- Email: [email protected]
Association between growth hormone therapy and mortality, cancer and cardiovascular risk: Systematic review and meta-analysis
YGHIR-01009; No of Pages 7 Growth Hormone & IGF Research xxx (2014) xxx–xxx
Contents lists available at ScienceDirect
Growth Hormone & IGF Research journal homepage: www.elsevier.com/locate/ghir
Review
Association between growth hormone therapy and mortality, cancer and cardiovascular risk: Systematic review and meta-analysis Annalisa Deodati a, Barbara Baldini Ferroli a, Stefano Cianfarani a,b,⁎ a b
D.P.U.O. “Bambino Gesù” Children's Hospital — “Tor Vergata” University, Rome, Italy Department of Women's and Children's Health, Karolinska Institutet, Stockholm, Sweden
a r t i c l e
i n f o
Article history: Received 19 February 2014 Accepted 23 February 2014 Available online xxxx Keywords: Growth hormone Growth hormone therapy IGF-I Metaanalysis
a b s t r a c t Objective: The potential involvement of growth hormone therapy in tumor promotion and progression has been of concern for several decades. Our aim was to assess systematically the association between growth hormone therapy and all-cause, cancer and cardiovascular mortality, cancer morbidity and risk of second neoplasm mainly in patients treated during childhood and adolescence. Design: A systematic review of all articles published until September 2013 was carried out. The primary efficacy outcome measures were the all-cause, cancer and cardiovascular standardized mortality ratios (SMR). The secondary efficacy outcome measures were the standardized incidence ratio (SIR) for cancer and the relative risk (RR) for second neoplasms. The global effect size was calculated by pooling the data. When the effect size was significant in a fixed model we repeated the analyses using a random model. Results: The overall all-cause SMR was 1.19 (95% CI 1.08–1.32, p b 0.001). Malignancy and cardiovascular SMRs were not significantly increased. Both the overall cancer SIR 2.74 (95% CI 1.18–5.41), and RR for second neoplasms 1.99 (95% CI 1.28–3.08, p = 0.002), were significantly increased. Conclusion: The results of this meta-analysis may raise concern on the long-term safety of GH treatment. However, several confounders and biases may affect the analysis. Independent, long-term, well-designed studies are needed to properly address the issue of GH therapy safety. © 2014 Elsevier Ltd. All rights reserved.
1. Introduction The growth promoting action of growth hormone (GH) is mainly mediated by IGF-I, which, in combination with the GH-independent IGF-II, exerts its actions on the cells in endocrine, paracrine and autocrine manner. The signaling transduction cascade induced by the binding of IGFs mainly to the IGF-I receptor, eventually leads to a potent stimulation of cell proliferation and survival [1]. Due to their antiapoptotic and mitogenic effects, the role of IGFs in cancer growth and development has been extensively investigated. While there is strong evidence based on experimental data obtained in cellular and animal models showing a role of the GH-IGF axis in the development, maintenance and spread of tumors, such evidence in humans is weak [2]. Epidemiological studies have shown an association between raised circulating levels of IGF-I and an increased risk of developing certain cancers such as prostate, breast and colorectal neoplasms [3–5]. The association between GH-IGF and carcinogenesis is also suggested by the observation that patients suffering from acromegaly, an endocrine
⁎ Corresponding author at: “Bambino Gesù” Children's Hospital, Piazza S. Onofrio 4, 00165-Rome, Italy. Tel.: +39 06 6859 3074; fax: +39 06 6859 2508. E-mail address: [email protected] (S. Cianfarani).
disorder characterized by sustained hypersecretion of GH and consequent increased levels of IGF-I, have a higher risk of developing colorectal and thyroid cancer [6–9]. In childhood, recombinant human growth hormone (rhGH) has been extensively used since 1985 to treat children with short stature secondary to a range of disorders including GH deficiency, Turner syndrome, chronic renal failure, small for gestational age (SGA), Prader– Willi syndrome, Noonan syndrome, SHOX deficiency and idiopathic short stature (ISS) [10]. The experience from many thousands of patient years of treatment demonstrates a good safety record for rhGH. Nevertheless, a few reports have raised concern about the long-term safety of GH therapy. In the 1980s, the potential link between GH treatment and malignancy was suggested by case reports linking GH therapy with leukemia risk [11]. A further detailed analysis of this cohort revealed that most of these patients had concomitant conditions predisposing them to cancer, thus leading to overestimation of the risk of malignancy following GH treatment. Reassuringly, the risk of leukemia was not increased in the National Cooperative Growth study, a large, ongoing cohort study, initiated in 1985, of children treated with GH in the USA [12]. In 2002, a long-term study of subjects treated with human pituitary GH during childhood and early adulthood showed an increased risk of mortality from cancer overall, and from colorectal cancer and Hodgkin disease in particular [13]. These conflicting data suggest that
http://dx.doi.org/10.1016/j.ghir.2014.02.001 1096-6374/© 2014 Elsevier Ltd. All rights reserved.
Please cite this article as: A. Deodati, et al., Association between growth hormone therapy and mortality, cancer and cardiovascular risk: Systematic review and meta-analysis, Growth Horm. IGF Res. (2014), http://dx.doi.org/10.1016/j.ghir.2014.02.001
2
A. Deodati et al. / Growth Hormone & IGF Research xxx (2014) xxx–xxx
2.2. Statistical analysis
Fig. 1. Search strategy for selection of mortality studies.
long-term surveillance remains crucially important, not only for those being treated with rhGH but also for those who have already discontinued this treatment. The aim of this systematic review and metanalaysis was to examine the evidence that GH treatment during childhood may be associated with a higher risk of all-cause, cancer and cardiovascular mortality and morbidity.
2. Methods We searched the Medline, EMBASE, ISI Web of Knowledge, and the bibliographic references from all retrieved articles describing such studies up to September 2013 using the search terms “growth hormone” and “rhGH therapy” and “mortality” and “cancer” and “incidence” and “morbidity” and “safety”. No language restrictions were applied. Inclusion criteria were treatment with rhGH therapy during childhood and adolescence and long-term follow-up.
2.1. Efficacy outcome measures and quality assessment The primary efficacy outcome measure was the all-cause, cancer and cardiovascular mortality rate, using the standardized mortality ratio (SMR), defined as the number of observed deaths divided by the number of expected deaths stratified for gender and age in the reference population. The secondary efficacy outcome measures were the cancer incidence rate, using the standardized incidence ratio (SIR), defined as the number of observed neoplasm divided by the number of expected cases and the risk of second neoplasms, using the relative risk (RR), defined as the incidence of second neoplasms among exposed to GH divided by the incidence among non-exposed. None of the studies, except one [14], provided raw data on single participants, therefore in the analysis we considered the average values for SMR, SIR and RR and their standard errors.
For primary analysis we calculated the effect size for each study. The effect size was computed as the SMR, SIR, and RR for each trial. We present these scores in a paired analysis with their 95% confidence intervals. Then we calculated the global effect size, pooling the data. When the effect size was significant in a fixed model we repeated the analyses using a random effects model [15]. The random effects model incorporates statistical heterogeneity (results, methods, and publication bias) and provides a more conservative estimate of the pooled effect size than a fixed model. We calculated I2 values for quantifying heterogeneity in the meta-analysis. I2 describes the percentage of variability in point estimates that is due to heterogeneity rather than to sampling error. Although no universal rule covers the definitions of mild, moderate, or severe heterogeneity, I2 values more than 50% indicate notable heterogeneity, whereas values less than 30% indicate mild heterogeneity. We assessed publication bias by funnel plot analysis (see web extras). Analyses were carried out using Review Manager 5 software for Windows package (Nordic Cochrane Centre, Copenhagen, Denmark) and double checked using STATA 12.0 statistical software (StataCorp, TX, USA). 3. Results 3.1. Mortality studies The search strategy identified 12 long-term studies concerning patient mortality. Eight studies were excluded; four reported observed deaths only [16–19] and four used indices different than SMR[20–23]. Four [13,24–26] out of the 12 studies used SMR rates to evaluate mortality (Fig. 1). These studies included overall 24,456 patients, with a mean chronological age at study enrolment of 32.6 ± 10.5 years. Mean GH dose in two studies [24,25] was 0.415 ± 0.28 mg/day, while only one study [26] reported a mean dose of 0.024 mg/kg/day. Duration of treatment was reported in only two studies [24,26], with an average of 4.8 ± 4.5 years (Table 1). 3.1.1. All-cause SMR Van Bunderen et al. [24] reported a significant increase of all-cause SMR in patients enrolled in the Dutch National Registry of Growth Hormone Treatment. The patients were retrospectively monitored and subdivided into three groups: a treatment group (n = 2229), a primary control group (who had not commenced treatment or who had discontinued it before 30 days, n = 109) and a secondary control group (who had endured treatment for more than 30 days but less than 90, n = 356). This cohort included both adult and childhood onset GH deficiency (about 80 vs 20%, respectively). The all-cause SMR was 1.27 (95% CI 1.04–1.56) for the treatment group, 1.42 (95% CI 0.79–2.56) for the primary control group and 1.17 (95% CI 0.82– 1.69) for the secondary control group. Gaillard et al. [25] reported data obtained from the analysis of KIMS (Pfizer International Metabolic Database) including 13,983 GH-deficient patients with 69,056 patient-years of follow-up. The all-cause SMR was 1.13 (95% CI 1.04–1.24). The cohort mainly consisted of patients with adult onset GH deficiency, about 20% showing childhood onset GH
Table 1 Mortality studies with SMR analysis. Study
Journal
Year
No. of patients
Age at study enrollment (years)
Dose of treatment
Duration of treatment
All-cause SMR
Cancer SMR
CVD SMR
Carel et al. [26] Gaillard et al. [25] Van Bunderen et al.[24] Swerdlow et al. [13]
JCEM EJE JCEM LANCET
2012 2012 2011 2002
6558 13,983 2229 1352
28.3 ± 5.3 26.9 ± 9.9 42.6 ± 16.3 –
0.024 μg/kg/day 0.42 ± 0.27 mg/day 0.41 ± 0.28 mg/day –
3.9 ± 2.6 – 5.7 ± 6.3 –
1.33 (1.08–1.649) 1.13 (1.04–1.24) 1.27 (1.04–1.56) –
1.02 (0.41–2.09) 0.88 (0.74–1.03) 0.86 (0.6–1.25) 2.3 (0.8–5)
3.07 (1.4–5.83) 0.83 (0.63–1.08) 1.35 (0.95–1.94) –
Please cite this article as: A. Deodati, et al., Association between growth hormone therapy and mortality, cancer and cardiovascular risk: Systematic review and meta-analysis, Growth Horm. IGF Res. (2014), http://dx.doi.org/10.1016/j.ghir.2014.02.001
A. Deodati et al. / Growth Hormone & IGF Research xxx (2014) xxx–xxx
3
Fig. 2. The all-cause SMR in GH treated subjects. Results of meta-analysis according to random model.
