Constipation describes symptoms such as hard stools, overstraining, infrequent bowel movements, digital manoeuvres, abdominal bloating and an incomplete evacuation (Rao et al, 2016). It is a frequent condition affecting about 20% of the population; incidence increases with age (Vazquez Roque and Bouras, 2015) causing discomfort, increased healthcare costs and loss of appetite (Çevik et al, 2018), an important symptom in older persons that can lead to impaired nutritional status (Bonetti et al, 2017; Bollo et al, 2019). Aetiology includes primary and secondary factors, the first being alterations of bowel function, which cause prolonged stool transit. Secondary factors include inadequate fluid and fibre intake, sedentariness, neurological conditions, neoplasms, and the use of medications such as opioids, antidepressants, and antipsychotics (Yurtdaş et al, 2020; Salari et al, 2023). According to the Rome IV Criteria (Lacy and Patel, 2017), functional constipation is the simultaneous presence of two or more factors among those listed in Box 1. In addition, these factors must have arisen at least 6 months before diagnosis (Simren et al, 2017; Russo et al, 2019).
Box 1.Rome IV Criteria for functional constipation
- Straining in at least 25% of bowel movements
- Hard or lumpy stools in more than 25% of bowel movements
- Sensation of incomplete evacuation in more than 25% of bowel movements
- Sensation of anorectal blockage in more than 25% of bowel movements
- Manual manoeuvres to facilitate more than 25% of bowel movements (eg, manual voiding)
- Fewer than three spontaneous bowel movements per week
Source: Lacy and Patel, 2017
Treatment of constipation involves pharmacological intervention with osmotic laxatives, which increase stool volume, as well as purgatives and stimulants. The recommended treatment, according to those consensus guidelines, is osmotic laxatives, specifically polyethylene glycol or lactulose; evidence level is 1A (Emmanuel et al, 2017) (polyethylene glycol is usually prescribed as ‘macrogol’ in the UK). However, this treatment leads to the onset of complications in the long term, including abdominal bloating and cramping, flatulence, irritation of the anal mucosa and dehydration (Çevik et al, 2018). Therefore, pharmacological treatments are usually accompanied by the indication to practice physical activity and increase fibre and water intake.
Bowel training is also commonly suggested in the literature. It consists of principles to follow for promoting correct evacuation: not ignoring the urge to defecate, avoiding staying in the bathroom for more than 10 minutes, and taking the right posture with a step for the feet while sitting on the toilet to straighten the anorectal angle (Takano and Sands, 2016; Çevik et al, 2018). Many studies have pointed out the importance of the anorectal angle and studied three different positions on the toilet: sitting with knees bent higher than the hips (semi-squatting) and squatting. The squatting and semi-squatting positions straighten the anorectal angle, thus allowing easier defecation. The sitting position, on the contrary, narrows the anorectal angle with the consequent difficulty of evacuation Sakakibara et al, 2010; Gregersen et al, 2022).
Several complementary interventions to pharmacologic treatments have been proposed over time (eg biofeedback therapy for dyssynergic defecation and sacral neuromodulation therapy for the levator ani syndrome (Sharma et al, 2022), aimed at reducing the side effects of drugs while respecting the scientific guidelines. In addition, several authors have proposed abdominal massage to stimulate the neurological reflexes of the bowel and improve motility. Historical sources point out that abdominal massage was an integral part of the nursing curriculum taught by Florence Nightingale in 1860 (Ruffin, 2011) and is still considered a nursing practice today (Yurtdaş et al, 2020). However, notwithstanding the number of studies on this topic, many conflicting indications exist on the techniques required for optimal results. Research is therefore needed to define a massage protocol compliant with current evidence. It is also necessary to verify if massage combined with laxatives can be more effective and have fewer side effects than laxatives alone. Finally, since the literature recommends bowel training for teaching correct bowel habits, it should be included in the treatment protocol. Therefore, the authors hypothesised that abdominal massage and bowel training combined with polyethylene glycol (PEG) could increase the frequency of evacuations, appetite, and food intake, compared with PEG alone. Considering the time and work required to apply this intervention, the team also sought to investigate the possibility of incorporating massage into nursing practice.
Design and method
Design
A pilot randomised controlled trial with parallel arms was conducted. A pilot study ‘asks whether something can be done, should the researchers proceed with it, and if so, how’–verifying the feasibility and acceptability of the intervention are among its purposes (In, 2017).
Sample
Over 4 weeks, the team enrolled a sample of older adults living in a facility for self-sufficient persons in Milan (healthcare residence). These residents live permanently in the facility, allowing them to receive basic care (eg, room cleaning and meals), participate in social activities organised by occupational therapists, attend church, spend time with their relatives, and so on. All 33 people living in the healthcare residence had constipation according to the Rome IV Criteria; all 33 were enrolled and randomised to receive abdominal massage plus PEG (intervention group) or the laxative alone (control group). The allocation ratio was 1:2, and block randomisation was used to assign patients to the groups. The 1:2 ratio was chosen because only one person was dedicated to assessing all patients and performing massages; for this reason, it was necessary to limit the number of patients in the intervention group, whereas the subjects of the control group could be more numerous as they did not undergo massage. In addition, all patients had a Mini-Mental State Examination (MMSE) (Stein et al, 2012) score higher than 25, excluding cognitive impairment, which was also an exclusion criterion for living in the facility where the study was conducted.
Procedure
A thorough literature review was conducted on PubMed, CINAHL, Cochrane Library and Google Scholar to retrieve articles published in the last 10 years. Boolean operators and related articles suggested by the database were used for this search. The keywords used to build the search strings were ‘abdominal massage’, ‘functional constipation’, and ‘nurs*’. There was conflicting evidence about timing and massage techniques. Only Celik and Bilik (2022) highlighted the importance of the environment in which the massage is given. Some authors (McClurg et al, 2018;Çevik et al, 2018) suggest a supine position, while others (Turan and Atabek, 2016) adopt the ‘semi Fowler’ position – supine, with the head of the bed raised 30–45° (legs can be straight or knees bent). Others (Ashton and Cassel, 2020) recommend using an unspecified oil to avoid friction of the hands on the abdomen during the massage. As regards massage techniques, some authors had not explained the details of the procedure; therefore, the researchers contacted them but have yet to receive a reply. To prepare the massage protocol, the team focused on the techniques the retrieved papers had in common: pressure, effleurage, petrissage, percussion, and vibration (Table 1).
Table 1. Detail of the techniques contained in the abdominal massage protocol
| Pressure Vigorous grazing performed with fingertips | |
| Effleurage Extensive skin grazing by hand | |
| Petrissage Gently pinch the skin between the thumb and the other fingers | |
| Vibration Rapid shaking performed with fingertips |
Indications about the time of the day when the massage should be performed were provided by one paper only (Celik and Bilik, 2022). Massage length varied from 10 minutes (McClurg et al, 2018) to 15 minutes (Birimoglu Okuyan and Bilgili, 2019) to as long as 45 minutes (Çevik et al, 2018). Following the literature review, the team created the protocol shown in Box 2.
Box 2.Abdominal massage protocol
- Prepare the room where the abdominal massage will be performed: if the patient is alone, the massage can be performed in his/her own room. In case the patient is in a double room or shared with other patients, it is advisable to perform the massage in a dedicated room according to the possibilities of the facility and the mobility of the patient. Attention should be paid to the microclimate, as the patient should remain with the abdomen uncovered for the whole massage duration
- Have the patient lie on the bed in the supine position and uncover the abdomen only
- Perform an objective abdomen examination to assess for abdominal pain, skin integrity, and abdominal tension; assess for flexure encumbrance
- Remove jewellery. Warm the hands by rubbing them together; place a small amount of massage oil on the cupped hands; rub them together to warm the oil and place them on the patient's abdomen for a few moments
- Apply gentle pressure with both hands running from the epigastric region to the hypogastric region, asking the patient if nausea occurs
- Perform clockwise effleurage for 2 minutes (=10 times), starting with the ascending colon and continuing with the transverse and descending colon
- Perform clockwise petrissage for 2 minutes (=10 times), starting with the ascending colon and continuing with the transverse and descending colon
- Perform pressure on the descending colon from top to bottom for 2 minutes (=20 times)
- Repeat the previous three steps (effleurage, petrissage and pressure)
- Perform vibrations on the abdominal wall for 1 minute (=10 vibrations) clockwise following the anatomy of the colon and ending with the mesogastric region; at this stage, pause on the flexures in case of faecal obstruction
Overall duration: one session of 15 minutes for 5 days a week for 4 weeks
In this study, the treatment was performed by a registered nurse and a student nurse in the final year of the bachelor's degree course, with formal education in massage and treatment of bowel dysfunction, under the supervision of the registered nurse. Feasibility was studied in terms of time required for performing patient assessment and abdominal massage, as time is a crucial variable in the complexity of everyday nursing practice. In addition, the team prepared a public YouTube channel and a video demonstrating the techniques as a resource for other nurses who might be interested in the future development of this project. The video (https://youtu.be/qXaaH5p8VDM) is currently in Italian, but subtitles are available with translation in other languages; to make access easier, a QR code was created for use with a smartphone.
All patients in the sample were instructed on bowel training according to the literature (Sakakibara et al, 2010; Baird et al, 2019) before beginning the treatment. Such training included the following principles: do not ignore the urge to defecate; avoid sitting in the bathroom for more than 5 to 10 minutes; use the correct posture, which involves using a step or a platform to rest one's feet to straighten the anorectal angle. Numerous studies deal with the anorectal angle, which shows that three positions are known for defecation: sitting, sitting with the knees bent higher than the hips (semi-squatting), and squatting. The best positions are squatting, which is preferable in terms of effectiveness, and semi-squatting, as they cause stretching and straightening of the anorectal angle, which thus allows effective defecation with less effort than the sitting position, which causes a constriction of the anorectal angle with consequent difficulty in evacuation (Sakakibara et al, 2010). All treatments took place in the individual rooms of the patients.
Measures
The Bristol Stool Form Scale (Lewis and Heaton, 1997)(Figure 1) was administered prior to beginning any intervention (Assmann et al, 2000) and prior to any subsequent massage session; data regarding bowel habits, most recent bowel movement, enemas, laxatives, medications influencing bowel function, digital rectal voiding, appetite, and food intake were collected with the same criteria. Baseline data included age and medical history. The treatment was considered successful in the presence of a statistically significant increase in Bristol scores and the number of weekly defecations.
The authors chose to focus on patients' satisfaction with the protocol of abdominal massage rather than health-related quality of life. Therefore, a Likert scale was used, ranging from 0 (‘I do not like it and it was not useful’) to 10 (‘I like it and I think it is beneficial’) which was proposed and explained to all the participants in the intervention group.
Data analysis
If normally distributed, data were described as mean and standard deviation, or median and IQR (interquartile range) otherwise. Normality was checked by using Kolmogorov-Smirnov, Anderson-Darling and Shapiro-Wilk tests. Mixed effects linear models were used to assess the treatment protocol's efficacy after assumption checking (normality and homoscedasticity). There was no need for multivariate analysis, as no statistically significant differences were detected between the groups regarding baseline characteristics representing potential biases. The models were retained with an R-squared value ≥0.80; residuals were analysed for all models. The effect size was calculated as w2 (omega-squared) due to the small sample size, with values above 0.25 indicating a ‘large’ effect of massage on the dependent variable (Cohen, 1992). Statistical significance was defined as P<0.05. All analyses were performed with SAS 9 (SAS Inc, Cary, NY).
The study was approved by the ethical committee of the University of Milan (no. 4519) and by the nursing management of the facility where the trial was conducted. The participants gave explicit consent to participate after being informed about the study methods and purposes of the data collection and that there was no detriment in case of withdrawal. Informed consent was obtained from all patients. The principles of the Declaration of Helsinki were respected. Data management was compliant with the Italian law in force.
Results
All 33 people living in the facility gave consent to participate and were randomised to the intervention (n=11) or control group (n=22). A drop-out of a patient in the intervention group is reported: the day after signing the consent form, he refused to start the sessions without giving reasons. Therefore, the number of patients analysed in the intervention group was 10. No intention-to-treat analysis was possible, because this person also refused to undergo any further data collection. No other drop-outs were recorded in either group during the study. The patients were mostly female (n=24), and the median age of the sample was 87 years, IQR=[80;93] without significant differences between genders (P=0.20) and study groups (P=0.34) (Table 2).
Table 2. Sample characteristics
| Female gender n (%) | Male gender n (%) | Age median [IQR] | ||
|---|---|---|---|---|
| Group | Intervention (n=10) | 6 (60.0) | 4 (40.0) | 86 [80;89] |
| Control (n=22) | 18 (81.82) | 4 (18.18) | 87 [81;93] | |
| Overall sample | 24 (75.0%) | 8 (25.0%) | 87 [80;93] | |
IQR=interquartile range
Baseline characteristics
At baseline, four participants had two weekly defecations, 13 had three, 14 patients had four; only 1 reported five. Notwithstanding their high number of weekly bowel movements, those who reported three or more all declared feeling of incomplete emptying, hard or lumpy stools, or straining. Thirty participants reported ‘medium’ or ‘high’ discomfort during defecation. All participants but one had moderate defecatory urgency; the remaining one used to void his rectum manually and had severe urgency. The preferred defecatory position was related to postural limitations: 22 patients evacuated in the supine position with a diaper, 8 adopted the sitting position on the toilet, and 2 used both positions described above. Twenty-eight patients scored 2 or 3 on the Bristol stool chart, but the frequency of their evacuations was still compliant with the Rome criteria; the remaining four had type 1 stools.
The number of monthly low-volume enemas at baseline ranged from 0 (n=11) to 6 (n=3). Twenty-two people were taking PEG on medical prescription; 10 used no medications but received low-volume enemas every third day without spontaneous defecation, on medical prescription. Twenty-four people used to eat half of the food provided by the facility, or even less; appetite was small or medium in 29 participants (only three reported ‘big appetite’).
Apart from gender, which was then controlled in statistical analysis, baseline characteristics were comparable in the two groups (P>0.05 for all variables). In addition, the baseline data of perceived distress related to defecatory difficulties (low, medium, high) did not differ significantly between the two groups (P=0.28). The median dosage of PEG in the sample was 125 mg/day, IQR[125-218.75] with no statistically significant differences between the two groups (P=0.32) and most patients (9 in the intervention group and 21 in the control group) taking 125 mg/day. All patients taking PEG had already received their prescription before beginning the study, therefore excluding a possible bias regarding the knowledge of patients by the physician. Enemas had also been prescribed for all patients by the physician before the study as part of the medication regimen of the patients and were administered by nurses.
The urgency levels reported before the study started were also comparable (P=0.13). Of note, the levels of urgency were ‘medium’ or ‘high’ for all patients; two reported a ‘low’ level of discomfort.
Results of the treatment: bowel emptying
The number of weekly defecations at the end of the study was significantly higher in the group undergoing an abdominal massage; the mean number in the control group did not change over the four weeks (31±3 defecations), while in the intervention group it increased from 31±2 to 51±4 (P<0.001, w2=0.68, 95%IC[0.59;0.77]). The Bristol scores remain unchanged in the controls but increased significantly in the interventions (median=4, IQR[3;4], P<0.0001, w2=0.20, 95%IC[0.19;0.21]). The number of low-volume enemas decreased significantly in the intervention group (median=1.5, IQR=[1.0-5.0] to median=1.0, IQR=[1.0-4.0], P=0.004, w2=0.75, 95%IC[0.71;0.78]). Discomfort and defecatory urgency perceived were studied as part of the first criterion of Rome IV (straining in 25% of evacuations or more). Urgency did not improve, but the feeling of medium or high discomfort shows a statistically significant decrease (P<0.0001). The person who used to perform manual emptying continued to use this manoeuvre. PEG administration was not stopped for any patient.
Food intake and appetite
The number of people eating more than half of their food increased from 14 to 15 in the intervention group, whereas those who ate 50% or less decreased from 10 to 7. One patient in the intervention group reported consuming the entire meal, whereas no patient had this characteristic at baseline. Appetite reported by the patients improved significantly in the intervention group: the number of participants with ‘small appetite’ decreased from 5 to 2 (P=0.01), and those with ‘medium appetite’ increased from 4 to 7 (P=0.01). The number of subjects who reported a ‘big appetite’ remained unchanged in both groups (Table 3).
Table 3. Main outcomes
| Outcome | T0 | T1 | P |
|---|---|---|---|
| Weekly defecations | Mean | Mean | |
| Control group | 31±3 | 31±3 | >0.05 |
| Intervention group | 31±2 | 51±4 | <0.001 |
| Medium/high discomfort referred during the evacuation | n (%) | n (%) | |
| Control group (n=22) | 21 (95) | 19 (86) | >0.05 |
| Intervention group (n=10) | 9 (90) | 4 (40) | <0.0001 |
| Monthly low-volume enemas | Median [IQR] | Median [IQR] | |
| Control group | 1.5 [1.0-4.5] | 1.5 [1.5-3.5] | >0.05 |
| Intervention group | 1.5 [1.0-5.0] | 1.0 [1.0-4.0] | 0.004 |
| Appetite | n | n | |
| Low level – Control group | 11 | 9 | >0.05 |
| Low level – Intervention group | 5 | 2 | 0.01 |
| Medium level – Control group | 9 | 11 | >0.05 |
| Medium level – Intervention group | 4 | 7 | 0.01 |
Satisfaction of the recipients
At the end of the study, a Likert scale ranging from 0 (‘I do not like it and it was not useful’) to 10 (‘I like it and I think it's beneficial’) was proposed and explained to all the participants in the intervention group. The scores had a median of 8, IQR[7;9].
