Baseline characteristics (Table 1)

Compared with non-AI patients, AI patients were more likely to be male, older, have a longer duration of GHD, have more adult-onset GHD and a diagnosis of non-functioning or corticotrophic adenoma, have benign non-pituitary neoplasms and more defects in other pituitary axes, and have had more pituitary surgery. They were more likely to have a history of coronary heart disease, to be taking anticoagulants and to be former smokers. The median dose of rhGH received was the same in AI and non-AI.

Cardiovascular risk profile at baseline (Table 2) Table 2 Baseline values for cardiovascular risk profile for both groups

At baseline, AI patients had significantly higher levels of total cholesterol and low-density lipoprotein (LDL) cholesterol (in the group of patients using lipid-lowering drugs; p = 0.004 and p < 0.001, respectively). Triglyceride levels were also higher in the AI subgroup, but this difference lost significance after adjustment for age, sex and GHD onset.

Cardiovascular risk profile during adult GHRT (Figs. 2, 3 and 4) Fig. 2

The course of blood pressure during adult GHRT (including 95% confidence intervals). Significant changes between groups (p < 0.05) : × = significant only in fully adjusted model (fully adjusted for age, sex, GHD onset, diagnosis of Cushing’s disease, LH/FSH—TSH—or AVP deficiency, history of coronary artery disease at baseline and former smoking)

Fig. 3

The course of body composition during adult GHRT (including 95% confidence intervals). Significant changes between groups (p < 0.05) : о = significant only in unadjusted model; × = significant only in fully adjusted model (fully adjusted for age, sex, GHD onset, diagnosis of Cushing’s disease, LH/FSH—TSH—or AVP deficiency, history of coronary artery disease at baseline and former smoking); * significant in both models

Fig. 4

The course of lipid profile and glucose metabolism during adult GHRT (including 95% confidence intervals). Significant changes between groups (p < 0.05) : × = significant only in fully adjusted model (fully adjusted for age, sex, GHD onset, diagnosis of Cushing’s disease, LH/FSH—TSH—or AVP deficiency, history of coronary artery disease at baseline and former smoking); * significant in both models

During adult GHRT, AI patients were more likely to use antihypertensives (33% vs. 28%; χ2 = 6.37, p = 0.01), lipid-lowering drugs (28% vs. 21%; χ2 = 11.42, p < 0.001) and anticoagulants (18% vs. 10%; χ2 = 25.46, p < 0.001) than non-AI patients. No significant differences were found in the use of antiarrhythmics or glucose lowering medication (oral diabetics and insulin).

Unadjusted modelsBlood pressure (Fig. 2)

No significant differences between AI and non AI patients were found in systolic or diastolic blood pressure.

Body composition (Fig. 3)

Body mass index (BMI): After year 2, BMI increased significantly in both groups and this increase was less in AI patients at years 10 and 15 (p = 0.01 and p < 0.001).

Waist circumference: Men with concomitant AI had a significantly smaller increase in waist circumference at year 10 (p = 0.003) than men without, while no differences were found between the groups in women.

Hip circumference: The increase in hip circumference was greater in women with concomitant AI in year 4 (p = 0.05) than in those without concomitant AI.

Waist-to-hip ratio (WHR): In men, no differences in WHR were found between the groups, whereas in women, a significantly smaller increase in WHR was found in AI patients at year15 (p = 0.001).

Lipid profile and glucose metabolism (Fig. 4)

Total cholesterol: During GHRT total cholesterol decreased in both groups, but this decrease was significantly greater in the AI group from year 2 to year 15 (all p < 0.05).

Triglycerides: No significant differences between AI and non AI patients were found.

Low-density lipoprotein (LDL) cholesterol: Results are shown separately for patients with and without using lipid-lowering medication at baseline. Greater reductions in LDL cholesterol were seen in AI patients in years 3 to 5 (all p < 0.05) in the group of patients using lipid-lowering drugs, and in year 10 (p = 0.01) in the group of patients without lipid-lowering drugs.

High-density lipoprotein (HDL) cholesterol: No differences in HDL were found between the groups in both men and women.

