β‐adrenergic sweat test in children with inconclusive cystic fibrosis diagnosis: Do we need new reference ranges?

Abstract

Background

Investigating inconclusive cystic fibrosis (CF) diagnosis in children is difficult without advanced cystic fibrosis transmembrane conductance regulator (CFTR) function tests. This study investigated the utility of beta (β)‐adrenergic sweat test to exclude CF in participants with inconclusive diagnosis (CF suspects) in South Africa.

Methods

β‐adrenergic sweat test and sweat chloride tests (SCT) were performed simultaneously in CF suspects and adult controls (healthy, CFTR heterozygotes and CF). Cholinergic and β‐adrenergic induced sweat rate was measured by evaporimetry (transepithelial water loss [TEWL]: g H₂O/m²/h) following intradermal injections. Next‐generation sequencing of CFTR was performed in CF suspects. CF diagnosis was defined by genotype.

Results

Thirty‐seven controls (10 healthy, 14 CF, 13 CFTR heterozygotes) and 32 CF suspects (26 children; 6 adults) were enrolled. Six were excluded from formal analyses due to β‐adrenergic sweat test failure. In adults, evaporimetry was superior to SCT for diagnosis of CF with β‐adrenergic:cholinergic ratio TEWL ≤ 0.05 achieving 100% sensitivity and specificity. Twenty‐two CF suspect children (age range: 3.4–15.6 years) completed β‐adrenergic sweat testing of which none had CF confirmed by genotyping: β‐adrenergic:cholinergic ratio > 0.05 successfully excluded CF in all but one child who was CFTR heterozygous. Median peak β‐adrenergic TEWL and β‐adrenergic:cholinergic ratio in CFTR negative and CFTR heterozygous children was significantly lower than adult controls.

Conclusion

β‐adrenergic sweat test is more accurate than SCT for excluding CF in children with inconclusive diagnosis. Established reference ranges for β‐adrenergic sweat test may not be suitable for children due to lower β‐adrenergic sweat secretion compared to adults.

1. INTRODUCTION

The diagnosis of cystic fibrosis (CF) is confirmed by demonstration of reduced or absent cystic fibrosis transmembrane conductance regulator (CFTR) function and/or identification of two pathogenic CFTR variants. ¹ Reduced or absent CFTR activity in sweat glands results in reduced chloride reabsorption by epithelial cells and increased sweat chloride concentration (≥60 mmol/L), which is the hallmark of CF diagnosis. Quantitative pilocarpine iontophoresis cholinergic sweat gland stimulation is the gold standard in clinical practice to diagnose CF. ¹ However, access to sweat chloride test (SCT) is scarce and increasingly unaffordable in low‐middle income countries (LMIC) where the clinical presentation of CF in children is poorly distinguished from more common conditions such as non‐CF bronchiectasis, malnutrition, tuberculosis and human immunodeficiency virus infection. ² Although the SCT in conjunction with supportive genotyping is accurate in most cases of CF, SCT in the intermediate range (30–59 mmol/L) without supportive genotyping may not always distinguish CF from CFTR‐related disorders (CFTR‐RD) associated with residual CFTR activity, from subjects without CF. ¹ Furthermore, genotyping alone is unable to confirm or exclude CF in the presence of CFTR variants of variable or unknown clinical significance. In such cases, adjunctive measurement of CFTR function can aid the diagnosis or exclusion of CF or CFTR‐RD.

The SCT has several limitations with respect to interpretation of borderline sweat chloride results associated with inconclusive CFTR genotyping. False‐positive and false‐negative results may be the result of insufficient sweat volume or erroneous collection techniques, or other rare underlying medical conditions. ³ Furthermore, age‐related variation of sweat electrolytes due to sweat gland maturation; and within and between test variability makes interpretation of sweat chloride levels in the intermediate or borderline zones difficult. ⁴ , ⁵ Several techniques to measure CFTR activity have been developed and include the electrophysiological tests nasal potential difference (NPD) and intestinal current measurement (ICM), and the β‐adrenergic sweat gland stimulation test. ³ , ⁶ , ⁷ Of these, NPD and ICM are most widely described in research settings and clinical practice. Both, however, are expensive and require highly specialized technical skill and expertize which are not available in LMIC, including South Africa.