deficiency. Interestingly, these authors reported a significant negative association between mortality risk and IGF-I levels during therapy, the less IGF-I the higher the SMR. Carel et al. [26] analyzed patients enrolled in the French populationbased register, as part of a European project aimed at evaluating safety and appropriateness of GH treatments in Europe (SAGhE). The initial cohort (n = 6,892) included patients with childhood onset idiopathic isolated GH deficiency (n = 5162), neurosecretory dysfunction (n = 534), idiopathic short stature (n = 871), or born short for gestational age (n = 335). Follow-up data were available for final analysis in 94.7% of the cohort (n = 6558). The reported all cause SMR was 1.33 (95% CI 1.08–1.64). The overall all-cause SMR in GH treated subjects, mostly children, included in these three studies was significantly increased: 1.19 (95% CI 1.08–1.32, p b 0.001) (Fig. 2). See web extra Supplemental Data 1 and 2 for the results in the analysis carried out according to a fixed model and funnel plot (Supplemental Data 1 and 2).
3.1.2. Malignancy SMR Four studies reported SMRs for deaths due to primary neoplasms in subjects treated with GH[13,24–26]. Swerdlow et al. [13]analyzed a cohort of patients (n = 1848) in the UK who were treated during childhood and early adulthood with human pituitary growth hormone from 1959 to 1985, compared to
that of the general population, controlling for age, sex and calendar period. They found a malignancy SMR of 2.3 (95% CI 0.8–5) after excluding patients with conditions predisposing to cancer (496 subjects). In the cohort of patients reported by Van Bunderen et al. [24] the SMR was 0.86 (95% CI 0.6–1.25) in the treated group, 0.39 (95% CI 0.05–2.75) in the primary control group and 1.14 (95% CI 0.63–2.05) in the secondary control group. In the French cohort [26], malignancy SMR for treated patients was 1.02 (95% CI 0.41–2.09). Finally, in the report from Gaillard et al. [25] the malignancy SMR was 0.88 (95% CI 0.74–1.03). The overall mean malignancy SMR for treated patients was not significantly increased: 0.94 (95% CI 0.74–1.19, p = 0.61) (Fig. 3, and Supplemental Data 3 and 4).
3.1.3. Cardiovascular (CVD) SMR Mortality due to cardiovascular events was analyzed in three studies [24–26]. In the Dutch study [24] the calculated CVD SMR was 1.35 (95% CI 0.95–1.94) for the treated group (n = 2229), 1.45 (95% CI 0.54–3.87) for the primary control group (n = 109), and 0.12 (95% CI 0.02–0.87) for the secondary control group (n = 356). In the French cohort (n = 6,558) [26], the CVD SMR resulted significantly increased: SMR = 3.07 (95% CI 1.40–5.83). In this study, an increase in mortality due subarachnoid or intracerebral hemorrhage (SMR 6.66, 95% CI 1.79–17.05) was also observed. Finally, data from KIMS (n = 13,983) [25] reported a CVD SMR of 0.83 (95% CI 0.63–
Fig. 3. The overall malignancy SMR in GH treated subjects. Results of meta-analysis according to random model.
Fig. 4. The overall cardiovascular SMR in GH treated subjects. Results of meta-analysis according to random model.
Please cite this article as: A. Deodati, et al., Association between growth hormone therapy and mortality, cancer and cardiovascular risk: Systematic review and meta-analysis, Growth Horm. IGF Res. (2014), http://dx.doi.org/10.1016/j.ghir.2014.02.001
4
A. Deodati et al. / Growth Hormone & IGF Research xxx (2014) xxx–xxx
(GHD) and revealed a significantly increased risk of malignancy: SIR = 2.74 (95% CI 1.18–5.41). Wilton et al. [14] reported data from a large cohort of patients (n = 58,603) treated with GH and recruited in KIGS—the Pfizer International Growth Database—without cancer or any other condition in medical history known to increase the risk of cancer. The reported malignancy SIR was 1.26 (95% CI 0.86–1.78). The above cited Swerdlow's study [13] found a malignancy SIR of 1.4 (95% CI 0.5–2.8) after excluding high risk groups. Bell et al. [12] reported data from the National Cooperative Growth Study (NCGS), including 54,996 children enrolled between 1985 and 2006 in USA. This study showed a malignancy SIR of 1.12 (95% CI 0.75–1.61). Taking all the reports together, the overall mean malignancy SIR was 1.36 (95% CI 1.00–1.85, p = 0.05), showing a slightly significantly increased risk of primary neoplasm. (Fig. 6 and Supplemental Data 7 and 8). Fig. 5. Search strategy for selection of morbidity studies for primary cancer.
3.3. Risk of second neoplasms 1.08). The mean cardiovascular SMR derived from all these studies was not significantly increased, SMR = 1.39 (95% CI 0.76–2.55, p = 0.28), (Fig. 4, and Supplemental Data 5 and 6).
3.2. Incidence studies The search strategy identified 7 long-term studies concerning primary cancer incidence. Three studies were excluded: two reported only the observed neoplasms [27] and [28] and one used rates different than SIR[29]. Four studies used SIR to evaluate the incidence of primary cancers [12–14,30] (Fig. 5). These studies included 121,791 patients, with a mean dose of treatment of 0.28 mg/day and a mean duration of treatment of 4.3 years (Table 2).
3.2.1. Malignancy SIR Child et al. [30] reported data from the Hypopituitary Control and Complications Study (Hypoccs) pharmaco-epidemiological database. They evaluated the incidence of de novo and second cancers in 6840 GH-treated and 940 non GH-treated adult patients, 20% with childhood onset GH deficiency. The reported SIR was 0.88 (95% CI 0.74–1.04) for the whole cohort treated with GH. However, the analysis focused only on patients from the USA with childhood onset (CO) GH deficiency
The search strategy identified 17 long-term studies concerning second cancer incidence. 12 studies were excluded; seven [16,31–36] reported only the absolute number of observed cases; 2 studies [37, 38] were excluded because they reported tumor recurrence for brain cancer in subjects with primary diagnosis of CNS tumor and three were excluded for reporting cancer incidence using the hazard ratio (HR) [39–41]. Five studies used relative risk (RR) to evaluate the incidence of secondary tumors [42–46] (Fig. 7). These studies included 908 patients, with a mean duration of treatment of 5 years (0.1–14). Mean age at primary cancer diagnosis of 13.06 years (0–45). The mean GH dose reported in only two studies was 0.26 mg/kg/week [44] (Table 3).
3.3.1. RR second neoplasms Ergun-Longmire et al. [42] reported data from a cohort of 361 patients enrolled in the Childhood Cancer Survivor Study. The RR was 2.15 (95% CI 1.33–3.47) compared with non-GH-treated survivors. Meningiomas were the most common second neoplasm among the GH-treated group. Mackenzie et al. [43] included in the analysis previously irradiated patients (n = 110), the patient group comprised a mixture of children and adult patients. Patients had received GH therapy for a median of 8 years (range 1–19) and were followed up for a median
Table 2 Cancer morbidity studies with SIR analysis.
Child et al. [30] Wilton et al. [14] Swerdlow et al. [13] Bell et al. [12]
Journal
Year
No. of patients
Age at study enrollment (years)
Dose of treatment
Duration of treatment (years)
Cancer SIR
EJE JP LANCET JCEM
2011 2010 2002 2010
6840 58,603 1352 54,996
46.4 (34.1–56.3) 10.3 ± 4 – 10.3
0.30 (0.20–0.46) mg/day 0.25 ± 0.1 mg/kg/week – –
– 3.6 (0.08–9.70) – 3.6
2.74 (1.18–5.41) 1.26 (0.86–1.78) 1.4 (0.5–2.8) 1.12 (0.75–1.61)
Fig. 6. The overall malignancy SIR in GH treated subjects. Results of meta-analysis according to random model.
Please cite this article as: A. Deodati, et al., Association between growth hormone therapy and mortality, cancer and cardiovascular risk: Systematic review and meta-analysis, Growth Horm. IGF Res. (2014), http://dx.doi.org/10.1016/j.ghir.2014.02.001
A. Deodati et al. / Growth Hormone & IGF Research xxx (2014) xxx–xxx
5
tumor in cancer survivors who underwent GH therapy during childhood (Fig. 9 and Supplemental Data 11 and 12). 4. Discussion
Fig. 7. Search strategy for selection of morbidity studies for second neoplasms.
