Very Elevated Serum Copeptin Concentrations Occur in a Subset of Healthy Children in the Minutes after Phlebotomy

Summary

Objective:

Although AVP and its surrogate, copeptin, are mainly regulated by osmotic and volume stimuli, their secretion is also elicited by stress and growth hormone (GH) stimulating agents. The aim of this report is to describe unusual patterns of copeptin response in a subset of children undergoing GH stimulation tests (GH-ST).

Methods:

We conducted a secondary analysis of a cohort of 93 healthy short children with no polydipsia, polyuria or fluid/electrolyte abnormalities, undergoing GH-ST with intravenous arginine, insulin; oral clonidine, L-Dopa/carbidopa; in various combinations.

Serum copeptin concentrations were measured 1-3 minutes after phlebotomy (0 minutes) and at 60, 90, 120 minutes during GH-ST.

Results:

In 85 subjects (normal response group, NRG) serum copeptin concentrations increased from a 0 minutes median of 9 pmol/L (IQR 6, 11.5) (all values ≤21) to a median peak between 60-120 minutes of 22 (IQR15, 38) pmol/L, which varied depending on the stimulating agent. Conversely, in the 8 outliers, copeptin concentrations decreased gradually from a median of 154 (IQR 61, 439) pmol/L (all≥40 pmol/L) to values as low as 14% of the basal value, by 120 minutes. Test-associated anxiety was described in 17 subjects in the NRG (20%) and 5 of the outliers (63%).

Conclusions:

A distinctive pattern of very elevated serum copeptin concentrations occurred in 8.6 % of children undergoing GH-ST, similar to reports in previous pediatric studies. Etiology may include pain or stress of phlebotomy. This phenomenon should be recognized for proper interpretation of copeptin values in children.

Introduction

Copeptin, the C-terminal of the pro-AVP molecule, which is secreted in equimolar amount to arginine vasopressin (AVP), has been measured in the last 2 decades as a surrogate marker for AVP, due to its immunoassay being more robust and reliable than the AVP assay. Copeptin, like AVP, is mainly regulated by osmotic and volume stimuli(1); however, its secretion is also stimulated by acute stress(2-4).Recent studies related to copeptin have identified a sporadic phenomenon of markedly elevated copeptin levels in children(5, 6) including our previous report(7) In all these studies, the reasoning for these very high copeptin levels was unclear. We thus extended our initial observation of this puzzling finding to investigate a larger number of short but otherwise normal pediatric subjects undergoing growth hormone stimulation tests, in order to provide a more in-depth description of cases of elevated copeptin concentrations, explore potentially associated factors, and characterize in detail the subsequent pattern in copeptin response during stimulation.

Materials and Methods

This is a secondary analysis of serum copeptin concentrations in 109 children undergoing growth hormone stimulation tests between May 2020 and December 2022. The primary goal of the parent study was to assess serum copeptin response to various agents known to stimulate the anterior pituitary and presumptively copeptin secretion in otherwise healthy children. In this analysis, we unexpectedly identified cases (“outliers”) of distinctly elevated serum copeptin levels prior to stimulation testing.

Subjects

In this study, we included children whose only clinical complaint was short stature. Subjects who were found to have anatomical abnormalities of the hypothalamic pituitary axis (such as pituitary stalk interruption syndrome, septo-optic dysplasia, or status/post pituitary surgery) and/or multiple anterior pituitary deficiencies were excluded a posteriori. We included subjects who had at least four serial samples of copeptin obtained during the stimulation test between 0 and 120 minutes, to explore the subsequent change in copeptin concentrations. A total of 16 subjects were excluded due to missing samples. Approximately half of the remaining 93 subjects (n=46) had an additional copeptin level obtained after insertion of the IV catheter but 20 minutes before the growth hormone stimulation test was started (−20-minute value). These subjects were allowed to rest in a 45-degree semi-recumbent position prior to stimulation testing. No subject had excessive fluid intake by history (which was verified by a review of fluid intake records obtained from the caregivers, after copeptin values were found to be unexpectedly elevated). All subjects had normal blood count and chemistry panel (including serum sodium, potassium, calcium, creatinine, BUN and glucose measurements). All subjects had been fasting (for both solids and liquids) for at least 8 hours at the time of testing.

