Assisted infant toilet training and the prevalence of functional gastrointestinal disorders up to the age of 9 months: a randomised, controlled trial

Abstract

Background

It is suggested that late initiation of toilet training contributes to functional gastrointestinal disorders in children. This study aims to evaluate if assisted infant toilet training can reduce the prevalence of functional gastrointestinal disorders during the first 9 months of life.

Methods

The ongoing Bowel And Bladder function in Infant Toilet Training (BABITT) study is a randomised, controlled trial that recruited infants aged 0–2 months at six Child Health Centres in Sweden. Infants were randomly assigned (1:1) to either start parent-assisted infant toilet training at the age of 0–2 months (intervention) or as controls. The intervention group was encouraged to practise at a moderate frequency of 1–3 times/day, 5–7 days/week. A validated web questionnaire was answered at ages 2, 3, 6 and 9 months.

Results

271 infants (145 males, 54%) were randomly assigned at median age of 34 days (136 intervention; 135 controls). The intention-to-treat analysis with complete cases (127 intervention; 132 controls) showed no difference in the total period prevalence of functional gastrointestinal disorders (infant colic, infant dyschezia and/or functional constipation, defined by Rome IV criteria) between the intervention group and controls (52.0% vs 49.6%, difference 2.4% (95% CI −9.8 to 14.7); p=0.6956) up to the age of 9 months. Parent-reported intervention adherence ranged between 53% and 63%.

Conclusions

This study does not support that assisted infant toilet training, practised at a moderate frequency, reduces the prevalence of gastrointestinal disorders during infancy. Long-term effects are being evaluated in the ongoing BABITT study up to the age of 4 years.

Trial registration number

NCT04082689.

WHAT IS ALREADY KNOWN ON THIS TOPIC

• It has been suggested that late initiation of toilet training contributes to functional gastrointestinal disorders in children. We found no previous studies evaluating assistant infant toilet training and the prevalence of functional gastrointestinal disorders.

WHAT THIS STUDY ADDS

• This randomised, controlled trial found no difference in the prevalence of functional gastrointestinal disorders (infant colic, infant dyschezia and/or functional constipation) up to the age of 9 months when a moderate frequency of assisted infant toilet training was introduced at 0–2 months of age. Parental stress and mother-to-infant attachment did not differ between groups. The intervention group used significantly fewer disposable nappies.

HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY

• More research is warranted, both on how to support parents practising assisted infant toilet training as well as introducing and completing toilet training, with different ages, settings and interventions, and on the potential effect on gastrointestinal disorders. We need evidence-based guidance for parents, child healthcare, day care services and societies.

Introduction

Since the introduction of disposable nappies, the age for onset and completion of toilet training has increased.^{1 2} During the 1970s in Sweden, children used the potty on a daily basis before the age of 1 year and completed toilet training after 2 years of age.¹ Research in the field is scarce and the latest published study from 2005 concludes that Swedish children completed toilet training at a median of 3.5 years.³ As a comparison in a global perspective, Vietnamese mothers start toilet training in early infancy by tradition.⁴ This practice is called ‘elimination communication’ or ‘assisted infant toilet training’.⁵ The Vietnamese children reached parent-assisted dryness at the age of 9 months.⁴ This implies that the former perspective of awaiting a child’s ‘readiness’ to toilet train is not dependent on a child’s physical and mental maturation, but a matter of training and communication. There is no consensus on how and when to initiate toilet training, studies are heterogeneous and depend on cultural traditions.⁶ Assistant infant toilet training has no known negative consequences reported in the literature⁷ but to further evaluate positive or negative effects, we included validated questionnaires on parental stress⁸ and attachment.⁹

Meanwhile, studies imply that postponed toilet training contributes to problems related to bowel and bladder control.^{2 10} Functional gastrointestinal disorders are common during childhood, cause suffering for the affected child and family and consume healthcare resources.^{11 12} In Sweden, prescription of drugs for functional constipation has increased more than sixfold during the last 18 years.¹³ Evaluating changes in prevalence of functional gastrointestinal disorders over time is challenging, since definitions are not uniform and have continuously been revised. There are many pathophysiological factors contributing to functional constipation. ¹⁴ Also parent’s prerequisites regarding parental leave, day care services and availability of disposable nappies differ between countries. We defined functional gastrointestinal disorders (infant colic, infant dyschezia and functional constipation) according to the Rome IV criteria for infants and toddlers.¹⁵

Infant colic has been shown to be associated with negative long-term effects on attachment as well as contributing to postpartum depression.¹⁶ For infant colic and infant dyschezia, evidence-based treatment recommendations are limited. In contrast, functional constipation has well-established treatment recommendations during childhood,¹⁷ but relapses are common and quality of life is negatively affected.¹²

A consequence of toilet training at a later age is a prolonged period of using disposable nappies. Disposable nappies are made out of plastic that cannot be recycled,¹⁸ the waste management is challenging and the cost can cause financial stress for families.

