Transabdominal ultrasound of rectal diameter in healthy infants: a prospective cohort study during the first year of life

Transabdominal rectal ultrasound (TRU) is used to measure transverse rectal diameter (TRD) in order to diagnose functional constipation (FC) and megarectum, and to evaluate treatment. The proposed cut‐off value is 3.0 cm. Currently, no standardised values exist for children below the age of 4. We used repeated TRUs to establish reference TRD values in healthy infants and to describe rectal diameter in infants with FC.

Transabdominal rectal ultrasound (TRU) measures transverse rectal diameter (TRD) in order to help diagnose childhood functional constipation (FC) and was first reported by Klijn et al. 1 and Singh et al. 2 The technique is well tolerated and often used in paediatric outpatient clinics.TRU is commonly used to investigate whether FC is a cause of bladder dysfunction, 1,3,4 and to strengthen the FC diagnosis.It is also used to evaluate treatment 1,2,5,6 and to diagnose or evaluate bowel management in children with megarectum, a condition often associated with anorectal malformation 7 and great morbidity. 8ublished data on TRD as used in diagnosing FC does not generally deal with infants younger than one and defining a cut-off value for TRD is difficult.The suggested limit of 3.00 cm diameter 5 produces a major overlap between children with FC and healthy controls, 9 and ultrasound scanning gives insufficient evidence of a diagnostic association between FC and TRD. 3,10One reason is the lack of age-related normative TRD data.Some reports 2,[11][12][13] show that rectal diameter correlates to age, but no standardised values exist for children younger than 4.
FC is a clinical diagnosis based on at least two of the five Rome IV criteria 14 and the dominating criteria in infants <2 years are hard bowel movements and faecal impaction. 14In infancy, both formula feeding and the change to solid feeds correlate with hard stools. 15The pathogenesis of FC probably starts with a painful defecation leading to stool retention. 16It is still unknown whether a dilated rectum is the result, or the cause, of FC. 17 The primary aim of this study was to provide a large dataset of TRDs in healthy children at 2 and 12 months of age, so presenting a normal reference value.Secondary aims were to report rectal diameters in children who developed FC and to investigate whether infants who later presented with FC had a greater rectal diameter before diagnosis.

Study design
This study comprised measurements of TRD in healthy term infants, at 2 and 12 months of age.This is part of an ongoing longitudinal birth-cohort study of healthy children from newborn to school age previously described in detail. 18,19At 2, 6 and 12 months of age, interviews were carried out using questionnaires to collect data on bowel habits and identify children with FC (according to Rome III criteria). 15A 3-day bowel diary was completed at every follow-up, reporting frequency and consistency of stools using a reference chart (permission from Weaver et al.). 20arents were asked to contact the study centre when defecation problems occurred.If questionnaires indicated FC, or parents reported problems, it was followed by a visit to the physician where a medical history and a physical examination were done where organic disease was excluded.If FC diagnosis was confirmed, treatment was given, and an additional TRU was offered at an outpatient clinic visit.Infants who had FC were regarded as constipated for the rest of the study.This was due to the short study period and relapses being common, and that no clear guidelines exist on when to be regarded as being free from FC.

Subjects
From September 2014 to September 2019, healthy babies were enrolled from a maternity department at a university hospital in Sweden.Inclusion criteria were term pregnancy (gestational age at birth > 36 weeks), mothers' first child and at least one member of the family able to understand Swedish.Exclusion criteria were surgery, disease or medication affecting defecation.Written informed consent was obtained from parents at inclusion.

Ultrasound protocol
The TRU was performed with the patient in the supine position using an Aloka Prosound 6 ultrasound system (SECMA AB Gothenburg, Sweden) with an SOP-Alpha 6-24 AIP; 3.5 MHz curved array transducer.The transducer was situated 1-2 cm above the symphysis at a 10-15 downward angle, as described by Klijn et al. 1 The bladder was identified, and volume was measured.Expected bladder capacity was calculated according to the ICCS standardisation document. 21The rectum was identified and the maximal TRD in cm was measured three times.The diameter was calculated as the mean of the three measurements, and intra individual SD of measurements was also calculated.Time since last void and bowel movement was noted.In babies with FC, enemas and laxatives were given by parents if prescribed and noted in the patient's chart.

Questionnaires and bowel diary
Questions included patient demographics, growth, feeding, bowel habits and FC symptoms according to the Rome III criteria. 15Defecation frequency was in numbers of stools per day or per week.Consistency was divided into: runny, soft, hard and very hard.The questionnaires used have previously been published. 19

Statistical analysis
The distribution of continuous variables is given as Mean, SD, Median, Quartile 1, Quartile 3, minimum and maximum.When creating reference values, 5, 10, 90 and 95% are also added.Distribution of categorical variables is given as numbers and percentages.To compare the two groups, Fisher's non-parametric permutation test is used for continuous variables, Fisher's exact test for dichotomous variables and Pearson's chi-square test for non-ordered categorical variables.Mean differences between the two groups regarding dichotomous variables and continuous variables are given with 95% confidence interval (CI).To analyse changes in continuous variables within groups, Fisher's non-parametric permutation test for paired observations is used.Mean difference within a group is given with 95% CI.All correlation analysis was performed using Spearman's correlations coefficient (r s ).Intra-individual SD is the most important measure regarding intra-observer agreement and is calculated as the square root of the mean values of intra-individual variances.All tests are two-tailed and conducted at 0.05 significance level.All analyses are performed using SAS v9.4 (Cary, NC, USA).

Ethical considerations
This study protocol was approved by the Local Committee of Ethics for Medical Research (No. 249-13).

Population characteristics
A total of 122 newborns were included.Dropouts were 17/122 (13.9%) at 12 months and the major reason was parent reported lack of time.At the 2-month follow-up, 109/117 (93.2%) of the infants had a TRU and at the 12-month follow up 91/105 (86.7%;Fig. 1; Table 2).
In total, 200 TRUs were done on 110 infants.Ninety infants had a TRU at both 2 and 12 months.Nineteen infants had only the 2-month TRU, and one infant had only the 12-month TRU.Of the TRUs, 197/200 (98.5%) were done by one nurse specialising in bladder/bowel ultrasound assessments and 3/200 (1.5%) by another nurse also specialising in bladder/bowel ultrasound assessments.Initially, every TRU was supervised by a paediatric radiologist.From this point, only data on infants with a TRU (n = 110) is presented.

Patient demographics
Of the infants, 62/110 (56.4%) were male, mean gestational age was 40.2 (SD 1.5) weeks and weight and length for age were normal (Table 1).The majority 92/97 (94.8%) passed meconium during the first 24 h.Infants passing meconium later were all healthy.The remaining 13 families had no recollection of the exact time for meconium passage.There were no differences in demographics between infants with or without FC.

Transverse rectal diameter
When analysing TRD in infants without FC anytime, the mean TRD at the 2-month TRU was 1.56 (SD 0.32) cm at a mean age of 2.24 (0.46) months, and at the 12-month TRU 1.78 (SD 0.47) cm at mean age of 11.91 (SD 0.48) months (Table 2, Fig. 2, Table S1).About 95% of infants had a TRD less than 2.26 cm at 2 months of age, and at 12 months less than 2.64 cm.In infants without an FC diagnosis anytime who had a TRU at both 2 and 12 months (n = 77), the TRD showed a mean increase from 2 to 12 months of 0.21 cm (95% CI: 0.099-0.318,P = 0.0062; Table 2, Table S1).
When analysing TRD in all infants, with or without FC, the mean TRD at the 2-month TRU was 1.56 (SD 0.33) cm at a mean age of 2.24 (SD 0.45) months, and at the 12-month TRU was 1.79 (0.49) cm at a mean (SD) age of 12.01 (0.75) months (Table 2, Fig. 2).There were similar results when analysing the infants with more than one ultrasound (n = 90).In all infants, with or without FC, who had a TRU at both 2 and 12 months (n = 90), the TRD showed a mean increase from 2 to 12 months of 0.22 cm (95% CI: 0.115-0.325,P = 0.0008; Table 2, Table S1).
No infant had a greater diameter than 3.00 cm at any age.When performing 2-month TRUs, there was no age difference between children with or without constipation.At 12-month TRUs, infants without FC at anytime were younger (P = 0.016)  with a mean difference (95% CI) of À0.64 (À1.03 to À0.11) months.At 2 and 12 months, TRUs revealed a low intra-individual standard deviation in the difference between the three measurements (Table 2).
At 2 and 12 months, there were no significant correlations between TRD and the following: time since last bowel movement, weight (Fig. 3), breast or formula feeding, need for stool softeners, need for assisted defecation, stool frequency and percentage of hard and very hard stools.
At 12-month TRUs, the mean TRD was lower in infants (n = 27) who had started potty training (1.74 (SD 0.73) cm) compared to infants who had not (1.83 (SD 0.52) cm), although the difference was not significant (P = 0.36).

