GASTROENTEROLOGY 1992;103:1523-1528
Manometry Can Predict Feeding Readiness in Preterm Infants CAROL
LYNN
BERSETH
and CYNTHIA
K. NORDYKE
Department of Pediatrics and Gastroenterology Research Unit, Mayo Clinic, Rochester, Minnesota
To assess the usefulness of duodenojejunal manometry in predicting when neonates will tolerate enteral feeding, small intestinal manometry was performed in 48 preterm infants. Characteristics of motor activity during fasting and feeding differed in 40 infants who were identified to be tolerant and 8 who were intolerant of feeding. Among infants who were feeding intolerant, motor quiescence was less pronounced and clustered motor activity more prominent than it was in infants who were able to tolerate feedings (P c 0.005). Those infants who tolerated feedings changed their pattern of motor activity in response to feeding (P < O.Ol), and infants who did not tolerate feedings had no “fed response. ” Infants who were initially feeding intolerant became tolerant of feedings coincident with the appearance of motor activity that was similar to that of infants who were initially tolerant to food. The sensitivity of manometry to predict feeding intolerance was 1.0, and its specificity was 0.13. he provision of adequate nutrition to the very preterm infant (4500 g birthweight) is a major challenge to the neonatologist. Although feeding practices vary widely among neonatal intensive care units,’ several recent reports have documented the efficacy of providing early enteral nutrition to preterm infants recovering from respiratory disease.‘+ However, preterm infants who receive enteral feedings are thought to be at risk for complications including necrotizing enterocolitis (NEC), aspiration, and/or intermittent abdominal distention. A method to assess functional maturation of the preterm intestine might help to determine the optimal time for feeding and thereby improve the clinical management of enteral nutrition. Until recently, no method has been available to directly assess the maturational function or responsiveness of the preterm intestine. We recently developed a low-compliance continuous-infusion manometry system to characterize motor activity in infants.s-B Motor activity is immature in preterm infants; complete interdigestive cycles are absent, and clustered phasic motor activity is
T
prominent.6*7 The characteristic clustered activity changes with gestational and postnatal age. Cluster activity decreases in duration and increases in amplitude with age. At 34-36 weeks’ gestation, the complete interdigestive cycle appears, but clustered phasic activity still represents three fourths of all motor activity.G-8 Because intestinal motor activity changes with age, we postulated that manometry might diagnostically predict whether enteral feeding would be tolerated by the preterm intestine. The feeding experience of infants who had been studied manometrically was followed up prospectively to identify those motor activity characteristics associated with feeding intolerance. Material and Methods Subjects All 48 infants were admitted to the Newborn Intensive Care Unit (NICU) for respiratory distress syndrome. They ranged in gestational age from 26 to 34 weeks, with birth weights of 780-2115 g. According to the NICU routine, no infant received enteral nutrition the first several days of life when ventilator support was being provided; rather, all infants received parenteral nutrition. Intestinal manometry was recorded by the fifth day after infants had been extubated or, if they still required intubation, when they required fewer than three ventilator changes per day. Approximately one third of the infants studied required ventilator support when they were studied. The studies were performed on the day that the attending neonatologist had planned to begin enteral feedings by orogastric or transpyloric tube. This protocol was approved by the Mayo Institutional Review Board, and all infants’ parents provided informed consent for entry into this study.
Study Design No infant received enteral feedings before entry into the study; rather, these infants received parenteral nutrition according to the NICU routine. None of the infants had clinical or radiographic evidence of NEC or in0 1992 by the American Gastroenterological Association 0016-5065/92/$3.00
1524
BERSETH AND NORDYKE
testinal abnormality at study entry. The feeding/motility tube was placed into the proximal duodenum, and small intestinal manometry was recorded for 4 hours fasting. Manometry was then recorded for an additional 4 hours, 2 hours during a continuous infusion of 4 mL/kg Similac formula (Sim 20; Ross, Columbus, OH) and 2 hours after completion of the feeding infusion. The attending neonatologist was not informed of the results of manometry, and all infants received enteral nutrition. All infants were fed Sim 20 in &hour cycles consisting of a 2-hour infusion by orogastric or transpyloric tube followed by 2 hours of fasting. Feedings were begun at 4 mL - kg-’ - 2 h-‘. The amount of formula was increased daily by 1-2 mL/2 h by the attending neonatologist until an intake of 120 cal - kg-’ - day-’ was achieved. To record data concerning feeding intolerance, the charts of all study infants were reviewed weekly by a technician who was not involved in patient care. Decisions concerning feeding were made prospectively by the attending neonatologist, who discontinued feedings if intolerance developed. Feeding intolerance was defined as vomiting, an increase in abdominal girth by more than 1.5 cm, or milk retained in the stomach 2 hours after the feeding infusion had been completed. For a single episode of vomiting, feedings were withheld 4-a hours. For increased abdominal girth, feedings were withheld until the abdominal girth decreased to baseline fasting girth. For milk retained in the stomach, feedings were withheld for an additional 2 hours. If no milk remained in the stomach at the end of the additional 2 hours, the next feeding was given. If milk was still present at the end of this additional 2 hours, feedings were withheld, and the presence of milk was reassessed every 2 hours. If milk was still present after an additional 4 hours, feedings were withheld for 24 hours. Manometry Motor activity was recorded using a neonatal lowcompliance continuous-perfusion unit. This unit has been previously described and validated.” It provides an infusion rate of 0.01 mL/min per recording port at 10 psi with a response rate of 57 mm Hg/s.’ In turn, this system infused a five-port manometric/feeding tube. This tube, manufactured in the Mayo Gastrointestinal Research Unit, was made from polyvinyl extrusion tubing with an outer diameter of 3.5 mm. The distal port was used for feeding, and the four proximal ports, spaced 2.5 cm apart, were used to record motor activity. For tube placement, the infant was positioned supine with the right side down. The manometric tube was perfused, and motor activity was recorded while the tube was advanced through the stomach into the duodenum. Motor activities characteristic of the antrum and intestine were noted to appear as each perfusion port passed distally from the stomach into the duodenum. All four ports were positioned in the duodenum, and the final position of the tube was documented by the presence of a motility pattern characteristic of the intestine. Our previous studies have shown that when these activities are localized at these two anatomic sites, the position of record-
GASTROENTEROLOGY
Vol. 103, No. 5
ing ports can be verified by fluoroscopy.“-a Motor activity was recorded for 4 hours of fasting. The infant was then given Similac 20 at 4 mL/kg for a 2-hour infusion through the distal feeding port. Motor activity was recorded throughout the &hour feeding and for 2 hours after the completion of the feeding infusion.