In adults, the squat position has been shown to be beneficial for defecation.¹⁹ This is explained by the anatomy of the anal canal; thus, the squat position straightens the anorectal angle, facilitating defecation. When practising assisted infant toilet training, the child is positioned in a squat position over a potty or basin for urinating and defecating. We therefore hypothesised that by facilitating defecation, assisted infant toilet training could reduce symptoms of infant colic, infant dyschezia and functional constipation.

In our ongoing randomised controlled trial (Bowel And Bladder function in Infant Toilet Training, BABITT), our overall aim is to evaluate whether assisted infant toilet training in healthy infants can reduce the prevalence of functional gastrointestinal and bladder disorders up to the age of 4 years. In this first part of the ongoing BABITT study, the aim is to evaluate if assisted infant toilet training can reduce the prevalence of infant colic, infant dyschezia and functional constipation up to the age of 9 months.

Methods

Study design and setting

The ongoing BABITT study was designed as an open-label, randomised, superiority and controlled trial with a 4-year follow-up. The infants were recruited from six Child Health Centres in the central part of Sweden. The BABITT study protocol is published²⁰ and the full clinical study protocol, adhering to the Standard Protocol Items: Recommendations for Interventional Trials 2013 statement,²¹ is available at ClinicalTrials.gov (NCT04082689). The study adheres to the Consolidated Standards of Reporting Trials guidelines.²²

Participants

Study participants were healthy full-term infants recruited before the age of 2 months and 2 weeks. Exclusion criteria were malformations or disorders that may affect the gastrointestinal or urinary tract, infants born small for gestational age (<2 SD birth weight), premature (<gestational week 37+0) or parent/caregiver with insufficient understanding of the Swedish language.

Randomisation

After completion of the baseline questionnaire, the investigators (TN or AL) randomly assigned infants with a 1:1 allocation ratio to either assisted infant toilet training or control group. The computer-generated sequence used stratified block randomisation with permuted blocks by the infant’s sex.

Intervention

The intervention comprises parents learning and practising assisted infant toilet training. Parents in the intervention group received oral instructions on how to conduct assisted infant toilet training and information about potential health benefits in a face-to-face video meeting with one of the two investigators (TN or AL). They also received a book on the topic,²³ a brief summary of the book in a brochure made by the investigators (TN and AL) and a potty. Parents in the intervention group were instructed to start directly after allocation at age 0–2 months. Furthermore, on the routine visits to the Child Health Centre, parents had the possibility of receiving support from child health nurses.

Parents in the control group were instructed not to start assisted infant toilet training until after 9 months of age. Thereafter, they received equivalent information and material about assisted infant toilet training and were instructed to start practising.

Adherence to the intervention of assisted infant toilet training was defined by parent-reported practising at least once daily (without the requirement of a subsequent urination or defecation) on at least 5 out of 7 days per week.

Outcomes

The primary outcome was the total period prevalence of parent-reported functional gastrointestinal disorders (infant colic, infant dyschezia and/or functional constipation), defined by Rome IV criteria.¹⁵ Parents answered validated web questionnaires at 2, 3, 6 and 9 months. If any of the three functional gastrointestinal disorders were reported at any of these four time points, it was considered as a primary outcome.

Secondary outcomes included parent-reported mother-to-infant attachment, assessed by the Maternal Postnatal Attachment Scale (MPAS)⁹ and parental stress assessed by the Swedish Parenthood Stress Questionnaire (SPSQ)⁸ at 3 and 9 months.²⁰

The web questionnaires used in the study were evaluated in a content validation procedure.²⁴

Statistical analysis

Sample size was calculated based on the expected reduction of the prevalence of the primary outcome (total period prevalence of functional gastrointestinal disorders; infant colic, infant dyschezia and/or functional constipation) up to the age of 9 months. Prevalence reduction was estimated from 15%^{15 25 26} to 4%, with a two-tailed significance level of 0.05 and 80% power. From this calculation (including an estimated 10% missing), a total of 268 children (134 in each group) were needed in the study.

Complete cases were analysed according to the intention to treat (ITT) principle with participants coded to their allocation status and the per protocol (PP) principle. In the primary analyses, the total period prevalence of functional gastrointestinal disorders (infant colic, infant dyschezia and/or functional constipation) up to the age of 9 months was analysed with logistic regression in a model that included terms for the intervention group (assisted infant toilet training vs control group) and the child’s sex (R V.4.2.3). Based on the model, we estimated the risk of outcome in the intervention group and controls and the risk differences with 95% CIs between groups (emmeans V.1.8.5 in R V.4.2.3). In addition, we conducted a best worst-case analysis to assess the impact of data attrition. In the secondary analyses, MPAS and SPSQ summary scores, as well as the use of laxatives and disposable nappies, were compared between the intervention group and controls with the Mann-Whitney U test (R V.4.2.3).

Results

Between 1 April 2020 and 24 September 2021, all 1041 newborns at the six Child Health Centres were assessed for eligibility, and a total of 271 healthy infants were included. Overall, 136 infants were randomly assigned to the intervention group and 135 infants to the control group (figure 1). In total, 127 children in the intervention group and 132 children in the control group were included in the primary outcome ITT analysis. Infants were included at a median age of 34 days (range 9–71) and 145 were males (54%). Baseline characteristics of the study population are shown in table 1.

Figure 1. Trial profile.

Table 1. Baseline characteristics of children in the BABITT study (n=271).

	Assisted Infant Toilet Training (n=136)	Controls (n=135)
Sex
Male	73 (54%)	72 (53%)
Female	63 (46%)	63 (47%)
Age at inclusion, days, median (range)	34 (9–71)	34 (9–70)
Weight at birth, grams, mean (±SD)	3618 (±501)	3672 (±413)
Social situation
Caregivers married/living together	132 (97%)	132 (98%)
Caregiver with sole custody	4 (3%)	3 (2%)
Siblings
Yes	68 (50%)	83 (62%)
Age of caregiver at inclusion
First caregiver, mother, years, median (range)	31 (21–47)	32 (22–45)
Second caregiver, partner, years, median (range)	33 (21–56)	33 (21–55)
Caregivers born in Sweden
First caregiver, mother	124 (91%)	124 (92%)
Second caregiver, partner	116 (87%)	115 (87%)
Highest level of education
First caregiver, mother
Upper secondary school or less	31 (23%)	31 (23%)
Higher education or postgraduate	105 (77%)	104 (77%)
Second caregiver, partner
Upper secondary school or less	51 (38%)	57 (43%)
Higher education or postgraduate	83 (62%)	75 (57%)
Caregivers’ history of constipation*
Yes	38 (28%)	22 (16%)
Exclusive breastfeeding†
Yes	116 (87%)	117 (88%)
Infant colic at inclusion
Yes	17 (13%)	15 (11%)

Data shown as n (%) = number of cases and per cent unless otherwise specified.

^*A history of seeking medical care for functional constipation by one or two parents.

^†Exclusive breastfeeding reported by caregiver in the 2-month web-questionnaire (n=266).

BABITT, Bowel And Bladder function in Infant Toilet Training.

There was no difference in the parent-reported total period prevalence of functional gastrointestinal disorders (infant colic, infant dyschezia and/or functional constipation, defined by Rome IV criteria) when comparing the intervention group and the control group (52.0% vs 49.6%, difference 2.4% (95% CI −9.8 to 14.7); p=0.6956) up to the age of 9 months (table 2, figure 2). Results remained consistent in the best worst-case analysis of the primary outcome.

Table 2. Parent-reported prevalences of functional gastrointestinal disorders up to the age of 9 months.

	Assisted infant toilet training (n=136)		Controls (n=135)		Difference % (95% CI)	P value
	n	%	n	%
Infant colic
2 months	6/132	4.50	5/132	3.80
3 months	1/127	0.80	3/133	2.30
Total period prevalence of infant colic*		4.80		3.80	1.0% (−4.0 to 6.0)	0.6995
Infant dyschezia
2 months	32/132	24.20	32/132	24.20
3 months	23/127	18.10	26/133	19.50
6 months	13/119	10.90	17/134	12.70
9 months	10/126	7.90	9/134	6.70
Total period prevalence of infant dyschezia*		37.40		39.00	−1.6% (−13.6 to 10.4)	0.7926
Functional constipation
2 months	3/132	2.20	2/132	1.50
3 months	8/127	5.90	6/133	4.40
6 months	11/119	8.10	8/134	5.90
9 months	21/126	15.40	18/134	13.30
Total period prevalence of functional constipation*		28.60		17.70	10.9% (0.4 to 21.3)	0.041
Total period prevalence of the primary outcome†		52.00		49.60	2.4% (−9.8 to 14.7)	0.6956

n=number of participants with parent-reported primary outcome followed by the total number of complete cases.

Total period prevalence data are analysed according to the intention to treat principle.

^*Estimates based on logistic regression.

^†Parent-reported functional gastrointestinal disorders (infant colic and/or infant dyschezia and/or functional constipation, according to ROME IV criteria) up to the age of 9 months.

Figure 2. Functional gastrointestinal disorders (primary outcome) shown as total period prevalence up to the age of 9 months in the Assisted Infant Toilet Training group (n=127) and controls (n=132) (A). Period prevalence of Infant Colic (B), Infant Dyschezia (C) and Functional Constipation (D).

Concerning infant colic, the total period prevalence did not differ between the intervention group and the control group (4.8% vs 3.8%, difference 1.0% (95% CI −4.0 to 6.0); p=0.6995) at 2 and 3 months of age (table 2). Already at inclusion, at the median age of 34 days, the distribution of prevalent infant colic was equal between groups (17 out of 136 (13%) of the intervention group and 15 out of 135 (11%) of the controls) (table 1).

The total period prevalence of infant dyschezia was similar between the intervention group and the control group (37.4% vs 39.0%, difference −1.6% (95% CI −13.6 to 10.4); p=0.7926) at 2, 3, 6 and 9 months of age (table 2).

The prevalence of functional constipation increased from 2 months to 9 months of age in both the intervention group (2.2% to 15.4%) and in the control group (1.5% to 13.3%) (table 2). However, up to the age of 9 months, the total period prevalence of functional constipation was slightly higher in the intervention group compared with the control group (28.6% vs 17.7%, difference 10.9% (95% CI 0.4 to 21.3); p=0.0410) (table 2). The use of laxatives showed no differences between groups (data not shown).

Parent-reported Mother-to-infant attachment (MPAS) and Parental stress (SPSQ) at ages 3 and 9 months showed no difference between the intervention group and the control group, as total scores, nor as scores in the subscales (online supplemental web appendix table 1).

Parent-reported intervention to adherence ranged between 53% and 63% up to the age of 9 months (online supplemental web appendix table 2) and the majority (49%–78%) practised assistant infant toilet training 1–3 times daily (online supplemental web appendix table 3). We did not observe that the frequency of practised assisted infant toilet training resulted in any meaningful differences on the prevalence of functional gastrointestinal disorders.

In the control group, 22 out of 132 (17%) did not follow their allocation (ie, they started practising assisted toilet training prematurely, before 9 months of age) (figure 1). Reasons given for starting ahead of time were to facilitate defecation and/or to ease flatulence of the child (n=10), constipation symptoms at food introduction (n=6), parents ‘felt ready’ to start (n=5), or the child having discomfort at urination or defecation (n=1).

PP analysis showed no difference between the groups on the total period prevalence of functional gastrointestinal disorders up to age of 9 months (p=0.4302), nor on secondary outcomes (MPAS, SPSQ) (online supplemental material tables 1 and 2, online supplemental material figure 2).

The number of used disposable nappies per day was significantly lower in the intervention group compared with the controls at 6 months of age (n=37 intervention vs n=42 controls, p=0.0412) and at 9 months of age (n=61 intervention vs n=66 controls, p<0.0001) (figure 3).

Figure 3. Parent-reported number of used disposable nappies per day in the Assisted Infant Toilet Training group and in the controls at age of 6 months (n=37 intervention vs n=42 controls) and 9 months age (n=61 intervention vs n=66 controls).

Discussion

This first part of the ongoing BABITT study does not support that assisted infant toilet training, practiced a moderate frequency, starting at the age of 0–2 months reduces the prevalence of functional gastrointestinal disorders up to the age of 9 months. To our knowledge, this is the first randomised, controlled trial evaluating the effect of assisted infant toilet training on functional gastrointestinal disorders, and therefore the study adds new important information in the research field.

We found a surprisingly high overall prevalence of functional gastrointestinal disorders in both groups, as compared with similar infant populations.¹⁵ This could reflect a truly high prevalence or might be explained by a surveillance bias. Notably, our present study evaluated a combined primary outcome with three different functional gastrointestinal disorders (infant colic, infant dyschezia and/or functional constipation).

Concerning infant colic, our study showed a low prevalence as compared with other studies.²⁵ This is likely explained by the fact that the 2-month and 3-month questionnaires were presented to the parents at a late stage in the natural course of infant colic, that is, spontaneously declining after 8–9 weeks of age.²⁵ An earlier inclusion of infants in the BABITT study would have been preferable but did not seem feasible. In addition, the low prevalence of infant colic in our study might have been a result of selection bias. However, randomisation at inclusion gave an equal distribution of already prevalent infant colic between the intervention group and the control group at baseline (table 1).

