Diagnostic Testing of Patients Suspected of Primary Ciliary Dyskinesia

Transcription

1 Diagnostic Testing of Patients Suspected of Primary Ciliary Dyskinesia Wendy A. Stannard 1, Mark A. Chilvers 1 *, Andrew R. Rutman 1, Chris D. Williams 1, and Chris O Callaghan 1 1 Division of Child Health, Department of Infection, Immunity, and Inflammation, and Institute of Lung Health, University of Leicester, and Leicester Royal Infirmary, Leicester, United Kingdom Rationale: Electron microscopy (EM) of ciliated epithelium is widely used to diagnose primary ciliary dyskinesia (PCD). Ciliary beat frequency (CBF) has been used to screen samples to determine whether EM is indicated. Beat pattern analysis has been advocated as an additional diagnostic test. Neither has been subject to formal review. Objectives: To determine the ability of CBF and beat pattern analysis to predict EM-diagnosed PCD. Methods: CBF calculation and beat pattern analysis, using high-speed video microscopy, and EM were performed on nasal tissue from 371 patients consecutively referred to the Leicester Royal Infirmary for diagnostic assessment for PCD. With EM as the gold standard, receiver operating characteristic (ROC) curves were constructed and sensitivity, specificity, and positive (PPV) and negative (NPV) predictive values were calculated for CBF less than 11 Hz, ciliary dyskinesia score equal to or exceeding 2, at least 90% of ciliated edges beating dyskinetically, and an immotility index equal to or exceeding 10%. Measurements and Main Results: PCD was excluded in 270 patients and confirmed in 70 by EM. The sensitivity, specificity, PPV, and NPV for CBF less than 11 Hz were 87.1, 77.2, 50.0, and 95.8%, respectively. These values were higher for ciliary dyskinesia scores equal to or exceeding 2 (92.5, 97.6, 91.2, and 98.0%) and when at least 90% of ciliated edges were dyskinetic (97.1, 95.3, 84.6, and 99.2%). ROCs confirmed that the ciliary dyskinesia score and percentage of dyskinetic edges were superior screening indices compared with CBF and the immotility index. Conclusions: The use of CBF alone to screen which biopsies should have EM will result in a significant number of missed diagnoses. Ciliary beat pattern analysis is a more sensitive and specific test for PCD with higher PPV and NPV. Keywords: cilia; primary ciliary dyskinesia; bronchiectasis; ciliary beat pattern; sinusitis Efficient clearance of mucus from the upper and lower respiratory tract by coordinated ciliary movement is a major component of host defense. This system relies on coordinated beating of cilia with a normal beat pattern and beat frequency (1, 2). The importance of functioning cilia is highlighted in patients with primary ciliary dyskinesia (PCD), whose cilia are either immotile or beat with an abnormal pattern that fails to (Received in original form March 26, 2009; accepted in final form November 11, 2009) Supported by Action Medical Research, Cystic Fibrosis Trust (UK). We are very sad to report the death, in a car accident, of Dr. Wendy Stannard, who was an outstanding pediatrician and researcher. * Present address: Division of Paediatric Respiratory Medicine, BC Children s Hospital, Vancouver, BC, Canada. Correspondence and requests for reprints should be addressed to Professor Chris O Callaghan, M.D., Ph.D., Division of Child Health, Institute of Lung Health, Department of Infection, Immunity, and Inflammation, University of Leicester, Robert Kilpatrick Clinical Sciences Building, P.O. Box 65, Leicester Royal Infirmary, Leicester LE2 7LX, UK. ajb64@le.ac.uk Am J Respir Crit Care Med Vol 181. pp , 2010 Originally Published in Press as DOI: /rccm OC on November 12, 2009 Internet address: AT A GLANCE COMMENTARY Scientific Knowledge on the Subject Electron microscopy of ciliated epithelium, is widely used to diagnose primary ciliary dyskinesia. Ciliary beat frequency (CBF) has been used to screen samples to determine whether electron microscopy is indicated and ciliary beat pattern analysis has been advocated as an additional test. Neither test has been subject to formal review. What This Study Adds to the Field The use of CBF alone to screen for biopsies requiring further assessment by electron microscopy will result in missed diagnoses. transport mucus. Many of those affected remain undiagnosed and many are diagnosed only after years of chronic respiratory symptoms. Severe lung disease is not uncommon in older patients (3 6). The diagnosis of PCD in patients with a typical clinical phenotype relies on a number of tests, most of which have not been subject to formal review. Diagnostic testing has traditionally been based on measurement of ciliary beat frequency (CBF) and electron microscopy (EM) (7). EM is used to detect defects in the ciliary axoneme and to determine the