Effect of GH Deficiency Caused by Nonfunctioning Pituitary Masses on Serum C-reactive Protein Levels

Abstract Context GH supplementation for GH deficiency (GHD) has been reported to decrease high-sensitivity C-reactive protein (hs-CRP), an inflammatory marker; however, the association between GHD and hs-CRP remains unclear. Objective We aimed to clarify the impact of impaired GH secretion due to pituitary masses on hs-CRP levels. Methods We retrospectively examined the association between GH secretion, assessed using GH-releasing peptide-2, and serum hs-CRP levels before and a year after the pituitary surgery in patients with nonfunctioning pituitary neuroendocrine tumor or Rathke cleft cyst. Results Among 171 patients, 55 (32%) presented with severe GHD (peak GH response to GH-releasing peptide-2 < 9 ng/mL). Serum hs-CRP levels were significantly higher in patients with severe GHD than in those without (P < .001) and significantly correlated with the peak GH (r = −0.50, P < .001). Multiple regression analyses showed that the peak GH significantly and negatively predicted hs-CRP levels (β = −0.345; 95% CI, −0.533 to −0.158) and the lowest quartile of the peak GH (<5.04 ng/mL) were significantly associated with increase in hs-CRP levels (exp [β] = 1.840; 95% CI, 1.209 to 2.801), after controlling for other anterior hormones and metabolic parameters. Postoperative change in the peak GH (N = 60) significantly predicted change in hs-CRP levels (β = −0.391; 95% CI, −0.675 to −0.108), independent of alterations in other anterior hormones and metabolic parameters. Conclusion The inverse association between GH secretion and hs-CRP levels highlights the protective role of GH in the increase in hs-CRP.

GH deficiency (GHD) is the most common endocrine deficit complicated by pituitary tumors [1] and is known to cause metabolic complications, such as obesity, dyslipidemia [2], and fatty liver [3], and premature atherosclerosis [4].The major causes of adult-onset GHD are nonfunctioning pituitary neuroendocrine tumors (NF-PitNETs) and Rathke cleft cysts (RCCs) [5].Patients with GHD reportedly present higher mortality and cardiovascular event risks than age-and sexmatched controls [6,7].Furthermore, an observational study reported a decreased risk of cardiovascular events after GH supplementation therapy [7].Therefore, GHD is believed to play a role in the pathogenesis of atherosclerosis, leading to cardiovascular events.
Atherosclerosis is characterized by a low-grade chronic inflammation of the arterial wall [8].C-reactive protein (CRP) is an established inflammatory biomarker and 1 of the most widely used biomarkers associated with the risk of cardiovascular events [9,10].Patients with hypopituitarism have been observed to exhibit elevated CRP levels [11].Serum high-sensitivity CRP (hs-CRP) levels decrease after GH supplementation in GH-deficient patients [12,13].These reports suggest an important role for GHD in the development of inflammation in patients with hypopituitarism; however, the association between impaired GH secretion and the development of inflammation remains unclear.
In the present study, we aimed to clarify the role of impaired GH secretion in the increase in CRP levels.We retrospectively investigated the association between GH secretion and serum hs-CRP levels in patients with NF-PitNET and RCC.

Study Population
In this retrospective cross-sectional study, we reviewed all available charts of adult patients with NF-PitNET or RCC who were admitted to the Department of Endocrinology and Hypertension at Tokyo Women's Medical University Hospital between January 2014 and March 2018.Patients with NF-PitNET or RCC who underwent GH-releasing peptide (GHRP)-2 testing and serum hs-CRP measurements were included in this study.Patients with an estimated glomerular filtration rate (GFR) of <30 mL/min/1.73m 2 , active inflammatory or malignant diseases, or a history of pituitary surgery, or under GH supplementation therapy were excluded.Background clinical characteristics including age, sex, body mass index (BMI), comorbidities, renal function, pituitary function, and pituitary mass size were retrieved.In patients who underwent pituitary surgery, serum hs-CRP levels, peak GH response to GHRP-2, and metabolic parameters a year after pituitary surgery were also retrieved.Obesity was defined as a BMI ≥25 kg/m 2 .This study was conducted in accordance with the 1964 Declaration of Helsinki and its amendments.Because the study was defined as one without human samples under the Japanese guidelines presented by the Ministry of Health, Labour and Welfare, written informed consent was not required, and we used our official institutional website as an opt-out method.This study was approved by the Ethics Committee of the Tokyo Women's Medical University Hospital (4856-R) and was designed and conducted in accordance with Strengthening the Reporting of Observational Studies in Epidemiology guidelines [14].

