Combined arginine and insulin stimulation elicits a robust and consistent copeptin response in short children

Abstract

Introduction:

Copeptin, co-secreted with arginine-vasopressin, is regulated by osmotic and volume stimuli but also responds to intravenous arginine and insulin-induced hypoglycemia. The serum copeptin response to the latter agents has been studied in adults, but only to a limited extent in children. The objective of this study was to describe the copeptin response to combined arginine and insulin in children with normal posterior pituitary function.

Methods:

We conducted a prospective, single-arm assessment of serum copeptin concentrations in children (age 7–16 years, n=38) undergoing growth hormone stimulation testing with an arginine-insulin tolerance test (AITT) for short stature or growth deceleration in a tertiary referral center. After overnight fasting, arginine (500 mg/kg) was administered between 0–30 minutes intravenously (IV) followed by insulin (0.1 units/kg IV) at 60 minutes. Copeptin serum concentrations were measured at baseline (0 minutes), at the post-arginine peak (60 minutes), and at the post-insulin peak (90 minutes; 30 minutes post-insulin), respectively. The main outcome was the peak copeptin concentration.

Results:

Mean±SD copeptin concentrations increased from 9.9±5.0 pmol/L at 0 minutes to 13.2±5.8 pmol/L at 60 minutes (p<0.0001 vs. 0 minutes) and 27.7±14.2 pmol/L at 90 minutes (p<0.0001 vs 0 and 60 minutes). There was no significant correlation between copeptin concentrations and age, BMI, pubertal status, cortisol, growth hormone, or glucose concentrations.

Discussion/Conclusion:

Arginine and insulin appear to have an additive and consistent effect resulting in significant stimulation of copeptin secretion in children. The AITT may be a useful tool to evaluate for normal posterior pituitary function in this age group, with potential implications for the evaluation of polyuria-polydipsia syndrome.

Introduction

The secretion of arginine vasopressin (AVP) is mainly regulated by osmotic stimuli and by intravascular volume changes [1, 2]. Although important in conditions of altered osmotic or volume homeostasis, the measurement of plasma AVP concentrations has been limited by pre-analytical and analytical difficulties of the AVP assay, as well as scarce commercial availability [3]. In the last 15 years, copeptin, the C-terminal of the pro-AVP molecule, which is co-secreted with AVP in equimolar amounts, has emerged as a surrogate marker of AVP secretion [4, 5]. Copeptin has the advantage of being very stable in serum or plasma, where it can be easily quantitated by a reliable immunometric assay [4].

Measurement of serum copeptin has been shown to be useful in the diagnostic evaluation of adult subjects with the polyuria-polydipsia syndrome (PPS), a condition characterized by large output of hypotonic urine (greater than 50 mL/kg in adults [6] or greater than 2 L/m²/day in children [7]) with similarly increased intake of fluids. The differential diagnosis of PPS includes primary polydipsia, central and nephrogenic diabetes insipidus. Traditionally, testing for these conditions has involved hypertonic saline or water deprivation testing. Both are cumbersome and require inpatient hospitalization, and the water deprivation test may have limited sensitivity to distinguish partial central diabetes insipidus (CDI) from primary polydipsia [8].

Prior studies have found that copeptin, like AVP, responds to stimuli unrelated to sodium or volume balance, including intravenous (IV) arginine and insulin-induced hypoglycemia. As these agents appear to evoke a consistent, yet quantitatively limited copeptin response, they have been proposed alternatives to traditional testing methods to evaluate for PPS. Stimulated copeptin levels greater than 3.8 pmol/L in response to arginine [9] or 4.7 pmol/L in response to insulin [10] have been suggested to exclude central diabetes insipidus in adults. Only limited normative values have been published for serum copeptin concentrations in children [11, 12]. The serum copeptin response to arginine stimulation alone appears to be more modest in children than in adults [9] and somewhat lower than the response to insulin in children or adults [10, 13]. As pharmacological stimulation of copeptin has great potential to replace routine water deprivation testing when evaluating posterior pituitary function in children, we require a better understanding of the stimulatory effect of different agents. The primary aim of this study was to describe the copeptin response to combined administration of arginine and insulin in clinically healthy, short children. A secondary objective was to describe reference values for pediatric fasting and stimulated copeptin concentrations to add to the existing body of literature.