Test Procedures

For the growth hormone stimulation testing, children had an IV catheter placed in an antecubital (in most patients) or other veins by trained nurses using distraction techniques and no topical analgesia.

We used the following agents for stimulation: arginine and insulin combination (arginine-insulin tolerance test or AITT, as described in March et al(7)); arginine and clonidine combination [arginine 500 mg/Kg infused intravenously (IV) between 0 and 30 minutes; clonidine 150 mcg/m² given PO at 0 minutes; serial sampling between 0 and 120 minutes]; arginine and Levodopa/carbidopa combination [arginine 500 mg/Kg infused IV between 0 and 30 minutes; L-Dopa/Carbidopa (10:1 ratio), 200 mg of L-Dopa/m² given PO at time 0 minutes; serial blood sampling between 0 and 120 minutes]; and insulin alone [Insulin tolerance test or ITT: regular insulin 0.1 Units/Kg given IV at time 0; serial blood sampling between 0 and 60 to 120 minutes]. While arginine alone is a fairly weak stimulus of copeptin secretion in children(8), we have found that arginine in combination with insulin(7) or with L-Dopa/carbidopa induces a more robust copeptin response.

All subjects were maintained in a semi-recumbent position throughout the test. The first blood sample for copeptin measurement either at −20 or 0 minutes was collected between at least one, and not more than three minutes after placement of the venous catheter. This time was considered sufficient to allow copeptin to reach the periphery from the posterior pituitary (with the assumption that secretion of copeptin could be induced rapidly by venipuncture). All subsequent samples were obtained from the indwelling IV line.

Covariates

We collected and analyzed the following data on all subjects: age, sex, anthropometric measurements, the indications for growth hormone testing, serum sodium levels, and stimulated growth hormone levels. We reviewed the notes that our nurses routinely collect during the stimulation test, for their assessment of anxiety in patients in the minutes before the test, and for evidence of multiple attempts at placement of the IV line. Subjects were either described as “calm” or “anxious” (the latter being described with terminology such as: anxious, very anxious, upset, screaming).

Laboratory Measurements

Serum copeptin was measured by an automated 2-site immunofluorescent assay (B·R·A·H·M·S copeptin pro-AVP KRYPTOR, Thermo-Fisher 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 Children’s Hospital of Pittsburgh Clinical Laboratories on FDA approved assays, to include serum sodium, calcium, creatinine, BUN and other chemistries on a standard laboratory analyzer, and growth hormone measured on the Immulite 2000 chemiluminescent immunometric assay (Siemens Healthliners).

Analysis

We categorized subjects by their baseline serum copeptin concentration into a “normal response” group (NRG) and a “high copeptin outlier” group (outliers). The NRG group was defined by copeptin concentrations which increased progressively from a baseline, comparable to what was previously reported(5) , to a peak which varied in response to different stimulating agents. Eight subjects were identified as outliers based on the following criteria: A) elevated baseline (at −20 or 0 minutes) copeptin concentrations ≥ 40 pmol/L (≥ 7 SDS of the NRG) and B) a pattern of response characterized by a decrease, rather than an increase of serum copeptin concentrations during the GH-ST.

We completed descriptive statistics using IBM SPSS software for Windows. Data are presented as Mean ± SD or Median (Interquartile range, IQR), depending on the distribution. Given the small number of outliers (N=8, 8.6%) differences between the NRG and outliers are summarized descriptively only. Limited comparison testing assessed changes in copeptin concentrations in the NRG only, using paired t-tests.