orientation of cilia within cells (8), and is the most widely used diagnostic test for PCD, although a phenotype of PCD with no obvious ultrastructural defect has been described (9). A CBF of less than 11 beats per second (,11 Hz) has been suggested as a cutoff value, with only those with lower beat frequency proceeding to EM (7). It has been shown, using slow motion video analysis, that specific ultrastructural defects of the ciliary axoneme seen in PCD result in characteristic abnormalities of ciliary beat pattern (10, 11). Beat pattern analysis has been adopted by a number of diagnostic centers as a method of screening biopsies in addition to CBF measurement (5). However, the ability of CBF and ciliary beat pattern analysis to predict an EM diagnosis of PCD has not been determined. Therefore, the primary aim of this article was to determine the sensitivity, specificity, positive (PPV) and negative (NPV) predictive values of CBF measurement and ciliary beat pattern analysis in predicting PCD diagnosed by EM. Our group receives referrals, almost exclusively from pediatric and adult respiratory specialists, for diagnostic testing of patients who are thought to have PCD on clinical and demographic grounds. In this article, we also describe the prevalence of clinical and demographic features among referrals who go on to receive positive and negative EM diagnoses. We have observed that many referrals who do not have EM evidence of PCD exhibit areas of ciliated epithelium with a dyskinetic beat pattern. In many of these dyskinetic areas, CBF is also low. We therefore report comparative ultrastructural

2 308 AMERICAN JOURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINE VOL and ciliary function data between our referrals and previously published data from healthy, asymptomatic volunteers. METHODS Recruitment Before the establishment of our nationally funded PCD diagnostic center (Leicester Royal Infirmary, Leicester, UK), 371 patients had nasal ciliary brush biopsies consecutively evaluated at our unit. Five patients underwent biopsy elsewhere and samples were sent by courier. The remaining patients underwent clinical interview at our center using standardized forms, followed by nasal brush biopsy. Ethics This report is based on an audit of data collected as part of the routine clinical care of our patients. No additional interventions were required and all data has been anonymized. Considering guidelines in the Declaration of Helsinki, formal ethics approval was not needed. Our earlier study with healthy volunteers, which is used for comparison, received ethics approval from the Leicestershire Ethics Committee. Clinical Assessment At clinical interview, data related to demographics, neonatal history, and reported respiratory tract symptoms were recorded. A persistent wet sounding cough was considered to represent excessive lower airway secretions in children unable to expectorate sputum because of age. Upper airway symptoms recorded included persistent rhinorrhea (nasal discharge on most days), glue ear (diagnosed by a doctor), previous grommet insertion, and hearing difficulty (tested objectively in an audiology department). Nasal Brush Biopsy At the time of nasal brush biopsy all patients reported that they had been free from a symptomatic acute upper respiratory tract infection for at least 6 weeks. Ciliated samples were obtained by brushing the nasal epithelium with a 2-mm cytology brush (Olympus KeyMed Ltd, Southend-on-Sea, UK) and placed in medium 199 (ICN Biomedicals Inc., Solon, Ohio) supplemented with antibiotics (penicillin, 50 IU/ml; streptomycin, 50 mg/ml) (12). Assessment of Ciliary Function Ciliary beat frequency and beat pattern analysis were performed as described previously (11, 12). Up to 10 ciliated edges greater than 50 mm in length were observed at 378C with an oil immersion 3100 objective lens. Beating cilia were recorded with a digital high-speed video camera (Kodak Motion Corder analyzer 1000; Eastman Kodak, Rochester, NY) at a rate of 400 frames per second, allowing video sequences to be played back at a reduced rate to directly observe the ciliary beat cycle and to determine CBF, ciliary dyskinesia score, and immotility index. The CBF was determined at 10 sites along each edge, using the following formula: CBF 5 400/(number of frames for 10 beats 3 10) (11). The ciliary dyskinesia score was calculated as previously described (12). Briefly, an epithelial edge with all cilia beating normally scored 0; an edge with all cilia beating dyskinetically scored 3. The percentage of edges exhibiting any areas of dyskinetic cilia was also recorded (11). Finally, the immotility index was calculated using the following formula: immotility index 5 number of immotile cilia readings per total number of readings for the