Using an artificial intelligence tool incorporating natural language processing to identify patients with a diagnosis of ANCA-associated vasculitis in electronic health records.

BACKGROUND: Because anti-neutrophil cytoplasmatic antibody (ANCA)-associated vasculitis (AAV) is a rare, life-threatening, auto-immune disease, conducting research is difficult but essential. A long-lasting challenge is to identify rare AAV patients within the electronic-health-record (EHR)-system to facilitate real-world research. Artificial intelligence (AI)-search tools using natural language processing (NLP) for text-mining are increasingly postulated as a solution. METHODS: We employed an AI-tool that combined text-mining with NLP-based exclusion, to accurately identify rare AAV patients within large EHR-systems (>2.000.000 records). We developed an identification method in an academic center with an established AAV-training set (n = 203) and validated the method in a non-academic center with an AAV-validation set (n = 84). To assess accuracy anonymized patient records were manually reviewed. RESULTS: Based on an iterative process, a text-mining search was developed on disease description, laboratory measurements, medication and specialisms. In the training center, 608 patients were identified with a sensitivity of 97.0 % (95%CI [93.7, 98.9]) and positive predictive value (PPV) of 56.9 % (95%CI [52.9, 60.1]). NLP-based exclusion resulted in 444 patients increasing PPV to 77.9 % (95%CI [73.7, 81.7]) while sensitivity remained 96.3 % (95%CI [93.8, 98.0]). In the validation center, text-mining identified 333 patients (sensitivity 97.6 % (95%CI [91.6, 99.7]), PPV 58.2 % (95%CI [52.8, 63.6])) and NLP-based exclusion resulted in 223 patients, increasing PPV to 86.1 % (95%CI [80.9, 90.4]) with 98.0 % (95%CI [94.9, 99.4]) sensitivity. Our identification method outperformed ICD-10-coding predominantly in identifying MPO+ and organ-limited AAV patients. CONCLUSIONS: Our study highlights the advantages of implementing AI, notably NLP, to accurately identify rare AAV patients within large EHR-systems and demonstrates the applicability and transportability. Therefore, this method can reduce efforts to identify AAV patients and accelerate real-world research, while avoiding bias by ICD-10-coding.

The role of intra- and interdialytic sodium balance and restriction in dialysis therapies.

The relationship between sodium, blood pressure and extracellular volume could not be more pronounced or complex than in a dialysis patient. We review the patients' sources of sodium exposure in the form of dietary salt intake, medication administration, and the dialysis treatment itself. In addition, the roles dialysis modalities, hemodialysis types, and dialysis fluid sodium concentration have on blood pressure, intradialytic symptoms, and interdialytic weight gain affect patient outcomes are discussed. We review whether sodium restriction (reduced salt intake), alteration in dialysis fluid sodium concentration and the different dialysis types have any impact on blood pressure, intradialytic symptoms, and interdialytic weight gain.

Bone Mineral Parameters in Peritoneal Dialysis Patients after Lowering Calcium Concentration in Dialysis Fluids: A Case Series in Patients Using Icodextrin.

In patients treated with peritoneal dialysis (PD), lowering the calcium level in PD fluids results in lower serum calcium levels and higher parathyroid hormone (PTH) levels. It is hypothesized that this effect is attenuated when patients are using icodextrin 7.5% for the once-daily long dwell (containing high calcium concentration). In this case series, we included 8 stable PD patients (mean age 68 ± 13 years, 7 male), all using icodextrin 7.5% (containing 1.75 mmol/L calcium) for the once-daily long dwell. The calcium content of the PD fluids for the remaining dwells was lowered from 1.75 mmol/L to 1.25 mmol/L. Bone mineral parameters and phosphate prescription at baseline, 6 weeks after this change, and after 6 months were compared. After lowering calcium concentration of the PD fluids - except for the icodextrin 7.5% - from 1.75 mmol/L to 1.25 mmol/L, calcium levels changed from 2.32 ± 0.11 to 2.29 ± 0.12 (p = NS); intact PTH (iPTH) from 39.6 ± 28.3 to 64.9 ± 34.5 pmol/L (p = 0.045); and alkaline phosphatase from 104.13 ± 48.75 to 101.38 ± 32.39 (p = NS). After 6 months, all bone mineral parameters were similar to baseline levels; however, slightly higher calcium-based phosphate binders were prescribed. Lowering calcium content from 1.75 mmol/L to 1.25 mmol/L in PD fluids in patients on icodextrin resulted in stable calcium values, a temporal increase in iPTH and a modest increase in calcium-based phosphate binder prescription. Using icodextrin for the long once-daily dwell appears to attenuate the effects on bone mineral parameters when lowering the calcium concentration of the short dwells.