The novel CFTR‐dependent β‐adrenergic sweat test (BAST) measures rate of sweat production by evaporimetry or sweat droplet imaging. ⁶ , ⁸ , ⁹ , ¹⁰ BAST is easier to perform and less invasive than NPD and ICM and demonstrates a greater ability than SCT to discriminate between CF, CFTR‐RD, CFTR heterozygotes and subjects without CFTR variants. ⁶ , ¹¹ The BAST is safe, validated, and well tolerated in adults but the diagnostic utility, feasibility and safety in children and populations with diverse ethnicity is not well established. ¹² We, therefore, conducted a pilot study of BAST in a South African cohort of predominantly children with inconclusive CF diagnosis to evaluate the safety, feasibility, and utility of BAST to exclude CF or CFTR‐RD in an LMIC setting where newborn screening for CF, NPD and ICM are not available.

2. METHODS

2.1. Study design and setting

We prospectively recruited a series of consenting participants fitting the following groups at Red Cross War Memorial Children's Hospital, Cape Town, South Africa, between February 2020 and April 2021:

• (1)Healthy adults ≥ 18 years (healthy controls)
• (2)Adults with confirmed CF diagnosis and two pathogenic CFTR variants (CF).
• (3)Adults who were a parent of a child with confirmed CF diagnosis and two pathogenic CFTR variants (CFTR heterozygous).
• (4)Children (≥3 years age) and adults ≥18 years with suspected CF based on clinical symptoms and inconclusive SCT and/or previous CFTR genotype (CF suspects).

2.2. Study procedures

Study participants each underwent simultaneous SCT on the right forearm and BAST on the left forearm (Supporting Information: Figures E1 and E2) according to standardized protocols (see online Supporting Information Material). SCT was conducted with the Macroduct® sweat stimulation and collection system (ELITechGroup) and sweat chloride analysis with Sherwood® Mk II Chloride Analyser 9265 (Sherwood Scientific Limited).

Sweat rate was measured using two evaporimetry probes (CyberDerm RG‐1; Dasylab) placed onto the forearm skin, using a previously described technique. ⁶ , ¹² The reference probe was positioned medially and the measuring probe laterally (Online Supporting Information Material: Figure E1). Evaporimetry was expressed as transepithelial water loss (TEWL, grams of water loss/m²/h). The β‐adrenergic sweat secretion test was performed as described by Quinton et al. (2012) by intradermal injection of atropine sulfate (0.2 ml, 44 µg/ml) below the reference probe, followed by sequential intradermal injection of carbachol (0.1 ml, 0.1 µg/ml), atropine sulfate (0.2 ml, 44 µg/ml), and finally 0.2 ml β‐adrenergic cocktail containing atropine sulfate (8.0 µg), isoproterenol (4.4 µg) and aminophylline (0.84 mg) below the measuring probe. ⁶ Stock solutions were prepared by a research pharmacist at least 48 h before participant testing and stored at 4°C to ensure drug stability. ¹³

Peak/Δcholinergic, peak/Δβ‐adrenergic TEWL and β‐adrenergic:cholinergic ratio were recorded and calculated in each participant. A previously described “modified” BAST protocol and analysis measuring only peak/Δβ‐adrenergic TEWL was applied in children <6 years age where the full protocol could not be followed, or if cholinergic secretion in response to carbachol was not detected for any reason. ¹² Details of the BAST technique, “modified” BAST protocol or analysis, standard operating procedures and output measurements are provided in the online Supporting Information Data Material.