Materials and Methods

Setting & Participants

We conducted a prospective assessment of serum copeptin concentrations during the arginine-insulin tolerance test (AITT) in a convenience sample of short children undergoing evaluation for growth hormone deficiency. Children had no known health issues aside from their primary growth concerns. They were recruited from the UPMC Children’s Hospital of Pittsburgh, an academic, tertiary referral center between 2019–2021. In our institution, the AITT is the test of choice to assess children for possible growth hormone deficiency. Our highly trained testing nurses, under physician supervision, have performed at least 1,500 such tests in the last 20 years in pediatric patients 3 years and older without any significant incident or adverse outcomes.

Included subjects had diagnoses of short stature, growth deceleration, or both for their testing indication. Children were excluded if there was prior history of polydipsia or polyuria, known neurological or psychiatric conditions (apart from attention-deficit hyperactivity disorder), known anatomical central nervous system abnormalities, if they were taking corticosteroids or other medications known to interfere with growth hormone, cortisol or copeptin responses, or they had an abnormal serum sodium value. Included subjects were euthyroid clinically and by thyroid function testing. They had a normal set of electrolytes obtained during their initial evaluation for short stature and a normal repeat serum sodium value at the time of the AITT. The study protocol was approved by the University of Pittsburgh Institutional Review Board (PRO19120068). Written informed consent was obtained from the participants’ parent/legal guardian/next of kin prior to participation in the study.

Test Procedures

AITTs were performed in the morning following an overnight fast for at least 8 hours. Children were permitted ad lib fluids the day prior. On the morning of the test, we assessed weight and vital signs to evaluate for any clinical signs of dehydration. Placement of an intravenous cannula (angiocath) was performed by trained nurses without the application of local anesthetics. Arginine-Hydrochloride (10% solution) 500 mg/kg was infused IV between 0 and 30 minutes. At 60 minutes, insulin 0.1 units/kg was administered IV over 1 minute. Blood samples were obtained at 0, 15, 30, 45, 60 minutes (just before insulin administration), 80, 90, 105 and 120 minutes for measurement of serum glucose, growth hormone and cortisol.

Children included in the study achieved hypoglycemia (serum glucose ≤ 50 mg/dl or 2.8 mmol/L); mean glucose values are shown in Table 1. During testing, nursing staff monitored children for signs and symptoms of nausea or hypoglycemia with an attending physician available on the premises. Nausea was reported at 60 minutes by 1 participant (3%). Blood glucose was monitored by bedside hospital meter. If the patient was experiencing symptomatic hypoglycemia (e.g., hunger, cold sweat, sleepiness, or impaired consciousness) at the 90-minute point, at the nadir of blood glucose, juice (120 ml or approximately 12 grams of sucrose) was given. This occurred in n=15/38 (40%) of subjects; in one subject, IV glucose (0.25 g/Kg, as a 25% solution administered over 2–3 minutes) was required at 90 minutes. The test was continued to completion, irrespective of the need for the glucose/sucrose rescue.

Table 1:

Background characteristics and AITT findings for included participants

Characteristic	Mean±SD, Median [IQR] or N(%)
Age, years	12.4±2.2
Male sex	27 (71)
Race/Ethnicity
Non-Hispanic white	31 (82)
Non-white	7 (18)
Presenting diagnosis
Short stature	34 (89)
Growth deceleration	4 (11)
Height z-score	−2.2 [−2.7, −1.8]
Weight z-score	−1.6 [−2.3, −0.5]
BMI z-score	−0.1 [−1.3, 0.8]
Tanner stage
Stage 1 (prepubertal)	18 (47)
Stage 2 (pubertal)	20 (53)
Primed for AITT*	3 (17)
Growth hormone, base, ug/L	0.3 [0.1,1.8 ]
Growth hormone, peak, ug/L	5.8 [3.4, 8.8]
Cortisol, base, nmol/L	314±154
Cortisol, peak, nmol/L	556±117
Glucose, base, nmol/L	4.7±0.9
Glucose, nadir, nmol/L	1.9±0.5
IGF-1 level, ug/L	130 [100,153]
IGF-1 z-score	−2.0±0.9
Diagnosed with GH deficiency (peak GH <10 ug/L)	29 (76)
Abnormal MRI findings^	5 (20)

