Health Care Analytics With Time-Invariant and Time-Variant Feature Importance to Predict Hospital-Acquired Acute Kidney Injury: Observational Longitudinal Study.

BACKGROUND: Acute kidney injury (AKI) develops in 4% of hospitalized patients and is a marker of clinical deterioration and nephrotoxicity. AKI onset is highly variable in hospitals, which makes it difficult to time biomarker assessment in all patients for preemptive care. OBJECTIVE: The study sought to apply machine learning techniques to electronic health records and predict hospital-acquired AKI by a 48-hour lead time, with the aim to create an AKI surveillance algorithm that is deployable in real time. METHODS: The data were sourced from 20,732 case admissions in 16,288 patients over 1 year in our institution. We enhanced the bidirectional recurrent neural network model with a novel time-invariant and time-variant aggregated module to capture important clinical features temporal to AKI in every patient. Time-series features included laboratory parameters that preceded a 48-hour prediction window before AKI onset; the latter's corresponding reference was the final in-hospital serum creatinine performed in case admissions without AKI episodes. RESULTS: The cohort was of mean age 53 (SD 25) years, of whom 29%, 12%, 12%, and 53% had diabetes, ischemic heart disease, cancers, and baseline eGFR <90 mL/min/1.73 m2, respectively. There were 911 AKI episodes in 869 patients. We derived and validated an algorithm in the testing dataset with an AUROC of 0.81 (0.78-0.85) for predicting AKI. At a 15% prediction threshold, our model generated 699 AKI alerts with 2 false positives for every true AKI and predicted 26% of AKIs. A lowered 5% prediction threshold improved the recall to 60% but generated 3746 AKI alerts with 6 false positives for every true AKI. Representative interpretation results produced by our model alluded to the top-ranked features that predicted AKI that could be categorized in association with sepsis, acute coronary syndrome, nephrotoxicity, or multiorgan injury, specific to every case at risk. CONCLUSIONS: We generated an accurate algorithm from electronic health records through machine learning that predicted AKI by a lead time of at least 48 hours. The prediction threshold could be adjusted during deployment to optimize recall and minimize alert fatigue, while its precision could potentially be augmented by targeted AKI biomarker assessment in the high-risk cohort identified.

Neutrophil-to-Lymphocyte Ratio Predicts Development of Immune-Related Adverse Events and Outcomes from Immune Checkpoint Blockade: A Case-Control Study.

The utility of neutrophil-to-lymphocyte ratio (NLR) and platelet-to-lymphocyte ratio (PLR) utility in predicting immune-related adverse events (irAEs) and survival have not been well studied in the context of treatment with immune checkpoint inhibitors (ICIs). We performed a case-control study of cancer patients who received at least one dose of ICI in a tertiary hospital. We examined NLR and PLR in irAE cases and controls. Logistic and Cox regression models were used to identify independent risk factors for irAEs, progression-free survival (PFS), and overall survival (OS). The study included 91 patients with irAEs and 56 controls. Multiple logistic regression showed that NLR < 3 at baseline was associated with higher occurrence of irAEs. Multivariate Cox regression showed that development of irAEs and reduction in NLR from baseline to week 6 were associated with longer PFS. Higher NLR values at baseline and/or week 6 were independently associated with shorter OS. A reduction in NLR from baseline to week 6 was associated with longer OS. In this study of cancer patients treated with ICIs, NLR has a bidirectional relationship with adverse outcomes. Lower NLR was associated with increased occurrence of irAEs while higher NLR values were associated with worse clinical outcomes.

Cumulative iodinated contrast exposure for computed tomography during acute kidney injury and major adverse kidney events.