2.3. CFTR variant analysis

All CF suspects had whole blood samples collected for next‐generation sequencing of CFTR by Invitae Laboratories. ¹⁴ CFTR variants in parents of children were determined by their child's genotype. Healthy adult volunteers were not tested for CFTR variants. Participants were stratified for analysis into three final CFTR genotype categories: (A) CF or CFTR‐RD; (B) CFTR heterozygous and (C) CFTR negative/healthy control (no CFTR variant identified or healthy adult control). CF and CFTR‐RD were grouped together based on previous reporting that BAST was unable to differentiate CF from CFTR‐RD in children. ¹² The Wong and Baker faces pain rating scale was used to measure self‐reported pain before and 2 min after the BAST for all participants. ¹⁵ Participants were monitored throughout and for at least 5 min after BAST and SCT for any adverse events.

2.4. Clinical data collection in CF suspects

Clinical information of CF suspects, including age, ethnicity, sex, and symptoms, were documented by the participant questionnaire and from the medical records. Physical examination of CF suspects including weight and height measurements was performed at the study visit and analysed with body mass index (BMI) or BMI‐z scores according to the World Health Organisation (WHO) reference equations. Routine CF‐related investigations including chest radiographs, computed tomography (CT) chest scan, the microbiology of respiratory samples, fecal elastase, and spirometry were extracted from the medical records where available. Prebronchodilator spirometry values were calculated according to Global Lung Initiative reference equations. ¹⁶ Previous SCT results collected by Gibson and Cooke method were documented and the mean sweat chloride value (mmol/L) for each participant was calculated.

2.5. Statistical analysis

Reported measures of centrality and spread were guided by whether distributions were approximately normal. Groups of individuals (by recruitment, diagnosis, ethnicity, sex, or age) were compared using χ ² or Fisher's exact (categorical variables), Kruskal–Wallis (medians of continuous variables) or analysis of variance (means of continuous variables) tests. Pearson correlation coefficients described associations of age and sex between SCT, and BAST results. Data in adults and children were analysed and presented separately due to the absence of BAST reference data in children. Receiver operating characteristic curves were plotted to describe and compare the diagnostic accuracy of BAST and SCT in adults ≥18 years to exclude CF/CFTR‐RD, and to select the optimal sensitivity and specificity cut‐offs for the diagnosis of CF/CFTR‐RD. For analysis and comparisons, final CFTR genotype diagnosis category was used as the diagnostic standard due to the absence of BAST reference data in the local study population. All data analyses were conducted using SPSS Windows 27.0 (IBM SPSS Inc.) and Statistica Windows Version 13.0 (TIBCO Software Inc.).

2.6. Ethical considerations

The study was approved by the University of Cape Town's Human Research Ethics Committee (HREC 032/2019) and the South African Health Products Regulatory Authority (SAHPRA trial ref. 20190602). Signed informed consent and assent (children aged 7–17 years) were obtained from all participants. Ethical approval to recruit healthy control children was not obtained.

3. RESULTS

3.1. Study population characteristics

Ten healthy adults, 14 adults with CF (10 pancreatic insufficient; four pancreatic sufficient), 13 adult CFTR heterozygotes, and 32 CF suspects (26 children and six adults) were recruited between February 2020 and April 2021. Detailed clinical and CF diagnostic investigations including genotype and evaporimetry results of all participants are presented in Table 1.

Table 1.

Demographic characteristics and investigations of recruited subjects with inconclusive CF diagnosis (CF suspects) ranked by age

