Growth Hormone Replacement Therapy in Adult Growth Hormone Deficiency and Risk of Cancer : A Meta-Analysis

The growth hormone (GH) replacement therapy in adult growth hormone deficiency (AGHD) is now well developed, nevertheless, the safety of GH replacement, especially the incidence of cancer in these patients remains to be further clarified. To summarize the evidence on the safety of using GH in AGHD, we conduct this meta-analysis to assess the relationship between the risk of cancer and GH replacement therapy. Randomized controlled trials and cohort studies involved in GH therapy for AGHD were selected. Meta-analysis was performed and risk ratio (RR) was pooled with 95% confidence interval (CI) to investigate the relationship between GH replacement and the risk of cancer. The result indicated that there was no evidence to draw a conclusion that GH replacement therapy will increase the risk of cancer (P = 0.001, RR = 0.77, 95% CI [0.65, 0.90]). Meanwhile, according to the calculated analysis, the replacement therapy might even reduce the risk of cancer. Furthermore, subgroup analysis demonstrated that there was no correlation between replacement therapy of GH and the risk of cancer both in prospective and retrospective cohort design research, and in prospective group, the risk of cancer even decreased (P = 0.0002, RR = 0.71, 95%CI [0.59, 0.85]). In conclusion, our study corroborates evidence from previous studies showing that GH replacement therapy in AGHD patients would not increase the risk of cancer; instead, it might be even decrease cancer risk. The results suggested that GH replacement therapy in AGHD patients was safe. #These authors contributed equally to this work. How to cite this paper: Zeng, L., Song, X.X., Lin, C.H., Ho, J.K., Yu, P.X., Jaiswal, S. and Xu, X.H. (2017) Growth Hormone Replacement Therapy in Adult Growth Hormone Deficiency and Risk of Cancer: A Meta-Analysis. Open Journal of Endocrine and Metabolic Diseases, 7, 173-189. https://doi.org/10.4236/ojemd.2017.79016 Received: August 22, 2017 Accepted: September 23, 2017 Published: September 26, 2017 Copyright © 2017 by authors and Scientific Research Publishing Inc. This work is licensed under the Creative Commons Attribution International License (CC BY 4.0). http://creativecommons.org/licenses/by/4.0/ Open Access


Introduction
Growth hormone (GH) was originally used to remedy short stature in childhood [1].As experience with the treatment improved, the range of indications was expanded.The GH replacement therapy for adults with hypopituitarism and GH deficiency is now well established, however, some reports have raised concern about the safety of GH therapy.The biological effect of GH is associated with the GH-GH receptor-insulin-like growth factor 1 (IGF-1) axis, while IGF-1 has significant effects on cell proliferation and differentiation [2].Epidemiological studies had shown a relationship between elevated circulating levels of IGF-1 and an increased risk of several types of cancers such as prostate, colorectal and breast neoplasms [3] [4] [5].The association between GH-IGF axis and carcinogenesis was also announced by the study in which patients enrolled suffering from acromegaly, which characterized by hypersecretion of GH and elevated levels of IGF-1.This research showed that the individual with GH replacement therapy has a higher risk of developing thyroid and colorectal carcinoma [6].A meta-regression analysis conducted by Renehan et al. drew a same conclusion [7].Data from the Pharmacia International Metabolic Surveillance (KIMS) study showed that the overall occurrence of the novo cancer or the rate of recurrence for primary pituitary adenomas were not increased [8].There is ongoing concern about the potential mitogenic property of GH-IGF-1 axis and the safety of GH replacement in adult because of currently widely used of GH replacement therapy.
To further definitely resolve this cancer-related safety issue in GH replacement therapy, specifically in AGHD patients, we therefore used meta-analysis methodology to assess published studies which examined incidence of cancer in AGHD patients who applied GH replacement therapy.

Literature Search Strategy
PUBMED, Web of Science, Ovid, Cochrane library, CBM, CNKI and Wan Fang database were systematically searched for articles published from the beginning of the database to July 2017.Keywords included in the search were "growth hormone deficiency/GHD/AGHD/adult growth hormone deficiency/hypopituitarism" AND "GH therapy/GH treatment/growth hormone treatment/growth hormone therapy/replacement" AND "cancer/tumor/neoplasm" All our work in this systematic review referred to the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) guidelines [9].All analyses were based on previous published studies, thus no ethical approval and patient consent are re-Open Journal of Endocrine and Metabolic Diseases quired.

Study Inclusion Criteria
Two investigators assessed the eligibility of all retrieved papers and disagreement was resolved by a thorough discussion.Studies were preserved if they satisfied the following criteria: 1) case-control or cohort study; 2) participants were aged at least 18 years old or above; 3) reported definitions and measurement of AGHD; 4) reported risk, odds or hazard ratios across different exposure categories of GH; 5) reported morbidity outcomes for cancer; 6) published in English or Chinese language.

