Reference Ranges for Respiratory Ciliary Function for Chinese Children May Not Be Extrapolated From European Data

Background:To aid in the diagnosis of Primary Ciliary Dyskinesia (PCD) and to evaluate the respiratory epithelium in respiratory disease, normal age related reference ranges are needed for ciliary beat frequency (CBF), beat pattern and ultrastructure. These reference data are not available for Asian or Chinese children. Our aim was to establish reference ranges for healthy Chinese children. Methods:Ciliated epithelial samples were obtained from 135 health Chinese children aged below 18 years by brushing the inferior nasal turbinate. CBF and beat pattern were analysedfrom high speed video recordings. Epithelial integrity and ciliary ultrastructure were assessed using transmission electronic microscopy. Results:The mean CBF from 135 children studied was 10.1 Hz (95% CI 9.8 to 10.4). Approximately 20% (ranged 18.0-24.2 %) of ciliated epithelial edges were found to have areas of dyskinetically beating cilia. Normal beat pattern was observed in ciliated epithelium from all subjects. We did not nd any effect of exposure to second hand smoke on CBF in our subjects. Microtubular defects were found in 9.3% of all of the cilia counted in these children while other ciliary ultrastructural defects were found in less than 3%. Conclusions: We established the reference range for CBF, beat pattern and ultrastructure in healthy Chinese children. Using similar methodology, we found a lower overall mean CBF than previously obtained European values. This study highlights the need to establish normative data for ciliary function in different populations.


Introduction
Primary ciliary dyskinesia (PCD) is a rare genetic condition with ciliary ultrastructural defects leading to ineffective ciliary movement and impaired mucociliary clearance. This results in sinusitis, recurrent chest infections leading to bronchiectasis and respiratory failure in affected individuals [1,2]. Majority of patients have persistent symptoms from birth or early infancy. However, a diagnosis is often made at a late stage, when signi cant bronchiectasis and permanent loss of lung function already occurred [3].Early diagnosis facilitates implementation of management strategies for new cases with the hope that morbidity can be reduced and lung structure and function preserved.
The European Respiratory Society (ERS) task force suggests a combination of nasal nitric oxide (NO) measurement and digital high speed video microscopy (HSVM) assessment of ciliary beat frequency (CBF) and pattern as the rst step in diagnostic testing for PCD [4]. Although CBF can be measured by photomultiplier and photodiode techniques, the values were signi cantly lower than by digital HSVM [5] and they do not allow assessment of ciliary beat pattern. HSVM has the additional advantage of allowing analysis of beat pattern that has been shown to be abnormal in PCD [6]. HSVM has been shown to have a high sensitivity and speci city for the diagnosis of PCD [7][8][9]. However, expertise is required to interpret ciliary function using HSVM as secondary epithelial damage following viral infection [10] and in severe asthma may result in signi cant secondary ciliary dyskinesia [11,12].
Transmission electron microscopy (TEM) is advocated as the next step if HSVM or nasal NO are abnormal and is regarded as a highly speci c test to con rm PCD. Yet, some patients with PCD have an apparently normal ultrastructure [13].As such, diagnostic armentarium should include both HSVM and TEM.
Normal ranges are essential to help interpret ndings and to date normal data on ciliary beat pattern has only been established for a European population [1]. The aim of this study was to establish normal age related reference ranges for CBF and pattern using digital HSVM and ultrastructure of the ciliated epithelium and ciliary axonemes using TEM in Chinese children. It also allowed comparison with normative data from a European study using the same methodology.

