Antidiuretic hormone release associated with increased intracranial pressure independent of plasma osmolality

Abstract Objective Introduce and evaluate a new model which explains the release of brain antidiuretic hormone (ADH) independent of plasma osmolality. Methods Systematic review and critical analysis of the professional literature. Results Primary electronic database searches using key terms revealed 57,432 hits. Secondary searches with application of specific inclusion and exclusion criteria and manual inspection for completeness reduced the total number of studies to fourteen (N = 14). Twelve (N = 12) studies investigated human subjects in the hospital settings, and two (N = 2) studies investigated animals (rhesus monkeys and dog) under invasive experimental conditions. All fourteen studies included direct or indirect indicators of intracranial pressure (ICP), measurements of plasma ADH, and plasma osmolality or urine osmolality. Findings, in brief, reveal a stable and positive association between increased intracranial pressure (ICP) and increased ADH release, in patients with low or normal blood osmolality. Findings are reliable and reproducible across human and animal populations. Conclusions Findings support the proposed model, which explains increase secretion of brain ADH when plasma osmolality is low or within normal limits. Mechanical pressures exerted on hypothalamic nuclei, especially paraventricular and supra‐optic nuclei, as a consequence of increased intracranial pressure, produce release of ADH, independent of plasma osmolality. The mechanical pressure model explains release of ADH previously unexplained by traditional plasma osmolality models. Findings have important clinical implications for the medical and surgical management of patients.

those who manage patients presenting with disturbance of electrolyte imbalances, for example hyponatremia, when blood osmolality values are low or within normal limits.
The existing body of experimental and clinical research literature fails to provide an adequate or unifying theory as to why increased intracranial pressure (ICP) is associated with increased secretion of ADH, when plasma osmolality is low or within normal limits. Our proposed model explains how increased ICP produces an increase in ADH release, offers a mechanism of action independent of plasma osmolality, and offers a theory based explanation as to why clinical conditions such as hyponatremia can develop in the presence of normal plasma osmolality.
Traditional models emphasize plasma osmolality as primary in initiating and maintaining ADH release. As plasma osmolality increases ADH is released, returning plasma osmolality to normal values. Over the past 70 years, investigators repeatedly report increases in ADH secretion when plasma osmolality is high. These findings are explained within the traditional osmolality models and are well understood.
However, in the presence of increased ICP, there is frequently a subgroup of patients who evidence increased secretion of ADH in the presence of low or normal plasma osmolality. Investigators periodically have offered explanations to account for these findings; however, none of the models venture far from the traditional osmolality model.
For example, nearly fifty years ago, Wise reported enlargement of the third ventricle is associated with increased ADH release. He attributes the increased ADH release to disruption of hypothalamic osmo-receptors (Wise, 1968).
Gaufin, Swowsky, and Goodman, working with rhesus monkeys and investigating nonosmotic triggers potentially responsible for ADH release, reported the magnitude of ADH release is directly associated with the magnitude of intracranial pressures. As intracranial pressures increase, ADH release increases. Gaufin, Swowsky, and Goodman are among the first to offer data implicating nonosmotic triggers and offer the foundation upon which the proposed model is based (Gaufin, Skowsky, & Goodman, 1977).
Current prevailing models explain the release of ADH as the result of increased osmotic pressure, produced by increased plasma osmolality. These models correctly and adequately explain the relationship between ADH release and plasma osmolality for healthy individuals and for most patients presenting with common electrolyte imbalances, such as hyponatremia. However, these models fail to adequately explain increased ADH release when plasma osmolality is low or within normal limits.
Our proposed model in brief states, mechanical pressure exerted on hypothalamic nuclei, specifically paraventricular and supra-optic nuclei, mechanically and directly stimulates firing of these neurons, resulting in the release of ADH. The mechanism of stimulation is completely independent of the typical and traditionally accepted mechanism of stimulation by way of increased blood osmolality and brain osmo-receptors.
Specific questions tested by this study are:

| Electronic databases
The scientific and medical literature was systematically searched electronically using the following four electronic databases: PubMed, Google Scholar, Embase, and OVID.

| Search strategy
An initial search was conducted using the specific key search terms.
A secondary search was then conducted with refined and specific inclusion and exclusion criteria. Results from the secondary search were independently reviewed by two investigators (EM,WK) for final selection. Reference lists in each article identified by the secondary search were examined manually and relevant articles identified in the reference lists and not identified by the electronic search were added to the final study list. Studies identified as meeting all inclusion and exclusion criteria and containing all essential information; for example, measurements of ICP, ADH, plasma osmolality, or urine osmolality were placed on the final study list and independently reviewed by two investigators (EM, WK) for completeness. Only studies meeting all inclusion and exclusion criteria and independently confirmed by both investigators were retained and subjected to statistical analysis and summary.

