Changes in renal function following administration of oral sodium phosphate or polyethylene glycol for colon cleansing before colonoscopy

GI Diseases Research Unit, Hotel Dieu Hospital, Queen’s University, Kingston, Ontario Correspondence: Dr Stephen Vanner, Hotel Dieu Hospital, 166 Brock Street, Kingston, Ontario K7L 5G2. Telephone 613-544-3310, fax 613-544-3114, e-mail vanners@hdh.kari.net Received for publication May 11, 2006. Accepted July 18, 2006 R Abaskharoun, W Depew, S Vanner. Changes in renal function following administration of oral sodium phosphate or polyethylene glycol for colon cleansing before colonoscopy. Can J Gastroenterol 2007;21(4):227-231.

O ral sodium phosphate (NaP) is one of the most effective colon cleansing agents for colonoscopy currently available (1-6) and is better tolerated than large-volume polyethylene glycol (PEG) solutions.Although numerous clinical studies (2,7,8) suggest that oral NaP has a good safety profile, it is a small volume, osmotically active solution that causes transient hyperphosphatemia (1,7) and has the potential to induce intravascular volume contraction (1,2,9).There have been reports (2,10) of serious adverse events, including death (11), in a small number of isolated cases with oral NaP, but these adverse events appeared to result from inappropriate dosing, improper patient selection (eg, use in patients with contraindications such as renal failure) and/or poor hydration during the cleansing period (2).
Recently, there have been case reports (12)(13)(14) of nephrocalcinosis with renal failure associated with the use of oral NaP.This relationship was first reported in a single case study (14) and subsequently in a case series of 21 patients at a large referral centre (12).These latter cases were identified by reviewing a large renal biopsy database and finding a temporal relationship between nephrocalcinosis and ingestion of oral NaP.The mechanism underlying this relationship is unclear, but these reports raise the concern that there may be more frequent systematic renal damage that went clinically undetected (12,14).
We have a large database of patients who underwent serial colonoscopies for screening and had serum creatinine (Cr) measurements obtained before their colonoscopies.This afforded the opportunity to examine changes in their renal function over time following use of oral NaP.Because some patients received PEG, we could compare changes in renal function following oral NaP use with changes occurring after the use of PEG.

METHOD
The database from the endoscopy unit at the Hotel Dieu Hospital, Queen's University (Kingston, Ontario), for outpatient colonoscopies performed between 1995 to 2004 was reviewed.Ethics approval was obtained from the Queen's University School of Medicine Institutional Review Board.All patients who underwent two colonoscopies, three months to nine years apart (with almost 80% of the subsequent procedures between the first and fifth year), and had a serum Cr measured before each colonoscopy were included.These measurements were routinely drawn within two weeks before the procedure in both the initial and follow-up procedure.Patient age, sex, medications, comorbidities and colonic preparation were recorded for each case (Table 1).
Patients received one of two commercially available bowel preparations: oral NaP or oral PEG.For both regimens, patients were instructed to drink clear fluids with their preparation the day before their procedure.The dose of oral NaP prescribed was 45 mL taken at 17:00 and at 22:00.In addition, patients were instructed to drink at least 2 L to 3 L of fluid that day.Patients receiving GoLYTELY (PEG) were required to drink 4 L of preparation in 3 h starting at 18:00 the day before their procedure.
The development of renal insufficiency following ingestion of a cleansing agent was defined in two ways: a serum Cr level above the upper limit of normal at repeat measurement in those patients who had normal baseline serum Cr level (ie, ratio of Cr/baseline Cr greater than one), or an estimated Cr clearance (CrCl) using the Cockroft-Gault equation ( 15): Over the course of the present study, the laboratory changed the normal reference value for the upper limit of normal from 80 μmol/L to 110 μmol/L.
Data analysis was performed using SPSS (SPSS Inc, USA) statistical software version 12.0.The comparison of continuous variables was performed using Student's t test.Categorical variables were compared using the Fisher's exact test.Multivariate logistic regression was used to determine which variables contributed to the development of renal insufficiency in the study population.

