The study was partly cross-sectional and partly designed as an uncontrolled, observational, single-group, single-center, longitudinal, and retrospective study of prospectively collected lung function measurements in a PCD cohort.

Patients with PCD are followed lifelong at the National Danish PCD Center and were all eligible for the study. Because of the changing standards of PCD work-up over the 30-year study period, patients in this cohort were not uniformly diagnosed. However, all patients had a verified abnormal function analysis of ciliary beat pattern and ciliary beat frequency in combination with a consistent history of symptoms of PCD and continuing clinical signs of PCD throughout follow-up. Additionally, electron microscopy (EM) of ciliary ultrastructure, pulmonary radioaerosol mucociliary clearance (10), and nasal nitric oxide measurements was performed in most patients. Nasal nitric oxide was measured using NIOX (Nitric Oxide Monitoring System, Aerocrine, Solna, Sweden) equipment with sampling of nasal nitric oxide during breathhold to ensure soft palate closure in alignment with American Thoracic Society and European Respiratory Society recommendations (11). Our own reference values for nasal nitric oxide levels were used with a cut off value of 175 ppb (unpublished data).

Patients diagnosed with PCD, as discussed previously, and at least 1.5 years of follow-up with acceptable spirometries were eligible for the study.

Patients with uncertain diagnosis, poor cooperation to spirometry, or nonvalid lung function measurements, and patients followed less than 1.5 years were all excluded.

The intentional treatment strategy for Danish patients with PCD is derived from cystic fibrosis (CF) care at our center (12), although patients with PCD are followed less closely (3-month versus monthly clinic visits) and consists of the following:

• Twice daily: Chest physiotherapy using a positive-expiratory pressure mask.

• Monthly: All patients are asked to send sputum samples once every month. Endolaryngeal suction (using a nasal suction catheter) is performed whenever a patient (child or adult) is unable to produce a sputum sample.

• Every 3 months: Clinical examination for all patients and lung function measurements by spirometry (from approximately 5 years of age).

• Yearly: Chest x-ray; visit to ear-nose-throat department.

Immediate antibiotic treatment is initiated whenever a sputum or endolaryngeal suction sample is positive, even with only minor or absent clinical symptoms. Haemophilus influenzae and Moraxella catarrhalis are treated with 14 days of amoxicillin with or without clavulanic acid and in some cases alternately macrolide for 14 days (M. catarrhalis) to 3 months (H. influenzae).

Pseudomonas aeruginosa is treated with colistin inhalation and oral ciprofloxacin for 30 days whenever P. aeruginosa is cultured positive in patients without chronic P. aeruginosa infection. When chronic infection exists, elective intravenous treatment is given for 14 days every third month.

Age, sex, clinical, and diagnostic parameters, combined with longitudinal data on height, FEV₁ (liter), and FVC (liter) were extracted from the PCD cohort database and from patient files. All PCD patients entered the PCD cohort at the time of their diagnosis and were then followed prospectively. Data were collected post hoc.

Spirometry was performed as soon as the child could cooperate consistently, usually from 5 to 6 years of age, and continuously performed approximately every third month hereafter. For each child every flow–volume curve was evaluated and excluded if technique was insufficient to minimize bias because of spirometry training. FEV₁ and FVC measurements were as per American Thoracic Society standards (13). Values of both z score and percent predicted for both FEV₁ and FVC were calculated (14).

We included 74 patients with PCD (39 males, 35 females) from the National Danish PCD cohort. The periods of follow-up ranged over three decades, with variable lengths of follow-up for each patient since PCD diagnosis (Figure 1). During the 3-decade follow-up, a total of 2,937 lung function measurements were performed. Median (range) age at diagnosis was 8.1 (0–43.7) years with a skewed distribution toward preschool age (Figure 2). Median (range) duration of follow-up after diagnosis was 9.5 (1.5–30.2) years (Figure 1). The median (range) age of the cohort increased during the 30 years between the first spirometry (1979) and the last spirometry (2009) from 9 (4.4–43.7) years to 19.3 (6–70) years. The cohort consisted mostly of children and young adults, as shown by the accumulated number of patient years (PY) and associated recordings of lung function that fell into each age intervals as follows: 0–10 years, 136 PY; 11–20 years, 410 PY; 21–30 years, 250 PY; 31–40 years, 102 PY; 41–50 years, 76 PY; and more than 51 years, 20 PY. The low number of accumulated PY at age 0–10 years is caused by the fact that spirometry is only performed from approximately 5–6 years of age.

