Effect of Infant Formula on Stool Characteristics of Young Infants

(1)

Effect

of Infant

Formula

on Stool

Characteristics

of Young

Infants

Jeffrey S. Hyams, MD*; William R. Treem, MD*; Nancy L. Etienne, RN*; Harry Weinerman, MD*;

Douglas MacGilpin, MD*; Peter Hine, MD*; Karin Choy, MD*; and Georgine Burke, PhD*

ABSTRACT. Background. Many infants are switched

between multiple formula preparations early in life

be-cause of perceived abnormalities in stooling pattern as

well as gastrointestinal symptoms.

Objective. To investigate the relationship between

the type of formula consumed and the stooling

charac-teristics and gastrointestinal symptoms of young infants.

Methods. Healthy 1-month-old infants were fed one

of four commercial formula preparations (Enfamil,

Enfa-mil with Iron, ProSobee, and Nutramigen) for 12 to 14

days in a prospective double-blinded (parent/physician)

fashion. Parents completed a daily diary of stool

charac-teristics as well as severity of spitting, gas, and crying for

the last 7 days of the study period. A breast-fed infant

group was studied as well.

Results. Two hundred eighty five infants were

en-rolled and 238 completed the study. Infants receiving

breast milk or Nutramigen had twice as many stools as

other formula groups (P < .001). Infants receiving

ProSobee had hard/firm stools more often than

breast-fed or other formula-fed groups (P < .00001). Watery

stools were more common in infants fed Nutramigen

than other formula groups (P < .04). Green stools were

more common in 12 mgfL iron preparations (Enfamil with

iron, ProSobee, Nutramigen) than in those with 1 mgIL

(Enfamil, breast milk) (P < .00001). Spitting, gassiness, and crying were of equal severity in all formula groups.

Conclusions. The interpretation of stool frequency,

color, and consistency must take into account the infant’s

formula type as significant variations in normal infants

occur. Parental education on the range of infant stooling

characteristics as well as the common occurrence of

spit-ting, gas, and crying may alleviate concern for formula

intolerance and underlying gastrointestinal disease.

Pediatrics 1995;95:50-54; confidence interval, ANOVA, analysis of variance.

ABBREVIATIONS. CI, confidence interval; ANOVA, analysis of variance.

In the United States approximately 50% of all

new-borns, and 87% of 3-month-old infants, are fed a

commercial formula either as their sole source of

nutrition or as a supplement to breast milk (data on

file, Mead Johnson Nutritionals, Evansville, IN). It is

common for many formula-fed infants to be

From the *paflent of Pediatrics, Hartford Hospital, Hartford, Connect-icut and The University of Connecticut Health Center, Farmington,

Connecticut.

Received for publication Dec 2, 1993; accepted Apr 6, 1994.

Reprint requests to (J.S.H.) Dept of Pediatrics, Hartford Hospital, P0 Box 5037, Hartford, CT 06102-5037.

Acad-emy of Pediatrics.

switched from one formula to another either by their parents or physicians. Although the reasons for such

frequent formula switching are sometimes elusive,

most of the changes occur because of perceived

ab-normalities in stooling patterns (too much/too little, too hard/too loose) or reports that the infant is

un-comfortable while consuming a specific formula.

Although there may be considerable variability in

the frequency with which infants pass their stools,5

our knowledge of the effects of various formulas on

stool characteristics is limited. Previous reports that

have evaluated the impact of formula content on

infant stool habits have been limited to cow’s milk

preparations with varying iron contents.69 The

pur-pose of the present study was to examine the effect of

a variety of standard infant formulas on the stool

characteristics as well as gastrointestinal symptoms

of a large group of healthy infants in a

double-blinded prospective manner.

Patients

PATIENTS AND METHODS

Infants were recruited from four private pediatric practices in the greater Hartford area. Eligibility for the study was determined by: term birth, no significant neonatal problems, no intolerance of breast milk or the original formula, and normal weight gain. After obtaining informed written parental consent, study infants were seen at 4 weeks of age, weighed, and the current type of formula or feeding was recorded. For those infants who were bottle-fed the parents received a 12 to 14 day supply of a study formula (deter-mined by random number) that was identified by a code letter only. Neither the parent nor the study nurses or physicians were aware of the composition of the formula until the entire study was completed. The compositions of the study formulas as marketed at the time of initiation of the study are shown in Table I.

