AJCP / Original Article
Risk Factors for Quantity Not Sufficient Sweat Collection
in Infants 3 Months or Younger
Matthew N. Collins,1 Cindy B. Brawley, MT(AMT),1 Courtney E. McCracken, PhD,2
Prabhu R. V. Shankar, MD, MS,2 Michael S. Schechter, MD, MPH,2,3
and Beverly Barton Rogers, MD1,2
From the Departments of 1Pathology and 2Pediatrics, Children’s Healthcare of Atlanta and Emory University School of Medicine, Atlanta, GA,
and 3Virginia Commonwealth University, Children’s Hospital of Richmond at VCU, Richmond, VA.
Key Words: Cystic fibrosis; Sweat collection; QNS; Infant
Am J Clin Pathol July 2014;142:72-75
DOI: 10.1309/AJCPLHG2BUVBT5LY
ABSTRACT
Objectives: The purpose is to identify demographic
characteristics associated with a quantity not sufficient
(QNS) sweat collection in infants 3 months or younger.
Methods: History of premature birth, infant race and sex,
gestational age at delivery, and weight of the infant were
compared with QNS collection.
Results: Of 221 sweat collections from 197 infants, 25 were
QNS. Infant weight less than 3 kg and history of prematurity
were associated with QNS collection (P < .001). Thirteen
(30.2%) of 43 infants weighing less than 3 kg had QNS
collections compared with 12 (7.9%) of 151 infants 3 kg
or more. Twelve (46.2%) premature infants had QNS
collections compared with 13 (7.6%) term infants. Lower
birth gestational age and corrected gestational age were
associated with QNS collections. Six (86%) of seven infants
who weighed less than 3 kg, had a history of prematurity, and
were more than 54 days old at testing had a QNS result. Sex
and race did not correlate with QNS collections.
Conclusions: Weight less than 3 kg and history of prematurity
are associated with an increased chance of QNS sweat
collections.
72
72
Am J Clin Pathol 2014;142:72-75
DOI: 10.1309/AJCPLHG2BUVBT5LY
Increased levels of chloride in sweat have been the principal means of diagnosing cystic fibrosis (CF) for more than
50 years, even with advances in genetic testing.1 In 2004, the
Centers for Disease Control and Prevention recommended
that states should consider including newborn screening for
CF due to the long-term benefits of early recognition. By
2010, newborn screening had been adopted in all 50 of the
United States, resulting in earlier intervention to decrease the
nutritional and pulmonary manifestations.2
CF newborn screening algorithms vary by state. All begin
by assessing immunoreactive trypsinogen (IRT) levels in the
blood, which are elevated in patients with CF, but a single
elevated IRT level is nonspecific, so a second-tier test is then
performed. Depending on the state, the second-tier test may be
a repeat IRT level or genetic screen. The screening procedure
has a relatively low positive predictive value, so confirmation via sweat test is necessary. Traditionally, concentrations
of sweat chloride below 40 mmol/L are considered negative
for CF, above 60 mmol/L indicates a positive test, and levels
between 40 and 60 mmol/L are indeterminate. However, with
increasing numbers of young infants being tested due to newborn screening, it has been learned that a sweat chloride concentration of 30 mmol/L is not uncommonly found in infants
with CF aged 3 months or younger.
The validity of sweat chloride testing to diagnose CF
requires that measurements be from sweat produced from
sweat glands that are maximally stimulated by pilocarpine
iontophoresis, so accurate determination requires that a sufficient quantity of sweat be collected within 30 minutes of the
start of stimulation. When this quantity cannot be collected
in the allotted time, the result is referred to as a quantity not
sufficient (QNS). As recommended by the Cystic Fibrosis
© American Society for Clinical Pathology
AJCP / Original Article
Foundation, institutions should maintain a QNS rate of 5% or
less for infants older than 3 months and 10% or less for infants
3 months or younger.3
Collecting sufficient sweat to meet the Cystic Fibrosis
Foundation guidelines can be difficult, particularly for infants
3 months or younger. Factors associated with an increased
QNS rate in young infants have been reported to include
weight at collection less than 2 kg, preterm birth, African
American race, corrected postmenstrual age less than 36 to
39 weeks,4,5 type of collection method,6,7 and location of collection center.5 To improve any process, it is imperative to
understand the factors affecting outcome. The studies done
to date to determine the patient demographic effect on QNS
rate were retrospective, were multicenter, or occurred over
many years; any of these factors can result in variability in
outcomes. To our knowledge, our study is the first populationbased, prospective study performed at one institution to assess
the impact of patient demographic information on the QNS
rate in infants 3 months or younger.