Patient characteristics
A total of 767 patients (n=618 for oral NaP, n=149 for PEG) with complete data were identified from the database at the Hotel Dieu Hospital between 1995 and 2004.Patients received oral NaP or PEG depending on the preference of the clinician, the patient, or if there was a contraindication to oral NaP.To mitigate against selection bias, 75 PEG patients were identified in 2002 and 2003, at a time when PEG was used almost exclusively in the endoscopy unit.Nonetheless, comparison of their baseline characteristics (Table 1) identified some differences between the groups.There were more PEG patients receiving antihypertensives, bisphosphonates and nonsteroidal anti-inflammatory drugs.Similarly, there were more patients with diabetes, coronary artery disease and hypertension in this group.The mean age of the oral NaP group was slightly younger than the PEG group (54.1±11.7 years versus 59.2±13.4years, respectively; P<0.001), although the age range was similar.Overall, 50.7% of patients were female.The mean time between colonoscopies in the oral NaP and PEG groups was statistically significant (3.7±1.8 years versus 1.0±1.5 years, respectively).The median time between colonoscopies was three years in the oral NaP group and one year in the PEG group (Figure 1).

Comparison of renal function in oral NaP and PEG groups
Renal function, assessed by serum Cr measurements and estimated CrCl determined at baseline (ie, before the first colonoscopy) and at the second measurement (ie, before the second colonoscopy), was compared for the two groups (Table 2).The baseline mean Cr level was slightly higher in the oral NaP group than in the PEG group, but the difference of 3 μmol/L is minimal and not significant.Moreover, CrCl calculated using the Cockroft-Gault equation was similar between oral NaP and PEG preparations (101.9 mL/min versus 101.1 mL/min, respectively; P=0.807).Repeat Cr measurements were obtained three months to nine years later (almost 80% were between the first and fifth year).After colon cleansing, the mean serum Cr level and CrCl did not differ between the two groups.The change in serum Cr between baseline and repeat measurement was significantly different, favouring stable renal function in the oral NaP group, but these changes were  small and clinically insignificant.Furthermore, the mean change in CrCl from baseline to repeat measurement was not statistically significant and remained within the normal range.

Comparison of the development of renal insufficiency in oral NaP and PEG groups
Of the 767 patients identified, 55 developed renal insufficiency based on our criteria.Only a small proportion of patients exhibited abnormal renal function in each group (6.8% of patients in the oral NaP group versus 8.7% of patients in the PEG group; P=0.382), and there was no difference between the groups (Figure 2).All these patients had a normal serum Cr level and estimated CrCl at baseline, and there were no differences between the two groups in mean serum Cr level or CrCl at baseline or following colon cleansing (Table 3).When medications were compared in this subset of oral NaP and PEG patients, only antihypertensive medication use was different (11.9% versus 46.2%, respectively; P=0.014).When this group of patients was compared with the cohort of patients who did not develop renal failure, there were significantly more patients in the renal insufficiency group who were on diuretics and antihypertensives.There were also significantly more patients with renal insufficiency who had hypertension or coronary artery disease.
A multiple logistic analysis was performed with development of renal insufficiency (ie, Cr ratio greater than one) as the dependant variable.The following factors were analyzed: preparation used, age, sex, comorbid conditions and medications, as listed in Table 1.The only predictors of the development of renal failure were age (P=0.014) and hypertension (P=0.001).) Proportion of patients developing renal insufficiency in the oral sodium phosphate (NaP) and polyethylene glycol (PEG) groups as defined by a repeat serum creatinine measurement above the upper limit of normal (6.8% versus 8.7%, respectively; P=0.382).There was no difference in the proportion of patients developing renal insufficiency with regard to the preparation used  No significant differences were found between oral sodium phosphate (NaP) and polyethylene glycol (PEG) groups with regard to their renal function both at baseline and at repeat measurement.NS Nonsignificant The variables in this model predicted the development of renal insufficiency in 93% of our cases.Preparation used was not an independent predictor of the development of renal insufficiency.

Comparison of the magnitude of renal insufficiency in the oral NaP and PEG groups
In the subset of patients (n=55) who developed renal insufficiency, the serum Cr increase above the upper limit of normal was small, and the distribution and magnitude did not differ between the oral NaP and PEG groups (Figure 3).Similar results were obtained when the changes in estimated CrCl from baseline were compared for the two groups (Figure 4).