period of 14.5 years. RR for second neoplasm was 2.50 (95% CI 0.495–12.61). Woodmansee et al. [44] assessed the incidence of second neoplasm in a cohort of childhood cancer survivors enrolled in the Eli Lilly and Company's pediatric (Genetics and Neuroendocrinology of Short Stature International Study (GeNeSIS)) and adult (Hypopituitary Control and Complications Study (HypoCCS)) observational studies (n = 394). The RR for second neoplasm was 0.83 (95% CI 0.201–3.457). In the pediatric cohort the incidence of second neoplasms in untreated patients was 0/27 whereas the incidence in the GH treated group was 15/394. In the calculation of RR we considered an incidence of 1/27 in the untreated cohort in order to perform the analysis. The RR for second neoplasms in the pediatric cohort was 1.3 (95% CI 0.14–7.49). Wing-Leung et al. [45] reported the risk of second neoplasms in cancer survivors with GH deficiency secondary to ALL treatment (n = 43). The RR was 1.58 (95% CI 0.38–6.50). The study of Sklar et al. [46] reported the risk of second neoplasms in different subgroups of cancer survivors according to different diagnoses: CNS tumors, leukemia and rhabdomyosarcoma (n = 361). The RR of second neoplasm in patients treated with rhGH after the initial diagnosis of CNS tumors was 2.34 (95% CI 0.96–5.70). The RR in patients with primary diagnosis of leukemia was 4.98 (95% CI 1.95–12.74) and the RR in subjects with initial diagnosis of rhabdomyosarcoma was 1.82 (95% CI 0.41–8.01). Initially, we assessed the recurrence of second neoplasms including all selected studies [42–46]. The mean RR for second neoplasms was 2.34 (95% CI 1.64–3.23, p b 0.00001) showing a higher risk of second tumor in cancer survivors who underwent GH therapy during childhood (Fig. 8 and Supplemental Data 9 and 10). Thereafter, we have considered only the studies reporting the general relative risk of second tumor [42–44]. The mean RR for second neoplasms was 1.99 (95% CI 1.28–3.08, p = 0.002), confirming a significantly higher risk of second
In this systematic review all the available studies reporting all-cause SMR, cancer SMR, CVD SMR, cancer SIR and second neoplasm RR in patients, mostly children, treated with GH were included. The results show a significant increase of all-cause mortality but no significant increase of the malignancy and CVD mortality. The morbidity assessment (SIR) showed a significant increase of cancer incidence. Finally, the risk for second neoplasms in patients treated with GH after a primary cancer was significantly increased. Although these data may raise concern about the long-term safety of GH treatment a number of potential confounders and biases should be considered. First of all the study populations were heterogeneous, comprising in many reports both adult and pediatric cohorts, but, even more importantly, including patients with different diagnoses. With the only exception of one study [26], all the other studies did not stratify the patients according to diagnostic groups. The potential different risk related to a pharmacological rather than a replacement therapy is intuitive. In Carel's [26] cohort, only patients defined as at low risk, were assessed. However, subjects born small for gestational age (SGA) were included in the analysis, thus potentially biasing the results not only because this diagnostic group may comprise undefined congenital syndromes, but also because there is strong evidence for a higher cardiovascular risk in subjects born SGA. The Swerdlow's cohort [13], including only patients treated with pituitary derived GH before 1985, in theory could be considered the most homogenous since the only approved indication at that time was GH deficiency. However, the detailed description of diagnoses revealed that GH deficiency was of idiopathic origin in 53% of patients and attributable to intracranial neoplasia in 26%. Many other disorders affected the remaining 21% of patients, including cases of leukemia or lymphoma, chromosome fragility syndromes, and glycogen storage diseases. Furthermore, all these studies were also affected by a relatively limited sample size and, especially, a low event rate. They also lack validated external reference in that no untreated “control” group was available for comparison and no reference mortality data corrected for the potential effect of height were available. The collected data were obtained from multiple centers throughout a country or even the world thus including many unidentified confounding factors and missing data. A further potential confounder is the lack of key data such as familial predisposition to certain diseases, exposure to environmental hazards, lack of local mortality and morbidity indices. Moreover, in most studies, the mean dose and duration of GH treatment as well as the follow-up length were not reported. Further limitations are inherent in the meta-analysis approach. One potential limitation of any meta-analysis is the pooling of studies from heterogeneous populations. Another potential bias is the inclusion of study populations mostly derived from sponsored post-marketing surveillance programs. A third potential limitation of any meta-analysis is the “file-drawer” effect, in which studies with negative or “unexpected”
Table 3 Studies reporting the relative risk (RR) of second neoplasm. Sklar's [46] and Ergun-Longmire's [42] studies report data from the same cohort. Study
Journal
Sklar et al. [46] JCEM Ergun-Longmire et al. [42] JCEM Mackenzie et al. [43] JCEM Wing Leung et al. [45] JCO Woodmansee et al. [44] EJE
Year
No. of patients
Age at primary diagnosis (yrs)
Dose of treatment
Duration of treatment
RR second neoplasm
RR second neoplasm after leukemia
RR second neoplasm after CSN tumor
2002 2006 2011 2002 2013
361 361 110 43 394
3.5 (0–17.2) 3.5 (0–17.2) 33 (14–45) 10.35 ± 1.7 5.4 (3–8.5)
– – – 0.3 mg/kg/week 0.22 ± 0.03 mg/kg/week
4.6 (0.1–14) 4.6 (0.1–14) 8 (4–10) 4.5 (1–8) 2.9 (1.4–4.8)
– 2.15 (1.33–3.47) 2.5 (0.495–12.61) – 1.3 (0.14–7.49)
4.98 (1.95–12.74) – – 1.58 (0.376–6.53) –
2.34 (0.96–5.7) – – – –
Please cite this article as: A. Deodati, et al., Association between growth hormone therapy and mortality, cancer and cardiovascular risk: Systematic review and meta-analysis, Growth Horm. IGF Res. (2014), http://dx.doi.org/10.1016/j.ghir.2014.02.001
6
A. Deodati et al. / Growth Hormone & IGF Research xxx (2014) xxx–xxx
Fig. 8. The overall RR of second neoplasm including GH treated subjects with specific primary diagnosis. Results of meta-analysis according to random model.
Fig. 9. The overall RR of second neoplasm in GH treated subjects. Results of meta-analysis according to random model.
results might remain unpublished thus biasing the literature towards positive findings. Ideally, an appropriately designed study should be characterized by an adequate sample size, detailed diagnostic characterization and stratification according to the potential risks, accurate documentation of GH treatment regimens, long-term follow-up and the availability of a control cohort. The realization of such ambitious study implies the involvement of multiple scientific societies as well as regulatory authorities and pharmaceutical industry. In conclusion, although the meta-analysis of available data has provided evidence for a statistically significant increase of all-cause mortality, primary cancers and second neoplasms in patients treated with GH, these results may be affected by several biases and confounders which need to be addressed in further large-scale controlled surveys. Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.ghir.2014.02.001. References [1] D. Le Roith, C. Bondy, S. Yakar, J.L. Liu, A. Butler, The somatomedin hypothesis: 2001, Endocr. Rev. 22 (2001) 53–74. [2] P.E. Clayton, I. Banerjee, P.G. Murray, A.G. Renehan, Growth hormone, the insulinlike growth factor axis, insulin and cancer risk, Nat. Rev. Endocrinol. 7 (2011) 11–24. [3] J.M. Chan, M.J. Stampfer, E. Giovannucci, P.H. Gann, J. Ma, P. Wilkinson, C.H. Hennekens, M. Pollak, Plasma insulin-like growth factor-I and prostate cancer risk: a prospective study, Science 279 (1998) 563–566. [4] S.E. Hankinson, W.C. Willett, G.A. Colditz, D.J. Hunter, D.S. Michaud, B. Deroo, B. Rosner, F.E. Speizer, M. Pollak, Circulating concentrations of insulin-like growth factor-I and risk of breast cancer, Lancet 351 (1998) 1393–1396. [5] J. Ma, M.N. Pollak, E. Giovannucci, J.M. Chan, Y. Tao, C.H. Hennekens, M.J. Stampfer, Prospective study of colorectal cancer risk in men and plasma levels of insulin-like growth factor (IGF)-I and IGF-bindingprotein-3, J. Natl. Cancer Inst. 91 (1999) 620–625. [6] D. Baris, G. Gridley, E. Ron, E. Weiderpass, L. Mellemkjaer, A. Ekbom, J.H. Olsen, J.A. Baron, J.F. Fraumeni Jr., Acromegaly and cancer risk: a cohort study in Sweden and Denmark, Cancer Causes Control 13 (2002) 395–400. [7] R. Kauppinen-Mäkelin, T. Sane, M.J. Välimäki, H. Markkanen, L. Niskanen, T. Ebeling, P. Jaatinen, M. Juonala, Finnish Acromegaly Study Group, E. Pukkala, Increased cancer incidence in acromegaly—a nationwide survey, Clin. Endocrinol. 72 (2009) 278–279. [8] S.M. Orme, R.J. McNally, R.A. Cartwright, P.E. Belchetz, Mortality and cancer incidence in acromegaly: a retrospective cohort study. United Kingdom Acromegaly Study Group, J. Clin. Endocrinol. Metab. 83 (1998) 2730–2734.
[9] E. Ron, G. Gridley, Z. Hrubec, W. Page, S. Arora, J.F. Fraumeni Jr., Acromegaly and gastrointestinal cancer, Cancer 68 (1991) 1673–1677. [10] J. Kirk, Indications for growth hormone therapy in children, Arch. Dis. Child. 97 (2012) 63–68. [11] S. Watanabe, S. Mizuno, L.H. Oshima, Y. Tsunematsu, J. Fujimoto, A. Komiyama, Leukemia and other malignancies among GH users, J. Pediatr. Endocrinol. 6 (1993) 99–108. [12] J. Bell, K.L. Parker, R.D. Swinford, A.R. Hoffman, T. Maneatis, B. Lippe, Long-term safety of recombinant human growth hormone in children, J. Clin. Endocrinol. Metab. 95 (2010) 167–177. [13] A.J. Swerdlow, C.D. Higgins, P. Adlard, M.A. Preece, Risk of cancer in patients treated with human pituitary growth hormone in the UK, 1959–85: a cohort study, Lancet 360 (2002) 273–277. [14] P. Wilton, A.F. Mattsson, F. Darendeliler, Growth hormone treatment in children is not associated with an increase in the incidence of cancer: experience from KIGS (Pfizer International Growth Database), J. Pediatr. 157 (2010) 265–270. [15] L.V. Hedges, T.D. Pigott, The power of statistical tests in meta-analysis, Psychol. Methods 6 (2001) 203–221. [16] S.L. Blethen, D.B. Allen, D. Graves, G. August, T. Moshang, R. Rosenfeld, Safety of recombinant deoxyribonuclei acid-derived growth hormone: The National Cooperative Growth Study Experience, J. Clin. Endocrinol. Metab. 81 (1996) 1704–1710. [17] C.A. Quigley, A.M. Gill, B.J. Crowe, K. Robling, J.J. Chipman, S.R. Rose, J.L. Ross, F.G. Cassorla, A.M. Wolka, J.M. Wit, L.T.M. Rekers-Mombarg, G.B. Cutler Jr., Safety of growth hormone treatment in pediatric patients with idiopathic short stature, J. Clin. Endocrinol. Metab. 90 (2005) 5188–5196. [18] M.L. Hartmann, R. Xu, B.J. Crowe, L.L. Robinson, E.M. Erfurth, D.L. Kleinberg, A.G. Zimmerman, W.W. Woodmansee, G.B. Cutler Jr., J.J. Chipman, S. Melmed, Prospective safety surveillance of GH-deficient adults: comparison of GH-treated vs untreated patients, J. Clin. Endocrinol. Metab. 98 (2013). [19] L. Savendahl, M. Maes, K. Albertsson-Wilkand, B. Borgstrom, J.C. Carel, S. Henrard, N. Speybroeck, M. Thomas, G. Zandwijken, A. Hokken-Koelega, Long-term mortality and causes of death in isolated GHD, ISS, and SGA patients treated with recombinant growth hormone during childhood in Belgium, The Netherlands, and Sweden: preliminary report of 3 countries participating in the EU SAGhE study, J. Clin. Endocrinol. Metab. 97 (2010) E213–E217. [20] K. Stochholm, C.H. Gravholt, T. Laursen, P. Laurberg, M. Andersen, L.O. Kristensen, U. Feldt-Rasmussen, J.S. Christiansen, M. Frydenberg, A. Green, Mortality and GHdeficiency: a nationwide study, Eur. J. Endocrinol. 157 (2007) 9–18. [21] S.P. Taback, H.J. Dean, Members of the Canadian Growth Hormone Advisory Committee. Mortality in Canadian children with growth hormone (GH) deficiency receiving GH therapy 1967–1992, J. Clin. Endocrinol. Metab. 81 (1996) 1693–1696. [22] K. Stochholm, S. Juul, J.S. Christiansen, C.H. Gravholt, Mortality and socioeconomic status in adults with childhood onset GH deficiency (GHD) is highly dependent on the primary cause of GHD, Eur. J. Endocrinol. 167 (2012) 663–670. [23] J. Svensson, B.A. Bengtsson, T. Rosen, A. Odén, G. Johannsson, Malignant disease and cardiovascular morbidity in hypopituitary adults with or without growth hormone replacement therapy, J. Clin. Endocrinol. Metab. 89 (2004) 3306–3312.
