The Human Experience With Ghrelin Administration

Abstract

Context:

Ghrelin is an endogenous stimulator of GH and is implicated in a number of physiological processes. Clinical trials have been performed in a variety of patient populations, but there is no comprehensive review of the beneficial and adverse consequences of ghrelin administration to humans.

Evidence Acquisition:

PubMed was utilized, and the reference list of each article was screened. We included 121 published articles in which ghrelin was administered to humans.

Evidence Synthesis:

Ghrelin has been administered as an infusion or a bolus in a variety of doses to 1850 study participants, including healthy participants and patients with obesity, prior gastrectomy, cancer, pituitary disease, diabetes mellitus, eating disorders, and other conditions. There is strong evidence that ghrelin stimulates appetite and increases circulating GH, ACTH, cortisol, prolactin, and glucose across varied patient populations. There is a paucity of evidence regarding the effects of ghrelin on LH, FSH, TSH, insulin, lipolysis, body composition, cardiac function, pulmonary function, the vasculature, and sleep. Adverse effects occurred in 20% of participants, with a predominance of flushing and gastric rumbles and a mild degree of severity. The few serious adverse events occurred in patients with advanced illness and were not clearly attributable to ghrelin. Route of administration may affect the pattern of adverse effects.

Conclusions:

Existing literature supports the short-term safety of ghrelin administration and its efficacy as an appetite stimulant in diverse patient populations. There is some evidence to suggest that ghrelin has wider ranging therapeutic effects, although these areas require further investigation.

Ghrelin, a 28-amino acid peptide with n-octanoylation at serine 3, was first identified in 1999 when isolated in rat stomach as an endogenous ligand specific for the GH secretagogue receptor (GHS-R) (1). GHS-Rs are present in the hypothalamus, heart, lung, pancreas, intestine, and adipose tissue (1). Ghrelin binds to the GHS-R and activates the phospholipase C signaling pathway, leading to release of intracellular calcium (2). The ghrelin octanoyl group is necessary for action at the GHS-R to release GH from the pituitary (1). The highest concentration of ghrelin is found in the X/A-like cells of the oxyntic glands in the gastric fundus (3); other organs with high concentrations of ghrelin mRNA include small intestine, lung, pancreas, colon, pituitary, breast, kidney, and ovary (4). It has been demonstrated in mice that ghrelin is able to cross the blood–brain barrier (5). Ghrelin has been shown in humans to be involved in prolactin and ACTH release (6, 7), hunger response (8), and glucose metabolism (9). Although acylated ghrelin is the form responsible for GH release, des-octanoyl (unacylated) ghrelin is found in higher serum concentrations and may have distinct metabolic effects (10).

Exogenous ghrelin administration was first noted to promote food intake and weight gain in rats (11, 12), and serum ghrelin concentrations were suppressed by refeeding (12). These effects are independent of ghrelin's effects on GH stimulation and the GHS-R, and may be related to neuropeptide Y (NPY) release (12). In humans, endogenous ghrelin levels peak just before eating and quickly fall to trough levels after eating a meal (13).

Since 2000, dozens of studies have been published in which ghrelin was administered to human subjects. In distinction from prior published review articles, we sought to review the complete human experience with exogenous ghrelin administration, with the main goal of presenting information on both the beneficial and adverse consequences of ghrelin administration to humans.

Materials and Methods

We searched the MEDLINE electronic database through November 1, 2012, using the PubMed search engine to find clinical studies in which ghrelin was administered to humans. Search terms included “ghrelin” and were limited to the English language. This initial search strategy yielded 3934 papers. Those publications considered potentially eligible based on the title and/or abstract were read fully, and of those, 121 were ultimately identified for inclusion. Reference lists of all included studies were manually searched for other potentially eligible studies. All identified studies in which ghrelin was administered to humans were included. Information was extracted from each study in the following categories: amount of ghrelin administered, manner of administration (bolus or continuous infusion), size and characteristics of the study population, study design, physiological outcomes, and documentation of adverse events. Studies were excluded from adverse event analysis if there was no comment on adverse events in the article (n = 51) or if adverse events were mentioned but the number of participants experiencing specific symptoms was not documented (n = 4).

