KEY POINTS
Question: Is preoperative hyperchloremia or hypochloremia associated with mortality and morbidity after noncardiac surgery?Findings: Preoperative hypochloremia and hyperchloremia were related to increased 90-day mortality after noncardiac surgery, and preoperative hypochloremia was related to an increased risk for postoperative acute kidney injury.Meaning: Preoperative serum chloride level is a valuable test variable that exhibits an independent relationship with mortality and morbidity after noncardiac surgery. Serum chloride is the most common anion in the plasma and interstitial fluid. It plays an important role in determining the tonicity of plasma and influences acid-base exchange.1 , 2 3 4–7 6 , 8 9 10 4 , 5
McCluskey et al11 −1 ).12 11 6 , 8
In this study, we aimed to evaluate the relationship between preoperative hyperchloremia or hypochloremia and postoperative 90-day mortality and morbidity in patients undergoing noncardiac surgery. We hypothesized that preoperative hyperchloremia and hypochloremia would both increase postoperative 90-day mortality and morbidity.
METHODS
The present investigation was a retrospective cohort study conducted with approval from the Institutional Review Board of Seoul National University, Bundang Hospital (Seoul, Korea) (approval number, B-1710/424-107; approval date, April 13, 2018). The requirement for informed consent was waived by the institutional review board, considering the retrospective cohort design of the study, which targeted patients who had previously undergone treatment, and the use of anonymized patient data.
Patients
Medical records of all patients >20 years of age who underwent surgical procedures at the Seoul National University, Bundang Hospital between January 2010 and December 2016 were included in the analysis. If a patient underwent >1 surgical procedure during the study period, only the most recent case was included in the analysis. Patients who underwent previous cardiac surgery (including major vascular surgery with intraoperative cardiopulmonary bypass) or those with incomplete or missing medical records regarding preoperative chloride levels were excluded.
Preoperative Normochloremia, Hyperchloremia, and Hypochloremia Groups (Independent Variable)
For most elective and emergency surgeries at Seoul National University, Bundang Hospital, the chloride (mmol·L−1 ) level is usually assessed in ≤1 month before the surgery in routine preoperative laboratory tests. Chloride levels determined in ≤1 month before surgery were categorized as preoperative normochloremia (97–110 mmol·L−1 ), hypochloremia (<97 mmol·L−1 ), or hyperchloremia (>110 mmol·L−1 ).
Measurements (Potential Covariates)
Demographic information (sex, age [year], body mass index [kg·m−2 ]), information regarding preoperative comorbidities (American Society of Anesthesiologists physical status, estimated glomerular filtration rate [mL·minute−1 ·1.73 m−2 ], hypertension, diabetes mellitus, liver disease [hepatocellular carcinoma, liver cirrhosis, and fatty liver], history of ischemic heart disease and cerebrovascular disease, known heart failure, cancer, history of diuretic use [furosemide and thiazide], preoperative serum sodium level [mmol·L−1 ]), and surgery-related information (emergency surgery, duration of surgery,11 13 −1 .154 × (age)−0.203 (× 0.742 if female). Total fluid input (mL) on the day of surgery was determined for balanced crystalloid (Ringer’s lactate solution and plasmalyte), 0.9% NaCl, 0.45% NaCl, hydroxyethyl starch, and 5% albumin. Like chloride, sodium (mmol·L−1 ) levels are usually assessed in ≤1 month before the surgery in preoperative routine laboratory tests. All patients were categorized as having preoperative normonatremia (135–145 mmol·L−1 ), hyponatremia (<135 mmol·L−1 ), or hypernatremia (>145 mmol·L−1 ). All of the information was obtained from a preanesthetic registry and a surgical registry by medical record technicians blinded to the purpose of the study.
Postoperative 90-Day Mortality and Acute Kidney Injury (Dependent Variable)
To determine whether 90-day mortality was reflective of not only surgery-related death but also perioperative quality of care,14 , 15
Second, the Kidney Disease: Improving Global Out comes criteria and grading were used for the diagnosis of acute kidney injury, which is a primary morbidity in postoperative patients.16 −1 ), measured within ≤1 month before surgery, an increase in the postoperative serum creatinine level by ≥1.5 times or ≥ 0.3 mg/dL within ≤1 week of surgery was defined as postoperative acute kidney injury. At Seoul National University, Bundang Hospital, the preoperative serum creatinine level is measured as a routine test, and postoperative serum creatinine testing was performed according to the nature of the surgery or the status of the patient.