Results

We categorized 85 individuals (~ 91%) as the NRG, and 8 subjects (~ 9%) met criteria as outliers. The NRG and the outliers had similar age (11.9± 3.0 vs. 12.1±2.9 years), sex distribution (73% vs 50% male), height deficit by standard deviation scores (height Z-score −2.2±0.77 vs.−2.4±0.46), serum sodium concentrations (139±1.6 vs 139±1.5 mEq/L), creatinine values(0.51+/−0.09 vs 0.54+/−0.07) and growth hormone response to stimulation (peak growth hormone 10.0±8.8 vs 10.5±6.6 mcg/L). A similar percentage of children with a growth hormone peak response of <10 mcg/L was seen in the two groups (57% in the NRG vs 63% in the outliers). Outliers tended to have more of a weight deficit, and a lower BMI (Weight Z-score: NRG −1.3±1.2 vs. outliers −2.0±1.5; BMI Z-scores NRG 0.03±1.08 vs. outliers −0.66±1.8). In addition to baseline blood samples, stimulated samples were collected at 60, 90, 120 minutes in 51 of 85 of the NRG subjects, and in all the outliers. In the remainder of the NRG, some of the stimulated blood samples were offset by 10-15 minutes, due to venous access issues or other technical problems, although at least three stimulated samples were obtained between 45 and 135 minutes.

NRG Group:

In the NRG cohort, copeptin values obtained at time 0 of the growth hormone stimulation test were normally distributed, with a mean of 9.1±4.0 pmol/L (range 3-21 pmol/L), median 9 (IQR 6,11.5). In the subset of 48 children in the NRG who had two serial baseline serum copeptin measurements, copeptin concentrations decreased significantly from 10.6±5.0 at −20 min to 9.1±4.0 pmol/L at 0 min (P<0.001, paired t-test). In every subject, the serum copeptin concentration increased from 0 min to a peak between 60-120 minutes (except for an earlier peak in subjects undergoing the ITT). The median peak was 22 pmol/L (IQR 15, 38), but the time and intensity of the peak in the NRG varied depending on the stimulating agents. Stimulated copeptin values in response to different agents are shown, for exemplification purposes, in a subset of 51 subjects of the NRG who had copeptin measured at the same time interval as the outliers (Table1). This subgroup had stimulated copeptin concentrations (median 21, IQR 13, 40) similar to those of the entire NRG.

Table 1:

Median copeptin response by stimulatory agent for a subset of the Normal Response Group

Stimulating Agent(s)	Copeptin (pmol/L)
	0 minutes	60 minutes	90 minutes	120 minutes
Arg-Clonidine (n=6)	9 [6,14]	11 [7,18]	12 [8,17]	14 [7,15]
Arg-Insulin (n=16)	9 [6,15]	12 [9,20]	23 [17,38]	16 [13, 24]
Insulin (n=8)a	8 [7,11]	10 [6,14]	-	-
Arg-L-Dopa/ Carbidopa (n=21)	9 [6,13]	17 [11,41]	24 [13,36]	30 [16,49]

Serum copeptin concentrations in the subset of the NRG (n=51) who had serial blood sampling at 0, 60, 90, 120 minutes (^aexcept for subjects tested with insulin alone, for whom only 0- and 60-minute values were available). Data displayed as median [interquartile range]. Abbreviations: Arg, arginine.

Outlier Group:

Among the outliers, baseline copeptin was measured, at 0 minutes only, in 5 subjects, and at −20 and 0-minute in 3 subjects. Serum copeptin concentrations at baseline (either −20 or 0 minute) spanned a wide range, from 40 to 1361 pmol/L, with a median value of 154 pmol/L (IQR 61, 439) (Figure 1). In the five outliers with a 0-minute value only, copeptin concentrations decreased gradually by 47-86% of the baseline values by 120 minutes (Table 2). In the three subjects who had additional serum copeptin measured at −20 minutes, the initial concentration declined by 67-95% of baseline by 0 minutes, and continued to decrease (two subjects), or remained steady (one subject), by 120 minutes. The largest incremental decrease (>80% of baseline values between the first and last sample) was observed in the 4 subjects with the highest (>200 pmol/L) copeptin values at baseline. Of note, in 3 of 8 outliers, there was a secondary small copeptin peak between 60 and 90 minutes, corresponding to the time points at which a copeptin peak response was observed in the NRG. In the NRG, 17 subjects (20%) were described as “anxious,” whereas a higher proportion of outliers (five subjects, 63%) were reported as “anxious”. The number of children with multiple IV attempts was small; this was described in six (7%) subjects categorized as NRG and only one (12.5%) subject in the outlier group. Though the latter subject was described as “calm”, his copeptin concentration was the highest at 1361 pmol/L on the blood sample obtained after the third attempt at phlebotomy. For comparison, the six children in the normal response group who underwent more than one IV attempt had non-elevated average copeptin values 5.8±2.0 pmol/L). Among the 16 exclusions, 14 subjects had baseline copeptin concentrations within 3 SD of the mean. Two subjects, however, had copeptin concentrations of 32 and 30 pmol/L at −20 minutes (>4SD of the NRG), which decreased slightly to 25 and 29 pmol/L, respectively, at 0 minutes.