Results

We found 48 published studies in which ghrelin was administered as an infusion to a total of 681 participants and 73 published studies in which ghrelin was administered as a bolus to a total of 1169 participants (as detailed in Supplemental Tables 1 and 2, published on The Endocrine Society's Journals Online web site at http://jcem.endojournals.org). The average study size was small, with 15 participants, and ranged from 4 to 117. Populations studied included healthy volunteers (both lean and obese), and several patient populations including those with congestive heart failure, several types of cancer, diabetes mellitus, pulmonary disease, anorexia nervosa, end-stage renal disease, Cushing's syndrome, gastroparesis, polycystic ovary syndrome (PCOS), hyperthyroidism, hyperparathyroidism, depression, acromegaly, and GH deficiency (Table 1). Participant ages ranged from 18 to 80 years, with the exception of 1 study that included children (14).

Table 1.

Summary of Study Characteristics

Study Population	Infusion	Bolus	Total
Healthy	19	42	61
Obese	5	7	12
Gastrectomy	7	2	9
Cancer	4	2	6
Hypopituitarism	1	5	6
Metabolic syndrome/diabetes	3	1	4
Anorexia nervosa/bulimia nervosa	2	2	4
Pulmonary disease	3	0	3
Gastroparesis	2	1	3
Hyperthyroid	0	3	3
Renal disease	0	2	2
Heart failure	2	0	2
Cushing's syndrome	0	2	2
PCOS	0	2	2
Functional dyspepsia	1	0	1
Acromegaly	0	1	1
Hyperparathyroid	0	1	1
Osteoarthritis	0	1	1
Major depression	0	1	1
Totala	49	75	124

^aTotal may include some studies more than once if more than 1 nonhealthy patient population was included in the study.

The range of iv infusion doses of acylated ghrelin administered varied from 0.003 to 1.33 μg/kg·min, with the most frequent dose being 0.017 μg/kg·min (a dose used in 14 studies). Infusion time ranged from 5 to 1440 minutes, although in most studies infusion time ranged between 30 and 300 minutes. Most studies administered single-dose infusions, although several studies infused ghrelin multiple times daily for several weeks (15–23). Three studies used intra-arterial infusions of acylated ghrelin (24–26). One study administered iv infusions of unacylated ghrelin in addition to acylated ghrelin (27), and 1 study administered only unacylated ghrelin (28).

The range of iv bolus doses of acylated ghrelin administered varied from 0.03 to 10.0 μg/kg, with the most frequent bolus dose being 1.0 μg/kg, used in 43 studies. Some studies administered a single bolus of ghrelin; however, many studies administered multiple boluses. Most investigators studied the effects of iv ghrelin; however, 5 studies employed sc ghrelin injections (29–33), and 1 study used an intra-arterial ghrelin bolus (34). Two studies included iv boluses of both acylated and unacylated ghrelin (35, 36). Seventy-two studies (60%) were crossover studies. Thirty-seven of the infusion studies were placebo controlled (77%), and 34 were randomized (71%). Forty-six of the bolus studies were placebo controlled (61%), and 53 were randomized (71%). Only 20 studies that were not crossover studies were randomized and placebo controlled (17%).

Pharmacokinetics

Intravenous ghrelin is rapidly cleared. Akamizu et al (37) found that 15 minutes after administration of a single 1 or 5 μg/kg iv bolus, total plasma ghrelin concentrations rose to 1058.7 and 6598.9 fmol/mL, respectively, from a mean baseline of 168.6 fmol/mL. Elimination half-life was 27 to 31 minutes (37). Nagaya et al (38) found a 61-fold increase in circulating total ghrelin 1 minute after a 10 μg/kg iv ghrelin bolus and a half-life of 10 minutes. Less is known about sc ghrelin. After a single bolus of 12.1 μg/kg sc ghrelin, plasma total and active ghrelin levels peaked at 15 and 30 minutes, respectively (31). Enomoto et al (29) noted that at 30 minutes after a 10 μg/kg dose of sc ghrelin, total ghrelin levels increased 12-fold and returned near pretreatment baseline by 180 minutes.