End Point
The primary end point of this study was differences in postoperative 90-day mortality among the preoperative chloride groups. The secondary end point was the difference in postoperative acute kidney injury incidence among the preoperative chloride groups.
Statistical Analysis
For comparison of characteristics among the 3 groups (ie, normochloremia, hypochloremia, and hyperchloremia), a 1-way ANOVA test was used for continuous variables, and the χ2 test was used for categorical variables. Next, overall survival time according to preoperative chloride groups was presented in the form of Kaplan-Meier curves, and χ2 tests were used to compare 90-day mortality between preoperative chloride groups. Before performing Cox regression analysis for postoperative 90-day mortality and logistic regression analysis for postoperative acute kidney injury, log odds of postoperative 90-day mortality and acute kidney injury according to preoperative chloride (main independent variable) and sodium (most important covariate) were analyzed using restricted cubic splines to assess the linearity of the variables as shown in Supplemental Digital Content, Figures 1–2, https://links.lww.com/AA/C671
Univariable Cox regression analysis was performed to determine individual relationships among all covariates excluding those in the preoperative chloride group with postoperative 90-day mortality. Then covariates with P < .2 were selected from the univariable Cox regression model and analyzed via multivariable Cox regression analysis with the main variable (preoperative chloride group). We tested the interaction of chloride level and a function of time, implemented as a time-dependent chloride level, to confirm that the proportional hazards assumption was satisfied (P = .078). Next, univariable logistic regression analysis was performed for postoperative acute kidney injury (the secondary end point), and variables with P < .2 were included in the final multivariable logistic regression analysis. Additionally, we performed uni- and multivariable logistic regression analysis for postoperative 90-day mortality using the same methods. There was no significant multicollinearity between variables in any of the multivariable Cox or logistic regression models (all variance inflation factors <2.5). For these multivariable Cox and logistic regression analyses, 2 different final multivariable models were made; these models differed with respect to the independent variables of main interest: the main independent variable was preoperative dyschloremia in model 1, while it was combined dyschloremia with dysnatremia in model 2.
The results of the Cox regression and logistic regression analyses are presented as hazard ratio with 95% CI and odds ratio with 95% CI, respectively. In addition, partial Pearson correlation analysis was performed to investigate the relationships between preoperative comorbidities and the preoperative chloride level, controlling for years of surgeries. To detect a 1% difference in the incidence of postoperative 90-day mortality between the normochloremia and hypochloremia groups with a 0.05 chance of type 1 error and 80% power, 33,511 patients were required. To detect a 1% difference in the incidence of postoperative 90-day mortality between the normochloremia and hyperchloremia groups with a 0.05 chance of type 1 error and 80% power, 112,735 patients were needed. The sample size calculations were performed using PASS software 15.0. SPSS version 24.0 (IBM Corporation, Armonk, NY) and R software (version 3.3.2 with R packages) were used for all other analyses. P values <.05 were considered statistically significant.
RESULTS Table 1.: Comparisons for Characteristics of Patients According to Preoperative Serum Chloride Group
Table 2.: Partial Pearson Correlation Analysis Between Preoperative Serum Chloride Level and Preoperative Comorbidities, Controlling for Year of Surgery
Figure 1.: Flow chart illustrating patient selection.
