Vandecruys A, Huysentruyt K, Van De Maele K, Vandenplas Y. How Best to Estimate Insertion Length of Multichannel Intraluminal Impedance pH Probes in Children. J Pediatr. 2023 Aug;259:113449. Epub 2023 May 6.
How Best to Estimate Insertion Length of Multichannel Intraluminal Impedance pH Probes in Children
Amber Vandecruys, MD1, Koen Huysentruyt, MD, PhD1, Kristel Van De Maele, RN1,
and Yvan Vandenplas, MD, PhD2
PMID: 37150291 DOI: 10.1016/j.jpeds.2023.113449
1 Vrije Universiteit Brussel (VUB), UZ Brussel, KidZ Health Castle, Brussels, Belgium.
2 Vrije Universiteit Brussel (VUB), UZ Brussel, KidZ Health Castle, Brussels, Belgium. Electronic address: email@example.com.
Objective: To assess the reliability of the KidZ Health Castle formula (KHC-F) to determine the correct probe position of a multichannel intraluminal impedance pH.
Study design: A retrospective cohort study was performed on 222 children between 1 month and 18 years of age undergoing multichannel intraluminal impedance pH. The primary outcome was the comparison of the pH sensor location determined by the KHC-F with the radiological target position. The margin of error was defined as 1 cm from the target position. Performance of the KHC-F and existing formulas was determined via the percentage with a correct position, mean error, 95% limits of agreement (Bland-Altman plots), and Spearman correlation. A post hoc analysis was performed with an updated KHC-F v2, subtracting -0.5 cm from the KHC-F.
Results: Positioning with KHC-F was correct in two-thirds of the participants, with a very strong correlation (ρ = 0.91) with the target position. Bland-Altman plots showed good agreement between KHC-F and target position (mean error of -0.44 cm, lower limit -3.2 cm, upper limit 2.3 cm). A post hoc analysis with the KHC-F v2 showed a correct positioning in 74% of patients. Comparison with other formulas showed a stronger performance of KHC-F and KHC-F v2 on correct positioning, mean error, and 95% limits of agreement.
Conclusions: The KHC-F leads to reliable results. KHC-F v2 outperforms all other existing formulas in children, thereby reducing the need for repositioning and the amount of x-ray exposure. The age distribution of the sample may be a limitation, as well as the retrospective nature of the study.
Keywords: Fluoroscopy; gastroesophageal reflux; pH monitoring; pH sensor; pH-metry; radiation; radiography.
ESPGHAN - European Society of Pediatric Gastroenterology, Hepatology and Nutrition
EURO-PIG - ESPGHAN European Pediatric Impedance Working Group
GER - Gastroesophageal reflux
HFA - Height for age
KHC-F (v2) - KidZ Health Castle Formula (version 2)
MII-pH - Multichannel intraluminal impedance pH monitoring
NASPGHAN - North American Society of Pediatric Gastroenterology, Hepatology and Nutrition
WFA - Weight for age
Gastroesophageal reflux (GER) is a frequent physiologic phenomenon in infants and children. The European Society of Pediatric Gastroenterology, Hepatology and Nutrition (ESPGHAN) and North American Society of Pediatric Gastroenterology, Hepatology and Nutrition (NASPGHAN) define GER disease as “when GER leads to troublesome symptoms that affect daily life and/or complications.” 1 Multichannel intraluminal impedance pH monitoring (MII-pH) is considered the gold standard to determine the prevalence of GER episodes and their relation with symptoms.1 A standardized esophageal position of the pH sensor and, as a consequence of the impedance rings, is needed for a correct, reproducible interpretation of the data and to allow comparability of results of investigations.
Historically, the position of the pH sensor was proposed to be at the second vertebra above the diaphragm.2 The second vertebra above the diaphragm corresponds to the position of the pH sensor in adults, which is 5 cm above the diaphragm.2 An easy-to-use formula predicting the correct location of the pH sensor would minimize the need for radiation and relocation of the probe. Different formulas were developed, but up to now, none of these formulas is adequate. A summary of existing formulas developed for the estimation of esophageal length and/or pH-metry probe insertion depth in children is given in Table I.
An easy-to-use formula was developed by the nursing team of the KidZ Health Castle: “the distance of the nose tip to the ear canal (cm) + the distance of the nose tip with head in neutral position to the nipple line (cm)” (KidZ Health Castle formula [KHC-F]), which is shown in Figure 1. The aim of this study was to (1) evaluate the accuracy of the KHC-F in comparison with the radiological target positioning in children referred for MII-pH, and (2) compare the performance of the KHC-F vs existing formulas.
MethodsA retrospective cohort study was performed based on data from 229 consecutively referred children between 1 month and 18 years of age for MII-pH, between February 25, 2020, and October 12, 2021 (Figure 2). Children younger than 1 month, those older than 18 years, and children with severe scoliosis or a history of esophageal surgery were excluded.