Data Extraction and Quality Assessment
Data pertaining to the study and participant details, intervention measures, follow-up duration as well as research outcomes were extracted independently by two reviewers using a standardized data extraction form and reconciled for accuracy.
Studies reported tumorigenesis outcomes of AGHD patients treated with GH were considered for inclusion in the analysis.For studies that were eligible but without detailed number of tumorigenesis, the investigators would contact with authors in order to get the additional information essential for analysis.
Quality of inclusion studies was assessed by the Newcastle-Ottawa Scale [10].A "star system" was used to evaluate data quality based on three broad aspects for studies: the selection, the comparability and the outcome.The scores ranged from 0 to 9. Researches with scores of 6 stars to 9 stars were regarded to be of high quality.The quality as well as the bias of included studies was cross-checked by two reviewers independently and disagreements between two reviewers would be settled by open discussion.

Statistical Analysis
Meta-analysis was performed with extracted data by statistical software Revman 5.0, the interesting outcome of this meta-analysis was the incidence of cancer.Risk ratio values and the corresponding 95% confidence interval were used to estimate the relationship between GH replacement therapy and risk of cancer in AGHD patients.
Statistical analysis was performed as previously described by Tsilid is et al. [11]; all analyses were conducted using the fixed effects model.Methods of fixed effect meta-analysis were based on the assumption that a single common (or "fixed") effect underlies every study in the meta-analysis without any heterogeneity between studies.We assessed heterogeneity between studies using the P value of the χ 2 based Cochrane Q test and the I 2 metric of inconsistency; this could reflect either genuine diversity or bias.The Q test is obtained by the weighted sum of the squared differences of the observed effect in each study minus the fixed summary effect.
The I 2 metric ranges between 0% and 100% and is the ratio of variance between studies over the sum of the variances within and between studies.We further conducted subgroup analyses stratified by study characteristics according to prospective study design or retrospective study design.Then, in order to assess the risk of bias, we drew a funnel plot using the software of Revman 5.0.As the funnel plot was symmetrical, it prompted that the literature publication bias was controlled passably.Finally, Sensitivity analysis was performed to quantify the effect proportion of each included studies (through a "leave-one-out" method) on the overall evaluation.

Study Identification
An online database search with the specified terms retrieved 3640 records presented or published.After removing duplicate records, 2993 items were left.Then 64 articles were selected for full-text screening for eligibility through reading the titles and abstracts.After final screening, data from 10 selected studies published from 2002 to 2017 were extracted for inclusion in the current analysis [12]- [21].
A flow diagram of literature search methodologies is provided in Figure 1.

Study Characteristics
No randomized, controlled studies comparing GH replacement therapy versus without GH replacement therapy in AGHD patients were identified.All studies included in this analysis were prospective or retrospective researches.Overall, the current meta-analysis included 13,404 patients and the largest studies which were reported by Child et al. (2015) accounted for more than half of these patients.Half of the studies defined pituitary tumors as the outcome.Four studies mentioned second neoplasms.The follow-up duration was no less than 10 years in five studies.The majority of patients were with an average age of around 50 years old, while the study conducted by Brignardello et al. did not reveal the age of patients.All studies but one described the proportion of gender distribution.Baseline characteristics for patients included in this analysis are outline in Table 1.The quality assessment showed that all of the selected studies obtained 6 or more scores evaluating by the NOS, which indicated the studies quality were generally high.

Publication Bias
Symmetrical funnel plot showed in Figure 2 proved little publication bias.

Quantitative Synthesis
Results of the meta-analysis about risk of cancer in AGHD patients treated with GH compared with untreated group are shown in Figure 3.No significant association was revealed between GH replacement therapy and cancer risk (P = 0.001, RR = 0.77, 95% CI [0.65, 0.90]).As the forest plot shown, the risk of cancer in AGHD patients treated with GH might even decrease compared with the untreated

Sensitivity Analyses
For sensitivity analyses, we left out each of the studies in turn, in order to investigate the effect of a single study on the overall risk estimation (Table 2).When we excluded the study "Buchfelder 2007" and "Olsson 2012" that measured at least 5-year tumor progression-free survival rate (PFSR) in pituitary adenomas and craniopharyngioma patients on long-term growth hormone replacement therapy, the pooled risk ratio was 0.74 (P = 0.0005, 95% CI [0.63, 0.88]) and 0.81 (P = 0.02, 95% CI [0.69, 0.96]) respectively (Figure 4 and Figure 5).The I 2 was 0% and P value was 0.45 after the exclusion of "Buchfelder", the I 2 was 0% and P value was 0.70 after the exclusion of "Olsson 2012", suggested that the study "Buchfelder 2007" and "Olsson 2012" caused heterogeneity mainly.