Data Analysis All tracings were analyzed in SO-minute segments. During fasting, three types of motor activity were identified as described previously. 68 Phase I was defined as motor quiescence and phase III as a burst of rhythmic phasic pressure waves with a duration longer than 2 minutes that migrated over three or more ports. This latter activity was defined as a migrating motor complex (MMC). Phase II was defined as irregular activity or phasic activity longer than 0.4 minutes in duration that did not migrate over three ports. The latter motor pattern was defined as a cluster. The duration of each type of activity was determined in minutes per hour of recording per recording lead. The occurrence of each type of activity was calculated as number of episodes per hour of recording per lead. The mean duration of each episode of phase I, II, or III was determined in minutes. The periodicity of phase III was determined in minutes and the migration velocity in centimeters per minute. Activity during feeding was characterized in a manner similar to that of adults as described previously.’ Motility index was calculated as log, (sum of pressure peaks X amplitude + l).’ Also determined were the number of pressure peaks in 30 minutes, the mean amplitude of pressure peaks in millimeters of mercury, and the duration of quiescence in minutes per hour of recording per lead. Feeding-intolerant babies were identified by an observer who was not involved in clinical care of the infants (C.K.N.). Feeding-intolerant babies were defined as those in whom feedings were withheld within the first week after manometry on three separate occasions or for more duration. Eight preterm inthan 48 hours of cumulative fants were defined to be feeding intolerant; three of these received a diagnosis of NEC by radiographic and clinical
Table 1. Personal Characteristics and Clinical Course of Preterm Infants Who Are Tolerant and Intolerant of Enteral Feedings Preterm infants
Tolerant
n Gestational age at study (wk) Birth weight (g) Postnatal age [days) Length of feeding intolerance (days) Time to establish oral feedings (days) Hospitalization (days)
40
a
30.4 f 0.2 1376 + 46 6.4 + 0.6
31.0 * 0.7 1145 + 137 a.7 + 1.2
“P < 0.01.
1.0 5
0.5
37 f 3 40 + 3
Intolerant
13.0 & 3.6” 56 ? 6” 61 + 5a
MANOMETRY PREDICTS FEEDING READINESS 1525
November 1992
Proximal duodenum
Mid duodenum
125 mm Hg Mid duodenum
-
Distal duodenum 1 min
Figure 1. Duodenal motor activity during fasting in a preterm infant who tolerated feedings. The tracing from the most proximal port is shown at the top of the figure and the most distal at the bottom. Note the periods of quiescence that are sustained for 3 minutes or more in the three more proximal leads. Also present are clusters of phasic contractions in the three more distal leads that range from 3 to 10 minutes in duration.
findings. Personal characteristics of the infants at the time of study entry did not differ (Table 1). A similar proportion of infants were ventilator dependent at the time of the manometric study in each group. Characteristics of motor activity were compared between babies who tolerated feedings and those who did not tolerate feedings by unpaired Student’s t test; motor activity during feeding and fasting was compared by paired Student’s t test.
Results Manometrics By visual inspection, babies who tolerated feedings had prolonged periods of quiescence that alternated with migrating activity or clusters (Figure 1). With feeding, quiescence became less pronounced, although still present, and the duration of clustered contractions became longer (Figure 2). In contrast, feeding-intolerant infants displayed more motor activity; quiescence occupied a minimal por-
tion of the recording (Figure 3). In response to feeding, motor activity did not change; episodes of quiescence were rare, and pressure waves were frequent and prominent during feeding (Figure 4). Quiescence occupied a significantly greater proportion of total recording time in babies who tolerated feedings than it did in feeding-intolerant infants (Table 2; P < 0.05). Conversely, feeding-tolerant babies had significantly less cluster activity than did babies who did not tolerate feedings, although they had more irregular activity (Table 2; P < 0.05 or less). Characteristics of clusters and quiescence also differed between these two groups of infants. Quiescence duration was 25% greater in feeding-tolerant compared with feeding-intolerant infants, and the mean duration of a single episode of quiescence was longer in feeding-tolerant than it was in feeding-intolerant infants (Table 3; P < 0.003). Cluster duration in minutes per hour per lead was almost 50% greater
Fed Proximal duodenum ”
I...
