An observational study of the association between microalbuminuria and increased N-terminal pro-B-type natriuretic peptide in patients with subarachnoid hemorrhage

Patients

The study protocol was approved by the Institutional Research and Ethics Committee.
In accordance with the Helsinki Declaration, informed consent was obtained from all
patients or the patients’ next of kin. Between July 2008 and June 2010, we prospectively
studied 61 consecutive patients who were admitted to the Nagasaki Rosai Hospital within
48 h after SAH onset. The exclusion criteria included the presence of renal dysfunction
(serum creatinine level 1.2 mg/dL), which affects ACR values regardless of changes
in glomerular permeability 10]; no treatment administered for a ruptured aneurysm, because of poor neurological
condition (Hunt and Hess grade V) or either because the aneurysm was not verified
by cerebral angiography; and continuous postoperative barbiturate coma, which did
not allow the determination of a Glasgow Coma Scale (GCS) score. An initial computed
tomography (CT) scan of the head was performed on admission. The diagnosis was established
on the basis of the CT scan findings or by observing xanthochromia of the cerebrospinal
fluid when results of the CT scan were negative. Thereafter, cerebral four-vessel
angiography and/or three-dimensional CT were performed. Surgical clipping of the aneurysm
or endovascular surgery was performed at the earliest possible time, as determined
by a team of neurosurgeons experienced in both treatment modalities. Patients underwent
surgery for ruptured aneurysms within 72 h after stroke. Anesthesia was induced with
propofol and fentanyl and maintained with sevoflurane, fentanyl, and vecuronium bromide.

Clinical management

Postoperatively, all patients received conventional brain-oriented intensive care
therapy according to clinical requirements. Normovolemia was maintained through the
systemic administration of an intravenous physiological electrolyte solution at 1000–2000 mL/day
and a colloid solution as needed. Symptomatic hypovolemia was corrected by the administration
of additional fluids, followed by a continuous infusion of dopamine, as needed—to
maintain a systolic arterial blood pressure of 100 mmHg. The systolic arterial blood
pressure was maintained either at 160 mmHg in normotensive patients or at 180 mmHg
in hypertensive patients via continuous infusion of nicardipine as needed. Caution
was taken to avoid hyperthermia (38.5 °C) and hyperglycemia (150 mg/dL) through
the administration of diclofenac sodium and insulin as needed. In mechanically ventilated
patients, the arterial carbon dioxide partial pressure was maintained at 35–40 mmHg,
and the peripheral oxygen saturation was maintained at 97 %—under sedation with midazolam
if required. Oral and enteral nutrition was initiated as soon as possible. A CT scan
was performed in clinically deteriorating patients to identify secondary complications
such as hydrocephalus or ischemia. Hypertensive hypervolemic hemodilution (triple-H)
therapy was not used for prophylaxis. Hydrocephalus was treated with ventriculostomy.
Delayed cerebral ischemia was suspected in the event of neurological deterioration
(two-point decrease in the GCS score and/or focal neurological deficits) and confirmed
by cerebral angiography 11]. Hypertensive therapy was reinforced with dobutamine in each vasospasm episode. The
target systolic blood pressure was 160–180 mmHg. When technically possible, angioplasty
was performed in patients who failed to respond to therapy.

Data collection

The following clinical data was collected: age, sex, previous medical history of hypertension,
ischemic heart disease, heart failure and diabetes mellitus, size and location of
the aneurysm, and systolic blood pressure. The location of the ruptured aneurysms
was classified according to their presence in the anterior or posterior circulation
12]. The aneurysm size was categorized as large (12 mm) or other 12]. The neurological condition on admission was scored according to the Hunt and Hess
classification 13]. The amount of blood observed on the initial CT scan was graded according to the
Fisher classification 14]. The GCS score 15] and systemic inflammatory response syndrome (SIRS) score 16] were calculated at admission and daily for seven postoperative days. All physiological
variables were recorded as the most abnormal value obtained in a day, except those
recorded on admission. Arterial blood was sampled at admission and for seven postoperative
days to determine the PaO
₂
/F
_I
O
₂
ratio, C-reactive protein (CRP) level, troponin I level, and NT-pro-BNP level. Serum
troponine I levels were assayed by using chemiluminescence immunoassay (Architect
Troponin-I, Abbott Laboratories, Abbott Park, IL) with an applied threshold of 0.30 ng/mL
for the diagnosis of myocardial necrosis (a lower limit of detection of 0.01 ng/mL;
coefficient variation, 10 %). Serum NT-pro-BNP levels were assayed by performing electrochemiluminescence
immunoassay (Roche Diagnostics, Indianapolis, IN) with an applied threshold of 125 pg/mL
(a lower limit of detection of 5 pg/mL; coefficient variation, 10 %). The ACR and
vanillylmandelic acid/creatinine ratio (VMACR) measurements were determined from urine
samples collected at admission and daily for seven postoperative days. We assayed
urine vanillylmandelic acid by using high-performance liquid chromatography with a
lower limit of detection of 0.05 mg/dL and a coefficient variation of 2.9 %. VMACR
values of 5 mg/g were considered normal. We quantitatively assessed urine albumin
by using an immunonephelometric method (N-antiserum albumin, Dade Behring, Liederbach,
Germany), and we quantitatively measured urine creatinine by performing an enzymatic
colorimetric test. The sensitivity limit for urine albumin was 2.3 mg/L; for statistical
analyses, values below this limit were considered to be 0 mg/L. Both inter-assay and
intra-assay coefficient variations were within 5 %. Generally, clinically relevant
proteinuria is defined as an ACR of ?300 mg/g and microalbuminuria, as an ACR of 30–299 mg/g;
values 30 mg/g are considered normal. A medical chart review was conducted to determine
the duration of mechanical ventilation, intensive care unit stay, and hospital stay.
The neurosurgeons used the Glasgow Outcome Scale (GOS) 17] to assess the neurological outcomes at hospital discharge. Neurological outcomes
were stratified as unfavorable (GOS score of 1–3: death, persistent vegetative state,
and severe disability) or favorable (GOS score of 4 or 5: moderate disability and
good recovery) 1], 18]. In the present study, the intensive care unit care providers (neurosurgeons and
nursing staff) were blinded to the ACR and NT-pro-BNP values. The authors were the
patients’ anesthesiologists who were responsible only for blood and urine sampling
and the aforementioned assays.

Statistical analysis

Data distributions for the quantitative variables were expressed as median (interquartile
range). Intergroup comparisons were performed using the Mann-Whitney U test. Dichotomous variables were analyzed with the Fisher exact probability test
or the chi-squared test. To assess the predictive value for neurological outcomes,
receiver operating characteristic (ROC) curves were constructed for each variable
and the area under the ROC curve (AUC) was determined. Predictive variables were considered
significant predictors when the AUC was 0.8. The threshold value, sensitivity, specificity,
and likelihood ratio of the ACR were compared with those of some significant predictors
of unfavorable neurological outcomes. Multivariate logistic regression analyses were
performed, including significant predictors and dichotomized by its threshold value
base on ROC analyses with stepwise elimination of valuables not contributing to the
model (p??0.1). Associations between the ACR or the NT-pro-BNP level and other variables
were evaluated using the Spearman rank correlation test.

The sample size of 22 patients with SAH per group (standard deviation, 64 mg/g) was
determined on the basis of our previous study 1], which reflected a power of 90 % to detect a 100 % difference in the highest ACR
between the two neurological outcomes at a significance level of 5 %.