Association of prescription of oral sodium polystyrene sulfonate with sorbitol in an inpatient setting with colonic necrosis: a retrospective cohort study.

BACKGROUND

Colonic necrosis has been reported after sodium polystyrene sulfonate (SPS)/sorbitol use, but the incidence and relative risk (RR) are not established.

STUDY DESIGN

Retrospective cohort study.

SETTING & PARTICIPANTS: 123,391 adult inpatients at a tertiary medical center.

PREDICTOR

Receipt of SPS prescriptions (exposed) or a prescription other than SPS (unexposed internal comparison group) between September 1, 2001, and October 31, 2010.

OUTCOMES

The main outcome measure was tissue-confirmed diagnosis of colonic necrosis, considered SPS-associated if SPS was prescribed 30 or fewer days before tissue accession date.

MEASUREMENTS

Demographics, serum chemistry test results, hospital location, and International Classification of Diseases, Ninth Revision diagnostic codes.

RESULTS

SPS was prescribed to 2,194 inpatients. 82 inpatient colonic necrosis cases were identified. 3 received oral SPS (1 gram per 4 milliliters of 33% sorbitol) 30 or fewer days before the colonic necrosis accession date (3.7% of inpatient colonic necrosis cases). The data were linked with 123,391 individuals who received inpatient prescriptions between the same dates. Colonic necrosis incidence was 0.14% (95% CI, 0.03%-0.40%) in those prescribed SPS versus 0.07% (95% CI, 0.05-0.08%) in those not prescribed SPS (RR, 2.10; 95% CI, 0.68-6.48; P = 0.2). The number needed to harm was 1,395 (95% CI, 298-5,100). Subgroup analysis (age >65 years; estimated glomerular filtration rate, <30 mL/min/1.73 m(2), intensive care unit admission, or surgical ward status) did not show significant associations. Sample-size analysis indicated that 4,974 SPS-treated individuals older than 65 years and a comparison group 10 times larger would be required for rigorous multivariate analysis of SPS-associated colonic necrosis risk.

LIMITATIONS

Individuals with colonic necrosis admitted to non-Department of Defense hospitals would not have been ascertained. Only individuals who had colonic biopsy or surgical tissue submitted for pathologic review could be ascertained as having colonic necrosis.

CONCLUSIONS

SPS-associated colonic necrosis is rare, and inpatient SPS/sorbitol prescription was not associated significantly with an increased RR of colonic necrosis in this retrospective cohort analysis. Multivariate analysis would require retrospective clinical cohorts from larger or more than one hospital system(s).