Sweat
Chloride
Concentrations
in Infants
Homozygous
or Heterozygous
for
F8
Cystic
Fibrosis
Philip M. Farrell, MD, PhD and Rebecca E. Koscik, MS
ABSTRACT. Objective. To determine whether an
ad-equate volume of sweat could be obtained routinely from
infants younger than 6 weeks old and to evaluate sweat
chloride levels in infants with known genotype statuses, including heterozygote carriers for cystic fibrosis (CF).
Methodology. Infants were evaluated using
pilo-carpine iontophoresis and measurement of sweat volume
and chloride concentration. The majority of these infants were referred because of newborn screening test results positive for CF based on immunoreactive trypsinogen analysis. DNA analyses for the 3-base pair deletion at codon 508 of the CF transmembrane regulator gene (F
mutation) were performed whenever possible, and
pa-tients with CF were categorized by genotype.
Results. Sweat tests were performed successfully
(5O mg of sweat) in 99.3% of the infants tested, and there was no difference in the proportion of unsuccessful tests in infants younger than or older than 6 weeks of age. The normal mean ± SD sweat chloride was 10.6 ± 5.2
mEq/L (95% confidence interval, 9.9-11.3). Patients with
CF who are F homozygotes or F compound heterozy-gotes or who have two other non-F mutant alleles were
shown to have similar sweat chloride levels, with mean
values of 99.9, 98.8, and 96.6 mEqIL, respectively. The group of infants who were found to be CF (F) hetero-zygote carriers, when compared with the healthy group, had mildly but significantly increased sweat chloride
concentrations, with a mean ± SD of 14.9 ± 8.4 mEq/L
(95% confidence interval, 13.4-16.4).
Conclusions. Quantitative pilocarpine iontophoresis can be used successfully in infants younger than 6 weeks of age who are undergoing routine diagnostic
evalua-tions to follow up newborn screening test results that are
positive for CF. The upper limit of normal sweat chloride in infants should be revised to 40 mEqIL (mean + 3 SD of the CF heterozygote carrier group). CF heterozygote car-rier infants with one F mutant allele show phenotypic manifestations of CF, including subclinical elevations of sweat chloride. Pediatrics 199697:524-528; cystic fibrosis,
sweat test, infants, genotype, newborn screening.
ABBREVIATIONS. CF. cystic fibrosis; CFTR, cystic fibrosis trans-membrane regulator; IRT, immunoreactive trypsinogen; F, 3-base pair deletion at codon 508 of the CFTR gene; QPIT,
quan-titative piocarpine iontophoresis (sweat) test; N, normal DNA test revealing no F alleles (designates the wild-type or normal gene).
Ever since the landmark discovery by di
Sant’Agnese et al,’ the diagnosis of cystic fibrosis
From the Department of Pediatrics, University of Wisconsin, Madison.
Received for publication Feb 27, 1995; accepted May 18, 1995.
Reprint requests to (P.M.F.) Professor of Pediatrics and Dean, University of Wisconsin Medical School, 600 Highland Aye, Madison, WI 53792-4108.
PEDIATRICS (ISSN 0031 4005). Copyright © 1996 by the American Acad-emy of Pediatrics.
