Prognosis and Serum Creatinine Levels In Acute Renal Failure at the Time of Nephrology Consultation: An Observational Cohort Study

Jose Ramon Perez-Valdivieso; Maira Bes-Rastrollo; Pablo Monedero; Jokin de Irala; Francisco Javier Lavilla


BMC Nephrology 

In This Article


The University Hospital of Navarra is a tertiary care academic teaching medical center with 400 beds in the city of Pamplona, Spain. After hospital ethics committe approval, a cohort of 1008 consecutive patients who had been diagnosed as having acute renal failure, and had been admitted in our hospital between 1996 and 2006, was prospectively entered in a computerized database. Explicit patient consent was not required by the hospital ethics committee due to the observational nature of the study, and because collecting the data was a part of our academical-hospital routine work. Patients' anonymity was always strictly preserved. We restricted our analysis to patients who had enough increase of serum creatinine during their admission to fulfill the RIFLE criteria for ARF,[20] and were older than 16 years. Three hundred and six patients did not fulfill the RIFLE serum creatinine criteria nor presented oliguria during their stay. Forty-nine patients presented oliguria, but did not have substantial increase of creatinine. Seven patients were younger than 16 years. Finally, 646 patients were included for subsequent analysis.

ARF was defined and diagnosed if patients had showed substantial increment in serum creatinine during their admission to qualify for the RIFLE criteria. After that, patients were classified to the maximum RIFLE class according to the peak creatinine (the highest) reached during their hospital stay. On the contrary, we could not collect full information for urine output, so we were not able to classify the patients by that RIFLE criteria. For patients without chronic renal failure reported, the baseline creatinine was calculated using the Modification of Diet in Renal Disease (MDRD) equation,[24] as recommended by the ADQI workgroup,[20] assuming a glomerular filtration rate of 75 ml/minute/1,73 m2. For patients with a history of renal failure the baseline creatinine was defined as the one measured at hospital admission.[4] The term community-acquired ARF was used when the patient had presented ARF on admission to the hospital.

Demographic data, etiology of acute renal failure, comorbid conditions, severity of illness, and laboratory data were all prospectively collected. The difference between serum creatinine value at time when nephrologist first saw the case and baseline value was calculated and expressed as a percentage: Creatinine change (%) = [(creatinine when nephrologist saw the case - basal creatinine)/basal creatinine] * 100. Severity of illness was measured through Liano score (0.032*age in decades - 0.086*male gender - 0.109*nephrotoxic + 0.109*oliguria + 0.116*hypotension + 0.122*jaundice + 0.150*coma - 0.154*consciousness + 0.182*assisted respiration+ 0.210),[21] Karnofsky score at home,[25] and prior food intake were also determined and computed at that point. The patient's clinical status and treatment ARF were recorded daily. Serum creatinine at hospital discharge was recorded as a measure of recovery of renal function; the difference between this value and the baseline value was calculated and expressed in percentage: Creatinine change (%) = [(creatinine at hospital discharge - basal creatinine)/basal creatinine] * 100. Food intake was defined as previous caloric ingestion, and this was classified as appropriate when it was optimal, light malnutrition when it had been inappropriate less than three days, moderate malnutrition when it had been inappropriate between three and seven days, and severe malnutrition when it had been inappropriate for more than seven days. We followed up the patients after they were discharged from hospital to study the mortality at 6 months from the beginning of nephrology consultation.

Serum creatinine concentration was measured using the kinetic Jaffe assay.

In our hospital, when renal function is impaired, nephrology consulation can be requested, no matter in which hospital ward the patient stays. A nephrologist is always the only physician who provides a renal replacement therapy, according to a strict protocol. However, opinion from other medical departments involved in the patient's care is always sought. In this study, all patients were treated after the consultation by the same nephrologist, and all the data were gathered by the same observer.

The primary endpoints were in-hospital mortality and recovery of renal function at hospital discharge, following the recommendations from the ADQI.[26] The secondary endpoint was overall mortality at 6 months.

The increase in the percentage of creatinine reached before the nephrologist first saw the case was considered as the main exposure variable. The patients were divided into two groups using the median (101%) as the cut-off value. In addition, this cut-off point was the most optimal point from a ROC analysis. We obtained one below the median (group 1 or < 101% increment, hereafter), and one above the median (group 2 or ≥ 101% increment, hereafter). Non-conditional logistic regression models were fitted to assess the relationship between the increase of percentage in creatinine and the risk of in-hospital mortality. Odds ratios (OR) and their 95% confidence intervals (CI) were calculated considering the < 101% increment group as the reference category. The OR represents an estimation of relative risks of death during the inpatient stay.

In addition, linear regression models were used to assess the association between the increment of creatinine at time when nephrologists saw the case (exposure), and the recovery of renal function considering the increase in the percentage of creatinine at hospital discharge with respect to the baseline values (outcome). After excluding inhospital dead patients, we estimated the regression coefficient and their 95% CI for the larger increment group compared to the lesser increment group considered as the reference category. This coefficient represents the absolute difference in creatinine increment at hospital discharge between the ≥ 101% increment group and the < 101% increment group.

In both non-conditional logistic regression and linear regression, we fitted a crude model (univariate, i.e., without any adjustment), an age and gender model, and a multivariate-adjusted model including the following variables: Liano scoring, Karnofsky scoring, prior food intake, chronic renal failure, diabetes, treatment of acute renal failure, causes of acute renal failures, community-acquired acute renal failure, basal hemoglobin, basal serum albumin, and RIFLE classes selected by the descriptive univariate analysis of potential confounders with a p value less than 0.10. The Liano scoring includes in its equation the variables for nephrotoxicity, oliguria, hypotension, jaundice, mental status, and assisted respiration. Continuous variables were expressed as medians (and interquartile ranges), and compared using Mann-Whitney U test. Categorical variables were expressed as proportions and compared with the Chi-squared test. We evaluated all first-order multiplicative interactions (effect modifications) through product terms.

Overall survival at 6 months across groups was analyzed using the Kaplan-Meier methods, and differences between groups were tested using the log-rank test. The time of origin was the date when the nephrology consultation started. The event defined was death whereas those cases alive at the end of follow-up and those lost to follow-up were censored at their last observation.

All p values presented are two tailed, p < 0.05 was considered statistically significant. Statistical analyses were conducted using SPSS v.12.0.1 (SPSS Inc., Chicago, IL, USA).


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.