^*If prepubertal, at the discretion of the endocrine provider

^{^}MRI was obtained if peak GH <10 ng/mL for most subjects (n=25)

The optimal sampling times for serum copeptin were determined by a preliminary study in the first 16 subjects undergoing the AITT (mean age±SD 12.3±2.6 years, 59% male). In this pilot group, copeptin was measured at frequent intervals of 0, 15, 30, 45, 60, 80, 90, 105, and 120 minutes (Fig 1, inset) in accordance with the protocol for the other laboratory studies measured in the AITT. Copeptin concentrations were found be consistently highest at 60 minutes post-arginine and at 80–90 minutes post-insulin (hereafter labeled the post-arginine peak and post-insulin peak, respectively). The timing and magnitude of the post-arginine peak was in keeping with what was previously reported in children [9]. Therefore, copeptin levels were measured at 0, 60, 80, and 90 minutes for the subsequent subjects and are reported for the entire cohort.

Figure 1:

Serum copeptin concentrations at 0, 60, 90 minutes during the AITT are shown in the columns on the left. The two columns on the right show the delta copeptin increase from 0–60, and 0–90 minutes, respectively. *** p< 0.0001; n=38. The inset indicates the copeptin response to the AITT in the first 16 subjects, which included more frequent blood sampling.

Laboratory Assays

Serum copeptin was measured by an automated 2-site immunofluorescent assay (B·R·A·H·M·S Copeptin proAVP KRYPTOR, ThermoFisher Scientific) at Quest Laboratories. The KRYPTOR assay is linear from 2 – 500 pmol/L with intra-assay coefficient of variation (CV) ≤ 5.2% and inter-assay CV ≤ 3.7%. Other laboratory tests were performed in the College of American Pathologists-accredited UPMC Clinical Laboratories on FDA approved assays, to include serum cortisol measured on the UniCel DxI 800 (Beckman Coulter, Brea, CA, USA), and growth hormone measured on the Immulite 2000 chemiluminescent immunometric assay (Siemens Healthineers). IGF-1 was measured by mass spectrometry (liquid chromatography-high resolution/accurate mass-based LC-MS, Quest Diagnostics, Chantilly, VA USA).

Clinical Variables

At the time of testing, the following variables for each patient were collected: age, sex, anthropometric measurements, Tanner stage by a pediatric endocrinologist, use of sex steroid priming pre-AITT, IGF-1 serum concentration and z-score, bone age z-score, baseline and nadir glucose, and baseline and peak cortisol and growth hormone concentrations during the AITT. Sex steroid priming, if indicated, was completed with oral estradiol 1 mg twice daily for two days prior to testing. Most of the individuals with peak growth hormone level <10 ng/mL underwent subsequent brain magnetic resonance imaging (MRI), and we recorded any identified anatomic abnormalities of their midline structures.

Analysis

Descriptive statistics of participant background characteristics, AITT test findings, and serum copeptin concentrations at all time points were completed. Median with inter-quartile range or mean and standard deviation are presented depending upon the distribution of the continuous variable. The main outcome measures were serum copeptin level at 60 minutes (post-arginine peak) and 90 minutes (post-insulin peak). Copeptin concentrations at 90 minutes were similar to (within 2 pmol/L), or higher than, values at 80 minutes, and thus 90-minute values were labeled as the post-insulin peak for analyses. To examine the change in serum copeptin from baseline, we compared the delta increase between 0 to 60 minutes, 0 to 90 minutes, and 60 to 90 minutes using paired t-tests. Unpaired t-tests were used to examine differences in baseline and stimulated copeptin levels by sex and pubertal status. Significant relationships were determined by a p-value of <0.05. We further examined whether the relative increase (from now on referred to as “delta increase” or ”Δ”) varied by baseline copeptin level quartile using one-way ANOVA, with our significance determination adjusted for multiple comparisons with the Bonferroni method with a significance threshold of p<0.0125.