The first 5 days on the study formula were considered a “wash-out” period. Breast-fed infants fed normally with no formula supplements allowed. Starting on the sixth study day a daily log was maintained for I week. Information recorded on this

pre-printed log included: number of stools each day, stool color, stool

consistency, number of stools passed with straining, amount of gas, spitting, and crying, as well as number of ounces of formula consumed for bottle-fed infants. Each parent was shown a series of photographs by a nurse coordinator depicting stool color and

consistency and was supplied with written criteria explaining the

various characteristics of stool consistency.

Guidelines to help determine the amount of gas, spitting, and

crying were given to the parents. No gas was defined as absent or

minimal gas (0 points), mild connoted gas passed several times during the day (1 point), and severe was defined by very frequent gas passage (2 points). Spitting was described as none (0 points), “few” if it occurred no more than twice daily and involved only a mouthful of milk (I point), and “many” connoted three or more episodes of effortless emesis or forceful vomiting (2 points). Mild crying was described as short periods of cry, <5 minutes in dura-tion, and easily consolable (0 points). Moderate crying was deter-mined by longer periods, up to 20 minutes, but still generally consolable (I point). Severe crying was described by periods

at Viet Nam:AAP Sponsored on September 1, 2020

www.aappublications.org/news

(2)

TABLE 1. Composition of Study Milks

Milk mOsmt/kg CHO (g/l00 mL) Fat (g/I00 mL) P rotein (g/I00 mL) Fe mg/qt

Breast* 273 6.5 (Lactose) 3.6 1 .4 I

Enfamil

300

6.9 (Lactose)

3.8

55% Coconut oil 45% Soy oil 55% Coconut oil

1 .5

1.5

60% Whey 40% Casein 60% Whey

1

12

with iron 45% Soy oil 40% Casein

ProSoBee

Nutramigen

200

320

6.7 (Corn syrup

solids)

8.8 (Corn syrup

solids, corn starch)

3.5

2.6

55% Coconut oil 45% Soy oil Corn oil

2.0

I .9 Soy

Casein

Hydrolysate

12

* Established from published sources.

t mOsm, milliosmole; CHO, carbohydrate; Fe, iron.

longer than 20 to 30 minutes several times per day, and not easily consolable (2 points). The infants were seen again at 6 weeks of age when they were weighed and the data log collected.

This study was approved by the Institutional Review Board at Hartford Hospital.

Statistical Analysis

Within each feeding type, the numbers of patients who dropped out of the study for reasons related to type of feeding were expressed as a rate per 100 enrolled. Confidence intervals were estimated for each rate by standard methodsio with 5% error.

The information from the daily logs of stool characteristics was summarized as follows. Totals for the 7-day period were obtained

by adding daily totals for number of stools, and the number each

of yellow, brown, or green stools, and hard, firm, mushy, or

watery stools. An average daily stool number was calculated.

Stool color and consistency totals were expressed as the

percent-age of all stools for 7 days. Group means for each feeding type

were calculated for the following: total stools, average stools per day, and the percentage of yellow, brown, green, hard, firm, mushy, and watery stools. Differences in means were evaluated with one-way analysis of variance (ANOVA), usingJMP statistical software (SAS Institute, Cary, NC). If the group mean differences were statistically significant with <5% error, post hoc comparisons

were made with either Dunnett’s test of comparisons with a reference group (breast-fed infants) or the Tukey-Kramer HSD (honestly sig-nificantly difference) test for multiple group comparisons. In addi-lion, categorical measures were calculated based on the presence of

watery, firm, or hard stools for 3 or more days. The Pearson was

used to evaluate group differences for all categorical variables and an

error of <5% was accepted as evidence of statistical significance. A composite score was obtained for each of the symptoms of crying, spitting, and gas by summing the daily scores recorded during the 1-week observation period. The Kruskal-Wallace test of

equality of ranks was used to compare scores among the feeding

groups. To account for the possible bias introduced by attrition from the study related to these specific symptoms, the composite scores were grouped into three categories: 0 to 4 was taken as the absence of a symptom or of a mild nature, 5 to 9 connoted moderate, and 10 or greater connoted severe. Patients who dropped out of the study for spitting, gas, or crying were assigned a score of 10 for that particular symptom and added to the “severe” group. The proportion of infants categorized as mild, moderate, or severe within each formula type was evaluated with Pearson . In all cases the results from the and

Kruskal-Wallace tests were in agreement.