Subgroup Analysis
Subgroup analysis based on study design demonstrated that the pooled risk ratio was statistically significant in prospective studies (P = 0.0002, RR = 0.71, 95% CI [0.59, 0.85]).Heterogeneity (I 2 = 3%, P = 0.4) was low indicating the consistency of the decreased risk of cancer for GH treated group in prospective studies.In the retrospective subgroup, there was no significant differences between the GH treated group and control group as the pooled statistic shown (P = 0.45, RR = 1.16, 95% CI [0.79, 1.69]) (Figure 6).The pooled risk ratio suggested that the risk of cancer was not increasing in GH treated group.There was no heterogeneity (I 2 = 0%, P = 0.93), indicating the consistency of no increasing risk of cancer for GH treated group in retrospective studies.

Discussion
A doubtful association between GH replacement therapy and risk of developing cancer or recurrence of tumor has long been speculated.A number of studies have been undertaken to investigate the safety of GH replacement therapy.More data about the safety of GH treatment are available in children, while the safety problem of GH treatment in adult especially for cancer was not fully established.
A systematic review and meta-analysis was previously carried out to assess the association between GH therapy and cancer risk in patients treated with GH during childhood and adolescence, the result showed that both the overall cancer standardized incidence ratio (SIR) 2.74 (95%CI 1.18 -5.42) and risk ratio (RR) 1.99 (95% CI 1.28 -3.08, P = 0.002) for second neoplasms were significantly increased [22].And another meta-analysis showed that the risk of recurrence for pediatric brain tumors was not increased in patients treated with GH [23].On the contrary, Liang Shen et al. indicated that GH replacement therapy decreased the risk of recurrence or progression in children with intracranial tumors, such as craniopharyngioma, medulloblastoma, astrocytoma, or glioma, but not for pituitary adenomas, non-functional pituitary adenomas, and ependymoma.There was no relation between GHRT and pituitary adenomas, non-functional pituitary adenomas [24].And a study of pediatric GH treatment which conducted by Child C.J et al. also showed that if there has no history of malignancy, the risk of primary cancer did not increase.[25].patients in hypopituitary control and complications study compared with cancer rates of general population, standardized by country, gender, and age [30].Bunderen et al. played a review to discuss the long-term efficacy and safety of GH treatment in AGHD patients with emphasis on morbidity [31].In a subgroup of the review, the risk of regrowth and recurrences of pituitary tumors were not increasing in AGHD patients treated with GH compared with untreated AGHD patients, while secondary brain tumors remained more prevalent.Generally, the conclusion showed that fatal and nonfatal malignancies were not more prevalent in GH-treated adults compared to the general population.In our meta-analysis, we compared the risk of cancer between GH-treated AGHD patients with the patients without receiving GH treatment.Furthermore, it differed from Bunderen et al. who focused on morbidity; the only outcome of our study was the risk of cancer in AGHD patients treated with GH.And unlike the review previously carried out by Kirstine Stochholm et al. showed that the GH replacement did not increase the cancer risk [32], the consequence of our study was more positive, we found the risk of cancer in AGHD patients treated with GH might even decrease.
Our meta-analysis showed that AGHD patients treated with GH demonstrated no increased risk of cancer compared with the untreated group; moreover, the result indicated that GH replacement therapy might even decrease the risk of cancer.The subgroup analysis revealed that GH treatment was associated with decreased risk of cancer in prospective design group, whereas no significant difference of cancer risk was observed between GH treated group and untreated group in retrospective design group.
The above observations led us to hypothesize that the possible mechanism might be related to the improvement of body composition, exercise performance and the individual GH dose adjusted according to the serum IGF-1 level regularly conducted by titration regimen.
It was well recognized that AGHD patients had an abnormal body composition with a decrease in body lean mass, whereas an increase in total and visceral fat mass which was related to insulin resistance [33].Adipose tissues release a variety of different small protein factors, including chemokines [monocyte chemotactic protein 1 (MCP-1)], IL-6, IL-1, TNF-α, adipokines [haptoglobin, leptin, visfatin, resistin, and vascular endothelial growth factor (VEGF)], on the other hand, the secretion of anti-inflammatory adipokines (adiponectin, IL-10, IL-1) decline [34] [35].TNF-α appears to contribute to the development of the tissue architecture necessary for tumor growth and metastasis [36] [37], while IL-6 regulates chronic inflammation, which can create a cellular microenvironment beneficial to cancer growth [38].Releasing of short-range cytokines boosts the local expansion of other immune cells and the regeneration of damaged tissue, these include strong inducers of proliferation, metastasis, and angiogenesis; releasing of long-range cytokines (IL-6, TNF-α) will boost insulin resistance in the liver and other metabolism-controlling organs, thus generating increased levels of insulin that may in turn Open Journal of Endocrine and Metabolic Diseases promote cancer growth [39].In contrast, adiponectin with anti-atherogenic, anti-inflammatory, and insulin-sensitivity effects might have the function to inhibit the growth of tumor [40].In consistent with this hypothesis, several case-control studies have verified that serum adiponectin levels were significantly decreased in breast cancer patients [41].Studies have reported that GH replacement was associated with a significant reduction in body fat and improvement in insulin resistance [33].Therefore, we believe that a decrease of total and visceral fat mass result from the replacement therapy of GH may reduce the risk of cancer in AGHD patients.
Aside from being associated with the improvement of body composition, there may be other potential mechanism.Study revealed that physical activity have the effect to protect against colon, breast and endometrial cancers, besides, insufficient physical activity levels were also estimated to cause 9% of breast cancer cases and 10% of colon cancer cases in Europe [42].Observational studies demonstrated that physical activity was independently related to decreased risk of all-cause mortality in cancer survivors [43].A great deal of researches has showed that GH replacement was associated with an improvement in exercise tolerance [18], and probably decreased the risk of cancer.
During the period of GH replacement treatment, the dose of GH was adjusted regularly according to serum level of IGF-1.Previously human population studies revealed that higher IGF-1 levels could be associated with increased risk of cancer [44].The study conducted by Popovic et al. concluded that IGF-1 levels targeted to within normal age-related reference ranges during GH replacement were not associated with the occurrence of malignancies [45].Current guidelines recommended that high inter-individual variability in both GH absorption and sensitivity makes the stepwise, individualized, upward titration method rather than standard weight-based dosing strategies, and individualized-dose titration regimens leads to similar beneficial effects and fewer side effects than weightbased regimens [46].Consequently, all the cohort studies included in our analysis managed the dose of GH according to the individual titration regimen, which might be the other reason for the outcome.
The "leave-one-out" analysis revealed no substantial changes after exclusion of each study from the pooled analyses except the study "Buchfelder 2007" and "Olsson 2012".When the study "Buchfelder 2007" was left out, the result (RR = 0.74, 95% CI [0.63, 0.88]) and when the study "Olsson 2012" was left out, the result (RR = 0.81, 95% CI [0.67, 1.05]), which showed that the risk of cancer for GH replacement therapy group was not increased compared with the untreated group.However, the indicator I² was declined to 0% and P value was 0.45 and 0.70 respectively.This result suggested study "Buchfelder 2007" and "Olsson 2012" were the main source of heterogeneity.None of the cross-sectional studies was of low quality according to the NOS; therefore, we did not perform sensitivity analyses based on study quality.
The result of the detail information about gender and age were unavailable in Open Journal of Endocrine and Metabolic Diseases There are some limitations in our study.First of all, available data were limited to non-randomized studies; all studies included were only retrospective or prospective cohort studies.Secondly, the study populations were heterogeneous, and the case number of the non GH-treated group was much smaller than the GH-treated group.Thirdly, for GH replacement therapy, there might have selection bias, untreated patients might be older, sicker, and more likely to have an intracranial tumor than treated patients.Additionally, the optimum duration of tumorigenesis or recurrence was probable beyond the follow-up period; similar analysis with longer follow-up is needed to confirm assurance of no increase, even decrease about the risk of cancer during GH therapy in adults.

