
- 13.Buchfelder M, van der Lely AJ, Biller BMK, et al. Long-term treatment with pegvisomant: observations from 2090 acromegaly patients in ACROSTUDY. Eur J Endocrinol 2018. 179 419–427. ( 10.1530/EJE-18-0616) [DOI] [PubMed] [Google Scholar]
- 14.Trainer PJ. ACROSTUDY: an overview. Horm Res 2007. 68 (Supplement 5) 68–69. ( 10.1159/000110480) [DOI] [PubMed] [Google Scholar]
- 15.World Health Organization . Global Health Estimates 2016: Disease Burden by Cause, Sex, by Country and by Region, 2000-2016. Geneva, Switzerland: WHO. (https://www.who.int/data/gho/data/themes/mortality-and-global-health-estimates) [Google Scholar]
- 16.Arosio M, Reimondo G, Malchiodi E, et al. Predictors of morbidity and mortality in acromegaly: an Italian survey. Eur J Endocrinol 2012. 167 189–198. ( 10.1530/EJE-12-0084) [DOI] [PubMed] [Google Scholar]
- 17.Mercado M, Gonzalez B, Vargas G, et al. Successful mortality reduction and control of comorbidities in patients with acromegaly followed at a highly specialized multidisciplinary clinic. J Clin Endocrinol Metab 2014. 99 4438–4446. ( 10.1210/jc.2014-2670) [DOI] [PubMed] [Google Scholar]
- 18.Sherin RPV, Vietor NO, Usman A, et al. Cardiovascular disorders associated with acromegaly: an update. Endocr Pract 2024. 30 1212–1219. ( 10.1016/j.eprac.2024.09.014) [DOI] [PubMed] [Google Scholar]
- 19.Arosio M, Sciannameo V, Contarino A, et al. Disease control of acromegaly does not prevent excess mortality in the long term: results of a nationwide survey in Italy. J Endocrinol Investig 2024. 47 1457–1465. ( 10.1007/s40618-023-02257-3) [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Arnardottir S, Jaras J, Burman P, et al. Long-term outcomes of patients with acromegaly: a report from the Swedish pituitary register. Eur J Endocrinol 2022. 186 329–339. ( 10.1530/EJE-21-0729) [DOI] [PubMed] [Google Scholar]
- 21.Orme S, McNally R, James PW, et al. Increased mortality in acromegaly is due to vascular and respiratory disease and is normalised by control of GH levels-A retrospective analysis from the UK acromegaly register 1970–2016. Clin Endocrinol 2024. 100 558–564. ( 10.1111/cen.15060) [DOI] [PubMed] [Google Scholar]
- 22.Sherlock M, Reulen RC, Alonso AA, et al. ACTH deficiency, higher doses of hydrocortisone replacement, and radiotherapy are independent predictors of mortality in patients with acromegaly. J Clin Endocrinol Metab 2009. 94 4216–4223. ( 10.1210/jc.2009-1097) [DOI] [PubMed] [Google Scholar]
- 23.Brada M, Burchell L, Ashley S, et al. The incidence of cerebrovascular accidents in patients with pituitary adenoma. Int J Radiat Oncol Biol Phys 1999. 45 693–698. ( 10.1016/s0360-3016(99)00159-5) [DOI] [PubMed] [Google Scholar]
- 24.Osorio RC, Oh JY, Choudhary N, et al. Pituitary adenomas and cerebrovascular disease: a review on pathophysiology, prevalence, and treatment. Front Endocrinol 2022. 13 1064216. ( 10.3389/fendo.2022.1064216) [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Riis P. Thirty years of bioethics: the Helsinki Declaration 1964–2003. New Rev Bioeth 2003. 1 15–25. ( 10.1080/1740028032000131396) [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.