sample 3 100% (13). Assessment of Ciliary Ultrastructure Each sample was then processed for transmission EM to allow ultrastructural analysis of the ciliary axoneme and ciliary orientation (8, 12). Sections were examined in a methodical grid-square search so that each square was examined only once. At low power (35,000) the health and condition of any cells were noted. Any cells were then assessed at medium power (366,000) for the quality of the crosssection in which they had been cut. Only sections that were cut by chance in the best of cross-sections were suitable for microtubular and dynein arm assessment. These were examined at high power (3250,000) and the cilia from each cell were assessed for the presence or absence of normal dynein arms and for microtubular structure. Data were recorded for all cross-sections seen on a cell-by-cell basis. The percentage of cilia with dynein arm and microtubular defects was recorded (8). Outer dynein arm defects were defined as the complete absence of all outer dynein arms or, where all dynein arms were present universally, as a very short, stublike projection as opposed to the fully developed arm. A similar definition was used for inner dynein arm defects. In a typical cross-section, we would expect to see seven to nine outer dynein arms and fewer (five or more) inner arms. Ciliary function assessment was blind to the results of EM analysis (as is standard practice within our unit). Analysis Statistical analyses were performed with Minitab version (Minitab Inc, State College, PA) and SPSS version 16 (SPSS, Chicago, IL). Demographic, neonatal, and clinical history were presented as percentage prevalence. The difference between two proportions was used to compare responses between EM-positive and EM-negative referrals. CBF, ciliary dyskinesia score, percentage of dyskinetic edges, and immotility index were presented as means (and standard deviation). Sensitivity, specificity, and positive predictive values (PPV) and negative predictive values (NPV) were calculated for CBF less than 11 Hz, ciliary dyskinesia score equal to or exceeding 2, and at least 90% of ciliated edges exhibiting dyskinetic movement, compared with EM as the gold standard. Receiver operating characteristic (ROC) curves were constructed for each screening test. Patients were excluded from analysis if EM did not yield a clear result. In a small number of early referrals, clinical data were incomplete and some beat pattern indices were not measured. We previously published the results of functional and ultrastructural analysis of nasal brush biopsy samples from 76 healthy and asymptomatic volunteers aged 6 months to 43 years (11). These data were used to draw comparisons between our referrals and asymptomatic individuals. Analysis of variance was used to compare EM-positive and EM-negative referrals and our normal population, and the post-hoc Tukey test was used to make pair-wise comparisons when necessary. RESULTS Three hundred and seventy-one patients were referred for ciliary assessment. Three hundred and forty-eight were children (0 18 yr) and 23 were adults (>19 yr). Inclusions and exclusions are summarized in Figure 1. Three hundred and forty biopsies (91.6%; 177 male; mean age, 8.0 yr; age range, birth to 65 yr) yielded sufficient healthy tissue for analysis. EM confirmed an ultrastructural abnormality consistent with PCD in 70 patients (18.6%: 36 male; mean age, 5 yr; age range, birth to 17 yr) and normal ultrastructure in 270 patients (72.7%: 141 male; mean age, 8.8 yr; age range, birth to 65 yr). It was not possible to make a definite EM diagnosis in 31 patients (8.1%), who were excluded from further analysis. In 25 samples, there was inadequate/unhealthy ciliated epithelium. Four patients were excluded because of abnormal ciliary ultrastructure, where we could not be confident that they had a phenotype consistent with PCD (ciliary aplasia [n 5 2], unusual inner arm defect [n 5 2]). A further two were excluded because of normal ciliary ultrastructure but consistently abnormal beat pattern and frequency, which were believed to represent a variant of PCD with abnormal beat pattern but normal ultrastructure (9). Demographic, neonatal, and clinical characteristics for the patients who received a clear positive or negative EM diagnosis are presented in Tables 1 and 2. Fully completed forms were available for 264 of these patients. No demographic/clinical data were available for five patients (one with PCD) who had samples sent by courier from elsewhere. Forms were partially incomplete in 69 patients (20 with PCD). Some items, such as neonatal history, were not included in early versions of our form. Some patients or parents did not know part of the medical history (e.g., older patients knowledge of their birth history).