OBJECTIVES: To determine if contrast-enhanced CT imaging performed in patients during their episode of AKI contributes to major adverse kidney events (MAKE). METHODS: A propensity score-matched analysis of 1127 patients with AKI defined by KDIGO criteria was done. Their mean age was 63 ± 16 years with 56% males. A total of 419 cases exposed to CT contrast peri-AKI were matched with 798 non-exposed controls for 14 covariates including comorbidities, acute illnesses, and initial AKI severity; outcomes including MAKE and renal recovery in hospital were compared using bivariate analysis and logistic regression. MAKE was a composite of mortality, renal replacement therapy, and doubling of serum creatinine on discharge over baseline; renal recovery was classified as early versus late based on a 7-day timeline from AKI onset to nadir creatinine or cessation of renal replacement therapy in survivors. RESULTS: Sixty-two patients received cumulatively > 100 mL of CT contrast, 143 patients had > 50-100 mL, and 214 patients had 50 mL or less; MAKE occurred in 34%, 17%, and 21%, respectively, as compared with 20% in non-exposed controls (p = 0.008 for patients with > 100 mL contrast versus none). More contrast-exposed patients experienced late renal recovery (27% versus 20%) and longer hospital days (median 10 versus 8) than non-exposed patients (all p < 0.01). On multivariate analysis, cumulative CT contrast > 100 mL was independently associated with MAKE (odds ratio 2.39 versus non-contrast, adjusted for all confounders, p = 0.005); cumulative CT contrast under 100 mL was not associated with MAKE. CONCLUSIONS: High cumulative volume of CT contrast administered to patients with AKI is associated with worse short-term renal outcomes and delayed renal recovery. KEY POINTS: • Cumulative intravenous iodinated contrast for CT imaging of more than 100 mL, during an episode of acute kidney injury, was independently associated with worse renal outcomes and less renal recovery. • These adverse outcomes including renal replacement therapy were not more frequent in similar patients who received cumulatively 100 mL or less of CT contrast, compared with non-exposed patients. • More patients with CT contrast exposure during acute kidney injury experienced delayed renal recovery.

Nitrates in combination with hydralazine in cardiorenal syndrome: a randomized controlled proof-of-concept study.

AIMS: Cardiorenal syndrome (CRS) is a common problem of great morbidity and mortality. Hydralazine-isosorbide dinitrate (H-ISDN) may be used in renal failure and may improve exercise capacity in heart failure (HF). Our proof-of-concept study aimed to evaluate early evidence of efficacy, safety, and feasibility of H-ISDN compared with standard of care in CRS. METHODS AND RESULTS: This multi-centre, single-blind, randomized trial in Singapore enrolled CRS patients, defined as chronic HF with concomitant renal failure [estimated glomerular filtration rate (eGFR) < 60 ml/min/1.73 m2 ]. The primary outcome was 6 min walk test (6MWT) distance measured at 6 months. Secondary outcomes included study feasibility; efficacy outcomes which included renal, cardiac, and endothelial functions, health-related quality of life using Short Form-36, clinical outcomes; and adverse events. Forty-four patients [71 ± 10 years; 75% male; median (inter-quartile range) N-terminal prohormone brain natriuretic peptide 1346 (481-2272) pg/mL] with CRS (left ventricular ejection fraction 42 ± 12% and eGFR 46 ± 15 ml/min/1.73 m2 ) were randomized into two equal groups. Of these, 39 (89%) had hypertension, 27 (61%) had diabetes mellitus, and 17 (39%) had atrial fibrillation. Six (27%) discontinued H-ISDN owing to intolerance and poor compliance. There was a trend towards improved 6MWT distance with H-ISDN compared with standard of care at 6 months (mean difference 27 m; 95% CI, -12 to 66), with little differences in secondary efficacy outcomes. Giddiness and hypotension occurred more frequently with H-ISDN, but HF hospitalizations and mortality were less. CONCLUSIONS: Our pilot study does not support the addition of H-ISDN on top of standard medical therapy to improve exercise capacity in patients with CRS.

High-sensitivity Troponin I Predicts Galectin-3 in Chronic Kidney Disease Patients.

PURPOSE: Plasma galectin-3 (pG3) regulates inflammation. B-type natriuretic peptide (BNP), high-sensitivity Troponin I (hsTnI), and pG3 concentrations are elevated in chronic kidney disease (CKD) patients. The associations of pG3 with hsTnI/BNP are unclear. We explored the relationship of hsTnI and BNP with pG3 in Asian CKD patients and healthy controls. METHODS: We retrieved prospectively collected frozen plasma samples from 163 stable CKD patients and 105 healthy controls. BNP, hsTnI and pG3 were assayed. pG3 was assessed for associations with age, gender, ethnicity, blood pressures; height, weight, body mass index (BMI), previously diagnosed CKD, diabetes, hypertension, coronary artery disease, estimated glomerular filtration rate (eGFR); C-reactive protein, beta-trace protein, 24 h urine protein, serum albumin, uric acid and cystatin C. We created two models predicting pG3 using multiple linear regression. Akaike Information Criterion (AIC) was used for comparison. Significance was taken at P < 0.05. RESULTS: CKD versus healthy participants: mean BMI (28.2 vs. 24.9 kg/m2), median serum creatinine (159 vs. 69 µmol/L; 1.8 vs. 0.78 mg/dL), median eGFR (49 vs. 104 mL/min/1.73m2), median pG3 (29.4 vs. 15.4 ng/mL), median BNP (136 vs. 23 pg/mL), and median hsTnI (12.5 vs. 2.6 pg/mL). By univariate analysis, all variables are associated with pG3 except weight, gender and diagnosis of cerebrovascular or peripheral vascular diseases. A parsimonious model selected for hsTnI, BMI, serum albumin, cystatin C and eGFR (AIC = 77.6). CONCLUSION: BNP and hsTnI are associated with pG3 in Asian CKD patients. hsTnI is a better predictor of pG3.