Mid duodenum
Mid duodenum
]25mmHg
Distal duodenum I,.,
1 min
Figure 2. Duodenal motor activity in the same preterm infant as in Figure 1 when the milk infusion was begun. Orientation is similar to that in Figure 1. Note that quiescent periods are still present, but the individual duration of episodes is approximately 0.5 minutes, and contractures occupy most of the recording.
1526
BERSETHAND NORDYKE
GASTROENTEROLOGYVol.103,No.5
Proximal duodenum
Mid duodenum
125 mm Hg
Mid duodenum
Distal duodenum
Figure 3. Duodenal motor activity during fasting in a preterm infant who did not tolerate feedings. Orientation is the same as in Figure 1. Quiescent episodes are approximately 0.5 minutes long, and contractures occupy most of the recording.
in feeding-intolerant infants, and the mean duration of an individual cluster was shorter than that in feeding-tolerant infants (Table 3;P < 0.003). MMC periodicity and velocity were similar in both groups. Feeding-tolerant infants showed a change in motor activity in response to feeding, but feeding-intolerant infants did not. In infants who tolerated the feedings, motility index and the number of pressure peaks in 30 minutes increased significantly, and quiescence decreased during feeding to fasting (Table 4; P < 0.007 or less). In infants who were intolerant of feedings, none of these three features differed between feeding and fasting. The lack of a fed response in preterm infants who were intolerant of feedings appeared to be the result of differences in activity during fasting. Motility index and the number of peaks in 30 minutes was greater in feeding-intolerant than feeding-tolerant infants during fasting (Table 4; both P < 0.05), but they were similar in the two groups during feeding. Similarly, the duration of motor quiescence during fasting was lower in feedingintolerant than feeding-tolerant infants during fast-
ing (Table 4; P < O.O5),but values were similar in the two groups during feeding. Long-Term
Outcome
All eight feeding-intolerant infants were evaluated for the presence of an underlying metabolic or anatomic cause for their intolerance; none was found. However, the differences in motor activity in preterm infants was reflected in the clinical outcomes. Babies who were feeding intolerant had significantly more days of feeding difficulty; these clinical problems resulted in a longer time to achieve full oral nutrition and discharge from the hospital (Table 1;all P < 0.01). Manometry was recorded again approximately 2 weeks later in all eight feeding-intolerant preterm infants. In six infants, cluster duration (18.6 f 1.6min/h per lead) and quiescence duration (25.6 * 2.7 min/h per lead) were similar to those for feeding-tolerant infants. Three of these infants showed a change in motor activity in response to feeding. Five of these infants had no further feeding problems during the second poststudy week; feed-
Proximal duodenum
Mid duodenum
125 mm Hg Mid duodenum
Distal duodenum
Figure 4. Duodenal motor activity in the same infant in Figure 3 approximately 60 minutes after the milk infusion began. Note that there is little difference in duration and occurrence of quiescence or contractures compared with that seen in Figure 3.
November
MANOMETRY
1992
Table 2. Motor Activity in Preterm Infants During Fasting Preterm infants Quiescence (% of recording time) Phase II Irregular (% of recording time) Clusters (% of recording time) MMCs (% of recording time)
Tolerant (n = 40)
Intolerant (n = 8)
43.4 * 2.3
27.0 f 2.9O
13.5 + 1.0 40.7 + 2.2 3.1 f 0.7
8.5 + 1.2” 58.6 + 3.2b 6.6 f 2.3
“P < 0.05. hP < 0.005.
ings were withheld for one infant on two single occasions. Two of the preterm infants who were restudied continued to show prolonged cluster duration (50.9 and 47.4 min/h per lead) and less prominent quiescence (6.4 and 7.4 min/h per lead) compared with feeding-tolerant infants. Neither of these two infants showed a change in motor activity in response to feeding. Feeding problems persisted in these two infants. Manometry was repeated in both of these infants 2 weeks later. At that time, the cluster duration had decreased to 13.4 and 14.8 min/h per lead, and the quiescence increased to 28.7 and 21.4 min/h per lead, similar to characteristics in feeding-tolerant infants. One of the two infants showed a change in motor activity in response to feeding. Neither baby showed feeding intolerance after this final study. Sensitivity
and Specificity
We assessed the sensitivity and specificity of manometrics in identifying infants at risk for feeding intolerance. All manometric records were reviewed and scored using two criteria. During fasting, the duration of motor quiescence in minutes per hour per lead was determined. Based on normative data for preterm infants of 28-32 weeks’ gestationional age,7 a value of ~20 min/h per lead (mean - 3 SD) was scored as abnormal. Abnormal feeding responses
Table 3. Characteristics of Quiescence, Clusters, and MMCs in Fasting Preterm Infants Preterm infants Quiescence Duration (min/h per lead) Duration of episode (min) Clusters Duration (min/h per lead) Duration of episode (min) MMCs Periodicity (min) Velocity (cm/min) “P < 0.005. bP < 0.05.
Tolerant
Intolerant
24.5 If: 1.2
15.6 + 1.7”
1.1 f 0.1
0.7 + O.lb
23.2 + 1.3 1.8 f 0.1
33.8 + 2.0a 2.6 + 0.3b
63 f 9 1.6 f 0.4
65 + 27 2.2 f 0.7
PREDICTS FEEDING READINESS
Table 4. Motor Activity
Response to Feeding
1527
in Preterm
Infants
Preterm
infants
Tolerant Motility index No. of peaks per 30 min Quiescence duration Intolerant Motility index No. of peaks per 30 min Quiescence duration “P < 0.05, unpaired
Fasting
Fed
P value (fasting vs. fed)
12.5 + 0.1
13.3 + 0.1
0.0001
165 f 9 24.5 f 1.3
235 f 9 19.5 f 1.7
0.0001 0.007
13.2 f 0.2”
13.0 + 0.3
NS
225 +_12” 15.6 + 1.7”
232 + 26 18.4 & 4.2
NS NS
Student’s t test, tolerant vs. intolerant.
were determined to be present if the motility index did not change by more than 10% of that seen during fasting. Records were defined to predict risk for feeding intolerance if both criteria were abnormal. Using these criteria, manometrics identified all 8 feedingintolerant infants correctly; additionally, 6 of the 39 feeding-tolerant infants were identified to be at risk for intolerance. Thus, the sensitivity of manometry recordings to predict feeding intolerance was 1.0, and specificity was 0.13. Discussion Feeding intolerance results in significant mortality and morbidity in preterm infants. However, no diagnostic test is currently available to identify preterm infants who are at risk for feeding intolerance. Manometry has been used extensively to evaluate adult disorders of gastrointestinal motility. Manometric abnormalities in such patients include an absence of MMCs, exaggerated cluster activity, absence of a “feeding pattern,” and low-amplitude contractions.5 These same motor activity patterns have been described in children with myopathic and neuropathic diseases.g-” However, preterm and term infants also show motor activity that is “abnormal” compared with that of adults and older children in that cluster activity is prominent, MMCs are often absent, and a “fed response” is inconsistently present. Hence, there was a need to control for these normal maturational characteristics of intestinal motor activity by studying age-matched feeding-tolerant infants. In this prospective study, preterm infants with feeding intolerance showed less quiescence and more clustered motor activity during fasting than age-matched infants who tolerated feedings. They also showed no change in motor activity in response to feeding. Thus, the abnormalities shown in the feeding-intolerant infants were similar
1528
BERSETH AND NORDYKE
GASTROENTEROLOGY
to those found in adults with impaired neuronal modulation of motility,5 even when age-related characteristics of motor activity were considered. Manometry provided excellent sensitivity in identifying preterm infants who developed feeding intolerance. Although Morriss et al.” have shown that duodenal contraction rate and amplitude are similar in infants who develop NEC and control infants, they did not evaluate the presence of motor quiescence or migrating activity, nor did they evaluate responses to feeding. Because these two aspects of motor function depend on neuronal inhibition and/or integration, our more extensive profile of motor activity may be necessary to identify babies at risk for feeding intolerance. In view of the high incidence of NEC in preterm infants, a specificity of 0.13 may be acceptable to neonatologists because it permits error in a manner that is less harmful to infants. Manometric “abnormalities” in feeding-intolerant infants were transient, and they resolved over 4 weeks concomitant with improvement of feeding tolerance. Hence, serial manometry may be useful to track and predict feeding readiness in preterm infants and to reduce feeding-related morbidity. Such information should provide valuable insight into the introduction and manipulation of enteral feeding regimens in preterm infants and an ability to test methods to enhance or potentiate the normal maturational processes of gut function in preterm infants. References I. Churella HR, Bachhuber WL, MacLean WC Jr. Survey: methods of feeding 76:243-249.
low birth
weight
infants.
Pediatrics
1985;
Vol. 103, No. 5
2. Brosivs KK, Ritter DA, Kenny JD. Postnatal growth curve of the infant with extremely low birth weight who was fed enterally. Pediatrics 1984;74:778-782. 3. Moyer-Mileur L, Chan GM. Nutritional support of very-lowbirth-weight infants requiring prolonged assisted ventilation. Am J Dis Child 1986;140:929-932. 4. Berseth CL. Early feeding enhances maturation of the preterm small intestine. J Pediatr 1992;120:947-953, 5. Malagelada J-R, Camilleri M, Stanghellini V. Manometric diagnosis of gastrointestinal motility disorders. New York: Thieme, 1986. 6. Amarnath RP, Berseth CL, Malagelada J-R, et al. Postnatal maturation of small intestinal motility in preterm and term infants. J Gastrointest Motil 1989;1:138-143. 7. Berseth CL. Gestational evolution of small intestinal motility in preterm and term infants. J Pediatr 1989;115:646-651. 8. Berseth CL. Neonatal small intestinal motility: the motor responses to feeding in term and preterm infants. J Pediatr 1990;117:777-782. 9. Hyman PE, Napolitano JA, Diego A, Pate1 S, Flores EF, Grill BB, Reddy SN, Garvey TQ, Tomomosa T. Antroduodenal manometry in the evaluation of chronic functional gastrointestinal symptoms. Pediatrics 1990;86:39-44. 10. Anuras S, Mitros FA, Soper RT, Pringle KC, Maves BV, Younoszai MK, Franken, Jr., EA, Whitington P. Chronic intestinal pseudo-obstruction in young children. Gastroenterology 1986;91:62-70. 11. Hyman PE, McDiarmid SV, Napolitano J, et al. Anteroduodenal motility in children with chronic intestinal pseudo-obstruction. J Pediatr 1988;112:899-905. 12. Morriss FH, Moore M, Gibson T, West MS. Motility of the small intestine in preterm infants who later have necrotizing enterocolitis. J Pediatr 1990;117:520-523. Received December 23, 1991. Accepted May 28, 1992. Address requests for reprints to: Carol Lynn Berseth, M.D., Pediatrics Newborn Section, Baylor College of Medicine, One Baylor Plaza, Houston, Texas 77030. Supported by National Institutes of Health grant HD 24558 and in part by the North American Pediatric Pseudo-obstruction Society.