3 Stannard, Chilvers, Rutman, et al.: Ciliary Dyskinesia 309 Figure 1. Flow diagram illustrating the demographic data available for referrals. PCD 1ve 5 electron microscopy (EM) confirmed PCD; PCD -ve 5 no EM evidence of PCD. For the remaining cases, information was not recorded by the assessing clinician. ROC curves (Figure 2) showed that ciliary dyskinesia score (area beneath the curve, 0.991; 95% confidence interval [95% CI], ) and percentage of edges dyskinetic (area beneath the curve, 0.982; 95% CI, ) were superior to immotility index (area beneath the curve, 0.903; 95% CI, ) and CBF (area beneath the curve, 0.914; 95% CI, ) as predictors of an EM diagnosis of PCD. The coordinates of the curve confirmed our choice of cutoff values to be reasonable. The sensitivity, specificity, PPV, and NPV for a CBF less than 11 Hz to predict an EM diagnosis of PCD were 87.1, 77.2, 50.0, and 95.8%, respectively. These values were higher when a ciliary dyskinesia score equal to or exceeding 2 (sensitivity, 92.5%; specificity, 97.6%; PPV, 91.2%; NPV, 98.0%) was used as a predictor of EM-confirmed PCD and also when at least 90% of the subject s ciliated edges were dyskinetic (sensitivity, 97.1%; specificity, 95.2%; PPV, 84.6%; NPV, 99.2%). An immotility TABLE 1. PREVALENCE OF CHEST, NASAL, AND EAR SYMPTOMS AND TYPE OF SYMPTOM IN PATIENTS REFERRED FOR CILIARY ASSESSMENT PCD Confirmed on EM (n 5 70) Yes Data Unknown PCD Not Confirmed on EM (n 5 70) Yes Data Unknown P Value Chest symptoms 67 (95.7) 1 (1.4) 253 (93.7) 7 (2.6) 0.7 Wet-sounding 61 (87.1) 2 (2.9) 198 (73.3) 11 (4.4),0.05 cough Nasal symptoms 62 (88.6) 1 (1.4) 188 (69.6) 7 (2.6),0.05 Persistent 47 (67.1) 1 (1.4) 117 (43.3) 9 (3.3),0.05 rhinorrhea Ear symptoms 37 (52.9) 1 (1.4) 113 (41.9) 6 (2.2) 0.11 Glue ear 32 (45.7) 1 (1.4) 71 (26.3) 6 (2.2),0.05 Grommets 21 (30.4) 2 (2.9) 45 (16.7) 7 (2.6),0.05 Hearing difficulty 25 (35.7) 1 (1.4) 57 (21.1) 9 (3.3),0.05 Definition of abbreviations: EM 5 electron microscopy; PCD 5 primary ciliary dyskinesia. Data are presented as number of patients (%). Significant probability values are presented in boldface. index equal to or exceeding 10% gave the following: sensitivity, 69.6%; specificity, 98.9%; PPV, 94.1%; NPV, 92.6%). We also performed calculations with tests that combined beat frequency and beat pattern indices: CBF, 11 Hz or ciliary dyskinesia score >2 and CBF, 11 Hz or percentage dyskinetic edges.90%. The sensitivity, specificity, PPV, and NPV of these tests were identical to those obtained for the corresponding beat pattern test alone. The 70 referrals with EM-confirmed PCD had defects that could be categorized into one of several ultrastructural phenotypes (14). The beat patterns and associated ultrastructural defects of 59 of these patients have already been published (10, 15). The majority had dynein arm defects (n 5 55, 78.6%); the remainder had a radial spoke defect (n 5 4, 5.7%), transposition defect (n 5 6, 8.6%), central microtubular agenesis (n 5 3, 4.3%), or ciliary disorientation (n 5 2, 2.9%). Results of ciliary function assessment for various ultrastructural defects are shown in Table 3. Of the patients with EM-confirmed PCD, 40% had situs inversus. The results of CBF and beat pattern analysis are shown in Figure 3. The mean CBF for referrals with EM-confirmed PCD (4.7 Hz; SD, 4.9; 95% CI, ) was significantly lower than TABLE 2. SUMMARY OF NEONATAL HISTORY FOR PATIENTS REFERRED FOR CILIARY ASSESSMENT PCD Confirmed on EM (n 5 70) Yes Data Unknown PCD Not Confirmed on EM (n 5 70) Yes Data Unknown P Value Neonatal cough 45 (64.3) 5 (7.1) 88 (32.5) 10 (3.7),0.05 SCBU admission 40 (57.1) 15 (21.4) 58 (21.5) 38 (14.1),0.05 Ventilated 8 (11.4) 2 (2.9) 24 (8.9) 2 (0.7),0.05 Oxygen 27 (38.6) 23 (8.5),0.05 Development Motor delay 5 (7.1) 4 (5.7) 29 (10.7) 13 (4.8) 0.4 Language delay 12 (17.1) 3 (4.3) 38 (14.1) 14 (5. 2) 0.5 Social delay 10 (14.3) 3 (4.3) 31 (11.5) 14 (5.2) 0.5 Hydrocephalus 1 (1.4) 5 (7.1) 1 (0.4) 12 (4.4) 0.3 Definition of abbreviation: SCBU 5 special care baby unit. Data are presented as number of patients (%). Significant probability values are presented in boldface.

