不同血清磷水平对综合性医院住院患者和慢性肾脏病患者病死率的影响

Impacts of serum phosphate on mortality in hospitalized patients and patients with chronic kidney disease in a general hospital

目的探讨不同血清磷水平对综合性医院住院患者和慢性肾脏病(CKD)患者病死率的影响。方法收集2014年10月1日—2015年9月30日复旦大学附属中山医院所有住院患者入院时的血清磷水平和临床资料(年龄、性别、身高、体重、合并症、血常规、尿常规、肝和肾功能、电解质、血糖、院内死亡、死亡原因等)。按照血清磷水平分为6组:≤0.46、0.47~0.79、0.80~1.12、1.13~1.45、1.46~1.78、≥1.79mmol/L组。采用Logistic回归分析影响住院患者和CKD患者死亡的危险因素。结果共纳入57 786例住院患者,总体病死率为1.2%(678例)。低磷血症(<0.80mmol/L)患者占6.5%,高磷血症(>1.45mmol/L)患者占6.6%,低磷血症和高磷血症患者的病死率均显著高于血清磷正常的患者(P值均<0.001)。低磷血症发病率前3位的科室分别为肝肿瘤外科(25.3%)、肝肿瘤内科(14.3%)和消化科(10.4%),高磷血症发病率前3位的科室分别为肾内科(34.6%)、内分泌科(13.6%)和血液科(13.5%)。血清磷水平≤0.46mmol/L占0.64%(367例)、0.47~0.79mmol/L占5.90%(3 408例)、0.80~1.12mmol/L占43.71%(25 256例)、1.13~1.45mmol/L占43.16%(24 942例)、1.46~1.78mmol/L占5.21%(3 009例)、≥1.79mmol/L占1.39%(804例),各组的病死率分别为7.63%、3.17%、1.05%、0.69%、1.23%、8.33%。以病死率最低的1.13~1.45mmol/L组为对照组,校正年龄、性别、收缩压、舒张压、空腹血糖、白蛋白、血红蛋白、白细胞计数、估算的肾小球滤过率、血电解质(血钠、血钾、血氯、血钙、血镁)和合并症(高血压、糖尿病、冠状动脉性心脏病、恶性肿瘤、脑卒中、肝硬化、心力衰竭、急性肾损伤、CKD)后,血清磷水平≤0.46mmol/L(OR=3.071,95%CI为1.797~5.247)、0.47~0.79mmol/L(OR=1.900,95%CI为1.410~2.561)、0.80~1.12mmol/L(OR=1.319,95%CI为1.056~1.648)和≥1.79mmol/L组(OR=2.292,95%CI为1.470~3.574)的死亡风险均显著高于对照组(P值均<0.05),而血清磷水平1.46~1.78mmol/L(OR=1.144,95%CI为0.755~1.736,P=0.525)并不与住院患者病死率独立相关。对住院患者中11 392例CKD患者进行同样的血清磷水平分组,各组的病死率分别为13.68%、6.80%、2.47%、1.84%、2.84%、9.16%。以病死率最低的1.13~1.45mmol/L组为对照组,同样校正除CKD外的上述因素后,血清磷水平≤0.46mmol/L(OR=2.739,95%CI为1.203~6.235)、0.47~0.79mmol/L(OR=1.794,95%CI为1.146~2.808)、0.80~1.12mmol/L(OR=1.539,95%CI为1.093~2.168)和≥1.79mmol/L组(OR=2.759,95%CI为1.673~4.548)的死亡风险均显著高于对照组(P值均<0.05),而血清磷水平1.46~1.78mmol/L(OR=1.247,95%CI为0.726~2.141,P=0.423)并不与CKD患者病死率独立相关。结论住院患者中血清磷水平异常较为常见,低磷血症和严重的高磷血症是住院患者和CKD患者死亡的独立危险因素。