Appetite/Energy Intake Outcomes

Published evidence in humans confirms initial findings in rats that exogenous ghrelin administration is an effective appetite stimulant. Acylated ghrelin infusion and bolus each have consistently been shown to increase appetite and/or food intake in healthy volunteers (31, 39–41), cancer patients (21, 22, 33, 42), obese participants (40, 43), and patients with pulmonary cachexia (16, 18), heart failure (15), end-stage renal disease on dialysis (30, 32), and gastric functional disorders (17). These endpoints have been assessed in multiple randomized, blinded, placebo-controlled trials and studies extending more than 1 day (Table 2). The increase in energy intake occurred without compensatory undereating (30, 31, 39, 40, 42) or evidence of tachyphylaxis (32). The hunger profile of obese participants became more typical of lean individuals after ghrelin infusion (43). The only patient populations in which a single infusion of ghrelin was not an effective appetite stimulant were postvagotomy or gastrectomy patients (43, 44) and patients with anorexia nervosa (45). However, twice daily repeated injections in patients with anorexia nervosa or immediately after gastrectomy successfully stimulated hunger and food intake (19, 20). Twice daily infusions also stimulated intake in patients with esophageal cancer either postoperatively or during cisplatin therapy (21, 22). Plasma NPY levels rose after iv ghrelin bolus (46, 47), implicating a role for NPY in ghrelin's effect on appetite.

Table 2.

Summary of the Major Outcomes Measured After Ghrelin Administration, Presented as the Number of Participants Studied (Number of Studies), in Rank Order for the 15 Outcomes With the Largest Total Number of Participants Assessed

Major Outcome Measure	Direction of Findinga	Study Design		Dosing Duration		Total
		Higher Qualityb	Lower Qualityc	<24 h	>24 h
GH	↑	422 (30)	574 (39)	898 (64)	98 (5)	996 (69)
	−	63 (2)	17 (2)	12 (1)	68 (3)	80 (4)
Cortisol	↑	202 (17)	229 (14)	431 (31)	0	431 (31)
	−	37 (3)	54 (4)	63 (5)	28 (2)	91 (7)
Glucose	↑	107 (6)	191 (12)	298 (18)	0	298 (18)
	−	39 (4)	33 (5)	53 (7)	19 (2)	72 (9)
Insulin	↑	22 (2)	0	22 (2)	0	22 (2)
	−	106 (6)	84 (9)	103 (10)	87 (5)	190 (15)
	↓	38 (3)	107 (8)	145 (11)	0	145 (11)
ACTH	↑	81 (8)	205 (12)	286 (20)	0	286 (20)
	−	0	52 (4)	47 (3)	5 (1)	52 (4)
Energy intake	↑	198 (11)	37 (4)	83 (6)	152 (9)	235 (15)
	−	64 (4)	6 (1)	35 (3)	35 (2)	70 (5)
Prolactin	↑	57 (6)	189 (14)	246 (20)	0	246 (20)
	−	0	52 (4)	47 (3)	5 (1)	52 (4)
Appetite	↑	161 (9)	28 (3)	86 (6)	103 (6)	189 (12)
	−	51 (4)	18 (1)	49 (4)	20 (1)	69 (5)
Weight/body mass	↑	135 (5)	49 (4)	0	184 (9)	184 (9)
	−	29 (1)	6 (1)	0	35 (2)	35 (2)
IGF-I	−	129 (8)	23 (4)	72 (8)	80 (4)	152 (12)
Leptin	−	83 (5)	30 (3)	54 (5)	59 (3)	113 (8)
Epinephrine	↑	58 (4)	0	58 (4)	0	58 (4)
	−	20 (2)	6 (1)	26 (3)	0	26 (3)
	↓	0	18 (1)	0	18 (1)	18 (1)
Gastric motility	↑	45 (5)	6 (1)	51 (6)	0	51 (6)
	−	47 (2)	0	47 (2)	0	47 (2)
Norepinephrine	−	52 (5)	14 (2)	66 (7)	0	66 (7)
	↓	0	25 (2)	0	25 (2)	25 (2)
FFAs	↑	75 (6)	8 (1)	83 (7)	0	83 (7)
	−	8 (1)	0	8 (1)	0	8 (1)

^a↑, increased, −, no change; ↓, decreased. The absence of a row with a specific direction indicates that there were no published reports of a finding in that direction for that outcome.

