Analyzing Chinese Patients with Post-Operative Acute Kidney Injury

Abstract

Background: To investigate clinical characteristics and risk factors of Chinese patients with post-operative acute kidney injury (PO-AKI). Methods: Patients with PO-AKI in Ruijin Hospital from December 1997 to December 2005 were retrospectively studied. Results: Patients’ mean age was 62.2 ± 18.1 years. There were 111 males and 57 females. The mean serum creatinine at diagnosis was 370.41 ± 320.92 μmol/L and the mean estimated glomerular filtration rate was 33.56 ± 24.24 mL/min. For the outcome of the patients, 38 died and the mortality rate was 22.6%. There were 17 patients (10.1%) with Acute Dialysis Quality Initiative-RIFLE (risk-injury-failure-loss-end classification) phase R, 21 (12.5%) with phase I, and 130 (77.4%) with phase F. There was no significant difference in mortality regarding patients who underwent different types of surgeries. For the risk factors related to PO-AKI, acute tubular necrosis (ATN) increased relative risk of mortality PO-AKI (odds ratio = 7.089, 95% confidence interval = 2.069–24.288, p < 0.001). Multivariate regression models showed that ATN had a positive correlation with mortality of PO-AKI. Conclusions: PO-AKI is one of the most common causes of AKI in patients who underwent operations. Special attention should be paid to risk factors related to PO-AKI in order to improve prognosis.

• acute kidney injury
• post-operative
• risk factors
• renal replacement therapy

INTRODUCTION

Post-operative acute kidney injury (PO-AKI) is a severe complication in patients who underwent operations, which is associated with prolonged hospitalization and increased mortality. It is one of the most common causes of hospital-acquired AKI (HA-AKI).¹ High-risk surgeries including cardiac or aortic surgeries could add burden to kidney and lead to renal dysfunction. Clinical factors that are closely associated with prognosis of PO-AKI include anesthetic drugs used in/after operations, hypovolemia, sepsis, multiple organ dysfunction syndrome (MODS), and congestive heart failure (CHF).² The mortality rate of PO-AKI is estimated to be more than 50%.³ However, few studies have been published focusing on Chinese patients with PO-AKI. In this study, we investigate the clinical characteristics of Chinese patients with PO-AKI and survival rate in our patients in order to further study the disease.

PATIENTS AND METHODS

Patients Selection and Definition

From December 1997 to December 2005 patients diagnosed with PO-AKI were enrolled in this study and data were retrospectively analyzed.

AKI was defined and classification of AKI was made according to the definition from Acute Dialysis Quality Initiative (ADQI)-RIFLE.⁴ As the definition of AKI was not introduced until 2004, the patients were enrolled according to the following criteria before ADQI-RIFLE was introduced: increase in serum creatinine (Scr) >50% or decrease in creatinine clearance >50%.⁵ The estimated glomerular filtration rate (eGFR) was calculated using Cockcroft Gault formula.⁶ PO-AKI as defined as AKI developed within 72 h after operations.^7,8 Only patients who met the definition of AKI were enrolled in the current study. Patients with previously reduced renal function suffering a deterioration were excluded.

Risk factors affecting mortality were selected according to the literatures.^9–11 Infection was defined as pathological process caused by invasion of normally sterile tissue or fluid or body cavity by pathogenic or potentially pathogenic micro-organisms.¹² Hypotension was defined as mean arterial pressure (MAP) <70 mmHg or any vasopressor except dopamine dosed at less than 5.0 μg/kg/min. CHF was defined as New York Heart Association grade III or IV heart failure. Nephrotoxin exposure was defined as amphotericin B, aminoglycosides, nonsteroidal anti-inflammatory drugs excluding aspirin, and radiocontrast given within 24 h.¹⁰ Chronic obstructive pulmonary disease (COPD) was defined as history of COPD requiring medical treatment. Hypertension was defined as systolic pressure ≥140 mmHg, or diastolic pressure ≥90 mmHg or history of hypertension requiring antihypertensive medication.¹³ Chronic kidney disease (CKD) was defined according to K/DOQI definition.¹⁴ Peripheral vascular disease was defined as evidence of peripheral vascular disease by angiography or noninvasive testing, history of vascular surgery, or symptomatic vascular occlusion.