4 310 AMERICAN JOURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINE VOL Figure 2. Receiver operating characteristic curves for ciliary beat frequency (CBF), ciliary dyskinesia score, percentage of dyskinetic edges, and immotility index versus electron microscopy (EM) as the gold standard. for referrals with normal EM (12.3 Hz; SD, 1.98; 95% CI, ) (P, 0.05) and our previously published data from healthy volunteers (12.4 Hz; SD, 2.33; 95% CI, ) (P, 0.05). There was no significant difference in mean CBF between referrals with normal EM and healthy volunteers (P ). The mean ciliary dyskinesia score of referrals with EMconfirmed PCD (2.8; SD, 0.4) was significantly higher than for patients with normal EM (0.4; SD, 0.4) (P, 0.05) and healthy volunteers (0.1; SD, 0.2) (P, 0.05). In addition, the mean ciliary dyskinesia score was significantly higher for referrals with normal EM compared with healthy volunteers (P, 0.05). The mean percentage of dyskinetic edges in patients with EM-confirmed PCD (99.4%; SD, 3.0) was significantly higher than for referrals with normal EM (26.9%; SD, 22.9) (P, 0.05) and healthy volunteers (10.5%; SD, 15.8) (P, 0.05). In addition, this percentage was significantly higher in referrals with normal EM compared with healthy volunteers (P, 0.05). The mean immotility index was significantly greater in patients with EM-confirmed PCD (46.0%; SD, 39.2) than for referrals with normal EM (0.31%; SD, 2.07) (P, 0.05) and healthy volunteers (0.028%; SD, 0.24) (P, 0.05). There was no significant difference in immotility index between healthy volunteers and referrals in whom the diagnosis was excluded. The results for ciliary ultrastructure are shown in Figure 4. Interestingly, the mean number of nonspecific central microtubular defects in referrals without EM-confirmed PCD was significantly greater than for the normal population (Figure 4B). Heterogeneous microtubular defects and compound cilia were observed in these patients, compared with a single microtubular defect typically observed in patients with PCD (16). DISCUSSION We found that a CBF less than 11 Hz as a predictor of EMconfirmed PCD had lower sensitivity, specificity, and positive and negative predictive values compared with ciliary dyskinesia score equal to or exceeding 2 and at least 90% of ciliated edges beating dyskinetically. The use of CBF alone to screen which samples should be examined by EM (7) would have resulted in 12.9% of EM-positive referrals having a missed diagnosis of TABLE 3. CILIARY BEAT FREQUENCY AND CILIARY BEAT PATTERN ANALYSIS RESULTS FOR VARIOUS ELECTRON MICROSCOPY PHENOYPES Screening Test (mean and standard deviation) Ciliary Dyskinesia Score Percentage of Dyskinetic Edges Immotility Index Ciliary Beat Frequency Inner arm defect (n 5 14) 2.20 (0.92) 90.7 (22.0) 12.5 (13.8) 9.39 (5.04) Outer arm defect (n 5 18) 2.94 (0.24) 100 (0) 54.9 (36.6) 2.46 (2.47) Inner and outer arm defects (n 5 23) 3.00 (0) 100 (0) 70.6 (35.8) 1.00 (1.73) Central microtubule agenesis (n 5 3) 2.13 (0.23) (0) 0.7 (1.2) (0.67) Radial spoke (n 5 4) 3.00 (0.00) (0) 45.0 (10.8) 4.33 (0.50) Transposition (n 5 6) 3.00 (0.00) (0) 16.7 (40.8) (1.91) Ciliary disorientation (n 5 2) 1.00 ( ) ( ) 0.00 ( ) (3.54)

5 Stannard, Chilvers, Rutman, et al.: Ciliary Dyskinesia 311 Figure 3. Scatter plots illustrating (A) the ciliary beat frequency (CBF), (B) the percentage of dyskinetic edges, (C ) the ciliary dyskinesia score, and (D) the immotility index for referrals with negative electron microscopy (EM -ve), positive electron microscopy (EM 1ve), and healthy volunteers (11). Small shaded diamonds, individual values; large solid diamonds, means; H-brackets, 95% confidence intervals. PCD, whereas ciliary dyskinesia score and percentage of dyskinetic edges would have missed 7.4 and 2.9% of EMpositive referrals, respectively. This is of relevance to centers devising diagnostic pathways for PCD. It may also be advisable to revisit the possibility of PCD in patients with ongoing symptoms if a historical negative diagnostic test was based on CBF alone. Our data are consistent with previous reports (3) that EM-confirmed PCD is significantly more common in patients who exhibit persistent wet cough, nasal symptoms, glue ear, grommets, hearing impairment, neonatal cough and neonatal unit admission than in other patients thought to have PCD by referring clinicians. Our data confirm that the mean CBF of patients with central microtubular abnormalities (transposition or central microtubular agenesis [15]) and ciliary disorientation is higher than for other defects and that these defects are particularly likely to be missed by screening on the basis of CBF alone. Interestingly, neither of these abnormalities typically results in situs inversus. This may be responsible