^bStudy design included the 3 following characteristics: randomization, blinding (single or double), and placebo control.

^cStudy lacked at least 1 of the 3 following characteristics in their reported study design: randomization, blinding (single or double), or placebo control.

Repeated ghrelin infusions or boluses may result in changes in weight and body composition, as assessed in multiple randomized, blinded, placebo-controlled trials of repeated doses (Table 2). Patients with chronic respiratory infections increase weight with repeated infusions of ghrelin (18). Weight and lean mass increased with repeated ghrelin infusions in patients with heart failure (15). Older participants who had undergone hip replacement increased lean body mass and decreased fat mass after repeated ghrelin boluses (48). Although administered the same dose as participants in other studies, severely ill patients with pulmonary cachexia did not have an increase in weight or lean mass after repeated ghrelin infusions (23).

Gastrointestinal Outcomes

Ghrelin exhibits local effects on gastric motility in rats (49), with similar effects noted in humans. There was no change in gastric emptying in healthy volunteers after low-dose ghrelin infusion or bolus (39, 50), but higher dose ghrelin infusion was effective in stimulating gastric emptying in healthy participants (51). Patients with diabetic, neurogenic, or idiopathic gastroparesis have been shown to be responsive to low doses of both ghrelin infusion and bolus (52–54). This endpoint has been assessed in 7 small, randomized, blinded, placebo-controlled trials but never as a repeated dose (Table 2). Increased gastric emptying may involve induction of a premature activity front starting in the stomach, increased proximal gastric tone, and decreased gastric accommodation in a motilin-independent mechanism (55, 56). Ghrelin infusion does not appear to affect intestinal transit time in healthy men (57). Ghrelin bolus may decrease gastric pH in healthy volunteers without a change in gastrin levels (58).

Cardiopulmonary and Vascular Outcomes

Cardiac effects of ghrelin are dose and route dependent, with greater potency from iv ghrelin compared to the sc route. In healthy men, low-dose ghrelin infusion was associated with increased mean peak myocardial systolic velocity and global longitudinal systolic contraction amplitude of the left ventricle but not change in vascular measurements or left ventricular ejection fraction (LVEF) (59). A large ghrelin iv bolus increased stroke volume and decreased systemic vascular resistance and mean arterial pressure (MAP) in healthy volunteers (38), but a similar dose of sc ghrelin increased LVEF without change in MAP (29). In participants with heart failure, higher and repeat dose infusions decreased MAP, pulmonary capillary wedge pressure, and systemic vascular resistance and increased cardiac index, stroke volume, and LVEF; these changes were associated with improved exercise capacity (15, 60). In GH-deficient patients, a small ghrelin bolus did not affect left ventricular ejection phase velocities or blood pressure (61). These endpoints have been assessed in a few small studies, which were frequently not randomized, blinded, or placebo controlled (Supplemental Table 3).

Ghrelin may have a therapeutic effect in patients with respiratory diseases. In patients with mild pulmonary cachexia from chronic obstructive pulmonary disease, ghrelin infusion increased respiratory muscle strength, Karnofsky performance status score, and 6-minute walk distance (6MWD), but not pulmonary function measured by spirometry (16). In more severely affected chronic obstructive pulmonary disease patients, there was improvement in symptoms and respiratory muscle strength but no difference in 6MWD (23). In patients with chronic respiratory infections, ghrelin administration resulted in increased 6MWD and decreased alveolar-arterial oxygen gradient (18). These endpoints have been assessed in only 3 small studies, which were not all randomized, blinded, or placebo controlled (Supplemental Table 3).