Study design and subjects
This cross-sectional study recruited healthy Chinese school children aged less than 18 years and adult volunteers in Hong Kong (HK). Subjects with a history of chronic respiratory or nasal disease, symptomatic upper respiratory tract infection during the previous 6 weeks, those requiring long-term medication or known cigarette smokers were excluded [1]. Exposure to second hand smoking (SHS) at home were explored. Subjects were examined to exclude obvious nasal defects with an auroscope. The inferior turbinate of subjects was brushed with a 3mm cytology brush to obtain nasal epithelial cells.
Analysis of CBF and beat pattern by HSVM [1,6] and ultrastructure by TEM [7] Ciliated strips of epithelium were suspended in a chamber created by glass slide and cover slips. The slide was then placed on a heated stage (37°C) of a Leica light microscope mounted on an anti-vibration table. Specimens were examined using a x100 interference contrast lens. Undisrupted ciliated strips of about 50μm in length and devoid of mucus were studied. Beating ciliated edges from a side-view pro le, were recorded using a digital high speed video camera (IDT, model NX4) at a rate of 500 frames per second. The recorded video sequences, projected onto a high resolution monitor were played back at reduced frame rates or frame by frame to determine CBF. The ciliated edge was divided into 5 adjacent areas measuring 10μm. Two measurements of CBF were recorded in each area, resulting in a total of 10 measurements for each edge. A maximum of 10 edges were analysed per subject. The number of frames for groups of beating cilia required to complete 10 cycles were recorded. CBF was converted using a calculation (CBF= 500/(number frames for 10 beats) x10). The reproducibility for inter-observer (measured independently by 2 observers) and intra-observer (measured by the same observer after 2 days) CBF measurements were evaluated [1]. Normal ciliary beat pattern denoted a coordinated cilium beating in a forward backward motion along the whole epithelial edge. Edges with dyskinetically beating cilia were counted and percentage of these edges was calculated.
Transmission electron microscopy analysis For TEM, nasal brushing samples were xed in 2% glutaraldehyde and processed through resin by standard techniques [1]. Ultrathin sections were cut at 100 nm, collected on 200 mesh thin bar copper grids, stained in 1% uranyl acetate, counterstained in Reynold's lead phosphate.

Evaluation of ciliary structure and function
The epithelial and ciliary ultrastructural changes of epithelium were assessed in a blinded fashion [7].
The number of ciliated cells, mucous cells, and dead cells of the epithelium were expressed as a percentage of total number of cells examined. Disruption and damage to the tissue were quanti ed using a scoring system that assess the degree of loss of cilia from ciliated cells, projection of cells from epithelial edge, cytoplasmic blebbing and mitochondrial damage with a summation epithelial integrity which incorporated all these measurements to assess the overall epithelial damage: 0=no damage, 1=minor, 2=mild, 3=moderate, 4=major, 5=severe damage [1]. Cilia in less than perfect cross-section but in which the microtubular arrangement could be recognised were recorded as being normal (9+2 microtubules) or defective. The results were recorded in batches of 100 counts, assembling as many such batches from a single section, and thus any pattern would be revealed. Dynein arms were assessed in any high-quality cross-sections encountered during the recording of microtubular arrangements and documented as showing the presence of both arms, only an outer or inner arm, or neither arm. If the presence of an arm was equivocal, it was counted as being present [14].

Data analysis
To calculate the sample size to establish age-related reference in children aged <18 years with a local population size of around 1,000,000 [15], allowing 9% margin of error from the mean at 5% con dence level, 120 subjects was required. This sample size of 120 would have a power of 80% and 5% level of signi cance (two-sided) to detect a 10% difference of mean CBF between the Chinese and European population [1]. Assuming 10% of sample insu ciency and loss during preparation, a total of 135 subjects were recruited. The mean and standard deviation (SD), 5th and 95th percentiles of CBF, the mean percentage, 5th and 95th percentiles of edges exhibiting dyskinetically beating cilia, cells with loss of cilia, cellular projections, cytoplasmic blebbing, mitochondrial damage and ultrastructural defects including microtubular, dynein arm or other defects of the 3 individual age groups and the whole group were calculated. A one way analysis of variance (ANOVA) was performed to detect a signi cant difference among individual groups. The mean CBF between children <18 years with adults between subjects with and without SHS exposure in each individual age group and between subjects in current study and the European study were compared using t-test.Inter-observer and intra-observer difference of CBF measurement were calculated by comparing the 95% con dence interval (CI).