Figure 1. Lexis diagram of follow-up periods in each of the 74 included PCD patients from diagnosis (left-hand time point) to last spirometry (right-hand time point). Open circles = year of diagnosis; dashed lines before circles = period of time after diagnosis with no spirometries; full lines = time since first acceptable spirometry.

Download Figure | Download Powerpoint

[More] [Minimize]

Figure 2. Linear regression of the first recorded lung function by spirometry after diagnosis versus age in 74 patients. (A) FEV₁% predicted. (B) FVC % predicted. Asterisks = patients with first spirometry performed less than 6 months after diagnosis; open circles = patients with first spirometry performed more than 6 months after diagnosis; dashed lines before open circles = period of time after diagnosis with no spirometries.

Download Figure | Download Powerpoint

[More] [Minimize]

Clinical and diagnostic features of the patients in the cohort are shown in Table 1. All 74 patients had a typical clinical phenotype suggestive of PCD and a ciliary function analysis showing abnormal ciliary beat pattern or ciliary beat frequency. Abnormal ciliary ultrastructure was demonstrated in all patients in whom EM was performed and conclusive (n = 45; 60.8%). Nasal nitric oxide, sampled during breathhold in 19 (median age [range]: 18 [7–71] yr) of the 29 patients without EM was below our cut off of 175 ppb (unpublished data), thus further supporting the PCD diagnosis. In 5 of the remaining 10 cases, pulmonary radioaerosol mucociliary clearance was performed in addition to ciliary function analysis and demonstrated abnormal mucociliary clearance. In the last 5 cases abnormal ciliary function served as the only criterion for diagnosis. Situs inversus was seen in 9 (31%) of 29 of the patients without EM result, which was not significantly different from the ratio of situs inversus among patients with verified EM defects (17 [∼38%] of 45 ). No patients with syndromic PCD variants were included in the study.

A cross-sectional analysis of the first recorded lung function measurement versus age is shown in Figure 2. The first measured FEV₁ in patients with preschool diagnosis (<6 yr of age; n = 28) varied widely between normal and subnormal (i.e., <80% predicted), but with a substantial number (n = 10; 36%) demonstrating subnormal values (59–79% predicted). Yet, another complementary and overlapping subgroup of 30 (41%) children younger than 10 years of age had their first spirometry performed more than 6 months, median (range) 4 (0.5–8.3) years, after diagnosis and initiated treatment because of unacceptable performances with spirometry. Still, this group did not exhibit better FEV₁ % predicted (P = 0.5) or FVC % predicted (P = 0.8) compared with children less than 10 years of age with spirometry performed within 6 months from diagnosis.

Lung function, expressed as the percentage of predicted, was considerably lower in patients with later diagnosis, and the mean annual decline across the cohort was 0.8% per year (P < 0.0001; 95% confidence interval, 1.2 to −0.4). However, there was a tendency to deviate from linearity at older ages, where the relation seemed to become less steep. Most patients with late diagnosis exhibited subnormal values, even as low as 35 to 50% of predicted in FEV_1.

A total of 2,937 lung function measurements were performed in the longitudinal study of the 74 patients during the 30-year observation period.

Time courses of lung function following diagnosis, described as slopes over 10 subsequent years for subjects with at least 5 years of follow-up varied greatly between individuals but with no significant association to age at diagnosis (n = 58; P = 0.22) (see Figure EA in the online supplement). Furthermore, no significant association between the slopes, recalculated by omitting the first spirometry for each subject and initial measurement, was found (n = 58; P = 0.055). However, we did observe a tendency to a negative association as expected from the positive correlation between observations of lung function close in time (see Figure EB).

