kidney stone predictive model

a recent study developed a clinical prediction rule for uncomplicated ureteral stones (see kidney stone predictive model bmj 2014 or doi: 10.1136/bmj.g2191).  there were 2 components to the study. first was a retrospective observational study to develop the screening tool, a random selection of 1040 adults (derivation cohort) who underwent non-contrast CT for suspected uncomplicated kidney stone from 2005 to 2010. their data was used to derive the top five factors associated with stones analysis and ascribe points reflecting their importance (see STONE score below). the second study was the validation study, 491 patients where the ER physicians felt that the patient presentation was consistent with ureteral stone.

    --for the observational component: the five key factors were -- male sex, short duration of pain, non-black race, presence of nausea or vomiting, and microsopic hematuria.

    --in the derivation and the validation cohorts (respectively)

            --STONE score of 0-5: 8.3% and 9.2% had stones (this reflected 19.8% and 15.5% of patients)

            --STONE score of 6-9: 51.6% and 51.3% had stones (this reflected 49.6% and 46.8% of patients)

            --STONE score of 10-13: 89.6% and 88.6% had stones (this reflected 30.6% and 37.7% of patients)

                   -- of note, in this high probability group, acutely important alternative findings were present in 0.3% and 1.6% (mostly diverticulitis, appendicitis, cholecystitis and a spattering of other diagnoses)

STONE score:

male sex =2

duration of pain: >24 hours =0, 6-24 hours =1, <6 hrs =3

race: nonblack=3

nausea alone=1, vomiting  alone=2

erythrocytes in urine: present =3

max total = 13

so, interesting study in that it could decrease the use of CT scans and the attendant ionizing radiation (average of 11.2 mSv), especially in the 1/3 of patients who are high risk. CT scans are the diagnostic procedure of choice in the US, though ultrasounds are preferred in Europe. the authors comment that "despite a 10-fold increase in the utilization of CT scanning for diagnosis of kidney stones from 1996-2007, the proportion of patients with a diagnosis of kidney stone, findings of significant alternative diagnoses, or hospital admission has not changed". one potential plus for CT in those in the high probability STONE score is to find the 20% or so with large stones who likely need an intervention. they offer the possibility of a substantially reduced dose CT scan (which could miss some important alternative findings, though present in <2% of their groups), or ultrasound (which sometimes cannot see the stone well). or possibly limiting the reduced dose CT to younger patients who are even less likely to have an important alternative finding.

but a few caveats: this is one study done in 2 centers at Yale, so needs to be repeated. 85% of patients were white (more common for stones in whites, but interestingly also more common in men and 46% of the patients were women). so not sure about generalizability to other regions, ethnicities. however, i am always concerned about the amount of ionizing radiation we subject patients to. the thought of low-dose CT is appealing, though ultrasound makes a lot of sense to me for initial imaging.