Since most of the patients with HA-AKI were in critical condition and renal biopsy might not be necessary in evaluation and therapy of patients with AKI,⁵ the cause of patients with AKI was made according to clinical evaluation. Prerenal AKI was diagnosed when Scr rises after periods of volume depletion, and etiologic treatment was rapidly successful in restoring renal function. A fractional excretion of sodium 1% was considered as additional evidence; acute tubular necrosis (ATN) was suspected if AKI followed a period of moderate or severe renal hypoperfusion and persisted despite restoration of systemic blood pressure and tissue reperfusion or history of nephrotoxic agents. A fractional excretion of sodium >2% was considered as additional evidence that diuretics had not been used for 48 h before; postrenal AKI was defined by radiologic investigations of obstruction.^7,15,16

Treatment

Treatment of AKI was made based on patients’ clinical situation and pathogenesis of disease. For patients presented with hypotension, hypovolemia, as well as AKI, such disorders were corrected before renal replacement therapy (RRT) could be performed. For RRT, indications were listed as follows: PO-AKI with nonobstructive oliguria (urine output <400 mL/day) or urine output <20 mL/h, systemic inflammation reaction syndrome (SIRS), septicemia, sepsis, CHF and pulmonary edema with unstable circulation, brain edema, acute respiratory distress syndrome that depended on mechanical ventilations, acute necrotizing pancreatitis, high catabolism, hepatic encephalopathy, requirement of intravenous nutrition support, MODS, hyperbilirubinemia, and lactic acidosis.^17–20 Different types of RRT including continuous veno-venous hemofiltration (CVVHF) or continuous veno-venous hemodiafiltration (CVVHDF) were performed according to patients’ clinical situations.

Statistical Analysis

The study population was divided into survival and death groups. These patients were also separated into groups who received RRT or who did not. Data were analyzed using SPSS software (version 11.0) (SPSS Inc., Chicago, IL, USA). Potential association between various clinical variables (drugs, infection, shock, hypotension, hypovolemia, ATN, history of COPD, CHF, gender, peripheral vascular disease, MODS) between groups was assessed by independent t-test (for data that were normally distributed) or nonparametric test (for data that were not normally distributed). Differences of statistical data were compared using χ² test. Values fitting with standard normal distribution were represented as mean ± standard deviation (X ± SD). Risk factors screened by t-test were analyzed by binary logistic regression models. p-Value <0.05 was considered statistically significant. Then, the patients were divided into three groups: cardiac/major vascular, trauma/burns, and general surgeries. One-way ANOVA test was used to analyze the numerical variables. The differences of the incidences between the three groups were judged by R × C crosstable.

Figure 1. Age distribution of post-operative acute kidney injury (PO-AKI). Patients were categorized in five groups according to different age intervals and their percentages are shown.

RESULTS

Characteristics of Patients

During December 1997 to December 2005, charts of 168 patients diagnosed with PO-AKI were reviewed in this study. They made up 34.3% of 490 HA-AKI patients in our hospital. The incidence of PO-AKI among total patients receiving surgeries was 0.13% (168/128,973) and the incidence among total inpatients was 0.08% (168/215,132). The incidence of PO-AKI among total patients receiving surgeries ranged from 0.079% to 0.17% from that time.

In the study, there were 111 males and 57 females with the male-to-female ratio of 1:0.51 and their average age was 62.2 ± 18.1 years (ranging from 7 to 94 years). Details of age distribution are shown in Figure 1. In our patients, more elderly patients (age 69–79 years) developed PO-AKI than other patients. The mean Scr at diagnosis was 370.41 ± 320.92 μmol/L and the mean eGFR was 33.56 ± 24.24 mL/min.

Figure 2. Types of operations. The percentage of patients with different types of operation is illustrated.

Table 1. Comparison of clinical variables of patients who underwent different types of surgeries.

	Trauma/burns surgeries (n = 15)	General surgeries (n = 89)	Cardiac/major vascular surgeries (n = 64)	p-Value
Scr, μmol/L	309.92 ± 222.01	400.98 ± 369.71	345.60 ± 261.17	NS
eGFR, mL/min	27.07 ± 24.62	35.65 ± 26.05	41.17 ± 29.86	NS
Sepsis, %	46.67(7/15)	42.70(38/89)	59.4(38/64)	NS
Hypotension, %	40.00(6/15)	11.24(10/89)	35.9(23/64)	<0.01
Hypovolemia, %	6.67(1/15)	21.35(19/89)	6.3(4/64)	<0.05
ATN, %	73.33(11/15)	62.92(56/89)	81.3(52/64)	<0.05
Chronic dialysis, %	6.67(1/15)	11.24(10/89)	3.1(2/64)	NS
Mortality^a, %	26.67(4/15)	20.22(18/89)	25.0(16/64)	NS
Survival, %	66.67(10/15)	75.28(67/89)	70.3(45/64)	NS

Notes: Scr, serum creatinine; eGFR, estimated glomerular filtration rate; ATN, acute tubular necrosis.

^aOne patient with trauma/burns surgeries, four with general surgeries, and three with cardiac/major vascular surgeries were lost for follow-up and their data were not included when we analyzed mortality/survival.