Study no.	Age (years)	Sex	Ethnicity	Sweat chloride tests		FEV1pp	Sputum microbiology	Fecal elastase mean (µg/g)	CFTR variants	CFTR genotype category	Evaporimetry
				Mean previous Cl (mmol/L)	Study visit Cl (mmol/L)						ΔCholinergic	Δβ‐adren	Ratio β‐adren:Chol
1	3.4	M	Mixed	65	14			500	p. Val520Ilea/_	CFTR Hetero	23.39	1.84	0.079
2	4.1	M	Mixed	65	44			393	_/_	CFTR negative	75.02	21.6	0.288
3b	5.7	M	Mixed	43		112			3272‐26A>G1/_	CFTR Hetero
4	6.0	F	Mixed	68	35	56	PA	321	c.1210‐34TG[11]T[5]/_	CFTR Hetero	12.21	0.22	0.018
5b	6.3	M	Mixed	70	32	91		500	p. Ala280Serd/_	CFTR Hetero
6	6.5	F	Black African	48	20	75			_/_	CFTR negative	20.98	8.93	0.426
7c	6.6	F	Caucasian	50	48	94		438	3272‐26A>G1/12789 + 2insAd	CFTR‐RD		2.94
8	7.5	F	Mixed	75	66	70		251	_/_	CFTR negative	43.74	2.64	0.060
9	7.7	M	Black African	83	31	53		133	_/_	CFTR negative	78.53	20.71	0.264
10	7.9	F	Black African	65	15	109			_/_	CFTR negative	46.98	23.32	0.496
11	8.1	F	Mixed	53	45	106		437	_/_	CFTR negative	25.87	14.35	0.555
12	8.4	M	Mixed	63	36	78		476	_/_	CFTR negative	81.28	53.7	0.661
13d	8.7	F	Black African	60	50	56		135	_/_	CFTR negative		19.8
14	8.9	F	Mixed	56	12	100		257	_/_	CFTR negative	29.34	2.03	0.069
15	9.5	F	Black African	43	37	60		138	_/_	CFTR negative	45.08	8.2	0.182
16	10.7	M	Mixed	45	50	61	H. Influenzae		_/_	CFTR negative	72.23	59.62	0.825
17	11.0	M	Black African	74	12	73		425	_/_	CFTR negative	85.98	36.71	0.427
18	11.0	F	Mixed	51	60	67		500	_/_	CFTR negative	71.52	35.1	0.491
19	11.4	F	Mixed	67	35	85		386	_/_	CFTR negative	29.9	38.49	1.287
20	11.6	M	Mixed	84	61	80		30	_/_	CFTR negative	22.85	7.06	0.309
21b	12.3	F	Mixed	71	47	76	MSSA, H. Influenzae	369	_/_	CFTR negative
22	12.8	M	Mixed	60	14	93		462	_/_	CFTR negative	78.19	32.19	0.412
23	13.7	F	Black African	20	22	23	PA, Aspergillus fumigatus		_/_	CFTR negative	16.08	2.63	0.164
24b	14.0	F	Mixed	74	22	47		141	_/_	CFTR negative
25	14.1	F	Mixed	59	62	87	MSSA	500	_/_	CFTR negative	27.41	22.98	0.838
26	15.6	F	Mixed	50	27	73		420	_/_	CFTR negative	71.06	42.99	0.605
27	18.1	F	Mixed	69	63	71		470	_/_	CFTR negative	25.54	29.38	1.150
28	20.0	M	Mixed	87	67	74		368	_/_	CFTR negative	60.88	25.51	0.419
29	20.8	M	Mixed	44	26	60			_/_	CFTR negative	91.92	59.17	0.644
30	24.2	M	Black African	51	44	56		233	_/_	CFTR negative	77.67	78.63	1.012
31	30.2	M	Black African	93	90	61	MSSA	222	_/_	CFTR negative	37.41	37.24	0.995
32	36.0	M	Mixed	47	47	58	PA		p. Phe508del/3849 + 10kbC‐>T	CF	64.62	−0.6	0.000

Abbreviatons: CF/CFTR‐RD, cystic fibrosis/Cystic fibrosis‐related disorder; CFTR, cystic fibrosis transmembrane regulator protein; FEV1_pp, forced expiratory volume in first‐second percent predicted; Hetero, heterozygous; MSSA, methicillin sensitive staphylococcus aureus; PA, Pseudomonas aeruginosa.

^a CFTR variant of uncertain significance.

^b Participants were excluded from analysis due to technical failure with the BAST.