Please cite this article as: A. Deodati, et al., Association between growth hormone therapy and mortality, cancer and cardiovascular risk: Systematic review and meta-analysis, Growth Horm. IGF Res. (2014), http://dx.doi.org/10.1016/j.ghir.2014.02.001
A. Deodati et al. / Growth Hormone & IGF Research xxx (2014) xxx–xxx [24] C.C. Van Bunderen, C. van Nieuwpoort, L.I. Arwert, M.W. Heymans, A.A.M. Franken, H.P.F. Koppeschaar, A.J. van der Lely, M.L. Drent, Does growth hormone replacement therapy reduce mortality in adults with growth hormone deficiency? Data from the Dutch National Registry of Growth Hormone Treatment in Adults, J. Clin. Endocrinol. Metab. 96 (2011) 3151–3159. [25] R.C. Gaillard, A.F. Mattsson, A.C. Akerblad, B.A. Bengtsson, J. Cara, U. FeldtRasmussen, M. Koltowska-Haggstrom, J.P. Monson, B. Saller, P. Wilton, R. Abs, Overall and cause-specific mortality in GH-deficient adults on GH replacement, Eur. J. Endocrinol. 166 (2012) 1069–1077. [26] J.C. Carel, E. Ecosse, F. Landier, D. Meguellati-Hakkas, F. Kaguelidou, G. Rey, J. Coste, Long-term mortality after recombinant growth hormone treatment for isolated growth hormone deficiency or childhood short stature: preliminary report of the French SAGhE study, J. Clin. Endocrinol. Metab. 97 (2012) 416–425. [27] D. Wyatt, Lessons from the national cooperative growth study, Eur. J. Endocrinol. 151 (2004) S55–S59. [28] Y. Nishi, T. Tnaka, K. Takano, K. Fujieda, Y. Igarashi, K. Hanew, T. Hirano, S. Yokoya, K. Tachibana, K. Saito, S. Watanabe, Recent status in the occurrence of leukemia in growth hormone-treated patients in Japan, J. Clin. Endocrinol. Metab. 84 (1999) 1961–1965. [29] V. Popovic, A.F. Mattson, R.C. Gaillard, P. Wilton, M. Koltowska-Häggström, M.B. Ranke, Serum insulin-like growth factor I (IGF-1), IGF-binding proteins 2 and 3, and the risk for development of malignancies in adults with growth hormone (GH) deficiency treated with GH: data from KIMS (Pfizer International Metabolic Database), J. Clin. Endocrinol. Metab. 95 (2010) 4449–4454. [30] C.J. Child, A.G. Zimmermann, W.W. Woodmansee, D.M. Green, J.J. Li, H. Jung, E.M. Erfurth, L.L. Robinson, Assessment of primary cancers in GH-treated adult hypopituitary patients: an analysis from the Hypopituitary Control and Complications Study, Eur. J. Endocrinol. 165 (2011) 217–223. [31] D.S. Olsson, M. Buchfelder, S. Schlaffer, B.A. Bengtsson, K.E. Jakobsson, G. Johannsson, A.G. Nilsson, Comparing progression of non-functioning pituitary adenomas in hypopituitarism patients with and without long-termGH replacement therapy, Eur. J. Endocrinol. 161 (2009) 663–669. [32] M. Buchefelder, P.H. Kann, C. Wuster, U. Tuschy, B. Saller, G. Brabant, A. Kleindienst, P. Nomikos, the German KIMS Board, Influence of GH substitution therapy in deficient adults on the recurrence rate of hormonally inactive pituitary adenomas: a case–control study, Eur. J. Endocrinol. 157 (2007) 149–156. [33] T.T. Chung, J. Evanson, D. Walker, S.A. Akker, G.M. Besser, J.P. Monson, W.M. Drake, Safety of GH replacement in hypopituitary patients with nonirradiated pituitary and peripituitary tumours, Clin. Endocrinol. 68 (2008) 965–969. [34] T.T. Chung, W.M. Drake, J. Evanson, D. Walker, P.N. Plowman, S.L. Chew, A.B. Grossman, G.M. Besser, J.P. Monson, Tumour surveillance imaging in patients with extrapituitary tumours receiving growth hormone replacement, Clin. Endocrinol. 63 (2005) 272–279.
7
[35] L.R. Meacham, K. Sullivan, Characteristics of growth hormone therapy for pediatric patients with brain tumours in the National Cooperative Growth Study (NCGS) and from a survey of pediatric endocrinologists, J. Pediatr. Endocrinol. Metab. 15 (2002) 689–696. [36] S.J. Howell, P. Wilton, A. Lindberg, S.M. Shalet, Growth hormone replacement and the risk of malignancy in children with neurofibromatosis, J. Pediatr. 133 (1998) 201–205. [37] A.L. Ogyvil-Stuart, W.D.J. Ryder, H.R. Gattamanemi, P.E. Clayton, S.M. Shalet, Growth hormone and tumour recurrence, BMJ 304 (1992) 1601–1605. [38] A.J. Swerdlow, R.E. Reddingius, C.D. Higgins, H.A. Spoudeas, K. Phipps, Z. Qiao, W.D.J. Ryder, M. Brada, R.D. Hayward, C.G.D. Brook, P.C. Hindmarsh, S.M. Shalet, Growth hormone treatment of children with brain tumors and risk of tumor recurrence, J. Clin. Endocrinol. Metab. 85 (2000) 4444–4449. [39] N. Karavitaki, J.T. Warner, A. Marland, B. Shine, F. Ryan, J. Arnold, Helen E. Turner, J.A. H. Wass, GH replacement does not increase the risk of recurrence in patients with craniopharyngioma, Clin. Endocrinol. 64 (2006) 556–560. [40] J.R. Arnold, F. Arnold, A. Marland, N. Karavitaki, J.A.H. Wass, GH replacement in patients with non-functioning pituitary adenoma (NFA) treated solely by surgery is not associated with increased risk of tumour recurrence, Clin. Endocrinol. 70 (2009) 435–438. [41] D.S. Olsson, M. Buchfelder, K. Wiendieck, N. Kremenevskaja, B.A. Bengtsson, K.E. Jakobsson, M. Jarfelt, G. Johannsson, A.G. Nilsson, Tumour recurrence and enlargement in patients with craniopharyngioma with and without GH replacement therapy during more than 10 years of follow-up, Eur. J. Endocrinol. 166 (2012) 1061–1068. [42] B. Ergun-Longmire, A.C. Mertens, P. Mitby, J. Qin, G. Heller, W. Shi, Y. Yasui, L.L. Robinson, C.A. Sklar, Growth hormone treatment and risk of second neoplasms in the childhood cancer survivor, J. Clin. Endocrinol. Metab. 91 (2006) 3494–3498. [43] S. Mackenzie, T. Craven, H.R. Gattamaneni, R. Swindell, S.M. Shalet, G. Brabant, Long-term safety of growth hormone replacement after CNS irradiation, J. Clin. Endocrinol. 96 (2011) 2756–2761. [44] W.W. Woodmansee, A.G. Zimmermann, C.J. Child, Q. Rong, E.M. Erfurth, P. BeckPeccoz, W.F. Blum, L.L. Robinson, on behalf of the GeNeSIS and HypoCCS International Advisory Boards, Incidence of second neoplasm in childhood cancer survivors treated with GH: an analysis of GeNeSIS and HypoCCS, Eur. J. Endocrinol. 168 (2013) 565–573. [45] W. Leung, S.R. Rose, Y. Zhou, M.L. Hancock, S. Burstein, E.A. Schriock, R. Lustig, R.K. Danish, W.E. Evans, M.M. Hudson, C.H. Pui, Outcomes of growth hormone replacement therapy in survivors of childhood acute lymphoblastic leukemia, J. Clin. Oncol. 20 (2002) 2959–2964. [46] C.A. Sklar, A.C. Mertens, P. Mitby, G. Occhiogrosso, J. Qin, G. Heller, Y. Yasui, L.L. Robison, Risk of Disease recurrence and second neoplasms in survivors of childhood cancer treated with growth hormone: a report from the childhood center survivors study, J. Clin. Endocrinol. Metab. 87 (2002) 3136–3141
Please cite this article as: A. Deodati, et al., Association between growth hormone therapy and mortality, cancer and cardiovascular risk: Systematic review and meta-analysis, Growth Horm. IGF Res. (2014), http://dx.doi.org/10.1016/j.ghir.2014.02.001
Contents lists available at ScienceDirect
Growth Hormone & IGF Research journal homepage: www.elsevier.com/locate/ghir
Review
Association between growth hormone therapy and mortality, cancer and cardiovascular risk: Systematic review and meta-analysis Annalisa Deodati a, Barbara Baldini Ferroli a, Stefano Cianfarani a,b,⁎ a b
D.P.U.O. “Bambino Gesù” Children's Hospital — “Tor Vergata” University, Rome, Italy Department of Women's and Children's Health, Karolinska Institutet, Stockholm, Sweden
a r t i c l e
i n f o
Article history: Received 19 February 2014 Accepted 23 February 2014 Available online xxxx Keywords: Growth hormone Growth hormone therapy IGF-I Metaanalysis
a b s t r a c t Objective: The potential involvement of growth hormone therapy in tumor promotion and progression has been of concern for several decades. Our aim was to assess systematically the association between growth hormone therapy and all-cause, cancer and cardiovascular mortality, cancer morbidity and risk of second neoplasm mainly in patients treated during childhood and adolescence. Design: A systematic review of all articles published until September 2013 was carried out. The primary efficacy outcome measures were the all-cause, cancer and cardiovascular standardized mortality ratios (SMR). The secondary efficacy outcome measures were the standardized incidence ratio (SIR) for cancer and the relative risk (RR) for second neoplasms. The global effect size was calculated by pooling the data. When the effect size was significant in a fixed model we repeated the analyses using a random model. Results: The overall all-cause SMR was 1.19 (95% CI 1.08–1.32, p b 0.001). Malignancy and cardiovascular SMRs were not significantly increased. Both the overall cancer SIR 2.74 (95% CI 1.18–5.41), and RR for second neoplasms 1.99 (95% CI 1.28–3.08, p = 0.002), were significantly increased. Conclusion: The results of this meta-analysis may raise concern on the long-term safety of GH treatment. However, several confounders and biases may affect the analysis. Independent, long-term, well-designed studies are needed to properly address the issue of GH therapy safety. © 2014 Elsevier Ltd. All rights reserved.