Traditional and disease-related cardiovascular risk factors in ANCA-associated vasculitis: A prospective, two-centre cohort study.

OBJECTIVES: ANCA-associated vasculitis (AAV) has been associated with increased risk of cardiovascular (CV) events. The aim was to assess traditional and disease-related CV risk determinants in a two-centre prospective cohort of AAV patients. METHODS: Patients were recruited from centres in the Netherlands and Canada. A comprehensive CV risk assessment was performed at inclusion. Subjects were followed up yearly for 3-5 years until the first CV event, death or end of follow-up. Cox proportional hazards analyses were performed to relate baseline characteristics to the first CV event. RESULTS: A total of 144 patients were included (mean age 62 years, female sex 44%, median Framingham risk score 14.3%). Insulin resistance was present in 73% of patients tested at inclusion, independent of concurrent prednisone therapy. After a median follow-up of 2.90 years, 16 patients (11%) experienced a CV event (14 non-fatal and 2 fatal). The incidence of CV events was 5.45 per 100 patient-years. Age, Framingham risk score, HbA1c level, Diabetes Mellitus (DM), and previous CV event were significantly associated with CV events. Other factors, such as sex, impaired renal function, dyslipidemia, hypertension, smoking history and microalbuminuria, or disease-specific variables, like ANCA serotype or disease activity, were not significantly related to CV events in univariable or age-adjusted cox regression analysis. CONCLUSIONS: Determinants of an increased CV risk were identified. Disease-related factors and treatments can further modify individual risk factors, such as for steroids causing chronic insulin resistance and DM. Treatment of risk factors is essential to optimize long-term outcomes in AAV patients.

Predictors of fatal and non-fatal cardiovascular events in ANCA-associated vasculitis: Data from the Toronto CanVasc cohort.

OBJECTIVES: Patients with ANCA-associated vasculitis are at increased risk of cardiovascular events. The aim of the present study was to assess predictors of cardiovascular events in patients with granulomatosis with polyangiitis or eosinophilic granulomatosis with polyangiitis. METHODS: This retrospective cohort study included patients from the Canadian Vasculitis Research Network cohort in Toronto. Characteristics at diagnosis were collected. During follow-up, non-fatal cardiovascular events were determined from the Vasculitis Damage Index; mortality and causes of death were recorded. Cox regression models were developed to determine predictors of cardiovascular events, defined as stroke or myocardial infarction. RESULTS: A total of 336 patients were included (231 [69%] granulomatosis with polyangiitis; 105 [31%] eosinophilic granulomatosis with polyangiitis). The mean age at diagnosis was 44 (±18) years and 44% were male. The incidence rate for the combined outcome of all fatal and non-fatal events was 7.2 events per 1000 patient-years. In a multivariate model, family history of cardiovascular events and a higher Birmingham Vasculitis Activity Score at diagnosis were predictive of cardiovascular events (hazard ratio and 95% confidence interval 3.46 [1.06-11.28] and 1.09 [1.02-1.16] respectively). In a subgroup analysis there was no association between cardiovascular or disease-specific characteristics and cardiovascular events in eosinophilic granulomatosis with polyangiitis. CONCLUSIONS: In this cohort of patients with granulomatosis with polyangiitis and eosinophilic granulomatosis with polyangiitis, both traditional and disease-related risk factors were predictive of cardiovascular events. Further prospective studies should elucidate the impact of these and other modifiable risk factors on cardiovascular risk in ANCA-associated vasculitis.

Role of Albumin Assay on Calcium Levels and Prescription of Phosphate Binders in Chronic Hemodialysis Patients.