Figure 1:

Serial serum concentrations of copeptin during growth hormone stimulation tests in 8 subjects with very elevated (≥40 pmol/L) serum copeptin values at baseline (outliers).

Table 2:

Characteristics and serum copeptin concentrations during GH testing for the Outlier group

Sex	Age (years)	Stimulating Agent(s)	Copeptin (pmol/L)					%decrease from 0 to 120 minutes	Anxiety
			−20 minutes	0 minutes	60 minutes	90 minutes	120 minutes
Male	11.5	Arg-Clon	-	62	42	39	33	47%	Not Present
Male	14.5	Arg-Insulin	-	471	130	140	71	85%	Present
Male	15.0	Arg-Insulin	-	362	65	87	48	86%	Present
Female	15.7	Insulin	-	202	51	40	29	86%	Present
Male	7.2	Arg-L-Dopa /Carbidopa	-	1361	282	355	194	86%	Not Present
Male	13.5	Arg-L-Dopa /Carbidopa	158	106	126	65	67	58%	Present
Female	11.9	Arg-L-Dopa /Carbidopa	71	61	33	26	28	61%	Not Present
Female	9.1	Arg-L-Dopa /Carbidopa	42	40	36	39	38	9%	Present

Abbreviations: Arg, arginine; Clon, clonidine

We also examined baseline cortisol levels (time 0 minutes), as a surrogate marker of stress. These were available for 44/85 subjects in the normal response group (NRG) and 6/8 subjects in the outlier group. The median cortisol value was 10.0 mcg/dL (IQR 7, 16) in the NRG and 23.0 mcg/dl (IQR 20,26.5) in the outlier group. We found a similar proportion of growth hormone deficiency in the NRG and the outliers (45/85 subjects in the NRG and 4/8 in the outliers).

We compared the creatinine values in the NRG (mean+/− SD 0.51+/−0.09) and the outliers (0.54 +/− 0.07). These creatinine values are clinically similar. This, along with the similar serum sodium values and absent history of polydipsia/polyuria, makes the possibility of nephrogenic diabetes insipidus unlikely.

Discussion

In the present series, we found that very elevated serum copeptin concentrations occurred within minutes after IV access in approximately 9% of children with no evidence of sodium or fluid disturbances following an overnight fluid restriction of at least 8 hours. Our findings are comparable to what has previously been reported on this phenomenon. Bonnet et al noted that 5 of 278 (2 %) of children had very elevated copeptin concentrations obtained after an 8-hour overnight fast, in the absence of fluid/sodium imbalances or acute illness(5). Additionally, Binder(6) identified 6 of 68 (~9%) children undergoing arginine stimulation tests who, at the beginning of the arginine stimulation test, had moderately elevated copeptin concentrations, which paradoxically decreased during arginine infusion. This suggests a consistent phenomenon which has clinical implications for the interpretation of serum copeptin concentrations in children.