^c A modified BAST protocol/analysis was used.

^d Failed cholinergic response.

The median age of CF suspects was 10.9 years (range: 3.5–36) with a similar distribution of sexes (47% male). The majority were of mixed ethnicity (69%) with fewer black Africans (28%) and Caucasians (3%).

The majority of CF suspects (29/32, 90%) had chronic respiratory symptoms with abnormal chest radiographs (13/31, 48%) or CT chest scans (12/13, 92%). Forced expiratory volume in first–second percent predicted (FEV₁pp) was <80 in 20/30 (67%) and a recognized CF‐associated bacterium was isolated from sputum in 7/31 (26%) participants. Low fecal elastase was recorded in 4/21 (19%), and 14/32 (44%) had a BMIz < −1.0 or BMI < 18.5. After extensive CFTR testing of CF suspects, CF was diagnosed genetically in one adult, CFTR‐RD in one child and four children identified as CFTR heterozygotes (Table 1).

No serious adverse events were recorded in any participants who completed the BAST. Minor adverse events occurred in five participants: one healthy adult had anxiety, one child had inconsolable crying, and three participants (one adult and two children) had redness at the injection site, which resolved without intervention.

3.2. BAST in adults

Forty‐three adults underwent BAST of which 41 were included in analyses and grouped into the following final CFTR genotype categories: 14 CF; 12 CFTR heterozygotes; and 15 CFTR negative/healthy controls. Receiver operator curve analysis in adults is presented in Table 2. Evaporimetry performed better than SCT ≥ 60 mmol/L for diagnosis and exclusion of CF. Δβ‐adrenergic ≤ 4.5 TEWL and β‐adrenergic:cholinergic ratio ≤ 0.05 achieved 100% sensitivity and specificity for genetic CF diagnosis, thus confirming the 0.05 cut‐off to be optimal for confirmation and exclusion of CF/CFTR‐RD in this adult cohort.

Table 2.

Receiver operator curve analysis for sweat chloride test and β‐adrenergic sweat test for CF/CFRD diagnosis in adults

	Sensitivity (%)	Specificity (%)	PPV (%)	NPV (%)
SCT ≥ 60 mmol/L	85.7	88.0	80.0	91.7
Δβ‐adrenergic ≤ 4.5 TEWL	100.0	96.3	93.3	100.0
β‐adrenergic:cholinergic ratio ≤ 0.00	85.7	100.0	100.0	92.3
β‐adrenergic:cholinergic ratio ≤ 0.002	92.9	100.0	100.0	96.0
β‐adrenergic:cholinergic ratio ≤ 0.050	100.0	100.0	100.0	100.0

Note: SCT AUC = 0.969 (95% CI: 0.923, 1.000) p < 0.001; Δβ‐adrenergic AUC = 0.997 (95% CI: 0.987, 1.00) p < 0.001; β‐adrenergic:cholinergic ratio AUC = 1.00 (95% CI: 1.00, 1.00) p < 0.001

Abbreviations: NPV, negative predictive value; PPV, positive predictive value; SVCT, sweat chloride test.

Two participants' (one CF and one CFTR heterozygote) BAST results were excluded due technical failure and the modified BAST analysis was applied in three females due to failed cholinergic secretion; two of these three females had insufficient sweat volume collected with the simultaneous SCT. BAST correctly diagnosed one CF suspect adult (participant no. 32) with intermediate SCT and p. Phe508del/3849 + 10kbC‐>T variants. False positive SCT was observed in three adults who were CFTR negative, including one male (participant no. 31) who was human immunodeficiency virus (HIV) infected. Median Δβ‐adrenergic TEWL and median β‐adrenergic:cholinergic ratio in CF, CFTR heterozygotes and CFTR negative/healthy control groups is presented in Figure 1. Median β‐adrenergic:cholinergic ratio in CFTR heterozygotes was half that of CFTR negative adults but the difference did not reach statistical significance.

Figure 1.

Peak β‐adrenergic and β‐adrenergic:cholinergic ratio indices plotted by CFTR diagnosis category for adults and children. Boxes indicate first to third quartiles, the dividing line the median, whiskers the nonoutlier range and markers the outliers. p Values indicate differences between adults and children with the Mann–Whitney U test. [Color figure can be viewed at wileyonlinelibrary.com]

3.3. BAST in children

Twenty‐six CF suspect children underwent BAST of which 22 (one CFTR‐RD; two CFTR heterozygotes and 19 CFTR negative) were included in BAST analyses. Four children (two CFTR negative and two CFTR heterozygotes) were excluded due to BAST failure: three due to movement artifact rendering evaporimetry tracing unreadable, and one child refused testing all together. The modified BAST protocol or analysis was applied in two children. No children were diagnosed with CF based on CFTR testing and one child (participant no. 7) was diagnosed with CFTR‐RD based on intermediate SCT and two CFTR variants (3272‐26A>G1 and 12789 + 2insA). False positive SCT was observed in three children and intermediate SCT was observed in 12 children of which one had CFTR‐RD and 11 were CFTR negative. Of the 22 children tested by BAST, Δcholinergic TEWL was <30 in nine (41%), Δβ‐adrenergic TEWL < 4.5 in six (27%) and β‐adrenergic:cholinergic ratio <0.05 in one. Of the six children with Δβ‐adrenergic TEWL < 4.5, one had CFTR‐RD (participant no. 7), two were CFTR heterozygotes and three were CFTR Negative (i.e., 5/6 were false positive results). However, the β‐adrenergic:cholinergic ratio in children was >0.05 in all but one child who was heterozygous with c.1210‐34TG[11]T[5] variant (participant no. 4).

3.4. Effect of age and sex on BAST in CFTR negative/healthy adult subjects

Median Δcholinergic TEWL was significantly higher in males compared to females but similar in children and adults. Median Δβ‐adrenergic TEWL was significantly higher in males compared to females and significantly higher in adults compared to children. Median β‐adrenergic:cholinergic ratio was similar between sexes but significantly lower in children compared to adults (online Supporting Information: Table E1). Scatterplots of BAST responses and age demonstrate no relationship with Δcholinergic TEWL, and a trend toward positive correlation with Δβ‐adrenergic TEWL and β‐adrenergic:cholinergic ratio (Figure 2).

Figure 2.

Scatterplots of Δcholinergic TEWL (A), Δβ‐adrenergic TEWL (g H₂O/m²/h) (B), β‐adrenergic:cholinergic ratio (g H₂O/m²/h) (C), and age for all CFTR negative participants, fitted with linear r ². TEWL, transepithelial water loss. [Color figure can be viewed at wileyonlinelibrary.com]

4. DISCUSSION

This pilot study in a South African cohort validates previous studies demonstrating the accuracy and safety of BAST in the diagnosis and exclusion of CF or CFTR‐RD where first‐tier routine investigations are inconclusive. BAST may therefore also be a suitable and more accurate alternative to SCT where SCT is not available or unaffordable as in many LMIC. We further demonstrated that BAST in young children can be successful, but that technical difficulty, interpretation and acceptability of the test are challenges in this age group. Our study findings suggest there is a need to establish reference ranges for healthy children due to lower β‐adrenergic secretion in children without CFTR variants compared to adult controls. BAST is a suitable and simpler alternative to NPD or other established advanced CFTR activity tests in clinical or research settings where CFTR genotyping is nondiagnostic or not available. Furthermore, in our experience, the technical aspects of BAST with evaporimetry were relatively easy to learn and implement, which is an advantage over other established CFTR activity tests.