GHRP-2 Test
All patients with nonfunctioning pituitary masses underwent GHRP-2 tests to determine GH secretion.Briefly, pralmorelin hydrochloride (0.1 mg) (Kaken Pharmaceutical Co. Ltd., Tokyo, Japan) was IV injected at 9 AM, and blood samples were collected before and 15, 30, and 45 minutes after injection.Severe GHD was defined as a peak GH concentration <9 ng/mL [15].A GH cutoff value of 9 ng/mL with GHRP-2 corresponded to a GH value of 1.8 ng/mL with an insulin tolerance test when the GH value was calibrated with the recombinant World Health Organization 98/574 standard [15].Although the GHRP-2 test is not included in the Endocrine Society Guidelines [16], it is considered safe [17] and is widely used to diagnose severe GHD in Japan because of its high sensitivity and specificity compared with those of an insulin tolerance test [18].

Statistical Analyses
Statistical analyses were performed using JMP Pro software (version 16.0; SAS Institute, USA).Data are presented as mean ± SD or median (interquartile range) for descriptive analyses.Data normality was analyzed by visual inspection of the Q-Q plots and histograms.Given the skewed distribution, the peak GH response to GHRP-2, serum levels of hs-CRP, AST, ALT, GGT, HDL-cholesterol, LDL-cholesterol, and triglyceride, and pituitary mass volume were log-transformed for the analyses that follow.Statistical significance between 2 groups was calculated using the unpaired Student t test.Statistical significance between the values before and a year after pituitary surgery was calculated using the paired Student t test.
Multiple comparisons among the groups were performed using the Tukey test.The significance of categorical data was assessed using Fisher exact test.Univariate regression analyses were performed using Pearson correlation coefficients.Analysis of covariance was used to determine the differences between NF-PitNET and RCC after adjusting for mass size.
Because background characteristics, including other pituitary hormone secretions may bias the association between GH secretion and hs-CRP levels in multiple regression analyses, to estimate serum hs-CRP levels, explanatory factors were selected from peak GH response to GHRP-2; elevated liver enzyme levels as a complication of GHD; other anterior pituitary hormone deficiencies; and known factors associated with serum hs-CRP levels such as conventional coronary heart disease risk factors including age, sex, BMI, hypertension, diabetes mellitus, and hyperlipidemia [22], and serum creatinine level [23].In the analyses, peak GH response to GHRP-2 as an explanatory variable was used as a continuous variable or as a quartile group and thyroid function as a tertile group.The standardized β coefficient was calculated to compare the strength of each independent variable.We conducted a sensitivity analysis for patients without corticotroph deficiency, hypothyroidism, or hypogonadotropic hypogonadism.The number of cases during the study period determined the sample size.No missing data were imputed.Statistical significance was set at P < .05.
Mass volumes in patients with NF-PitNETs were significantly higher than those in patients with RCCs (4596  volumes in patients with 2 (P = .001),3 (P < .001),or 4 (P = .01)deficits in the anterior pituitary hormones were significantly higher than those in patients without anterior pituitary hormone deficiencies (Fig. 2B).
a Values were log-transformed for analyses.
with NF-PitNET and 17 with RCC) were included in the longitudinal analyses (Fig. 1).Table 4 shows the postoperative changes in clinical characteristics.
Postoperative changes in serum hs-CRP levels were significantly and inversely correlated with those in peak GH response to GHRP-2 (r = −0.46,P < .001)(Fig. 3F).In addition, the postoperative changes in serum hs-CRP levels were also significantly correlated with those in serum HDL-cholesterol levels (r = −0.40,P = .002)but not in BMI, systolic blood pressure, or serum levels of AST, ALT, GGT, HbA1c, LDL-cholesterol, and triglyceride (Table 5).Multiple linear regression analyses revealed that the postoperative change in the peak GH response to GHRP-2 (β = −0.392;95% CI, −0.673 to −0.111; standardized β = −0.354)was a significant factor in determining the change in serum hs-CRP levels after adjusting for changes in anterior pituitary hormone secretion other than GH (Table 6).When the model was further extended to include BMI, systolic blood pressure, and serum levels of ALT and HDL-cholesterol, the change in the peak GH response to GHRP-2 (β = −0.391;95% CI, −0.675 to −0.108; standardized β = −0.353)was also a significant factor in determining the change in serum hs-CRP levels.