Since some of our patients were diagnosed with acute renal failure (ARF) before ADQI-RIFLE definition was introduced, we re-evaluated all our patients and the results showed that all our patients met the definition of AKI according to the diagnosing criteria. For the classification of our patients with PO-AKI, there were 17 patients (10.1%) with AKI-RIFLE phase R, 21 (12.5%) with phase I, and 130 (77.4%) with phase F.

Types of Surgeries

Regarding types of surgeries in our patients, 64 patients (38.1%) had thoracic or cardiovascular surgeries, the rest 104 patients (61.9%) had other types of surgeries including general surgeries (53 cases), urological surgeries (14 cases), gynecology and obstetrics surgeries (11 cases), orthopedic surgeries (10 cases), cerebral surgeries (7 cases), burn surgeries (5 cases), and neck and dentofacial surgeries (4 cases). Details are summarized in Figure 2.

We categorized patients into cardiac/major vascular surgeries, trauma/burns surgeries, and general surgeries groups. In our study, hypotension, hypovolemia, and ATN were significantly different among the three groups. There was no significant difference regarding mortality in patients with different types of surgeries. Details are summarized in Table 1.

Outcome and Renal Replacement Therapy in Patients with PO-AKI

While 122 patients survived and 38 died, 8 cases were lost for follow-up (3/8 received RRT while 5/8 did not receive RRT). The mortality in the current study was 22.6% (38/168). Forty-three patients underwent RRT, which made up 25.6% of all PO-AKI patients. All those patients received continuous renal replacement therapy (CRRT) during their first hospitalization. In our study, 13 patients did not recover from renal insufficiency; they received intermittent hemodialysis afterwards. The mortality rate among RRT group was 27.9% while it was 20.8% in patients who did not receive RRT. There was no significant difference regarding mortality between patients who received RRT and those who did not (p > 0.05).

As RRT was an important type of therapy for patients with AKI, we then compared the data between patients who received RRT and those who did not. Data are summarized in Table 2. In our study, there was no significant difference regarding mortality between the two groups of patients. However, patients on RRT had lower percentage of infection (p < 0.01).

Table 2. Comparison of patients who received and who did not receive RRT.

	Patients who did not receive RRT (n = 125)	Patients who received RRT (n = 43)	p-Value
Age	52.15 ± 20.55 years	54.03 ± 23.40 years	NS
BUN	16.84 ± 12.16 mmol/L	21.25 ± 17.88 mmol/L	NS
Scr	355.26 ± 306.43 μmol/L	412.36 ± 358.53 μmol/L	NS
Infection, %	59.2 (74/125)	20.9 (9/43)	<0.01
Chronic dialysis, %	8.0 (10/125)	7.0 (3/43)	NS
Survival^a, %	75.2 (94/125)	65.1 (28/43)	NS
Mortality, %	20.8 (26/125)	27.9 (12/43)	NS

Notes: Scr, serum creatinine; RRT, renal replacement therapy; BUN, blood urea nitrogen.

^aFive patients in without-RRT group and three in with-RRT group were lost for follow-up and their data were not included when we analyzed mortality/survival.

Clinical Variables Related to PO-AKI

Clinical and demographic variables were assessed as univariate predictors of PO-AKI. The details of univariate variables and confidence intervals (CI) are listed in Table 3.

Table 3. Clinical variables associated with prognosis of patients with PO-AKI.

Clinical variables	OR	95% CI	p-Value
ATN	7.525	2.203–25.707	<0.001
Nephrotoxin exposure	1.234	0.625–2.438	NS
Hypotension and/or CHF		−0.0530.294	NS
Infection	0.858	0.435–1.692	NS
Hypertension	1.495	0.687–3.252	NS
COPD	0.872	0.216–3.528	NS
Gender (>70 years)	0.643	0.319–1.296	NS
CKD	0.746	0.132–4.216	NS
Peripheral vascular disease	2.328	0.273–19.878	NS
MODS	0.505	0.208–1.223	NS

Note: PO-AKI, post-operative acute kidney injury; OR, odds ratio; CI, confidence interval; ATN, acute tubular necrosis; CHF, chronic heart failure; COPD, chronic obstructive pulmonary disease; CKD, chronic kidney disease; MODS, multiple organ dysfunction syndrome.

Table 4. Results of multifactorial logistic regression analysis of mortality.

			95% CI
	p-Value	OR	Lower	Upper
ATN	0.001	7.089	2.069	24.288

Note: OR, odds ratio; CI, confidence interval; ATN, acute tubular necrosis.

In our study, presence of ATN increased relative risk of mortality of PO-AKI [odds ratio (OR) = 7.089, 95% CI = 2.069–24.288, p < 0.001]. Other variables including drug, hypotension, infection, hypertension, history of COPD, CHF, gender, peripheral vascular disease, and MODS did not predict mortality of PO-AKI. Multivariate regression models showed that ATN had a positive correlation with mortality of PO-AKI (Table 4). Other factors had no impact on the mortality rate.

