The prevalence of thanatophoric dysplasia and lethal osteogenesis imperfecta type II in Northern Ireland - a complete population study.

The minimum prevalence of lethal Osteogenesis imperfecta type II, thanatophoric dysplasia and achondroplasia were derived following detailed case note review of all perinatal lethal skeletal dysplasias (SD) in Northern Ireland over a 12 year period. Multiple sources of ascertainment, including genetic notes, radiological reports and post mortem findings, were used. 39 cases were identified. Thanatophoric dysplasia was the commonest diagnosis made (22), followed by osteogenesis imperfecta type II (four children) and achondroplasia (two children). Eleven other diagnoses each occurred once in the 12 year period. The minimum prevalence range, per live births, of each of the common skeletal dysplasias in Northern Ireland has been calculated; thanatophoric dysplasia 0.80/10,000, osteogenesis imperfecta type II 0.15/10,000 and achondroplasia 0.07/10,000. The prevalence range for thanatophoric dysplasia is much higher than reported in previous studies. We discuss reasons for the prevalence figures obtained.

BACKGROUND

The genetic skeletal dysplasias are a large group of disorders that can present at any age and with a very varying phenotype. A proportion of these conditions are lethal due to significant dysplasia of the thoracic cavity, leading to pulmonary hypoplasia. Phenotypic variation remains, however, with some fetuses dying in the early second trimester, while some continue to full term and may live for a few months after birth. Lethal skeletal dysplasias can be picked up on antenatal ultrasound scan, with many having specific diagnostic features that can be detected from the second trimester onwards1,2. All have short ribs, with small chest circumference and abnormal chest to abdominal circumference ratios for gestational stage. Thanatophoric dysplasia (OMIM 187600 and 187601) is characterised by ‘French telephone receiver’ shaped femora in type I, and a cloverleaf-shaped skull in addition in type II. Osteogenesis imperfecta type II (OMIM 166210) causes multiple fractures, particularly in the long bones, which leads to bony deformity. Although some fetuses in our review had achondroplasia (OMIM 100800), it should be noted that the vast majority of infants with this diagnosis survive.

Accurate figures on incidence and prevalence of lethal skeletal dysplasias are difficult to estimate. Not many studies have published prevalence figures3–5, and only one of these was published in the past 20 years. Our population of 1.75 million, is homogeneous, relatively static, with low immigration and emigration, has a defined geographical area and a health care system where our regional genetics service covers the entire geographical area, allowing prevalence to be calculated precisely. Ascertainment was from multiple sources to improve accuracy.

METHODS

We carried out a review of all cases of lethal skeletal dysplasias, defined as perinatal deaths due to skeletal dysplasias occurring up to 6 months postnatally. The study period covered 12 years, from 1st January 1995 to 31stDecember 2006. Patients were identified from a variety of sources. A search was carried out using our departmental genetic database for all patients diagnosed with a skeletal dysplasia during this period. The records database from the radiology department in the Royal Belfast Hospital for Sick Children, where all regional and post-mortem paediatric radiographs are sent for reporting, provided a list of all children/fetuses with a diagnosis of a skeletal dysplasia within the study period. Post mortem reports were obtained for all patients with a lethal skeletal dysplasia. Genetics charts were reviewed retrospectively and data was collected on diagnosis, genetic and non-genetic test results, parental details and family history.

STATISTICS

Descriptive statistics were carried out on data stored in an Excel spreadsheet. The prevalence was calculated based on total population, and on numbers of total and live births, obtained from the Northern Ireland Statistics and Research Agency [http://www.nisra.gov.uk].

RESULTS

39 patients were identified who died from a skeletal dysplasia in the study period. Of these, 37 (94.9%) were an intra-uterine death or died in the immediate post-natal period. One child with achondroplasia survived to six months of age. Another, with suspected Pacman dysplasia, died at 11 weeks of age.

The diagnosis was thanatophoric dysplasia, type I or II, in the majority: 22 cases (56.4%), followed by osteogenesis imperfecta type II, 4 cases (10.3%) and achondroplasia, 2 cases (5.1%). There were 11 other fetuses that died from conditions with a significant skeletal component, only five of whom had a definite diagnosis (Table I).