Eccrine sweat glands are stimulated independently by cholinergic and β‐adrenergic conductive pathways but cholinergic responses after maximal stimulation are greater than β‐adrenergic responses. ¹⁷ Cholinergic stimulation is CFTR‐independent and mediated by calcium and potassium‐dependent conductance. In contrast, β‐adrenergic stimulation is mediated only by cAMP‐dependent chloride conductance and is thus CFTR‐dependent. Recognition of CFTR‐dependent β‐adrenergic pathways in sweat electrophysiology and early observations of reduced sweat secretory responses in CFTR carriers and absent responses in people with CF led to the development of the BAST as a diagnostic instrument. ¹⁸ , ¹⁹ The original validation study using evaporimetry in adults by Quinton et al. reported 100% sensitivity and specificity for CF, CF heterozygotes and healthy subjects, with a TEWL cut‐off at 4.5 g H₂O/m²/h, whereas TEWL in subjects with CFTR‐RD was variably low, or absent. ⁶ Other studies with adult subjects employing bubble imaging‐based ratiometric sweat rate assays after β‐adrenergic stimulation report similar accuracy in distinguishing CF, CFTR heterozygotes and healthy controls with β‐adrenergic:cholinergic ratio cut‐off values of 0.0055 or less. ⁸ , ⁹ Sweat bubble imaging techniques are technically difficult to operate but are more sensitive than evaporimetry at detecting small but physiologically important differences that distinguish CF from CFTR‐RD or measuring the effects of CFTR‐modulating drugs on CFTR activity. ⁸ , ¹⁰ , ¹³ , ²⁰ In our study the clinical objective was to exclude CF or CFTR‐RD in people presenting with clinically suspected CF for which evaporimetry performed well with optimal cut‐off values in adults of Δβ‐adrenergic 4.5 TEWL and β‐adrenergic:cholinergic ratio 0.05. We found a high number of false positive SCT in this cohort, in which BAST with evaporimetry successfully excluded CF/CFTR‐RD in all cases, and correctly diagnosed two adults with CF who had intermediate sweat chloride levels and residual function mutations known to be associated with lower or normal sweat chloride levels. ²¹ Reasons for falsely elevated sweat chloride levels in the intermediate and positive range in this population warrants further investigation and may include several known biological factors including malnutrition, environmental deprivation, acquired CFTR dysfunction from exposure to cigarette smoke or mineralocorticoid deficiency. Significant intra‐individual variability of SCT over time may also explain discrepant SCT results observed in our cohort. ²² , ²³ , ²⁴ , ²⁵  One adult male in this study with repeatedly high SCT values (90 mmol/L) was newly diagnosed with HIV‐infection and had an ichthyosis‐like skin disorder. False positive SCT in HIV infection has been reported in a child but the mechanism is unknown. ²⁶

A novel finding of our study was an age‐dependent effect on β‐adrenergic secretion, with lower β‐adrenergic secretion and β‐adrenergic:cholinergic ratios in CFTR negative children compared to adults who were CFTR negative or healthy controls. Lower cholinergic secretion sweat volume and β‐adrenergic secretion rate in females compared to males have been previously reported. ⁶ , ¹² , ²⁷ We were unable to extensively investigate or report BAST in children with CF or were CFTR heterozygotes due to small numbers and study protocol restrictions to only adult control subjects. Most CF suspect participants in this cohort were children with CF‐like symptoms who did not have genetically diagnosed CF/CFTR‐RD. Our findings in these participants thus provides valuable “non‐CF” control data for BAST in children, which has not been previously reported. Our data suggest that due to lower β‐adrenergic secretion driven likely by difference in sweat gland maturation, peak β‐adrenergic secretion using a cut‐off of 4.5 TEWL as per adult references does not accurately discriminate between CF, CFTR‐RD, CFTR heterozygous or CFTR Negative status in young children. However, the β‐adrenergic:cholinergic ratio, which is less influenced by sex, may be more useful than peak β‐adrenergic secretion in young children as supported by our data of children with Δβ‐adrenergic < 4.5 TEWL, where β‐adrenergic:cholinergic ratio > 0.05 excluded CF/CFTR‐RD in all but one child. However, further studies are needed to establish optimal age and sex‐dependent reference ranges of the BAST in children as conflicting data in older children has been published suggesting established reference ranges are valid. ¹¹

