MARC J. M. BONTEN, CARLO A. GAILLARD, RENE van der HULST, PETER W. ... Correspondence and requests for reprints should be addressed to Marc J. M..
Intermittent Enteral Feeding: The Influence on Respiratory and Digestive Trad Colonization in Mechanically Ventilated Intensive-Care-Unit Patients MARC J. M. BONTEN, CARLO A. GAILLARD, RENE van der HULST, PETER W. de LEEUW, SIEBE van der GEEST, ELLEN E. STOBBERINGH, PETER B. SOETERS Departments of Internal Medicine, Surgery, and Medical Microbiology, University Hospital Maastricht, Maastricht, the Netherlands; and Department of Internal Medicine, Eemland Hospital, Amersfoort, the Netherlands
Continuous enteral feeding (CEF) has been associated with decreased gastric acidity, thereby stimulating gastric colonization and ventilator-associated pneumonia (VAP). Intermittent enteral feeding (IEF) could induce a temporary increase in gastric acidity and decrease the risk of VAP. We studied the influence of IEF (18 hid) and CEF (24 hid) on gastric and oropharyngeal colonization. Sixty patients were randomized to receive either IEF or CEF, and continuous intragastric pH monitoring was performed in 50 patients. Median intragastric pH levels were similar before enteral feeding was instituted (pH 2.5 for CEF and pH 2.4 for IEF), and median pH values increased slightly after institution of nutrition (NS). In patients receiving IEF, median pH decreased from 3.5 to 2.2 (p = 0.0002) when enteral feeding was discontinued. However, despite this, 80% of the patients in both study groups were colonized in the stomach after 7 d in study. In addition, colonization rates of the oropharynx and trachea, the incidence of YAP, and mortality were similar in both study groups. IEF was less well tolerated than CEF.We conclude that almost all patients receiving enteral feeding are colonized in the stomach with gram-negative bacteria. IEF resulted in a slight decrease in intragastric pH without influencing rates of colonization and infection of the respiratory tract. Bonten MJM, Gaillard CA, van der Hulst R, de Leeuw PW, van der Geest 5, 5tobberlngh EE, Soeters PB. Intermittent enteral feeding: the Influence of respiratory and digestive tract colonization In mechanically ventilated Intensivecare-unit patients. AM J RESPIR CRIT CARE MED 1996;154:J94-9.
Among mechanically ventilated intensive-care-unit (lCU) patients, ventilator-associated pneumonia (VAP) is the most frequently observed nosocomial infection (I). The gastropulmonary route of colonization has been assumed to be instrumental in the pathogenesis of this infection (2). In this hypothesis, bacterial colonization of the stomach plays a pivotal role. Under physiologic circumstances, gastric acidity prevents bacterial growth in the stomach (3). During enteral feeding, dilutional alkalinazation may lead to gastric bacterial overgrowth (4). Moreover, an increase in gastric volume may stimulate gastroesophageal reflux, colonization of the upper respiratory tract, and eventually infection of the lungs. Therefore, gastric feeding has been associated with an increased risk of YAP (4). Intermittent discontinuation of gastric feeding has been proposed as a measure to induce a temporary increase in intragas-
(Received in originol form May 22, 1995 and in revised form January 3, 1996)
Supported by a grant from the Dutch Institute of Internal Medicine. Theresults of this studywere presentedat the 34th Interscience Conference on Antimicrobial Agents and Chemotherapy, 1994, Orlando, Florida, abstract number J175. Correspondence and requests for reprints should be addressed to Marc J. M. Bonten, Department ofInternal Medicine, University Hospital Maastricht, the Netherlands. Am J Resplr Crit Care Med Vol 154. pp 394-399, 1996
tric acidity, thereby preventing gastric bacterial colonization as much as possible (5). However, the use of intermittent enteral feeding (lEF) has never been subjected to a prospective controlled trial with accurate monitoring of intragastric acidity and bronchoscopic techniques for diagnosing YAP. Therefore, we tested the hypothesis that IEF has a beneficial effect on intragastric acidity and bacterial colonization of the stomach and respiratory tract, in a prospective randomized trial comparing IEF and continuous enteral feeding (CEF).
METHODS Study Population The study was performed in a university hospital in three ICUs (two mixed and one cardiosurgicallCU) between August 1993 and August 1994. All mechanicallyventilated patients admitted to one of thesewards wereeligible for this clinicaltrial wheneverenteral nutrition wasstarted and when a remaining period of mechanical ventilation of at least 3 d was to be expected. Patients werenot eligible if they were less than 15 yr of age, when they wereadmitted for massive gastric hemorrhage or bleeding esophageal varices, after esophageal or gastric surgery, when enteral nutrition was administered via gastrostomy, or when Hj-antagonists or H+K+-ATPase inhibitors were indicated. Study Design Randomization betweenCEF and IEF was performed with sealedenvelopes. Demographic characteristics, diagnosis on admission, underlyingdiseases, vital signs,medications, and laboratory valueswere recorded on admission. The acute physiology and chronic health evaluation
Bonten, Gaillard, van der Hulst, et 0/.: Enteral Feeding in Mechanically Ventilated (APACHE II) scoring system was used to assess the severity of illness (6). Clinical and laboratory values were recorded daily. Samples of gastric and oropharyngeal swabs were collected on admission and subsequently on a daily basis at 8:00 A.M. After obtaining gastric aspirates, the first 15 ml were discarded. Tracheal aspirates were collected on admission and subsequently twice weekly. Additional tracheal aspirates were analyzed when required by the attending physician. Patients were studied for a maximum period of 14 d, or until extubation or death. In general, patients were in the supine position during controlled mechanical ventilation and, if possible, in the semirecumbent position during weaning. There were no changes in body position during the course of the day, and promotility agents were not used. The protocol was approved by the institutional review board of the hospital, and informed consent was obtained from all participants, or, if this was not possible, from a representative of the family. Enteral Feeding Enteral feeding (Nutrison; Nutricia, Loehermeer, the Netherlands; pH = 6.0) was administered intragastrically. The formula was delivered in sterile containers containing volumes sufficient for 24 h. Infusion of enteral feeding via the nasogastric tube was performed by means of an electric infusion pump system (IVAC591;San Diego, CA). CEF was administered at a continuous rate over a 24-h period, whereas IEF was administered at a continuous rate over an 18-h period. In the latter group the administration was discontinued between 2:00 and 8:00 A.M. This time period was chosen for practical reasons. In each patient, enteral feeding was begun at 6:00 P.M. and 500 ml was administered on theTirst day, 1,000 ml on the second day, 1,500 ml on the third day, and 2,000 ml from the fourth day on. The amounts of enteral feeding administered and the residual volumes were monitored daily. In case of gastric retention of > 400 ml per day, enteral feeding was diminished or discontinued. Intragastric pH Monitoring Intragastric pH monitoring was performed continuously as described previously (7). During monitoring the pH was recorded every 4 to 16 s, depending on the total length of the recording. The duration of monitoring and median pH value were calculated from each measurement. In the analyses, results wereexpressed as median values, with the interquartile range (IQR; i.e., 25070 and 75070 percentiles) of all individual median pH values. Three periods were analyzed: (I) before enteral feeding; (2) during enteral feeding; and (3) between 2:00 and 8:00 A.M. To exclude any influence of diurnal changes in intragastric acidity, the values measured between 2:00 and 8:00 A.M. were analyzed for both study groups. Bacteriologic Examination Colonization was analyzed semiquantitatively using the four-quadrant streak method on sheep blood (with and without cefuroxime 25 ug/rnl), CLED, chocolate, sabouraud, and Slanetz and Baerley (SBA) plates. Enteric gram-negative bacterial included Enterobacteriaceae (Escherichia coli, Klebsiella species, Enterobacter species, Proteus species, Morganella morganii, Citrobacter freundiii, and Pseudomonadaceae (Pseudomonas aeruginosa, Xanthomonas maltophila). Definitions
Ventilator-associatedpneumonia. In cases of clinical suspicion of pneumonia, bronchoscopy with a protected specimen brush (PSB) and bronchoalveolar lavage (BAL) were performed as described elsewhere (8). The criteria for the diagnosis of YAP are listed in Table l. Colonization. Colonization was defined as the isolation of microorganisms in the absence of infection from gastric aspirate, oropharyngeal swab, or tracheal aspirate specimens. Colonization on entrance into the study was established by a positive culture obtained on the day on which enteral feeding was started. Colonization was designated as "acquired" if there were positive cultures of the same species of microorganism from at least two consecutive samples without colonization with this species on entrance into the study. Statistical Analysis The sample size-of30 patients per group was calculated to detect a reduction in gastric colonization from an assumed 80070 in the CEF group to an expected incidence of 40070 in the IEF group (a = 0.05 and 13 =
leu Patients
395 TABLE 1
CRITERIA USED FOR THE DIAGNOSIS OF VENTILATOR-ASSOCIATED PNEUMONIA
Criteria '" three positive of the following four: Rectal temperature above 38.0· C or below 35.5· C 8100d leukocytosis (> 10.10 3/mm 3 ) and/or left shift or blood leukopenia « 3.10 3/mm 3) 10 leukocytes per high-power field in gram stain of tracheal aspirate Positive culture from tracheal aspirate and II New, persistent, or progressive infiltrate on chest radiograph and III '" one positive of the following three: Positive quantitative culture of a sample obtained by 8Al (cutoff point> 104 du/ml) or PSB (cutoff '" 10 3 du/ml) Positive blood culture unrelated to another source and obtained within 48 h before and after respiratory sampling Positive pleural fluid culture in the absence of previous pleural instrumentation Definition of abbreviations: BAL = bronchoalveolar lavage; PSB brush. Pneumonia was diagnosed if I, and II. and III were positive.
= protected
specimen
0.1). The percentages of colonization were based on data available at the time the study was designed. Groups were compared by Student's t test and the chi-square test for categorical variables. Median intragastric pH values were compared by Mann-Whitney U test or Wilcoxon's rank test for paired samples. A p value < 0.05 was deemed statistically significant.
RESULTS Study Population
Sixty patients were eligible for this trial, and weredivided equally into both treatment groups, which were similar in terms of demographic characteristics and severity of illness on admission (Table 2). Unfortunately, because the study was performed in three different ICU wards, not all patients received the same agents for stress-ulcer prophylaxis. Twenty patients (seven CEF and 13 IEF) received antacids (15ml six times daily) and 30 patients (23 CEF and 17 IEF) received sucralfate (1 g six times daily) (Table 2). However, demographic characteristics were similar for patients receiving stress-ulcer prophylaxis with antacids or sucralfate, and for all patients receiving either of these agents (data not shown). Intragastric pH Monitoring
Continuous intragastric pH monitoring was performed in 50 patients, 24 receiving CEF and 26 receiving lEE For safety reasons, intragastric pH monitoring was not performed in patients with trauma of the head and neck region (n = 2) or intracerebral bleeding (n = 2), or in patients with coagulation disorders (n = 2). In addition, three patients (or relatives) refused intragastric pH monitoring, and data for one patient were lost due to technical failure. The durations of continuous intragastric pH measurements are listed in Table 2. The median pH values(lQR) in the periods before enteral feeding was started were 2.5 (1.3 to 5.1) and 2.4 (1.2 to 6.6) for patients receiving CEF and IEF, respectively (p = NS). In both study groups the intragastric pH was not different before and after institution of enteral feeding (Figure 1). In patients receiving IEF the median intragastric pH decreased from 3.5 (1.8 to 5.2) to 2.2 (1.3 to 3.9) (p = 0.0002, Wilcoxon's rank test for matched pairs) when feeding was discontinued. Because 17 patients received antacids for stress-bleeding pro-
396
AMERICAN JOURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINE VOL 154
1996
TABLE 2 CHARACTERISTICS OF THE PATIENTS STUDIED Continuous EF Number of patients Male/female Age, median + IQR APACHE II, median + IQR Medical specialty Surgery Trauma Medical Pulmonology Neurologic Cardiac surgery Surgical procedures None Laparotomy Thoracic surgery Neurosurgery Stress-ulcer prophylaxis Antacids Sucralfate Colonization with EGB upon study entry Stomach Oropharynx Trachea Durations of pH-measurement (minutes ± SD) Before enteral feeding During enteral feeding During discontinuation of enteral feeding Duration of ICU-stay before start of EF, median + IQR Duration of hospitalization before admission to ICU, median + IQR
Interm ittent EF
30
30
19/11 65 (55-70) 19 (13-23)*
16/14 68 (61-73) 17 (13-22)t
3
5
4
2
8
4 8 3
6 2 7
8
17 4
18
8
10
1
o
7 23
13 17
6 14 4
10 19 6t
279 ± 349 4,918 ± 1,006
229 ± 318 3,046 ± 1,139 1,236±518 4.5 (3-8) 3 (1-8)
2
4 (2-7)t 2 (1-28)
Definition of abbreviations: IQR = interquartile range; EGB = enteric gram-negative bacteria . • n = 29. t n = 2B.
phylaxis, the results of pH monitoring were analyzed with regard to this potentially confounding variable (Table 3). Before the start of enteral feeding, median pH values tended to be lower in patients receiving sucralfate (p = NS). In contrast, patients receiving sucralfate tended to have higher median pH values both during the administration of enteral feeding and between 2:00 and 8:00 A.M. than did patients receiving antacids (p = NS). Acquired Colonization
On the first day of study, six patients (210/0) receiving CEF and 10 patients (36%) receiving IEF (p = NS) were colonized with enteric gram-negative bacteria in the stomach (Table 2). It should be noted that patients had had a median ICU stay of 4 and 4.5 d, respectively, before enteral feeding was begun. Remarkably, the proportion of colonized patients in both study groups increased gradually to 80% on Day 6 (Figure 2). Among the 16 patients colonized on the first day of study, colonization peristed in 14 patients (five receiving IEF and nine receiving CEF), and nine patients acquired colonization with other gram-negative bacteria (three receiving IEF and six receiving CEF). Moreover, among patients without gastric colonization on study entrance, 18(75%) of 24 receiving IEF and 15 (670/0) of 20 receiving CEF acquired colonization with enteric gram-negative bacteria. Cultures of tracheal aspirates in addition to the study schedule were frequently obtained. The total number of tracheal aspirates (n = 258) analyzed, divided by the total number of patient-days studied (n = 516), was 0.50 ± 0.14 (mean ± SD). For the oropharynx and trachea, the number of patients colonized on admission (Table 2) and the number of patients acquiring colonization (Table 4) were similar in both study groups. Enteral Feeding
Adjustments of the amounts of enteral feedings because of gastric residua, were necessary in two patients receiving CEF and
Intragastric pH
a
Continuous enteral feeding
a4 c
:6
-
~ 3
0
ns
test Wilcoxon-rank
before EF
B"
___
during EF
2.00-6.00 am
Intragastric pH
b
Intermittent enteral feeding 7
6
~B~[J
a4 Ii
'5 ~ 3
0
test Wilcoxon-rank
before EF
during EF
2.00-6.00 am
Figure 7. Median intragastric pH values (with interquartile ranges) in patients with CEF(a) and IEF(b) in the periods before enteral feeding, during enteral feeding, and between 2:00 and 8:00 A.M. (enteral feeding was discontinued in patients receiving IEF).
Bonten, Gaillard, van der Hulst, et 01.: Enteral Feeding in Mechanically Ventilated leu Patients
397
TABLE 3 ANALYSIS OF pH MONITORING IN PATIENTS RECEIVING SUCRALFATE AS COMPARED WITH PATIENTS RECEIVING ANTACIDS FOR STRESS-ULCER PROPHYLAXIS Continuous Enteral Feeding
Before enteral feeding During enteral feeding Enteral feeding discontinued
Intermittent Enteral Feeding
Sucralfate (n = 17)
Antacids (n = 7)
p Value
Sucralfate (n = 16)
Antacids (n = 10)
p Value
1.7 (1.4--5.0) 3.2 (1.7-4.7) 4.0 (1.4-4.8)
3.2 (1.1-5.5) 1.0 (0.8-5.5) 1.0 (0.8-4.5)
0.8 0.2 0.2
1.7 (1.0-7.1) 3.9 (1.7-5.0) 1.9 (1.3-3.9)
3.3 (2.0.-4.4)' 2.7 (1.7-5.4) 2.2 (1.3-4.2)
0.5 0.9 0.9
Values in median (interquartile range [IQR]). * n = 9.
in five patients receiving lEE Moreover, in one of these latter five patients the administration of enteral feeding had to be discontinued completely. As a result, patients in the CEF group received larger volumes of enteral feeding (Figure 3), despite the uniform scheme of administration and increase in volume during the first 4 d, which applied to all of the patients studied. Infections and Mortality Ten patients developed an episode of YAP during the administration of enteral feeding, with a similar incidence among patients receiving IEF and CEF (Table 4). Fifteen patients succumbed during their ICU stay, six of them after the 14-d study period. DISCUSSION
The results of this study strongly suggest that discontinuation of enteral feeding does not influence the incidence of digestive and respiratory tract colonization. Almost all patients receiving enteral feeding acquire gastric colonization with enteric gramnegative bacteria. Moreover, IEF was less welltolerated than CEE We therefore conclude that IEF is not effective as a measure to prevent YAP. Institution of enteral feeding did not substantially decrease intragastric acidity. Our data confirm previous results reported by others and ourselves, also using computerized continuous intragastric pH monitoring (7, 9). Although some investigators reported a decrease in gastric acidity during gastric feeding, the methodology of measuring intragastric acidity applied in these studies can be questioned (10). For instance, gastric pH was in almost all studies measured with color-scaled indicator papers in aspirates of gastric juice (4, 5, 11), a method that has been criticized (12, 13). Moreover, the frequency of these measurements
100 , - - - - - -- - --
-
-
-
-
-
-
-
- - --
-
-
was only once or twice daily (5, 11, 14), which may be too low to adequately detect changes in intragastric acidity (10). Although in our hands gastric feeding had no significant influence on intragastric acidity, with median intragastric pH values < 3.5 in both study groups, 80% of the patients were colonized with enteric gram-negative bacteria in the stomach when enteral feeding had been administered for 7 d. A relation between the institution of gastric nutrition and gastric colonization has been reported (4, 7, 15). Pingleton and coworkers, for instance, described 18mechanically ventilated patients receivingenteral feeding, all of whom were colonized in the stomach at some time during the study (4). In a previous study we also demonstrated that enteral feeding was associated with gastric colonization, and that enteral feeding counteracted the preventive effects of topical antimicrobial prophylaxis (15). How do we account for the high gastric colonization rates in the present study when these cannot be explained by decreased intragastric acidity? Some speculations can be made about the stimulating effects of enteral feeding on gastric colonization. For instance, the intragastric pH may not be the same in each part of the stomach. On the other hand, enteral feeding may provide the bacteria with extra nutrients, allowing them to survive in a surrounding in which under normal circumstances growth would be inhibited. However, the potential relationship between enteral feeding and gastric colonization should be established in studies that more specifically address this issue. Discontinuation of enteral feeding resulted in a slight increase in intragastric acidity, which could not be explained by physiologic diurnal changes. However, the temporary increase in acidity did not influence colonization and infection rates, and intermittent enteral feeding was less well tolerated. Therefore, we do not recommend its use as a preventive measure for YAP. Two previous studies of the effects of IEF on intragastric acidity, gastric
--, 3000 , - - - - - - - - - - - --
- - - --
-
-
-
-
-
--,
80
/