TABLE I

Breakdown of ‘Other Diagnoses’

	Number
Undiagnosed SD, dysmorphic	1
Spondyloepiphyseal dysplasia congenitaa	1
Hypophosphatasiab	1
Osteocraniostenosis / atenolol teratogenesis	1
Severe costovertebral dysplasia, Jarcho levinc	1
Weyer's syndrome, Acrofacial dysostosisd	1
Sirenomelia	1
Apert syndromee	1
Pseudoachondroplasia, Hypochondroplasiaf	1
Pacman dysplasiag	1
Larsen syndromeh	1

type II collagen disorder

bent bone dysplasias

dysostosis with predominantly costal or vertebral involvement

dysostosis with predominantly craniofacial involvement

craniosynostosis syndrome

FGFR3 group of dysplasias

Filamin group of dysplasias

GENETIC TESTS

A karyotype result was available in 20 (51.3%) babies. Karyotype analysis failed in a further nine (23.1%), though three had trisomy 13, 18 and 21 ruled out by FISH (fluorescent in situ hybridization). Thus, a total of 29 (74.3%) had a karyotype result requested (tables II and III). There were equal numbers of males and females. There were two coincidental associations; one child with trisomy 21 (Down's syndrome, later confirmed by karyotype) and achondroplasia6, and one with 22qll.2 deletion syndrome (Di George syndrome) and thanatophoric dysplasia. This was the only fetus with thanatophoric dysplasia that had a confirmed FGFR3 mutation as this diagnosis is commonly made radiologically. The two patients tested for 22qll.2 deletion syndrome both had characteristic dysmorphic facies, along with cardiac abnormalities; truncus arteriosus in the fetus who tested positive and dextroposition in the fetus who tested negative. The fetus tested for PTPN11 was thought clinically to have Noonan's syndrome, and had bilateral pleural effusions, polyhydramnios and widely spaced nipples.

TABLE II

Karyotype results

	Thanatophoric Dysplasia	O.I. Type II	Achondroplasia	Other	Total
46XX	2	1	0	6	9
46XY	6	1	1	1	9
47XX	0	0	1	0	1
Normal, sex not reported	1	0	0	0	1
FISH T13/18/21 normal	3	0	0	0	3
Failed growth	5	0	0	1	6
Not tested	5	2	0	3	10

TABLE III

Results of other genetic tests

	Thanatophoric Dysplasia	O.I. Type II	Achondroplasia	Other
22ql 1.2 deletion detected	1	0	0	0
FISH 22ql 1.2 normal	0	0	0	1
FGFR3 mutation detected	0	0	1	0
PTPN11 normal		0	0	1

NON-GENETIC TESTS

These consisted of antenatal ultrasound scans, radiographs (either ante or post-mortem), and post mortem studies (Table IV). Radiographs were the most accurate diagnostic test, and were carried out in 38 cases (97.4%). They confirmed the diagnosis in the majority (84.6% of total, 86.8% of those with radiology), while all remaining cases showed generalised abnormalities. One baby did not have radiology; a diagnosis of thanatophoric dysplasia was made on the basis of a characteristic antenatal ultrasound appearance, the parents declining further investigation.

TABLE IV

Non-genetic diagnostic testing

	Thanatophoric Dysplasia	O.I. Type II	Achondroplasia	Other	Total
Total	22	4	2	11	39
Radiology:
4 Confirmed diagnosis	21	4	2	6	33
Generalized abnormality	0	0	0	5	5
Not carried out	1	0	0	0	1
PM:
Yes	20	3	1	9	33
No	2	1	1	2	6
Scan:
Normal	0	0	0	1	1
Abnormal	22	4	2	10	38

There were 20 patients (84.6%) who had a post mortem. Dysmorphic features were frequently documented. Non-skeletal defects were rare. Cardiac defects occurred in three fetuses as follows: Fallot's tetralogy associated with trisomy 21 and achondroplasia, truncus arteriosus associated with 22qll.2 deletion syndrome and thanatophoric dysplasia, and dextroposition associated with possible Jarcho Levin syndrome. Hypospadias and an inguinal hernia occurred in a fetus with possible Pacman dysplasia. Hydrocephalus and cleft lip and palate were found in a fetus with thanatophoric dysplasia. These associations were considered to be coincidental. One child, who had a normal antenatal ultrasound scan, and died at 39 weeks gestation, had a diagnosis of severe costovertebral dysplasia, possibly due to Jarcho Levin syndrome.

None of the fetuses were conceived by assisted reproductive techniques. Gestational ages at diagnosis and birth details were not accessible for every case. Although most abnormalities were picked up at the 20-week anomaly scan, some were seen as early as 15 weeks gestation. Most were diagnosed less than a week from birth (53.8%), though five fetuses (12.8% of total, 25% of those in whom gestation known) survived to full term (data not shown).

In total, 22 families (56.4%) were seen by our genetics service. There was no family history of skeletal dysplasia in any of these. Prevalence was calculated in three ways, based on total population, total number of births and total number of live births (Table V). As our numbers of cases of skeletal dysplasias fluctuate each year, a prevalence range was obtained, along with total prevalence for the whole study period (Tables VI, VII).