Prevalence of FC and its relation to TRD
Thirteen infants were diagnosed with FC at a mean age of 5.9 (SD 3.9) months.At the 2-and 12-month TRUs, there were no TRD differences between infants with or without FC (Table 2, Fig. 2).Regarding the FC children, time from symptom onset to medication was mean 1.7 (SD 1.6) months in 10.All infants received treatment according to national guidelines.Two infants were diagnosed at 1 month and one infant at 1.5 months.All received oral lactulose and were medicated for approximately 1 month before the 2-month TRU.Another five infants were diagnosed after the 6-month questionnaire and the remaining five shortly afterwards.At the 12-month TRU, 5/13 (38.5%) infants had oral laxatives, two of them in combination with enemas, while 8/13 (61.5%) had no medication.Seven of the latter children were put on laxatives after diagnosis but could come off medication before the 12-month TRU.One infant was diagnosed just before the 12-month TRU and had dietary treatment when the TRU was performed.TRD mean (SD) in the non-medicated group was 1.75 (0.62) cm and in the medicated group was 2.03 (0.56) cm.There was no significant difference in TRD (P = 0.41) or in age for FC diagnosis (P = 0.21) between groups.The mean TRD at 2 months in infants with an FC diagnosis between 0-2 months (n = 3) and the mean TRD at 12 months in infants with an FC diagnosis between 2-12 months (n = 10), did not differ from infants without FC diagnosis anytime at the same ages (Table S2).
Infants with FC were offered and additional TRU at an outpatient clinic visit after FC diagnosis.Some declined, mainly due to lack of time or pandemic reasons.Figure S4 presents data on seven infants with FC who had additional TRUs at a mean age of 6.2 months.Three showed a decrease in TRD over time and the three were all off medication by 1 year of age (nos 1-3).The remaining infants showed an increase in TRD over time.At follow-up after the age of 1, they were all still on medication or had relapsed.

Feeding and bowel habits
The first solid feed was at a mean age of 4.4 (SD 0.6) months, while all had some solid feeds by 6 months.During the study, potty training started and could be dated in 27/110 (24.5%), at a mean (SD) age of 8.8 (2.6) months.No child stopped using nappies during the study.