Ghrelin has also been implicated in vascular changes in humans. In healthy men, intra-arterial ghrelin bolus increased forearm vasodilation independent of nitric oxide (34). Intra-arterial ghrelin infusion in metabolic syndrome improved endothelium-dependent vasodilator responses (24), likely by increased availability of nitric oxide (25). These endpoints have been assessed in 4 studies, 2 of which were randomized, blinded, and placebo controlled (Supplemental Table 3). Unacylated, but not acylated, ghrelin infusion protects endothelial cells necessary for vascular remodeling in patients with type 2 diabetes mellitus (27).

Neuroendocrine Outcomes

The GH response to exogenous ghrelin was independent of gender, did not vary with the menstrual cycle (62), and occurred in a dose-dependent manner (6, 63, 64). There was no tachyphylaxis of GH response during a 24-hour infusion of ghrelin in healthy participants (65). Ghrelin infusion did not change the circulating 20-kDa GH isoform to total circulating GH ratio (66). GH secretion in response to ghrelin bolus occurred in malnourished dialysis patients (32), type 1 diabetes patients (67), and gastrectomized patients (68, 69). It was blunted in obesity (70–72), PCOS (72), anorexia nervosa (73), hyperthyroidism (74, 75), Cushing's disease (76), and primary hyperparathyroidism (77). The GH response to ghrelin infusion occurred in gastrectomized patients (78), obesity (79), and anorexia nervosa (45). Damage to the pituitary stalk or pituitary reduced GH response to ghrelin bolus (14, 80), suggesting the potential for ghrelin bolus administration as a GH stimulation test in diagnosing GH deficiency (81). The GH response to ghrelin bolus was more robust than the response after GHRH bolus (7, 82, 83) or hexarelin (7) and was synergistic with the GHRH response (7, 64, 84). Somatostatin (SS; also known as GH release-inhibiting hormone) infusion (85), oral glucose tolerance test, and lipid infusion only weakly blunted the GH response to ghrelin bolus, possibly due to ghrelin antagonization of SS action (83). Additionally, L-arginine (a SS inhibitor) potentiated the effect of ghrelin bolus on cortisol secretion (86), and ghrelin bolus resulted in an increase in circulating SS concentrations (87, 88). High/normal total T improved GH response to ghrelin bolus during SS infusion but not GHRH infusion (84). The GH response to ghrelin bolus was reduced by centrally acting cholinergic antagonism (89), but was not affected by peripherally acting cholinergic blockade (90), cholinergic agonist (89, 90), oxytocin (91), dopamine receptor blockade (92), or a β-adrenergic agonist (93). In healthy postmenopausal women, estradiol or combination estradiol-progestin replacement increased GH secretion in response to a ghrelin bolus (94, 95), and estradiol replacement increased basal, but not pulsatile, GH secretion in response to a ghrelin infusion (96). Effects of acylated ghrelin on GH have been assessed in over 1000 participants, including in multiple randomized, blinded, placebo-controlled trials and studies of > 24 hours in duration (Table 2). Unacylated ghrelin bolus did not induce GH secretion in obese patients (36).

A single ghrelin bolus or ghrelin infusion increases circulating ACTH, cortisol, and prolactin in healthy subjects (6, 7, 38, 41, 64, 97–99); only 1 study failed to show an effect on cortisol levels (100). Most but not all studies showed that ghrelin decreased LH and FSH in healthy men (101–103) and possibly women (104, 105), without an effect on TSH (6, 38, 64, 106). These pituitary hormone responses to ghrelin bolus were demonstrated in obesity (70), type 1 diabetes (67), anorexia nervosa (73), and bulimia (98). In patients with heart failure, infusion of ghrelin increased GH, prolactin, ACTH, and cortisol, but not FSH, LH, or TSH (60). Patients with hyperthyroidism had an exaggerated ACTH response but normal cortisol response to ghrelin (75, 107). Patients with acromegaly had normal GH, ACTH, and prolactin responses to ghrelin with a blunted cortisol response (88). Although GH response was blunted in patients with Cushing's disease, ACTH and cortisol responses were normal (76) or elevated (108). The effect of ghrelin bolus on pituitary hormones has been shown to be dose-dependent (6) and not affected by SS infusion (85) or oxytocin injection (91). Prolactin responses to ghrelin in normal women were not enhanced by TRH administration (109) or dopamine receptor blockade (92). ACTH, cortisol, and prolactin have been assessed in multiple randomized, blinded, placebo-controlled trials, but only 2 studies of repeated dosing (Table 2). The other pituitary hormones are less well studied (Supplemental Table 3).

