Added value of non-contrast CT for the diagnosis of acute pyelonephritis in older patients with suspected infection with an unknown focus: a retrospective diagnostic study

Abstract

Objectives: In older patients, the diagnosis of acute pyelonephritis (APN) is challenging. The aim was to evaluate the added value of CT to history, physical examination and urinalysis for the diagnosis of APN in older patients with suspected infection with an unknown focus.

Design: Retrospective diagnostic study.

Setting: Department of General Medicine in an acute care hospital in Japan.

Participants: Patients aged ≥65 years who underwent blood cultures, a urine culture, and chest and abdominal CT to detect the focus of infection were included.

Primary outcome measures: Two radiologists independently reviewed four non-contrast CT signs: perirenal fat stranding, pelvicalyceal wall thickening, enlargement of the kidney and thickening of Gerota's fascia. Findings on contrast-enhanced CT could not be evaluated due to an insufficient number of patients in whom contrast-enhanced CT was performed. An expert panel was used as the reference standard for APN. The added value of CT findings was quantified by comparing the diagnostic performance between a model based on 10 predictors available before CT and an extended model including the CT findings.

Results: Of 473 patients, 61 (14.8%) were diagnosed with APN. When the laterality of the CT findings was taken into account, the model fit was not improved by adding them. In the laterality-insensitive analysis, the model performance was significantly improved by adding the CT signs (likelihood-ratio test p=0.03; c-index 0.89 vs 0.91, p=0.03). However, their clinical utility was only to improve the classification of 11.5% of patients with APN.

Conclusions: The added value of non-contrast CT findings to history, physical examination and urinalysis was limited for the diagnosis of APN in older patients with a suspected infection with an unknown focus.

Keywords: Computed tomography; Pyelonephritis; Urinary tract infections.

Figure 1

Findings of CT in acute pyelonephritis. (A) An 86-year-old man with acute pyelonephritis. Axial non-contrast CT scan of the abdomen showing perirenal fat stranding (white arrowheads). (B) An 86-year-old woman with acute pyelonephritis. Axial non-contrast CT scan of the abdomen showing pelvicalyceal wall thickening (white arrowheads). (C) An 84-year-old woman with acute pyelonephritis. Axial non-contrast CT scan of the abdomen showing renal enlargement (white arrowheads). (D) A 66-year-old woman with acute pyelonephritis. Axial non-contrast CT scan of the abdomen showing thickening of Gerota’s fascia (white arrowheads).

Figure 2

Calibration plots of the base and extended models. Ideally, all decile groups of predicted probabilities fit close to the dashed diagonal line (perfect calibration). The base model shows underprediction around the estimated risk of 0.10, which is improved by adding CT findings in the laterality-insensitive analysis.