Materials and Methods
The study was approved by the institutional review board
of Children’s Healthcare of Atlanta (CHOA). For a 1-year
period beginning July 1, 2011, and ending June 30, 2012, we
collected data on infants 3 months or younger undergoing
sweat chloride determination. Sweat collection occurred at
two CHOA hospitals in the health care system, with technicians who had received uniform training in sweat collection.
The sweat collection procedure used was pilocarpine gel iontophoresis collected in Macroduct coils (Wescor Biomedical
Systems, Logan, UT) in accordance with the National Committee for Clinical Laboratory Standards. Sweat collection
was attempted in two separate locations on the infant, usually
the arms but occasionally on the legs. A sweat collection was
deemed QNS if less than 15 µL was obtained from both collection sites.
At the time of sweat collection, the phlebotomists
weighed the infant and asked the parent or guardian standardized questions. Variables recorded were age, history of prematurity, weight, and race. In addition to information obtained
at the time of sweat collection, the gestational age at birth and
the birth weight of those infants who were referred through
the newborn screening program were provided by the Georgia
Department of Public Health.
Statistical analysis was conducted using SAS version
9.2 (SAS Institute, Cary, NC) and R (www.R-project.org).8
Statistical significance was assessed using a significance
level of .05. Patient demographics were summarized using
mean and standard deviations for continuous variables (age
[days], adjusted age [weeks], gestational age [weeks], and
© American Society for Clinical Pathology
birth weight [kg]) or counts and percentages for categorical
variables (sex, prematurity status, race, weight category,
and age category). The weight and age cutoff values used to
define the weight and age category variables were derived
using recursive partitioning and regression trees (rpart) in
R. For weight, less than 3 kg was identified as a splitting
point, and for age (in days), 54 days or less was identified as
a splitting point. If a continuous variable was nonnormally
distributed, the median and range were presented instead.
Two-sample t tests were used to compare the mean values
between infants who had sufficient sweat collection for chloride analysis with those who were reported as sweat QNS.
If the continuous demographic was nonnormal, a Wilcoxon
rank sum test was used instead. Categorical variables were
compared across QNS groups using c2 tests or the Fisher
exact test when the cell counts were small (n ≤ 5). Multivariate logistic regression was used to identify predictors
of failed sweat tests (QNS). All predictors were initially
included in the base model (weight, age, sex, race, prematurity status, etc), and backward elimination with the inclusion criterion of P < .10 was used to select the final reduced
model. Statistical analysis was performed on the first infant
visit; repeated visits were not analyzed statistically.
Results
Data on a total of 221 sweat collections were analyzed
on 197 infants 3 months or younger. Weight and sex were
not recorded for three infants. Of the 197 infants on whom
sweat collection was performed, 20 infants had collections
on two separate occasions and one infant had three collections. The median age of infants with a QNS collection was
20 days, compared with 28 days when sweat collection was
sufficient (P = .09) ❚Table 1❚. The median gestational age at
birth in infants with QNS collections was 36 weeks (range,
24-40 weeks) compared with 39 weeks (range, 27-42
weeks) for infants without a QNS collection (P < .001).
When the age at sweat collection was assessed in relation
to the gestational age at delivery, the median adjusted (or
corrected) age of the infants who had a QNS result was 39.7
weeks’ gestation compared with 41.9 weeks for the infants
who did not have QNS collections (P < .001). The adjusted
age of the infants is defined as gestational age at delivery
plus postdelivery age. Lower birth weight and weight less
than 3 kg at the time of collection were also associated with
QNS collections (P < .001 for each parameter). Of the 43
infants who were less than 3 kg at sweat collection, the
QNS rate was 30.2% compared with 7.9% for those infants
weighing 3 kg or more. Infants with a history of prematurity provided by the parents were more likely to have a
QNS collection (P < .001), with a QNS rate of 46% in the
73
Am J Clin Pathol 2014;142:72-75
DOI: 10.1309/AJCPLHG2BUVBT5LY
73
73
Collins et al / Risk Factors for QNS Sweat Collection in Infants
❚Table 1❚
Patient Demographics by Quantity Not Sufficient (QNS) Status
Not QNS (n = 172)
QNS (n = 25)
P Value
Demographic
Overall (N = 197)
Age, median (range), d
Gestational age at birth, median (range), wkª
Adjusted age at collection, median (range), wkª
Birth weight, median (range), kgª
Weight, mean ± SD, kg
Weight category, No. (%)ª
<3 kg
≥3 kg
Sex, No. (%)ª
Female
Male
Prematurity, No. (%)
No
Yes
Race, No. (%)
African American
Not African American
21.0 (4.0-90.0)
28.0 (10.0-90.0)
20.0 (4.0-90.0)
.086
39.0 (24.0-42.0)
39.0 (27.0-42.0)
36.0 (24.0-40.0)
<.001
41.6 (35.9-51.9)
41.9 (36.4-51.9)
39.7 (35.9-43.0)
<.001
3.5 (2.0-6.4)
3.6 (0.9-6.4)
3.0 (2.0-4.5)
<.001
3.6 ± 0.9
3.7 ± 0.9
3.0 ± 0.7
<.001
43 (22.2)
30 (69.8)
13 (30.2)
<.001
151 (77.8)
139 (92.1)
12 (7.9)
93 (47.9)
84 (90.3)
9 (9.7)
.201
101 (52.1)
85 (84.2)
16 (15.8)
171 (86.8)
158 (92.4)
13 (7.6)
<.001
26 (13.2)
14 (53.9)
12 (46.2)
60 (30.5)
50 (83.3)
10 (16.7)
.267
137 (69.5)
122 (89.0)
15 (11.0)
ª Indicates missing data: gestational age and birth weight missing on 61 (31%) of 197.
