The baseline demographics and clinical characteristics of the 20 subjects in this study are summarized in Table . We studied 12 cases who developed at least Stage I AKI and 8 control subjects who did not. Cases and controls were similar with regard to sex and race. On average, cases were younger than controls (57 ± 12 years versus 70 ± 17 years, p = 0.05) and had lower APACHE II scores (27 ± 11 versus 17 ± 8, p = 0.04). There was no statistically significant difference for in-hospital mortality rates between the two groups.
Baseline characteristics of patients with and without acute kidney injury
Subjects with AKI had a baseline serum creatinine of 67 ± 15 μmol/L with a peak inpatient serum creatinine of 217 ± 86 μmol/L compared with a baseline serum creatinine of 69 ± 20 μmol/L and a peak serum creatinine of 81 ± 17 in non-AKI subjects (p < 0.001 for peak levels). As noted earlier, we excluded potential study subjects with a baseline creatinine greater than 97 μmol/L to avoid patients with underlying CKD. No patient had AKI attributable to rhabdomyolysis. Eight patients had acute tubular necrosis, two patients had AKI after orthotopic liver transplantation, one had cardiorenal syndrome, and one had multifactorial AKI. Two patients had Stage I AKI by the AKIN criteria [41
], five patients had Stage II AKI, and five patients had Stage III AKI. Four of 12 (33%) of the AKI subjects were treated with dialysis.
Subjects with and without AKI had mean ionized calcium levels of 1.19 ± 0.1 mmol/L and 1.15 ± 0.08, respectively (p
= 0.41). Serum phosphorus levels were significantly higher in AKI subjects compared with controls (4.5 ± 1 mmol/L versus 3.3 ± 1.1 mmol/L, p
= 0.02). The median intact PTH level was 63 mg/dL (25-75% interquartile range (IQR), 38-213) in AKI subjects and 70 mg/dL (25-75% IQR, 58-126) in controls (Figure , p
= 0.73). When severe hyperparathyroidism was defined as an intact PTH >250 mg/dL, a level that has been associated with increased cardiovascular disease risk in prior studies [43
], none of the control subjects had a severe hyperparathyroidism but 3 of 12 (25%) of the AKI subjects did (although this result did not meet conventional levels of statistical significance, p
PTH and FGF-23 levels in non-AKI and AKI subjects.
FGF-23 levels were significantly higher in critically ill AKI cases compared with critically ill non-AKI subjects, with a median FGF-23 level of 1948 RU/mL (IQR, 437-4369) in AKI cases compared with 252 RU/mL (IQR, 65-533) in critically ill controls (p = 0.01; Figure ). After adjusting for age and APACHE II as potential confounders, AKI remained a significant predictor of log-transformed FGF-23 levels (Table ). Among patients with AKI, FGF-23 levels were higher in nonsurvivors (n = 4) compared with survivors (n = 8), with respective median levels of 4446 RU/mL (IQR, 3455-5443) versus 544 RU/mL (IQR, 390-1948; p = 0.02). Although serum phosphorus and FGF-23 levels were both elevated in AKI subjects, no correlation was observed between the two variables, as shown in Figure (r = 0.08, p = 0.74). There was a correlation between PTH and FGF-23 levels (r = 0.55, p = 0.02); when this analysis was restricted to patients with AKI, this correlation only had borderline statistical significance (r = 0.58, p = 0.05), likely due to the small size of the cohort.
Association of log-transformed FGF-23 levels with AKI (multivariable linear regression)
Correlation of serum phosphorus (PO4) and FGF-23 levels in patients with and without acute kidney injury. No correlation was observed between PO4 and FGF-23 levels (r = 0.08, p = 0.74).