Analysis of ciliary beat frequency and beat pattern measurements
We recruited 164 children (88 males, age range 2-17 years) and 50 adult volunteers from December 2015 to November 2016. We excluded 34 subjects because of known underlying diseases or inadequate samples and 1 adult subjects who was an active smoker. Samples from the remaining 135 children subjects (67 males, age range 3-17 years) and 44 adults (25 males, age range 18-60 years) were analysed. Forty-six children were exposed to SHS at home. Ten adults were exposed to SHS at home while 34 did not. The mean CBF and the percentage of dyskinetically beating edges for all subjects were summarized in Table 1. No signi cant difference was found in mean CBF between individual age groups (ANOVA, p=0.542) and dyskinetically beating edges between individual age groups (ANOVA, p=0.212).The normal ciliary beat pattern from a subject showed a coordinated ciliary beating in a forward backward motion along the whole epithelial edge (see Video 1) and the ciliated edges analysed that exhibited areas of dyskinetically beating cilia ranged from 18.0 to 24.2% (representative images of ciliated edge exhibited static cilia, seeVideo 2). There was also no signi cant difference in mean CBF between children with and without exposure to SHS for the 3 different age groups (ANOVA, p=0.89 for children aged 2-6, p=0.29 for children aged 7-12, p=0.58 for children aged 13-17) but there was slightly higher mean CBF in adults with SHS exposure compared to adults without (ANOVA, p=0.04) ( Table 2). The mean CBF for children aged < 18 years was slightly higher than the adult group but it did not reach statistical signi cance [10.1Hz (95% CI 9.8 to 10.4) versus 9.5Hz (95% CI 8.9 to 10.0), ANOVA, p=0.05]. Yet, it was lower than that of the European study [12.8Hz (95% CI 12.3 to 13.3, ANOVA, p<0.05) [1].  To establish a reference range, mean CBF was plotted against age of each subject ( Figure 1A). A weak negative correlation was found between mean CBF and increasing age (r 2 = 0.021). As cilia were found to beat at different frequencies within each sample, we also plotted sample variation in CBF, the ciliated edges with the highest and lowest CBF against age. The highest mean CBF of edges ranged from 7.3 to 24.3Hz ( Figure 1B) with 93% of subjects having a maximal CBF of >10 Hz. The lowest mean CBF of edges ranged from 2.3 to10.4Hz( Figure 1C) with 45% of subjects having a minimum CBF of >6 Hz.
Transmission electron microscopy examination of cell types Some subjects had an inadequate sample for ultrastructural analysis as tissue might have been lost during initial HSVM assessment performed before sample processed for TEM [1]. Among 179 subjects analysed with CBF and beat pattern, 121 subjects had su cient tissue for epithelial integrity measurements and 159 subjects had tissue processed for ultrastructure analysis. The percentages of different cell types observed in the ciliated epithelial strips were summarized in Table 3. There was no signi cant difference between the percentages ofdifferent cells types across age groups (ANOVA, p>0.1). Ciliated cells formed about 50% of the cell population.  Results are expressed as the mean percentage (5th, 95th percentiles) for each age group.

Assessment of ciliary epithelial integrity and ultrastructure by transmission electron microscopy
The integrity of ciliated epithelium was assessed by examining factors including loss of cilia, cellular extrusion, cytoplasmic blebbing, and mitochondrial damage in Figure 2. Normal with a normal healthy mitochondrion (arrow, bar=1µm) in Figure 2A. Loss of cilia, grade 3, and a cell with a damaged mitochondrion (arrow, bar=1µm) in Figure 2B. Cellular extrusion, grade 2, (bar=2µm) in Figure 2C.
Cytoplasmic blebbing, grade 2 (arrow, bar=2µm) in Figure 2D. Evidence of minor epithelial damage was observed and the analysis was summarised in Table 4.There was no signi cant difference of epithelial integrity score across age groups (ANOVA, p=0.07). Normal nasal epithelium with an intact ciliated surface and minimal disruption (epithelial integrity score=0, bar=2µm) was shown in Figure 3Aand abnormal epithelium with severely disrupted cell surface and marked loss of cilia (epithelial integrity score=4, bar=2µm) in Figure 3B.  Results are expressed as the mean percentage (5th, 95th percentiles) for each age group.
Ultrastructural analysis was summarised in Table 5. Abnormal cilia were observed in some subjects. The commonest defect identi ed was single micro-tubule ( Figure 4E).
The mean microtubule defects in 13-17 year group (14.4%) were signi cantly higher than those in the other age groups (8.3% in 2-6 year group, 5.3% in 7-12 year group; ANOVA, p<0.01). Central microtubules defects, compound cilia and other microtubule defects referring to a combination of defects were found in less than 3% of all of the cilia counted in children aged <18 years (Table 5). There was no signi cant difference between groups for the central microtubules defects, compound cilia and other microtubule defects (ANOVA, p>0.05).  Results are expressed as the mean percentage (5th, 95th percentiles) for each age group.