(CF) has relied on detecting elevated chloride or
sodium concentrations in sweat samples. The sweat
electrolyte abnormality in CF reflects the
fundamen-tal disturbance, because it is now well established
that the underlying defect in the cystic fibrosis
trans-membrane regulator
(CFTR)
gene primarily altersthe chloride channel on the apical surface of CF
epithelial cells.2 For routine diagnostic purposes,
sweat is obtained after stimulation with piocarpine
iontophoresis,3 using standardized sweat test
meth-ods that are well established and well described4’
and are mandated in the CF centers of the United
States that are sponsored by the national Cystic
Fi-brosis Foundation. Despite the technical,
labor-inten-sive demands of the quantitative pilocarpine
ionto-phoresis sweat test, this method has been very
successful because of its high sensitivity and
speci-ficity attributable to the wide separation of
electro-lyte concentrations between individuals with and
without CF. This is especially true for chloride
val-ues, which overlap less than those of sodium4’6 and
discriminate accurately at 60 mEq/L between the
populations with and without CF. On the other
hand, patients have been reported with characteristic
manifestations of CF who have had “borderline”7 or
“intermediate range”8 sweat chloride levels. In
addi-tion, some evidence9”0” suggests that “obligate” CF
heterozygote carriers (ie, parents of patients with
proven CF) may have intermediate-range chloride
values; however, whether CF heterozygote carriers
show any phenotypic manifestations has been a
con-troversial issue for decades.7”2”3 Also, recent
obser-vations indicate that some isolated individuals with
two CF mutant alleles have typical CF respiratory
disease and “normal” chloride levels.’4
Although CF is the most common severe
autoso-mal recessive disease of the white population, its
diagnosis is often delayed. The median and mean
ages of diagnosis in the United States are
approxi-mately 5 months and 4 years, respectively, based on
data available recently from the Cystic Fibrosis
Foun-dation Patient Registry.’5”6 There has been limited
experience with sweat testing young infants (ie,
pa-tients younger than 2 months of age), and questions
have been raised about the usefulness and reliability
of sweat tests in this population.’7 Screening
pro-grams for CF in newborns have had to face this
challenge and have had to use sweat tests in
asymp-tomatic 3- to 8-week-old infants.’8”9 As we proceeded
with such a program, combining measurement of
anal-ysis for the major CFTR mutation, namely the 3-base
pair deletion at codon 508 of the CFTR gene (F
allele),2#{176}we were presented with an opportunity to
close several gaps concerning infant sweat tests by
addressing the following objectives: (1) determining
whether an adequate volume of sweat could be
ob-tained routinely at less than 6 weeks of age using piocarpine iontophoresis; (2) assessing the chloride
levels in infants with CF who are homozygous for
F compared with F compound heterozygotes (ie,
patients with one F allele and one other mutation) and a group of infants with CF who have two other
(non-F) CFTR mutant alleles; (3) determining the
normal range of chloride values in healthy young
infants without the F mutation; and (4)
determin-ing sweat electrolytes in infants who are
heterozy-gote carriers for the F mutation to ascertain
whether they differ from healthy infants and do,
indeed, have intermediate levels of sweat chloride. The last two objectives have become particularly
fin-portant in CF neonatal screening programs using the
IRT and DNA method#{176} because of the need to
evaluate sweat electrolytes in asymptomatic infants who have one F allele.
METHODS
A total of 725 infants were evaluated in this study, including 481 with newborn screening test results positive for CF; the other infants had been referred to the University of Wisconsin since 1985 for sweat tests. Those with positive screening test results had blood DNA analysis for the F allele whenever sufficient blood was available; this analysis was performed as described else-where, either as part of a two-tiered screening procedure or
subsequently for research purposes. Quantitative piocarpine
iontophoresis tests (QPITs) were performed with measurements of sweat weight and chloride concentrations. Sweat specimens were
collected concurrently from the right and left arms using the
following procedures. The skin of the midforearm was first cleansed with deionized, distilled water. A 2-in-square gauze pad (salt-free) was applied to the skin, soaked with 0.5% piocarpine solution, and covered with a 33-mm-diameter metal electrode. lontophoresis was then accomplished with a current gradually
increased over 30 seconds to I .5 mA for 5 minutes. After a
4-minute pause, the skin was rinsed with deionized, distilled water and dried thoroughly; a 33-mm-diameter disk of pre-weighed Whatman (Clifton, NJ) No. 42 filter paper was then
applied over the stimulated area and covered with Parafilm
(American National Can, Chicago, IL) held in place by tape over
the edges, thereby adhering to and sealing the cover. After 30
minutes, the Parafilin was lifted, and the filter paper was removed
with forceps and then rapidly placed in a preweighed, stoppered
Erlenmeyer flask. The weight of the sweat-impregnated filter pa-per was then determined with an analytical balance, and the electrolytes were eluted with deionized, distilled water. A
mini-mum sweat weight of 50 mg was required for analysis, in
accor-dance with guidelines existing at the time of this study. Chloride concentrations were analyzed with a digital chloridometer. Each
analysis was performed in duplicate and was accompanied by
standards, including one with 100 mEq/L of chloride; as a quality oontrol procedure, this standard had to yield a value of 100 ± 4 mEq/L.