Metabolism Outcomes

Ghrelin modulates metabolism and glucose homeostasis in a GH-independent manner, suggesting a link between ghrelin and the endocrine pancreas. One early study found no relationship between ghrelin infusion and fasting glucose, glucose production or utilization rates, insulin, glucagon, or C-peptide in healthy subjects (100), and another failed to find glucose changes in obese patients after ghrelin injection (110). However, these results are discrepant from multiple later studies. Glucose and insulin have been assessed in multiple (10 or more) randomized, blinded, placebo-controlled trials but few studies of repeated dosing (Table 2). Both ghrelin bolus and ghrelin infusion have consistently been shown to increase glucose levels in healthy young men and women (9, 37, 87, 111) and in obese (70, 112), thyrotoxic (74), and bulimic patients (98) and individuals with cancer (113) or PCOS (72). After ghrelin administration, patients with anorexia nervosa had a decrease in C-peptide levels (114) but blunted hyperglycemia response (73, 114). Similar blunting occurred in malnourished dialysis patients (32). Patients with acromegaly had an exaggerated rise in glucose and decrease in insulin in response to ghrelin compared to healthy controls (88), and patients with Cushing's disease had an absent glucose response (76). The hyperglycemic response to ghrelin is independent of GH (115) and the vagus nerve (79); instead, direct effects on both insulin resistance and insulin secretion occur. After ghrelin administration, insulin levels in healthy subjects either remained stable (58, 70, 98) or decreased (9, 37, 87, 111) in the face of increasing glucose. Insulin resistance was found in healthy participants (99, 116) and gastrectomized patients (117) in hyperinsulinemic euglycemic clamp studies after ghrelin infusion. In healthy volunteers, a long ghrelin infusion blunted early postprandial insulin responses without a significant increase in glucagon (118). Ghrelin bolus blunted arginine-induced insulin secretion in healthy participants (9). Although ghrelin bolus failed to blunt oral glucose tolerance test-induced insulin secretion (9), a similar dose of ghrelin infusion suppressed iv glucose tolerance test-stimulated insulin and C-peptide (119). The glucose and insulin responses to ghrelin were not affected by cholinergic agonist or antagonist administration (90). Differences between postghrelin insulin concentrations were not clearly related to route of administration or dose in these studies, but may have been related to varying acylated to unacylated ghrelin ratios across study populations.

In contrast to acylated ghrelin infusion, unacylated ghrelin infusion decreased fasting glucose and increased postprandial insulin in healthy humans (28). In obese patients, unacylated ghrelin alone did not cause changes in glucose or insulin levels during fasting or postprandial conditions, but when unacylated ghrelin was administered in combination with acylated ghrelin, it improved insulin sensitivity (36). In GH-deficient men, administration of acylated and unacylated ghrelin each increased fasting and postprandial glucose levels, whereas coadministration of acylated and unacylated ghrelin increased insulin sensitivity (35).

Ghrelin administration increased lipolysis in healthy humans (99, 100) and potentiated a β-agonist-induced increase in free fatty acids (FFAs) (120). Lipolysis was ghrelin dose-dependent and independent of circulating insulin and GH levels (26). Acylated ghrelin infusion was associated with a rise in FFA before each meal in healthy, obese, and gastrectomized patients (79). Lipolysis has been assessed in 7 small randomized, blinded, placebo-controlled trials but no repeated dose studies (Table 2).

Unacylated ghrelin infusion, in contrast, decreased circulating FFA levels in healthy volunteers (28). Acylated plus unacylated ghrelin coadministration decreased postprandial FFA levels in GH-deficient men (35). Unacylated ghrelin alone or in combination did not affect FFA levels in obese patients (36).