Discussions
We established the reference ranges for ciliary function and structure in healthy Chinese children using HSVM and TEM. We found an overall mean CBF of 10.1 Hz in children aged <18 years. CBF did not vary with age in these children.
We did not nd a signi cant effect of SHS exposure on CBF of nasal cilia in our children but adults with SHS exposure had a signi cantly higher mean CBF compared to adults without. Previous studies had been carried out in adults only and the results were con icting. Nasal CBF in active or passive smokers was signi cantly lower than that in non-smoke exposed in one study [16]. However, this study included subjects with middle ear diseases and healthy control. Our nding concurred with more recent studies that showed higher mean CBF in nasal epithelial biopsies or air liquid interface cultures among active smokers and non-smokers exposed to SHS than non-smokers [17,18].
We found a lower overall mean CBF in healthy Chinese children compared to the European study using the same methodology [1]. The HK team was trained by the European study team in England. A member of that team, who developed normal ranges for the UK, visited the HK laboratory to ensure the methods used were similar, with samples observed using similar slide chambers and with a heated stage that would negate differences due to laboratory temperature and humidity. It was possible that the difference was due to the different status of the subjects as the samples were obtained from awake Chinese children and from European children immediately after anaesthetic induction for elective surgery.
However, Propofol used in the anaesthetic induction was not shown to affect CBF [19].Subclinical laboratory-con rmed respiratory infections due to respiratory syncytial virus (RSV) and in uenza are common in children during in uenza seasons in HK [20,21]and they might have affected ciliary function in some cases. Yet, similar exclusion criteria were used to the European study to minimise this effect. The other possibility was an observer bias but the difference of CBF of our intra-observer and inter-observer mean was smaller than the European study. Thus, our nding may represent a genuine genetically determined difference between European and Chinese children.
One patient had a mixed ciliary beat pattern with dyskinetic cilia, cilia with normal pattern and cilia with circular beat pattern on occasional ciliated cells. Circular beat pattern is observed in ciliary trans-position defect and central microtubular agenesis [6,22], but the cilia of all the ciliated cell show dyskinetic beat pattern when viewed from above in these cases. We invited the subject to repeat the test, which showed normal CBF and normal beat pattern in 3 different beating planes including sideway pro le, towards the observer and view from above [23]. The percentage of dyskinetic edge decreased from 33.3% to 10%. Thus, we believed that the initial abnormal nding was secondary in nature, possibly due to a preceding subclinical respiratory infection.
Our results of ultrastructural analysis were in agreement with other reports on the number of outer and inner dynein arms visible, number of ciliated cells and epithelial integrity [1,24].Nevertheless, we found a higher percentage of microtubular defects, especially in [13][14][15][16][17] year group compared to the European study. The reasons for this are unclear. As discussed above, some children may have had unrecognized subclinical or be in the recovery phase of a mild respiratory infection. Microtubular defects have been reported in the recovery phase of mild respiratory infection in some patients up to 10 weeks after acute phase [25,26]. The damage to the nasal ciliated epithelium might also occur during sampling. Culture of ciliated epithelial cells at air-liquid interface, with re-differentiation and re-analysis of ciliary function and ultrastructure may eliminate secondary damage but was not feasible during this study [27].
We had a large sample size of healthy subjects and none were on medication. Our methodology was robust as we followed the European study [1] and were closely supervised. Our study was conducted to avoid peak seasons for viral infections. However, laboratory surveillance data showed in uenza was still prevalent in March 2016 with adenovirus, RSV and rhinovirus circulating in our community for that whole year [28]. We might have included some subjects with subclinical infections or recovering from very mild infection. We also identi ed subjects with undiagnosed allergic rhinitis and asthma. We excluded them in the analysis but might still include some with very mild allergic rhinitis with local in ammation leading to secondary ciliary dyskinesia.
In conclusion, we established normal reference range for respiratory ciliary function and ultrastructure in healthy Chinese children. Unexpectedly, we found a lower mean CBF in our children aged <18 years compared to a Europeanstudy. The reasons remain uncertain but could include potential genetic basis.
Our results suggest the necessity to establish normal reference ranges for different populations.