BACKGROUND/AIMS: In hemodialysis (HD) patients, the bromcresol green (BCG) assay overestimates, whereas the bromcresol purple (BCP) assay underestimates albumin concentration. Since corrected calcium concentrations depend on albumin, the albumin assay may have implications for the management of bone mineral disorders. METHODS: A subset of patients from CONTRAST, a cohort of prevalent HD patients, was analyzed. Bone mineral parameters and prescription of medication were compared between patients in whom albumin was assessed by BCP versus BCG. RESULTS: Albumin was assessed by BCP in 331 patients (9 of 25 centers) and by BCG in 175 patients (16 of 25 centers). Albumin was the lowest in the BCP group (34.5 ± 4.2 vs. 40.3 ± 3.1 g/L; p < 0.0005). Measured calcium levels and the prescription of calcium-based phosphate binders were similar in both groups. Corrected calcium levels, however, were markedly higher in the BCP group (2.45 ± 0.18 vs. 2.33 ± 0.18 mmol/L; p < 0.0005). CONCLUSION: These findings suggest that calcium levels are not corrected for albumin in clinical practice when considering the prescription of calcium-free or calcium-based phosphate-binders in dialysis patients.

Effect of Change in Fluid Status Evaluated by Bioimpedance Techniques on Body Composition in Hemodialysis Patients.

OBJECTIVE: This prospective study uses calf bioimpedance spectroscopy (cBIS) to guide the attainment of dry weight (DWcBIS) in chronic hemodialysis (HD) patients. The primary aim of this study was to evaluate whether body composition is altered when fluid status is reduced to DWcBIS. METHODS: Target post-HD weight was gradually reduced from baseline (BL) until DWcBIS was achieved. DWcBIS was defined as the presence of both flattening of the curve of extracellular resistance and the attainment calf normalized resistivity in the normal range during the dialysis treatment. Extracellular volume (ECV), intracellular volume, and total body water (TBW) were measured using whole body BIS (Hydra 4200). Fluid overload, lean body mass, and fat mass were calculated according to a body composition model. RESULTS: Seventy-three patients enrolled and 60 completed the study (55 ± 13 years, 49% male). Twenty-eight patients (25% diabetes) achieved DWcBIS, whereas 32 patients (47% diabetes) did not. Number of treatment measurements were 16 ± 10 and 12 ± 13 studies per patient in the DWcBIS and non-DWcBIS groups, respectively. Although significant decreases in body weight and ECV were observed, lean body mass and FM did not differ significantly in both groups from BL to the end of study. ECV, ECV/TBW, and fluid overload were higher in the non-DWcBIS than in the DWcBIS group both at BL and at the end of study. Ratios of intradialytic changes in calf normalized resistivity, ECV, and ECV/TBW to ultrafiltration volume were significantly lower in diabetic than in non-diabetic patients. CONCLUSIONS: This study shows that decreasing fluid status by gradual reduction of post-HD weight in both DWcBIS and Non-DWcBIS groups did not affect body composition significantly over a period of about 4 weeks.

Cardiovascular events in anti-neutrophil cytoplasmic antibody-associated vasculitis: a meta-analysis of observational studies.

OBJECTIVES: Several chronic inflammatory diseases are associated with cardiovascular disease, but the risk in ANCA-associated vasculitis is poorly quantified. The aim of the present study was to review the evidence for an increased cardiovascular risk, including ischaemic heart disease, cerebrovascular accidents and peripheral arterial disease, in patients with ANCA-associated vasculitis. METHODS: A comprehensive systematic review was conducted in accordance with guidelines of preferred reporting items for systematic reviews and meta-analyses. The databases PubMed, Embase.com and the Cochrane Library (Wiley) were searched for original observational studies comparing vasculitis patients with at least one control group. Summary estimates were derived with a random-effects model and reported as relative risks. RESULTS: One thousand three hundred and seventy-five studies were identified. Seven studies were included, comprising almost 14 000 ANCA-associated vasculitis patients vs general population controls in six studies and chronic kidney disease patients in one study. ANCA-associated vasculitis carried a relative risk of 1.65 (95% CI: 1.23, 2.22) for all cardiovascular events, 1.60 (95% CI: 1.39, 1.84) for ischaemic heart disease and 1.20 (95% CI: 0.98, 1.48) for cerebrovascular accidents. We did not find studies that addressed the risk for peripheral arterial disease separately. No heterogeneity was seen in the estimates. CONCLUSION: This meta-analysis of observational studies supports an increase in cardiovascular risk in patients with ANCA-associated vasculitis of ∼65%, similar to that found in other chronic inflammatory diseases. Hence, there is a clear need for active cardiovascular risk management in patients with ANCA-associated vasculitis.