Data on sweat chloride concentrations, demographic
character-istics, and dinical information were analyzed using SAS (Cary, NC) statistical software. Statistical analyses included Fisher’s exact and tests for categorical data and t tests, analysis of variance, and linear regression for continuous data. The significance level for all tests was P < .05.
RESULTS
Sweat tests were performed successfully in 99.3%
of the infants, for whom successfully indicates a
weight of sweat of 50 mg or greater for at least one
arm. To our knowledge, the 720 infants with
success-ful sweat tests represent the largest series reported
for an infant population.
We used two methods to determine whether an
adequate volume of sweat could be obtained in
pa-tients seen for QPIT as a diagnostic assessment after
newborn screening test results positive for CF. First,
we compared the proportion of tests leading to quan-tities not sufficient for sweat analysis in infants
younger than 6 weeks of age with those of infants
older than 6 weeks; there was no difference between the two groups (P = .576, Fisher’s exact test). Second,
we compared the mean sweat volumes between the
young (younger than 6 weeks of age) and older
infants. These results are shown in Table I. The t tests
revealed a statistically significance difference for both the right and left arms. Although the total vol-ume of sweat was significantly lower in the younger
infants, the lower limit of the 95% confidence interval
for each arm is well above our minimum
require-ment of 50 mg (see Table 1). We also examined by
linear regression the sweat volume as a function of
age in weeks and found no significant relationship.
Two populations of infants without CF were used
to calculate normal concentrations of chloride. The
results are shown in Table 2. The first group of
infants found to be free of the F mutation were
found to have mean ± SD chloride levels of 10.6 ±
5.2 mEq/L. The mean ± SD age of this group was 9.3
± 5.3 weeks. The second group of infants referred to
rule out CF sweat tests (generally because of
recur-rent respiratory symptoms) were investigated at an
average age of 21.2 ± 13.1 weeks and were found to have chloride levels that were very similar at 11.7 ± 5.4 mEq/L.
Table 2 also provides sweat chloride data listed
according to genotype categories for the infants who
underwent both sweat tests and DNA analyses. The
data demonstrate no significant differences in sweat
chloride values between the three categories of
pa-tients shown to have CF by both sweat analysis and
the eventual development of characteristic clinical
signs and symptoms.
The group of the 128 CF heterozygote carriers with
one F mutant allele had a mean ± SD sweat
chlo-ride level of 14.9 ± 8.4 mEq/L. This value is
signif-TABLE 1. Amount (Milligrams) of Sweat Obtained by Piocarpine lontophoresis
Age Right Arm Left Arm
n Mean (SD) 95% Confidence Interval
n Mean (SD) 95% Confidence Interval
<6 wk 6 wk
P
41 334
157.1 (78.2) 133.2-181.0
194.4 (79.2) 185.9-202.6
TABLE 2. Infant Sweat Chloride Levels by Category of Infant Sweat Tested*
Subjects (Genotype* Age weks
Mean (SD)
Sweat Chloride mEq/L
Mean (SD) 95% Confidence Interval
Normal (N/N) 184 9.3 (5.3) 10.6 (5.2) 9.9-11.3t
Normal (no gene data) 280 21.2 (13.1) 11.6 (5.3) 11.0-12.2
CF heterozygote carrier (N/F) 128 8.8 (5.0) 14.9 (8.4) 13.4-16.4t
Infants with CF,(F/F) 61 10.5 (9.9) 100.0 (9.2) 97.6-102.4
Infants with CF, (cf/F) 47 9.3 (9.3) 97.6 (15.3) 93.1-102.1
Infants with CF (cf/cf) 7 16.5 (17.3) 99.6 (14.1) 86.5-112.7
* N, normal DNA test revealing no F alleles; CF, cystic fibrosis; F, 3-base pair deletion at codon 508 of the CF transmembrane regulator
gene; cf, other CF transmembrane regulator mutation, ie, mutant allele other than the F mutation.