Linear regressions of full-length follow-up for all 74 patients are shown in Figure 3A (FEV₁ % predicted) and Figure 3B (FVC % predicted) and as z-scores in Figure EC (FEV₁) and Figure ED (FVC). The regression lines are dashed according to the behavior seen in the full-length follow-up period (i.e., increasing [n = 7; 10%] FEV₁ % predicted more than 10 percentage points, stable [n = 42; 57%] with change within ±10 percentage points, and decreasing [n = 25; 34%] more than 10 percentage points). Altogether, the change in lung function ranged from −37 to +42 percentage points in FEV₁ and −28 to +28 percentage points in FVC, across the cohort.

Figure 3. Linear regressions on full-length follow-up (years) on courses of lung function in 74 patients with PCD from their first measured lung function. Each regression line represents one patient. Lines are divided into three categories according to whether the patients followed an increasing (>10% percentage points), stable (change over time within ± 10 percentage points), or decreasing (>10 percentage points) course of lung function. (A) FEV₁ % predicted. (B) FVC % predicted.

Download Figure | Download Powerpoint

[More] [Minimize]

Figure 4 shows the cross-sectional relation from Figure 2 now with an overlay showing the median and interquartile range of the individual slopes for the time course of follow-up spirometries. It illustrates that although the median slope after diagnosis is somewhat less steep than the cross-sectional line, there is also a large fraction of subjects exhibiting an even larger decrease with age. The median follow-up slope was −0.246, the corresponding average was very close at −0.242, and a 95% confidence interval was −0.630 to 0.146 (i.e., not significantly different from 0 [P = 0.22], but just about significantly different from the general decline according to the cross-sectional regression line [P = 0.05]).

Figure 4. Linear regression line (full line) of the first recorded FEV₁ % predicted after diagnosis versus age in 74 patients with an overlay of dashed lines showing the median and interquartile range of the individual slopes for the time course of follow-up spirometries with origins at different ages and lung function levels.

Download Figure | Download Powerpoint

[More] [Minimize]

The study suggests PCD to be a serious threat to evolution of lung function even early in life: lung function was shown to decline very rapidly with age, starting already among preschool children. Additionally, lung function measured at diagnosis (or as soon as possible hereafter) was below 80% of predicted in more than one-third of children with preschool diagnosis despite initiated standard monitoring and control. This is indeed worrying, although selection bias toward early diagnosis of children with more severe symptoms may contribute to some extent to such a large fraction. Still, little is known to date of the progression of lung disease in early childhood PCD. A concern is that even the youngest patients already have irreversible lung damage. A recent case report to support this concern showed two out of three children with PCD less than 3 years of age to have bronchiectasis verified by high-resolution CT scan and three of three to have decreased lung function by infant pulmonary lung function testing despite the very early age at diagnosis (15).

Thirty children (41%) less than 10 years of age were too young to have spirometry performed when diagnosed, and therefore had their first spirometry postponed for more than 6 months (median, 3.2 yr) after diagnosis. This group did not show significantly different first FEV₁ or FVC compared with children less than 10 years (n = 13) with spirometry performed when newly diagnosed (<6 mo). Again, our data did not support the existence of a possible short-term effect of initial treatment.

First measured lung function declined with age. Because this decline was expressed by percent of predicted values (FEV₁) in which formula the natural decline with age is already included, the demonstrated loss of lung function probably results from delay of diagnosis. We found an annual decline of 0.8% in FEV₁ % predicted related to age at first measured spirometry after PCD diagnosis. This is completely comparable with another large cross-sectional study of spirometry data on patients with PCD in which Noone and colleagues (16) found an overall decrease of 0.8% in FEV₁ per year across their cohort.

Course of lung function varied greatly in the study. The course of lung function after diagnosis was related neither to age at diagnosis nor to level of lung function at first measurement after diagnosis. This is new knowledge and in contradiction with previous studies suggesting that lung function can be held stable after diagnosis and initiated treatment (4–6).