TABLE V

Population details

Year	Total population	Total number of births	Total number of live births
95	1649100	23838	23693
96	1661800	24535	24382
97	1671300	24218	24087
98	1677800	23790	23668
99	1679000	23089	22957
00	1682900	21605	21512
01	1698300	22074	21962
02	1696600	21507	21385
03	1702600	21756	21648
04	1710300	22431	22318
05	1724400	22417	22328
06	1741600	23361	23272
Total	20295700	274621	273212

Data from Northern Ireland Statistics & Research Agency ref [http://www.nisra.gov.uk]

TABLE VI

Prevalence calculations per 10,000 population and total births

Year	Thanatophoric Dysplasia			O.I. Type II			Achondroplasia
	No.	PTP	PTB	No.	PTP	PTB	No.	PTP	PTB
95	0	0	0	0	0	0	0	0	0
96	0	0	0	1	0.006	0.41	0	0	0
97	1	0.006	0.41	1	0.006	0.41	0	0	0
98	3	0.018	1.26	0	0	0	1	0.006	0.42
99	1	0.006	0.43	0	0	0	0	0	0
00	1	0.006	0.46	1	0.006	0.46	0	0	0
01	2	0.012	0.91	1	0.006	0.45	0	0	0
02	2	0.012	0.93	0	0	0	0	0	0
03	3	0.018	1.38	0	0	0	0	0	0
04	3	0.018	1.34	0	0	0	0	0	0
05	2	0.012	0.89	0	0	0	1	0.006	0.45
06	4	0.023	1.71	0	0	0	0	0	0
Total	22	0.011	0.80	4	0.002	0.15	2	0.001	0.073

Key: PTP = prevalence based on total population, PTB = prevalence based on total number of births

TABLE VII

Prevalence calculations per 10,000 births

	Thanatophoric Dysplasia	O.I. type II	Achondroplasia
Prevalence range	0-1.71	0-0.46	0-0.45
Donnelly (this study)	0.80	0.15	0.07
Orioli4	0.09	0.43	0.46
Stall5	0.28	0.64	0.64
Waller3 - Atlanta	0.25		0.39
Waller3 - Iowa	0.30		0.41
Waller3 - Oklahoma	0.21		0.60
Waller3 - Texas	0.21		0.39

DISCUSSION

The relative frequencies of the three commonest dysplasias in our population study are as follows; 22 cases (56%) of thanatophoric dysplasia, four cases (10%) of osteogenesis imperfecta type II and two cases (5%) of achondroplasia. Overall these three diagnoses make up 72% of our population. This breakdown differs from that previously observed1, and our proportion of babies with thanatophoric dysplasia is significantly higher.

Our prevalence rates were calculated according to three different population statistics. As expected, prevalence based on total population was much lower than prevalence based on total and live births, which were very similar. However, most studies in the literature state prevalence based on total births3–5. Our minimum prevalence rates per 10,000 total births are as follows; 0.80 (range 0-1.72) for thanatophoric dysplasia, 0.15 (range 0-0.46) for osteogenesis imperfecta type II and 0.073 (range 0-0.45) for achondroplasia. The range is higher than the overall figure for osteogenesis imperfecta type II and achondroplasia as the fact that, in most years, no children died perinatally with these diagnoses, brings the overall average prevalence down. Our prevalence ofthanatophoric dysplasia is higher than that previously reported3, when our paternal age is taken into account; (average age 33.4 years, range 29 to 40 years). We expect our data to be more accurate than the prevalence rates previously published. Firstly, almost complete ascertainment should be possible in Northern Ireland due to the static nature of our population, contained in a well-defined geographical area, which is served by a single genetics unit. A central paediatric hospital reports on all X-rays and carries out all post mortems. Secondly, our figure for those in whom the diagnosis was uncertain, 20%, is much lower than in previous studies4. The prevalence of osteogenesis imperfecta type II and achondroplasia were much lower than previous reports. All of our cases represent new mutations. Unfortunately, detailed analysis of parental age was not possible as this information was not available in every case. Information on birth details, such as proportion of terminations, was also incomplete. However, this should not affect our prevalence rates as termination would have been carried out after the diagnosis of a lethal skeletal dysplasia had been made.

A karyotype was requested in almost 75% of cases. Six out of the 29 cultures (20%) failed to grow and this would not be unexpected from post mortem samples. It should be noted that amniocentesis is much more accurate7. Although the majority of karyotypes were normal, some interesting coincidental findings occurred. One child had trisomy 21 and achondroplasia, which has rarely been observed6,8. One had 22ql 1.2 deletion syndrome and thanatophoric dysplasia; this association does not appear to have been previously reported.

Radiology was the best investigative tool, confirming the diagnosis in the vast majority of cases (87%). There were 33 (85%) fetuses that had a post mortem. Non-skeletal defects were extremely uncommon and were much more likely to occur in the setting of a second genetic diagnosis.

Here, we publish the most accurate prevalence figures to date for lethal skeletal dysplasias. We have shown that thanatophoric dysplasia is more common than previously thought. By raising awareness of this important group of disorders, we hope to improve antenatal diagnostic accuracy and genetic counselling.