t Nonparametric 95% confidence intervals were 8.90-10.1 for the N/N group and 11.3-13.7 mEq/L for the N/F CF heterozygote carriers. The distributions of chloride values were skewed (not bimodal) in the two groups and were different (P <.0001) by the Wilcoxon rank sum test.
icantly different from both the healthy populations
and the patients with CF by either parametric or
nonparametric statistical analysis. Although the
in-crease in chloride concentration may seem relatively
modest, there was no overlap in the 95% confidence
intervals compared with either the healthy group
proven to have the F mutant allele or the healthy
group with no gene data (40% and 28% increases,
respectively). In addition, nine healthy infants with
the F allele had sweat chloride values greater than
30 mEQ/L, whereas only one infant lacking this
mutation had a value above this level. The
distribu-tion of sweat chloride concentrations is skewed in
the CF heterozygote carrier group, as in the case of
healthy infants, but it is not bimodal; the
distribu-tions are different
(P
< .0001) when CF heterozygotecarriers (N/F) are compared to the healthy (N/N)
group using the Wilcoxon rank sum test. It also
should be noted that the group of 280 infants with
negative sweat test results but no DNA analyses is
likely to include at least 6 CF (F) heterozygote
carriers, based on the determined frequency (1 in 44)
of this allele in Wisconsin infants.2#{176}Therefore, the
group of CF heterozygote carriers is truly
intermedi-ate with regard to sweat electrolyte concentrations.
DISCUSSION
Our 9-year investigation of neonatal screening for
CF presented us with a special opportunity to
an-swer long-standing questions about the QPIT,
espe-cially during infancy. Indeed, the major advantage
we exploited in this study of sweat volumes and
chloride levels was the narrow, controlled age of the
infants tested. This is an important consideration
because of data indicating that sweat electrolyte
con-centrations increase with age after puberty?
Unfor-tunately, previous assessments of sweat chloride
1ev-els in CF heterozygote carriers, which studied
obligate carrier parents, did not control for the
con-founding effect of age. This variable has also skewed
the determination of the upper limit of the normal
value and accounts in part for the traditional
60-mEq/L cutoff, despite the fact that it is extremely
unusual for a sweat chloride value in a child to be
greater than 30 mEq/L, according to our data. Some
authors, in fact, have proposed two upper limit of
normal values, namely, one for children and another
for adu1ts’4’4; however, imprecision in the data
avail-able has made it difficult to designate an age for
using differential cutoff values.
The results of this study clearly indicate that the
QPIT can be performed routinely in young infants
being evaluated because of newborn screening test
results positive for CF. This confirms and extends
earlier reports involving smaller numbers of infants
seen because of either neonatal intestinal
obstruc-tion’3’17 or as part of other newborn screening
pro-grams.18 Oil’ study also differs from previous reports
by virtue of including genotyped infants at a
younger age and in a larger number than has been
reported previously.