Serial Quantification of Urinary Protein Biomarkers to Predict Drug-induced Acute Kidney Injury.

BACKGROUND: Drug-induced Acute Kidney Injury (AKI) develops in 10-15% of patients who receive nephrotoxic medications. Urinary biomarkers of renal tubular dysfunction may detect nephrotoxicity early and predict AKI. METHODS: We prospectively studied patients who received aminoglycosides, vancomycin, amphotericin, or calcineurin inhibitors, and collected their serial urine while on therapy. Patients who developed drug-induced AKI (fulfilling KDIGO criteria) were matched with non-AKI controls in a 1:2 ratio. Their urine samples were batch-analyzed at time-intervals leading up to AKI onset; the latter benchmarked against the final day of nephrotoxic therapy in non- AKI controls. Biomarkers examined include clusterin, beta-2-microglobulin, KIM1, MCP1, cystatin-C, trefoil-factor- 3, NGAL, interleukin-18, GST-Pi, calbindin, and osteopontin; biomarkers were normalized with corresponding urine creatinine. RESULTS: Nine of 84 (11%) patients developed drug-induced AKI. Biomarkers from 7 AKI cases with pre-AKI samples were compared with those from 14 non-AKI controls. Corresponding mean ages were 55(±17) and 52(±16) years; baseline eGFR were 99(±21) and 101(±24) mL/min/1.73m2 (all p=NS). Most biomarker levels peaked before the onset of AKI. Median levels of 5 biomarkers were significantly higher in AKI cases than controls at 1-3 days before AKI onset (all µg/mmol): clusterin [58(8-411) versus 7(3-17)], beta-2-microglobulin [1632(913-3823) versus 253(61-791)], KIM1 [0.16(0.13-0.76) versus 0.07(0.05-0.15)], MCP1 [0.40(0.16-1.90) versus 0.07(0.04-0.17)], and cystatin-C [33(27-2990) versus 11(7-19)], all p<0.05; their AUROC for AKI prediction were >0.80 (confidence intervals >0.50), with average accuracy highest for clusterin (86%), followed by beta-2-microglobulin, cystatin-C, MCP1, and KIM1 (57%) after cross-validation. CONCLUSION: Serial surveillance of these biomarkers could improve the lead time for nephrotoxicity detection by days.

Transitional care program to facilitate recovery following severe acute kidney injury.

BACKGROUND: Patients with acute kidney injury needing prolonged renal replacement therapy (AKI-RRT) may benefit from a structured care process in form of an AKI transitional care program (ATCP), to facilitate RRT weaning and recovery. METHODS: We examined outcomes following ATCP implementation in adults with AKI-RRT from a tertiary institution (versus pre-ATCP controls), including mortality, cumulative hospital days, and renal function over one year; RRT and haemodialysis catheter days in initial 90 days. RESULTS: We studied 89 patients with age 62 ( ± 15) years. 47% had septic AKI, 20% cardiorenal syndrome, and 29% had baseline eGFR < 30 mL/min/1.73 m2. Comparing 45 ATCP patients with 44 controls: 64% and 45% received continuous RRT (CRRT) (p = 0.07), with comparable rates of heart failure (24% versus 25%), ICU care (67% versus 70%), RRT successfully weaned (71% versus 75%), respectively; corresponding mortality rates were 24% and 32% (p = 0.44), hospital days of 205 (197-213) and 223 (215-232) per 1000 patient-days alive over one year (p = 0.002); with comparable RRT and catheter days. Serial serum creatinine in months following RRT cessation were comparable between either survivor-group. On multivariate analysis, heart failure or having received CRRT independently predicted mortality and longer hospital days (p < 0.05); ATCP was independently associated with reduced hospital days (p < 0.001). 17 ATCP patients and 14 controls required outpatient RRT weaning, with catheter days of 607 (568-648) and 683 (638-731) per 1000 patient-days in initial 90 days, respectively (p = 0.01). CONCLUSIONS: Implementing a structured care pathway in patients with AKI-RRT may help reduce hospitalization, and reduce haemodialysis catheter days in the subgroup for outpatient RRT weaning.

