Age and Ageing 2003; 32: 388–393 Age and Ageing Vol. 32 No. 4 # 2003, British Geriatrics Society.

ics Society. All rights reserved.

A comparison of two feeding methods

in the alleviation of diarrhoea in older
tube-fed patients: a randomised
controlled trial
J ENNY S HUN W AH L EE, T UNG W AI A UYEUNG
Department of Medicine and Geriatrics, Tuen Mun Hospital, Tuen Mun, New Territories, Hong Kong, China

Address correspondence to: J. S. W. Lee. Fax: (q852) 246 85800. Email: jleesw@hotmail.com

Downloaded from http://ageing.oxfordjournals.org by on May 15, 2010

Abstract
Objective: to compare the effect of two feeding methods on older tube-fed patients suffering from diarrhoea.
Design: randomised controlled study.
Setting: geriatric wards in a regional hospital.
Subjects: 105 patients over the age of 60 who developed diarrhoea while on intermittent bolus tube feeding.
Methods: subjects were randomised either to switch to continuous tube feeding or to continue with intermittent
bolus tube feeding for 3 days, with some extended to 5 days. Incontinence pads were inspected 6 times per day by
independent observers. Each motion was evaluated using a validated score according to its volume and consistency.
Stools were sent for pathogen culture and Clostridium difficile cytotoxin assay. Those with positive assays were excluded
from analysis. Gastric aspirations were performed at regular intervals and significant residual volumes recorded.
Results: 86 patients completed the study. Twelve patients (14%) were excluded due to positive CD stool assays.
Seventy-four patients were entered into analysis. The median diarrhoea scores of the continuous infusion group and
the intermittent bolus group were: day 0 (12 versus 10, P=0.18), day 3 (4 versus 6, P=0.23) and day 5 (5 versus 4, P=0.83)
respectively. There was no statistical difference between the two groups. Two patients in the continuous infusion group
(n=37) and 4 patients in the intermittent bolus group (n=37) had significant gastric residual volumes. The incidence
did not differ significantly.
Conclusion: contrary to common belief and recommendation, we could not find a significant difference between the
continuous feeding method and the intermittent feeding method in the alleviation of diarrhoea in tube-fed patients.
We do not recommend routine switch to continuous pump feeding in the management of tube-fed patients suffering
from diarrhoea.

Keywords: older adults, tube feeding, diarrhoea, randomised

Introduction such as Clostridium difficile (CD) cytotoxin [4, 5], recent or

Diarrhoea is one of the commonest gastrointestinal
problems seen in tube-fed patients. In studies, the inci-
dence of diarrhoea in tube-fed patients varies from 2 to
68% [1, 2]. This wide discrepancy arises from the fact
that there is no standardised definition of diarrhoea;
some define it by stool weight, others by frequency of
bowel openings, and still others by the consistency of
the stool [3].
A large number of factors have been studied in
the past for an association with diarrhoea in tube-fed
patients. Established culprits include infective causes
concurrent antibiotic usage [1, 4], syrup medication
constituents, namely sorbitol [5], and of course spurious
diarrhoea secondary to faecal impaction, another common
gastrointestinal problem in tube-fed patients. So far there
are no conclusive results regarding factors such as fibre
content [6–10], osmolality of the feeding formula [11],
or the method of feeding, i.e. intermittent bolus or
continuous infusion [12, 13], though general recom-
mendations were to start continuous infusion if tube-fed
patients experience diarrhoea.
There has been a paucity of studies of older tube-fed
patients despite the large number of older adults put on

388

tube feeding for various reasons, especially neuro-
dysphagia after stroke or in late stage dementia. Most
clinical trials in tube-feeding have involved patients
in an intensive care unit (ICU) setting and frequently in
post-operative conditions.
There have been two randomised trials which com-
pared intermittent and continuous tube feeding in terms
of the incidence of complications including diarrhoea,
aspiration pneumonia, clogged tubes, and the time to
achieve the required caloric goal. Hiebert et al. [12] studied
76 adult burn patients in a burn centre and randomised
them to receive either mode of feeding. Among the 62
patients (age 12–81 years) who completed the study
period of 1 week, those fed continuously had signi-
ficantly less frequent stools than intermittently fed
patients. The volume and consistency of stools were
not studied.
In the randomised trial done by Kocan and Hickisch
[13], 34 adult patients in a neurological intensive care unit
were randomised to receive either mode of tube feeding.
The number and consistency of stools per day, the inci-
dence of aspiration as evidenced by presence of blue dye
indicator in pulmonary secretions, the amount of residual
gastric contents, the amount of nursing time required,
and the time required to meet caloric goals were com-
pared. Only 17 patients completed the entire study period
of 10 days, the mean study period being 7.6 days in both
groups. No significant differences were found in the
number or consistency of stools per day, incidence of
aspiration, or caloric intake as a percentage of the
nutritional goal.
There is only one randomised prospective trial that
has studied the effect of the mode of feeding on the
incidence of diarrhoea exclusively in older patients [14].
In this study, which involved 30 elderly patients in each
group, continuous feeding was associated with a 30%
reduction in the incidence of diarrhoea (96% versus 66%,
P-0.008). This study randomised patients with no
diarrhoea to begin with and looked for the incidence of
diarrhoea after receiving different methods of feeding.
No study has ever been performed to study the effect of
switching to continuous pump feeding when patients
developed diarrhoea. Our study aims to see whether
changing to the continuous infusion mode of feeding
would be useful in the management of diarrhoea.
Patients and methods
One hundred and five patients receiving intermittent
bolus feeding and suffering from diarrhoea were recruited.
Forty-two patients suffered from advanced dementia; 33
patients had cerebrovascular disease; 25 patients were
tube-fed for nutritional support in acute illness; and 1
patient had oral malignancy. Diarrhoea was defined as
three or more bowel openings in one day. All subjects
were 60 years or older. Patients with laxative use in the
previous 72 hours were excluded, as were those with
Feeding methods for older tube-fed patients
spurious diarrhoea or with a history of frequent self-
extubation. The study was approved by the ethics
committee of the hospital and subjects or their next
of kin gave informed consent.
Subjects were randomised into two groups. Block
randomisation in groups of 6 (3 for intermittent bolus
feeding and 3 for continuous infusion feeding) was used
to avoid high discrepancy between the subject number in
each group. The first subject was assigned by drawing a
sealed envelope out of six. The second subject was
assigned by drawing a second envelope out of the
remaining five, and so on. When all the six envelopes had
been drawn the cycle was repeated again.
In the intermittent bolus group (I) the patients
continued to receive the same mode of feeding. In
the continuous infusion group (C) the patients were
switched to continuous infusion of the same daily
Downloaded from http://ageing.oxfordjournals.org by on May 15, 2010
volume of feeding over the next 3 days, with some
extended to 5 days. Continuous infusion was delivered
via a pump set at a rate of 1–2 ml per minute for 16
hours per day. Intermittent bolus feeding was delivered
via a hanging feeding funnel at a rate of 10 ml per
minute, 250 ml per bolus, with 4–6 bolus per day
depending on the prescribed caloric intake. Incontinence
pads were inspected every 4 hours by staff nurses who
were independent observers. The frequency of bowel
openings, volume and consistency of stool were recorded
daily. The volume of stool was graded as small, moderate
and large, and the consistency of stool was graded as
watery, loose and soft, according to standard nursing
practice in our hospital. Interrater reliability of observa-
tions by nurses had been established in a previous study
[15]. Aspiration of gastric residual contents was per-
formed every 4 hours during continuous infusion in
the C group and before each feeding in the I group.
Significant aspirated volumes of )100 ml were recorded
daily.
Factors which may be related to diarrhoea or nutri-
tional status of the subjects were recorded. These included
the Body Mass Index (BMI), with body height derived
from fibula length measurements in bed bound patients
[16], serum albumin level, total lymphocyte count,
concurrent use of antibiotics, syrup medications, fibre
content of the feeding formula, total daily volume and
osmolality of the feeding. The use of anti-diarrhoeal
drugs was also recorded. The clinicians in charge were
free to dilute the formula or prescribe anti-diarrhoeal
drugs as needed.
Stool specimens from each subject were sent for
pathogen culture as well as for CD cytotoxin assay.
Those who were subsequently found to have infective
causes of diarrhoea were excluded from data analysis.
Diarrhoea score
Severity of diarrhoea was graded according to a pre-
viously validated diarrhoea scoring system [15]. The
389
J. S. W. Lee, T. W. Auyeung
original scoring system was expanded to accommodate a
moderate grade for stool volume in accordance with the
standard nursing practice in our wards. The grades of 1,
2 and 3 were assigned to small, moderate and large stool
volumes respectively. The stool consistency was des-
cribed as soft, loose and watery, and the grades of 1, 2
and 3 were then also assigned respectively. The score for
each bowel motion is the product of the grading for both
consistency and volume. The total diarrhoea score for
each day was the sum of scores for all bowel motions
during the day. Hence the highest possible score for any
day would be 54 (9 for each motion and maximum 6
times per day), and the lowest possible score would be 0.
Higher scores meant more severe diarrhoea.
Statistical analysis
Our study was designed to detect a 10% difference
between the diarrhoea scores of the 2 groups, at a power
of 0.9. Within a score range of 0–54, any difference of
-10% was not likely to be clinically significant. The
required sample size was calculated to be 36 in each group.
The Statistical Package for Social Sciences (SPSS)
version 9.0 was used for statistical analysis. Student’s
unpaired t-test was used for parametric data analysis
while Chi-square test was used for dichotomous data.
Fisher’s Exact Test was used when the expected number
in any cell fell below 5. The Mann-Whitney U test was
used for the comparison of the diarrhoea scores in the
2 groups.
Figure 1. Trial profile.
390
Results
A total of 105 patients older than 60 years of age were
recruited. Of these, 86 patients completed the 3-day
study. Reasons for dropouts included discharge from
hospital (8), transfer to another facility (2), being kept
nil by mouth (3), resumed oral intake (2), or death (4).
Twelve patients (14%) were subsequently found to carry
CD cytotoxin in their stools and hence were excluded
from data analysis. A total of 74 patients were entered
into the final analysis (Figure 1).
There were 37 patients in the C group and 37 in the
I group. The mean age of patients was 81.3 years (C) and
82.8 years (I). The baseline characteristics and other
possible factors that might be related to diarrhoea
were compared and are tabulated in Table 1. None of
these baseline values differed significantly between the
Downloaded from http://ageing.oxfordjournals.org by on May 15, 2010
2 groups (Table 1).
The median diarrhoea scores at the baseline were 12
(C) and 10 (I) (P=0.175). The median scores on day 3
were 4 (C) and 6 (I) (P=0.230). Among those who were
able to have extended trial days (28 in the C group
and 20 in the I group), the diarrhoea scores were 5 (C)
and 4 (I) (P=0.833). There was no statistical difference
between the 2 groups in terms of the median diarrhoea
scores at baseline, day 3 and day 5 (Table 2).
Two of 37 subjects in the continuous feeding group
were detected to have significant gastric residual volume,
defined as )100 ml in this study. In the intermittent
group, the number of subjects was 4 out of 37 (P=0.674,
Fisher’s Exact Test).
 Feeding methods for older tube-fed patients

Table 1. Comparison of baseline characteristics

Continuous infusion (n = 37) Intermittent bolus (n = 37) P 95% CI
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Age (years) 81.3"1.5 82.8"1.5 0.484 5.60–2.68

Sex (female/male) 32/5 31/6 0.744 –
BMI (kg/m2) 22.0"0.8 21.0"0.9 0.425 1.50–3.51
Albumin 27.1"0.9 26.4"0.9 0.582 1.87–3.31
Total lymphocyte count (109/mm) 1.2"0.1 1.2"0.1 0.914 0.33–0.30
Volume of formula (litre/day) 1.2"0.0 1.2"0.0 0.809 0.01–0.01
Hyperosmolar formula (%) 7 (18.9) 4 (10.8) 0.327 –
Fibre-enriched formula (%) 26 (70.3) 22 (59.5) 0.330 –
Full strength formula (%) 31 (83.8) 33 (89.2) 0.496 –
Syrup use (%) 8 (21.6) 13 (35.1) 0.197 –
Antibiotic use (%) 25 (68) 26 (70) 0.802 –
Antidiarrhoeal use (%) 1 (2.7) 0 (0.0) 1.000 –

Plus-minus values denote means"standard error.

BMI = Body Mass Index.

Downloaded from http://ageing.oxfordjournals.org by on May 15, 2010

Table 2. Median diarrhoea scores at baseline, day 3 and day 5 in both groups
Diarrhoea scores
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Feeding method Baseline P = 0.175 Day 3 P = 0.230 Day 5 P = 0.833

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Continuous infusion 12 (4–33) n = 37 4 (0–29) n = 37 5 (0–19) n = 28

Intermittent bolus 10 (3–40) n = 37 6 (0–27) n = 37 4 (0–23) n = 20

Range given in brackets.

Discussion as the patients were dispersed in both space and time. It

is probably related to the frequent antibiotic usage in this
Continuous infusion feeding of milk formula has been frail elderly group suffering from recurrent infections.
recommended in textbooks and guidelines [17–20] as a Other bacterial pathogens were found to be a negligible
means of reducing the incidence of diarrhoea, but cause of diarrhoea in tube-fed patients in this study, as is
published data are limited. In our trial, which is the consistent with studies elsewhere. It is prudent to con-
largest at present, we could not find enough evidence clude, therefore, that CD is an important cause of dia-
to support the idea that changing to continuous pump rrhoea in tube-fed patients, and all patients with this
feeding can significantly alleviate diarrhoea in tube-fed problem should be screened for the cytotoxin.
patients. The volume, frequency and consistency of the There was a high prevalence of antibiotic use in both
stool as reflected by the diarrhoea scores in either group groups of subjects, 68% in the continuous feeding
did not differ significantly after 3 or 5 days despite group and 70% in the bolus feeding group. This is
different modes of feeding. consistent with the findings in previous studies where
Our sample was able to detect a 10% difference broad spectrum antibiotics were found to be the main
between the two groups at the power of 0.9. As any cause of diarrhoea in tube-fed patients [1, 4, 22, 23]. In
reduction of -10% in the score would be clinically one study, all antibiotic-related diarrhoea was controlled
insignificant, we therefore conclude that changing to with syrup codeine phosphate 30–150 mg daily. It has
continuous feeding does not appear to alter the course also been recommended that broad spectrum antibiotics
of diarrhoea. should be withheld or stopped if diarrhoea does occur
However, it is obvious from the data in this study that during tube feeding [19, 24].
diarrhoea in these patients tends to subside after 3 days The incidence of significant gastric residual volume
(Table 2). This may reflect the natural course of dia- ()100 ml) was higher in the intermittent bolus feeding
rrhoea in tube feeding. In that case we may conclude that group in this study, although due to the small number of
diarrhoea in tube-fed patients is a self-limiting disorder events statistical significance could not be reached. It is
and may require simple supportive measures only. an attractive theory that a smaller residual gastric volume
Twelve out of the 86 patients (14%) with diarrhoea would lead to a lower chance of regurgitation and hence
were found to carry CD cytotoxin in their stool speci- tracheal aspiration of gastric contents. As aspiration
mens. However, no patient had a positive culture for pneumonia is one of the commonest causes of hospital
stool bacterial pathogen. This is similar to the incidence admission and death in these patients, the use of
of CD related diarrhoea in other series [21]. This high continuous feeding with a low rate of feed delivery
incidence could not be accounted for by a ward outbreak may be desirable in this aspect.

391
J. S. W. Lee, T. W. Auyeung
This study was conducted in a hospital setting and
the results may therefore not be applicable to nursing
home residents. Our results may not be applicable to
younger patients in other settings, such as in intensive
care or burns units, or in post-operative states.
Conclusions
Contrary to common belief and recommendations, we
could not find evidence that a change to the continuous
mode of feeding can result in any additional reduction in
diarrhoea in elderly tube-fed patients, as compared with
those who continued with intermittent feeding. CD
cytotoxin has been found to be a cause of diarrhoea in
a significant proportion of patients in our study, and
routine stool assay of the toxin is recommended in these
situations. A higher proportion of patients on inter-
mittent bolus feeding were found to have high gastric
residual volumes on routine aspirations. However, whe-
ther this finding is related to the incidence of aspiration
pneumonia awaits further elucidation.
Key points
. Changing to continuous pump feeding does not provide
additional benefit to tube-fed patients with diarrhoea.
. The diarrhoea in tube-fed patients is usually self-limiting
and abates in a few days.
. Clostridium difficile related diarrhoea is an important
cause of diarrhoea in tube-fed patients and the
cytotoxin should always be screened for in such
circumstances.
Acknowledgements
We would like to thank the nursing staff of the geriatrics
wards for their support and patience, without which this
project would have been impossible. This study was
supported by a research grant from the Hong Kong
Geriatrics Society.
Disclaimer
There is no conflict of interest involved in the study.
The authors have not received any support from any
commercial organisations.
References
1. Gottschlich MM, Warden GD, Michel M et al. Diarrhoea in
tube-fed burn patients: incidence, aetiology, nutritional impact,
and prevention. J Parenter Enteral Nutr 1988; 12: 338–45.
2. Benya R, Layden TJ, Mobardan S. Diarrhoea associated with
tube feeding: the importance of using objective criteria. J Clin
Gastroenterol 1991; 13: 167–72.
392
3. Bliss DZ, Guenter PA, Settle RG. Defining and reporting
diarrhoea in tube-fed patients – what a mess! Am J Clin Nutr
1992; 55: 753–9.
4. Guenter PA, Settle RG, Perlmutter S, Marino PL,
DeSimone GA, Rolandelli RH. Tube feeding-related diarrhoea
in acutely ill patients. J Parenter Enteral Nutr 1991; 15: 277–80.
5. Edes TE, Walk BE, Austin JL. Diarrhoea in tube-fed
patients: feeding formula not necessarily the cause. Am J Med
1990; 88: 91–3.
6. Homann KK, Kemen M, Fuessenich C, Senkal M,
Zumtobel V. Reduction in diarrhoea incidence by soluble
fibre in patients receiving total or supplemental enteral
nutrition. J Parenter Enteral Nutr 1994; 18: 486–90.
7. Shankardass K, Chuchmach S, Chelswick K et al. Bowel
function of long-term tube-fed patients consuming formulae
with and without dietary fiber. J Parenter Enteral Nutr 1990;
14: 508–12.
Downloaded from http://ageing.oxfordjournals.org by on May 15, 2010
8. Dobb GJ, Towler SC. Diarrhoea during enteral feeding in
the critically ill: a comparison of feeds with and without fibre.
Intensive Care Med 1990; 16: 252–5.
9. Khalil L, Ho KH, Png D, Ong CL. The effect of enteral
fibre-containing feeds on stool parameters in the post-surgical
period. Singapore Med J 1998; 39: 156–9.
10. Lin HC, Zhao XT, Chu AW, Lin YP, Wang L. Fibre-
supplemented enteral formula slows intestinal transit by
intensifying inhibitory feedback from the distal gut. Am J
Clin Nutr 1997; 65: 1840–4.
11. Pesola GR, Hogg JE, Eissa N, Matthews DE, Carlon GC.
Hypertonic nasogastric tube feedings: do they cause diarrhoea?
Crit Care Med 1990; 18: 1378–82.
12. Hiebert JM, Brown A, Anderson RG, Halfacre S,
Rodeheaver GT, Edlich RF. Comparison of continuous vs
intermittent tube feedings in adult burn patients. J Parenter
Enteral Nutr 1981; 5: 73–5.
13. Kocan MJ, Hickisch SM. A comparison of continuous and
intermittent enteral nutrition in NICU patients. J Neurosci
Nurs 1986; 18: 333–7.
14. Ciocon JO, Galindo-Ciocon DJ, Tiessen C, Galindo D.
Continuous compared with intermittent tube feeding in the
elderly. J Parenter Enteral Nutr 1992; 16: 525–8.
15. Guenter PA, Sweed MR. A valid and reliable tool to
quantify stool output in tube-fed patients. J Parenter Enteral
Nutr 1998; 22: 147–51.
16. Auyeung TW, Lee JSW. Estimation of height in older
Chinese adults by measuring limb length. J Am Geriatr Soc
2001; 49: 684–5.
17. Reuben DB, Herr K, Pacala JT, Potter JF, Semla TP, Small
GW. Geriatrics at your Fingertips. 2000 Edition. USA: Excerpa
Medica, Inc., 2000: 82–4.
18. Grant A, Todd E. Enteral nutrition. In Grant A, Todd E eds.
Enteral and Parenteral Nutrition. 2nd Edition. Oxford: Blackwell
Scientific Publications, 1987: 74–95.
19. Thomas B. Manual of Dietetic Practice. 2nd Edition.
Oxford: Blackwell Scientific Publications, 1994: 65–75.

20. Stuart S, Stuart M, Unger LD. Enteral and parenteral nutrition
support. In Morrison G, Hark L eds. Medical Nutrition and
Disease. New York: Blackwell Science, 1996: 339–47.
21. Bliss DZ, Johnson S, Savik K, Clabots CR, Willard K,
Gerding DN. Acquisition of Clostridium difficile and
Clostridium difficile-associated diarrhoea in hospitalized
patients receiving tube feeding. Ann Int Med 1998;
129: 1012–19.
22. Woolfson AMJ, Saour JN, Ricketts CR, Pollard BJ,
Hardy SM, Allison SP. Prolonged nasogastric tube feeding in criti-
cally ill and surgical patients. Postgrad Med J 1976; 52: 678–82.
Feeding methods for older tube-fed patients
23. Heimburger DC, Sockwell DG, Geels WJ. Diarrhoea with
enteral feeding: Prospective reappraisal of putative causes.
Nutrition 1994; 10: 392–6.
24. Weinsier RL, Heimgurger DC, Butterworth CE Jr.
Handbook of Clinical Nutrition: Clinician’s Manual for the
Prevention, Diagnosis, and Management of Nutritional
Problems. 2nd Edition. St Louis: The CV Mosby Company,
1989: 209–20.
Received 12 March 2002; accepted in revised form 27 January
2003
Downloaded from http://ageing.oxfordjournals.org by on May 15, 2010
393