Placement of the probe was done by the nursing staff according to a standardized protocol, in use for years. The type of probes used were ZIN-BS-51 and ZPN-BG-07 (Diversatek Healthcare). The probe was inserted transnasally with the application of lubricating ointment on the tip. No anesthesia was used. Nurses inserted the MII-pH catheter and positioned the latter according the KHC-F, developed by experience. The target position was 2 vertebrae above the diaphragm (or the seventh posterior rib), as defined by the ESPGHAN European Pediatric Impedance (EURO-PIG) working group.2 After the initial insertion, the nurses always checked the position according to the target position using fluoroscopy. If necessary, the probe was repositioned. Both insertion lengths — the one according to the formula and the one after radiological control — were noted consecutively in all patients.
The following data were collected: age (months), height (cm), weight (kg), insertion length of the probe obtained by the formula (cm), and difference with insertion length after radiological control (cm). Children with missing data were excluded. Weight for age (WFA) z-score and height for age (HFA) z-score were calculated according to the Flemish reference charts.21
Statistical analysis was done using R version 4.0.3 (The R Foundation). Distribution of the probe positions was determined visually with QQ plots and tested formally using classical distribution tests. Spearman correlations were calculated between the distance to the correct position and clinical factors. The agreement between the insertion depth obtained by the formula and the target insertion depth was determined based on Bland-Altman plots, with the calculation of the mean error and the 95% limits of agreement.
Based on the mean difference, a post hoc correction to the KHC-F was developed, labelled the “KHC-F v2”, by subtracting 0.5 cm of the insertion length obtained with the original KHC-F.
A margin of error between the position according to the KHC-F and fluoroscopy up to 1 cm above or below the target position was accepted, because this difference is considered to be similar to the variation owing to breathing and changes of head position. The aforementioned tests of agreement were also calculated for previously published formulas applied to our study cohort. All formulas developed to determine the esophageal length from the nares to the lower esophageal sphincter were corrected by a multiplication factor of 0.87 to determine the correct position of the pH sensor, as suggested by Euler and Ament22 and applied by former authors.4-6
The association of individual factors such as age, height, HFA, and WFA with the insertion depth according to the KHC-F was determined using nonparametric testing and χ2 tests where appropriate. A P value of .05 was set as a significance level for all tests. Approval from the local ethical committee for the use of this information was obtained.
The initial study sample consisted of 229 participants. Seven children were excluded from analysis owing to missing data (Figure 2). The final study sample that was used for analysis consisted of 222 patients, of whom 106 were girls (47.7%). The median age was 6.5 months (Q1; Q3, 2.8; 15.7 months); 70.3% of patients were under the age of 1 year. The mean WFA and HFA z-score were, respectively, -0.37 ± 2.0 and -0.42 ± 1.3. A total of 186 children (83.8%) were shorter than 100 cm.
Evaluation of the KHC-F and Comparison with Existing Formulas
The probe insertion depth according to the KHC formula and the radiologically determined target position correlated strongly (r = 0.91; P < .001). The mean difference between the KHC-F and the target position was 0.44 cm caudally, with a lower limit of confidence of -3.2 cm caudally and an upper limit of confidence of +2.3 cm cranially (Table II). For 64.9% of the children (144/222), the KHC insertion length fell within the accepted difference of ±1 cm from the target position. This percentage increased further when limiting the formula to children younger than 1 year of age (69.9%) or shorter than 100 cm (68.3%). The KHC-F v2 (distance nose tip to ear canal [cm] + distance nose tip with head in neutral position to nipple line [cm] -0.5 cm) was also included, based on the mean difference of -0.44 cm on the Bland-Altman plot. The KHC-F v2 resulted in 74% of the children having a correct probe position.
An overview of the mean errors with their respective upper and lower limits of agreement and the percentages of children with a correct position of the probe according to the KHC-F and other existing formulas with our data is listed in Table II. Formulas to determine esophageal length other than nares to lower esophageal sphincter were excluded from Table II, because our data could not be used in these formulas.
Among the existing formulas, the Jolley14 and Moreau6,20 formulas came closest to the overall KHC-F performance, with 56.8% and 54.1% of the probes in a correct position, respectively. All other formulas performed sufficiently accurate in less than 50% of our study cohort. Mean errors of the existing formulas were under 1 cm in the Jolley formula (-0.76 cm), Staiano15 formula (0.78 cm), and Moreau formula (-0.88 cm) and both the original KHC-F (-0.44 cm) as well as the KHC-F v2 (0.05 cm), but none of the formulas resulted in upper and lower limits of agreements below 1 cm.