Sleep Outcomes

Ghrelin increased slow-wave sleep and stage 2 non-REM sleep and decreased REM in young and elderly healthy men, but not women, when a ghrelin bolus was administered in the late evening (121–124), but did not affect sleep when injected in early morning (125). Depressed men had an increase in non-REM but no change in REM sleep with ghrelin administration, whereas women with major depression had decreased REM sleep in response to ghrelin (126). Sleep outcomes have been assessed in a 6 small randomized, blinded, placebo-controlled trials but not for more than 1 night (Supplemental Table 3).

Other Outcomes

Ghrelin may have an inhibitory effect on the sympathetic nervous system (127–129). In both healthy and gastrectomized patients, ghrelin infusion did not affect type 1 collagen β C-telopeptide or procollagen type-1 amino-terminal propeptide bone turnover markers after a fixed meal (78). Ghrelin administration decreased airway neutrophils and inflammatory cytokines in chronic respiratory infection (18). Ghrelin administration increased sniff magnitudes in response to odorants (130). There was no evidence of body surface temperature decrease after ghrelin infusion (131).

Adverse Events

At the doses evaluated in the 66 published studies with adverse event reporting, ghrelin demonstrated an excellent short-term safety profile with few adverse effects (Table 3). Serious adverse events such as pneumonia, enteritis, and lung cancer were extremely rare and difficult to attribute biologically to ghrelin administration. Most of the severe adverse events derived from 1 study of ghrelin-vs-placebo administration in severely ill patients with pulmonary cachexia, a group that is vulnerable to developing additional medical problems (23). Mild adverse events occurred in approximately 20% of participants receiving ghrelin. The most common effect was transient flushing, which occurred in 10% of volunteers, but resulted in discontinuation of study medication in only 3 of the 939 participants in whom adverse event collection was reported (20, 22, 129). There was no difference in the percentage of participants experiencing flushing between bolus and infusion routes of administration. Larger ghrelin doses may increase the risk of flushing, as indicated by the higher rate of flushing in the 2 ghrelin bolus studies that employed the largest tested dose, 10.0 μg/kg. The most common gastrointestinal side effect was gastric rumbles, which occurred in 22 (2.3%) participants and was never severe enough to lead to ghrelin discontinuation. Gastrointestinal side effects and increased thirst were more common in volunteers who received ghrelin infusions, perhaps due to the longer duration of exposure to ghrelin. Few participants developed neurocognitive effects including somnolence, fatigue, vertigo, or change in mood (2.8%; 26 subjects), but these effects were more common in subjects who received ghrelin bolus, potentially due to the rapid ghrelin delivery.

Table 3.

Adverse Events by Administration Type, in the Subset of 66 Studies Reporting on Adverse Events

Adverse Event	Bolus		Infusion		Total Adverse Events
	n	%	n	%	n	%
Flushinga	65/619	10.5	33/320	10.3	98/939	10.4
Gastrointestinal
Gastric rumbles	4/619	0.6	18/320	5.6	22/939	2.3
Emesis/nausea	8/619	1.3	0	0	8/939	0.9
Abdominal pain/discomfort	2/619	0.3	5/320	1.6	7/939	0.7
Frequent bowel movements	4/619	0.6	1/320	0.3	5/939	0.5
Neurocognitive
Somnolence/fatigue	15/619	2.4	2/320	0.6	17/939	1.8
Vertigo	5/619	0.8	0	0	5/939	0.5
Mood elevation	3/619	0.5	0	0	3/939	0.3
Depression	0	0	1/320	0.3	1/939	0.1
Other
Thirst/dry mouth	0	0	8/320	2.5	8/939	0.9
Glucosuria	1/619	0.2	0	0	1/939	0.1
Worsening neuropathy	0	0	1/320	0.3	1/939	0.1
Shortness of breath	0	0	1/320	0.3	1/939	0.1
Decrease in blood pressure	0	0	1/320	0.3	1/939	0.1
Liver function test elevation	0	0	1/320	0.3	1/939	0.1
Increase in total cholesterol	0	0	1/320	0.3	1/939	0.1
Hypoproteinemia	0	0	1/320	0.3	1/939	0.1
Pneumonia	0	0	1/320	0.3	1/939	0.1
Infective enteritis	0	0	1/320	0.3	1/939	0.1
Lung cancer	0	0	1/320	0.3	1/939	0.1
Totalb	107/619	17.3	77/320	24.1	184/939	19.6

^aIncludes warm/sleepy feeling, facial warmth, sweating, perspiration.

