We conducted an observational and prospective parallel study in two groups of patients that compared the safety and efficacy of a PEG laxative with electrolytes and another without for the treatment of chronic constipation.

The researchers obtained the consent of the patents before starting the period of observation.

PEG without electrolytes (Casenlax® powder for oral solution, 4g and 10g packets).

• Faecal impaction: 1.5–2g/kg/day in two doses until resolution for a maximum of six days (fixed dose).

• Constipation: 0.4–1g/kg/day in two doses to a maximum of 20g/day. The duration of treatment was of at least 12 weeks.

PEG with electrolytes (Movicol®, powder for oral solution, 6.9 and 13.9g packets).

• Faecal impaction: 1.5–2g/kg/day in two doses until resolution for a maximum of six days (fixed dose).

• Constipation: 0.4–1g/kg/day in two doses to a maximum of 27.8g/day. The duration of treatment was of at least 12 weeks.

We included the patients referred to the Department of Paediatric Gastroenterology of the Hospital Sant Joan de Déu, between November 2011 and March 2013. The inclusion criteria were: age between 1 and 17 years, failure of educational and dietary interventions implemented for a minimum of one month, having had faecal impaction as part of the ongoing process, and meeting the Rome III criteria for the diagnosis of chronic functional constipation (presence of at least two of the following prior to diagnosis, for at least one month in children aged less than 4 years13 or two months in older children5: 2 defecations or fewer per week, at least one episode of faecal incontinence per week, history of retentive posturing [age>4 years], History of painful or hard bowel movements, presence of faecal masses in the rectum, and history of large diameter stools that may obstruct the toilet).

The exclusion criteria included patients that had used medications that can affect intestinal motility such as laxatives, probiotics or prebiotics in the four weeks preceding the first visit; patients with defecation problems of an organic aetiology, such as Hirschsprung disease, spina bifida (occult form), hypothyroidism, coeliac disease, metabolic or renal disorders, and patients with suspected gastrointestinal obstruction or stenosis.

All patients included in the study were assigned to one treatment or another in an alternating fashion in order to obtain a similar number of patients in each group. The administered dose was determined based on whether the patient presented with or without faecal impaction (as specified above). The possibility of changing the dose was considered if the patient did not respond to treatment after three days or developed any adverse effects, such as diarrhoea or abdominal pain.

To improve patient acceptance of both products, we allowed their suspension in fruit juice or milk.

We requested that parents keep a diary during the followup, documenting the frequency of depositions per week, their shape and their consistency according to the Bristol scale, which rates the consistency of faeces on a scale from 1 to 7.14 Appointments were scheduled at weeks 1, 4 and 12 from initiation of treatment, and subsequently every 8 weeks until treatment could be discontinued. Each visit included a full history taking and physical examination, including a rectal exam. Potential adverse effects of the treatment were also documented: abdominal pain, diarrhoea, flatulence, nausea and product acceptance. The need to increase or decrease the dose of the laxative based on the patient's clinical condition and bowel movements recorded in the journal was evaluated during each visit.

The primary outcome measures were the number of bowel movements per week recorded at weeks 6 and 12 of treatment, and the presence of electrolyte imbalances at week 6. Secondary outcomes included the time required to achieve disimpaction, change in the consistency of the stools, use of laxatives and tolerance (episodes of nausea or abdominal pain).

A blood sample was collected at the beginning of the study, six weeks of treatment and whenever the clinician deemed it appropriate during follow-up visits. All were fasting samples, and were collected in the laboratory of the hospital. The tests performed included complete blood count, electrolyte panel, serum osmolality, glucose, urea, creatinine, total serum protein and albumin. A first-void urine sample was also collected for electrolyte and osmolality testing. Table 1 presents the laboratory result ranges.

No changes were made to the usual diet of patients during the entire period under study.

Based on previous studies,15,16 we decided that a difference of more than 30% between the treatments would be considered clinically relevant. Assuming an average frequency of seven bowel movements per week with a standard deviation of 3.5, we calculated that a sample of 60 (30 in each group) would be needed to require a 30% difference between the two groups in the number of bowel movements per week with a power of 80% (α=0.05).