Conclusion
This meta-analysis suggests that in AGHD patients treated with GH, there are no major safety concerns associated with cancer risk; moreover, GH replacement therapy might even decrease the risk of cancer for AGHD patients.Similar results were also reflected from the analysis conducted by Liang Shen et al. in children with intracranial tumors, craniopharyngioma, medulloblastoma, astrocytoma, or glioma [24].Ideally, this result should be confirmed by a large-scale multicenter prospective randomized study with sufficiently long follow-up duration.

Figure 1 .
Figure 1.Flow chart of identification process for eligible studies.

Figure 6 .
Figure 6.Forest plot of subgroup analysis based on study design.
study conducted by Buchfelder et al. and Brignardello et al., the subgroup analyses stratified by gender and age were unable to implement.Meanwhile, the inconsistent definition of event outcomes covered tumor recurrence and second neoplasms made it difficult to precise classification, so we were unable to perform the subgroup analyses according to classification of tumor.Finally, the doses of GH were individually conducted by titration regimens instead of specific dose, which induced the implementation of dose-response analyses difficulty.

Table 1 .
Characteristic of studies included in the meta-analysis.
"None" means no study was left out.
Study performed by Chung T.T. et al. observed that GH treatment was not associated with increased tumor recurrence or second neoplasms [29].Child et al. demonstrated no increased risk of primary cancers in GH-treated AGHD Open Journal of Endocrine and Metabolic Diseases