^bTotal may include some subjects more than once if they experienced multiple adverse events.

Conclusions

Originally discovered as a ligand for the GHS-R, ghrelin has since been found to have wide-reaching functions within humans, many of which are independent of GH. This has afforded scientists many opportunities to test ghrelin as a therapeutic agent for a wide array of disorders. This review sought to catalog the beneficial and adverse effects of ghrelin administration in humans, as documented in a number of small studies. We also chose to evaluate the strength of current evidence regarding each of these documented effects. We found that there is strong evidence that ghrelin is an effective appetite stimulant, resulting in increased energy intake. This effect has been demonstrated in many populations where weight gain might be beneficial, including cancer, renal disease, and pulmonary cachexia. There is less evidence that ghrelin can cause positive changes in body composition and almost no evidence of increase in muscle strength and performance due to a paucity of studies with this outcome measure. There is some evidence that ghrelin may increase gastric motility, especially in patients with gastroparesis, and improve cardiac function in subjects with heart failure, although the number of participants in these studies was small. There is strong evidence that ghrelin causes secretion of GH, prolactin, ACTH, and cortisol, but the evidence that ghrelin affects FSH, LH, and TSH levels is conflicting and requires further assessment. There is strong evidence that ghrelin causes hyperglycemia, perhaps due to GH-independent effects on insulin, although the mechanism is not yet fully elucidated. There is less evidence regarding effects of ghrelin on the vasculature, pulmonary function, lipolysis, or sleep, and these areas could use confirmation by additional studies. Finally, there is some evidence that unacylated ghrelin can reverse the metabolic changes induced by acylated ghrelin, although this area is not yet fully evaluated.

This review is not a meta-analysis. The patient populations and methodologies of each study are too disparate to combine into a single statistical result. Instead, we dissected the differences in study populations and research approaches to report on the effects of different study doses and routes of administration on widely varied patient groups. The data suggest that many patient populations benefit from the effects of ghrelin. Another limitation to our review is the potential for bias in the estimates of adverse events, due to differences in thresholds for reporting the presence or absence of adverse events across studies, resulting in exclusion of 55 studies. In addition, many studies lacked randomization or placebo control, and the duration of ghrelin administration was generally limited to a single dose. In the studies conducted in patient populations with serious illness, who are at baseline vulnerable to developing additional medical problems, numbers were too small to ascertain whether adverse events were truly ghrelin-related. In addition, long-term safety and efficacy data are not available. For brevity we chose not to incorporate studies of oral ghrelin receptor agonists, such as MK-677 and TZP-102, in the present review although several of these drugs are currently under development. Of note, although easier to administer, these drugs have substantially longer half-lives and greater risk of hyperglycemia. Although we acknowledge these limitations, the aggregate data provide a reassuring view of ghrelin as a safe short-term therapeutic agent with few serious adverse effects, even in populations with serious comorbidities.

Existing information supports several therapeutic prospects for ghrelin administration. Ghrelin clearly increases GH secretion, but, due to its short half-life, does not offer specific benefits over administration of GH itself for this purpose. Ghrelin shows promise in stimulating appetite and energy intake in populations with negative energy balance such as cancer, renal disease, and the elderly (32, 33). Although these studies were too short to evaluate outcome measures such as hospitalizations or mortality, the improvements in intermediate outcome measures are encouraging. A small number of published studies indicate that ghrelin may improve gastric motility in gastroparesis, although this has not been assessed in repeated doses.

At present, a full evaluation of the therapeutic prospects of ghrelin is incomplete. Despite the large number of studies in which ghrelin was administered to humans, most studies were small and of short duration. Further research is required to examine the clinical applications of ghrelin, including its physiology, its full complement of effects on the human body, and its potential as an extended-use therapeutic agent in multiple patient populations.