Relation between duration of the prodromal phase and renal damage in ANCA-associated vasculitis.

BACKGROUND: In ANCA-associated vasculitis the acute phase of the disease is often preceded by prodromal symptoms. The aim of the present study was to analyze the relation between the duration of the prodromal phase and renal damage. METHODS: Patients with ANCA-associated vasculitis and renal involvement from a retrospective single-center cohort were divided into two equal groups based on the duration of the prodromal phase. The prodromal phase was defined as the time between first vasculitis related symptoms and the date of diagnosis. Clinical characteristics at diagnosis and renal items on the vasculitis damage index at 6 months were compared between the two groups. In addition, the relation between a long prodromal phase and 3-year end-stage renal disease and mortality as a composite outcome was studied. RESULTS: A total of 72 patients were included (age 64 ± 12 years; 74% male; 96% Caucasian). At diagnosis, in patients with a prodromal phase ≤22 weeks versus >22 weeks estimated glomerular filtration rate and proteinuria did not differ significantly (35 (interquartile range 50) versus 30 (50) ml/min p = 0.84; 75% versus 87%, p = 0.21 respectively). Furthermore, Birmingham Vasculitis Activity Scores were comparable (7 (3), p = 0.71). At 6 months, a long prodromal phase was associated with proteinuria (odds ratio 5.38, 95% confidence interval (CI) 1.47-19.62), but not with an estimated glomerular filtration rate ≤ 50 ml/min (odds ratio 0.89, 95% CI 0.33-2.37) in multivariable analyses. In addition, a long prodromal phase was associated with end-stage renal disease/mortality (hazard ratio 5.22, 95% CI 1.13-24.20). CONCLUSIONS: A long prodromal phase was associated with proteinuria and 3-year end-stage renal disease/mortality, but not with a reduced renal function at 6 months. These results underline the importance of an early diagnosis in ANCA-associated vasculitis patients in order to improve renal outcomes.

Optimization of anti-infective dosing regimens during online haemodiafiltration.

Online haemodiafiltration (HDF) is increasingly used in clinical practice as a routine intermittent dialysis modality. It is well known that renal impairment and renal replacement therapy can substantially affect the pharmacokinetic behaviour of several drugs. However, surprisingly few data are available on the need for specific dose adjustments during HDF. Due to convection, drug clearance may be increased during HDF as compared with standard haemodialysis. This may be of particular interest in patients undergoing anti-infective therapy, since under-dosing may compromise patient outcomes and promote the emergence of bacterial resistance. Drug clearance during HDF is determined by (i) dialysis characteristics, (ii) drug characteristics and (iii) patient characteristics. In this review, we will discuss these different determinants of drug clearance during HDF and advise on how to adjust the dose of antibacterial, antimycotic and antiviral agents in patients undergoing HDF. In addition, the possible added value of therapeutic drug monitoring is discussed. The review provides guidance for optimization of anti-infective dosing regimens in HDF patients.

Diagnosing ANCA-associated vasculitis in ANCA positive patients: A retrospective analysis on the role of clinical symptoms and the ANCA titre.