The Pearson correlation coefficient was used to examine the potential relationships between baseline and stimulated copeptin concentrations and the copeptin concentrations with clinical variables, including age, height, weight and BMI z-scores, glucose, growth hormone, and cortisol serum concentrations. Baseline values in the AITT were tested with the 0-minute copeptin level, whereas peak growth hormone and cortisol levels and the glucose nadir were tested with the 60- and 90- minute levels. Given the number of clinical variables tested, significant correlations were adjusted for multiple comparisons using the Bonferroni method with a significance threshold of p<0.007. All analyses were completed in STATA v. 15.0 (StataCorp LLC, College Station, TX).

Results

Forty-seven youth undergoing AITT were consented for copeptin measurements. We excluded subjects with missing copeptin measurements (n=4) or extremely high baseline copeptin concentrations > 50 pmol/L (n=2). Such high levels have been reported in relation to significant dehydration, AVP resistance, stress, or other unclear etiologies (11). We also excluded two subjects who were later found to have multiple anterior pituitary hormone deficiencies and one subject who was found to have an elevated serum sodium value at the time of AITT (>145 mmol/L), despite a normal sodium six months earlier. A final sample of 38 subjects was thus available for analysis, mean age 12.4±2.2 years (range 7.4–16.0), 71% male, 82% non-Hispanic white, 47% prepubertal (Tanner 1, n=18; Tanner 2, n=13; Tanner 3, n=7).

Background characteristics and hormone measurements from the AITT are included in Table 1. There were no significant differences in mean age and sex distribution between the excluded and included subjects. As shown in Figure 1, mean copeptin concentrations were 9.9±5.0 pmol/L (range 3–25) at baseline, 13.2±5.8 pmol/L (range 4–28) at 60 minutes (post-arginine peak), and 27.7±14.2 pmol/L (range 8–68) at 90 minutes (post-insulin peak). The post-arginine and post-insulin peak copeptin levels were significantly higher compared to baseline (p<0.001 for both) and the post-insulin peak was significantly higher than the post-arginine peak (p<0.001). The post-arginine peak was relatively modest compared to baseline (~30% increase). The post-insulin peak was notably greater, with ~100% increase in copeptin from 60 to 90 minutes and ~180% increase in copeptin when comparing baseline to 90 minutes.

There was no significant difference in baseline or 60-minute stimulated copeptin levels by sex. At 0 minutes, mean copeptin level was 9.5±1.7 pmol/L in girls compared to 10.0±4.9 pmol/L in boys (p=0.80) and at 60 minutes, mean copeptin level was 13.4 ±6.8 pmol/L in girls compared to 13.1±5.4 pmol/L in boys (p=0.90). However, at 90 minutes, girls had a significantly higher stimulated copeptin level to 35.0±18.0 pmol/l compared to 24.7±7.9 pmol/L in boys (p=0.04). Notably, there was a greater range of values for the 90-minute copeptin level in girls (22.9–47.1 pmol/L) compared to boys (19.6–26.3 pmol/L). There were no significant differences in 0, 60, or 90 minute copeptin levels by pubertal status comparing prepubertal (Tanner 1) to pubertal (Tanner 2 or 3) subjects (all p>0.05).

Twenty-nine (76%) of children had a peak growth hormone level <10 ng/mL, meeting U.S. criteria for growth hormone deficiency. Of these, 25 (86%) underwent a brain MRI to evaluate their midline structures. Five children (20%) were found to have an abnormality, including small anterior pituitary (n=2), Rathke’s cleft cyst (n=1), partial empty sella (n=1), and ectopic posterior pituitary/pituitary stalk interruption syndrome (n=1). Excluding these children, there were no significant changes in serum copeptin concentrations at all time points (data not shown). Though all five of these children had a peak growth hormone level <10 ng/mL, they had normal peak cortisol levels (≥18 μg/dL or 500 nmol/L) and 4 out of 5 (including the subject with ectopic posterior pituitary) had a robust copeptin response. One individual with a small anterior pituitary had a slightly lower simulated copeptin concentration of 8 pmol/L at 90 minutes (baseline concentration 4 pmol/L); removing this individual, the minimum 90-minute stimulated copeptin concentration was 11 pmol/L. Thus, 97% of subjects in our cohort had a copeptin response > 10 pmol/L.