Discussion
To our knowledge, this is the first prospective study presenting longitudinal data on TRDs in healthy infants, as measured by TRU.After infancy, several studies show that transrectal diameters correlate to age, 2,11,12 a finding in line with ours since we demonstrated that TRD increased slightly with age.We suggest that the upper limit for a normal TRD is 2.3 cm at 2 months of age and 2.6 cm at 12 months.
There are few other studies which measure TRD using ultrasound during the first year of life.However, two studies of infants have similar results to ours.They use air and barium enemas, respectively, in order to measure rectal diameter (median 1.27 cm and mean 2.24 cm). 22,23There are other studies including healthy children aged less than 3 years where TRD was measured by ultrasound [11][12][13] (Table S3).Dogan et al. 13 reported diameters (mean 1.68 cm) very similar to what we found at 12 months.An interesting observation of healthy children acting as controls in these studies was the limited increase in rectal diameter with age.Another interesting observation was that children with FC often had a larger TRD than controls, especially when a faecal mass was present.However, in line with our results, Dogan et al. showed no differences in TRD between healthy versus FC in the youngest children, that is, below the age of 3 years. 13In our cohort, there was no difference in TRD at baseline between infants without FC and those who later developed it.This may be explained by the short period between the onset of FC and treatment, but it speaks against FC being the result of a dilated rectum.However, our cohort, and the number of infants with FC, are both small.
Rectal diameters in our study vary only slightly during the first year of life despite changes from fluid to solid feeds, changes in stool consistency and in stool frequency. 20,24These intriguing findings emphasise the great compliance of the rectum, although much remains unknown.Our study found no correlation between TRD and weight, which was suggested in a study of older children, 12 nor between TRD and time since the last bowel movement as reported in Modin et al. 25 Modin et al. studied children older than 4 years of age and found that TRD decreased after defecation, then rapidly returning to the original size.We studied infants, who have a defecation frequency much greater than older children and with runny and soft stool consistency in a greater extent, which both could explain the lack of correlation between TRD and time since the last bowel movement. 19,26,27A finding that is important when implementing TRU in clinical practice.
This study is a part of an ongoing birth cohort study, designed with the primary aim of describing bowel habits in healthy children and finding the true prevalence of FC in the cohort during childhood.We recently published data that stated the prevalence of FC in the cohort during the first year of life to be 14.3%, 19 which is consistent with the literature with an estimated prevalence of 15%. 9one of the constipated infants in our study had a TRD larger than 3.00 cm, indicating that a cut-off value of 3.00 cm for diagnosing FC is not applicable before 1 year of age.In Pop et al., 12 half of the children under 4 years old with FC had a diameter > 3.00 cm.However, the children in our study were younger and the duration of FC prior to treatment was shorter.Pop et al. 12 could see a moderate correlation between larger diameters and long-standing constipation.When suggesting a cut-off value for TRD, one should consider symptom duration as well as age.
A few studies show that effective treatment for FC reduces TRD. 5 We saw a similar tendency in three of the seven FC patients who had an extra TRU at 6 months of age.These three patients were off medication at the age of one and showed a decrease in TRD, whereas four still needed medication and had increased TRD.In fact, Singh et al. 2 suggested that for clinicians rectal diameters, not medical history, should determine when to stop treatment.
Overall, ultrasound measurements of rectal diameters were of no benefit in diagnosing constipation in this study, since TRD did not differ between children with and without FC.One probable explanation is the short time between the symptom start of FC and treatment, along with the limited number of patients with FC.To evaluate the usefulness of TRD in children with FC in this age group, we need a larger study, recruiting infants who have had FC symptoms for longer.In this, we would use the present TRD results as normative values.Megarectum, seen in anorectal malformations, is another possible area of the use of TRU during infancy.In these cases, treatment is extensive bowel management and TRU may be used to follow treatment response by measuring TRD.
A study weakness is that one investigator performed most ultrasounds; therefore, we cannot calculate inter-observer variability.The intra-observer variability shows small variations as mentioned above.The investigator also had access to the patients' chart and could have had knowledge about the presence or absence on FC, which could influence the measurement.Other weaknesses are the small sample of children with FC, FC treatment given quickly and infants with FC regarded as constipated for the rest of the study period, which probably affects TRD measurements on infants with FC.Study strengths are the prospective design, questionnaires conducted as interviews, and suspected FC being confirmed by a doctor's examination.

Conclusion
We present TRD data in healthy infants at 2 and 12 months of age.There was a slight increase in diameter from 2 to 12 months.We suggest that the upper limit of normality for TRD should be 2.3 cm at 2 months of age and 2.6 cm at 12 months, and that the cut-off value of 3.00 cm should not be used to diagnose constipation under 1 year of age.No correlations could be found between TRD and weight, time since last bowel movement, defecation frequency or stool consistency during the first year.Infants of any age in this cohort who presented with FC did not have a greater TRD than infants who did not, although larger studies are needed to verify this.Ultrasound measurements of rectal diameters were of no benefit in diagnosing constipation in infants in this study.

Fig. 1
Fig. 1 Figure showing a study flow chart.n = participants at inclusion at 2 weeks of age and at 2, 6 and 12 months of age.n^= numbers (%) of completed ultrasounds.n^^= numbers (%) of completed questionnaires.n^^^= numbers (%) completing a follow-up visit at 6 months.

Fig. 2
Fig. 2 Figure showing transrectal diameters at 2-and 12-month ultrasound in infants with or without functional constipation at any time during the study (left) and in all infants (right).
90) n = 13 n = 13 n = 13 † is the 95% CI for mean change.Table showing transrectal diameters at 2-and 12-month ultrasound in all children, including children without functional constipation at any time during the study and children diagnosed with functional constipation during the study.For continuous variables, mean (SD)/(P5; P95)/n = is presented.II-SD means Intra Individual SD, expressed in standard deviation of the mean difference between the three measurements.

Fig. 3
Fig. 3 Figure showing transrectal diameters and Spearman correlations between hours since last bowel movement and weight at 2-and 12-month ultrasounds.

Table 1
Participant demographics