The data shown in Table 2 have enabled us to
delineate the normal range of sweat electrolyte
con-centrations for specimens obtained during infancy
by the QPIT. Specifically, based on our data in the
184 infants with no F alleles detected, we suggest
that 30 mEq/L (calculated by rounding the mean +
2 SD) could be used as the upper limit of normal,
rather than the 60-mEq/L value traditionally
used.4’”4’ Infants who are CF heterozygote carriers,
however, may have chloride levels greater than 30
mEq/L. Therefore, we recommend that 40 mEq/L or
greater be used to distinguish infants with CF; this
represents the mean ± 3 SD value for the group of
128 CF heterozygote carriers identified by DNA
anal-ysis. During infancy, any sweat chloride value
greater than the 40 mEq/L level has a low
probabil-ity of being a true normal, assuming that the QPIT
was performed properly. Accordingly, if a young
infant has a sweat chloride level between 40 and 60
mEq/L, a diagnosis of CF is likely, and the infant
should be followed expectantly for symptoms of the
disease. Chloride levels greater than 60 mEq/L are
almost invariably diagnostic of CF.4’12’
Our results also settle the issue about CF
hetero-zygote carriers, an area of controversy for more than
3 decades. Specifically, we have demonstrated after
controffing the age factor that CF heterozygote
car-riers with the F mutation have significantly
in-creased sweat electrolyte concentrations, although
they are not high enough to be in the range
diagnos-tic of the disease. This was first proposed by di
Sant’Agnese and Powell,9 but their study and
oth-ers’#{176}”3were challenged because of overlapping
chloride levels and the potentially confounding
Sant’Agnese and Powell9 reveal an impressive
differ-ence between 97 obligate CF carriers and 117
“unselected adult controls” (mean sweat chloride
values of 32 and 17 mEq/L, respectively;
P
< .01).Therefore, the sweat electrolyte abnormality should
be considered a subclinical phenotypic manifestation
in the CF heterozygote. There is other evidence from
investigations of sweat glands1’ that supports the
conclusion that functional abnormalities occur in CF
heterozygote carriers. In addition, we have
demon-strated recently that CF heterozygote carriers have
transient hypertrypsinogenemia during the neonatal
period,2#{176}and this may be regarded as another
phe-notypic manifestation. These abnormalities are
prob-ably attributable to minor degrees of impaired
chlo-ride channel functioning in the heterozygote carrier
and the secondary consequences in electrolyte and
water transport.
A practical implication emerging from this study is
our clear indication that screening programs for CF
in newborns can perform the QPIT for diagnostic
assessment successfully when infants are a few
weeks old. This is an opportune time, because
ap-proximately 2 postnatal weeks are needed to obtain
and transport the dried blood specimen, to measure
IRT and subsequently analyze DNA, and then to
report positive test results to primary-care
physi-cians and to communicate follow-up
recommenda-tions. Another week is often needed to schedule
and complete the QPIT. It should be pointed out that
previous experience by Hammond et al.18 in the
Col-orado newborn screening program suggested that
QPIT could be performed successfully in 99% of
infants, but their mean age at diagnostic testing was
7 weeks; in addition, 22% of infants with abnormal
IRT tests were lost to follow-up. Therefore, our
re-suits confirm and extend the observations in
Cob-rado and indicate that more expeditious follow-up is
feasible. This should be advantageous for at least
three reasons: (1) in general, our experience shows
that the sooner we attempt to contact postpartum
parents about a positive newborn screening test
re-suit the more likely we are to be successful (after 1 to
2 months, infants and parents can be lost to
fob-low-up for a variety of reasons); (2) most
primary-care physicians seem to be scheduling the first
rou-tine office visits by 2 to 3 weeks of age and, therefore,
can communicate information about the screening
test results and follow-up recommendations at that
time; and (3) some clinical manifestations of CF are
evident by 4 to 6 weeks of age, such as malnutrition,
including serious hemolytic anemia associated with
vitamin E deficiency.
Our finding that infant CF heterozygote carriers
have increased levels of sweat chloride can lead to
potential difficulties when the QPIT is performed in
an infant with a high IRT level and one F8 mutant
allele. Because of the limitations and expense of
mul-timutation analysis, the sweat test is critical in
screening programs for CF in neonates. Although
false-positive results occur less commonly with CF
screening compared with other neonatal tests such as
phenylketonuria and congenital hypothyroidism, the
positive predictive value of IRT and DNA testing is
only 17%;fl consequently, there will be six infants
with false-positive results for every infant with CF
identified. Sweat test methods, technical expertise,
and quality control procedures are especially
impor-tant in this situation. It is particularly challenging to
interpret QPIT results when the infant being tested is
asymptomatic-a feature evident in 58% of the
in-fants with diagnosed CF since 1985 in our screening
research program (P.M.F. and R.E.K., unpublished
observations, 1995). In these circumstances,
evapora-tive loss of sweat liquid must be avoided to eliminate
artifactual increases in electrolyte concentrations.