Currently no validated diagnostic system for antineutrophil cytoplasmic antibody (ANCA)-associated vasculitis (AAV) is available. Therefore, diagnosing AAV is often challenging. We aimed to identify factors that lead to a clinical diagnosis AAV in ANCA positive patients in a teaching hospital in The Netherlands.In this study, all patients that tested positive for ANCA proteinase 3 (PR3) and/or myeloperoxidase (MPO) between 2005 and 2015 were analysed. Patients with a clinical diagnosis of AAV were compared with patients without a clinical diagnosis of AAV. Clinical symptoms and laboratory variables at presentation, including the ANCA titre, were collected for both patients with and without AAV. Clinical and laboratory variables related with AAV were investigated, using multivariable logistic regression.Two hundred thirty seven consecutive patients with a positive ANCA were included, of whom 119 were clinically diagnosed with AAV. Of the 118 ANCA positive patients without AAV, 87 patients had an alternative diagnosis, including inflammatory bowel disease (n = 24), other rheumatic diseases (n = 23), infection (n = 11), malignancy (n = 4), and other diagnoses (n = 25). In a multivariable regression model, a high ANCA titre (odds ratio [OR] 14.16, 95% confidence interval [CI] 6.93-28.94) and a high number of affected organ systems (OR 7.67, 95% CI 3.69-15.94) were associated with AAV.MPO and PR3 ANCA can be positive in a variety of diseases that mimic AAV. A higher ANCA titre and multiple affected organ systems may help to discriminate between AAV and other systemic illnesses in anti-PR3 and anti-MPO positive patients. A diagnostic scoring system incorporating these factors should be considered.

Protein malnutrition and hypoalbuminemia as predictors of vascular events and mortality in ESRD.

BACKGROUND: It is proposed that chronic inflammation is common to the pathogenesis of malnutrition and vascular disease, both frequently observed in patients with end-stage renal disease. However, previous studies were unable to differentiate between true protein malnutrition and hypoalbuminemia. METHODS: This study was undertaken to determine the associations between malnutrition, measured by total-body nitrogen (TBN), and albumin, a marker of both nutritional status and chronic inflammation, with mortality and morbidity. One hundred nine patients starting dialysis therapy underwent nutritional assessment (TBN level and anthropometric measurements), vascular risk assessment (hypertension, hypercholesterolemia, diabetes mellitus, and smoking status), and serum albumin measurement. Subsequent patient mortality and new vascular events were recorded. RESULTS: Survival was associated independently with both TBN (hazard ratio [HR], 1.6; 95% confidence interval [CI], 1.1 to 2.5; P = 0.02 for every 10% decrease in nitrogen index) and serum albumin levels (HR, 1.1; 95% CI, 1.0 to 1.2; P = 0.004 for every 0.1-g/dL (1-g/L) decrease in serum albumin level) adjusted for other variables. Only low serum albumin level predicted subsequent vascular morbidity (HR, 2.2; 95% CI, 1.0 to 4.9; P = 0.049), as did increasing age (HR, 2.0; 95% CI, 1.4 to 3.0; P = 0.0004 for every 10-year increase in age) adjusted for other important risk factors. CONCLUSION: This study dissociates the effect of protein malnutrition and hypoalbuminemia on morbidity and mortality in patients starting dialysis therapy. Protein malnutrition and hypoalbuminemia are independently predictive of mortality, whereas hypoalbuminemia is predictive of vascular morbidity. The study supports the hypothesis that hypoalbuminemia is pathogenically associated with vascular disease, but dissociates this effect from protein malnutrition.

Prevalence of atherosclerotic renal artery stenosis in patients starting dialysis.

BACKGROUND: Atherosclerotic renal artery stenosis (ARAS) can lead to end-stage renal failure (ESRF). We determined the prevalence of ARAS in patients 45 years of age or older starting renal replacement therapy. METHODS: Forty-nine of 80 consecutive patients (37 males, 12 females) starting renal replacement therapy in our centre gave informed consent and underwent spiral computed tomographic angiography of their renal arteries. A renal artery diameter reduction of 50% or more assessed by two radiologists was considered as a significant stenosis. RESULTS: Twenty of 49 patients (41%) had an ARAS, and in eight cases (16%) this was bilateral or unilateral with a single kidney. Women were more likely to have an ARAS than men; 75 (9/12) vs 30% (11/37, P<0.01). However, relatively more women declined participation. Non-participants and participants did not differ in respect to other relevant clinical data. Nonetheless, findings in these patients would be negative, the prevalence of ARAS would still be 31% in women and 22% in men (NS). In 13 patients with ARAS the registered diagnosis of ESRF either was hypertension, renovascular disease or unknown. Assuming that in these patients atherosclerotic renovascular disease was the cause of renal failure, a total of 13 patients (13/49, 27%) entered the dialysis programme because of this problem. CONCLUSIONS: These results suggest that ARAS is an important cause of ESRF.