original article correlation study on postoperative ...comparison of preoperative renal function and...

Int J Clin Exp Med 2016;9(2):1027-1038www.ijcem.com /ISSN:1940-5901/IJCEM0015791

Original ArticleCorrelation study on postoperative changes in renal hemodynamics and renal function of patients with traumatic brain surgery by color doppler flow imaging (CDFI)

Feiya Hu1,2*, Anmin Shao1*, Shigang Qiao3, Ying Sun2, Liuhui Chang3, Chen Wang2

1Department of Anesthesiology, The Traditional Chinese Medicine Hospital of Kunshan, Kunshan 215300, China; 2Department of Anesthesiology and Critical Care, The Second Affiliated Hospital of Soochow University, Suzhou 215004, China; 3Department of Anesthesiology and Critical Care, Suzhou Science & Technology Town Hospital, Suzhou 215000, China. *Equal contributors and co-first authors.

Received September 7, 2015; Accepted November 23, 2015; Epub February 15, 2016; Published February 29, 2016

Abstract: To investigate the correlation of renal hemodynamic changes with renal function following traumatic brain surgery in order to provide clinical evidences for early diagnosis and treatment of postoperative acute kidney injury (AKI). 51 cases of adult with craniotomy evacuation of hematoma were divided into AKI group and non-AKI group. Hemodynamic parameters in interlobar arteries of bilateral kidneys including peak systolic velocity (PSV), end dia-stolic velocity (EDV), resistance index (RI) were detected by CDFI at different time points before and after operation. Blood sample was collected for detection of creatinine (Crea), blood urea nitrogen (BUN), uric acid (UA). The sensitiv-ity and specificity of postoperative 1 h RI, Crea, BUN, UA for diagnosis of AKI following traumatic brain surgery were evaluated by receiver operating characteristic (ROC) curve and the area under ROC curve. Compared with non-AKI group, patients in AKI group were older and underwent longer operative time (P < 0.05). Compared with preopera-tive parameters, postoperative 9-48 h Crea and BUN were increased in non-AKI group, while postoperative 9-48 h Crea, 4-48 h BUN and 9-16 h UA was increased in AKI group (P < 0.05). Compared with preoperative parameters, postoperative 1-4 h PSV was increased, postoperative 1-16 h EDV was decreased and postoperative 1-48 h RI was increased in non-AKI group while postoperative 1 h PSV was increased, postoperative 1-48 h EDV was decreased and postoperative 1-48 h RI was increased in AKI group. Furthermore, correlation analysis showed that EDV was negatively correlated with Crea, BUN and UA, while RI was positively correlated with Crea, BUN and UA in both groups (P < 0.05). Univariate analysis showed that age, operative time, postoperative 1 h Crea, EDV and RI was correlated with AKI (P < 0.05). Multivariate analysis showed that age and postoperative 1 h RI was correlated with AKI (P < 0.05). The area under ROC curve of postoperative 1 h RI for predicting AKI after traumatic brain surgery was 0.805, and the sensitivity was 90.00%, the specificity 67.70%, the optimal judgment value 0.70. In conclusion, after trau-matic brain surgery, RI of interlobar renal artery can reflect the changes in renal function earlier than Crea, BUN and UA, it can be an important index for early prediction of AKI.

Keywords: Ultrasonography, brain injury, resistance index, acute kidney injury (AKI)

Introduction

Acute kidney injury (AKI) is one of serious com-plications following traumatic brain surgery [1], and its early prediction and effective treatment has important clinical significance and practi-cal value. Currently, there is still a lack of high sensitivity and specificity of early diagnostic index for this disease in clinic. Previous studies

showed that Creatinine (Crea) was considered as the gold standard for the diagnosis of AKI, but it was influenced by many factors and its increase has hysteresis, so it might not well pre-dict early AKI [2, 3].

Color Doppler flow imaging (CDFI) diagnostic technique is known as non-invasive angiogra-phy and it is an effective means to detect kid-

Early diagnosis and treatment of postoperative AKI

1028 Int J Clin Exp Med 2016;9(2):1027-1038

ney diseases, especially renal injury [4]. Pre- vious studies showed that, in early AKI, chang-es in hemodynamics of renal artery were earlier than that in blood urea nitrogen (BUN), Crea, urinary albumin excretion rate, especially the increase in resistance index (RI) was the signal of early renal function impairment, so RI was a reliable index for evaluation of renal impair-ment [5, 6]. With the development of ultra-sound technology, the value of RI in the diagno-sis of AKI becomes more and more important. The distribution of renal blood flow can be directly reflected and the changes of each index can be monitored via ultrasonography, so it has important application value as the same as cre-atinine clearance, renal failure index and frac-tional excretion of filtration sodium [7]. However, there are still few clinical studies on postopera-tive renal hemodynamics of patients with trau-matic brain surgery.

In this study, changes in hemodynamic param-eters of interlobar arteries of bilateral were detected by CDFI at different time points before and after operation. Meanwhile, renal function indices were also detected. The order of chang-es in renal hemodynamic parameters and renal function indices at each time point was com-pared and the correlation between these parameters and indices were analyzed. Regre- ssion analysis of risk factors for AKI and ROC curve for comparison of the sensitivity and specificity of each index for early diagnosis of AKI in patients with traumatic brain surgery were performed in this study, which provided clinical evidences for early diagnosis and treat-ment of postoperative AKI.

Subjects and methods

Subjects

51 cases of subdural hemorrhage treated with craniotomy evacuation of hematoma in Kun- shan Chinese Medicine Hospital ICU room from

Inclusion criteria for patients [8]: (1) no abnor-mality of preoperative BUN, UA and Crea; (2) no renal and ureteral lithiasis as well as no hydro-nephrosis, renal cysts, renal artery stenosis and renal masses via CDFI examination; (3) no previous history of heart and lung diseases, diabetes and systemic lupus erythematosus as well as routine examination without arrhythmia; (4) no other trauma.

This study was conducted in accordance with the declaration of Helsinki. This study was con-ducted with approval from the Ethics Committee of Kunshan Chinese Medicine Hospital. Written informed consent was obtained from all par- ticipants.

Diagnostic criteria for AKI

In line with one of the following criteria, the patient was diagnosed as AKI according to the staging and standards recommended by KDIGO [9]. (1) The increase of Crea ≥ 26.5 μmol/L with-in 48 h; (2) creatinine increasing to 1.5 times the amount of the base value; (3) Urine output < 0.5 ml·kg-1·h-1 for more than 6 h. All patients were divided into AKI group and non-AKI group according to above criteria.

Collection and testing of blood samples

3 ml venous blood samples were taken at pre-operative and postoperative 1, 4, 9, 16, 24, 48 h respectively and were centrifuged for 10 min at 3500 rpm/min (Anhui USTC Zonkia scientific instruments Co., Ltd, KDC-2046 centrifuge). Next serum was separated for cryopreserva-tion. Cobas 8000 automatic biochemical ana-lyzer was used to detect blood biochemical indices of renal function including Crea, BUN, UA at each time point.

Color Doppler flow imaging (CDFI) assays

Hemodynamic changes in interlobar renal arteries of both kidneys were detected by CDFI at

Table 1. Comparison of the general situation of patients in both groups

Group Age (male/female) Gender Operative time (h)

Non-AKI group (n = 31) 59.00 (50.00, 65.00) 22:9 2.41 ± 0.51AKI group (n = 20) 73.15 ± 5.05* 14:6 3.71 ± 0.42*

P value < 0.001 0.941 < 0.001Note: Compared with non-AKI group, *P < 0.001.

August 2013 to July 2014 were selected in this study. After opera-tion, all patients needed mechani-cal ventilation persistently in ICU and the blood pressure should remained stable (mean arterial pressures of all patients ≥ 65 mmHg for more than 6 h without the use of vasopressors or ex- panding the capacity).


1029 Int J Clin Exp Med 2016;9(2):1027-1038

preoperative and postoperative 1, 4, 9, 16, 24, 48 h respectively. The center frequency of CDFI probe was 3.5 HZ. Preoperative measurement position of interlobar artery was recorded to maintain the same position for detection after operation, and the measurement depth of the probe was adjusted depending on each patient situation, however, each patient had the con-sistent depth before and after operation. PSV and EDV of interlobar arteries in both kidneys were measured at both sides of renal vertebra for three times and the averages were calcu-lated. If the difference of measurement values at both sides was more than 5%, this experi-ment was excluded. Finally, RI value was calcu-lated [RI = (PSV-EDV)/PSV].

Statistical analysis

All statistical analyses were performed using SPSS version 20.0 software (SPSS Inc, Chicago, IL, USA). Measurement data of normal distribu-tion was shown as the mean ± SD (

_x ± s) and

comparison of data between the two groups was analyzed using t-test or one-way ANOVA. Measurement data of non-normal distribution was shown as median and quartiles (P25, P75) and comparison of data between the two groups was analyzed using nonparametric rank sum test. Linear correlation and linear regres-sion analysis was used to show the relationship between variables. Multivariate analysis was performed by logistic regression method. ROC curve and the area under ROC curve were used for comparison of the sensitivity and speci-ficity of each index for early diagnosis of AKI in patients with traumatic brain surgery. P < 0.05 was considered statistically significant.

Results

Comparison of the general situation of pa-tients in both groups

As shown in Table 1, the difference in age and operative time between AKI group and non-AKI

group was statistically significant (P < 0.001); while there was no statistical difference in gen-der (P > 0.05). The results showed that patients in AKI group were older and underwent longer operative time compared with non-AKI group (P < 0.05; Table 1).

Comparison of preoperative renal function and hemodynamic parameters of patients in both groups

As shown in Table 2, there was no statistical difference in preoperative renal function indi-ces of Crea, BUN, UA and hemodynamic param-eters of PSV, EDV, RI in non-AKI group (P > 0.05).

Comparison of renal function and hemody-namic parameters of patients at each time point in non-AKI group

Renal function: At postoperative 9 h, Crea and BUN began to increase until postoperative 48 h, and the difference was statistically signifi-cant (P < 0.05). There was no significant differ-ence between preoperative and postoperative UA (P > 0.05).

Renal hemodynamic parameters: There was significant difference between preoperative and postoperative 1 h, 4 h PSV, EDV, RI (P < 0.05); there was significant difference between preoperative and postoperative 9 h, 16 h EDV, RI (P < 0.01); there was significant difference between preoperative and postoperative 24 h, 48 h RI (P < 0.01). Imaging data was shown in Figure 1, and the results showed that renal hemodynamic parameters changed earlier than biochemical indices of renal function in non-AKI patients (Figure 1; Table 3).

Comparison of renal function and hemody-namic parameters of patients at each time point in AKI group

Renal function: At postoperative 9 h, Crea began to increase until postoperative 48 h; at postoperative 4 h, BUN began to increase until

Table 2. Comparison of preoperative renal function and hemodynamic parameters of patients in both groupsGroup Crea (μmol/L) BUN (mmol/L) UA (mol/L) PSV (cm/s) EDV (cm/s) RINon-AKI group 57.55 ± 8.45 5.33 ± 0.91 302.41 ± 50.03 24.85 ± 1.54 8.23 ± 0.48 0.66 ± 0.01AKI group 59.56 ± 10.59 5.41 ± 0.68 329.75 ± 53.08 24.91 ± 1.12 8.12 ± 0.53 0.67 ± 0.01P value 0.457 0.734 0.267 0.891 0.448 0.056Note: Compared with non-AKI group, *P > 0.05.


1030 Int J Clin Exp Med 2016;9(2):1027-1038

Figure 1. Spectrum of blood flow in renal interlobar artery blood flow in non-AKI group. A. Preoperation; B. Postoperative 1 h; C. Postoperative 4 h; D. Postoperative 9 h; E. Post-operative 16 h; F. Postoperative 24 h; G. Postoperative 48 h.


1031 Int J Clin Exp Med 2016;9(2):1027-1038

postoperative 48 h; postoperative 9, 16 h UA increased, and the difference was statistically significant (P < 0.05).

Renal hemodynamic parameters: There was significant difference between preoperative and postoperative 1 h PSV, EDV, RI (P < 0.01); there was significant difference between preopera-tive and postoperative 4, 9, 16, 24, 48 h EDV, RI (P < 0.01). Imaging data was shown in Figure 2, and the results showed that renal hemody-namic parameters changed earlier than bio-chemical indexes of renal function in AKI patients (Table 4).

Results of linear correlation and linear regres-sion analysis for renal function and hemody-namic parameters

Correlation analysis showed that there was no correlation between PSV and Crea, BUN, UA (r = -0.081, -0.068, -0.049 respectively; P > 0.05); EDV was negatively correlated with Crea, BUN, UA (r = -0.446, -0.201, -0.31 respectively; P < 0.01) while RI was positively correlated with Crea, BUN, UA in both groups (r = 0.460, 0.185, 0.314 respectively; P < 0.01).

Crea was seen as the dependent variable y, and EDV, RI of renal interlobar artery was respectively seen as the independent variable x, so the corresponding regression equation was: yˆ = -11.114x + 157.783, yˆ = 329.273x-157.569 through linear regression analysis; BUN was seen as the dependent variable y, and EDV, RI of renal interlobar artery was respec-tively seen as the independent variable x, so the corresponding regression equation was: yˆ = -0.348x + 8.984, yˆ = 9.180x-0.100 through linear regression. UA was seen as the depen-dent variable y, and EDV, RI of renal interlobar

artery was respectively seen as the indepen-dent variable x, so the corresponding regres-sion equation was: yˆ = -23.504x + 495.862, yˆ = 682.955x-161.654 through linear regres-sion (Figures 3-5).

Results of logistic regression analysis

As shown in Table 5, univariate analysis showed that age, operative time, postoperative 1 h Crea, EDV, RI were correlated with AKI (P < 0.05).

As shown in Figure 6, the significant variables in univariate analysis was included into multi-variate analysis, and the results showed that age and postoperative 1 h RI were correlated with AKI (Table 6; P < 0.05).

ROC curve of each variable for prediction of postoperative AKI following traumatic brain injury

As shown in Figure 6 and Table 7, the area under ROC curve of postoperative 1 h RI for predicting AKI after traumatic brain surgery was 0.805 (> the area under ROC curve of Crea, BUN and UA), and the sensitivity was 90.00%, the specificity 67.70%, the optimal judgment value 0.70. This indicated that RI had a better predictive value for postoperative AKI compared with Crea, BUN, UA.

Discussion

AKI is one of the serious complications after traumatic brain surgery, and it is not to be ignored to prevent postoperative renal dam-age. Bahar et al. [10] reported that the risk of postoperative AKI was 1.243 for every addi-tional 10 years of age, and the relative risk of postoperative AKI for above 50-year-old was

Table 3. Comparison of renal function and hemodynamic parameters of patients at each time point in non-AKI groupTime point Crea (µmol/L) BUN (mmol/L) UA (mol/L) PSV (cm/s) EDV (cm/s) RIPre-operation 57.55 ± 8.45 5.33 ± 0.91 302.41 ± 50.03 24.85 ± 1.54 8.23 ± 0.48 0.66 ± 0.01Postoperative 1 h 57.75 ± 8.13 5.57 ± 1.27 309.04 ± 50.37 25.89 ± 1.52* 7.88 ± 0.48* 0.69 ± 0.02*

Postoperative 4 h 61.47 ± 9.29 5.60 (5.02, 6.43) 317.34 ± 55.02 25.65 ± 1.56* 7.67 ± 0.49* 0.70 ± 0.02*

Postoperative 9 h 69.90 ± 9.65* 7.15 ± 0.97* 303.70 ± 57.83 25.39 ± 1.59 7.50 ± 0.49* 0.70 ± 0.02*

Postoperative 16 h 70.66 ± 9.15* 6.33 ± 1.10* 297.75 ± 41.86 25.09 ± 1.61 7.72 ± 0.63* 0.69 ± 0.02*

Postoperative 24 h 66.82 ± 9.02* 6.23 ± 1.27* 291.15 ± 42.02 25.33 ± 1.43 8.13 ± 0.48 0.68 ± 0.02*

Postoperative 48 h 64.18 ± 8.61* 6.06 ± 1.23* 287.36 ± 35.13 25.37 ± 1.47 8.39 ± 0.50 0.67 ± 0.02*

Note: Compared with pre-operation, *P < 0.05.


1032 Int J Clin Exp Med 2016;9(2):1027-1038

Figure 2. Spectrum of blood flow in renal interlobar artery blood flow in AKI group. A. Pre-operation; B. Postoperative 1 h; C. Postoperative 4 h; D. Po- stoperative 9 h; E. Postopera-tive 16 h; F. Postoperative 24 h; G. Postoperative 48 h.


1033 Int J Clin Exp Med 2016;9(2):1027-1038

2.8, which might be due to the decrease of multi-organ function and intolerance to invasive surgery and to loss of volume in elderly patients. This study also suggested that there was a correlation between pro-longed operative time and postoperative AKI, which may be because the hemody-namic instability led to kidney hypoperfu-sion during a longer operative time. There- fore, in the operation, patients should be strictly underwent hemodynamic testing to maintain hemodynamic stability and re- duce AKI caused by volatility of renal blood perfusion.

At present, patient’s renal function was mainly determined by some biochemical indices of the amount of urine, Crea, BUN and UA in clinic, but above indices were vulnerable to many non-renal factors. More importantly, these indices began to rise until the irreversible damage in renal function occurred [11]. For example, only when severe renal impairment (> 50% of renal units) occurred did serum Crea sig-nificantly increased; only when renal lesion happened and glomerular filtration rate decreased to 1/3 to 2/3 of the normal value did serum Crea began to increase [2, 3], so it was difficult to timely and accu-rately reflect renal function, and patients could not receive timely treatment. Recent studies have shown that a slight increase in Crea value can also make a significant increase in mortality in critically ill patients [12, 13]. In this study, there were changes in preoperative and postoperative renal function indices of patients with brain inju-ry in AKI group, and postoperative Crea

Table 4. Comparison of renal function and hemodynamic parameters of patients at each time point in AKI groupTime point Crea (µmol/L) BUN (mmol/L) UA (mol/L) PSV (cm/s) EDV (cm/s) RIPre-operation 59.56 ± 10.59 5.41 ± 0.68 329.75 (254.85, 357.15) 24.91 ± 1.12 8.12 ± 0.53 0.67 ± 0.01

Postoperative 1 h 70.00 (55.00, 110.25) 5.75 ± 0.74 332.28 ± 48.60 26.37 ± 1.40* 7.53 ± 0.46* 0.71 ± 0.02*

Postoperative 4 h 80.70 (68.20, 124.25) 6.29 ± 0.70* 357.55 (306.80, 395.25) 25.70 ± 1.35 7.10 ± 0.54* 0.72 (0.71, 0.73)*

Postoperative 9 h 120.20 (93.95, 136.60)* 6.13 (5.41, 6.90)* 365.20 (320.73, 403.65)* 25.13 ± 1.18 6.94 ± 0.47* 0.72 ± 0.01*

Postoperative 16 h 126.5 (109.45, 159.95)* 7.27 ± 0.77* 355.24 ± 49.19* 24.98 ± 1.51 6.84 ± 0.61* 0.73 ± 0.02*

Postoperative 24 h 115.6 (104.85, 174.15)* 7.24 ± 1.01* 340.70 (281.80, 389.85) 25.31 ± 1.58 7.09 ± 0.71* 0.72 ± 0.02*

Postoperative 48 h 99.80 (93.70, 149.35)* 7.13 ± 1.31* 329.39 ± 61.71 25.26 ± 1.58 7.30 ± 0.81* 0.71 (0.69, 0.73)*

Note: All data was expressed as the mean ± SD (_x ± s) or median (P25, P75). Compared with pre-operation, *P < 0.05.

Figure 3. Scatterplot of postoperative Crea, BUN, UA and PSV in renal interlobar artery.


1034 Int J Clin Exp Med 2016;9(2):1027-1038

and BUN tended to increase compared with preoperative Crea and BUN with a statistically

ated with renal dysfunction. Therefore, mea-surement of these parameters in interlo-

Figure 4. Scatterplot of postoperative Crea, BUN, UA and EDV in renal interlobar artery.

significant change at postopera-tive 9, 16 h.

Color Doppler flow imaging (CDFI) was used to detect renal blood flow. We can see renal vessels clearly and measure renal blood velocity accurately as well as pre-dict renal vascular bed resistance via CDFI. Studies have indicated that the abnormal changes of renal hemodynamics were earlier than that of biochemical parame-ters of renal function [14], CDFI can be used to detect many renal hemodynamic parameters, of which PSV, EDV, Pulsatility Index (PI) and RI have a greater diagnostic significance and good reproduc-ibility [15]. PSV mainly reflects the intensity of vascular filling and blood supply; EDV mainly reflects vascular compliance and vascular bed resistance; PI and RI mainly reflect the state of vascular bed resistance. In recent years, many sc- holars have recommended that RI can be used to assess renal blood perfusion in critically ill patients [16-18]. Renal artery RI, reflecting the information of the systolic and diastolic blood flow, depends on the ratio of PSV to the difference value between the PSV and EDV. It not only reflects changes in blood vessel elasticity but also better suggests blood flow in small ves-sels, Especially, it reflects the compliance and blood flow resis-tance of a section of the artery, and is an objective and quantita-tive assessment of the degree of renal parenchymal damage [19]. Therefore, in this study, PSV and EDV were selected and measured at different time points before and after surgery. Petersen et al. [20] also believed that hemody-namic parameters in interlobar artery could easily reflect changes in blood flow because interlo-bar artery was close to renal parenchyma and closely associ-


1035 Int J Clin Exp Med 2016;9(2):1027-1038

bar artery was a perfect selection in this research.

ly and quantitatively. It is related to age and central hemodynamic parameters [23, 24], but

Figure 5. Scatterplot of postoperative Crea, BUN, UA and RI in renal interlobar artery.

In our study, PSV in interlobar renal arteries of both kidneys began to increase at postoperative 1 h in both non-AKI group and AKI group, which suggested renal vessels were in hypoperfusion. The correlation between PSV and renal function was analyzed, and the results showed that there was no obvious correlation between them. PSV mainly reflects the supply level of systolic blood vessel and blood sup-ply. It is widely used in the detection of renal artery stenosis, and studies have shown that PSV > 200 cm/s indicated renal artery stenosis [21, 22]. EDV mainly reflects diastolic perfusion of the kidney, and it de- clined obviously with renal blood perfusion reducing, so low velocity of renal flow was shown along the baseline level via CDFI. Patients with traumatic brain injury might present renal hypoperfusion for a long time. There was no statistical difference in preoperative EDV of interlobar renal arteries between non-AKI group and AKI group, so it was considered that preoperative EDV values of patients were at the same level. Compared with pre-operation, postoperative 1-16 h EDV of non-AKI group was decreased, while postoperative 1-48 h EDV of AKI group was decreased, and the difference was statistically signifi-cant. In addition, linear correlation and linear regression analysis showed that EDV was negatively cor-related with each index of renal function. Therefore, we inferred that the function of the kidney could be indirectly reflected by the EDV value of interlobar renal artery.

RI value is calculated from PSV and EDV, which may reacts arterial cro- ss-section compliance and blood flow elastic resistance, especially the blood flow in small vessels and vascular bed resistance, and assess the degree of renal lesion objective-


1036 Int J Clin Exp Med 2016;9(2):1027-1038

not affected by the angle of the ultrasonic probe and blood flow volume. Bude et al. [25] showed that color doppler ultrasound to detect

linear regression analysis showed that RI was positively correlated with each index of renal function. Therefore, we inferred that the func-

Table 5. Univariate Logistic regression analysis for the correlation of renal function and hemodynamic parameters at postoperative 1 h with AKI

Items Standardized partial re-gression coefficient (β) OR (95% CI) P value

Age 0.410* 1.51 (1.17, 1.93) 0.001Gender -0.047 1.33 (0.23, 7.83) 0.750Operative time 1.364* 3.91 (1.24, 12.30) 0.020Postoperative 1 h Crea 0.101* 1.11 (1.03, 1.20) 0.007Postoperative 1 h BUN 0.153 1.17 (0.69, 1.96) 0.564Postoperative 1 h UA 0.010 1.01 (0.998, 1.02) 0.112Postoperative 1 h PSV 0.231 1.26 (0.84, 1.89) 0.262Postoperative 1 h EDV -1.564* 0.21 (0.06, 0.80) 0.022Postoperative 1 h RI 63.709* 46.61 (6.25, 347.46) 0.002Note: *P < 0.05.

Figure 6. ROC curves of dif-ferent indices at postopera-tive 1 h.

renal artery RI was a holis-tic approach to assess renal small vessel circulation and renal parenchymal damage. Le Dorze et al. [24] showed that RI was in the normal range between 0.50 and 0.70, and 0.70 was generally considered to be the upper limit for normal adults, while RI ≥ 0.70 was considered the sign of acute kidney injury. In this study, preoperative RI values of AKI group and non-AKI group were less than 0.70, and there was no significant difference between the two groups. The change in RI was detected at postoperative 1 h in two groups, RI value of patients in AKI group was more than 0.07 at each time point. This might be associated with inten-sive vasoconstriction of the kidney caused by the release of vasoactive sub-stances produced by renin-angiotensin system, which was due to sympathetic nerve excitation during renal ischemia and hypoper-fusion after traumatic brain injury. With the conti- nuous supplement of blood volume, the kidney re- storing normal perfusion led to reperfusion injury, which caused the increase in oxygen free radical and microvascular damage, so peripheral vascular resis-tance increased signifi-cantly through the role of a variety of inflammatory mediators, and finally RI value increased. In addi-tion, linear correlation and

Table 6. Multivariate Logistic regression analysis for the correlation of renal function and hemodynamic parameters at postoperative 1 h with AKI

Standardized partial re-gression coefficient (β) OR (95% CI) P value

Age 0.41 1.51 (1.17, 1.93) 0.001Postoperative 1 h RI 3.265 38.20 (11.41, 127.84) 0.032Note: *P < 0.05.


1037 Int J Clin Exp Med 2016;9(2):1027-1038

tion of the kidney could be indirectly and early reflected by the RI value of interlobar renal artery.

ROC curve reflects dynamic change in sensitiv-ity and specificity with diagnostic threshold change, and it is recognized as the best way to evaluate diagnostic information and diagnostic decision quality [26]. The area under ROC curve is the most common parameter for evaluating ROC curve characteristics, and as a objective indicator for comparison of several diagnostic systems, its evaluation result are not subject to the change in diagnostic threshold [27]. This area can be considered the probability of cor-rect decision. It is generally considered that the area under ROC curve being between 0.50 and 0.70 indicates a lower accuracy; between 0.70 and 0.90 indicates a moderate accuracy and more than 0.90 indicates a higher degree of accuracy. In addition, in the ROC curve, the best diagnostic threshold can be chosen, and it is the most left top point of ROC curve, which has the best sensitivity and specificity and can improve the accuracy of diagnostic system. In this study, ROC curve was used for evaluation and comparison of the predictive value of post-operative 1 h RI, Crea, BUN, UA for diagnosis of acute kidney injury (AKI) following traumatic brain surgery. In ROC curve, the area under ROC curve of postoperative 1 h RI for predicting AKI was 0.805 (> the area under ROC curve of Crea, BUN and UA) with the sensitivity 90.00% and the specificity 67.70%, which suggested that RI has the predictive value for early diagno-sis of AKI. Darmon et al. [8] supported the results of our experiment and showed that RI value, with high sensitivity and specificity, was better than other renal function indices in the diagnosis of renal injury. In addition, its optimal judgment value was 0.70, indicating that post-operative 1 h RI > 0.70 increased the probabil-ity of AKI.

In this study, color Doppler flow imaging (CDFI) was used for detection of hemodynamic param-eters of interlobar renal arteries of both kid-neys, and the results showed that RI value had changed when serum Crea was within normal limits in the early period of AKI. At postopera-tive 1 h, RI value of AKI group was significantly higher, and it was still in a higher state at post-operative 48 h. Therefore, RI provided the basis for the early diagnosis of AKI.

Acknowledgements

This project is subject to the Second Affiliated Hospital of Soochow University Preponderant Clinic Discipline Group Project Funding and the Priority Academic Program Development of Jiangsu Higher Education Institutions.

Disclosure of conflict of interest

None.

Address correspondence to: Dr. Chen Wang, De- partment of Anesthesia and Critical Care, Suzhou Science & Technology Town Hospital, No. 1 Lijiang Road, Suzhou 215000, China. E-mail: chenwang- [email protected]

References

[1] Schirmer-Mikalsen K, Vik A, Gisvold SE, Skand-sen T, Hynne H, Klepstad P. Severe Head injury: control of physiological variables, organ failure and complications in the intensive care unit. Aeta Anaesthesiol Seand 2007; 51: 1194-1201.

[2] Bonventre JV, Vaidya VS, Schmouder R, Feig P, Dieterle F. Next-generation biomarkers for de-tecting kidney toxicity. Nat Biotechnol 2010; 28: 436-440.

[3] Bossard G, Bourgoin P, Corbeau JJ, Huntzinger J, Beydon L. Early detection of postoperative acute kidney injury by Doppler renal resistive index in cardiac surgery with cardiopulmonary bypass. Br J Anaesth 2011; 107: 891-898.

Table 7. ROC results of different indexes at postoperative 1 h

Variable value The optimal judgment value

The area under ROC curve

Standard error

Sensitivity (%)

Specificity (%)

95% confidence interval

Lower limit Upper limitPostoperative 1 h RI 0.70 0.805 0.064 90.00 67.70 0.679 0.930Postoperative 1 h Crea 62.35 0.727 0.074 70.00 67.70 0.582 0.873Postoperative 1 h BUN 4.89 0.588 0.08 90.00 41.90 0.431 0.744Postoperative 1 h UA 335.75 0.625 0.082 55.00 67.70 0.465 0.785

mailto:[email protected]

mailto:[email protected]


1038 Int J Clin Exp Med 2016;9(2):1027-1038

[4] Bouglé A, Duranteau J. Pathophysiology of sep-sis-induced acute kidney injury: the role of global renal blood flow and renal vascular re-sistance. Contrib Nephrol 2011; 174: 89-97.

[5] Kidher E, Harling L, Ashrafian H, Naase H, Chukwuemeka A, Anderson J, Francis DP, Atha-nasiou T. Pulse wave velocity and neutrophil gelatinase-associated lipocalin as predictors of acute kidney injury following aortic valve re-placement. J Cardiothorac Surg 2014; 9: 89.

[6] Corradi F, Brusasco C, Vezzani A, Palermo S, Altomonte F, Moscatelli P, Pelosi P. Hemorrha-gic shock in polytrauma patients: early detec-tion with renal Doppler resistive index measu-rements. Radiology 2011; 260: 112-118.

[7] Capotondo L, Nicolai GA, Garosi G. The role of color Doppler in acute kidney injury. Arch Ital Urol Androl 2010; 82: 275-279.