There was no significant difference in the delta increase in copeptin concentrations between 0 and 60 minutes when categorizing the subjects by baseline copeptin quartile (Figure 2). There was a trend toward a higher delta increase in serum copeptin from 0 to 90 minutes (p=0.07) in quartiles 1 and 4, likely driven by three individuals (one in quartile 1 and two in quartile 4) who stimulated to a copeptin concentration >50 pmol/L. Excluding these children, there was a slight upward trend in the change in copeptin between 0 and 90 minutes with increasing quartile, but this was not significant (p=0.27). Our findings were similar excluding those subjects with an abnormal brain MRI. Baseline copeptin was highly correlated with the concentration at 60 minutes (r=0.95, p<0.0001) and moderately correlated with the concentration at 90 minutes. (r=0.49, p=0.002) (shown in Figure 3).

Figure 2:

Baseline and AITT-stimulated serum copeptin concentrations, divided by quartiles of baseline copeptin values. Quartile 1 (n=7) 0–5 pmol/L; Quartile 2 (n=9), 6–8 pmol/L; Quartile 3 (n=12), 9–12 pmol/L; Quartile 4 (n=10), 13–25 pmol/L. There was no significant difference in the delta increase in serum copeptin concentration at 60 or 90 minutes by baseline copeptin level among the 4 groups.

Figure 3:

Correlation between baseline and stimulated copeptin concentrations with significance testing using Pearson coefficient. There was a strong correlation between baseline and 60-minute copeptin concentrations, r=0.95, p<0.0001. There was a moderate correlation between baseline and 90-minute copeptin concentrations, r=0.49, p=0.0002.

Bivariate correlations between serum copeptin level at baseline, 60, and 90 minutes with other clinical variables did not identify any significant correlations with age, height z-score, weight z-score, BMI z-score, or growth hormone, cortisol, or glucose measurements (shown in Table 2). Of note, we specifically focused on six subjects who had a GH response of <3 ng/mL, as such low GH response was associated with impaired copeptin response to insulin-induced hypoglycemia in a previous study [13]. This was not the case in the present subgroup, whose mean copeptin levels during the AITT were 9.8±4.3 at baseline, 13.8±4.4 at 60 minutes (post-arginine peak) and 35.6±21.0 (range 16–68) at 90 minutes (post-insulin peak), not significantly different from the values obtained in the remaining subjects in the entire cohort.

Table 2:

Bivariate correlations between clinical factors and copeptin levels

	Copeptin (pmol/L)
Factor	0 min	P-value	60 min	P-value	90 min	P-value
Age	−0.10	0.54	−0.12	0.48	−0.28	0.09
Height z-score	0.16	0.35	0.11	0.50	0.15	0.36
Weight z-score	−0.232	0.16	−0.15	0.36	−0.11	0.50
BMI z-score	−0.32	0.05	−0.19	0.25	−0.11	0.52
GH, baseline	0.02	0.90	-	-	-	-
GH, peak	-	-	0.12	0.46	0.08	0.65
Cortisol, baseline	0.37	0.02	-	-	-	-
Cortisol, peak	-	-	0.21	0.21	0.23	0.15
Glucose, baseline	0.07	0.69	-	-	-	-
Glucose, nadir	-	-	0.23	0.17	0.01	0.94

Pearson’s correlation coefficients, rho, presented.

Significance determined by p<0.007 to account for multiple comparisons.

Discussion

This is the first study to describe the stimulated copeptin response to combined arginine and insulin. Related to our primary objective, our findings indicate several key points. First, sequential arginine-insulin stimulation yielded a copeptin response which was both consistent and robust, resulting in a mean serum copeptin concentration at 90 minutes which was nearly triple the baseline value. Second, the substantial delta increase in serum copeptin in response to arginine and insulin occurred irrespective of the baseline copeptin concentration. Lastly, stimulated copeptin levels did not correlate with BMI z-score, pubertal status, or anterior pituitary hormone levels.