| Inclusion criteria
Studies appearing in the medical literature between the years 1954 and 2015 were included. Studies must have reported data containing measurements of ICP, ADH, plasma osmolality, or urine osmolality. It was not necessary for the study to specifically investigate the relationship between ICP, ADH, plasma osmolality, or urine osmolality, only that these measurements appear in the study and could be extracted. Studies were searched and extracted independent of subject's age, sex, or underlying patient pathology. Studies were searched and extracted independent of the language of the written article. Relevant non-English articles were translated into English using Google Translate. Human and animal studies were included in all searches and study extractions. Unpublished studies presented at professional meetings, which appear as published abstracts, and published articles appearing in the professional medical literature between the years 1954 and 2015 were included in all searches and study extractions.

| Exclusion criteria
Studies that did not report measurements of ICP, ADH, plasma osmolality, or urine osmolality were excluded.

| Data extraction
Studies meeting inclusion and exclusion criteria were compiled. The following data were extracted from each study: article title, article authors, year of publication, subject characteristics for example age, sex, and identified pathology if any, number of subjects in each study, number of subjects evidencing increased release of ADH, number of subjects demonstrating increased release of ADH when plasma osmolality were low or within normal limits, measurements of ICP, measurements of ADH, measurements of plasma osmolality, and measurements of urine osmolality.

| Data summary and statistical analysis
Data extracted from each study were summarized into two tables.

All fourteen (N = 14) studies included direct or indirect indicators
of increased intracranial pressure, for example direct measurement, computerized tomography, and clinical assessment. All studies included space occupying cerebral lesions, for example aneurysms of the anterior communicating artery, intracerebral tumor, cerebral hemorrhage, controlled balloon inflation, cerebral edema and affecting directly or indirectly the hypothalamic nuclei, particularly paraventricular and supra-optic nuclei. Studies included measurements taken immediately upon initial onset of increased intracranial pressure, as well as measurements taken several months following sustained, increased intracranial pressure.    (Gupta, 2015).
Empty cells indicate data were not reported or could not be extracted from the original study.
TA B L E 1 (Continued)   (Sørensen et al., 1985a). Data from the original study have been extracted, reorganized, and reanalyzed here.
See Table 1 (Study 6) and Figure 2.  F I G U R E 2 Plasma antidiuretic hormone as a function of intracranial pressure in eight adults with hydrocephalus. Mean ± SD.
Original data extracted and reanalyzed from Sorensen, Gjeeris, and Hammer (Sørensen et al., 1985a). See Table 1, Study 6. *Norm reference values from Merck Manual Professional Edition (Wians, 2015) Data originally reported by Brown and MacDonald examines increased ADH secretion associated with increased ICP in nine (N = 9) pediatric patients, presenting with Reye's syndrome. The study employed a repeated measure within group design. Plasma ADH was measured pretreatment high ICP conditions and posttreatment normal ICP conditions (Brown & MacDonald, 1981). Data from the original study have been extracted, reorganized, and reanalyzed here.

Does antidiuretic hormone secretion increase under conditions
of normal or low plasma osmolality? Yes.
Data originally reported by Sorensen, Gjerris, and Hammer examine the relationship between plasma ADH and plasma osmolality in adult patients, with and without increased ICP. The study employed a between independent group research design.
Plasma ADH and plasma osmolality were measured in 243 patients with assorted neurological disorders, psychiatric disorders, and patient controls (Sørensen et al., 1985b). Data from the original study have been extracted, reorganized, and reanalyzed here. See Table 1 (Study 5) and Figure 4.
The original data were reorganized into four independent groups, patient controls with no central nervous system or endocrine disorders (n = 52), psychiatric patient controls (n = 62), patients with no space occupying brain lesions (n = 38), and patients with space occupying brain lesions (n = 62). All patients included in the space occupying brain lesion group revealed increased ICP >18 mmHg, at the time plasma ADH sample were taken. The patient control group, psychiatric patient group, and nonspace occupying brain lesion group revealed normal ICP <15 mmHg, at the time plasma ADH samples were taken. Twenty-nine (n = 29) patients, presenting with normal pressure hydrocephalus and ICP <12 mmHg were excluded from the analysis, reducing the total number of patients and controls from 243 in the original study to 214. The normal pressure hydrocephalus patients were excluded from our analysis due to our inability to establish with certainly the underlying cause producing the NPH, for example nonspace occupying lesion or space occupying lesion. Plasma osmolality remained within normal limits for all four groups, at the time plasma ADH was sampled.
Statistical analyses reveal the space occupying brain lesion group releases significantly higher amounts of plasma ADH when Data originally reported by Yoshino, Yoshimi, Taniguchi, Nakamura, and Ikeda, a case study report of idiopathic normal pressure hydrocephalus, revealed high ADH secretion during a condition of increased ICP (blocked ventricular shunt), while plasma osmolality remained low and below normal limits (256 mOsm/kg) (Yoshino et al., 1999). See Table 1 (Study 11).
During the pretreatment (blocked shunt) condition, the patient's plasma ADH was significantly elevated (14.0 pg/ml) and above F I G U R E 4 Plasma antidiuretic hormone as a function of intracranial pressure, across three patient groups, when plasma osmolality remains within normal limits in all patient groups. Mean ± SD. Data extracted and reanalyzed from Sorensen, Gjerris, and Hammer (Sørensen et al., 1985b). See Table 1, Study 5. *Norm reference values from Merck Manual Professional Edition (Wians, 2015) F I G U R E 3 Plasma antidiuretic hormone as a function of intracranial pressure in nine children with acute encephalitis (Reye's syndrome) before and after treatment. Mean ± SD. Original data extracted and reanalyzed from Brown and MacDonald (Brown & MacDonald, 1981 (Cotton et al., 1993). Data from the original study have been extracted, reorganized, and reanalyzed here. See Table 1  The study employed an experimental within subject repeated measures research design (Gaufin et al., 1977). Data from the original study have been extracted, reorganized, and reanalyzed here. See Table 1  Mechanical pressure exerted on the hypothalamic nuclei results

| D ISCUSS I ON
in an increased release of ADH independent of plasma osmolality.
These results, independently and in combination support our hypothesis. Increased ICP stimulates specific hypothalamic nuclei, particularly paraventricular and supra-optic nuclei, resulting in the release of ADH independent of plasma osmolality. This is substantially different from traditional models, which rely exclusively on plasma osmolality as a stimulus for ADH release. The F I G U R E 5 Changes in urinary antidiuretic hormone as a function of normal, nonlethal, and lethal levels of intracranial pressures in four Rhesus monkeys. Mean ± SEM. Each animal served as its own control. Original data extracted ad reanalyzed from Gaufin, Showsky, and Goodman (Gaufin et al., 1977). See Table 2, Study 13 mechanical pressure model explains release of ADH, previously unexplained by traditional, plasma osmolality models. The proposed model has practical implications and applications for the practicing physician. For example, patients evidencing hyponatremia typically reveal high plasma osmolality. Interventions, for example water restriction, typically reduce ADH secretion and returns plasma osmolality to normal levels. However, in cases of increased ICP, water restriction often returns plasma osmolality to normal levels, while ADH continues to be released and the hyponatremia persist.
Patients evidencing increased ICP, low or normal plasma osmolarity, and continuing hyponatremia can now be understood, within a new evidence based model and necessary modifications made to standard treatment protocols.
This study has limitations. Specifically, many of the studies reviewed failed to report ICP and ADH within the same article. Not all the articles reviewed provided a direct measurement of ICP, ADH in plasma, ADH in CSF, ADH in urine, or plasma osmolarity. Some studies failed to report findings in standard units, choosing instead to report findings as a percentage change, making direct comparison between studies more difficult. While our study has limitations, the data and evidence revealed are persuasive.
Where do we go from here? In order to further advance our understanding and the professional knowledge base, it will be of value, whenever possible, to measure increased ICP, plasma ADH, and plasma osmolality simultaneously. It will of particular interest to investigate and measure plasma ADH and plasma osmolality in the specific patient population evidencing hyponatremia in the presence of increased ICP and normal plasma osmolality. Collecting and reporting these measurements in combination should contribute to further understanding the relationship in ADH secretion associated with changes in ICP and contribute to the effective clinical management and treatment of patients presenting with specific electrolyte imbalances resulting from increased ICP whether the cause is brain disease or brain trauma.

| CON CLUS IONS
The data extracted and reanalyzed from the professional literature, spanning the past 70 years, support the newly proposed model. ADH secretion increases in the presence of increased ICP.
Mechanical pressure as the primary mechanism of action producing ADH secretion is offered as an alternative to more traditional models emphasizing dysregulation of plasma osmo-receptors.
ADH secretion can and does occur independently of plasma osmolality regulatory mechanisms. Treatment interventions based upon prevailing medical dogma and proving ineffective can now be rationally modified.

ACK N OWLED G M ENTS
The authors thank the original investigators of the fourteen exemplary studies highlighted in this critical review.

CO N FLI C T O F I NTE R E S T S
None declared.