Constipation defined as changes in the frequency, volume, weight, consistency and ease of passage of the stool occurs in any age group. The most important factors known to promote constipation are reduced physical activity and inadequate dietary intake of fibres, carbohydrates and fluids. Fluid losses induced by diarrhoea and febrile illness alter water balance and promote constipation. When children increase their water consumption above their usual intake, no change in stool frequency and consistency was observed. The improvement of constipation by increasing water intake, therefore, may be effective in children only when voluntary fluid consumption is lower-than-normal for the child's age and activity level. In the elderly, low fluid intake, which may be indicative of hypohydration, was a cause of constipation and a significant relationship between liquid deprivation from 2500 to 500 ml per day and constipation was reported. Dehydration is also observed when saline laxatives are used for the treatment of constipation if fluid replacement is not maintained and may affect the efficacy of the treatment. While sulphate in drinking water does not appear to have a significant laxative effect, fluid intake and magnesium sulphate-rich mineral waters were shown to improve constipation in healthy infants. In conclusion, fluid loss and fluid restriction and thus de-or hypohydration increase constipation. It is thus important to maintain euhydration as a prevention of constipation.
The word constipation comes from the Latin word ‘constipare’, meaning to crowd together. The term constipation includes a range of symptoms and signs that either suggest that there is undue difficulty with defaecation, or that abnormal faecal retention, ‘impaction’, is present. The simplest definition of constipation is the difficulty of passage of large or hard stools, irrespective of frequency (Lewis & Rudolph, 1997). However, constipation is usually defined in terms of changes in the frequency, volume, weight, consistency or ease of passage of the stool. It may be defined variously as the occurrence of dry, hard and/or small stools, infrequent stools, stools that are difficult to expel, or sensation of incomplete evacuation.
In children, stool frequency depends on age. There is a decline from more than 4 stools per day during the first week of life (Nyhan, 1952) to a mean of 1.7 stools per day at 2 y of age, and to 1.2 stools per day at 4 y of age with a corresponding increase in stool size. It is considered that bowel movements occur from three times per day to three times per week in 96% of children 3–4 y of age as well as between 94 and 98% in adults (Weaver, 1988). No difference in bowel movements is found between sexes, but movement frequency is higher in children younger than 3 y than in older children (Corazziari et al, 1985). By the age of four, an adult pattern of bowel movement is achieved (Weaver, 1988). In prepubertal children, constipation is more common in boys than in girls (3:1 ratio) and this ratio reverses in adolescence with constipation being three times more common in postpubertal females. In adult subjects on a Western diet at least three bowel movements per week are reported. A lower bowel frequency indicates constipation and was observed in 1 and 5% of a healthy British and American population, respectively (Read & Timms, 1986). In healthy subjects fed a normal mixed diet, the daily stool output varies between 50 and about 400 g. The mean value was significantly lower in women 149 and 112 g than in men, 229 and 166 g, respectively (P<0.001 and <0.05) (Cann et al, 1983; Davies et al, 1986). As 90% of the total gut transit time of food residue takes place in the colon, the use of ingested markers is an accurate measurement of colonic transit (Read & Timms, 1986). The difference in gastrointestinal transit time between men and women was significant with median values of 1.9 and 2.4 days, respectively, confirming previous observations showing that women had significantly longer transit for the whole colon than did men (Abrahamsson et al, 1988). In constipated patients matched for age and sex with healthy controls, a prolonged gastrointestinal transit time of 4.7 days instead of 2.4 days as well as transit dysfunction in separate colonic segment were reported (Abrahamsson et al, 1988).
There are multiple conditions and physiopathologic states that lead to difficult or infrequent passage of stools (Smith et al, 1955; McClung et al, 1995). As an example, since the first report by Hirschsprung on the constipation in newborns due to dilatation and hypertrophy of the colon (Hirschsprung, 1887), several review papers have described the incidence, aetiology and diagnostic of this disease (Ehrenpreis, 1971). However, the large number of definitions of the symptom of constipation has led to considerable confusion rather than to a common understanding and treatment (McClung et al, 1995).