The finding of elevated copeptin concentrations in up to 9 % of healthy children after phlebotomy(5-7) may be a consequence of pain, stress or anxiety related to the procedure. Preclinical data support this hypothesis. In one study, experimental rats were subjected to a painful stimulus not dissimilar to phlebotomy (pricking of the tail with multiple needlesticks over a minute). Following this stimulus, different groups of rats were sacrificed at time intervals from 30 seconds to 3 hours in order to assess the effect of pain on the supraoptic-neurohypophysial unit. (9) Within a few minutes of the painful stimuli, there was vasodilation and mobilization of neurosecretory material, with degranulation of vesicles from the hypothalamus-neurohypophysis into the lumen of the blood vessels. Further supporting this hypothesis, AVP/copeptin secretion is known to be triggered by a different group of pathologic conditions inducing physical stress, such as myocardial infarction, sepsis, pneumonia and acute neurological events.(2-4, 10) Although the etiology of the rise in copeptin during stressful conditions not associated with osmotic or volume imbalances is probably multifactorial, acute onset of pain is a likely contributor. Pain has been known to induce antidiuresis in rodents for decades, via an antidiuretic factor(11), likely AVP(12). In these animal studies, the increment in the antidiuretic agent/hormone occurs within minutes of the painful stimulus and is substantially more robust than the increase induced by osmotic stimulation. An increase in AVP has also been shown in humans, such as in patients presenting to an emergency department with acute pain(13) .The above studies suggest that pain per se may trigger AVP/Copeptin release, an effect that might be amplified by the inflammatory/hemodynamic changes associated with more complex pain situations, such as surgery(14), myocardial infarction(10), vaso-occlusive crisis in sickle cell disease(15). The relevance of copeptin as a marker of stress is discussed in detail in a recent review paper(16)

The issue of venipuncture as a stressful stimulus has been debated with relation to optimal blood drawing conditions for the diagnosis of pheochromocytoma/ paraganglioma due to the risk of a stress- related increase in catecholamines associated with phlebotomy(17-19). In adults, plasma epinephrine, metanephrine and normetanephrine concentrations were found to be higher in blood samples obtained by venipuncture, as compared to values obtained a few minutes earlier in the same subjects from an indwelling catheter(20). In a small study of children, epinephrine and norepinephrine decreased from the time of initial venipuncture with a butterfly needle to 30 min later, when blood was obtained from the indwelling IV line(21). In a preliminary report of children undergoing venipuncture from our work, we noted that both norepinephrine and copeptin values were significantly higher in the blood samples obtained 1-3 minutes after the initial venous access (−20 minutes), as compared to values obtained 20 minutes later (0 minutes) from an indwelling catheter(22) , which we have confirmed, with regard to copeptin concentrations, in the present, expanded series. These previous observations are in keeping with the concept that phlebotomy induces a noticeable stress and/or pain-related response in children in the ambulatory setting(23). However, the strikingly elevated serum copeptin increase (>40 pmol/l, and often >100 pmol/l) we observed in ~9% of children and adolescents undergoing phlebotomy for growth hormone stimulation tests appeared to be out of proportion to the pain or stress-related triggers associated with the procedure, yet suggests that in a subset of children, the stress related to the procedure of venipuncture can be so intense as to induce a rapid and exaggerated release of copeptin (and AVP) by the posterior pituitary.

The reasons subserving this uncommon phenomenon, which has been reported in other cohorts(5, 6), are difficult to characterize from the data from present series, also given the limited number of subjects. For example, though an extraordinarily high copeptin value, above 1000 pmol/L, was noted in a child who had blood obtained at the third attempt at IV placement, non-elevated copeptin values were measured in six children in the NRG undergoing multiple IV attempts. Possible variables promoting an exaggerated copeptin response in a minority of pediatric subjects include the intensity and perception of pain and psychological factors such as fear and anxiety about the procedure(24, 25). We noted anxiety to be slightly more prevalent in our group of outliers than in the NRG, but the data should be interpreted with caution given the small number of outliers and the subjective nature of this reporting by nursing staff. We also noted higher cortisol values in the outliers compared to the NRG. Cortisol is known to increase with acute stress via stimulation of the hypothalamic pituitary adrenal axis. Even though the difference in cortisol vales between the 2 groups cannot be interpreted statistically due to the small number of outliers, this may suggest that the subjects in the latter group were experiencing a high level of stress related to the procedure. The slow decline in serum copeptin in the 2 hours following the initial spike, not unlike what Binder et al. also noted in their “high-copeptin outliers”(6), is likely a function of the copeptin half-life (estimated to be longer than the AVP half-life, probably up to 20-30 minutes)(26). Interestingly, we observed a second smaller copeptin peak during the course of the 2 hours of the test in some subjects, which corresponded temporally to the time of the peak response to the stimulating agent in the NRG and suggested continuing ability of the AVP-secreting neurons to rapidly manufacture pre-pro-AVP and replete the pituitary (AVP and)-copeptin stores in soon after a large release of these compounds at the beginning of the test.