Discussion
The cornerstone of the study was the choice to maintain any oral therapy prescribed by the physician at the facility, as a guideline supported by high-level evidence. As a general consideration, the intervention programme was warmly accepted by the recipients, who were glad to receive the massage and discuss their bowel problems. They described the abdominal massage as ‘a caress’, a source of general wellbeing, tranquillity, and pleasure. It is important to point out that the intervention group patients considered the abdominal massage session an integral part of their daily routine, a fixed appointment, greatly looking forward to the massage.
No one experienced pain or discomfort. Due to postural limitations, not all participants could benefit from the correct defecatory position; however, significant improvements were found in all major outcomes.
Regarding practical criteria for applying the protocol, abdominal assessment sometimes led to slight adjustments of the intervention: faecal obstruction was found in two patients in the first week of the study (one patient had obstruction at the level of the ascending colon and another patient had this issue at the level of the descending colon). It was therefore decided that, whatever the site of faecal obstruction, the practitioner should proceed by performing 10 effleurages. This finding supports the need for a preliminary assessment of bowel conditions.
Of note, administration of low-volume enemas and days without defecations in the intervention group often corresponded to Saturday or Sunday, when the protocol was not applied (data collection took place from Monday to Friday due to the availability of the person in charge of data collection).
As regards comparison with the literature, no studies have so far investigated the effects of abdominal massage on food intake and appetite; as the recipients of this programme are older adults, this is an important aspect to consider. Furthermore, it is known that malnutrition is related to insufficient intake of calories and proteins (Bollo et al, 2019) and persons aged 85 or more, like those in the sample here, have a higher risk (Bonetti et al, 2017). For these reasons, the findings in this study deserve further investigation on larger samples.
The authors also investigated the feasibility of daily routine nursing care; the massage only takes 10 minutes, but assessment, preparation and data collection increases this time by 20 minutes. However, many informal caregivers (eg, sons/daughters) have expressed their interest in learning the basic abdominal massage techniques to help their older relatives. It would be interesting to have nurses assessing bowel conditions, while the massage could be taught by nurses to the caregivers by using the YouTube video and giving a practical demonstration. This possibility needs to be investigated to verify the actual ability of non-professional caregivers to perform the procedure correctly and safely.
The main limitation of this study is represented by the reduced sample size and small block randomisation size, which increases the risk that the allocation process may be predictable with risk of study bias. It would be interesting to continue the study, including a wider sample, to support the results of this study. An underpowered study increases the probability of results attributable to chance. On the other hand, one of the reasons behind a pilot study is to obtain the required preliminary data to calculate a sample size for the primary outcome (In, 2017). For this reason, future investigations on this topic will include sample size calculation based on the data from this sample.
It must be emphasised that a maximum response rate was obtained; thus, the data collected can be considered representative of the population studied.
Conclusions
This study was intended as a pilot investigation; the main results of this project are an evidence-based protocol, which appears promising (although on a small sample of patients), educational material, and findings about food intake, which were unavailable in the literature. For these reasons, the study deserves further development on larger samples and possibly in multicentre investigations.
KEY POINTS
- Treating constipation in older adults involves using polyethene glycol or lactulose. However, these drugs lose effectiveness and have contraindications when used in the long term
- This study looked at the effects of abdominal massage and bowel training combined with polyethylene glycol, compared with laxative alone, on the frequency of evacuations, use of laxatives, appetite, and food intake of older people
- The number of evacuations and Bristol stool scores increased significantly. In addition, the number of enemas was significantly reduced. All the participants receiving the massage perceived it as pleasant; a clinically relevant improvement in appetite was reported
- The treatment requires time; an instructional video and a QR code for easy access to this online were prepared for caregiver training
CPD reflective questions
- How do you assess a patient with constipation in your daily practice?
- What are the problems that can result from the overuse of laxatives? How would you identify these in a patient?
- What are the strategies for treating constipation?