DISCUSSION

Risk factors and prognosis of PO-AKI are receiving more and more attention nowadays. According to the study by Liano and Pascual,⁷ PO-AKI made up 27% of all HA-AKI patients. In the reports published elsewhere, the incidence of PO-AKI was about 1.2% in patients who had surgeries²¹ and about 1–31% in patients who had post-cardiac surgeries.²² In our study, PO-AKI patients made up 20% of all AKI patients. Our results suggest the epidemiological data of Chinese patients with PO-AKI were similar to studies reported elsewhere. The numbers of PO-AKI patients and AKI patients in our study were higher than in our previous studies (PO-AKI in 16.6% and AKI in 14.1% in all patients),^23,24 which might be related to increased numbers of complicated and high-risk surgeries in our hospital.

In our study, the occurrence of ATN was closely related to the mortality of PO-AKI by the regression model, which indicated that ATN was important for the outcome.²⁵ There are some differences regarding the risk factors to prognosis between our study and other post-operative studies in the literature. In the study by Fertmann and colleagues² advanced age, peritonitis, disease severity, and dependence on invasive ventilation are closely associated with outcome. Peritonitis is not a risk factor in our study because the number of our patients with abdominal surgery and peritonitis was small. In Fertmann’s study, the percentage of abdominal surgery was more than 50%; therefore, prevalence of peritonitis, which was a complication of abdominal surgery, was high accordingly. Considering that the constitution of types of surgeries was different between our study and that of Fertmann’s, the risk factors might vary as a result. In the study by Chawla et al.,¹⁰ the prognosis of AKI was closely related to CKD, hypovolemia during or after the operations, hypotension, advanced age, CHF, elevated bilirubin, mechanical ventilation, oliguria, valve surgery, and atherosclerotic cardiovascular disease. Chertow et al.²⁶ investigated their patients who had cardiac surgeries and found that 1.1% of their patients had AKI and required RRT while the mortality rate was 63.7%. Advanced age, histories of CKD or chronic renal failure, jaundice, septicemia, oliguria, mechanical ventilation, and occurrence of two or more organ failures were the risk factors for poor prognosis. Some other studies demonstrated that preoperative factors and intraoperative factors of PO-AKI included advancing age, sepsis, hepatic failure, and hypovolemia.^3,27 However, we did not find such factors were related to poor prognosis in our patients. Since the diagnosing criteria of AKI/ARF were not unanimous between the literature and ours, the risk factors might be different accordingly. Furthermore, different clinical situations (including type of surgeries, study subjects, definition of AKI/ARF) and integrative therapy might also contribute to different results. Therefore, further studies might be necessary to investigate the details.

RRT is an important clinical strategy to treat patients with AKI. To remove the excessive volume and inflammatory mediators, RRT might be necessary to improve prognosis.^28–33 It is known that RRT is beneficial to patients with sepsis-related MODS because of its capacity to remove cytokines; early use of RRT might improve prognosis in patients with SIRS and sepsis.^34–37 It was demonstrated that early initiation of CRRT could improve outcome in AKI patients who underwent cardiac surgery.³² Furthermore, it was shown that early initiation of RRT in patients with AKI might be associated with improved survival.^32,38 In our study, there was no significant difference regarding mortality between patients who received RRT and those who did not. Our results were not contradictory to established findings. One possible explanation might be patients who received RRT had more severe clinical situations. Although the beneficial effects of early initiation of CRRT are noticeable, the mechanism remains largely unclear. Further study is necessary to investigate the details.

This study has certain limitations and constraints. It is a retrospective study, and limited information regarding treatment and prognosis could be obtained. As patients with PO-AKI are usually in critical situation, it is difficult to perform prospective study to evaluate different types of RRT on prognosis. Furthermore, there is no protocol for RRT in PO-AKI patients and analysis of risk factors and treatment should depend on large-scale, well-designed studies. Future investigation is required to more rigorously investigate relationship between RRT, prognosis, and risk factors.

In conclusion, the results of our study indicated that PO-AKI was a common complication among patients who underwent operations and it was one of the most common cause of AKI and HA-AKI. The prognosis of patients with PO-AKI was poor. Special attention should be paid to risk factors related to PO-AKI in order to improve prognosis.

ACKNOWLEDGMENTS

This study was supported by a grant from the Leading Academic Discipline Project of Shanghai Health Bureau (no. 05III001), a grant from the Shanghai Leading Academic Discipline Project (no. T0201), a grant from National Natural Science Foundation (no. 81070586), and grants from Shanghai Scientific committee (nos. 10411965900, 08dz1900502).

Declaration of interest: The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper.