Impact of thyroid function on cystatin C in detecting acute kidney injury: a prospective, observational study

This prospective observational study was conducted in the general ICUs in Guangdong General Hospital. All consecutive patients aged 18 years or elder admitted to ICUs from October 2014 to December 2016 were enrolled for the study. The exclusion criteria included refusal of consent, preexisting hyperthyroidism or hypothyroidism, medical history of hormone replacement therapy except insulin use, preexisting end-stage renal disease (ESRD) or renal dysfunction requiring RRT before admission, preexisting renal transplantation or nephrectomy, or missing admission data. The primary outcome of this study was the detection of AKI within 1 week after ICU admission, and the secondary outcome comprised length of ICU and hospital stay, RRT during ICU stay, as well as ICU and hospital mortality.

Each patient’s baseline characteristics, including age, gender, body mass index (BMI), preexisting clinical conditions, sepsis, admission type, baseline SCr, baseline estimated glomerular filtration rate (eGFR), acute physiology and chronic health evaluation (APACHE II) score, length of ICU stay, length of hospital stay, RRT during ICU stay, ICU mortality, and hospital mortality, were collected from the electronic medical system. The serum sample for measuring SCr, Cys C, total triiodothyronine (TT3, normal value range: 1.34–2.73 nmol/L), total thyroxine (TT4, normal value range: 78.40–158.40 nmol/L), free triiodothyronine (FT3, normal value range: 3.80–6.00 pmol/L), free thyroxine (FT4, normal value range: 7.50–21.10 pmol/L), thyroid-stimulating-hormone (TSH; normal value range: 0.34–5.60 μIU/mL), blood urea nitrogen (BUN), and albumin was collected at admission, and thereafter SCr was measured daily between 7 a.m. to 9 a.m. as a part of routine clinical care until ICU discharge. The eGFR was calculated using the Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) creatinine equation [21].

A baseline SCr was determined using the following rules ranked in the descending order of preference [22]: (1) the most recent pre-ICU value between 30 and 365 days before ICU admission (n = 143); (2) a stable pre-ICU value > 365 days for patients aged < 40 years (stable defined as within 15% of the lowest ICU measurement) before ICU admission (n = 6); (3) pre-ICU value > 365 days before ICU admission and less than the initial SCr at ICU admission (n = 51, 4) a pre-ICU value (between 3 and 39 days before ICU admission) less than or equal to the initial SCr on admission to ICU and not distinctly in AKI (n = 638); or (5) the lowest SCr upon initial admission to ICU (n = 173), the last ICU value (n = 216), or the minimum value at follow-up up to 365 days (n = 112). According to the KDIGO (Kidney Disease: Improving Global Outcomes) criteria, patients with an increase in SCr by ≥0.3 mg/dL (≥26.5 μmol/L) within 48 h, or increase in SCr to ≥1.5 times baseline within 1 week, or urine output < 0.5 mL/(kg ·h) for 6 h were diagnosed with AKI. However, since the urine output criteria has limited sensitivity when diuretics are administrated [23], the AKI diagnosis was based on serum creatinine in this study. Established AKI was defined as the diagnosis of AKI on entry, and later-onset AKI indicated no AKI diagnosis on entry but meeting the KDIGO criteria during the following 1 week after admission. Low T3 syndrome, also named non-thyroidal illness syndrome, is characterized by decreased serum T3 and, in severe illness, decreased serum T4, increased serum reverse T3 (rT3) concentrations, and normal or slightly decreased concentration of TSH [24, 25]. In this study, patients with an FT3 level below normal lower bound, and FT4 and TSH lower or within the normal range were diagnosed with low T3 syndrome according to the manufacturer’s standard value, as described in a previous study [26].

All samples were collected within 1 h after ICU admission and measured within 24 h after collection. SCr, serum Cys C, albumin, and BUN were measured using a UniCel DxC 800 Synchron System (Beckman Coulter, CA, USA). Cys C was assayed by immunoturbidimetry, the coefficients of interassay and intraassay variations for which were ≤ 5% and ≤ 10%, respectively. The thyroid function test was conducted via chemiluminescent immunoassay using Unicel DxI800 Synchron System (Beckman Coulter, CA, USA).

All statistical analyses were conducted using SPSS version 16.0 (SPSS, IL, USA; 2007), and MedCalc version 15.8 (MedCalc Software bvba, Ostend, Belgium; 2015) software programs. A two-tailed P value < 0.05 was considered statistically significant. The characteristics of patients with continuous and categorical data were presented with median (interquartile range, IQR) and absolute value (percentage), respectively. For continuous variables, normally distributed variables were compared using the t test and non-normally distributed variables were compared using the Wilcoxon rank-sum test. Categorical variables were compared using the chi-square test or Fisher’s exact test. The bivariate correlation analysis was used to evaluate the association between two variables. A multivariable linear regression analysis was conducted to identify the independent factors for Cys C and determine the representative index of thyroid function for further analysis. A receiver operating characteristic (ROC) curve analysis was performed, and the area under the curve (AUC) was calculated to demonstrate the diagnostic and predictive value of Cys C in detecting AKI. The difference between AUCs in each group was calculated using the Hanley–McNeil method [27], and the optimal cutoff value for AKI detection was determined with the Youden’s index using the MedCalc software.