As described later, our results indicate a great variation in course of lung function after diagnosis, because course of lung function could present with three different patterns that were all unrelated to age at diagnosis and unrelated to level of lung function at diagnosis: (1) a declining course of lung function after diagnosis, (2) increasing lung function after diagnosis, and (3) a course of stabilized function. Hence, in our study, lung function was not maintained in all patients despite their being enrolled to follow the standard PCD treatment regimen at our Center. These data emphasize that treatment, predominantly with antibiotics, and monitoring alone are not sufficient to maintain lung function in a newly diagnosed patient with PCD at all times. However, median decline in 5 to 10 years of follow-up was less steep than the cross-sectional regression, indicating that there was some effect of treatment. Still, a large fraction of subjects exhibited an even larger decrease with age, as is shown by the interquartile ranges given in Figure 4, underlining the consistent variation in course of lung function after diagnosis for patients with early and late PCD diagnosis.

Previous published data on part of this cohort (6) have shown that patients with early diagnosis had better lung function at age at diagnosis compared with patients with late diagnosis. This is in alignment with our present study. However, in contrast to these previous findings (6), our study did not support that lung function is stabilized because of diagnosis and initiated treatment. On the contrary, our study suggests that the further course of lung function after diagnosis can decrease, stabilize, or even increase regardless of age at diagnosis and regardless of level of lung function when diagnosed. Our study includes more patients, an overall longer period of follow-up after diagnosis, and full data sets of spirometry as opposed to mean values per year. Also, the present applied statistical analysis of data allows each age at a PCD diagnosis to be included in our calculations instead of grouping according to fixed age groups.

Only a few other studies on lung function in patients with PCD have been reported and two of these included only a small number of patients: Hellinckx and colleagues (5) included 10 patients with PCD for longitudinal follow-up 3 to 20 years after diagnosis and found stabilized lung function in 8 out of 10 of these patients. As opposed to these results, Corkey and colleagues (4) included seven 12- to 25-year-old patients with PCD who all remained stable in lung function in the observed periods of 4 to 14 years, regardless of the level of lung function, when the patients entered the cohort.

Our study brings new knowledge to the field in two ways: previous longitudinal studies have only shown capability to stabilize lung function (4–6), whereas capability to improve lung function, as shown in a minority in the present study, has not previously been reported; and from the previous studies (4–6) decrease in lung function after diagnosis and initiated treatment was only very rarely observed. We found approximately one-third of the patients to have a decrease in lung function of more than 10 percentage points in predicted values over time, some even as much as −37 percentage points in FEV₁.

We suggest that the implications of these new findings are that any patient with PCD, regardless of age at diagnosis and level of lung function when diagnosed, could harbor the potential to improve, stabilize, or deteriorate in lung function, that it is highly unpredictable which route each patient will follow, and that early diagnosis in itself should not be considered a guarantee for a good prognosis.

We found the course of lung function after diagnosis to be independent of both age at diagnosis and level of lung function at diagnosis. Thus, we were unable to predict from these two parameters alone if a patient would have a beneficial increasing course of lung function or a further deterioration after entering the cohort and after initiating treatment. Possibly, these latter patients could have benefited from a more individualized and perhaps intensified treatment regimen, because the given treatment has not been satisfactory. However, the remaining and problematic future task will definitely be how to pinpoint these patients at an early stage to intervene.

This complicates the strategies in PCD care for preventing lung function decline. Of similar importance, however, is that our findings also suggest a challenge in pinpointing patients that can actually improve lung function after diagnosis. Therefore, PCD care should not just aim for stabilizing lung function in a patient.

Chronic infection with Pseudomonas aeruginosa has been reported to be expected in adult patients with PCD (16), which was also an observation from our cohort (data not reported). However, the potential impact of P. aeruginosa infection on lung function in patients with PCD remains to be investigated.