Manometry Can Predict Feeding Readiness in Preterm Infants CAROL
LYNN
BERSETH
and CYNTHIA
K. NORDYKE
Department of Pediatrics and Gastroenterology Research Unit, Mayo Clinic, Rochester, Minnesota
To assess the usefulness of duodenojejunal manometry in predicting when neonates will tolerate enteral feeding, small intestinal manometry was performed in 48 preterm infants. Characteristics of motor activity during fasting and feeding differed in 40 infants who were identified to be tolerant and 8 who were intolerant of feeding. Among infants who were feeding intolerant, motor quiescence was less pronounced and clustered motor activity more prominent than it was in infants who were able to tolerate feedings (P c 0.005). Those infants who tolerated feedings changed their pattern of motor activity in response to feeding (P < O.Ol), and infants who did not tolerate feedings had no “fed response. ” Infants who were initially feeding intolerant became tolerant of feedings coincident with the appearance of motor activity that was similar to that of infants who were initially tolerant to food. The sensitivity of manometry to predict feeding intolerance was 1.0, and its specificity was 0.13. he provision of adequate nutrition to the very preterm infant (4500 g birthweight) is a major challenge to the neonatologist. Although feeding practices vary widely among neonatal intensive care units,’ several recent reports have documented the efficacy of providing early enteral nutrition to preterm infants recovering from respiratory disease.‘+ However, preterm infants who receive enteral feedings are thought to be at risk for complications including necrotizing enterocolitis (NEC), aspiration, and/or intermittent abdominal distention. A method to assess functional maturation of the preterm intestine might help to determine the optimal time for feeding and thereby improve the clinical management of enteral nutrition. Until recently, no method has been available to directly assess the maturational function or responsiveness of the preterm intestine. We recently developed a low-compliance continuous-infusion manometry system to characterize motor activity in infants.s-B Motor activity is immature in preterm infants; complete interdigestive cycles are absent, and clustered phasic motor activity is
T
prominent.6*7 The characteristic clustered activity changes with gestational and postnatal age. Cluster activity decreases in duration and increases in amplitude with age. At 34-36 weeks’ gestation, the complete interdigestive cycle appears, but clustered phasic activity still represents three fourths of all motor activity.G-8 Because intestinal motor activity changes with age, we postulated that manometry might diagnostically predict whether enteral feeding would be tolerated by the preterm intestine. The feeding experience of infants who had been studied manometrically was followed up prospectively to identify those motor activity characteristics associated with feeding intolerance. Material and Methods Subjects All 48 infants were admitted to the Newborn Intensive Care Unit (NICU) for respiratory distress syndrome. They ranged in gestational age from 26 to 34 weeks, with birth weights of 780-2115 g. According to the NICU routine, no infant received enteral nutrition the first several days of life when ventilator support was being provided; rather, all infants received parenteral nutrition. Intestinal manometry was recorded by the fifth day after infants had been extubated or, if they still required intubation, when they required fewer than three ventilator changes per day. Approximately one third of the infants studied required ventilator support when they were studied. The studies were performed on the day that the attending neonatologist had planned to begin enteral feedings by orogastric or transpyloric tube. This protocol was approved by the Mayo Institutional Review Board, and all infants’ parents provided informed consent for entry into this study.
Study Design No infant received enteral feedings before entry into the study; rather, these infants received parenteral nutrition according to the NICU routine. None of the infants had clinical or radiographic evidence of NEC or in0 1992 by the American Gastroenterological Association 0016-5065/92/$3.00
1524
BERSETH AND NORDYKE
testinal abnormality at study entry. The feeding/motility tube was placed into the proximal duodenum, and small intestinal manometry was recorded for 4 hours fasting. Manometry was then recorded for an additional 4 hours, 2 hours during a continuous infusion of 4 mL/kg Similac formula (Sim 20; Ross, Columbus, OH) and 2 hours after completion of the feeding infusion. The attending neonatologist was not informed of the results of manometry, and all infants received enteral nutrition. All infants were fed Sim 20 in &hour cycles consisting of a 2-hour infusion by orogastric or transpyloric tube followed by 2 hours of fasting. Feedings were begun at 4 mL - kg-’ - 2 h-‘. The amount of formula was increased daily by 1-2 mL/2 h by the attending neonatologist until an intake of 120 cal - kg-’ - day-’ was achieved. To record data concerning feeding intolerance, the charts of all study infants were reviewed weekly by a technician who was not involved in patient care. Decisions concerning feeding were made prospectively by the attending neonatologist, who discontinued feedings if intolerance developed. Feeding intolerance was defined as vomiting, an increase in abdominal girth by more than 1.5 cm, or milk retained in the stomach 2 hours after the feeding infusion had been completed. For a single episode of vomiting, feedings were withheld 4-a hours. For increased abdominal girth, feedings were withheld until the abdominal girth decreased to baseline fasting girth. For milk retained in the stomach, feedings were withheld for an additional 2 hours. If no milk remained in the stomach at the end of the additional 2 hours, the next feeding was given. If milk was still present at the end of this additional 2 hours, feedings were withheld, and the presence of milk was reassessed every 2 hours. If milk was still present after an additional 4 hours, feedings were withheld for 24 hours. Manometry Motor activity was recorded using a neonatal lowcompliance continuous-perfusion unit. This unit has been previously described and validated.” It provides an infusion rate of 0.01 mL/min per recording port at 10 psi with a response rate of 57 mm Hg/s.’ In turn, this system infused a five-port manometric/feeding tube. This tube, manufactured in the Mayo Gastrointestinal Research Unit, was made from polyvinyl extrusion tubing with an outer diameter of 3.5 mm. The distal port was used for feeding, and the four proximal ports, spaced 2.5 cm apart, were used to record motor activity. For tube placement, the infant was positioned supine with the right side down. The manometric tube was perfused, and motor activity was recorded while the tube was advanced through the stomach into the duodenum. Motor activities characteristic of the antrum and intestine were noted to appear as each perfusion port passed distally from the stomach into the duodenum. All four ports were positioned in the duodenum, and the final position of the tube was documented by the presence of a motility pattern characteristic of the intestine. Our previous studies have shown that when these activities are localized at these two anatomic sites, the position of record-
GASTROENTEROLOGY
Vol. 103, No. 5
ing ports can be verified by fluoroscopy.“-a Motor activity was recorded for 4 hours of fasting. The infant was then given Similac 20 at 4 mL/kg for a 2-hour infusion through the distal feeding port. Motor activity was recorded throughout the &hour feeding and for 2 hours after the completion of the feeding infusion.
Data Analysis All tracings were analyzed in SO-minute segments. During fasting, three types of motor activity were identified as described previously. 68 Phase I was defined as motor quiescence and phase III as a burst of rhythmic phasic pressure waves with a duration longer than 2 minutes that migrated over three or more ports. This latter activity was defined as a migrating motor complex (MMC). Phase II was defined as irregular activity or phasic activity longer than 0.4 minutes in duration that did not migrate over three ports. The latter motor pattern was defined as a cluster. The duration of each type of activity was determined in minutes per hour of recording per recording lead. The occurrence of each type of activity was calculated as number of episodes per hour of recording per lead. The mean duration of each episode of phase I, II, or III was determined in minutes. The periodicity of phase III was determined in minutes and the migration velocity in centimeters per minute. Activity during feeding was characterized in a manner similar to that of adults as described previously.’ Motility index was calculated as log, (sum of pressure peaks X amplitude + l).’ Also determined were the number of pressure peaks in 30 minutes, the mean amplitude of pressure peaks in millimeters of mercury, and the duration of quiescence in minutes per hour of recording per lead. Feeding-intolerant babies were identified by an observer who was not involved in clinical care of the infants (C.K.N.). Feeding-intolerant babies were defined as those in whom feedings were withheld within the first week after manometry on three separate occasions or for more duration. Eight preterm inthan 48 hours of cumulative fants were defined to be feeding intolerant; three of these received a diagnosis of NEC by radiographic and clinical
Table 1. Personal Characteristics and Clinical Course of Preterm Infants Who Are Tolerant and Intolerant of Enteral Feedings Preterm infants
Tolerant
n Gestational age at study (wk) Birth weight (g) Postnatal age [days) Length of feeding intolerance (days) Time to establish oral feedings (days) Hospitalization (days)
40
a
30.4 f 0.2 1376 + 46 6.4 + 0.6
31.0 * 0.7 1145 + 137 a.7 + 1.2
“P < 0.01.
1.0 5
0.5
37 f 3 40 + 3
Intolerant
13.0 & 3.6” 56 ? 6” 61 + 5a
MANOMETRY PREDICTS FEEDING READINESS 1525
November 1992
Proximal duodenum
Mid duodenum
125 mm Hg Mid duodenum
-
Distal duodenum 1 min
Figure 1. Duodenal motor activity during fasting in a preterm infant who tolerated feedings. The tracing from the most proximal port is shown at the top of the figure and the most distal at the bottom. Note the periods of quiescence that are sustained for 3 minutes or more in the three more proximal leads. Also present are clusters of phasic contractions in the three more distal leads that range from 3 to 10 minutes in duration.
findings. Personal characteristics of the infants at the time of study entry did not differ (Table 1). A similar proportion of infants were ventilator dependent at the time of the manometric study in each group. Characteristics of motor activity were compared between babies who tolerated feedings and those who did not tolerate feedings by unpaired Student’s t test; motor activity during feeding and fasting was compared by paired Student’s t test.
Results Manometrics By visual inspection, babies who tolerated feedings had prolonged periods of quiescence that alternated with migrating activity or clusters (Figure 1). With feeding, quiescence became less pronounced, although still present, and the duration of clustered contractions became longer (Figure 2). In contrast, feeding-intolerant infants displayed more motor activity; quiescence occupied a minimal por-
tion of the recording (Figure 3). In response to feeding, motor activity did not change; episodes of quiescence were rare, and pressure waves were frequent and prominent during feeding (Figure 4). Quiescence occupied a significantly greater proportion of total recording time in babies who tolerated feedings than it did in feeding-intolerant infants (Table 2; P < 0.05). Conversely, feeding-tolerant babies had significantly less cluster activity than did babies who did not tolerate feedings, although they had more irregular activity (Table 2; P < 0.05 or less). Characteristics of clusters and quiescence also differed between these two groups of infants. Quiescence duration was 25% greater in feeding-tolerant compared with feeding-intolerant infants, and the mean duration of a single episode of quiescence was longer in feeding-tolerant than it was in feeding-intolerant infants (Table 3; P < 0.003). Cluster duration in minutes per hour per lead was almost 50% greater
Fed Proximal duodenum ”
I...
Mid duodenum
Mid duodenum
]25mmHg
Distal duodenum I,.,
1 min
Figure 2. Duodenal motor activity in the same preterm infant as in Figure 1 when the milk infusion was begun. Orientation is similar to that in Figure 1. Note that quiescent periods are still present, but the individual duration of episodes is approximately 0.5 minutes, and contractures occupy most of the recording.
1526
BERSETHAND NORDYKE
GASTROENTEROLOGYVol.103,No.5
Proximal duodenum
Mid duodenum
125 mm Hg
Mid duodenum
Distal duodenum
Figure 3. Duodenal motor activity during fasting in a preterm infant who did not tolerate feedings. Orientation is the same as in Figure 1. Quiescent episodes are approximately 0.5 minutes long, and contractures occupy most of the recording.
in feeding-intolerant infants, and the mean duration of an individual cluster was shorter than that in feeding-tolerant infants (Table 3;P < 0.003). MMC periodicity and velocity were similar in both groups. Feeding-tolerant infants showed a change in motor activity in response to feeding, but feeding-intolerant infants did not. In infants who tolerated the feedings, motility index and the number of pressure peaks in 30 minutes increased significantly, and quiescence decreased during feeding to fasting (Table 4; P < 0.007 or less). In infants who were intolerant of feedings, none of these three features differed between feeding and fasting. The lack of a fed response in preterm infants who were intolerant of feedings appeared to be the result of differences in activity during fasting. Motility index and the number of peaks in 30 minutes was greater in feeding-intolerant than feeding-tolerant infants during fasting (Table 4; both P < 0.05), but they were similar in the two groups during feeding. Similarly, the duration of motor quiescence during fasting was lower in feedingintolerant than feeding-tolerant infants during fast-
ing (Table 4; P < O.O5),but values were similar in the two groups during feeding. Long-Term
Outcome
All eight feeding-intolerant infants were evaluated for the presence of an underlying metabolic or anatomic cause for their intolerance; none was found. However, the differences in motor activity in preterm infants was reflected in the clinical outcomes. Babies who were feeding intolerant had significantly more days of feeding difficulty; these clinical problems resulted in a longer time to achieve full oral nutrition and discharge from the hospital (Table 1;all P < 0.01). Manometry was recorded again approximately 2 weeks later in all eight feeding-intolerant preterm infants. In six infants, cluster duration (18.6 f 1.6min/h per lead) and quiescence duration (25.6 * 2.7 min/h per lead) were similar to those for feeding-tolerant infants. Three of these infants showed a change in motor activity in response to feeding. Five of these infants had no further feeding problems during the second poststudy week; feed-
Proximal duodenum
Mid duodenum
125 mm Hg Mid duodenum
Distal duodenum
Figure 4. Duodenal motor activity in the same infant in Figure 3 approximately 60 minutes after the milk infusion began. Note that there is little difference in duration and occurrence of quiescence or contractures compared with that seen in Figure 3.
November
MANOMETRY
1992
Table 2. Motor Activity in Preterm Infants During Fasting Preterm infants Quiescence (% of recording time) Phase II Irregular (% of recording time) Clusters (% of recording time) MMCs (% of recording time)
Tolerant (n = 40)
Intolerant (n = 8)
43.4 * 2.3
27.0 f 2.9O
13.5 + 1.0 40.7 + 2.2 3.1 f 0.7
8.5 + 1.2” 58.6 + 3.2b 6.6 f 2.3
“P < 0.05. hP < 0.005.
ings were withheld for one infant on two single occasions. Two of the preterm infants who were restudied continued to show prolonged cluster duration (50.9 and 47.4 min/h per lead) and less prominent quiescence (6.4 and 7.4 min/h per lead) compared with feeding-tolerant infants. Neither of these two infants showed a change in motor activity in response to feeding. Feeding problems persisted in these two infants. Manometry was repeated in both of these infants 2 weeks later. At that time, the cluster duration had decreased to 13.4 and 14.8 min/h per lead, and the quiescence increased to 28.7 and 21.4 min/h per lead, similar to characteristics in feeding-tolerant infants. One of the two infants showed a change in motor activity in response to feeding. Neither baby showed feeding intolerance after this final study. Sensitivity
and Specificity
We assessed the sensitivity and specificity of manometrics in identifying infants at risk for feeding intolerance. All manometric records were reviewed and scored using two criteria. During fasting, the duration of motor quiescence in minutes per hour per lead was determined. Based on normative data for preterm infants of 28-32 weeks’ gestationional age,7 a value of ~20 min/h per lead (mean - 3 SD) was scored as abnormal. Abnormal feeding responses
Table 3. Characteristics of Quiescence, Clusters, and MMCs in Fasting Preterm Infants Preterm infants Quiescence Duration (min/h per lead) Duration of episode (min) Clusters Duration (min/h per lead) Duration of episode (min) MMCs Periodicity (min) Velocity (cm/min) “P < 0.005. bP < 0.05.
Tolerant
Intolerant
24.5 If: 1.2
15.6 + 1.7”
1.1 f 0.1
0.7 + O.lb
23.2 + 1.3 1.8 f 0.1
33.8 + 2.0a 2.6 + 0.3b
63 f 9 1.6 f 0.4
65 + 27 2.2 f 0.7
PREDICTS FEEDING READINESS
Table 4. Motor Activity
Response to Feeding
1527
in Preterm
Infants
Preterm
infants
Tolerant Motility index No. of peaks per 30 min Quiescence duration Intolerant Motility index No. of peaks per 30 min Quiescence duration “P < 0.05, unpaired
Fasting
Fed
P value (fasting vs. fed)
12.5 + 0.1
13.3 + 0.1
0.0001
165 f 9 24.5 f 1.3
235 f 9 19.5 f 1.7
0.0001 0.007
13.2 f 0.2”
13.0 + 0.3
NS
225 +_12” 15.6 + 1.7”
232 + 26 18.4 & 4.2
NS NS
Student’s t test, tolerant vs. intolerant.
were determined to be present if the motility index did not change by more than 10% of that seen during fasting. Records were defined to predict risk for feeding intolerance if both criteria were abnormal. Using these criteria, manometrics identified all 8 feedingintolerant infants correctly; additionally, 6 of the 39 feeding-tolerant infants were identified to be at risk for intolerance. Thus, the sensitivity of manometry recordings to predict feeding intolerance was 1.0, and specificity was 0.13. Discussion Feeding intolerance results in significant mortality and morbidity in preterm infants. However, no diagnostic test is currently available to identify preterm infants who are at risk for feeding intolerance. Manometry has been used extensively to evaluate adult disorders of gastrointestinal motility. Manometric abnormalities in such patients include an absence of MMCs, exaggerated cluster activity, absence of a “feeding pattern,” and low-amplitude contractions.5 These same motor activity patterns have been described in children with myopathic and neuropathic diseases.g-” However, preterm and term infants also show motor activity that is “abnormal” compared with that of adults and older children in that cluster activity is prominent, MMCs are often absent, and a “fed response” is inconsistently present. Hence, there was a need to control for these normal maturational characteristics of intestinal motor activity by studying age-matched feeding-tolerant infants. In this prospective study, preterm infants with feeding intolerance showed less quiescence and more clustered motor activity during fasting than age-matched infants who tolerated feedings. They also showed no change in motor activity in response to feeding. Thus, the abnormalities shown in the feeding-intolerant infants were similar
1528
BERSETH AND NORDYKE
GASTROENTEROLOGY
to those found in adults with impaired neuronal modulation of motility,5 even when age-related characteristics of motor activity were considered. Manometry provided excellent sensitivity in identifying preterm infants who developed feeding intolerance. Although Morriss et al.” have shown that duodenal contraction rate and amplitude are similar in infants who develop NEC and control infants, they did not evaluate the presence of motor quiescence or migrating activity, nor did they evaluate responses to feeding. Because these two aspects of motor function depend on neuronal inhibition and/or integration, our more extensive profile of motor activity may be necessary to identify babies at risk for feeding intolerance. In view of the high incidence of NEC in preterm infants, a specificity of 0.13 may be acceptable to neonatologists because it permits error in a manner that is less harmful to infants. Manometric “abnormalities” in feeding-intolerant infants were transient, and they resolved over 4 weeks concomitant with improvement of feeding tolerance. Hence, serial manometry may be useful to track and predict feeding readiness in preterm infants and to reduce feeding-related morbidity. Such information should provide valuable insight into the introduction and manipulation of enteral feeding regimens in preterm infants and an ability to test methods to enhance or potentiate the normal maturational processes of gut function in preterm infants. References I. Churella HR, Bachhuber WL, MacLean WC Jr. Survey: methods of feeding 76:243-249.
low birth
weight
infants.
Pediatrics
1985;
Vol. 103, No. 5
2. Brosivs KK, Ritter DA, Kenny JD. Postnatal growth curve of the infant with extremely low birth weight who was fed enterally. Pediatrics 1984;74:778-782. 3. Moyer-Mileur L, Chan GM. Nutritional support of very-lowbirth-weight infants requiring prolonged assisted ventilation. Am J Dis Child 1986;140:929-932. 4. Berseth CL. Early feeding enhances maturation of the preterm small intestine. J Pediatr 1992;120:947-953, 5. Malagelada J-R, Camilleri M, Stanghellini V. Manometric diagnosis of gastrointestinal motility disorders. New York: Thieme, 1986. 6. Amarnath RP, Berseth CL, Malagelada J-R, et al. Postnatal maturation of small intestinal motility in preterm and term infants. J Gastrointest Motil 1989;1:138-143. 7. Berseth CL. Gestational evolution of small intestinal motility in preterm and term infants. J Pediatr 1989;115:646-651. 8. Berseth CL. Neonatal small intestinal motility: the motor responses to feeding in term and preterm infants. J Pediatr 1990;117:777-782. 9. Hyman PE, Napolitano JA, Diego A, Pate1 S, Flores EF, Grill BB, Reddy SN, Garvey TQ, Tomomosa T. Antroduodenal manometry in the evaluation of chronic functional gastrointestinal symptoms. Pediatrics 1990;86:39-44. 10. Anuras S, Mitros FA, Soper RT, Pringle KC, Maves BV, Younoszai MK, Franken, Jr., EA, Whitington P. Chronic intestinal pseudo-obstruction in young children. Gastroenterology 1986;91:62-70. 11. Hyman PE, McDiarmid SV, Napolitano J, et al. Anteroduodenal motility in children with chronic intestinal pseudo-obstruction. J Pediatr 1988;112:899-905. 12. Morriss FH, Moore M, Gibson T, West MS. Motility of the small intestine in preterm infants who later have necrotizing enterocolitis. J Pediatr 1990;117:520-523. Received December 23, 1991. Accepted May 28, 1992. Address requests for reprints to: Carol Lynn Berseth, M.D., Pediatrics Newborn Section, Baylor College of Medicine, One Baylor Plaza, Houston, Texas 77030. Supported by National Institutes of Health grant HD 24558 and in part by the North American Pediatric Pseudo-obstruction Society.