Our findings demonstrated a substantially greater copeptin response to sequential arginine and insulin stimulation than what has been previously reported in the literature to either agent alone using the same assay. In a prior study of healthy short children (mean age 8.5 years, 57% male), arginine alone evoked a copeptin increase from a mean of 4.3 to 6.5 pmol/L at the 60 minute peak (a 2.2 pmol/L delta increase) [9] similar to the copeptin response to arginine in our patients (from 9.9 to 13.2 pmol/l, a 3.3 pmol/L delta increase). In another series of short children (mean age 11.7 years, 75% male), insulin-induced hypoglycemia resulted in serum copeptin increasing from a baseline of 5.2 to a peak of 9.7 pmol/L at 45 minutes (a 4.5 pmol/L delta increase) [13]. The copeptin response to arginine alone [8] or insulin alone [10, 14] appears to be greater in adults, with peak concentrations of 9.8 pmol/L (arginine), 8.5–11.1 pmol/L (insulin), and delta increments ranging from 4.4 to 7.4 pmol/L. Though the copeptin concentrations in response to arginine were modest in our study, insulin-induced hypoglycemia resulted in a near doubling of the copeptin concentration achieved at 60 minutes. Given the post-arginine copeptin peak is maximal or submaximal at 60 minutes in children and adults [9], the magnitude of the copeptin rise to subsequent insulin administration in our study suggests a possible intensification in copeptin stimulation in response to sequential administration of these agents.

The mechanisms by which arginine and insulin stimulate copeptin secretion is unclear. Nitric oxide has been proposed as the mediator of AVP/copeptin response to arginine [9, 15, 16] though this is controversial [17, 18]. In the case of insulin, older studies have postulated that insulin-induced hypoglycemia stimulates AVP secretion as part of the stress response, possibly mediated by direct activation of pre-pro-AVP cells in both the magnocellular and parvocellular neurons [19, 20]. More recent studies in rodents suggest that the A1/C1 neurons in the ventrolateral portion of the medulla oblongata are activated by hypoglycemia and project to AVP-secreting neurons of the supraoptic and paraventricular nuclei [21].

Another possibility related to somatostatin provides a theoretical framework to support our observation that sequential arginine and insulin stimulation exerts a robust effect on copeptin secretion. Somatostatin has been reported to inhibit the AVP response to insulin-induced hypoglycemia [22]. Indirect evidence points to arginine as having an overall suppressive effect on somatostatin [23], even though somatostatin concentrations increase during arginine infusion [24]. Thus, a post-infusion decrease in the somatostatinergic tone (“rebound effect”) could explain, at least in part, the stimulating effect of arginine and/or its priming action on the posterior pituitary response to hypoglycemia. A priming effect could also explain the earlier timing of the copeptin peak response to insulin (20–30 minutes after IV insulin) in our cohort, in comparison to the peak response at 45 minutes reported after insulin alone [10, 13, 14].

We found no correlation between growth hormone and copeptin concentrations in response to the combined stimulation of the AITT, including documentation of similar copeptin response in a subgroup of children with a growth hormone response of less than 3 ng/ml. Prior reports in the literature have found conflicting results regarding stimulated copeptin in the presence or absence of growth hormone deficiency [9, 13, 14]. Additionally, we found no correlation between the cortisol peak or the glucose nadir with the copeptin response. The lack of correlations among these parameters associated with the acute stress of hypoglycemia in our cohort need to be interpreted with caution, as the range of both cortisol and glucose concentrations during the AITT were narrow in most tested subjects, with few individuals showing a low cortisol response.