A total of 198,206 noncardiac procedures were performed between January 2010 and December 2016. Of these cases, the following were sequentially removed from analysis: patients who underwent ≥2 procedures in the study period (n = 41,559), those with incomplete or missing medical records (n = 22,627), patients <20 years of age (n = 25,257), and those with previous cardiac surgery (n = 2258). Ultimately, 106,505 patients who underwent noncardiac surgery were included in the final analysis, with 2147 (2.0%) in the preoperative hypochloremia group and 617 (0.6%) in the hyperchloremia group (Figure 1 ). Ninety-day mortality occurred in 871 (0.8%) patients, and acute kidney injury within ≤1 week after surgery occurred in 3204 (3.0%). Table 1 lists differences in the characteristics of the groups and postoperative mortality. Postoperative 90-day mortality rates were the highest in the preoperative hypochloremia group (8.5%) and second highest in the hyperchloremia group (3.7%). They were the lowest in the normochloremia group (0.6%; P < .001). The incidence of postoperative acute kidney injury was highest in the hypochloremia group (16.1%), second highest in the hyperchloremia group (8.3%), and the lowest in the normochloremia group (2.7%; P < .001). The results of the partial Pearson correlation analysis between the preoperative chloride level and preoperative comorbidities controlling for the year of surgery are presented in Table 2 .
Ninety-Day Mortality After Noncardiac Surgery According to Preoperative Chloride Level Table 3.: Cox Regression Analysis and Logistic Regression Analysis for Postoperative 90-d Mortality According to Status of Preoperative Serum Chloride After Noncardiac Surgery in 2010–2016
Figure 2.: Kaplan-Meier curve of 90-d mortality after noncardiac surgery, according to preoperative serum chloride group.
The Kaplan-Meier curve of 90-day mortality according to preoperative chloride group is shown in Figure 2 . Cumulative 90-day survival rates after noncardiac surgery were 99.4%, 91.5%, and 96.3% in preoperative normochloremia, hypochloremia, and hyperchloremia groups, respectively (P < .001). Supplemental Digital Content, Table 1, https://links.lww.com/AA/C671 Table 3 presents the 90-day mortality data according to preoperative hypochloremia and hyperchloremia before and after adjustment for covariates. Multivariable Cox regression analysis revealed that the preoperative hypochloremia group exhibited significantly increased 90-day mortality (hazard ratio, 1.46; 95% CI, 1.16–1.84; P = .001) when compared with the normochloremia group. The preoperative hyperchloremia group also exhibited significantly increased 90-day mortality (hazard ratio, 1.76; 95% CI, 1.13–2.73; P = .013) when compared with the normochloremia group. However, there was no significant difference in 90-day mortality between the hyperchloremia and hypochloremia groups (P = .459). In addition, univariable and multivariable logistic regression analyses for 90-day mortality in Supplemental Digital Content, Tables 2–3, https://links.lww.com/AA/C671
Postoperative Acute Kidney Injury According to Preoperative Chloride Levels Table 4.: Logistic Regression Analysis for Acute Kidney Injury After Noncardiac Surgery in 2010–2016
The results of univariable logistic regression analysis, revealing the factors associated with the incidence of postoperative acute kidney injury, are listed in Supplemental Digital Content, Table 3, https://links.lww.com/AA/C671 P < .001; Table 4 ). However, there was no significant difference between the odds of acute kidney injury in the preoperative hyperchloremia group and the normochloremia group (P = .879). In addition, there was a significant increase in the odds of acute kidney injury (1.78-fold) in the hypochloremia group when compared to that in the hyperchloremia group (odds ratio, 1.78; 95% CI, 1.20–2.63; P = .004).
DISCUSSION
Here we report that preoperative hypochloremia and hyperchloremia are both independently related to increased 90-day mortality after noncardiac surgery when compared with preoperative normochloremia. This association with increased mortality was not significantly different between the hypochloremia group and the hyperchloremia group. In terms of the incidence of postoperative acute kidney injury, only preoperative hypochloremia demonstrated a significant relationship. Additionally, patients exhibiting preoperative dyschloremia with dysnatremia demonstrated higher mortality or acute kidney injury after noncardiac surgery, compared with patients exhibiting normochloremia with normonatremia, as shown in model 2 of Tables 3 and 4 . The results of this study suggest that the preoperative chloride level is a valuable test variable that exhibits an independent relationship with mortality or morbidity after noncardiac surgery.