1. Introduction The growth promoting action of growth hormone (GH) is mainly mediated by IGF-I, which, in combination with the GH-independent IGF-II, exerts its actions on the cells in endocrine, paracrine and autocrine manner. The signaling transduction cascade induced by the binding of IGFs mainly to the IGF-I receptor, eventually leads to a potent stimulation of cell proliferation and survival [1]. Due to their antiapoptotic and mitogenic effects, the role of IGFs in cancer growth and development has been extensively investigated. While there is strong evidence based on experimental data obtained in cellular and animal models showing a role of the GH-IGF axis in the development, maintenance and spread of tumors, such evidence in humans is weak [2]. Epidemiological studies have shown an association between raised circulating levels of IGF-I and an increased risk of developing certain cancers such as prostate, breast and colorectal neoplasms [3–5]. The association between GH-IGF and carcinogenesis is also suggested by the observation that patients suffering from acromegaly, an endocrine
⁎ Corresponding author at: “Bambino Gesù” Children's Hospital, Piazza S. Onofrio 4, 00165-Rome, Italy. Tel.: +39 06 6859 3074; fax: +39 06 6859 2508. E-mail address: [email protected] (S. Cianfarani).
disorder characterized by sustained hypersecretion of GH and consequent increased levels of IGF-I, have a higher risk of developing colorectal and thyroid cancer [6–9]. In childhood, recombinant human growth hormone (rhGH) has been extensively used since 1985 to treat children with short stature secondary to a range of disorders including GH deficiency, Turner syndrome, chronic renal failure, small for gestational age (SGA), Prader– Willi syndrome, Noonan syndrome, SHOX deficiency and idiopathic short stature (ISS) [10]. The experience from many thousands of patient years of treatment demonstrates a good safety record for rhGH. Nevertheless, a few reports have raised concern about the long-term safety of GH therapy. In the 1980s, the potential link between GH treatment and malignancy was suggested by case reports linking GH therapy with leukemia risk [11]. A further detailed analysis of this cohort revealed that most of these patients had concomitant conditions predisposing them to cancer, thus leading to overestimation of the risk of malignancy following GH treatment. Reassuringly, the risk of leukemia was not increased in the National Cooperative Growth study, a large, ongoing cohort study, initiated in 1985, of children treated with GH in the USA [12]. In 2002, a long-term study of subjects treated with human pituitary GH during childhood and early adulthood showed an increased risk of mortality from cancer overall, and from colorectal cancer and Hodgkin disease in particular [13]. These conflicting data suggest that
http://dx.doi.org/10.1016/j.ghir.2014.02.001 1096-6374/© 2014 Elsevier Ltd. All rights reserved.
Please cite this article as: A. Deodati, et al., Association between growth hormone therapy and mortality, cancer and cardiovascular risk: Systematic review and meta-analysis, Growth Horm. IGF Res. (2014), http://dx.doi.org/10.1016/j.ghir.2014.02.001
2
A. Deodati et al. / Growth Hormone & IGF Research xxx (2014) xxx–xxx
2.2. Statistical analysis
Fig. 1. Search strategy for selection of mortality studies.
long-term surveillance remains crucially important, not only for those being treated with rhGH but also for those who have already discontinued this treatment. The aim of this systematic review and metanalaysis was to examine the evidence that GH treatment during childhood may be associated with a higher risk of all-cause, cancer and cardiovascular mortality and morbidity.
2. Methods We searched the Medline, EMBASE, ISI Web of Knowledge, and the bibliographic references from all retrieved articles describing such studies up to September 2013 using the search terms “growth hormone” and “rhGH therapy” and “mortality” and “cancer” and “incidence” and “morbidity” and “safety”. No language restrictions were applied. Inclusion criteria were treatment with rhGH therapy during childhood and adolescence and long-term follow-up.
2.1. Efficacy outcome measures and quality assessment The primary efficacy outcome measure was the all-cause, cancer and cardiovascular mortality rate, using the standardized mortality ratio (SMR), defined as the number of observed deaths divided by the number of expected deaths stratified for gender and age in the reference population. The secondary efficacy outcome measures were the cancer incidence rate, using the standardized incidence ratio (SIR), defined as the number of observed neoplasm divided by the number of expected cases and the risk of second neoplasms, using the relative risk (RR), defined as the incidence of second neoplasms among exposed to GH divided by the incidence among non-exposed. None of the studies, except one [14], provided raw data on single participants, therefore in the analysis we considered the average values for SMR, SIR and RR and their standard errors.
For primary analysis we calculated the effect size for each study. The effect size was computed as the SMR, SIR, and RR for each trial. We present these scores in a paired analysis with their 95% confidence intervals. Then we calculated the global effect size, pooling the data. When the effect size was significant in a fixed model we repeated the analyses using a random effects model [15]. The random effects model incorporates statistical heterogeneity (results, methods, and publication bias) and provides a more conservative estimate of the pooled effect size than a fixed model. We calculated I2 values for quantifying heterogeneity in the meta-analysis. I2 describes the percentage of variability in point estimates that is due to heterogeneity rather than to sampling error. Although no universal rule covers the definitions of mild, moderate, or severe heterogeneity, I2 values more than 50% indicate notable heterogeneity, whereas values less than 30% indicate mild heterogeneity. We assessed publication bias by funnel plot analysis (see web extras). Analyses were carried out using Review Manager 5 software for Windows package (Nordic Cochrane Centre, Copenhagen, Denmark) and double checked using STATA 12.0 statistical software (StataCorp, TX, USA). 3. Results 3.1. Mortality studies The search strategy identified 12 long-term studies concerning patient mortality. Eight studies were excluded; four reported observed deaths only [16–19] and four used indices different than SMR[20–23]. Four [13,24–26] out of the 12 studies used SMR rates to evaluate mortality (Fig. 1). These studies included overall 24,456 patients, with a mean chronological age at study enrolment of 32.6 ± 10.5 years. Mean GH dose in two studies [24,25] was 0.415 ± 0.28 mg/day, while only one study [26] reported a mean dose of 0.024 mg/kg/day. Duration of treatment was reported in only two studies [24,26], with an average of 4.8 ± 4.5 years (Table 1). 3.1.1. All-cause SMR Van Bunderen et al. [24] reported a significant increase of all-cause SMR in patients enrolled in the Dutch National Registry of Growth Hormone Treatment. The patients were retrospectively monitored and subdivided into three groups: a treatment group (n = 2229), a primary control group (who had not commenced treatment or who had discontinued it before 30 days, n = 109) and a secondary control group (who had endured treatment for more than 30 days but less than 90, n = 356). This cohort included both adult and childhood onset GH deficiency (about 80 vs 20%, respectively). The all-cause SMR was 1.27 (95% CI 1.04–1.56) for the treatment group, 1.42 (95% CI 0.79–2.56) for the primary control group and 1.17 (95% CI 0.82– 1.69) for the secondary control group. Gaillard et al. [25] reported data obtained from the analysis of KIMS (Pfizer International Metabolic Database) including 13,983 GH-deficient patients with 69,056 patient-years of follow-up. The all-cause SMR was 1.13 (95% CI 1.04–1.24). The cohort mainly consisted of patients with adult onset GH deficiency, about 20% showing childhood onset GH
Table 1 Mortality studies with SMR analysis. Study
Journal
Year
No. of patients
Age at study enrollment (years)
Dose of treatment
Duration of treatment
All-cause SMR
Cancer SMR
CVD SMR
Carel et al. [26] Gaillard et al. [25] Van Bunderen et al.[24] Swerdlow et al. [13]
JCEM EJE JCEM LANCET
2012 2012 2011 2002
6558 13,983 2229 1352
28.3 ± 5.3 26.9 ± 9.9 42.6 ± 16.3 –
0.024 μg/kg/day 0.42 ± 0.27 mg/day 0.41 ± 0.28 mg/day –
3.9 ± 2.6 – 5.7 ± 6.3 –
1.33 (1.08–1.649) 1.13 (1.04–1.24) 1.27 (1.04–1.56) –
1.02 (0.41–2.09) 0.88 (0.74–1.03) 0.86 (0.6–1.25) 2.3 (0.8–5)
3.07 (1.4–5.83) 0.83 (0.63–1.08) 1.35 (0.95–1.94) –
Please cite this article as: A. Deodati, et al., Association between growth hormone therapy and mortality, cancer and cardiovascular risk: Systematic review and meta-analysis, Growth Horm. IGF Res. (2014), http://dx.doi.org/10.1016/j.ghir.2014.02.001
A. Deodati et al. / Growth Hormone & IGF Research xxx (2014) xxx–xxx
3
Fig. 2. The all-cause SMR in GH treated subjects. Results of meta-analysis according to random model.
deficiency. Interestingly, these authors reported a significant negative association between mortality risk and IGF-I levels during therapy, the less IGF-I the higher the SMR. Carel et al. [26] analyzed patients enrolled in the French populationbased register, as part of a European project aimed at evaluating safety and appropriateness of GH treatments in Europe (SAGhE). The initial cohort (n = 6,892) included patients with childhood onset idiopathic isolated GH deficiency (n = 5162), neurosecretory dysfunction (n = 534), idiopathic short stature (n = 871), or born short for gestational age (n = 335). Follow-up data were available for final analysis in 94.7% of the cohort (n = 6558). The reported all cause SMR was 1.33 (95% CI 1.08–1.64). The overall all-cause SMR in GH treated subjects, mostly children, included in these three studies was significantly increased: 1.19 (95% CI 1.08–1.32, p b 0.001) (Fig. 2). See web extra Supplemental Data 1 and 2 for the results in the analysis carried out according to a fixed model and funnel plot (Supplemental Data 1 and 2).