Ac-cordingly, we strongly support the recent
recom-mendation by the Cystic Fibrosis Foundation to
re-quire a minimum of 75 mg of sweat,5 rather than the
50-mg criterion advised previously.5”8 Using 75 mg
as the required weight does not change any of our
conclusions, because 95% of the infants reported
herein had at least this amount. The Cystic Fibrosis
Foundation also has recommended that a second
sweat test be performed whenever the electrolyte
levels are high. Our experience would question the
necessity of this in newborn screening programs that
use DNA analysis and follow infants closely to
iden-tify and treat the characteristic manifestations of CF.
ACKNOWLEDGMENTS
This research has been supported by grant A001 5-01 from the
Cystic Fibrosis Foundation and grants DK34108 and RR03186
from the National Institutes of Health.
We thank Linda Makholm and the CF Neonatal Screening
Study Group, consisting of Elaine H. Mischler, MD; Norman Fost, MD, MPH; Ronald Gregg, PhD; Rebecca Koscik, MS; Anita Laxova; Mari Palta, PhD; Michael Rock, MD; Audrey Tluczek, RN;
Lynn Feenan, RN; Miriam Block, RN; L. J. Wei, PhD; Michael
Kosorok, PhD; and Benjamin Wilfond, MD, from the University of
Wisconsin-Madison Medical School, Madison; and W. Theodore Bruns, MD; Holly Colby, RN; Wffliam Gershan, MD; Catherine McCarthy, RN; Lee Rusakow, MD; and Mark Splaingard, MD.
REFERENCES
I. di Sant’Agnese PA, Darling RC, Perera GA, et al. Sweat electrolyte disturbances associated with childhood pancreatic disease. Am IMed.
1953;15:777-784
2. Anderson MP, Rich DP, Gregory RJ, Smith AE, Welsh MJ. Generation of
cAMP-activated chloride currents by expression of CFTR. Science. 1991;
251:679-682
3. Gibson LE, Cooke RE. A test for concentration of electrolytes in sweat in cystic fibrosis of the pancreas utilizing piocarpine by iontophoresis.
Pediatrics. 1959;23:545-549
4. di Sant’Agnese PA, Farrell PM. Neonatal and general aspects of cystic fibrosis. In: Young OS, Hicks JM, eds. The Neonate. Current Topics in
Clinical Chemistry. New York: John Wiley & Sons, 1976:199-217 5. LeGrys VA, Barlow WK, Bracey A, et al Sweat testing: sample
collec-tion and quantitative analysis. A proposed guideline. Nat! Comm Clin
Lab Stand. 1993;13:19-21
6. di Sant’Agnese PA, Davis PB. Research in cystic fibrosis. N Engi J Med. 1976;295:481-485
7. Shwachman H, Cluck L, Cone TE Jr. Dodge PR, Falkner F, Green M, eds. Cystic fibrosis. In: Current Problems in Pediatrics. Chicago, IL: Year
Book Medical Publishers, Inc; 1978;8:1-72
8. Stern RC, Boat TF, Abramowsky CR, Matthews LW, Wood RE, Doer-shuk CF. Intermediate-range sweat chloride concentration and pseudo-monas bronchitis. A cystic fibrosis variant with preservation of exocrine
pancreatic function. JAMA. 1978;239:2676-2680
9. di Sant’Agnese PA, Powell CF. The eccrine sweat defect in cystic fibrosis of the pancreas (mucoviscidosis). Ann N Y Acad Sci. 1962;93:555 10. Sproul A, Huang N. Diagnosis of heterozygosity for cystic fibrosis by
discriminatory analysis of sweat chloride distribution. J Pediatr. 1966; 69:759-770
cystic fibrosis heterozygotes. ILab Clin Med. 1988;111:511-518
12. Boat TF, Welsh MJ, Beaudet AL Cystic fibrosis. In: Scriver CR, Beaudet AL, Sly WS, Valle D, eds. The Metabolic Basis oflnherited Disease. 6th ed. New York: McGraw-Hill, mc; 1989:2649-2680