Objective To explore the relationship between serum phosphate concentration and mortality in hospitalized patients and patients with chronic kidney disease （CKD） in a general hospital. Methods All admissions to Zhongshan Hospital, Fudan University from October 1, 2014 to September 30, 2015 were screened. Serum phosphate level and relevant clinical data including age, sex, height, weight, comorbidities, blood routine, urine routine, hepatic-renal function, electrolytes, fast blood glucose, hospital deaths and death causes were extracted from the electronic database. All admissions were divided into six groups by admission phosphate levels as follows：0.46 mmol/L or less, 0.47 to 0.79 mmol/L, 0.80 to 1. 12 mmol/L, 1. 13 to 1.45 mmol/L, 1.46 to 1.78 mmol/L, and 1.79 mmol/L or higher. A multivariable logistic regression model was applied to assess the association between serum phosphate level and in-hospital mortality in hospitalized patients and OKD patients. Results A total of 57 786 patients were included in this study. The overall mortality was 1.2% （678 patients）. Hypophosphatemia （serum phosphate level was less than 0.80 mmol/L） was found in 6.5% of all patients and hyperphosphatemia （serum phosphate level was higher than 1.45 mmol/L） was noted in 6,6% of all patients. The mortality in patients with hypophosphatemia and hyperphosphatemia was significantly higher than that in patients with normal serum phosphate （P〈0. 001 ）. The top three departments of hypophosphatemia were surgery department of liver neoplasms （25. 3%）, internal medicine department of liver neoplasms （14. 3%） and department of gastroenterology （ 10.4 %）. The top three departments of hyperphosphatemia were department of nephrology （ 34. 6%）, department of endocrinology （ 13. 6%） ~nd department of hematology （13.5%）. Distribution of phosphate levels was as follows.. 0. 46 mmol/L or less in 367 patients （0. 63%）, 0. 47 to 0.79 mmol/Lin 3 408 patients （5. 90%）, 0. 80 to 1. 12 mmol/L in 25 256 patients （43. 71%）, 1. 13 to 1.45 mmol/L in 24 942 patients （43116 %）, 1.46 to 1.78 mmol/L in 3 009 patients （5.21%）, 1.79 mmol/L or higher in 904 patients （1.39%）. The mortality rates of each group were 7.63%, 3. 17%, 1.05%, 0.69%, 1.23% and 8.33%, respectively. The lowest in-hospital mortality （0.69%） occurred in patients with phosphate levels between 1.13 and 1.45 mmol/L, which was defined as a reference group. Adjusting for age, sex, systolic blood pressure, diastolic blood pressure, fasting blood glucose, albumin, hemoglobin, white blood cells, estimated glomerular filtration rate （eGFR）, admission electrolytes （including serum sodium, potassium, chlorine, calcium, magnesium）and comorbidities （including hypertension, diabetes mellitus, coronary disease, malignancy, stroke, liver cirrhosis, heart failure, CKD and acute kidney injury）, it was revealed that, compared with the reference group, there were higher death risks in the patients with phosphate level of 0.46 mmol/L or less （OR = 3.071, 95%CI 1. 797-5. 247）, 0.47 to 0.79 mmol/L （OR = 1. 900,95%C/ 1. 410 -2. 561）, 0.80 to 1.12 mmol/L （OR=1.319, 95%CI 1.056-1.648） and 1.79 mmol/L or higher （OR=2.292, 95%Cl 1.470- 3. 574） ~ while the phosphate level of 1.46 to 1.78 mmol/L was not independently associated with in-hospital mortality in hospitalized patients （OR = 1. 144, 95% CI O. 755- 1. 736）. There were 11 392 patients with OKD. They were divided into six groups by the same distribution of phosphate levels in all the hospitalized patients, and the mortality rates of each group were 13.68%, 6.80%, 2.47%, 1.84%, 2.84% and g. 16%, respectively. The lowest in-hospital mortality （1.84%） occurred in the patients with phosphate level of 1. 13 to 1.45 mmol/L which was defined as a reference group. Adjusting for the same factors above,the death risk in the groups of the phosphate level of 0.46 mmol/L or less （OR = 2. 739, 95% CI 1. 203 - 6. 235）, 0.47 to 0.79 mmol/L （OR = 1.794, 95%CI 1. 146-2.808）, 0.80 to 1.12 mmol/L （OR= 1.539, 95%CI 1.093-2. 168） and 1.79 mmol/L or higher （OR = 2. 759, 95% CI 1. 673- 4. 548） were significantly higher than that in the reference group, but aphosphate level of 1.46 to 1. 78 mmol/L was not independently associated with in-hospital mortality in OKD patients （OR = 1.247, 95% Cl 0. 726 - 2. 141）. Conclusion Abnormal serum phosphate is common in hospitalized patients. Hypophosphatemia and severe hyperphosphatemia are independent risk factors of death in hospitalized patients and CKD patients.