There has likely been an improvement in overall population health during the study period. Also, antibiotic treatment has changed over time, especially the more aggressive anti-P. aeruginosa treatment that has been drawn from studies in CF. The introduction of elective intravenous antibiotics for 14 days every third month in patients with CF and chronic P. aeruginosa in 1976 was followed by introduction of inhaled colistin and ad hoc oral ciprofloxacin between elective intravenous therapy. This resulted in a marked increase in survival rate in Danish patients with CF (12). Later, the immediate treatment with inhaled colistin and oral ciprofloxacin, whenever P. aeruginosa was cultured from sputum in patients with noncolonized CF, was implemented in 1989. This is still the main strategy today for postponing chronic P. aeruginosa in patients with CF and PCD. These changes are potential confounders that may have biased our data toward a poorer outcome in patients diagnosed early in the study period compared with those enrolled in the cohort in more recent years and have not been addressed in this study. However, most patients were diagnosed after 1990 (as illustrated in Figure 1) where treatment strategies and general is comparable to a large extent with present standards.

Diagnostic tools and standards have also changed over 30 years, and hence, some of our patients received the diagnosis with less work-up than the present standard with the risk of including patients in the cohort that did not have PCD. However, because these patients are also the ones followed for the longest time with continuous clinical symptoms of PCD, their diagnoses seem likely to be correct.

Because data were collected post hoc, compliance with treatment was not recorded in our study. Obviously, lacking compliance may influence outcome of lung function. However, we believe that through close monitoring of patients, asking for monthly sputum samples or endolaryngeal suction, and seeing them at 3-month clinical visits permanently after diagnosis, we had the opportunity to keep a high level of compliance.

Being a genetic disease (17), various genetic defects may well cause different lung function outcomes. Unfortunately, we were not able to include data on genetic PCD variants in this study.

Situs inversus was found in only 35.1% of patients, which was surprisingly few. Also, an unexpected high incidence of combined outer dynein arm and inner dynein arm defect was found (18.9%). Most of these outer dynein arm–inner dynein arm defect results are from early in the study period where the EM pictures were of poorer quality compared with the present time; some of these may be outer dynein arm defects misinterpreted as outer dynein arm–inner dynein arm defects because of difficulty in identifying the inner dynein arms.

With increasing focus on PCD across the United States and Europe, a growing population of adult patients with PCD may be expected. The established European Respiratory Society PCD Task Force in 2008 is a welcomed initiative with the opportunity to collect and compare data from many PCD centers in Europe (1, 18, 19).

Hopefully, with such initiatives, standards will follow for evidence-based treatment protocols, for how often patients with PCD should be monitored and how aggressive they should be treated with antibiotics and physiotherapy if lung function starts to decline, for how to handle P. aeruginosa infections, and for the impact on prognosis of these infections.

This study demonstrated PCD to be of serious threat to lung function already in early childhood, because one-third of preschool children had a first measured FEV₁ below 80% of predicted. Across the entire cohort, one-third of the patients lost more than 10 percentage points of FEV₁ during follow-up, despite treatment. Surprisingly, early diagnosis was not found to be protective against further decline in lung function, and the initial level of lung function did not predict the further course of lung function. Finally, no short-term effect of initial treatment on first FEV₁ or FVC could be demonstrated in children less than 10 years of age when comparing those that received treatment months to years before first lung function measurements with those less than 10 years of age who were newly diagnosed at their first lung function. This definitely raises the concern of a more pessimistic perspective on the prognosis of lung function in PCD, compared with previous knowledge (6). However, PCD is a heterogeneous disease, and the implications of this in terms of prognosis are still largely unexplored. Moreover, treatment guidelines for PCD are mainly derived from CF care and so far lack evidence of the needed intensity of care. Hence, one may speculate whether closer monitoring or additional therapeutics, such as anti-inflammatory drugs, could have a place in this disease, because our strategy primarily based on antibiotics is unsatisfactory in preventing deterioration. Moreover, the exact lung pathology of this disease has yet to be clearly established, and illumination of probable genotype-phenotype relations may be warranted.

Given the large fraction of young children with decreased lung function and the highly unpredictable variability in further course of lung function, we believe this study emphasizes that prognosis of lung function in PCD is indeed very complex, despite management in a single national center. We therefore found our study to stress that PCD is a disease that needs close attention in every patient and at any age, including patients with early diagnosis, because neither age nor initial level of lung function at diagnosis were determining factors on the course of lung function.