3.1.2. Malignancy SMR Four studies reported SMRs for deaths due to primary neoplasms in subjects treated with GH[13,24–26]. Swerdlow et al. [13]analyzed a cohort of patients (n = 1848) in the UK who were treated during childhood and early adulthood with human pituitary growth hormone from 1959 to 1985, compared to
that of the general population, controlling for age, sex and calendar period. They found a malignancy SMR of 2.3 (95% CI 0.8–5) after excluding patients with conditions predisposing to cancer (496 subjects). In the cohort of patients reported by Van Bunderen et al. [24] the SMR was 0.86 (95% CI 0.6–1.25) in the treated group, 0.39 (95% CI 0.05–2.75) in the primary control group and 1.14 (95% CI 0.63–2.05) in the secondary control group. In the French cohort [26], malignancy SMR for treated patients was 1.02 (95% CI 0.41–2.09). Finally, in the report from Gaillard et al. [25] the malignancy SMR was 0.88 (95% CI 0.74–1.03). The overall mean malignancy SMR for treated patients was not significantly increased: 0.94 (95% CI 0.74–1.19, p = 0.61) (Fig. 3, and Supplemental Data 3 and 4).
3.1.3. Cardiovascular (CVD) SMR Mortality due to cardiovascular events was analyzed in three studies [24–26]. In the Dutch study [24] the calculated CVD SMR was 1.35 (95% CI 0.95–1.94) for the treated group (n = 2229), 1.45 (95% CI 0.54–3.87) for the primary control group (n = 109), and 0.12 (95% CI 0.02–0.87) for the secondary control group (n = 356). In the French cohort (n = 6,558) [26], the CVD SMR resulted significantly increased: SMR = 3.07 (95% CI 1.40–5.83). In this study, an increase in mortality due subarachnoid or intracerebral hemorrhage (SMR 6.66, 95% CI 1.79–17.05) was also observed. Finally, data from KIMS (n = 13,983) [25] reported a CVD SMR of 0.83 (95% CI 0.63–
Fig. 3. The overall malignancy SMR in GH treated subjects. Results of meta-analysis according to random model.
Fig. 4. The overall cardiovascular SMR in GH treated subjects. Results of meta-analysis according to random model.
Please cite this article as: A. Deodati, et al., Association between growth hormone therapy and mortality, cancer and cardiovascular risk: Systematic review and meta-analysis, Growth Horm. IGF Res. (2014), http://dx.doi.org/10.1016/j.ghir.2014.02.001
4
A. Deodati et al. / Growth Hormone & IGF Research xxx (2014) xxx–xxx
(GHD) and revealed a significantly increased risk of malignancy: SIR = 2.74 (95% CI 1.18–5.41). Wilton et al. [14] reported data from a large cohort of patients (n = 58,603) treated with GH and recruited in KIGS—the Pfizer International Growth Database—without cancer or any other condition in medical history known to increase the risk of cancer. The reported malignancy SIR was 1.26 (95% CI 0.86–1.78). The above cited Swerdlow's study [13] found a malignancy SIR of 1.4 (95% CI 0.5–2.8) after excluding high risk groups. Bell et al. [12] reported data from the National Cooperative Growth Study (NCGS), including 54,996 children enrolled between 1985 and 2006 in USA. This study showed a malignancy SIR of 1.12 (95% CI 0.75–1.61). Taking all the reports together, the overall mean malignancy SIR was 1.36 (95% CI 1.00–1.85, p = 0.05), showing a slightly significantly increased risk of primary neoplasm. (Fig. 6 and Supplemental Data 7 and 8). Fig. 5. Search strategy for selection of morbidity studies for primary cancer.
3.3. Risk of second neoplasms 1.08). The mean cardiovascular SMR derived from all these studies was not significantly increased, SMR = 1.39 (95% CI 0.76–2.55, p = 0.28), (Fig. 4, and Supplemental Data 5 and 6).
3.2. Incidence studies The search strategy identified 7 long-term studies concerning primary cancer incidence. Three studies were excluded: two reported only the observed neoplasms [27] and [28] and one used rates different than SIR[29]. Four studies used SIR to evaluate the incidence of primary cancers [12–14,30] (Fig. 5). These studies included 121,791 patients, with a mean dose of treatment of 0.28 mg/day and a mean duration of treatment of 4.3 years (Table 2).
3.2.1. Malignancy SIR Child et al. [30] reported data from the Hypopituitary Control and Complications Study (Hypoccs) pharmaco-epidemiological database. They evaluated the incidence of de novo and second cancers in 6840 GH-treated and 940 non GH-treated adult patients, 20% with childhood onset GH deficiency. The reported SIR was 0.88 (95% CI 0.74–1.04) for the whole cohort treated with GH. However, the analysis focused only on patients from the USA with childhood onset (CO) GH deficiency
The search strategy identified 17 long-term studies concerning second cancer incidence. 12 studies were excluded; seven [16,31–36] reported only the absolute number of observed cases; 2 studies [37, 38] were excluded because they reported tumor recurrence for brain cancer in subjects with primary diagnosis of CNS tumor and three were excluded for reporting cancer incidence using the hazard ratio (HR) [39–41]. Five studies used relative risk (RR) to evaluate the incidence of secondary tumors [42–46] (Fig. 7). These studies included 908 patients, with a mean duration of treatment of 5 years (0.1–14). Mean age at primary cancer diagnosis of 13.06 years (0–45). The mean GH dose reported in only two studies was 0.26 mg/kg/week [44] (Table 3).
3.3.1. RR second neoplasms Ergun-Longmire et al. [42] reported data from a cohort of 361 patients enrolled in the Childhood Cancer Survivor Study. The RR was 2.15 (95% CI 1.33–3.47) compared with non-GH-treated survivors. Meningiomas were the most common second neoplasm among the GH-treated group. Mackenzie et al. [43] included in the analysis previously irradiated patients (n = 110), the patient group comprised a mixture of children and adult patients. Patients had received GH therapy for a median of 8 years (range 1–19) and were followed up for a median
Table 2 Cancer morbidity studies with SIR analysis.
Child et al. [30] Wilton et al. [14] Swerdlow et al. [13] Bell et al. [12]
Journal
Year
No. of patients
Age at study enrollment (years)
Dose of treatment
Duration of treatment (years)
Cancer SIR
EJE JP LANCET JCEM
2011 2010 2002 2010
6840 58,603 1352 54,996
46.4 (34.1–56.3) 10.3 ± 4 – 10.3
0.30 (0.20–0.46) mg/day 0.25 ± 0.1 mg/kg/week – –
– 3.6 (0.08–9.70) – 3.6
2.74 (1.18–5.41) 1.26 (0.86–1.78) 1.4 (0.5–2.8) 1.12 (0.75–1.61)
Fig. 6. The overall malignancy SIR in GH treated subjects. Results of meta-analysis according to random model.
Please cite this article as: A. Deodati, et al., Association between growth hormone therapy and mortality, cancer and cardiovascular risk: Systematic review and meta-analysis, Growth Horm. IGF Res. (2014), http://dx.doi.org/10.1016/j.ghir.2014.02.001
A. Deodati et al. / Growth Hormone & IGF Research xxx (2014) xxx–xxx
5
tumor in cancer survivors who underwent GH therapy during childhood (Fig. 9 and Supplemental Data 11 and 12). 4. Discussion
Fig. 7. Search strategy for selection of morbidity studies for second neoplasms.