We performed an intention-to-treat analysis of efficacy and biological and clinical safety that included all the data of enrolled patients that had received at least one dose of the laxatives under study and provided follow-up data for one or more of the analysed variables.

We performed comparisons between groups and within groups (before and after treatment) with the appropriate tests. We used the Kolmogorov–Smirnov test to ascertain the normality of the distributions. Quantitative parameters were compared by means of the Wilcoxon signed-rank test for paired samples and the Mann–Whitney U test for independent samples, and we described the data using the mean±standard deviation (SD) and the median. We summarised qualitative variables as frequencies and percentages, and compared them by means of the χ2 test or Fisher's exact test. The statistical analysis was performed with the software SPSS version 21. The level of statistical significance was 5%, and P-values of less than .05 were considered significant.

The number of bowel movements per week was similar in both groups at weeks 6 and 12. The mean (SD) number of bowel movements per week in the PEG+EL and the PEG groups were 6.0 (2.7) and 6.1 (2.5) at 6 weeks, and 5.4 (1.8) and 4.6 (2.2) at 12 weeks, respectively. The mean number of days elapsed until patients first passed soft stools was similar in both groups: 2.7 (1–10) days in the PEG group and 2 (1–7) days in the PEG+EL group, a difference that was not statistically significant (Table 3).

The proportion of patients with hard stools (Bristol ≤3) decreased considerably in the PEG group and to a lesser degree in the PEG+EL group, a difference that was not statistically significant. At 6 weeks, 21% (7 out of 32) had hard stools, compared to 6% (2 out of 30), respectively, but the differences between groups disappeared by 12 weeks, with hard stools found in 9% (9 out of 30) and 5% (5 out of 18) of patients, respectively (Table 4).

Faecal impaction at the time of enrolment was present in 37% of the PEG+EL group (12 out of 32) and 30% of the PEG group (9 out of 30). At 6 weeks of treatment, this symptom persisted in a higher percentage of patients in the PEG+EL group (21% [7/32]) compared to the PEG group (10% [3/30]), a difference that was not statistically significant. At 12 weeks, the difference had disappeared, with impaction found in only 6% (3 out of 31) and 9% (2 out of 30) of patients, respectively. The number of days after which patients first passed soft stools was very similar: it was 2 days on average in the PEG group, and 2.3 days in the PEG+EL group.

The mean doses at 6 weeks were 0.61g/kg/day (PEG) and 0.63g/kg/day (PEG+EL) (P=.8), and at 12 weeks, they were 0.56g/kg/day (PEG) and 0.61 (PEG+EL) (P=.49), with no statistically significant differences.

After six weeks of treatment, 56% of patients in the PEG+EL group (18 out of 32) and 83% in the PEG group (25 out of 30) had at least one abnormal laboratory result, a difference that was statistically significant (P=.02). This difference persisted when we analysed patients with more than two abnormal parameters, found in 28% of patients in the PEG+EL group (9 out of 32) and in 50% of the PEG group (15 out of 30), although the difference was not statistically significant (Table 5).

None of the patients had abnormal results in the complete blood count, total serum protein, albumin or transaminase levels.

Of the laboratory values that we analysed (Table 1), serum sodium was the parameter that was most frequently disturbed with treatment, and we found hyponatraemia (sodium≤135mEq/L) in up to 15% of the PEG+EL group (5 out of 32) and 36% of the PEG group (11 out of 30), a difference that was statistically significant (P=.05). The second abnormal result corresponded to urine potassium, found in 21% (5 out of 23) and 32% (9 out of 28), respectively. Finally, changes in urine osmolality were found in 17% of the PEG+EL group (4 out of 23) and 28% of the PEG group (8 out of 28). We did not find statistically significant differences in the baseline parameters of both groups.

Only three patients reported adverse effects that did not improve upon reducing the usual dose. The reported adverse effects were diarrhoea (one patient in the PEG group) and abdominal pain and flatulence (two patients in the PEG+EL group), and treatment was discontinued and switched in all of them. Only one patient in the PEG+EL reported that the product had a bad taste, leading to a change in treatment. None of the laboratory abnormalities were clinically relevant. All reported cases of hyponatraemia were mild, with a mean sodium level of 134mEq/L (133–135mEq/L).