There are several limitations to this study that we considered when interpreting our results. First, we did not do extensive genetic testing in healthy adult controls due to budgetary constraints, thus it is possible that some were asymptomatic CFTR heterozygotes, which could have lowered their β‐adrenergic secretion. However, the study objective was to confirm or exclude CF/CFTR‐RD, which would not have influenced related analyses. Second, several participants were excluded from analysis due to failure of BAST for different reasons including accidental needle repuncture of the skin during injection of BAST drugs or technical factors due to movement artifact. Absent or poor cholinergic secretion sweat rate has been previously reported and could be caused by inadvertent subcutaneous injection of carbachol, which is too deep from dermal sweat glands. ⁶ Our protocol and regulatory requirements did not allow recruitment of healthy children controls and permitted only one repeat carbachol injection if it failed and repeating the full BAST protocol in study participants was not permitted. Third, we applied at our discretion a “modified” protocol in young children where the full protocol was not possible and similarly calculated Δβ‐adrenergic TEWL in participants where β‐adrenergic secretion was recorded but cholinergic secretion failed. Although Δβ‐adrenergic secretion calculations with this approach differ slightly, they are unlikely to change substantially as the reference TEWL measurement is within a 5 gH₂O/m²/h margin of error. Furthermore, documenting Δβ‐adrenergic secretion above the CF range remains useful to exclude CF, even in the absence of recording cholinergic secretion. Cholinergic prestimulation with pilocarpine is shown to potentiate β‐adrenergic secretion, therefore omitting or failure of this step could have altered β‐adrenergic secretion calculations which must be taken into consideration when β‐adrenergic secretion is within CF range. ⁸ , ²⁸ Increasingly iontophoresis as drug delivery method in place of intradermal injection is being explored as a less invasive approach, which is desirable in children. ⁸ , ²⁹ , ³⁰ Although iontophoresis of pilocarpine and atropine is successful, iontophoresis of a β‐adrenergic cocktail solution has had variable success. ⁸ , ³⁰

In summary, BAST accurately excluded CF or CFTR‐RD in the majority of children with inconclusive CF diagnosis in this ethnically diverse South African cohort. BAST is feasible and safe in children but interpretation of BAST in younger children needs further investigation due to lower β‐adrenergic secretion observed in children. β‐adrenergic:cholinergic ratio performed better than peak β‐adrenergic TEWL for excluding CF in children.

AUTHOR CONTRIBUTIONS

Marco Zampoli conceptualized and designed the work and the principal investigator of this study performed study procedures and led the author to draft the manuscript. Janine Verstraete made substantial contributions to data acquisition, analysis, and interpretation of the data, drafting, and revising the manuscript, and approving the final version. Thao Nguyen‐Khoa helped oversee the implementation of the study procedure, made substantial contributions to the acquisition, analysis, and interpretation of the data, drafted, and revised the manuscript, and approved the final version. Isabelle Sermet‐Gaudelus made substantial contributions to conceptualizing the study, study design and revising intellectual content, data interpretation, and revising and approving the manuscript. Heather J. Zar made substantial contributions to conceptualizing the study, study design and revising intellectual content, data analysis, data interpretation, and revising and approving the manuscript. Tanja Gonska made substantial contributions to data analysis, data interpretations, and revising and approving the manuscript. Brenda M. Morrow made substantial contributions to conceptualizing the study, study design and revising intellectual content, data analysis, data interpretation, and revising and approving the manuscript.

Supporting information

Supporting information.

Zampoli M, Verstraete J, Nguyen‐Khoa T, et al. β‐adrenergic sweat test in children with inconclusive cystic fibrosis diagnosis: do we need new reference ranges? Pediatric Pulmonology. 2023;58:187‐196. 10.1002/ppul.26179

DATA AVAILABILITY STATEMENT

All data collected from participants is provided in the online Supporting Information Material.