premature group compared with 8% in infants without a
history of prematurity.
Six (86%) of seven infants who weighed less than 3 kg
had a history of prematurity and were more than 54 days old
at the time of collection had QNS results. Infant age of more
than 54 days did not maintain statistical significance as an
independent predictor of QNS collection following multivariate logistic regression controlling for prematurity and weight
(data not shown). Of this subgroup of infants with very high
QNS rates, the gestational age at birth available for three
infants was 25, 29, and 31 weeks. Demographic features not
associated with a QNS collection were infant’s sex and race
(including African American).
Of the 21 infants with repeat sweat collections and complete information, 10 had a QNS result on the first collection,
with a sufficient collection on the second. The number of days
between the collections varied from 3 to 56 (median, 24 days),
and the weight gain between the QNS collection and sufficient collection increased by 0.09 to 2.27 kg (median, 0.84
kg). Eight infants had repeat sweat collections that remained
QNS on the second attempt. The number of days between the
collections of these infants varied from 3 to 49 days (median,
20 days), and the weight gain between QNS collections was
0.03 to 1.62 kg (median, 0.34 kg). There were no differences
in the collection technique or the weight and age delineated
in the chart or by the phlebotomists. Specifically, edema and
skin disease were absent for all sweat collections regardless
of whether they were sufficient. While the median weight
gain was greater for infants who had a QNS result on the first
attempt and a sufficient collection on the second compared
with those who were QNS on both, there was sufficient variability in infant weight to indicate that weight, by itself, was
not the sole predictor of QNS in this group of patients.
74
74
Am J Clin Pathol 2014;142:72-75
DOI: 10.1309/AJCPLHG2BUVBT5LY
Discussion
To our knowledge, this is the first prospective study
performed at one institution to assess the demographic factors
affecting sweat collection for sweat chloride determination in
infants 3 months or younger. We found that an infant weight
less than 3 kg, a history of prematurity given by the parents
at the time of sweat collection, younger gestational age at
delivery, and a younger corrected gestational age at sweat
collection were the primary predictors of QNS collection. In
addition, infants having sweat tests at more than 54 days of
age were more likely to have had a QNS collection than those
who were younger, but this appears to be due to the extreme
prematurity that led to their delay in testing. Our results both
support and differ from those of others. Eng et al4 performed
a retrospective study assessing the characteristics of infants 6
weeks and younger associated with sufficient or insufficient
sweat collections. The study assessed 119 sweat tests on 103
infants performed over a 10-year period using pilocarpine
iontophoresis, which was the method used in our study.
Insufficient sweat was obtained in 26.2% of patients, with
demographic risk factors being weight less than 2 kg, African
American race, preterm birth, and postmenstrual age less than
36 weeks. Our study differs from that of Eng et al4 in several
aspects. The first is that the population we chose to examine
was 3 months or younger rather than 6 weeks or younger. This
is because the Cystic Fibrosis Foundation highlights this age
cutoff with an associated increase in the expected QNS rate.
Kleyn et al5 assessed predictors of QNS results on patients
from a retrospective, population-based analysis of results of
four testing centers using two different methods across the
state of Michigan. In total, 315 infants had an attempted sweat
collection, and 97% were 3 months or younger. The authors
found the following demographics to be associated with a
© American Society for Clinical Pathology
AJCP / Original Article
QNS result: birth weight less than 2.5 kg, gestational age at
birth less than 37 weeks, and corrected gestational age less
than 39 weeks. The specific CF care center where the collection was performed was also associated with a risk of a QNS
result. Our prospective study design, which was conducted
at one health care system and assessed infants 3 months or
younger, is very different from that of Kleyn et al,5 but some
of the same themes repeat. The repeating themes are a history
of prematurity and low birth weight as risk factors for a QNS
sweat collection. Kleyn et al5 also identified that, while black
race was a risk factor for a QNS result, when combined with
birth weight using a multivariate analysis, black race was not
an independent predictor of a QNS result. We found a QNS
rate of 16.7% in African American infants compared with
11.0% in others; we did not find this difference to be statistically significant but may have been underpowered to do so.