Exploration of Influencing Factors
The proportion of children with a correct position of the probe according to the KHC-F was significantly higher in children less than 1 year of age (69.9% vs 53.0%; P = .016) and in children shorter than 100 cm (68.3% vs 47.2%; P = .015). There was no significant difference in the mean HFA z-score between children with a correct or incorrect position of the probe based on the KHC method (95% CI of the mean difference, -0.14 to 0.61; P = .219), but the WFA z-score was significantly lower in the group with a correct probe position (95% CI of the mean difference, 0.05-1.04; P = .030).The KHC-F v2 reached 78.2% of correct positions under the age of 1 year and 77.9% of correct positions under 100 cm of height and outperformed every existing formula. Under the age of 1 year, the Jolley formula scored better than the original KHC-F, but not the KHC-F v2. The increasing scatter and tendency of the KHC-F to overestimate in children more than 1 year old is also clearly shown in Figure 3. The KHC-F v2 scores best for all ages.
Figure 3. Bland-Altman plot of KHC-F vs radiological positioning. The red solid line with surrounding blue dotted lines denotes mean difference with its 95% confidence interval. The red dashed line with surrounding blue dotted lines denotes the upper and lower limits of agreement with respective 95% confidence intervals. Negative values represent a probe that was positioned too much caudally and positive values a probe that was positioned too much cranially.
DiscussionThe KHC-F resulted in a correct positioning of the pH sensor in two-thirds of all patients, and a mean error of 0.44 cm caudally between the KHC-F and the radiologically determined target position. The pH sensor position was correct in 68.3% of the children with a height of less than 100 cm and 69.9% of the infants under the age of 1 year. The Bland-Altman plot (Figure 3) showed that 7 probes (3.1%) would have been placed 5 cm or more too caudally, which would increase the risk of false-positive outcomes.
The KHC-F v2 resulted in a mean error of 0.05 cm too cranially. It has an overall correct position in 74.3% of the entire cohort and 78.2% of the infants less than 1 year old, which outperforms all existing formulas.
Historically, several formulas were developed to determine either the esophageal length or the correct position of the pH probe, an overview can be found at Table I. Before 2012, most formulas for pH probe positioning in children were developed to determine the esophageal length. The right position of the pH sensor was determined at 0.87% of the esophageal length, as suggested by Euler and Ament.22 This changed when the ESPGHAN EURO-PIG working group recommended the position of 2 vertebrae above diaphragm.2 Formulas to determine the correct pH sensor positioning as suggested by the ESPGHAN EURO-PIG working group are scarce.
The criteria used to consider correct positioning of the pH sensor differ in most studies, going from stringent criteria (±1 cm) to one vertebra. However, the majority of the authors do not report the accepted margin of error.3,5,7,14,16-18,20 An overview is provided in Table I. The size of one vertebra is historically determined as 1/12th of the thoracic spine length, which is calculated by 2 formulas (thoracic spine length for children under the age of 2 years = 9.299 + 5.309 × age in years + 1.351 × age in years, thoracic spine length for children over the age of 2 years = 13.067 + 0.9989 × age in years), both developed by Currarino et al.23 It should be noted with the size of one vertebra as margin of error, the older the child the greater the margin of error (size from 1 cm in a 6-month-old to 2.4 cm in a 16-year-old). As seen in Table II, several formulas that were reported with a good performance4,6,8 had a poor outcome when applied to our data. An explanation of this difference in performance would be the stringent margin of error we apply (1 cm).
The major strength of this study lies in the stringent criteria of allowing the pH sensor position to deviate only 1 cm of the target position. Despite these strict criteria, the KHC-F and especially the updated KHC-F v2 performed well. The ease of use, without the need for extensive calculations or measuring the height of the child, is another major advantage of the formula. Moreover, because all calculations for the other formulas were done post hoc, this did not account for possible errors that can occur when those formulas would have been used in a clinical setting. A limitation of this study is the distribution of the study cohort because children less than 1 year of age were over-represented. GER disease in children is, however, most frequently seen under the age of 12 months. A second limitation of this study is inherent to the retrospective character of the study. This nature could have led to an inclusion bias, although only 7 of the 229 children were excluded owing to missing data.
The updated KHC-F v2 has a good agreement with radiological target positioning and is easy to calculate. The formula outperforms all other currently available formulas in children, thereby decreasing the need for repositioning of the probe. Radiological confirmation of the correct position remains recommended. However, the KHC-F v2 decreases the need for a second radiological control after repositioning. We suggest applying the updated KHC-F v2 as a standard formula to estimate the insertion length of MII-pH probes in pediatric patients, especially in children up to 1 year old and/or less than 100 cm in height, in a setting where radiological control probe positioning is not feasible. Further research is needed for children more than 1 year old and/or more than 100 cm in height, and to validate the updated KHC-F v2.
Declaration of Competing InterestThe authors declare no conflicts of interests.
Назад в раздел
Популярно о болезнях ЖКТ читайте в разделе "Пациентам"
Информация на сайте www.GastroScan.ru предназначена для образовательных и научных целей. Условия использования.