period of 14.5 years. RR for second neoplasm was 2.50 (95% CI 0.495–12.61). Woodmansee et al. [44] assessed the incidence of second neoplasm in a cohort of childhood cancer survivors enrolled in the Eli Lilly and Company's pediatric (Genetics and Neuroendocrinology of Short Stature International Study (GeNeSIS)) and adult (Hypopituitary Control and Complications Study (HypoCCS)) observational studies (n = 394). The RR for second neoplasm was 0.83 (95% CI 0.201–3.457). In the pediatric cohort the incidence of second neoplasms in untreated patients was 0/27 whereas the incidence in the GH treated group was 15/394. In the calculation of RR we considered an incidence of 1/27 in the untreated cohort in order to perform the analysis. The RR for second neoplasms in the pediatric cohort was 1.3 (95% CI 0.14–7.49). Wing-Leung et al. [45] reported the risk of second neoplasms in cancer survivors with GH deficiency secondary to ALL treatment (n = 43). The RR was 1.58 (95% CI 0.38–6.50). The study of Sklar et al. [46] reported the risk of second neoplasms in different subgroups of cancer survivors according to different diagnoses: CNS tumors, leukemia and rhabdomyosarcoma (n = 361). The RR of second neoplasm in patients treated with rhGH after the initial diagnosis of CNS tumors was 2.34 (95% CI 0.96–5.70). The RR in patients with primary diagnosis of leukemia was 4.98 (95% CI 1.95–12.74) and the RR in subjects with initial diagnosis of rhabdomyosarcoma was 1.82 (95% CI 0.41–8.01). Initially, we assessed the recurrence of second neoplasms including all selected studies [42–46]. The mean RR for second neoplasms was 2.34 (95% CI 1.64–3.23, p b 0.00001) showing a higher risk of second tumor in cancer survivors who underwent GH therapy during childhood (Fig. 8 and Supplemental Data 9 and 10). Thereafter, we have considered only the studies reporting the general relative risk of second tumor [42–44]. The mean RR for second neoplasms was 1.99 (95% CI 1.28–3.08, p = 0.002), confirming a significantly higher risk of second
In this systematic review all the available studies reporting all-cause SMR, cancer SMR, CVD SMR, cancer SIR and second neoplasm RR in patients, mostly children, treated with GH were included. The results show a significant increase of all-cause mortality but no significant increase of the malignancy and CVD mortality. The morbidity assessment (SIR) showed a significant increase of cancer incidence. Finally, the risk for second neoplasms in patients treated with GH after a primary cancer was significantly increased. Although these data may raise concern about the long-term safety of GH treatment a number of potential confounders and biases should be considered. First of all the study populations were heterogeneous, comprising in many reports both adult and pediatric cohorts, but, even more importantly, including patients with different diagnoses. With the only exception of one study [26], all the other studies did not stratify the patients according to diagnostic groups. The potential different risk related to a pharmacological rather than a replacement therapy is intuitive. In Carel's [26] cohort, only patients defined as at low risk, were assessed. However, subjects born small for gestational age (SGA) were included in the analysis, thus potentially biasing the results not only because this diagnostic group may comprise undefined congenital syndromes, but also because there is strong evidence for a higher cardiovascular risk in subjects born SGA. The Swerdlow's cohort [13], including only patients treated with pituitary derived GH before 1985, in theory could be considered the most homogenous since the only approved indication at that time was GH deficiency. However, the detailed description of diagnoses revealed that GH deficiency was of idiopathic origin in 53% of patients and attributable to intracranial neoplasia in 26%. Many other disorders affected the remaining 21% of patients, including cases of leukemia or lymphoma, chromosome fragility syndromes, and glycogen storage diseases. Furthermore, all these studies were also affected by a relatively limited sample size and, especially, a low event rate. They also lack validated external reference in that no untreated “control” group was available for comparison and no reference mortality data corrected for the potential effect of height were available. The collected data were obtained from multiple centers throughout a country or even the world thus including many unidentified confounding factors and missing data. A further potential confounder is the lack of key data such as familial predisposition to certain diseases, exposure to environmental hazards, lack of local mortality and morbidity indices. Moreover, in most studies, the mean dose and duration of GH treatment as well as the follow-up length were not reported. Further limitations are inherent in the meta-analysis approach. One potential limitation of any meta-analysis is the pooling of studies from heterogeneous populations. Another potential bias is the inclusion of study populations mostly derived from sponsored post-marketing surveillance programs. A third potential limitation of any meta-analysis is the “file-drawer” effect, in which studies with negative or “unexpected”
Table 3 Studies reporting the relative risk (RR) of second neoplasm. Sklar's [46] and Ergun-Longmire's [42] studies report data from the same cohort. Study
Journal
Sklar et al. [46] JCEM Ergun-Longmire et al. [42] JCEM Mackenzie et al. [43] JCEM Wing Leung et al. [45] JCO Woodmansee et al. [44] EJE
Year
No. of patients
Age at primary diagnosis (yrs)
Dose of treatment
Duration of treatment
RR second neoplasm
RR second neoplasm after leukemia
RR second neoplasm after CSN tumor
2002 2006 2011 2002 2013
361 361 110 43 394
3.5 (0–17.2) 3.5 (0–17.2) 33 (14–45) 10.35 ± 1.7 5.4 (3–8.5)
– – – 0.3 mg/kg/week 0.22 ± 0.03 mg/kg/week
4.6 (0.1–14) 4.6 (0.1–14) 8 (4–10) 4.5 (1–8) 2.9 (1.4–4.8)
– 2.15 (1.33–3.47) 2.5 (0.495–12.61) – 1.3 (0.14–7.49)
4.98 (1.95–12.74) – – 1.58 (0.376–6.53) –
2.34 (0.96–5.7) – – – –
Please cite this article as: A. Deodati, et al., Association between growth hormone therapy and mortality, cancer and cardiovascular risk: Systematic review and meta-analysis, Growth Horm. IGF Res. (2014), http://dx.doi.org/10.1016/j.ghir.2014.02.001
6
A. Deodati et al. / Growth Hormone & IGF Research xxx (2014) xxx–xxx
Fig. 8. The overall RR of second neoplasm including GH treated subjects with specific primary diagnosis. Results of meta-analysis according to random model.
Fig. 9. The overall RR of second neoplasm in GH treated subjects. Results of meta-analysis according to random model.
results might remain unpublished thus biasing the literature towards positive findings. Ideally, an appropriately designed study should be characterized by an adequate sample size, detailed diagnostic characterization and stratification according to the potential risks, accurate documentation of GH treatment regimens, long-term follow-up and the availability of a control cohort. The realization of such ambitious study implies the involvement of multiple scientific societies as well as regulatory authorities and pharmaceutical industry. In conclusion, although the meta-analysis of available data has provided evidence for a statistically significant increase of all-cause mortality, primary cancers and second neoplasms in patients treated with GH, these results may be affected by several biases and confounders which need to be addressed in further large-scale controlled surveys. Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.ghir.2014.02.001. References [1] D. Le Roith, C. Bondy, S. Yakar, J.L. Liu, A. Butler, The somatomedin hypothesis: 2001, Endocr. Rev. 22 (2001) 53–74. [2] P.E. Clayton, I. Banerjee, P.G. Murray, A.G. Renehan, Growth hormone, the insulinlike growth factor axis, insulin and cancer risk, Nat. Rev. Endocrinol. 7 (2011) 11–24. [3] J.M. Chan, M.J. Stampfer, E. Giovannucci, P.H. Gann, J. Ma, P. Wilkinson, C.H. Hennekens, M. Pollak, Plasma insulin-like growth factor-I and prostate cancer risk: a prospective study, Science 279 (1998) 563–566. [4] S.E. Hankinson, W.C. Willett, G.A. Colditz, D.J. Hunter, D.S. Michaud, B. Deroo, B. Rosner, F.E. Speizer, M. Pollak, Circulating concentrations of insulin-like growth factor-I and risk of breast cancer, Lancet 351 (1998) 1393–1396. [5] J. Ma, M.N. Pollak, E. Giovannucci, J.M. Chan, Y. Tao, C.H. Hennekens, M.J. Stampfer, Prospective study of colorectal cancer risk in men and plasma levels of insulin-like growth factor (IGF)-I and IGF-bindingprotein-3, J. Natl. Cancer Inst. 91 (1999) 620–625. [6] D. Baris, G. Gridley, E. Ron, E. Weiderpass, L. Mellemkjaer, A. Ekbom, J.H. Olsen, J.A. Baron, J.F. Fraumeni Jr., Acromegaly and cancer risk: a cohort study in Sweden and Denmark, Cancer Causes Control 13 (2002) 395–400. [7] R. Kauppinen-Mäkelin, T. Sane, M.J. Välimäki, H. Markkanen, L. Niskanen, T. Ebeling, P. Jaatinen, M. Juonala, Finnish Acromegaly Study Group, E. Pukkala, Increased cancer incidence in acromegaly—a nationwide survey, Clin. Endocrinol. 72 (2009) 278–279. [8] S.M. Orme, R.J. McNally, R.A. Cartwright, P.E. Belchetz, Mortality and cancer incidence in acromegaly: a retrospective cohort study. United Kingdom Acromegaly Study Group, J. Clin. Endocrinol. Metab. 83 (1998) 2730–2734.
[9] E. Ron, G. Gridley, Z. Hrubec, W. Page, S. Arora, J.F. Fraumeni Jr., Acromegaly and gastrointestinal cancer, Cancer 68 (1991) 1673–1677. [10] J. Kirk, Indications for growth hormone therapy in children, Arch. Dis. Child. 97 (2012) 63–68. [11] S. Watanabe, S. Mizuno, L.H. Oshima, Y. Tsunematsu, J. Fujimoto, A. Komiyama, Leukemia and other malignancies among GH users, J. Pediatr. Endocrinol. 6 (1993) 99–108. [12] J. Bell, K.L. Parker, R.D. Swinford, A.R. Hoffman, T. Maneatis, B. Lippe, Long-term safety of recombinant human growth hormone in children, J. Clin. Endocrinol. Metab. 95 (2010) 167–177. [13] A.J. Swerdlow, C.D. Higgins, P. Adlard, M.A. Preece, Risk of cancer in patients treated with human pituitary growth hormone in the UK, 1959–85: a cohort study, Lancet 360 (2002) 273–277. [14] P. Wilton, A.F. Mattsson, F. Darendeliler, Growth hormone treatment in children is not associated with an increase in the incidence of cancer: experience from KIGS (Pfizer International Growth Database), J. Pediatr. 157 (2010) 265–270. [15] L.V. Hedges, T.D. Pigott, The power of statistical tests in meta-analysis, Psychol. Methods 6 (2001) 203–221. [16] S.L. Blethen, D.B. Allen, D. Graves, G. August, T. Moshang, R. Rosenfeld, Safety of recombinant deoxyribonuclei acid-derived growth hormone: The National Cooperative Growth Study Experience, J. Clin. Endocrinol. Metab. 81 (1996) 1704–1710. [17] C.A. Quigley, A.M. Gill, B.J. Crowe, K. Robling, J.J. Chipman, S.R. Rose, J.L. Ross, F.G. Cassorla, A.M. Wolka, J.M. Wit, L.T.M. Rekers-Mombarg, G.B. Cutler Jr., Safety of growth hormone treatment in pediatric patients with idiopathic short stature, J. Clin. Endocrinol. Metab. 90 (2005) 5188–5196. [18] M.L. Hartmann, R. Xu, B.J. Crowe, L.L. Robinson, E.M. Erfurth, D.L. Kleinberg, A.G. Zimmerman, W.W. Woodmansee, G.B. Cutler Jr., J.J. Chipman, S. Melmed, Prospective safety surveillance of GH-deficient adults: comparison of GH-treated vs untreated patients, J. Clin. Endocrinol. Metab. 98 (2013). [19] L. Savendahl, M. Maes, K. Albertsson-Wilkand, B. Borgstrom, J.C. Carel, S. Henrard, N. Speybroeck, M. Thomas, G. Zandwijken, A. Hokken-Koelega, Long-term mortality and causes of death in isolated GHD, ISS, and SGA patients treated with recombinant growth hormone during childhood in Belgium, The Netherlands, and Sweden: preliminary report of 3 countries participating in the EU SAGhE study, J. Clin. Endocrinol. Metab. 97 (2010) E213–E217. [20] K. Stochholm, C.H. Gravholt, T. Laursen, P. Laurberg, M. Andersen, L.O. Kristensen, U. Feldt-Rasmussen, J.S. Christiansen, M. Frydenberg, A. Green, Mortality and GHdeficiency: a nationwide study, Eur. J. Endocrinol. 157 (2007) 9–18. [21] S.P. Taback, H.J. Dean, Members of the Canadian Growth Hormone Advisory Committee. Mortality in Canadian children with growth hormone (GH) deficiency receiving GH therapy 1967–1992, J. Clin. Endocrinol. Metab. 81 (1996) 1693–1696. [22] K. Stochholm, S. Juul, J.S. Christiansen, C.H. Gravholt, Mortality and socioeconomic status in adults with childhood onset GH deficiency (GHD) is highly dependent on the primary cause of GHD, Eur. J. Endocrinol. 167 (2012) 663–670. [23] J. Svensson, B.A. Bengtsson, T. Rosen, A. Odén, G. Johannsson, Malignant disease and cardiovascular morbidity in hypopituitary adults with or without growth hormone replacement therapy, J. Clin. Endocrinol. Metab. 89 (2004) 3306–3312.
