| Objectives Serum sodium,calcium and magnesium at admission were associated with the poor outcome in acute ischemic stroke(AIS)patients,but the results were inconsistent from previous studies.Moreover,the impact of serum potassium,phosphate and chloride on clinical outcomes after AIS are still not understood.We investigated the association between various serum electrolytes(serum sodium,calcium,magnesium,potassium,phosphate and chloride)upon hospital admission and in-hospital mortality in AIS patients.Methods Data of 3,450 AIS patients enrolled from December 2013 to May 2014 across 22 hospitals in Suzhou city were used for the present studies.Laboratory variables,including serum electrolytes were assayed at local laboratories and we collected data on all-cause in hospital patient mortality.I divided patients into 4 groups according to their level of admission calcium,magnesium,potassium,and phosphate levels.Patients were also divided into with and without hypochloremia,and with and without hyponatremia according to admission serum sodium and chloride.Log-rank test and Cox proportional hazard model were used to examine the effect of serum electrolytes and in-hospital mortality.The predictive ability of different electrolytes on in-hospital mortality was compared by calculating C-statistics(areas under receiver operating characteristic [ROC] curves).Results(1)A total of 3,450 AIS patients(1993 male and 1457 female)were enrolled in the study.The mean age was 68.5 years(±12.9),and the median baseline NIHSS score was 4.0(2.0–7.0).During hospitalization,124 patients(3.6%)died from all causes.(2)2485 patients with serum magnesium data within 24 hours after hospital admission were included in our analysis.We divided patients into 4 groups according to their level of admission magnesium: Q1(<0.82 mmol/L),Q2(0.82-0.89 mmol/L),Q3(0.89-0.98 mmol/L)and Q4(?0.98 mmol/L).The in-hospital mortality from Q1 to Q4 of magnesium were 7.7%,2.8%,1.9% and 3.0%,and the mortality rates differed between the groups(log-rank P <0.001).After adjusting for age,sex,baseline National Institutes of Health Stroke Scale score,and other potential covariates,the lowest serum magnesium level(Q1)was associated with a 2.03-fold increase in the risk of in-hospital mortality in comparison to Q4(hazard ratio [HR] 2.03;95% confidence interval [CI],1.11–3.70;P-trend=0.014).Sensitivity and subgroup analyses further confirmed a significant association between lower magnesium levels and a high risk of in-hospital mortality.(3)3288 patients with available serum calcium data within 24 hours after hospital admission were enrolled in our analysis.I divided patients into 4 groups according to their level of admission calcium: Q1(<2.11 mmol/L),Q2(2.11–2.21 mmol/L),Q3(2.21–2.31 mmol/L)and Q4(≥2.31 mmol/L).The in-hospital mortality from Q1 to Q4 were 3.9%,3.9%,3.4%,2.8%,and no significant association was found between serum calcium and in-hospital mortality(log-rank P =0.663).The result remained similar after adjusting for age,sex,baseline National Institutes of Health Stroke Scale score,and other potential covariates(P-trend=0.986).(4)3314 patients with serum sodium and chloride data within 24 hours after hospital admission were enrolled in analysis.Hypochloremia was defined as having a serum chloride concentration<98 mmol/L and hyponatremia as having a serum sodium concentration<135 mmol/L.In-hospital mortality differed significantly in patients with(9.3%)and without(3.1%)hypochloremia.After adjusting for age,sex,baseline National Institutes of Health Stroke Scale score,and other potential covariates,patients with hypochloremia was associated with a 2.43-fold increase in the risk of in-hospital mortality in comparison to patients without hypochloremia(HR 2.43,95% CI: 1.41–4.19;P-value =0.001)and sensitivity and subgroup analyses further confirmed such association.However,the relationship between hyponatremia and in-hospital mortality was not significant in multivariable analysis(P-value =0.905).(5)3333 patients with serum potassium data within 24 hours after hospital admission were enrolled in the present analysis.I divided patients into 4 groups according to their level of admission potassium: Q1(<3.59 mmol/L),Q2(3.59–3.85 mmol/L),Q3(3.85–4.11 mmol/L)and Q4(≥4.11 mmol/L).The in-hospital mortality from Q1 to Q4 were 4.2%,2.4%,3.3%,4.4% respectively,and there was no significant association between admission potassium and in hospital mortality(log-rank P =0.113).The result remained the same after adjusting for age,sex,baseline National Institutes of Health Stroke Scale score,and other potential covariates.(6)3034 patients with serum phosphate data within 24 hours after hospital admission were enrolled in analysis.I divided patients into 4 groups according to their level of admission phosphate: Q1(<0.94 mmol/L),Q2(0.94–1.08 mmol/L),Q3(1.08–1.21 mmol/L)and Q4(≥1.21 mmol/L).The in-hospital mortality rates from Q1 to Q4 were 5.1%,4.0%,2.1%,3.7%,and were significantly different between the groups(log-rank P =0.021).After adjusting for age,sex,baseline National Institutes of Health Stroke Scale score,and other potential covariates,a lowest serum phosphate(Q1)(HR 2.00,95% CI: 1.10–3.63)but not highest serum phosphate(Q4)(HR 1.70,95% CI: 0.91–3.15)was associated with increased in hospital mortality in comparison to Q3.(7)Serum magnesium,chloride and sodium were all associated with in-hospital mortality and the AUCs were 0.623,0.604 and 0.580 respectively.The predict ability of in-hospital mortality of serum magnesium appears better than other electrolytes.Conclusions Decreased serum magnesium levels and hypochloremia at admission were independently associated with in-hospital mortality in AIS patients.There was a U-shaped association between phosphate and all-cause mortality with significantly increased risk among patients with lower phosphate levels.However,no significant associations between serum sodium,calcium and potassium and in hospital mortality were found.The predict ability of in-hospital mortality of serum magnesium was better than other electrolytes. |
PDF Full Text Request