Discussion
In this retrospective study, serum hs-CRP levels were significantly correlated with the peak GH response to GHRP-2, as a marker of GH secretion, in patients with NF-PitNET and RCC.Although other pituitary hormone deficiencies including corticotroph deficiency, hypothyroidism, and hypogonadotropic hypogonadism were significantly associated with increased serum hs-CRP levels, the peak GH response to GHRP-2 was a significant variable for estimating serum hs-CRP levels after adjusting for other pituitary hormone secretions as well as atherosclerotic risk factors, and elevated liver enzymes as possible confounders of serum hs-CRP levels in multiple linear regression analyses.The lowest quartile group of the GH secretion was significantly associated with serum hs-CRP levels after the adjustment.Moreover, alterations in the peak GH response to GHRP-2 after pituitary surgery were significantly and inversely correlated with changes in serum hs-CRP levels, independent of changes in other anterior pituitary hormone secretions and metabolic parameters.These findings revealed a significant and inverse relationship between GH secretion and serum hs-CRP levels, independent of atherosclerotic risk factors, metabolic consequences of GHD including obesity and fatty liver, and other anterior pituitary hormone deficiencies.
There are 2 important findings regarding the association between impaired GH secretion and the increase in the inflammatory marker.First, impaired GH secretion may be a key factor associated with the increase in hs-CRP levels in patients with nonfunctioning pituitary mass lesions.Previous studies have established that patients with hypopituitarism exhibit increased CRP levels [11].However, it was not clear which anterior pituitary hormone affected CRP levels.Our data show a significant association between the number of anterior pituitary hormone deficiencies and increased hs-CRP levels.Although adrenal [24], thyroid [25], and gonadal [26,27] insufficiencies have been reportedly associated with increased inflammation, the results of this study suggest that GH secretion, as indicated by the peak GH response to GHRP-2, is an independent factor affecting serum hs-CRP levels, independent of other anterior hormone secretions.A previous in vivo study reported that IGF-1 administration decreased vascular inflammatory cytokines [28].GHD, which leads to low IGF-1 levels, may also contribute to increase in inflammatory cytokines.Thus, impaired GH secretion resulting from nonfunctioning pituitary masses may play a pivotal role in the increase in CRP independent of other anterior pituitary hormones.
Second, GH-deficient states may be associated with an increase in the inflammatory marker, independent of its complications.GHD can cause obesity [29] and fatty liver [3].A previous report demonstrated that white adipose tissue inflammation may contribute to elevated CRP levels in obese patients [30].In addition, hepatic steatosis and obesity are independently associated with increased odds of high hs-CRP levels [31].In the present study, high serum hs-CRP levels were associated with GHD independent of obesity and elevated liver enzyme levels.Similarly, serum CRP levels were decreased by GH supplementation therapy, independent of changes in body composition [12,29].Because the monocyte/macrophage-specific IGF1 receptor knockout model showed increased pro-inflammatory cytokine production in macrophages [32], a decrease in IGF-1 receptor signaling by GHD might lead to inflammation in macrophages and increase serum hs-CRP levels, at least in part, independent of the development of obesity and fatty liver.The lack of a significant correlation between IGF-1 SD scores and serum hs-CRP levels in our study may be explained by the fact that the main source of circulating IGF-1 is the liver [33] and its low diagnostic accuracy in estimating GH secretion [4].Thus, impaired GH secretion may contribute to the increase in the inflammatory marker, independent of the metabolic complications of GHD.To the best of our knowledge, this is the first study to reveal a direct relationship between GH secretion and CRP levels independent of other pituitary hormone secretions and the complications of GHD.
Our findings indicate that assessing GH secretion in patients with pituitary disease is important for understanding the risk of cardiovascular events.Although observational studies have shown that patients receiving GH replacement have reduced mortality compared with patients without GH replacement, there is no convincing evidence that treatment of GHD improves mortality rates [34].Glucocorticoid overreplacement and radiation therapy may bias the mortality of patients with GHD [34].Consequently, adult GHD is perceived as having a low priority in some countries [35].The direct association between impaired GH secretion and the inflammatory marker demonstrated in our study suggests an important protective role for GH in the development of cardiovascular diseases through inflammation.
Our study has some limitations.First, the mechanism linking the decrease in GH secretion to CRP production was not investigated.Pro-inflammatory cytokines such as IL-1 and IL-6 induce CRP gene expression [36]; however, the effect of GH supplementation on these cytokines has been contradictory [37].Further studies on the association between GH secretion and cytokine levels are warranted.Second, the diagnosis of GHD was confirmed using the GHRP-2 test, which has only been approved in Japan [38].However, a high correlation between the GH response to a GHRP-2 test and that to an insulin tolerance test has been reported [18].The nonpeptide GH secretagogue macimorelin, approved in the United States and Europe, also showed a high diagnostic accuracy comparable to that of the insulin tolerance test [39].The significant association between GH secretion and serum hs-CRP levels shown in our study should be confirmed in further studies using macimorelin to assess GH secretion.
In conclusion, this retrospective study revealed a significant and inverse association between GH secretion and serum hs-CRP levels, independent of atherosclerotic risk factors; metabolic complications of GHD, including obesity and fatty liver; and other anterior pituitary hormone deficiencies in patients with NF-PitNET and RCC.Postoperative changes in serum hs-CRP levels were significantly and inversely correlated with those in the peak GH response to GHRP-2.These results suggest an important protective role of GH secretion in increasing cardiovascular disease risk through inflammation, independent of the metabolic complications of GHD.