The most important characteristic of this study is the inclusion of preoperative sodium level as an important covariate. Preoperative dysnatremia is a well-known risk factor for increased 30-day mortality or morbidity after surgery.17 18 https://links.lww.com/AA/C671 −1 (Supplemental Digital Content, Figure 1A, https://links.lww.com/AA/C671 https://links.lww.com/AA/C671
Second, the potential-related factors of preoperative dyschloremia were considered in this study. The results of the partial Pearson correlation analysis shown in Table 3 suggest potential factors related to preoperative dyschloremia in this study. First, higher body mass index and older age were correlated with an increase in the preoperative chloride level. Considering that hypochloremia is a symptom of malnutrition,19 20 Table 3 were very close to 0 because of the large sample sizes, the clinical impact of potential factors related to preoperative dyschloremia might be controversial in clinical practice. Therefore, further studies should be performed to characterize the factors that cause preoperative dyschloremia.
Another interesting finding of this study is that, while preoperative hypochloremia is independently related to the increased incidence of acute kidney injury, hyperchloremia was not. Recent studies have reported that hyperchloremia in critically ill patients potentially increased the risk for acute kidney injury.21 , 22 23 7 , 24
The present study had a few limitations. First, there may have been selection bias due to the retrospective cohort design. To overcome this bias, medical records technicians blinded to the purpose of the study collected all medical records for the analysis. Second, the generalizability of the study may be in question because it was conducted at a single tertiary care hospital. Third, there was significant (>20,000) exclusion of data due to the absence of medical information regarding preoperative chloride or sodium levels. Despite these limitations, this study is significant because, to our knowledge, it is the first to establish a relationship between preoperative chloride levels and mortality and morbidity after noncardiac surgery. Furthermore, our results show that preoperative dyschloremia might also be associated with increased length of hospital stay, which leads to increased hospital costs. Future studies are needed to clarify this relationship.
In conclusion, we report that preoperative hypochloremia and hyperchloremia were both related to increases in 90-day mortality after noncardiac surgery. In addition, preoperative hypochloremia was related to an increased risk for postoperative acute kidney injury.
ACKNOWLEDGMENTS
We are particularly grateful to the statistician Boram Park at the Biometrics Research Branch, Research Institute and Hospital, National Cancer Center in Korea for her contribution to the statistical analysis.
DISCLOSURES
Name: Tak Kyu Oh, MD.
Contribution: This author helped design the study and draft the first manuscript and read and approved the final manuscript.
Name: Sang-Hwan Do, MD, PhD.
Contribution: This author helped with the acquisition of data, critically revised the manuscript, and read and approved the final manuscript.
Name: Young-Tae Jeon, MD, PhD.
Contribution: This author helped with the acquisition of data, critically revised the manuscript, and read and approved the final manuscript.
Name: Jinhee Kim, MD, PhD.
Contribution: This author helped with the acquisition of data, critically revised the manuscript, and read and approved the final manuscript.
Name: Hyo-Seok Na, MD, PhD.
Contribution: This author helped with the acquisition of data, critically revised the manuscript, and read and approved the final manuscript.
Name: Jung-Won Hwang, MD, PhD.
Contribution: This author helped with study design, critically revised the manuscript, and read and approved the final manuscript.
This manuscript was handled by: Tong J. Gan, MD.
REFERENCES
1. Berend K, van Hulsteijn LH, Gans RO. Chloride: the queen of electrolytes? Eur J Intern Med. 2012;23:203–211.
2. Wolfe LA, Kemp JG, Heenan AP, Preston RJ, Ohtake PJ. Acid-base regulation and control of ventilation in human pregnancy. Can J Physiol Pharmacol. 1998;76:815–827.
3. Yunos NM, Bellomo R, Story D, Kellum J. Bench-to-bedside review: chloride in critical illness. Crit Care. 2010;14:226.
4. Kimura S, Matsumoto S, Muto N, et al. Association of serum chloride concentration with outcomes in postoperative critically ill patients: a retrospective observational study. J Intensive Care. 2014;2:39.
5. Neyra JA, Canepa-Escaro F, Li X, et al.; Acute Kidney Injury in Critical Illness Study Group. Association of hyperchloremia with hospital mortality in critically ill septic patients. Crit Care Med. 2015;43:1938–1944.