The strength of our study is that we have characterized, for the first time, the pattern of copeptin secretion over 120 min in “high-copeptin responders”, thus addressing what has been until now a puzzling observation(5, 6).

Our study has limitations. Despite the fairly large (>100) number of subjects we studied, the outliers represented a small group, which prevented us from performing meaningful statistical comparison testing between the NRG and outlier groups (an undertaking that would probably require doubling or tripling the current number of control subjects). With the limitation of the small N, however, the pattern of copeptin response was qualitatively very consistent in our outliers, albeit with quantitative differences. Another limitation to our study is that subjects were fluid-restricted overnight for 8 hours or longer and were predominantly males. Both factors may account for the slightly higher baseline copeptin concentrations in our NRG than those noted under similar conditions of overnight water restriction by Bonnet et al(5) (median 7.6 pmol/L for the entire cohort, 8.3 pmol/L for males) and Binder et al (median of 6 pmol/L)(6). However, they are similar to those reported by Tuli et al. in children (10.6 pmol/L), also after overnight fasting(27). The slight differences across these studies (in which copeptin was measured y the same assay) may relate to the duration of overnight fasting and/or the timing of blood sampling in relation to venipuncture (ranging from less than 1 to 30 minutes). Additionally, normative values for baseline copeptin concentrations in children depending upon fluid status are limited(28). Whether the stimulus of overnight water restriction may have favored greater storage and/or release of AVP/copeptin than would otherwise occur in subjects allowed ad libitum fluid intake until the time of venipuncture, cannot be inferred from the present study, given its design. Conversely, we do not know whether subjects with primary polydipsia and (possibly) chronically suppressed AVP are likely to show less more or less pronounced copeptin increments in the event of an exaggerated response to phlebotomy. These and other considerations may be relevant to the differential diagnosis of hypotonic polyuria. In this context, randomly elevated (>21.4 pmol/l) serum copeptin values have been considered diagnostic of nephrogenic diabetes insipidus (NDI) in adults(28). However, our data caution against relying on a single elevated copeptin value for the diagnosis of NDI in children with the polydipsia-polyuria syndrome, as a spurious elevation of copeptin following phlebotomy in subjects with primary polydipsia, a more common entity than NDI, could lead to an incorrect conclusion. Finally, an additional limitation is that we were unable to obtain a more standardized assessment of pain and stress, as this study materialized after several serendipitous and initially puzzling observations of very elevated copeptin concentrations, during the course of an unrelated (primary) study evaluating the copeptin response to GH-stimulating agents. Although our nurses have been documenting accurately the patients’ condition and behavior during these tests for many years, the notes provided only limited qualitative, and somewhat subjective, information at this regard.

In summary, substantially elevated copeptin values may be seen in a small subset (fewer than 10%) of healthy children following phlebotomy. Though a similar observation has been made in previous studies, our analysis provides a more detailed evaluation of the different pattern of copeptin response dynamics in this “high responder” group vis-a vis “average” responders. Though pain or acute stress are possible triggers for the exaggerated copeptin response, the available data are insufficient to elucidate the underlying mechanism. Regardless, recognition of this occurrence is important to avoid misinterpreting serum concentrations of copeptin in the pediatric age.

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.