13. Anderson CM, Freeman M. Sweat test results in normal persons of
different ages compared with families with fibrocystic disease of the
pancreas. Arch Dis Child. 196035:581-587
14. Highsmith WE, Burch UI, Zhou Z, et al. A novel mutation in the cystic
fibrosis gene in patients with pulmonary disease but normal sweat
chloride concentrations. N Engl JMed. 1994331:974-980
15. Farrell PM, Mischler EM. Newborn screening for cystic fibrosis. Mv
Pediatr. 199Z39:35-70
16. FitzSimmons S. The changing epidemiology of cystic fibrosis. IPediatr.
1993;122:1-9
17. Elian E, Shwachman H, Hendren WH. Intestinal obstruction of the newborn infant. Usefulness of the sweat electrolyte test in differential diagnosis. N Engl JMed. 1961;264:13-16
18. Hammond KB, Abman SH, Sokol RJ, Accurso FJ. Efficacy of statewide neonatal screening for cystic fibrosis by assay of trypsinogen
concen-trations. N Engi JMed. 199L325:769-774
19. Rock MJ, Mischler EH, Farrell PM, Bruns WT, Hassemer DJ, Laessig
RH. Immunoreactive trypsinogen screening for cystic fibrosis: charac-terization of infants with a false-positive screening test. Pediatr
Pulmonol. 1989;6:42-48
20. Gregg RC, Wilfond BS, Farrell PM, Laxova A, Hassemer D, Mischler
EH. Application of DNA analysis in a population-screening program for neonatal diagnosis of cystic fibrosis (CF): comparison of screening
protocols. Am IHum Genet. 199352:616-626
21. Rock MJ, Mischler EH, Farrell PM, et al. Newborn screening for cystic fibrosis is complicated by age-related decline in immunoreactive trypsinogen levels. Pediatrics. 199085:1001-1007
22. Farrell PM, Aronson RA, Hoffman C, Laessig RH. Newborn screening
for cystic fibrosis in Wisconsin: first application of population-based
molecular genetics testing. Wis Med I.1994;93:415-421
23. Lobeck CC. Cystic fibrosis. In: Stanbury JB, Wyngaarden JB,
Fredrickson DS, eds. The Metabolic Basis oflnherited Disease. 3rd ed. New York: McGraw-Hill, mc; 1972:1605-1626
24. Davis PB. Sweat chloride concentration in adults with pulmonary dis-eases. Am Rev Respir Dis. 1983;128:34-37
25. Wilfond BS, Farrell PM, Laxova A, Mischler E. Severe hemolytic anemia
associated with vitamin E deficiency in infants with cystic fibrosis. Clin Pediatr (Phila). 199433:2-7
KELLOGG’S FLAKERY
. ..the zenith of America’s long obsessive coupling of food with moral rectitude
came with a Seventh-Day Adventist doctor named John Harvey Kellogg who in
1876 introduced a regime of treatments that was as bizarre as it was popular.
Possibly the two were not unconnected.
Patients who were underweight were confined to their beds with sandbags on
their abdomens and forced to eat up to twenty-six meals a day. They were not
permitted any physical exertion. Even their teeth were brushed by an attendant lest
they needlessly expend a calorie. The hypertensive were required to eat grapes and
nothing else-up to fourteen pounds of them daily. Others with less easily
dis-cernible maladies were confined to wheelchairs for months on end and fed
exper-imental foods such as gluten wafers and “a Bulgarian milk preparation known as
yogurt.”
Kellogg himseif was singular in his habits. It was his practice to dictate long
tracts on the evils of meat-eating and masturbation (the one evidently led to the
other) while seated on the toilet or while riding his bicycle in circles around the
lawn. Despite-or very possibly because of-these peculiarities, Kellogg’s
“Tem-ple of Health” thrived and grew into a substantial complex with such classy
amenities as elevators, room service, and a palm house with its own orchestra.
Among its devoted and well-heeled patrons were Teddy Roosevelt and John D.
Rockefeller.
Bryson B. Made in Ami’.rica. New York: Wm Morrow; 1994.