6. Oh HJ, Kim SJ, Kim YC, et al. An increased chloride level in hypochloremia is associated with decreased mortality in patients with severe sepsis or septic shock. Sci Rep. 2017;7:15883.
7. Suetrong B, Pisitsak C, Boyd JH, Russell JA, Walley KR. Hyperchloremia and moderate increase in serum chloride are associated with acute kidney injury in severe sepsis and septic shock patients. Crit Care. 2016;20:315.
8. Tani M, Morimatsu H, Takatsu F, Morita K. The incidence and prognostic value of hypochloremia in critically ill patients. Sci World J. 2012;2012:474185.
9. Lavie CJ, Crocker EF Jr, Key KJ, Ferguson TG. Marked hypochloremic metabolic alkalosis with severe compensatory hypoventilation. South Med J. 1986;79:1296–1299.
10. O’Dell E, Tibby SM, Durward A, Murdoch IA. Hyperchloremia is the dominant cause of metabolic acidosis in the postresuscitation phase of pediatric meningococcal sepsis. Crit Care Med. 2007;35:2390–2394.
11. McCluskey SA, Karkouti K, Wijeysundera D, Minkovich L, Tait G, Beattie WS. Hyperchloremia after noncardiac surgery is independently associated with increased morbidity and mortality: a propensity-matched cohort study. Anesth Analg. 2013;117:412–421.
12. Williams EL, Hildebrand KL, McCormick SA, Bedel MJ. The effect of intravenous lactated Ringer’s solution versus 0.9% sodium chloride solution on serum osmolality in human volunteers. Anesth Analg. 1999;88:999–1003.
13. Kilbride HS, Stevens PE, Eaglestone G, et al. Accuracy of the MDRD (Modification of Diet in Renal Disease) study and CKD-EPI (CKD Epidemiology Collaboration) equations for estimation of GFR in the elderly. Am J Kidney Dis. 2013;61:57–66.
14. Damhuis RA, Wijnhoven BP, Plaisier PW, Kirkels WJ, Kranse R, van Lanschot JJ. Comparison of 30-day, 90-day and in-hospital postoperative mortality for eight different cancer types. Br J Surg. 2012;99:1149–1154.
15. Talsma AK, Lingsma HF, Steyerberg EW, Wijnhoven BP, Van Lanschot JJ. The 30-day versus in-hospital and 90-day mortality after esophagectomy as indicators for quality of care. Ann Surg. 2014;260:267–273.
16. Kellum JA, Lameire N; KDIGO AKI Guideline Work Group. Diagnosis, evaluation, and management of acute kidney injury: a KDIGO summary (Part 1). Crit Care. 2013;17:204.
17. Lien YH. Dysnatremia: a clear and present risk for surgery. Am J Med. 2013;126:849–850.
18. Rassam SS, Counsell DJ. Perioperative electrolyte and fluid balance. Contin Educ Anaesth Crit Care Pain. 2005;5:157–160.
19. Jacocks MA. Claudet RR. Fluids. Electrolytes. Nutrition. In: General Surgery. 1993:New York, NY: Springer; 1–10.
20. Jentsch TJ. Chloride transport in the kidney: lessons from human disease and knockout mice. J Am Soc Nephrol. 2005;16:1549–1561.
21. Patel N, Baker SM, Walters RW, et al. Serum hyperchloremia as a risk factor for acute kidney injury in patients with ST-segment elevation myocardial infarction undergoing percutaneous coronary intervention. Proc (Bayl Univ Med Cent). 2016;29:7–11.
22. Sadan O, Singbartl K, Kandiah PA, Martin KS, Samuels OB. Hyperchloremia is associated with acute kidney injury in patients with subarachnoid hemorrhage. Crit Care Med. 2017;45:1382–1388.
23. Toyonaga Y, Kikura M. Hyperchloremic acidosis is associated with acute kidney injury after abdominal surgery. Nephrology (Carlton). 2017;22:720–727.
24. Shaw AD, Raghunathan K, Peyerl FW, Munson SH, Paluszkiewicz SM, Schermer CR. Association between intravenous chloride load during resuscitation and in-hospital mortality among patients with SIRS. Intensive Care Med. 2014;40:1897–1905.