HbA1c: A smaller increase in HbA1c at year 15 was seen in AI patients (p = 0.01).

Fully adjusted models

The differences compared to the unadjusted models are described below.

Blood pressure (Fig. 2)

Among patients using antihypertensive drugs at baseline or during follow-up, the increase in diastolic blood pressure at year 10 was lower in AI patients (p = 0.04).

Body composition (Fig. 3)

Hip circumference: among women, the difference in year 4 disappeared and no more differences were found in hip circumference between AI and non AI patients, while in men the increase in hip circumference was smaller in patients with concomitant AI in year 2 (p = 0.05).

Waist-to-hip ratio (WHR): in addition to a smaller increase in WHR in AI patients at year 15 among women, this was also found at year 10 (p = 0.05).

Lipid profile and glucose metabolism (Fig. 4)

Triglycerides: The decrease in triglycerides was greater in AI patients at year 15 (p = 0.05).

Low-density lipoprotein (LDL) cholesterol: in the group of patients without lipid-lowering drugs, a greater decrease in LDL cholesterol was found in AI patients in year 15 (p = 0.05) in addition to year 10.

High density lipoprotein (HDL) cholesterol: in women, a significantly smaller increase in HDL was found in AI patients at year 5 (p = 0.05).

Development of new-onset diabetes mellitus type 2, non-fatal cardiovascular- and cerebrovascular events (Fig. 5; Table 3) Fig. 5

Kaplan–Meier event-free survival curves for non-fatal cardiovascular disease, non-fatal cerebrovascular disease and new-onset type 2 diabetes mellitus

Table 3 The risk of developing non-fatal cardiovascular events, non-fatal cerebrovascular events and new-onset type 2 diabetes mellitus according to Cox hazard ratio for both groups

The Cox hazard ratio’s (HR) showed an increased risk of developing peripheral artery disease (HR: 2.22 [1.06–4.65]) and cerebrovascular disease (HR: 3.47 [1.60–7.52]) in the group of patients with concomitant AI compared with those without (Table 3, Kaplan–Meier curves shown in Fig. 5). However, this increased risk disappeared in the adjusted models. No differences were found in the risk of developing coronary artery disease, arrhythmias, other cardiac disease, any cardiovascular disease, or new-onset diabetes mellitus. There were also no differences in mean time to event (all p > 0.05).

Subanalysis of concomitant glucocorticoid medication (indications other than AI)Total group (AI and non-AI patients)

In the total group, 299 concomitant glucocorticoids were used in 227 (9%) patients during adult GHRT. Of these 227 patients, about a quarter (22%) used ≥ 2 glucocorticoids. Regarding route of administration, 136 patients used inhalers, 38 patients used transcutaneous forms, 35 patients took oral glucocorticoids, 33 intra-nasal and 4 used ocular glucocorticoids. Duration of glucocorticoid use was known for 73 of the 299 medications and had a mean of 696 ± 1163 [2–5717] days.

Cox Hazard ratios showed an increased risk of developing any non-fatal cardiovascular event (coronary heart disease, peripheral arterial disease, arrhythmia and other cardiac disease) in the group of patients using concomitant glucocorticoids compared to patients not using this medication, HR 1.57 [1.10–2.22], p = 0.01. No differences were found in the development of the subgroups of non-fatal cardiovascular events (coronary heart disease, peripheral arterial disease, arrhythmia and other cardiac disease), non-fatal cerebrovascular events, or new-onset diabetes mellitus type 2 (all p > 0.05, data not shown).

Comparisons between subgroups (AI versus non-AI)

There was no statistical difference in the number of patients in the AI (170/1836 = 9%) versus non-AI (57/633 = 9%) subgroup using concomitant glucocorticoids, χ2 = 0.04, p = 0.85.

Within the non-AI group, there were no differences in the development of any non-fatal cardiovascular events, non-fatal cerebrovascular events, or new-onset diabetes mellitus type 2 between patients taking concomitant glucocorticoids and those not taking this medication (all p > 0.05, data not shown)