DISCUSSION
The present study examined changes in renal function in patients who had received oral NaP as a colon cleansing agent before colonoscopy and compared them with a group who received PEG.Two major findings emerged from the present study -first, the proportion of patients developing abnormal Cr values over a three-month to nine-year period in a large cohort of patients after receiving oral NaP was small (6.8%) and did not differ from those receiving PEG (8.7%); second, the magnitude of the abnormalities was small (ie, the peak Cr level in both groups was less than 160 μmol/L).Furthermore, Cr measurements were drawn farther apart in the oral NaP group than in the PEG group (median three years versus one year).This would bias the results toward worsening renal function in the oral NaP group given the longer interval to develop renal insufficiency in those patients with predisposing comorbidities (eg, hypertension).Nonetheless, this worsening renal function was not observed in the oral NaP group.Our analysis of the medications and comorbidities of the oral NaP and PEG groups demonstrated that some differences existed between these groups (Table 1).Given that hypertension and associated medications, coronary artery disease and diabetes mellitus are risk factors for renal failure, it raises the possibility that these differences may confound the findings (ie, increase the rate of renal insufficiency in the PEG group).We did, however, find that the estimated baseline CrCl rates were not different between the two groups, which would argue that renal function at baseline was, in fact, similar between the groups.Nonetheless, given the potential confounders, we performed multivariate analysis to determine independent risk factors.This demonstrated that preparation alone cannot predict the development of renal failure.In view of these findings, as well as the relatively small numbers of patients affected, we concluded that oral NaP does not result in frequent, preparation-specific renal damage that had gone clinically undetected.However, given the limitations of our data set, we cannot exclude the possibility that oral NaP does lead to subclinical renal damage in a very small proportion of patients.A recent preliminary report (16) provides further support for the notion that oral NaP does not result in widespread, subclinical renal damage.In this study, the relationship between ingestion of oral NaP and coincident use of angiotensin-converting enzyme inhibitors (ACEIs) or angiotensin receptor blockers (ARBs) on renal function was examined prospectively in 191 patients.They did not find any abnormalities in serum Cr levels following ingestion of oral NaP, regardless of whether the patients were taking ACEIs or ARBs.However, patients were strongly encouraged to drink fluids during the cleansing period, as were patients in our study, to maintain hydration.The hydration status of patients in previous case reports (2,12,14) of renal failure is almost always unknown.It seems prudent to encourage ingestion of fluids, including oral rehydration solutions, given that oral NaP has the potential to induce intravascular volume contraction (17).
Oral hydration may be particularly important in patients with underlying renal disease or in those on medications that may alter renal blood flow.The mechanism linking the recent reports of nephrocalcinosis, renal failure and ingestion of oral NaP ( 12) is currently unknown.However, most of the patients (16 of 21) were hypertensive, and 14 were either taking ACEIs or ARBs, while four were on diuretics.It is possible that these patients were predisposed because of their medications and that inadequate oral hydration was a link in at least some of these cases.Our multivariate analysis also identified that age and hypertension were independent predictors of renal failure.The variables in our logistic regression model correctly predicted the presence or absence of renal insufficiency in 93% of the cases.These are well-established risk factors for renal disease (5,18,19).However, when taken together with recent reports of nephrocalcinosis and renal failure, it seems prudent to recommend that particularly careful attention be applied to hydration in patients with risk factors and drugs that may compromise renal function.

SUMMARY
The present retrospective study demonstrated that ingestion of oral NaP before colonoscopy had not resulted in widespread renal damage that had gone clinically undetected.Nonetheless, careful attention must be taken with patient selection (contraindicated in patients with renal failure, symptomatic ischemic heart disease, congestive heart failure, ascites and ileus; caution in those at extremes of age and those with hypertension or multiple serious comorbidities) and appropriate dosing of oral NaP (45 mL dose taken twice daily, 6 h to 12 h apart) to prevent renal toxicity.There is growing evidence that adequate hydration (at least 2 L to 3 L of clear fluid or oral rehydration solution throughout the cleansing period) is important during colon cleansing with oral NaP.

CONFLICT OF INTEREST DISCLOSURE:
The present investigator-driven study was supported by an unrestricted educational grant from Fleet Laboratories, USA. Dr Vanner has served on the speaker bureau, and both Dr Vanner and Dr Depew have a patent pending with this company on oral rehydration solutions and oral sodium phosphate.

YearsFigure 1 )
Figure 1) Box plot showing median interval between colonoscopies in the oral sodium phosphate (NaP) and polyethylene glycol (PEG) groups.Bold line represents the median.The median interval was three years (range three months to nine years) in the oral NaP group and one year (range three months to eight years) in the PEG group

Figure 3 )Figure 4 )
Figure3) Elevation of creatinine above the upper limit of normal in oral sodium phosphate (A) and polyethylene glycol (B) groups developing renal insufficiency.The magnitude of the change in creatinine concentration in both groups was small and insignificant (13.9 μmol/L versus 10.3 μmol/L in the oral sodium phosphate and polyethylene glycol groups, respectively; P=0.403)

TABLE 1
Patient demographics and risk factors (n=767)