for the low prevalence of situs inversus in our population, 40%, which is lower than the usually quoted prevalence of 50% (7). Although EM was used in this report as a gold standard and is widely considered to be the most reliable method to diagnose PCD, it has been shown (since we began data collection) that a phenotype of PCD exists where no ultrastructural abnormality can be detected. The CBF of these patients is high, but the beat pattern is abnormal. This defect is therefore likely to be detected only on beat pattern analysis (9). We excluded two patients with consistently abnormal ciliary beat pattern but normal ultrastructure, which would be consistent with this PCD phenotype. Clearly, these reports require a reappraisal of the gold standard test in future. The major advantage of using digital high-speed imaging is that it enables movement of a cilium to be observed directly throughout its beat cycle at a high resolution and magnification. A cilium can be viewed in slow motion, or frame by frame, with 40 to 50 frames per ciliary beat cycle. Using this technology, we previously found that respiratory cilia actually beat with a simple backward and forward motion (17) and have developed a scoring system to assess the degree of dyskinesia and the percentage of edges that have cilia beating in a dyskinetic fashion (11, 12). We have also developed reference ranges for ciliary beat pattern in healthy adults and children (11) and have shown that specific ultrastructural defects causing PCD result in predictable beat patterns (10). A disadvantage is that evaluation of beat pattern takes considerably more time than CBF measurement. In this article, we used this system to determine CBF and ciliary beat pattern of a large cohort of patients with chronic respiratory symptoms but without EM evidence of PCD. This group had a mean percentage of edges with dyskinetic cilia, 2.5 times greater than healthy volunteers, and a mean ciliary dyskinesia score four times greater than that of healthy volunteers. There was no significant difference, however, in CBF

6 312 AMERICAN JOURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINE VOL Figure 3. (Continued). between this group and healthy volunteers. These results confirm our previous observation that cilia may beat in a dyskinetic pattern despite a normal ciliary beat frequency and would be missed by measurement of CBF alone (12). This is of concern as many research studies have used only CBF to determine ciliary function. Although the nasal epithelium has been shown to be similar in function and ultrastructure to bronchial epithelium (18), it remains to be determined whether the findings of ciliary dyskinesia from the nasal epithelium of these patients are reflected by changes in the lower respiratory tract. Patients with bronchiectasis have been found to have impaired tracheal and bronchiolar clearance (19). It is of interest that a normal CBF has been reported in patients with bronchiectasis due to conditions other than PCD (20). The cause of secondary ciliary dyskinesia in the nasal samples of referrals with no EM evidence of PCD is unclear. We attempt to minimize the effects of secondary viral damage by taking ciliary biopsies only when patients have been free of symptomatic upper respiratory tract infection or respiratory exacerbation for at least 6 weeks. Acute infections have been shown to delay mucociliary clearance (21). However, most studies on ciliary activity during acute viral infections have focused on beat frequency, which has not been found to change (12, 22). Using beat pattern analysis, we have shown that acute symptomatic coronaviral infection did not have a significant effect on CBF but did result in a marked increase in ciliary dyskinesia (12). In addition to ciliary dyskinesia, cytopathic epithelial damage with loss of cilia has been described after acute viral upper respiratory tract infections (12, 21 24) that may also contribute to poor mucociliary clearance. Such epithelial damage can take a number of weeks to fully resolve (23, 24). A factor that we did not control for was the possibility of secondary bacterial infection, as bacteria and bacterial toxins have also been shown to disrupt human respiratory epithelium in vitro. Common respiratory tract bacteria including nontypeable Haemophilus influenzae, Pseudomonas aeruginosa, and Streptococcus pneumoniae have been shown to reduce CBF and damage respiratory epithelium in vitro (25 29). The consistency of ciliary function assessments is an important consideration. We have previously published data confirming high levels of intraobserver and interobserver agreement for ciliary function measurements at our center (11, 30). Virtually all patients were referred from experienced pediatricians