Figure 2 .
Figure 2. Characteristics of the pituitary function of the patients.(A) Association between the prevalence of GHD, hypogonadotropic hypogonadism, hypothyroidism, and corticotroph deficiency, and the deficit number of anterior pituitary hormones.(B) Association between the pituitary mass volume and the number of anterior pituitary hormone deficits.Abbreviations: CD, corticotroph deficiency; GHD, GH deficiency; HH, hypogonadotropic hypogonadism; HT, hypothyroidism; NF-PitNET, nonfunctioning pituitary neuroendocrine tumor; RCC, Rathke cleft cyst.

Figure 3 .
Figure 3. Association between serum hs-CRP levels and background characteristics.(A) Serum hs-CRP levels in the patients with NF-PitNET and RCC.(B) Correlation between serum hs-CRP levels and pituitary mass volumes in patients with NF-PitNET (open circles and dotted line) and RCC (closed circles and solid line).The bold solid line represents the regression slope in all the patients.(C) Association between serum hs-CRP levels and the number of anterior pituitary hormone deficits.(D) Serum hs-CRP levels in patients with and without severe GHD.(E) Correlation between serum hs-CRP levels and peak GH response to GHRP-2 in patients with NF-PitNET (open circles and dotted line) and RCC (closed circles and solid line).The bold solid line represents the regression slope in all the patients.(F) Correlation between changes in serum hs-CRP levels and peak GH response to GHRP-2 before and a year after pituitary surgery.The solid line represents the regression slope.Abbreviations: GHD, GH deficiency; GHRP, GH-releasing peptide; hs-CRP, high-sensitivity C-reactive protein; NF-PitNET, nonfunctioning pituitary neuroendocrine tumor; RCC, Rathke cleft cyst.

Table 4 . Postoperative changes in pituitary function and metabolic parameters in the 60 patients with postoperative assessments Baseline A year after pituitary surgery P
Figure 4. Association between serum hs-CRP levels and GH secretion.Forest plot demonstrating exponential transformation of β coefficients (exp [β]) and 95% CIs of quartiles of peak GH response to GHRP-2 levels in the multivariable regression analyses to estimate serum hs-CRP levels, with the highest quartile (Q4) as reference.Because the dependent variable was log-transformed, the β coefficients were exponentially transformed.Exponential transformation of β coefficients (Exp [β]) implies an interpretation on a multiplicative scale.Exp (β) implies an interpretation on a multiplicative scale.The serum hs-CRP level changes by 100 × (exp [β])% in the quartiles of peak GH response to GHRP-2 levels compared with Q4.Model 1: adjusted for age, sex, obesity, hypertension, diabetes mellitus, dyslipidemia, and creatinine level.Model 2: adjusted for factors used in model 1 + corticotroph deficiency, thyroid function, and hypogonadotropic hypogonadism.Model 3: Adjusted for the factors used in model 2 + elevated liver enzymes.Abbreviations: GHRP, GH-releasing peptide; hs-CRP, high-sensitivity C-reactive protein.

Table 5 . Correlation between the postoperative changes in serum hs-CRP levels a and those in metabolic parameters characteristics
aValues were log-transformed for analyses.

Table 6 . Multiple linear regression analyses with the postoperative changes in serum hs-CRP levels a
Model 1: the model includes postoperative changes in peak GH response to GHRP-2 and number of anterior hormone deficiencies (excluding GH) as explanatory factors.Model 2: adjusted for factors used in model 1 + postoperative changes in BMI, SBP, ALT, and HDL-cholesterol.Abbreviations: ALT, alanine aminotransferase; BMI, body mass index; GHRP, GH-releasing peptide; HDL, high-density lipoprotein; hs-CRP, high-sensitivity C-reactive protein; SBP, systolic blood pressure.a Values were log-transformed for analyses.