Please cite this article as: A. Deodati, et al., Association between growth hormone therapy and mortality, cancer and cardiovascular risk: Systematic review and meta-analysis, Growth Horm. IGF Res. (2014), http://dx.doi.org/10.1016/j.ghir.2014.02.001
A. Deodati et al. / Growth Hormone & IGF Research xxx (2014) xxx–xxx [24] C.C. Van Bunderen, C. van Nieuwpoort, L.I. Arwert, M.W. Heymans, A.A.M. Franken, H.P.F. Koppeschaar, A.J. van der Lely, M.L. Drent, Does growth hormone replacement therapy reduce mortality in adults with growth hormone deficiency? Data from the Dutch National Registry of Growth Hormone Treatment in Adults, J. Clin. Endocrinol. Metab. 96 (2011) 3151–3159. [25] R.C. Gaillard, A.F. Mattsson, A.C. Akerblad, B.A. Bengtsson, J. Cara, U. FeldtRasmussen, M. Koltowska-Haggstrom, J.P. Monson, B. Saller, P. Wilton, R. Abs, Overall and cause-specific mortality in GH-deficient adults on GH replacement, Eur. J. Endocrinol. 166 (2012) 1069–1077. [26] J.C. Carel, E. Ecosse, F. Landier, D. Meguellati-Hakkas, F. Kaguelidou, G. Rey, J. Coste, Long-term mortality after recombinant growth hormone treatment for isolated growth hormone deficiency or childhood short stature: preliminary report of the French SAGhE study, J. Clin. Endocrinol. Metab. 97 (2012) 416–425. [27] D. Wyatt, Lessons from the national cooperative growth study, Eur. J. Endocrinol. 151 (2004) S55–S59. [28] Y. Nishi, T. Tnaka, K. Takano, K. Fujieda, Y. Igarashi, K. Hanew, T. Hirano, S. Yokoya, K. Tachibana, K. Saito, S. Watanabe, Recent status in the occurrence of leukemia in growth hormone-treated patients in Japan, J. Clin. Endocrinol. Metab. 84 (1999) 1961–1965. [29] V. Popovic, A.F. Mattson, R.C. Gaillard, P. Wilton, M. Koltowska-Häggström, M.B. Ranke, Serum insulin-like growth factor I (IGF-1), IGF-binding proteins 2 and 3, and the risk for development of malignancies in adults with growth hormone (GH) deficiency treated with GH: data from KIMS (Pfizer International Metabolic Database), J. Clin. Endocrinol. Metab. 95 (2010) 4449–4454. [30] C.J. Child, A.G. Zimmermann, W.W. Woodmansee, D.M. Green, J.J. Li, H. Jung, E.M. Erfurth, L.L. Robinson, Assessment of primary cancers in GH-treated adult hypopituitary patients: an analysis from the Hypopituitary Control and Complications Study, Eur. J. Endocrinol. 165 (2011) 217–223. [31] D.S. Olsson, M. Buchfelder, S. Schlaffer, B.A. Bengtsson, K.E. Jakobsson, G. Johannsson, A.G. Nilsson, Comparing progression of non-functioning pituitary adenomas in hypopituitarism patients with and without long-termGH replacement therapy, Eur. J. Endocrinol. 161 (2009) 663–669. [32] M. Buchefelder, P.H. Kann, C. Wuster, U. Tuschy, B. Saller, G. Brabant, A. Kleindienst, P. Nomikos, the German KIMS Board, Influence of GH substitution therapy in deficient adults on the recurrence rate of hormonally inactive pituitary adenomas: a case–control study, Eur. J. Endocrinol. 157 (2007) 149–156. [33] T.T. Chung, J. Evanson, D. Walker, S.A. Akker, G.M. Besser, J.P. Monson, W.M. Drake, Safety of GH replacement in hypopituitary patients with nonirradiated pituitary and peripituitary tumours, Clin. Endocrinol. 68 (2008) 965–969. [34] T.T. Chung, W.M. Drake, J. Evanson, D. Walker, P.N. Plowman, S.L. Chew, A.B. Grossman, G.M. Besser, J.P. Monson, Tumour surveillance imaging in patients with extrapituitary tumours receiving growth hormone replacement, Clin. Endocrinol. 63 (2005) 272–279.
7
[35] L.R. Meacham, K. Sullivan, Characteristics of growth hormone therapy for pediatric patients with brain tumours in the National Cooperative Growth Study (NCGS) and from a survey of pediatric endocrinologists, J. Pediatr. Endocrinol. Metab. 15 (2002) 689–696. [36] S.J. Howell, P. Wilton, A. Lindberg, S.M. Shalet, Growth hormone replacement and the risk of malignancy in children with neurofibromatosis, J. Pediatr. 133 (1998) 201–205. [37] A.L. Ogyvil-Stuart, W.D.J. Ryder, H.R. Gattamanemi, P.E. Clayton, S.M. Shalet, Growth hormone and tumour recurrence, BMJ 304 (1992) 1601–1605. [38] A.J. Swerdlow, R.E. Reddingius, C.D. Higgins, H.A. Spoudeas, K. Phipps, Z. Qiao, W.D.J. Ryder, M. Brada, R.D. Hayward, C.G.D. Brook, P.C. Hindmarsh, S.M. Shalet, Growth hormone treatment of children with brain tumors and risk of tumor recurrence, J. Clin. Endocrinol. Metab. 85 (2000) 4444–4449. [39] N. Karavitaki, J.T. Warner, A. Marland, B. Shine, F. Ryan, J. Arnold, Helen E. Turner, J.A. H. Wass, GH replacement does not increase the risk of recurrence in patients with craniopharyngioma, Clin. Endocrinol. 64 (2006) 556–560. [40] J.R. Arnold, F. Arnold, A. Marland, N. Karavitaki, J.A.H. Wass, GH replacement in patients with non-functioning pituitary adenoma (NFA) treated solely by surgery is not associated with increased risk of tumour recurrence, Clin. Endocrinol. 70 (2009) 435–438. [41] D.S. Olsson, M. Buchfelder, K. Wiendieck, N. Kremenevskaja, B.A. Bengtsson, K.E. Jakobsson, M. Jarfelt, G. Johannsson, A.G. Nilsson, Tumour recurrence and enlargement in patients with craniopharyngioma with and without GH replacement therapy during more than 10 years of follow-up, Eur. J. Endocrinol. 166 (2012) 1061–1068. [42] B. Ergun-Longmire, A.C. Mertens, P. Mitby, J. Qin, G. Heller, W. Shi, Y. Yasui, L.L. Robinson, C.A. Sklar, Growth hormone treatment and risk of second neoplasms in the childhood cancer survivor, J. Clin. Endocrinol. Metab. 91 (2006) 3494–3498. [43] S. Mackenzie, T. Craven, H.R. Gattamaneni, R. Swindell, S.M. Shalet, G. Brabant, Long-term safety of growth hormone replacement after CNS irradiation, J. Clin. Endocrinol. 96 (2011) 2756–2761. [44] W.W. Woodmansee, A.G. Zimmermann, C.J. Child, Q. Rong, E.M. Erfurth, P. BeckPeccoz, W.F. Blum, L.L. Robinson, on behalf of the GeNeSIS and HypoCCS International Advisory Boards, Incidence of second neoplasm in childhood cancer survivors treated with GH: an analysis of GeNeSIS and HypoCCS, Eur. J. Endocrinol. 168 (2013) 565–573. [45] W. Leung, S.R. Rose, Y. Zhou, M.L. Hancock, S. Burstein, E.A. Schriock, R. Lustig, R.K. Danish, W.E. Evans, M.M. Hudson, C.H. Pui, Outcomes of growth hormone replacement therapy in survivors of childhood acute lymphoblastic leukemia, J. Clin. Oncol. 20 (2002) 2959–2964. [46] C.A. Sklar, A.C. Mertens, P. Mitby, G. Occhiogrosso, J. Qin, G. Heller, Y. Yasui, L.L. Robison, Risk of Disease recurrence and second neoplasms in survivors of childhood cancer treated with growth hormone: a report from the childhood center survivors study, J. Clin. Endocrinol. Metab. 87 (2002) 3136–3141
Please cite this article as: A. Deodati, et al., Association between growth hormone therapy and mortality, cancer and cardiovascular risk: Systematic review and meta-analysis, Growth Horm. IGF Res. (2014), http://dx.doi.org/10.1016/j.ghir.2014.02.001