Correspondence *Department of Colorectal Surgery, Cleveland Clinic Florida, 2950 Cleveland Clinic Boulevard, Weston, FL 33331, USA

Email deanh@ccf.org

Commentary

The study by Singal et al.¹ was conducted in a retrospective manner which entails many obvious limitations. As serum assays were not routinely taken, the authors had to indirectly derive this information by retrospectively identifying those patients whose conditions required the measurement of serum creatinine before and after bowel preparation. This methodology resulted in a restricted study population: the cohort was relatively advanced in age (mean age: sodium phosphate [NaP] group 65.6 0.9 years and polyethylene glycol [PEG] group 68.8 0.9 years) with significant comorbidity; therefore, the study population was not representative of all patients who receive bowel preparation. Secondly, serum creatinine was measured at different time intervals among patients in both groups (time interval precolonoscopy: NaP group 102 80 days, PEG group 84 81 days; time interval postcolonoscopy: NaP group 97 87 days, PEG group 78 76 days). This design flaw weakens the correlation that was found between change in serum creatinine level and the ingestion of the bowel preparation. Furthermore, changes in serum creatinine could be attributed to other factors such as the timing of taking the blood sample, which is hard to determine in a retrospective study.

We previously conducted a prospective, randomized study that compared the safety of oral NaP with that of PEG for bowel preparation in 450 individuals undergoing colonoscopy.² All patients were weighed, and serum electrolytes and osmolarity were measured before and after patients received bowel preparation and within a fixed time frame to assure homogeneity. We found that changes in serum electrolytes associated with NaP ingestion were transient, not clinically significant, and returned to within normal limits less than 24 h afterwards. This study design has been implemented in other studies that have reinforced our results.^{3, 4}

A meta-analysis that compared NaP with PEG for bowel preparation identified 25 randomized studies that met its inclusion criteria. Among these studies the occurrence of serious adverse events with oral NaP was zero.⁵ However, a study by Markowitz et al.⁶ identified 21 cases of acute phosphate nephropathy in patients with a recent history of receiving oral NaP. These 21 cases were derived from 7,349 renal biopsies. A literature search revealed 15 case reports that described severe adverse events occurring with NaP bowel preparation. These cases presented to their medical care providers with impaired renal function and elevated serum creatinine from a few days up to a few weeks after the ingestion of NaP. All of these patients had histopathologic evidence of phosphate nephropathy.

In the study by Singal and colleagues,¹ approximately 55% of the patients in the oral NaP group had hypertension, and about 28% had diabetes; these conditions are certainly potential contributing factors to nephropathy. In addition, owing to the study's retrospective design, there was a lack of information about the volume of water intake associated with NaP ingestion. This could be one of the main causes of potential subclinical dehydration resulting in a higher probability of renal function impairment following colonoscopy, especially in older patients.

The change in serum creatinine in the oral NaP group was an increment from 1.0 0.02 to 1.1 0.02 mg/dl (P = 0.07), and in the PEG group was a decrease from 1.1 0.02 to 1.0 0.03 mg/dl (P = 0.03). In general, these findings are in agreement with those of prior studies,^{2, 3} but the inconsistent nature of the timing of taking these measurements makes these data potentially unreliable. However, such findings could have been supported by objective evidence for an association between impaired renal function and NaP preparation intake had associated histopathology been identified that revealed phosphate nephropathy at the time of serum creatinine measurement. Alternatively, a lack of correlation might have been shown if other types of nephropathy, such as diabetic or hypertensive nephropathy, had been identified.

In a logistic regression analysis, NSAID use was the only factor in the study that was significantly associated with both a 25% increase (P = 0.05) and a 50% increase (P <0.001) in serum creatinine. Other variables including age, diabetes mellitus, hypertension, coronary artery disease, use of diuretics and use of angiotensin-converting-enzyme inhibitors, were not associated with either a 25% or a 50% increase in serum creatinine (P >0.1). By contrast, these variables were significantly associated with increased serum creatinine in the study by Markowitz et al.⁶ In addition, in most of the 15 case reports at least one of these variables was significantly associated with increased serum creatinine levels. The FDA⁷ and the joint guidelines of ASGE, ASCRS and SAGES⁸, therefore, recommend agents other than NaP to be used when one or more of the previously mentioned variables exist. Furthermore, serious adverse renal events occurred in patients with normal serum creatinine baseline results but with comorbidities, such as hypertension, diabetes mellitus, congestive heart failure, liver failure and impaired kidney function, and in those receiving medications such as NSAIDS, angiotensin-converting-enzyme inhibitors, diuretics and angiotensin-receptor blockers.

Singal and colleagues are to be commended for having provided in their discussion a very strong and very critical appraisal of their own work that emphasized all the weak points of their study.

In conclusion, this study provides somewhat weak evidence for any meaningful conclusion because of the inconsistent measurement of serum creatinine throughout the study. Reliable evidence could be produced by conducting a prospective study in which serum creatinine and electrolytes were measured at predetermined times, including prior to colonoscopy and perhaps 1 day and 3 days after colonoscopy. This analysis could be performed in three main patient groups: young age, old age but otherwise normal (control) and old age with nephropathy-predisposing comorbidities.

References

1. Singal AK et al. (2008) The renal safety of bowel preparations for colonoscopy: a comparative study of oral sodium phosphate solution and polyethylene glycol. Aliment Pharmacol Ther 21: 41–47
2. Cohen SM et al. (1994) Prospective, randomized, endoscopic-blinded trial comparing precolonoscopy cleansing methods. Dis Colon Rectum 37: 689–696 | Article | PubMed | ChemPort |
3. Marshall JB et al. (1993) Prospective randomized trial comparing sodium phosphate solution with polyethylene glycol electrolyte lavage for colonoscopy preparation. Gastrointest Endosc 39: 631–634 | PubMed | ChemPort |
4. Golub RW et al. (1995) Colonoscopic preparation-which one? A blinded prospective randomized trial. Dis Colon Rectum 58: 594–597 | Article |
5. Belsey J et al. (2007) Systematic review: oral bowel preparation for colonoscopy. Aliment Pharmacol Ther 25: 373–384 | PubMed | ChemPort |
6. Markowitz GS et al. (2005) Acute phosphate nephropathy following oral sodium phosphate bowel purgative: an under recognized cause of chronic renal failure. Am Soc Nephrol 16: 3389–3396 | Article | ChemPort |
7. U.S. Food and Drug Administration (FDA). FDA alert report 5/2006. Oral sodium phosphate (OSP) products for bowel preparation. [http://www.fda.gov/cder/drug/infopage/osp_solution/default.htm] (accessed 23 June 2008)
8. Wexner SD et al. (2006) A consensus document on bowel preparation before colonoscopy: Prepared by a task force from the American Society of Colon and Rectal Surgeons (ASCRS), the American Society for Gastrointestinal Endoscopy (ASGE), the Society of American Gastrointestinal and Endoscopic Surgeons (SAGES). Dis Colon Rectum 49: 792–809 | Article | PubMed |

Contact the journal about this article

Subject areas under which this article appears: Large intestine | Endoscopy