Though there was no difference in 0- or 60-minute copeptin level by sex, we found that girls had a significantly higher 90-minute copeptin level following arginine and insulin stimulation. One prior study noted significantly higher copeptin levels in boys compared to girls [11], although this has not been noted in others [12, 13]. The observed difference in our 90-minute copeptin levels may be related to the smaller number of female participants and the greater range in their stimulated copeptin levels; this finding would need to be validated and further explored in a larger cohort. We found no associations between copeptin levels and age, pubertal status, or BMI. Similar to our findings, prior studies have found that neither pubertal status [11, 12] nor sex steroid priming [13] influenced copeptin measurements. However, the small number of subjects who underwent sex steroid priming (n=3) limited our ability to study the effect on copeptin response to stimulation. Though obesity may be associated with higher copeptin [12], BMI in other ranges may not affect copeptin levels. As no subjects were obese in our participant sample, we could not assess the relationship between obesity and copeptin levels.

Our findings have potential implications for the clinical use of copeptin measurements, including contributing to establishing reference intervals for children. In our study, baseline serum copeptin concentrations were measured after an overnight fast, which included water restriction. The mean baseline copeptin concentration of 9.9±5.0 pmol/L was comparable to that found in a similarly fluid-restricted cohort of children [12]. However, the mean value was higher than the median baseline copeptin levels (ranging from 4.3–7.6 pmol/L) in children with free access to water until 0–2 hours before stimulation testing [9, 11–13]. Consequently, the reference range for copeptin values may need adjustment to the interval of fluid limitation before blood sampling.

Though this study does not provide any insight into the diagnostic utility of AITT for PPS, we would make two points. Firstly, if tolerated by children, simply obtaining a morning copeptin level following overnight water restriction may be a reasonable first step to rule out CDI, as suggested for adults [8]. Prior studies have proposed a cutoff random copeptin value of 3.5 pmol/L as suspicious for CDI in polyuric children [11, 25]. Secondly, children may benefit from using sequential arginine and insulin stimulation to evoke a greater copeptin response given their modest response to arginine alone. Cutoff values of 3.5–3.8 pmol/L after arginine stimulation have been proposed in adults [9], though no pediatric cutoff values have been suggested. The possible benefit of sequential arginine and insulin stimulation is contingent, however, on the hypothesis that the AITT-induced robust copeptin secretion in subjects with primary polydipsia and/or intact posterior pituitary function occurs without a parallel “overstimulation” of copeptin secretion in subjects with (partial) central DI. Future studies will be needed to establish the diagnostic accuracy of the AITT for PPS and compare its performance to other stimuli, including arginine alone.

This study has several limitations. The use of AITT is controversial given the concern for risk of hypoglycemia. With close supervision, the AITT has been very safe at our center, though symptomatic hypoglycemia may occur. Given the nature of this testing, the potential risks can be minimized by always providing oral (or IV) glucose at 90 minutes and ceasing the test, as stimulation of copeptin has already occurred. Second, the children and adolescents in our study were not completely “normal” as they were being evaluated for short stature and many were diagnosed with isolated growth hormone deficiency. However, they had no evidence of polydipsia, polyuria or posterior pituitary dysfunction by history, electrolyte measurement, and pituitary MRI (the latter in a subset of children). It would have been difficult to justify a test with small, yet identifiable risks, in children of normal stature. We also only have data for children in a discrete age range (7–16 years) and cannot comment on the performance of an AITT to stimulate copeptin in younger children.

Additionally, and as discussed above, the AITT was performed after overnight fluid restriction, which likely accounts for higher baseline copeptin concentrations [12] than reported in other studies with single stimulation agents [9, 13] in which children were allowed free access to water until 0–2 hours before the time of testing. Therefore, the performance of the AITT to stimulate copeptin will also need to be examined in the absence of water restriction. Lastly, priming with sex steroids was at the discretion of the clinician ordering the test. Although we did not observe any differences by pubertal status, the sample size was too small to discern whether sex steroid priming may affect the copeptin response to stimulation, though this has not been seen in a prior study [13].

Conclusion

In summary, we present the first study examining the serum copeptin response to AITT in clinically healthy, short children. This test achieved a strong and reproducible stimulation of serum copeptin concentrations which was greater than the previously reported copeptin response to either agent alone in the literature using the same assay. Future work will be needed to evaluate the diagnostic utility of this test, both in conditions of free and limited access to fluids, for the differential diagnosis of the PPS in children.

Data Availability Statement:

All data generated or analyzed during this study are included in this article. Further enquiries can be directed to the corresponding author.