Intravenous maintenance fluid tonicity and hyponatremia after major surgery- a cohort study.

BACKGROUND: Intravenous maintenance fluid (IMF) tonicity and composition influence plasma electrolyte balance. OBJECTIVE: To determine if hypotonic IMF therapy contributes to post-surgical hyponatremia. SETTING: Single-center tertiary institution. PARTICIPANTS: Adults who underwent major surgery and received peri-surgical IMF, with exclusive administration of hypotonic pre-mixed 0.33% saline, 5% dextrose and potassium chloride (DK0.33%S), or isotonic 0.9% saline with or without 5% dextrose (NS/DNS). OUTCOMES AND MEASURES: We examined post-surgical hyponatremia, hypokalemia and acute kidney injury (AKI), associated with use of either IMF. RESULTS: We studied 659 patients, of whom 161 patients (24%) developed post-surgical hyponatremia. DK0.33%S (versus NS/DNS) IMF was administered in 52% of patients who developed hyponatremia, compared to 38% of patients with stable natremia (p = 0.001). More patients with hyponatremia underwent gastrointestinal-hepatobiliary or abdominal (GI/HBS/Abd) surgery versus other surgical-sites (p = 0.001). Hypokalemia developed in 1% versus 10% of patients who received DK0.33%S and NS/DNS IMF respectively (p< 0.001), with corresponding AKI rates of 3% versus 7% (p = 0.02). On multivariate analysis, adjusted for timing of biochemistry post-surgery, IMF infusion rate and volume; independent factors associated with post-surgical hyponatremia included DK0.33%S administration, GI/HBS/Abd surgery (versus other sites), and post-surgical AKI (p < 0.05). Subgroup analysis by surgical sites showed that association of DK0.33%S administration with hyponatremia was most evident in GI/HBS/Abd surgery. CONCLUSIONS: Administration of DK0.33%S IMF, compared with NS/DNS, is associated with post-surgical hyponatremia in adults after major surgery, but with less hypokalemia. The higher rate of AKI observed with NS/DNS IMF requires further evaluation.

Electronic health records accurately predict renal replacement therapy in acute kidney injury.

BACKGROUND: Electronic health records (EHR) detect the onset of acute kidney injury (AKI) in hospitalized patients, and may identify those at highest risk of mortality and renal replacement therapy (RRT), for earlier targeted intervention. METHODS: Prospective observational study to derive prediction models for hospital mortality and RRT, in inpatients aged ≥18 years with AKI detected by EHR over 1 year in a tertiary institution, fulfilling modified KDIGO criterion based on serial serum creatinine (sCr) measures. RESULTS: We studied 3333 patients with AKI, of 77,873 unique patient admissions, giving an AKI incidence of 4%. KDIGO AKI stages at detection were 1(74%), 2(15%), 3(10%); corresponding peak AKI staging in hospital were 61, 20, 19%. 392 patients (12%) died, and 174 (5%) received RRT. Multivariate logistic regression identified AKI onset in ICU, haematological malignancy, higher delta sCr (sCr rise from AKI detection till peak), higher serum potassium and baseline eGFR, as independent predictors of both mortality and RRT. Additionally, older age, higher serum urea, pneumonia and intraabdominal infections, acute cardiac diseases, solid organ malignancy, cerebrovascular disease, current need for RRT and admission under a medical specialty predicted mortality. The AUROC for RRT prediction was 0.94, averaging 0.93 after 10-fold cross-validation. Corresponding AUROC for mortality prediction was 0.9 and 0.9 after validation. Decision tree analysis for RRT prediction achieved a balanced accuracy of 70.4%, and identified delta-sCr ≥ 148 µmol/L as the key factor that predicted RRT. CONCLUSION: Case fatality was high with significant renal deterioration following hospital-wide AKI. EHR clinical model was highly accurate for both RRT prediction and for mortality; allowing excellent risk-stratification with potential for real-time deployment.