Figure 1 presents the protocol and flow diagram of screening process. Among 1463 critically ill adult patients enrolled for the study, 124 were excluded for the following reasons: refused to consent (n = 45), complicated with thyroid dysfunction (hyperthyroidism or hypothyroidism, n = 24), history of nephrectomy (n = 10) or renal transplantation (n = 5), and ESRD (n = 40) prior to ICU admission. Finally, 1339 patients got involved in the study; AKI occurred in 402 patients, of which 225 were diagnosed with AKI at admission and 177 developed AKI during the following 1 week.

Basic clinical data and outcomes of the patients were demonstrated in Table 1. Compared with non-AKI patients, the patients with AKI were elder and had a significantly higher rate of complications, including hypertension (HTN), diabetes mellitus (DM), chronic kidney disease (CKD), coronary artery disease (CAD), chronic heart failure (CHF), stoke, and chronic liver disease. Sepsis was significantly associated with AKI, with an occurrence rate of 33.1%. The thyroid hormones concentration was significantly lower in the AKI group than in the non-AKI group, and patients in AKI group tended to be complicated with low T3 syndrome. Emergency patients and medical patients admitted to ICU were predisposed to be complicated with AKI. Patients with AKI showed significantly higher hospital and ICU mortality, and longer hospital and ICU stay.

As indicated in Table 2 showing the bivariate correlation analysis between Cys C and other factors, patients with a higher Cys C concentration were elder and had higher APACHE II score, SCr at admission, baseline SCr, and BUN. Moreover, albumin, and thyroid hormone concentrations, including FT3, TT3, and TT4 were negatively associated with Cys C. The multiple linear regression analysis shown in Table 3 indicated that among the abovementioned thyroid hormones, only FT3 (standardized β = − 0.100, P < 0.001) and TT3 (standardized β = 0.059, P = 0.007) were independently associated with Cys C. In addition, sex (standardized β = − 0.059, P < 0.001), age (standardized β = 0.136, P < 0.001), APACHE II score (standardized β = 0.087, P < 0.001), SCr at admission (standardized β = 0.578, P < 0.001), baseline SCr (standardized β = 0.066, P = 0.004), and BUN at admission (standardized β = 0.166, P < 0.001) were independent factors related to Cys C.

Cys C was significantly higher in the AKI group than the non-AKI group (Table 1). Figure 2 demonstrated the ROC analysis for the diagnostic and predictive ability of Cys C in detecting AKI. The AUCs calculated for Cys C was 0.753, 0.797, and 0.669 in detecting total AKI, established AKI, and later-onset AKI, respectively.

The patients were divided into five quintiles based on FT3 (Table 4), TT3 (Table 5), and FT4 (Additional file 1: Table S1), and AUC-ROC analysis was conducted to evaluate the impact of thyroid hormones on Cys C in detecting AKI. In the stratification of FT3 (Table 4), AUCs for Cys C in detecting total AKI, established AKI, and later-onset AKI was 0.777, 0.786, and 0.667, respectively, in quintile I; 0.740, 0.799, and 0.658, respectively, in quintile II; 0.721, 0.774, and 0.617, respectively, in quintile III; 0.727, 0.751, and 0.694, respectively, in quintile IV; and 0.700, 0.733, and 0.661, respectively, in quintile V. For quintiles of TT3 (Table 5), Cys C demonstrated its diagnostic value as indicated by AUCs of 0.784, 0.738, 0.730, 0.708, and 0.707, respectively, to detect total AKI; 0.804, 0.781, 0.786, 0.706, and 0.757, respectively, to detect established AKI; and 0.651, 0.652, 0.657, 0.690, and 0.670, respectively, to detect later-onset AKI. In summary, despite a tendency of decreasing AUCs across the quintiles by FT3 and TT3 in total and established AKI, no significant difference between AUCs was observed in any two groups for total AKI, established AKI, and later-onset AKI in different stratification of thyroid hormones.

Table 6 shows the subgroup analysis for the impact of low T3 syndrome on detecting AKI using Cys C. The patients were further separated into two categories: 650 with low T3 syndrome (FT3 < 3.8 pmol/L, and FT4 and TSH within or lower than the normal range), and 689 without low T3 syndrome. Compared with non-low T3 syndrome patients, the AUCs of Cys C in patients with low T3 syndrome were higher (0.762 vs 0.725, 0.789 vs 0.784, and 0.669 vs 0.663, respectively) in detecting total AKI, established AKI, and later-onset AKI. Furthermore, the optimal cut-off values of Cys C for AKI detection were markedly different between low T3 syndrome and non-low T3 syndrome patients. However, similar result was observed that AUCs for discrimination ability of AKI detection by Cys C did not differ significantly between groups.