with a specialist interest in thoracic medicine or from thoracic physicians due to clinical suspicion of PCD. The vast majority had undergone extensive investigations including a sweat test and basic immunological investigations before referral, although these were not routinely documented. It is likely that the high diagnostic rate seen may be due to highly selective referrals from experienced physicians after other causes of chronic respiratory symptoms had been excluded. In addition, many of our referrals come from areas of the United Kingdom with a highly consanguineous Asian population, where the incidence of PCD may be as high as 1:2,000 (31). Of the total number of patients referred to our diagnostic service, one in five was diagnosed with PCD. Among those

7 Stannard, Chilvers, Rutman, et al.: Ciliary Dyskinesia 313 Figure 4. Scatter plots illustrating (A) the percentage of dynein arm defects in referrals with negative electron microscopy (EM -ve), referrals with a positive EM diagnosis (EM 1ve) affecting dynein arms only, and healthy volunteers; (B) the percentage of microtubular defects in patients with referrals with negative EM, referrals with a positive EM diagnosis affecting the microtubules only, and healthy volunteers (11). Small shaded diamonds, individual values; large solid diamonds, means; H-brackets, 95% confidence intervals. referred, the mean age of diagnosis of PCD was similar to that described by other centers (3) and highlights problems with diagnostic delay. Patients remained under the care of the referring clinician and we have not collected data prospectively to determine whether clinical symptoms persisted after assessment at our center, or whether referring clinicians still agree with the diagnosis on clinical grounds. Nasal nitric oxide levels are usually low in PCD (6, 32 34) and this measurement is now used to screen patients for whom a diagnosis of PCD is suspected. The main disadvantage of nasal nitric oxide measurements is cooperation among younger children. This tool was not available during the study period, although since the establishment of the national diagnostic service for PCD in the United Kingdom, we have routinely measured nasal nitric oxide levels in line with recent European guidelines (35). Despite assessment of both ciliary function and ultrastructure, the diagnosis of PCD can remain difficult for some patients. New defects that cause PCD have been reported (9, 15) and it is likely that a number remain to be determined. In summary, the use of CBF as a laboratory screening test to determine which patients should undergo EM will result in a number of patients with PCD being missed. The use of beat pattern analysis appears to be a more sensitive and specific test, with higher PPV and NPV. Beat pattern analysis may also be helpful in detecting patients with PCD due to DNAH11 mutations and who have an abnormal beat pattern but a normal ciliary ultrastructure (9). Conflict of Interest Statement: W.A.S. is deceased and is unable to provide a financial disclosure statement. M.A.C. received up to $1,000 from GlaxoSmithKline in lecture fees. A.R.R. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript. C.D.W. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript. C.O. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript. Acknowledgment: The authors acknowledge Dr. John Beckett for expert guidance with the statistics. The authors acknowledge funding received from Action Medical Research. Dr. Wendy Stannard was supported by Action Medical Research and also received support from the CF Trust (UK). References 1. Knowles MR, Boucher RC. Mucus clearance as a primary innate defense mechanism for mammalian airways. J Clin Invest 2002;109: Widdicombe JH. Regulation of the depth and composition of airway surface liquid. J Anat 2002;201: Coren ME, Meeks M, Morrison I, Buchdahl RM, Bush A. Primary ciliary dyskinesia: age at diagnosis and symptom history. Acta Paediatr 2002;91: Ellerman A, Bisgaard H. Longitudinal study of lung function in a cohort of primary ciliary dyskinesia. Eur Respir J 1997;10: Bush A, O Callaghan C. Primary ciliary dyskinesia. Arch Dis Child 2002; 87: ; discussion Noone PG, Leigh MW, Sannuti A, Minnix SL, Carson JL, Hazucha M, Zariwala MA, Knowles MR. Primary ciliary dyskinesia: diagnostic

