Sex differences in clinical characteristics and outcomes after intracerebral haemorrhage: results from a 12-month prospective stroke registry in Nanjing, China

The Nanjing First Hospital stroke registry (NFHSR) was a hospital-based registry of stroke patients with 1-year prospective follow-up. Nanjing First Hospital is the only tertiary hospital in the Qinhuai District in Nanjing City (the Southeast urban region of China), there are 700,000 citizens in this district, and more than 80% stroke patients in this area are admitted to our hospital. All consecutive patients, admitted to the neurology department with acute stroke (ischemic stroke, intracerebral haemorrhage or subarachnoid haemorrhage) between August 2004 and August 2008, were invited to participate. In Nanjing First Hospital, more than 95% ICH patients are admitted to Neurology wards, and less than 5% ICH patients with larger hematoma volume and requirement of surgery were admitted to Neurosurgery ward. Patients were prospectively registered and followed up for 12 months after verbal or written informed consent were gained from each participant or their legal surrogate. Clinical diagnosis of stroke was made according to the WHO criteria and was confirmed by brain CT or MRI scans [16]. The present analyses included patients with acute ICH within 14 days of symptom onset who had a complete 12-month follow-up. In addition, primary intraventricular haemorrhage was excluded in these analyses, but intraventricular haemorrhage secondary to ICH was included. NFHSR was approved by the Nanjing First Hospital Ethics Committee for studies involving human subjects.

Detailed information on patient demographics (age and gender), risk factors, pre-stroke dependency, onset-to-door time (ODT), stroke severity, laboratory tests, and brain imaging data were collected. Risk factors included smoking and alcohol consumption status, presence of hypertension, diabetes and hypercholesterolemia, history of stroke, ischemic heart disease (including myocardial infarction and angina pectoris) and atrial fibrillation. Patients who had smoked >1 cigarette/day for more than one year were defined as current smokers. Patients who had consumed alcohol >50 ml/day for at least one year were classified as current drinkers. Hypertension was defined as systolic blood pressure (SBP) >140 mm Hg and/or diastolic blood pressure (DBP) >90 mm Hg or current use of antihypertensive agents. Diabetes was defined as fasting serum glucose level >7.0 mmol/L or current use of antidiabetic drugs. Hypercholesterolemia was defined as fasting serum cholesterol level >5.72 mmol/L or current use of lipid-lowering medicines. Pre-stroke dependency was defined as a modified Rankin Scale (mRS) score of 3-5 [17]. Severity of stroke was evaluated using the Glasgow Coma Scale (GCS) [18] and the National Institutes of Health Stroke Scale (NIHSS) [19]. Location of bleeding was defined as deep cerebral (periventricular white matter, caudate, globus pallidus, putamen, internal capsule, and thalamus), lobar, cerebellar and brain stem [20].

Participants were followed up at 3, 6 and 12 months after ICH by telephone interview or questionnaire. Death was defined as the cumulative all-cause death. Dependency was defined as a mRS score of 3-5 [17].

The primary outcomes were death or dependency at 3, 6 and 12 months after ICH. The secondary outcomes were each component of the primary outcome (death and dependency separately) at 3, 6 and 12 months.

Figure 1 shows the flow chart of patients' selection. Over the 5-year study period, 3412 consecutive participants with acute stroke were included in the NFHSR. After exclusion of patients who presented after 14 days of symptom onset and those with missing baseline data, 651 ICH patients were followed up for 12 months. Thirty six (5.5%) cases were lost to follow up, and the remaining 615 (94.5%) who had a complete 12 months follow-up were enrolled for the present analysis.

Table 1 shows the baseline clinical characteristics of the 615 ICH patients (220 [35.8%] women and 395 [64.2%] men). There was no significant difference in age (63.5 ± 14.0 vs. 62.7 ± 12.7, p = 0.500) between women and men. However, compared to the male group, the female group had a lower proportion of current smokers (3.6% vs. 46.1%, p < 0.001) and current drinkers (3.6% vs. 35.4%, p < 0.001). In addition, women had higher NIHSS scores (10 [4-17] vs. 8 [3-15], p = 0.039) than men. There were no statistical differences in pre-stroke dependency (11.8% vs. 9.1%, p = 0.286), ODT (6 [3-24] vs. 7 [3-48], p = 0.508), location of bleeding (P = 0.566) and other baseline clinical characteristics between women and men.

The outcomes at 3, 6, and 12 months after ICH are shown separately for women and men in Table 2.

Compared with men, women had higher rates of death or dependency at 3, 6 and 12 months after ICH (61.2% vs. 46.8%, P = 0.001; 56.7% vs. 45.3%, P = 0.009; and 51.8% vs. 44.1%, P = 0.065; respectively), but the gender differences in death or dependency attenuated gradually over the period of 12 months (Table 2 and Figure 2). After adjustment for potential confounding factors such as age, existing hypertension and diabetes, prior stroke, previous ischemic heart disease, previous atrial fibrillation, current smoking and alcohol consumption status, pre-stroke dependency, ODT, admission SBP and location of bleeding, the sex difference in death or dependency remained significant at 3 months [odds ratio (or): 1.85; 95% confidence interval (CI): 1.22-2.81] and 6 months (OR: 1.64; 95% CI: 1.09-2.47), but did not reach statistical significance at 12 months (OR: 1.34; 95% CI: 0.90-2.00) (Table 2). However, when additionally adjusted for the admission NIHSS, these associations remained significant only at 3 months (OR: 1.76; 95% CI: 1.07-2.89).

There were no clear differences in mortality between women and men (Table 2). In contrast, dependency was more frequently observed in women than in men at 3, 6 and 12 months after ICH (45.5% vs. 29.8%, 38.9% vs. 25.3% and 32.1% vs. 21.1%, respectively). These differences in dependency remained significant after adjustment for the same potential confounders such as age, existing hypertension and diabetes, prior stroke, previous ischemic heart disease, previous atrial fibrillation, current smoking and alcohol consumption status, pre-stroke dependency, ODT, admission SBP and location of bleeding at 3 months (OR:2.19; 95% CI:1.31-3.65), at 6 months (OR:2.17; 95% CI:1.28-3.66), and at 12 months (OR:1.76; 95% CI:1.03-3.01). However, when additionally adjusted for admission NIHSS, these associations remained significant only at 3 months (OR: 1.94; 95% CI: 1.11-3.42) and 6 months (OR: 2.02; 95% CI: 1.16-3.53).

In our study, the female group had a lower proportion of current smokers and current drinkers than the male group, which is consistent with previous reports on the Chinese general population [21] and other hospital-based ICH study [11]. However, in contrast to prior investigations, our study demonstrated that women had more serious forms of ICH [11],[14]. These discrepancies between the results of studies may be attributable to difference in ethnicity, study setting and inclusion criteria.

NFHSR has a number of strengths including the prospective design, 12-month follow-up, high follow-up rate (94.5%) and relatively large sample size (615 ICH patients). These strengths enabled us to explore the sex differences in outcomes among Chinese patients with ICH, which were not well-explored previously.

Some limitations should also be noted. First, NFHSR is a single-centre, urban setting, and teaching hospital-based stroke registry, not all ICH patients in the study area and in the hospital were covered. Furthermore, some patients with missing data or lost to follow up were excluded from the analysis. Therefore, there may be some selection bias in our registry and our findings may not be generalizable to the general ICH patients in China, especially those in rural areas. Second, there was no information on ICH hematoma volume in this registry. So we did not know if the hematoma volume had also influenced our results. Finally, there was a paucity of data on the management and care of ICH patients, which may have influenced the gender differences in ICH outcomes.