Faecal impaction can occur in any age group (Dresen & Krattzer, 1959). A survey performed in the United Kingdom indicated that 10% of adults reported constipation in the preceding 2 weeks (Murtag, 1990). Its occurrence appears to be more frequent in incapacitated or institutionalised elderly people, where 42% of the patients admitted to geriatric wards were diagnosed (Read et al, 1985). Faecal impaction was the main reason for admission in 18% of acutely ill and in 27% of chronically ill patients (Read et al, 1985).
Chronic constipation is also a very common clinical complaint in childhood, where it presents a significant diagnostic and management problem for the practising clinician and represents a reason for parents to seek medical care for their child. Estimates of constipation have varied from 3% of the paediatric population to as high as 8% (Loening-Baucke, 1993a). Constipation accounts for approximately 20–25% of visits to paediatric gastroenterology clinics (Loening-Baucke, 1993b). In a study, 16% of the parents of 22-month-old children reported constipation in their toddlers (Issenmann et al, 1987). However, poor patients/parent compliance has been blamed for less than ideal results in a major series on constipation in children (Loening-Baucke, 1993b). These investigators showed that when children have recovered, their parents still reported that 5% of the recovered children had stool withholding, 9% had abdominal pain and 24% occasionally passed large stools.
The colon is the primary organ involved in the physiopathology of impaction or constipation. Impaction can occur anywhere in the colon. Normal colonic transit time from caecum to rectum occurs in 24–100 h in adults and within 48 h in healthy children (Lewis & Rudolph, 1997) or between 19 and 33 h in 5-y-old children (Corazziari, 1985). Impaction typical symptoms are anorexia, nausea, vomiting, abdominal pain and abdominal tenderness, and distension. Particular attention should be paid to the general appearance of the patient, the state of hydration and evidence of weight loss.
It has been reported that a 20% difference in water content of the digesta from the caecum (91%) to the rectum (71%) in control pigs resulted in a 240-fold increased viscosity. These data demonstrate that once water content falls below 75%, a small decrease in water content results in a relatively large increase in digesta/stool viscosity (McRorie et al, 1998). In subjects of the same age and eating patterns, the water content of stools appears to remain constant, between 70 and 75% (Weaver, 1988). Lowering the viscosity of digesta, therefore, should have the effect of increasing the rate of flow or transit rate. The transit is described mathematically by using a rheological value known as the Deborah number. The formula for the Deborah number is the time constant of the substance (proportional to viscosity) divided by the duration of the event. A Deborah number greater than 1.0 would exhibit a more solid-like behaviour. A Deborah number less than 1.0 would exhibit a more fluid-like behaviour. The Deborah number predicts that high viscosity substrates, such as hard stool, would only be propelled by a few slowly propagating, high-amplitude, long-duration contractions, resulting in relatively slow colonic transit rates (McRorie et al, 1998).
Water flux, including secretion and absorption, through the gastrointestinal tract is shown in Figure 1 If we consider the physiology of the human intestinal tract, most liquid in the human diet (max. 1 l) in combination with approximately 8 l of gastric and intestinal secretions is absorbed by the small intestine which has an absorptive capacity of about 15 l/day (Figure 1). The remaining 1.5 l passing through the ileocaecal valve into the adult colon is then absorbed by the right portion of the colon before desiccation. The colon is capable of absorbing up to 5 l/day. Reabsorbed fluid goes back into the circulatory and lymphatic systems, where it is removed by the kidneys. As a consequence, drink and water intake appears to have little input into the enterosystemic circulation of fluid so that minor modifications in liquid intake or hypertonic beverages will not significantly alter stool consistency in the constipated adult and child. However, hypohydration may affect all secretions and water flux and, as a consequence, stool consistency.
The colon serves as a site of fluid and electrolyte absorption as well as a conduit and storage reservoir for faeces. The majority of the 1.5 l of fluid entering an adult colon daily are absorbed in the ascending and transverse colon, yielding a stool fluid volume of approximately 100 ml. Although rapid emptying occurs from the caecum and ascending colon, faecal matter generally is retained several hours in the transverse colon. The descending colon then propels faeces into the sigmoid colon where they are stored. High-amplitude colonic contractions propagate from the proximal to distal sigmoid colon, pushing the stool mass into the rectum (gastrocolic reflex). Using ingested markers, normal colonic transit time from caecum to rectum occurs in 24–100 h in adults and within 48 h in healthy children. However, as faecal material remains in the colon or rectum, continuing reabsorption of water results in hard and small stools that are difficult to expel.
These effects may be explained by prostaglandins that promote solute and water secretion and affect intestinal motility and thus the transit time, so that either diarrhoea or episode of constipation may be observed (Branski et al, 1986). Excess of endogenous opioids or altered opiate receptor affinity may also contribute to hypomotility in severe idiopathic constipation and a specific antagonist such as naloxone was shown to increase passage of faeces and increase wet and dry faecal weight (Kreek et al, 1983).
Factors affecting constipation
The major causes of constipation are listed in Table 1.
Nonorganic causes of constipation in children
Nonorganic causes of constipation in children include the ‘Grunting baby’, the infant dyschezia and functional faecal retention. In children, a usual neglect of the urge to defaecate because it interferes with playing may promote impaction. In children with earlier constipation, parent should pay attention to the child's stool frequency during holidays, travel, or significant changes in the family routine, as this is a common time of relapse of faecal retention.
Systemic and local functional factors
There are both systemic and local factors that influence colon excretory and motility functions. Simple constipation has been suggested to be primarily a colonic motility disorder. The three types of muscular activity, segmentation waves, peristaltic waves and mass movements, are affected by systemic influences such as sympathetic/parasympathetic activity, hormone secretions and metabolic factors. A wide range of values for frequency of contractions, duration, amplitude, percent of activity and surface area under the contraction curves were shown in children (Loening-Baucke & Younoszai, 1984). Stool accumulates in the left colon, where excess desiccation occurs, with the stool becoming hard, firm and difficult to pass.
Anal hypertonus shown by an increase in the maximal anal resting closure pressure (MARCP) was found in 46% of the constipated children. Decreased rectal sensitivity threshold was found in 68% of the constipated children (Meunier et al, 1979a; Loening-Baucke, 1984). Lower rectal pressures were also recorded during rectal distension in the constipated group than in the control (Read et al, 1985). The mean values for percent of activity and surface area under the contraction curves were significantly lower in the constipated than in the control children (Loening-Baucke & Younoszai, 1984). Therefore, it appears that the hypomotility was the result of the chronic faecal impaction and rectal distension and while it was not the cause of constipation, it may contribute to its severity.
Significant impairment of rectal sensory threshold was apparent in constipation and total gastrointestinal transit times (TGITT) were prolonged in the constipation group, mainly distally due to rectal stasis. Evaluation of TGITT in the adult population shows an upper value of 96 h and does not exceed 33 h in normal children (Corazziari et al, 1985). However, constipation in the elderly is not merely due to delayed transit and neurogenic deficits of sacral spinal cord function but also due to abnormalities in rectal motor and sensory function (Varma et al, 1988).
A most important local factor is the reduced physical activity. Upright posture and exercise have been shown to promote colonic motility.
The relationship of diet to the onset of constipation is important in adults as well as in infants. The introduction of solid foods, change from breast feeding to bottle feeding, change from formula to cow's milk or excessive intake of cow's milk may result in constipation in babies.
In a study performed on 49 constipated adult patients (15–72 y) and compared with 17 healthy volunteers (19–33 y), it was shown that the meal increased the motility when compared to fasting subjects (Meunier et al, 1979b). In constipated patients, the mean meal motility index (577±549) was comparable with that of controls (538±215) with a large distribution of individual values. The patients were segregated into three groups: hypomotor, normomotor and hypermotor patients. Only the meal test is able to identify three significant patterns of sigmoid activity and a large number (68%) of constipated patients were hypermotor (Meunier et al, 1979b). Adequate intake of fibres, carbohydrates and fluid are known to prevent constipation.
Diets containing fibres result in large, soft stools going rapidly through the intestine. By contrast, the refined low-fibre foods of developed countries produce small firm stools, which pass through the gut slowly. Many slimming diets lead to constipation due to inadequate fibre content (Clark & Godfrey, 1981). There is an inverse relationship between daily stool weights and the transit time. The consistency of faeces is related to bulk and viscosity. The bulk is largely determined by dietary fibre intake, which varies widely in the population, regardless of race (Burkitt et al, 1972). Viscosity also varies widely and is usually inversely related to bulk (Clark & Godfrey, 1981).
Infants may become constipated if there is little bulk in their diets, usually originating from milk constituents (eg Ca salts of long-chain saturated fat acids) (Olney, 1976). Children 2–12 y of age with chronic constipation treated daily with a fibre supplement in a dose of 0.25 g/kg during a 6-month period were shown to consume less than one-fourth of the recommended fibre intake (McClung et al, 1995). With this supplement, the speed of the intestinal transit was not changed but the dry weight of the children's stools increased approximately 40%, whereas the wet weight of their stools increased approximately 300% (McClung et al, 1995). In contrast, when 30 g of fibre as wheat bran and vegetable fibre were administered in healthy men and women, the mean transit time was faster and stool weights greater for men than women (Lampe et al, 1993). In a crossover design study, 12 men consumed either a low- or a high-fibre diet corresponding to 13 or 166 mg/kg a body weight crude fibre from fruits and vegetables. Significant changes were observed in the wet faecal weight (89 and 209 g/day) and water in faeces (65.6 and 156 g/day), while the percentages of water in faeces were not significantly different, 73 and 74.6%, respectively (Kelsay et al, 1978).
Bran is a bulking agent with capacity to hold water in the stool, thereby improving bowel function. Particle size is important to determine the effect of bran in increasing stool weight (Brodribb & Groves, 1978).
It is, however, difficult to maintain high-fibre diet (McClung et al, 1995). The mechanisms involve active electrolyte secretion, decreased water and electrolyte absorption, increased intraluminal osmolarity, and an increased hydrostatic pressure in the gut. The average weight of stools has been calculated in a controlled study performed between 1950 to 1953 on 115 healthy individuals in a prison in Texas. The diet was adequate and well balanced nutritionally and consisted of 14 600 kJ/day. With the exception of meat, free choice of quantity and variety of food was allowed. The average weight of stools was 123.6±40.2 g based on 8267 stools analysed (Rendtorff & Kashgarian, 1967).
Provision of carbohydrate promotes the colonisation of the colon by lactobacilli, the production of organic anions and an acid pH in the lumen (Calloway, 1964), all of which stimulate motility. In infants, the absorptive capacity of carbohydrates such as lactose in the small intestine is so limited that its malabsorption may result in osmotic diarrhoea. Any imbalance in the colon normal absorption of salt and water will contribute to the hardening of the stool.
Electrolyte abnormalities, such as hypercalcaemia, hyponatraemia or hypokalaemia may be associated with impaction. These disorders induce water depletion or alter smooth muscle function and are suggested to play a role in constipation. Iron supplement may induce constipation. However, in a study performed on 90 children ranging in age from 2 to 12 y with a history of chronic constipation, calcium and iron intake appears to have no effect on constipation, while increased fat intake above 100% RDAs did have an adverse effect on constipation (Anonymous, 1999). There are unusual cases of calcium-induced constipation demonstrated by the calcium provocation test. It was also shown with a magnesium-competition test that the effect of calcium on constipation was abolished by magnesium. The effects of these minerals are most probably exerted locally in the intestine or colon (Frithz et al, 1991).
The most common type of constipation in children is simple constipation. As first intervention steps, it requires behavioural training and dietary changes such as an increased liquid and fibre intake and considerations of restriction of constipating foods (Tedesco, 1985; Felt et al, 1999). Inadequate fluid intake or excessive fluid loss from diarrhoea, vomiting or febrile illness may cause hardening of the stool and is considered to be an important cause of constipation, especially in infants (Leung et al, 1996). Increasing liquid intake is commonly recommended for constipated children, adults and elderly subjects (Waggener, 1966; Tasman-Jones, 1973; Hyams, 1974; Bank & Marks, 1977; Klein, 1982; Read & Timms, 1986; Fitzgerald, 1987; Tremaine, 1990; Clayden, 1992; Squires & Doerck, 1993; Dardaine et al, 1994). Although the effects of fluid intake on constipation have never been fully studied (Young et al, 1998) or understood, the recommendation has remained mostly out of tradition (Anonymous, 1999).
In 90 acceptably euhydrated children (59 girls and 31 boys) aged 2–12 y, a 50% increase of water intake or an additional intake of hyperosmolar liquid had no effect on daily stool frequency, consistency and difficulty with passage daily assessed with the Constipation Assessment Scale (Young et al, 1998).
In the elderly, low liquid intake may be indicative of hypohydration, often thought to be a cause of constipation. Poor dentition, a decreased production of saliva and dehydration, especially in the warmer months makes the problem more complex (Wrenn, 1989). In 796 participants of the New Mexico Elderly Health Survey, daily fluid intake of less than three glasses, three to five glasses and six or more glasses showed no significant associations with the frequency of chronic constipation. However, a trend of lower frequency of constipation with higher intake was observed (20, 15 or 11%, respectively; Lindeman et al, 2000). In 18 constipated elderly and 18 controls measures included a 1-week food diary, a bowel diary and a colonic transit study (Towers et al, 1994). Slow colonic transit was significantly related to low caloric intake, higher percent of protein in the diet and a low fluid intake. Elderly home residents (n=21 012) were assessed at admission and 3 months later (Robson et al, 2000). By 3 months, 1291 residents had developed constipation. Stepwise logistic regression analysis showed poor consumption of fluids as second independent risk factor. These results support the observation that constipation is a ‘summer disease’ that may be treated by increasing the intake of fluid (Read & Timms, 1986). However, it appears more relevant to associate constipation with hypohydration status rather than fluid intake (Welch et al, 1980; Klein, 1982; Wright & Staats, 1986; Portenoy, 1987).
In a recent study, eight healthy adults received in randomised order more than 2500 ml fluid per day for 1 week and less than 500 ml/day in another week. A period of washout of 1 week was in-between (Klauser et al, 1990). Stool frequency and stool weight were significantly decreased during the liquid deprivation week, despite no change in the mean oral–anal transit time. Six healthy volunteers underwent extra intake of water and nine volunteers of isotonic fluids (Chung et al, 1999). Compared to baseline periods, additional intake of fluid did not result in a significant change of stool output.
A literature review provided only one study that showed a relationship between liquid deprivation and chronic constipation (Young et al, 1998; Anonymous, 1999) and no studies have been published showing that increasing liquid volume is effective as a treatment in euhydrated subjects with chronic constipation.
Only small minorities of patients who complain about constipation have underlying organic disorders including anorectal malformations, neurologic disorders, colonic neuromuscular disorders, and metabolic and medication-induced constipation.
Diseases such as hypothyroidism (De Lorenzo et al, 2000), diabetes (Ishikawa 2000), cystic fibrosis (Borovnicar et al, 2000), chronic renal failure (Alvarezlara et al, 2001) and cancer (Portenoy, 1987) may modify hydro-electrolytic balance. Faecal impaction was also observed in patients with spinal cord injury and brainstem lesion as explained by diminished visceral sensitivity (Gore et al, 1981; Weber et al, 1985).
Drugs inducing constipation
In adult and in elderly people, the importance of the use of drugs such as narcotics, antihypertensive medications, diuretics, antacids containing aluminium and the long-term use of stimulants or bulk-forming laxatives may cause impaction (Tedesco, 1985; Read & Timms, 1987). More recently, calcium channel blockers have been added to the list of compounds that can cause constipation (Nyska et al, 1994). Nifedipine, verapamil and diltiazem have been demonstrated to have a profound inhibitory effect on intestinal motility. Constipation induced by these drugs is due to delayed colonic transit time without any effect on gastric emptying or distal ileal filling. A delay in transit time in the colon may permit increased local fluid absorption of intestinal contents due to the extended mucosal contact duration. Constipation was not explained by dehydration, as water was available at any time throughout the study periods (Krevsky et al, 1992).
For the treatment of impaction with saline laxatives (Kamm & Speakman, 1993), the active ingredients, magnesium hydroxide (Kinnunen & Salokannel, 1989) or magnesium sulphate (Epsom salts), phosphate and citrate ions are relatively fast acting (American College of Gastroenterology's Committee, 1985). Laxatives have been generally classified according to their chemical properties rather than their mechanisms by which they increase stool water (Fingl, 1983). However, in 1975, the FDA used five overlapping categories: (1) stimulant of intestinal motor function, (2) saline or osmotic, (3) stool softener, (4) bulk and (5) lubricant (Donowitz, 1979; Thompson, 1980). As with all laxatives, dehydration may occur if fluid replacement is not maintained (American College of Gastroenterology's Committee, 1985). The daily ingestion of 3.5 g magnesium in healthy subjects produced a reduction of urinary volume from 1537±162 to 1173±72 ml, reflecting faecal loss of water related to the nonabsorbed magnesium acting as osmotic cathartic (Lembcke & Fuchs, 1984).
Laxative properties of hard waters have been suggested in consumers, particularly in sensitive population such as infants (Chien et al, 1968). These results were, however, not confirmed and milk appears to have a more significant laxative effect than does sulphate in water (Esteban et al, 1997). Laxative effect of sulphate in drinking water was suggested from case studies and anecdotal reports when transient populations such as tourists experienced an abrupt change in concentration. However, most published data concern uncontrolled studies performed in spa (Grossi & Scalabrino, 1996) or investigations performed for ethical reasons on animal models, particularly piglets (Veenhuizen et al, 1992: Gomez et al, 1995) and swine (Veenhuizen, 1993). A field investigation with a recruitment of more than 800 infants exposed to naturally occurring high levels of sulphate in drinking water failed because the number of mothers who intend to feed their babies using tap water to mix the infant milk formula was too low (Federal Register Notice, US EPA, 1999). Another intervention trial in adults subjects drinking water with sodium sulphate added at concentrations from 0, 250, 500, 800 and 1200 mg shows no effect on bowel movements as well as no diarrhoea after these acute exposures of sodium sulphate (Federal Register Notice, US EPA, 1999). With similar concentration range, a decrease in the mouth-to-anus appearance time was observed. For the highest concentration, a significant increase in stool mass was observed, but none of the study subjects reported diarrhoea (Heizer et al, 1997). In another recent clinical study performed on 60 young healthy male and female infants fed exclusively infant formulas (67%) with breast milk occasionally, constipation was diagnosed on the basis of examination by their paediatricians. The criterion was either the presence of hard stools for at least 3 days or the presence of a single stool every 36 h for 5 days. The study shows a significant improvement of constipation evaluated from stool frequency and stool consistency (p<0.05) due to an increase in water consumption. A significant inverse correlation was observed between total fluid ingested and hard stool production. This study confirms the recommendation to simply increase intake of fluids such as water to prevent constipation (Constant et al, 1999).
The therapeutic regimen begins with disimpaction: hyperosmolar milk and molasse enemas, mineral oil, lactulose, polyethylene glycol-balanced electrolyte solutions, milk of magnesia and citrate of magnesia can be used to mobilise the faecal mass. However, severe magnesium toxicity with cardiorespiratory arrest has been reported after magnesium sulphate enema administration in a chronically constipated child.
In the management of faecal impaction, long-range programmes are often established to prevent recurrence of the problem. Increased fluid intake is frequently mentioned (Waggener, 1966; Tasman-Jones, 1973; Hyams, 1974; Bank & Marks, 1977; Klein, 1982; Read & Timms, 1986; Fitzgerald, 1987; Tremaine, 1990; Clayden, 1992; Squires & Doerck, 1993; Dardaine et al, 1994).
A recent study showed that increased fluid intake is not beneficial in treating adequately euhydrated children with constipation and should not be recommended unless history suggests a lower-than-normal fluid intake for the child's age and activity level (Young et al, 1998). However, Young et al indicate that most children studied already consume liquids in excess. The absence of tools to evaluate hydration status may explain why dietary factors and fluid intake have been considered to have little input into the enterosystemic circulation of fluids. Consequently, minor modifications in liquid intake, or even the administration of excess hypertonic beverages were believed to be unable to significantly alter stool consistency in the constipated children (Young et al, 1998).
In contrast to these explanations, dehydration of the colonic contents will harden stool consistency. In adult and elderly subjects, poor intake of water or excessive loss of fluid and electrolytes from the body such as during vomiting, high sweat rates or from renal disease can reduce water content of the stools and lead to constipation. This statement is now supported by a study showing clearly that fluid restriction and thus dehydration increased constipation (Klauser et al, 1990). When fluid intake was increased, the discrepancies of the results may be explained by a beneficial effect limited to subjects with dehydration, while fluid overload will not improve stool consistency.
A simple, noninvasive and accurate test of hydration status would be necessary to better demonstrate the specific role of hydration status and to propose individual recommendations to either increase the fluid intake or not, to prevent constipation.