8 314 AMERICAN JOURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINE VOL and phenotypic features. Am J Respir Crit Care Med 2004;169: Bush A, Cole P, Hariri M, Mackay I, Phillips G, O Callaghan C, Wilson R, Warner JO. Primary ciliary dyskinesia: diagnosis and standards of care. Eur Respir J 1998;12: Rayner CF, Rutman A, Dewar A, Greenstone MA, Cole PJ, Wilson R. Ciliary disorientation alone as a cause of primary ciliary dyskinesia syndrome. Am J Respir Crit Care Med 1996;153: Schwabe GC, Hoffmann K, Loges NT, Birker D, Rossier C, de Santi MM, Olbrich H, Fliegauf M, Failly M, Liebers U, et al. Primary ciliary dyskinesia associated with normal axoneme ultrastructure is caused by DNAH11 mutations. Hum Mutat 2008;29: Chilvers MA, Rutman A, O Callaghan C. Ciliary beat pattern is associated with specific ultrastructural defects in primary ciliary dyskinesia. J Allergy Clin Immunol 2003;112: Chilvers MA, Rutman A, O Callaghan C. Functional analysis of cilia and ciliated epithelial ultrastructure in healthy children and young adults. Thorax 2003;58: Chilvers MA, McKean M, Rutman A, Myint BS, Silverman M, O Callaghan C. The effects of coronavirus on human nasal ciliated respiratory epithelium. Eur Respir J 2001;18: Greenstone M, Rutman A, Dewar A, Mackay I, Cole PJ. Primary ciliary dyskinesia: cytological and clinical features. QJMed1988;67: Afzelius B, Mossberg B, Bergstrom S. Immotile cilia syndrome (primary ciliary dyskinesia), including Kartagener syndrome. New York: McGraw-Hill Medical Publishing Division; Stannard W, Rutman A, Wallis C, O Callaghan C. Central microtubular agenesis causing primary ciliary dyskinesia. Am J Respir Crit Care Med 2004;169: Cornillie FJ, Lauweryns JM, Corbeel L. Atypical bronchial cilia in children with recurrent respiratory tract infections: a comparative ultrastructural study. Pathol Res Pract 1984;178: Chilvers MA, O Callaghan C. Analysis of ciliary beat pattern and beat frequency using digital high speed imaging: comparison with the photomultiplier and photodiode methods. Thorax 2000;55: Rutland J, Griffin WM, Cole PJ. Human ciliary beat frequency in epithelium from intrathoracic and extrathoracic airways. Am Rev Respir Dis 1982;125: Currie DC, Pavia D, Agnew JE, Lopez-Vidriero MT, Diamond PD, Cole PJ, Clarke SW. Impaired tracheobronchial clearance in bronchiectasis. Thorax 1987;42: de Iongh RU, Rutland J. Ciliary defects in healthy subjects, bronchiectasis, and primary ciliary dyskinesia. Am J Respir Crit Care Med 1995; 151: Pedersen M, Sakakura Y, Winther B, Brofeldt S, Mygind N. Nasal mucociliary transport, number of ciliated cells, and beating pattern in naturally acquired common colds. Eur J Respir Dis Suppl 1983;128: Wilson R, Alton E, Rutman A, Higgins P, Al Nakib W, Geddes DM, Tyrrell DA, Cole PJ. Upper respiratory tract viral infection and mucociliary clearance. Eur J Respir Dis 1987;70: Carson JL, Collier AM, Hu SS. Acquired ciliary defects in nasal epithelium of children with acute viral upper respiratory infections. N Engl J Med 1985;312: Rautiainen M, Nuutinen J, Kiukaanniemi H, Collan Y. Ultrastructural changes in human nasal cilia caused by the common cold and recovery of ciliated epithelium. Ann Otol Rhinol Laryngol 1992;101: Feldman C, Read R, Rutman A, Jeffery PK, Brain A, Lund V, Mitchell TJ, Andrew PW, Boulnois GJ, Todd HC, et al. The interaction of Streptococcus pneumoniae with intact human respiratory mucosa in vitro. Eur Respir J 1992;5: Rayner CF, Jackson AD, Rutman A, Dewar A, Mitchell TJ, Andrew PW, Cole PJ, Wilson R. Interaction of pneumolysin-sufficient and -deficient isogenic variants of Streptococcus pneumoniae with human respiratory mucosa. Infect Immun 1995;63: Tsang KW, Rutman A, Kanthakumar K, Belcher J, Lund V, Roberts DE, Read RC, Cole PJ, Wilson R. Haemophilus influenzae infection of human respiratory mucosa in low concentrations of antibiotics. Am Rev Respir Dis 1993;148: Tsang KW, Rutman A, Tanaka E, Lund V, Dewar A, Cole PJ, Wilson R. Interaction of Pseudomonas aeruginosa with human respiratory mucosa in vitro. Eur Respir J 1994;7: Wilson R, Read R, Cole P. Interaction of Haemophilus influenzae with mucus, cilia, and respiratory epithelium. JInfectDis1992;165:S100 S Thomas B, Rutman A, O Callaghan C. Disrupted ciliated epithelium shows slower ciliary beat frequency and increased dyskinesa. Eur Respir J 2009;34: O Callaghan C, Moya E, Chetcuti P. Incidence of primary ciliary dyskinesia in the Asian population of Bradford (personal communication). 32. Wessels MW, den Hollander NS, Willems PJ. Mild fetal cerebral ventriculomegaly as a prenatal sonographic marker for Kartagener syndrome. Prenat Diagn 2003;23: Karadag B, James AJ, Gultekin E, Wilson NM, Bush A. Nasal and lower airway level of nitric oxide in children with primary ciliary dyskinesia. Eur Respir J 1999;13: Lundberg JO, Weitzberg E, Nordvall SL, Kuylenstierna R, Lundberg JM, Alving K. Primarily nasal origin of exhaled nitric oxide and absence in Kartagener s syndrome. Eur Respir J 1994;7: Barbato A, Frischer T, Kuehni CE, Snijders D, Azevedo I, Baktai G, Bartoloni L, Eber E, Escribano A, Haarman E, et al. Primary ciliary dyskinesia: a consensus statement on diagnostic and treatment approaches in children. Eur Respir J 2009;34: