Polyuria and cerebral vasospasm after aneurysmal subarachnoid hemorrhage

We performed a retrospective chart review of all patients admitted with aSAH to our institution between January 2012 and December 2013. The study was approved by the Hartford Hospital Institutional Review Board (IRB). Patients were included if they had an aneurysmal subarachnoid hemorrhage and were admitted within 48 h of aneurysmal rupture (i.e. onset of symptoms). In order to avoid the situation where patients develop polyuria, but die before the onset of vasospasm, patients were excluded if they died within 5 days of admission. Patients were also excluded if they received mannitol after the first 24 h of admission or had a history of or developed diabetes insipidus as these both are independent causes of polyuria. Finally, patients were also excluded if they developed a significant acute kidney injury (AKI) during their hospital course as this could mask the development of polyuria due to decreased glomerular filtration rate. We defined AKI as a 2-fold increase in serum creatinine as per both the RIFLE criteria and stage 2 in the Kidney Disease Improving Global Outcomes (KDIGO) grading system [14, 15].

Patients with aSAH were treated according to the current management guidelines including early aneurysm securing procedures, enteral nimodipine, and maintenance of a euvolemic state. All patients had frequent measurement of their intake and output and a negative fluid balance was avoided by intravenous fluid boluses and titration of intravenous fluid maintenance rates unless symptoms of hypervolemia indicated otherwise as per the treating the team. Central venous pressures as well as pulse contour analysis were utilized to ensure euvolemia if clinically indicated. Patients were regularly evaluated for DCI and cerebral vasospasm by frequent neurologic evaluation and Transcranial Doppler Ultrasonography (TCD). TCDs were performed at baseline (by post-bleed day 2) and then every 1–2 days (as indicated) until at least day 10. Patients suspected of having DCI or cerebral vasospasm were evaluated by computerized tomography (CT) or catheter angiography, and if present underwent hemodynamic augmentation therapy. Patients not responding to medical therapy underwent endovascular therapy including intra-arterial calcium channel blockers and balloon angioplasty.

Data were obtained from an IRB approved patient registry as well as the patients’ electronic medical records. 24 h urine output was obtained from the nursing flow sheets. Polyuria was defined as urine output >6 liters in a day. Cerebral vasospasm was defined on TCDs as a mean velocity >120 m/s plus a Lindegaard ratio >3 and on angiography or CT angiography based on the attending neuroradiologist’s interpretation. For consistency, any documentation of mild, moderate, or severe vasospasm was interpreted as positive for vasospasm however, terminology such as “minimal narrowing” was not. The onset day for cerebral vasospasm was the first day any test was consistent with cerebral vasospasm as defined above. Urine output was obtained between days 2 and 10 of admission. However, if cerebral vasospasm was diagnosed on or after day 10, urine outputs were recorded up to 24 h after the day of vasospasm onset. The onset of hyponatremia was also defined as the first day the serum sodium was below 135 meq/L excluding days 1 or 2 of admission. As the main objective of this study was to identify a temporal relationship with cerebral vasospasm, we excluded sodium or urine outputs on the first day of admission to eliminate pre-admission factors or factors related to early resuscitation efforts, surgical procedures or osmotic agents.

In this study, we show that polyuria is common after aSAH and it is significantly associated with cerebral vasospasm. Polyuria tends to occur slightly earlier than cerebral vasospasm and the onset of vasospasm is clustered around the onset of polyuria. However, polyuria did not impact mortality after controlling for age and indicators of disease severity.

The incidence of vasospasm in this cohort is similar to prior studies [6, 22]. The incidence of polyuria was also high and this has not been previously reported. The majority of patients (66 %) developed polyuria on one specific day (excluding day 1) after aSAH. While we defined polyuria as a 24 h urine output above 6 liters, outside of the neurological intensive care unit polyuria has been relatively arbitrarily defined as a urine volume greater than 3 L within a 24 h period [23]. Using that definition would have been prodigiously liberal in our cohort and would have included 100 % of patients. Patients with aSAH are frequently maintained on intravenous fluids often in excess of 100 cc per hour. Given the additional enteral or oral intake, total fluid intake is rarely below 3 liters such that a urine output below 3 liters in the NICU is the exception rather than norm. Therefore, our definition of polyuria was more restrictive. However, our intention was not to show that a specific urine volume was associated with cerebral vasospasm, but rather to demonstrate the concept that an increase in or excessive volume of urine over baseline in any individual patient is a predictor of cerebral vasospasm. Based on our findings polyuria can serve as an easily obtainable clinical predictive factor to identify patients at risk for development of vasospasm.

Our study did not distinguish whether polyuria and vasospasm are independent phenomena that have similar etiologies or whether the onset of vasospasm contributes to polyuria or vice versa. The pathophysiology of excessive natriuresis and polyuria after aSAH is not completely understood however numerous studies have demonstrated an increase in natriuretic peptides including atrial natriuretic peptide (ANP) and brain natriuretic peptide (BNP) [12, 13, 24, 25]. Other less known peptides have been identified such as c-type natriuretic peptide and dendroaspis peptide [25, 26]. Hypothalamic dysfunction may be related to the severity of the SAH, both clinically and radiographically and could contribute to changes in the amount of these peptides released after SAH. A previous study found a higher incidence of hyponatremia with anterior communicating artery aneurysms and the putative mechanism was an increased incidence of hypothalamic dysfunction [27, 28]. We did not find an association between aneurysm location and the incidence of polyuria in this study. Regardless of the exact cause, the excessive secretion of natriuretic peptides could result in increased urinary sodium and volume loss leading to the clinical syndrome of cerebral salt wasting. However, there is considerable controversy regarding this pathophysiology. For example, the increase in natriuretic peptides may be a result of an associated cardiac injury which occurs in about 20 % of aSAH patients and is associated with vasospasm and DCI [29]. Importantly the association between polyuria and vasospasm held after controlling for well accepted measures of severity such as Hunt and Hess score, admission GCS and Fisher score in a multivariable model.

CSW is usually considered in the differential diagnosis of hyponatremia in aSAH patients, and is distinguished from SIADH by a clinical examination consistent with hypovolemia [30, 31]. However, as hypovolemia in aSAH patients is a known risk factor for cerebral infarction, aggressive measures are employed to prevent hypovolemia. Unlike SIADH in which isotonic fluids often exacerbate hypoanatremia, in CSW, replacement of renal losses with isotonic fluids may mitigate hyponatremia such that the only manifestation of this process is polyuria. Clinicians who frequently treat patients with aSAH are well-aware of sudden rises in urine output and anecdotally have associated this with cerebral vasospasm. While prior studies temporally linked hyponatremia with the onset of cerebral vasospasm, no study has tried to temporally link polyuria to cerebral vasospasm [8, 9]. Clinically, polyuria per se may be more useful than hyponatremia given that urine output is generally measured hourly in aSAH patients while sodium is usually measured daily. An increase in urine output may occur at any time during the day and may alert clinicians to the possibility of concomitant cerebral vasospasm. Furthermore, most centers try to avoid a negative fluid balance by titrating maintenance fluid rates in order to avoid hypovolemia. Therefore, making a diagnosis of cerebral salt wasting has become challenging for several reasons. First, hyponatremia is mitigated by the avoidance of a negative fluid balance and aggressive administration of saline. Second, even in patients who do develop hyponatremia, distinguishing CSW from SIADH is difficult given that the main distinguishing feature is volume status. Finally, even if attempts to avoid hypovolemia fail, as studies using pulse dye densitometry have shown, the bedside determination of volume status is often inaccurate [4, 32]. For these reasons, we believe that linking CSW to cerebral vasospasm is less clinically relevant than linking urine volume to vasospasm.

Our study is limited by its retrospective design. As stated above, it was not possible to completely rule out the possibility that some patients’ polyuria was a result of iatrogenic effects such as excessive fluid administration. While we excluded patients who received mannitol, we did not exclude patients who received hypertonic saline which could still have a mild diuretic effect although the relationship of polyuria to vasospasm persisted after controlling for this in multivariate analysis. Additionally, while urine output is a continuous parameter, the onset of cerebral vasospasm was limited to the timing of investigative tests. TCDs may not be performed daily in low risk patients without clinical changes and CTAs and conventional angiograms are often ordered only if vasospasm is suspected. Therefore, it is possible that vasospasm was missed in some patients or that the actual day of onset was earlier than recorded. It is also important to note that we chose to link polyuria to cerebral vasospasm rather than to delayed cerebral ischemia (DCI) which may be the more clinically relevant outcome than cerebral vasospasm [33]. This was not an oversight, but was intentional given the retrospective nature of this study, as linking polyuria to DCI would make it difficult to distinguish whether polyuria was a response to treatment. Once a patient is diagnosed with DCI, vasopressors are initiated and patients may receive additional fluid boluses, both of which may increase urine output. However, asymptomatic cerebral vasospasm is generally not treated with vasopressors or a change in fluid strategy. While our results may still be confounded by the patients who did develop DCI, by linking polyuria to mild vasospasm, which usually is not yet associated with ischemia, we believe the onset of polyuria coinciding with the onset of cerebral vasospasm was largely independent of iatrogenic issues. The retrospective design also made it difficult to determine who and when DCI was diagnosed as this is mainly an interpretation of chart notes whereas vasospasm was defined objectively by TCD changes and radiological findings. This study was also limited to the association with daily urine outputs. Future studies will examine hourly outputs and determine if precipitous changes in urine output are predictive of cerebral vasospasm. Future studies will also investigate whether the severity of vasospasm is associated with the incidence and degree of polyuria. Finally, while we did not detect an association of polyuria with mortality, clinical or radiological grade, or aneurysm location, a larger study may have detected an association as this initial study has limited power, despite being the largest to date on this topic.

In summary, we have shown that polyuria is significantly associated with cerebral vasospasm and that polyuria onset is clustered with the onset of cerebral vasospasm. This may be useful in determining which patients are at high risk of developing vasospasm and developing future recommendations for the monitoring and disposition of these patients. A prospective study is necessary to further evaluate this association.