Study Population

RESULTS

The characteristics of the study infants are shown

in Table 2. A total of 285 infants were enrolled and

238 completed the study. Twenty infants were

dropped from the study because of nonmilk related

problems including antibiotic exposure (9), deviation

from feeding protocol (four breast-fed infants who

received formula), failure of parent to adequately complete diary (6), and diagnosis of cystic fibrosis

(1). Twenty-three formula-fed infants were

with-drawn from the study during the data collection

phase by their parents because of perceived difficul-ties with the formula. These included: Enfamil-five infants (three, irritability; two, spitting), Enfamil with iron-three infants (all with irritability and spit-ting), ProSobee-four infants (three, gas; one, irrita-bility), and Nutramigen-1 I infants (six, irritability;

TABLE 2. Characteristics of Stu dy Population

Breast Enfamil Enfamil ProSollee Nutramigen

with Iron

Number enrolled 66 56 47 58 58

Number dropout total 12 6 7 7 15

Number related to milk 4 5 3 4 II

Number completing study 54 50 40 51 43

Male/Female 26/28 24/26 25/15 26/25 22/21

Birth weight (kg) 3.5 ± 0.1 3.6 ± 0.1 3.5 ± 0.1 3.5 ± 0.1 3.6 ± 0.1

Entry weight (kg) 4.3 ± 0.1 4.5 ± 0.1 4.3 ± 0.1 4.4 ± 0.1 4.4 ± 0.1

Finish weight (kg) 4.9 ± 0.1 5.0 ± 0.1 4.8 ± 0.1 4.9 ± 0.1 4.9 ± 0.1 Weight gain (kg) 0.51 ± 0.03 0.52 ± 0.03 0.55 ± 0.04 0.49 ± 0.03 0.49 ± 0.03

Birth order (%)

First 48 34 48 31 44

Second 39 54 38 49 35

Third 13 12 14 20 21

Ounces consumed 28.5 ± 0.7 28.5 ± 0.8 26.2 ± 0.7 24.5 ± 0.8*

Pre-study formula (%)

Breast 100 0 0 0 0

Cow 92 90 94 91

Soy 8 10 6 9

Data expressed as mean ± SEM.

* P < 0.05 compared to Enfamil, Enfamil with iron.

(3)

TABLE 3. Stooling Frequency in Study Groups 41

I

Enfamil with iron Prosobee 6.7 8.6

* P < .00001 breast or Nutramigen versus Enfamil, Enfamil with iron, Prosobee.

two, gas; two, poor intake; one, spitting). The drop-out rates for formula-related reasons were not statis-tically different: Nutramigen (19%, confidence

inter-val (CI) 10 to 32%), Enfamil (9%, CI 3 to 20%),

Enfamil with iron (6%, CI 2 to 19%), and ProSobee

(7%, CI 2 to 18%). However, when the dropout rate

for Nutramigen-fed infants was compared with the

other formulas combined, the Nutramigen-fed group

was almost three times as likely to dropout for

for-mula-related reasons (odds ratio 2.9, 95% CI I .1 to

7.6). A significant difference in total formula intake

per day was noted between the four formula groups

(ANOVA, P < .003), with infants in the Nutramigen

group taking the least amount per day. However,

mean weight gain during the study period was not

statistically different for all feeding groups.

Stooling Frequency

The relationship of milk type to stooling frequency is seen in Table 3. A significantly increased number

of stools were observed in those infants receiving

Nutramigen or breast milk compared with the other

three formulas (P < .001). Stool numbers were similar

in the infants fed Enfamil, Enfamil with iron, and

ProSobee.

Stool Consistency

The relationship of milk type to stool consistency is seen in Table 4. The data reflect the mean

percent-age of stools of a particular consistency passed by

infants consuming each of the respective formulas or

breast milk. Hard stools and firm stools were signif-icantly more frequent in those infants fed ProSobee

than in the other three formula groups or in the

breast-fed infants (P < .025, hard, and P < .00001,

firm). A total of 53% of infants fed ProSobee had

hard stools on at least 3 of 7 days during the 1-week

observation period compared with 0% breast, 4%

Enfamil, 8% Enfamil with iron, and 2% Nutramigen

(P < .001).

Significant differences were noted in the percent of

watery stools in the feeding groups (ANOVA, P <

.00001). Watery stools were significantly more

fre-quent in the Nutramigen group compared with all

others (P < .04). Seventy-six percent of infants fed Nutramigen had watery stools for at least 3 of 7 days

compared with 20% ProSobee, 23% Enfamil with

iron, 51% Enfamil, and 59% breast (P < .00001).

Wa-Number Total Stools/7 Days Daily Stools Subjects Range Mean ± SEM Range Mean ± SEM

Breast 54 2-67 29.4 ± 2.2* 0.3- 4.2 ± 0.3* 9.6

Enfamil 51 3-47 15.8 ± 1.2 0.4- 2.3 ± 0.2

40 6-29 14.8 ± 1.0 0.9- 2.1 ± 0.1 4.1

51 5-29 15.1 ± 0.9 0.7- 2.2 ± 0.1

4.1

Nutramigen 42 8-60 25.4 ± 1.7* I.1 3.6 ± 0.2*

TABLE 4. Stool Consi stency in Stud y Groups

Hard Firm Mushy Watery Breast 0% <1% 62 ± 5% 38 ± 5%

Enfamil <1% (0-17) 2±1% (0-100) 58±5% (0-100) 40±5% Enfamil with iron (0-14) <1% (0-48) 4±1% (0-100) 78±4% (0-100) 18±4% Prosobee (0-11) 5±2%* (0-33) 33±5%t (0-100) 51±5% (0-100) 12±3% Nutramigen (0-73) <1% (0-100) 1± 1% (0-25)

(0-100) 41 ± 6%

(0-100)

(0-100) 58±61 (0-100)

Data represent mean (%) of stools of respective consistency passed by infants on a particular milk source, as well as range of values

for each study group.

* < .025 Prosobee versus all others. t P < .00001 Prosobee versus all others.

I

< .04 Nutramigen versus all others.

§

P < .006 Enfamil versus Enfamil with iron.

tery stools were more common in the Enfamil group

compared with the Enfamil with iron group (P <

.006).

Straining at the time of defecation was similar

among the formula preparations with the exception

of being more common in those infants fed ProSobee

compared with those fed Nutramigen (P < .007). No

differences were noted between infants fed Enfamil

and those fed Enfamil with iron.

Stool Color

The relationship of stool color to type of infant

feeding is shown in Fig I. Breast-fed infants had an

average of 90 ± 5% yellow stools over the 7-day

observation period compared with 46 ± 5% Enfamil,

36 ± 5% Enfamil with iron, 35 ± 6% ProSobee, and

28 ± 5% Nutramigen (P < .00001). Brown stools were

more common in infants fed Enfamil (34 ± 4%) than

either Enfamil with iron (1 1 ± 5%), ProSobee (14 ±

4%), or breast (5 ± 4%) (P < .003). Green stools were

observed an average of 50 ± 5% among infants fed

Enfamil with iron, ProSobee, or Nutramigen

com-pared with 20 ± 5% of infants fed Enfamil or 5 ± 4%

of breast-fed infants (P < .00001).

I

0 YELLOW D BROWN

#{149}

GREEN

STOOL COLOR

Fig 1. Relationship of infant formula to stool color. Yellow stools were more common in the breast-fed than formula-fed infants

(P < .00001). Brown stools were more common in Enfamil than Enfamil/Fe, ProSobee, or breast milk (P < .003). Green stools were more common in Enfamil/Fe, ProSobee, and Nutramigen than Enfamil or breast milk (P < .00001).

(4)

Spitting, Gas, Crying

The relationships of type of milk ingested to spit-ting, gas, and crying are shown in Fig 2. Spitting was

commonly observed and was of equal severity in all

study groups. For all study groups combined

spit-ting was rated none/mild in 28%, moderate in 45%,

and severe in 27% of infants.

Parents rated their infants degree of gassiness as

well as crying similarly in all the formula groups. For

all study groups combined, crying was rated none/

mild in 69%, moderate in 26%, and severe in 5% of

infants.

DISCUSSION

In the present prospective blinded study we have

defined a broad range of stooling frequency,

consis-tency, and color for normal infants fed a variety of

commercial infant formulas as well as breast milk. In

addition, we have described the frequency of

com-monly observed gastrointestinal symptoms and

cry-ing in this patient population.

Our data show that breast-fed infants as well as

those fed Nutramigen have, on average, twice as

many stools as those fed a cow’s milk or soy

prepa-ration. The reasons for this difference are unclear and

similar observations have been made with another

protein hydrolysate preparation.11

BREAS

-_--ENFAMII

ENFAMUJFI

PROSOBEI

NUTRA14IGEt .:::.:::::

S I I I I I I I I I

0 10 20 30 40 50 60 70 80 90 100

SPIT(% of Patients)

BREftS :::::::::::::::

ENFAMII

ENFAMILJFI

PROSOBEI :.::::::::::::

NL1TRft,.lIGE : :: ::::::::::::::::

I I I I I I I 1 I 1

0 10 20 30 40 50 60 70 80 90 100

GAS

(%

of Patients)

BREPS1

EtIFAIIL : : : : : : : : : : : : : : : : :: ::::. : : : : : : : : : : : : :

ENFAMII../FE .:::::: .: : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : :

PROSOBEE

NI.J1RAMIGEN

I I I I I I I I I -I

D 10 20 30 40 50 60 70 80 90 100

CRY

(%

of Patients)

I

0

NONE-MIW MOOERATE

#{149}

SEVERE

Fig 2. Relationship of infant formula to spitting, gas, and crying.

No statistical difference in spitting, gas, or crying between formula groups was noted. Decreased moderate/severe gas in breast milk versus Enfamil (P < .01) or Enfamil/Fe (P < .02). Decreased mod-erate/severe crying in breast milk versus ProSobee (P < .03) and Enfamil (P < .001).

Stool consistency also varied between the study

groups with those fed the soy preparation having a

significantly greater number of hard and firm stools. Soy protein isolate formula contains a small amount of fiber while the other formula preparations do not.

Previous observations have suggested that

increas-ing the fiber content of infant formula may be

asso-ciated with firmer stools.12 No difference was noted in the frequency of firm or hard stools in infants fed

Enfamil or Enfamil with iron. Watery stools were

noted more commonly in infants who were fed

Nu-tramigen than the other formula-fed groups as well

as those fed breast milk. It is possible that the lower fat and higher carbohydrate content of Nutramigen

compared with the other milk preparations might be

associated with more rapid intestinal transit and

looser stools.

Wide variations were noted in stool color. Green

stools, often perceived by many parents as abnormal,

were noted in almost 50% of infants fed Enfamil with

iron, ProSobee, and Nutramigen and were less

fre-quent in breast-fed or Enfamil-fed babies. A previous study7 demonstrated that green stools predominated

in infants fed a whey predominant formula with 12

mg/L iron whereas yellow stools were most frequent

when the same formula had I.5 mg/L iron. A

casein-based formula, with 12 mg/L or 1.5 mg/L iron, was

associated with primarily yellow or brown stools. In

our study, the association of 12 mg/L iron with

either a 60% whey/40% casein protein content, soy

protein isolate or casein hydrolysate preparation was associated with primarily green stools. Green stools

predominated in infants fed Enfamil with iron

com-pared with yellow stools in those receiving Enfamil. No significant differences were noted in the

sever-ity of spitting, gas, and crying between the four

formula groups. Our study did not address the

pos-sibility that “colic-like” symptoms might have been

associated with either cow’s milk or soy protein

in-tolerance in infants fed these types of formulas. The

frequency of “colic-like” symptoms was similar in

the infants fed intact protein containing formulas

compared with those fed a protein hydrolysate

for-mula. However, we cannot exclude the possibility

that a Type II error might have limited detection of a difference of “colic-like” symptoms or severe crying

among the formula groups. It is of note that the

dropout rate for formula-related reasons was

great-est with the protein hydrolysate formula. Although

no specific reason reached statistical significance, a

combination of irritability and poor feeding was

cited by most parents.

The results of our study underscore the

impor-tance of parental education in the interpretation of

stooling patterns and gastrointestinal symptoms

during the administration of various infant formulas.

Although most formula-fed infants are given

stan-dard cow’s milk protein preparations with 12 mg/L

iron to start, they may be switched to reduced iron

cow’s milk products or soy or protein hydrolysate

formulas because of a perception of allergy or

intol-erance. Although true hypersensitivity to cow’s milk

or soy protein may occur, it is uncommon and many

infants are often mislabeled as being “allergic” to a

(5)

particular formula when their symptoms such as

loose stools, gas, spitting, and crying probably fall

within the normal range of variability observed with

all infant formulas. Previous observations have

sug-gested that the effect of changing formula

composi-tion in infants with colicky symptoms diminishes

with time and is often not reproducible.13 The

ten-dency to participate in “formula jumping” without

defined therapeutic goals needs to be avoided since

the switch from a cow’s milk preparation to a soy

protein or protein hydrolysate preparation may be

associated with their distinctive features as well.

ACKNOWLEDGMENTS

This study was supported in part by a grant from Mead John-son Nutritionals, Evansville, IN.

The authors thank the following individuals for their help in the completion of this study: Marie Shoup, RN, Liz Blankschen, RN, Joan Nowicki, LPN, Joanne Quaranto, RN, Patricia Bush, RN, Lon Testa, RN, Maureen Kelley, LPN, Rita Fazzino, RN, Joan Jarvis, RN, and Patricia Esposito, RD.

REFERENCES

1. Weaver LT, Steiner H. The bowel habit of young children. Arc/i Dis Child. 1984;59:649-652

2. Wolman Ii. Laboratori,i Applications in Clinical Practice. New York, NY:

McGraw-Hill; 1957:696-697

3. Nyan WL. Stool frequency of normal infants in the first week of life.

Pediatrics. 1952;10:414-425

4. Weaver LT, Ewing C, Taylor LC. The bowel habit of milk-fed infants. / Pediatr Gastroenterol Nutr. 1988;7:568-571

5. Lemoh IN, Brooke 0G. Frequency and weight of normal stools in

infancy. Arch Dis Child. 1979;54:719-720

6. Oski FA. Iron-fortified formulas and gastrointestinal symptoms in infants: a controlled study. Pediatrics. 1980;66:168-170

7. Malacaman EE, Abbousy FK, Crooke D, Nauyok Jr C. Effect of protein source and iron content of infant formula on stool characteristics. / Pediatr Gastroenterol Nutr. 1985;4:771-773

8. Nelson SE, Ziegler EE, Copeland AM, Edwards BB, Fomon SJ. Lack of

adverse reactions to iron-fortified formula. Pediatrics. 1988;81 :360-364 9. Bradley CK, Hillman L, Pennridge Pediatric Associates, et al.

Evalua-tion of two iron-fortified, milk based formulas during infancy. Ped

fat-rics. 1993;91:908-914

10. Fleiss JL. Statistical Method for Rates and Proportions. 2nd ed. New York,

NY: John Wiley & Sons, mc, 1981

I I. Muller-Teichen G, Passian K. The feeding of newborn and young in-fants in the first trimester with a hypoallergenic bottle formula: a report of experiences. Extracta Paediatr. 1988;12:170-186

12. Treem WR, Hyams J5, Blankschen E, Etienne N, Paule CL, Borschel

MW. Evaluation of the effect of a fiber-enriched formula on infant colic.

IPediatr. 1991;119:695-701

13. Forsyth BWC. Colic and the effect of changing formulas: a double-blind,

multiple-crossover study. / Pediatr. 1989;115:521-526

RONALD McDONALD CHILDREN’S CHARITIES

1994 AWARD OF EXCELLENCE WINNER

AUDREY EVANS, MD

Twenty years ago, Dr Evans’ idea-Ronald McDonald House-became a reality

in Philadelphia.

She was, and is, a doctor at The Children’s Hospital of Philadelphia, a specialist in children’s oncology, who dreamed of a house where the families of her seriously

ill young patients could stay . . . an inexpensive home-away-from-home within

walking distance of the hospital, where families could care for one another in a

supportive environment. Now there are more than 160 Ronald McDonald Houses

in I 1 countries around the world and I .5 million family members have called them

home.

Today, Dr Evans divides her workday between the bedsides of her patients and

the research lab, where she is getting closer to a better understanding of

neuro-blastoma, a deadly childhood cancer. She and her team of researchers are zeroing

in on the reasons why tumors of the nervous system will sometimes spontaneously regress.

Meanwhile, survival rates continue to rise. When she began her career in the

early 1950s, only one child in ten survived. “Now,” she says, “we cure more than

75% of the children with cancer.”

Throughout her career, Dr Evans has been just as concerned with nourishing the

spirits of her patients as she has been in stabilizing their illnesses. She believed so strongly in the importance of a hospital chaplain that she paid the chaplain’s salary

until the hospital administrators agreed. She did the same for the hospital’s first

social worker.

As a doctor, scientist, and humanitarian, Audrey Evans knows how to get the

right things done. Twice a year, she organizes a Celebration of Life, an emotional service with music, laughter, and tears, that celebrates the lives of the children who

have died. She also helped create an ongoing healing program for entire families.

Several months after the death of a child, the families are invited back to talk to

other families “to work through what it was like then and how they’re doing now.”

Dr Evans will donate her $100 000 award to The Cancer Center at The Children’s Hospital of Philadelphia.

(6)

1995;95;50

Pediatrics

MacGilpin, Peter Hine, Karin Choy and Georgine Burke

Jeffrey S. Hyams, William R. Treem, Nancy L. Etienne, Harry Weinerman, Douglas