When diagnosing functional constipation according to The Rome IV criteria, the third criteria ‘history of painful or hard bowel movements’ is troublesome when translated into Swedish. As part of the validation progress of our web questionnaires,²⁴ it came to our notice that informants required aid in the definition of painful or hard stools. Therefore, we chose to visualise ‘hard bowel movements’ with the Bristol stool scale²⁷ in our questionnaire. In accordance with previous publications, Bristol stool scale type 1–2 was chosen to define hard stools.²⁸ When infants defecate in a potty instead of a nappy, parents in the intervention group might have different prerequisites determining defecation discomfort and stool consistency compared with parents in the control group. Thus, functional constipation in the two groups may have been at risk of differential misclassification, which might have influenced results in our study. Admittedly, this could be an explanation of the higher prevalence of parent-reported functional constipation in the intervention group in our study. More studies in other study settings and with different study populations are needed.

Assisted infant toilet training has been shown to cause a more regular defecation pattern in infants.²⁹ Since the squat position is of importance for facilitating passage of stools, it is possible that a higher frequency of assisted infant toilet training could influence the prevalence of functional gastrointestinal disorders. In our present study, the intervention adherence up to the age of 9 months was rather low (53%–63%). However, per-protocol analyses did not change our results (online supplemental materials). Furthermore, a hypothesis-generating subgroup analysis of parents practising assisted infant toilet training more frequently (≥4 times per day vs 1–3 times per day) (online supplemental web appendix table 3) did not reduce the prevalence of functional gastrointestinal disorders.

Assisted infant toilet training has previously been associated with an earlier age of assisted dryness, especially if using cloth nappies, and more time spent nappy free.^{29 30} In our study, we decided not to include these requirements in the intervention because of concerns of recruitment difficulties or low acceptability of the intervention. Furthermore, the COVID pandemic was an unforeseen complicating reality during the study period. Social supportive meetings for participating parents were planned as part of the intervention but had to be cancelled due to the pandemic restrictions. Consequently, the only support on demand that could be offered to parents was telephone support from the investigators (TN or AL). Future studies on assisted infant toilet training might benefit from providing more support to parents with reminding, habit-stacking, nudging and enhancing self-efficacy in coached parent education as well as maintenance of practice.

The use of disposable nappies, with super absorbency and surface barriers, protects the child from the wet sensation of urine and faeces.¹⁸ Nevertheless, disposable nappies are an ecological challenge since they are fossil fuel products. In addition, the use of disposable nappies seems to increase the age at which toilet training is completed and may be associated with bladder dysfunction symptoms.^{2 31} Surprisingly, despite the low frequency of assisted infant toilet training and the rather low intervention adherence in our study, the intervention group significantly reduced the number of disposable nappies used at the age of 6 and 9 months (figure 3). At the same time, neither parental stress nor mother-to-infant attachment differed between groups (online supplemental web appendix table 1) and there were no adverse events or harmful side effects reported. Considering this, advocating assisted infant toilet training seems to be a safe practice in order to reduce the number of used disposable nappies, ameliorate waste management, lower the overall carbon footprint and reduce family expenses.

One strength of this study is the design; a randomised, controlled trial, reducing both known and unknown confounding factors that might influence results. Furthermore, the study used validated questionnaires²⁴ for parent-reported outcomes, that is, functional gastrointestinal disorders based on the strict criteria set by the leading scientific consensus panel of the Rome Foundation.¹⁵

As in all randomised controlled trials, a limitation of our study was the selection of families recruited and enrolled, which might have influenced the generalisability of the results.

Conclusion

This randomised controlled trial does not support that assisted infant toilet training, practised at a moderate frequency, reduces the prevalence of gastrointestinal disorders up to the age of 9 months. Parental stress did not differ between groups. The intervention group significantly reduced the use of disposable nappies, an important finding in the context of sustainability as well as the perspective of family economics. Long-term evaluation of bowel and bladder disorders will be further analysed in the ongoing BABITT study. More research is warranted on how to support parents practising assisted infant toilet training as well as introducing and completing toilet training. Studies are also needed to evaluate different ages, settings, frequencies and interventions, and on the potential effect on gastrointestinal disorders. We need evidence-based guidance to parents, child healthcare, day care services and societies.

Data availability statement

Data are available upon reasonable request.