[8] Darmon M, Schortgen F, Vargas F, Liazydi A, Schlemmer B, Brun-Buisson C, Brochard L. Di-agnostic accuracy of Doppler renal resistive index for reversibility of acute kidney injury in critically ill patients. Intensive Care Med 2011; 37: 68-76.

[9] Mehta RL, Kellum JA, Shah SV, Molitoris BA, Ronco C, Warnock DG, Levin A; Acute Kidney Injury Network. Acute kidney injury network: report of an initiative to improve outcomes in acute kidney injury. Crit Care 2007; 11: R31.

[10] Bahar I, Akgul A, Ozatik MA, Vural KM, Demir-bag AE, Boran M, Tasdemir O. Acute renal fai-lure following open heart surgery: risk factors and prognosis. Perfusion 2005; 20: 317-322.

[11] Fischer UM, Weissenberger WK, Warters RD, Geissler HJ, Allen SJ, Mehlhorn U. Impact of cardiopulmonary bypass management on postcardiac surgery renal function. Perfusion 2002; 17: 401-406.

[12] Vandijck DM, Oeyen S, Decruyenaere JM, An-nemans L, Hoste EA. Acute kidney injury, le- ngth of stay, and costs in patients hospitalized in the intensive care unit. Acta Clin Belg Suppl 2007; 2: 341-345.

[13] Chertow GM, Soroko SH, Paganini EP, Cho KC, Himmelfarb J, Ikizler TA, Mehta RL. Mortality after acute renal failure: models for prognostic stratification and risk adjustment. Kidney Int 2006; 70: 1120-1126.

[14] Franklin SS, Gustin W 4th, Wong ND, Larson MG, Weber MA, Kannel WB, Levy D. Hemody-namic patterns of age-related changes in blood pressure. The Framingham heart study. Circulation 1999; 96: 308-315.

[15] Imuta N, Kinoshita N, Etani H, Yagita Y, Nukada T. There producibility of color Doppler Dupl- ex sonography in the measuerement of renal arterial blood velocity. Ultrasound Med Biol 1997; 23: 813-819.

[16] Deruddre S, Cheisson G, Mazoit JX, Vicaut E, Benhamou D, Duranteau J. Renal arterial re-sistance in septic shock: effects of increasing mean arterial pressure with norepinephrine on the renal resistive index assessed with Doppler ultrasonography. Intensive Care Med 2007; 33: 1557-1562.

[17] Duranteau J, Deruddre S, Vigue B, Chemla D. Doppler monitoring of renal hemodynamics: why the best is yet to come. Intensive Care Med 2008; 34: 1360-1361.

[18] Lauschke A, Teichgraber UK, Frei U, Eckardt KU. ‘Low-dose’ dopamine worsens renal perfu-sion in patients with acute renal failure. Kindey Int 2006; 69: 1669-1674.

[19] Tublin ME, Bude RO, Platt JF. The resistive in-dex in renal Doppler sonography: where do we stand? AJR Am J Roentgenol 2003; 180: 885-892.

[20] Petersen LJ, Petersen JR, Talleruphuus U, La-defoged SD, Mehlsen J, Jensen HA. The pulsa-tility index and the resistive index in renal arte-ries. Associations with long-term progression in chronic renal failure. Nephrol Dial Trans-plant 1997; 12: 1376-1380.

[21] Chi YW, White CJ, Thornton S, Milani RV. Ultra-sound velocity criteria for renal in-stent reste-nosis. Vasc Surg 2009; 50: 119-123.

[22] Zeller T, Rastan A, Rothenpieler U, Müller C. Restenosis after stenting of atherosclerotic re-nal artery stenosis: is there a rationale for the use of drug-eluting stents. Catheter Cardiovasc Interv 2006; 68: 125-130.

[23] Murphy ME, Tublin WE. Understanding the Doppler RI: impact of renal arterial distensibil-ity on the RI in a hydronephrotic ex vivo rabbit kidney model. Ultrasound Med 2000; 19: 303-314.

[24] Le Dorze M, Bougle A, Deruddre S, Duranteau J. Renal Doppler ultrasound: a new tool to as-sess renal perfusion in critical illness. Shock 2012; 37: 360-365.

[25] Bude RO, Rubin JM. Relationship between the resistive index and vascular compliance and resistance. Radiology 1999; 211: 411-417.

[26] Metz CE, Herman BA, Roe CA. Statistical com-parison of two ROC. Curve estimation obtained from partially-paired datasets. Med Decis Making 1998; 18: 110-121.

[27] Metz CE. ROC analysis in medical imaging: a tutorial review of the literature. Radiol Phys Technol 2008; 1: 2-12.

original article correlation study on postoperative ...comparison of preoperative renal function and...

Documents