We began our study with two practical questions: (1) If
an infant 3 months or younger is referred to our organization
for sweat chloride collection, which infants are most likely to
have a successful collection compared with those who have
a QNS result? (2) Are there demographic data that can be
incorporated by the Cystic Fibrosis Foundation to essentially
normalize the QNS rate for infants 3 months or younger such
that a center’s QNS rate can be reported in conjunction with
the patient population? In other words, all patient populations
may not be made equal, so should the expected QNS result
reflect the difference in patient population?
To answer the first question, we have identified the
following parameters associated with an increased risk for
a QNS result that can easily be identified by the collection
site or referring physician: weight less than 3 kg at collection
(30.2% vs 7.9% QNS if ≥3 kg) and history of prematurity
(46.2% vs 7.6% QNS if not premature). The combination of
weight less than 3 kg, history of prematurity, and age more
than 54 days resulted in a QNS result for six (86%) of seven,
which reflects an infant who is severely undergrown.
The second question, whether the QNS rate should be
“corrected” in some manner for patient demographics before
being reported to the Cystic Fibrosis Foundation, is one that
could use deliberation. Those centers that test patients with
low rates of prematurity and relatively higher weights at testing should have lower QNS rates compared with centers that
are enriched for those patient populations. Another option
would be to exclude patients who have a very high risk of
QNS collection from the QNS data reports.
One of the main limitations of our study is that it was
done in one center, which reduces the demographic and testing differences that are present throughout the United States.
Generalizability to systems that are more heterogeneous and
complex, such as described in Michigan,5 may therefore be
© American Society for Clinical Pathology
limited. However, we believe that this reduction in test technique variation allows us to focus on differences in QNS rates
that result from variations in patient demographics, and the
Georgia population is a relatively diverse one.
In summary, we confirm what others have found—that
there is a risk of a QNS sweat collection for patients undergoing diagnostic testing for CF if the infant weighs less than 3 kg
and has a history of prematurity, which in our study resulted in
a QNS collection approximately 30% and 46% of collections,
respectively. In addition, the “trifecta” of weight less than 3
kg, history of prematurity, and age at collection more than 54
days resulted in an 86% rate of QNS. Race is not a predictor
of a QNS result using the Macroduct collection methods currently in use. We recommend that these parameters be taken
into consideration when referring an infant for sweat chloride
testing and when comparisons are made between CF center
laboratories performing sweat testing.
Address reprint requests to Dr Rogers: Dept of Pathology,
Children’s Healthcare of Atlanta, 1405 Clifton Rd NE, Atlanta,
GA 30322; [email protected].
Acknowledgments: The authors acknowledge the Georgia
Department of Public Health, which provided information about
gestational age at delivery. The authors also acknowledge Brigitte
Thompson and Craig Jordan, phlebotomy supervisors, and the
phlebotomists who performed sweat collections and compiled data
so this study could be done.
References
1. LeGrys VA, Yankaskas JR, Tuittell LM, et al. Diagnostic
sweat testing: the Cystic Fibrosis Foundation Guidelines.
J Pediatr. 2007;151:85-89.
2. Wagener JS, Zemanick ET, Sontag MK. Newborn screening
for cystic fibrosis. Curr Opin Pediatr. 2012;24:329-335.
3. LeGrys VA, McColley SA, Li A, et al. The need for quality
improvement in sweat testing infants after newborn screening
for cystic fibrosis. J Pediatr. 2010;157:1035-1037.
4. Eng W, LeGrys VA, Schechter MS, et al. Sweat-testing in
preterm and full-term infants less than 6 weeks of age. Pediatr
Pulmonol. 2005;40:64-67.
5. Kleyn M, Korzeniewski S, Grigorescu V, et al. Predictors
of insufficient sweat production during confirmatory testing
for cystic fibrosis. Pediatr Pulmonol. 2011;46:23-30.
6. Laguna TA, Lin A, Wang Q, et al. Comparison of
quantitative sweat chloride methods after positive newborn
screen for cystic fibrosis. Pediatr Pulmonol. 2012;47:736-742.
7. LeGrys VA. Assessment of sweat-testing practices for
the diagnosis of cystic fibrosis. Arch Pathol Lab Med.
2001;125:1420-1424.
8. R Development Core Team. R: A Language and Environment
for Statistical Computing. Vienna, Austria: R Foundation for
Statistical Computing; 2011.
75
Am J Clin Pathol 2014;142:72-75
DOI: 10.1309/AJCPLHG2BUVBT5LY
75
75