Heart disease in patients with osteogenesis imperfecta — A systematic review

International Journal of Cardiology 196 (2015) 149–157

Contents lists available at ScienceDirect

International Journal of Cardiology journal homepage: www.elsevier.com/locate/ijcard

Review

Heart disease in patients with osteogenesis imperfecta — A systematic review Hamoun Ashournia a,b,1, Frank Ted Johansen a,c,1, Lars Folkestad a,d,e,⁎, Axel C.P. Diederichsen f, Kim Brixen a,d a

Department of Endocrinology, Odense University Hospital, Odense, Denmark Department of Neurosurgery, Aarhus University Hospital, Aarhus, Denmark Department of Plastic Surgery, Aalborg University Hospital, Aalborg, Denmark d Institute of Clinical Research, University of Southern Denmark, Odense, Denmark e Department of Endocrinology, Hospital of Southwest Denmark, Esbjerg, Denmark f Department of Cardiology, Odense University Hospital, Denmark b c

a r t i c l e

i n f o

Article history: Received 27 March 2015 Received in revised form 2 June 2015 Accepted 12 June 2015 Available online 14 June 2015 Keywords: Osteogenesis imperfecta Cardiovascular disease Systematic review

a b s t r a c t Introduction: Osteogenesis imperfecta (OI) is a rare, inherited systemic connective tissue disease that causes decreased bioavailability of collagen type 1. Collagen type 1 is the most abundant connective tissue in the body and a key part of many organs. While the bone phenotype in OI is well described, less is known about the effects of decreased collagen on other organs. In the heart, collagen type 1 is present in the heart valves, chordae tendineae, annuli fibrosi and the interventricular septum. It is thus likely that the heart is affected in OI. Objectives: The aim of this systematic literature review was to investigate whether patients with OI have an increased risk of cardiovascular disease compared to healthy adults. Data sources: PubMed, Embase and key scientific meetings were searched for publications fulfilling the inclusion criteria. Study selection: Studies were selected if at least one patient with OI was described as having cardiovascular disease. The articles should be written in English, French, Italian, Spanish, German, Norwegian or Danish or have an English abstract. Data extraction: Data were extracted by HA, FTJ and LF using a predefined protocol. Results: A total of 68 studies were included in the review, comprising 51 case reports, 8 small case series (n b 10 patients), 4 large case series (n ≥ 10 patients) and 5 cross-sectional studies comparing patients and controls. Together, the papers comprised 499 patients and covered 45 years of medical literature. The most commonly reported heart diseases amongst the patients with OI were valvulopathies and increased aortic diameter. Findings in the large case series and the cross-sectional studies were broadly similar to each other. Conclusion: The findings support the hypothesis that patients with OI have increased risk of heart disease compared to healthy controls. It is biologically plausible that patients with OI may have an increased risk of developing heart disease, and valve disease in particular. © 2015 Elsevier Ireland Ltd. All rights reserved.

1. Introduction Osteogenesis imperfecta (OI) is a rare, inherited systemic connective tissue disease characterized by varying degrees of bone fragility and thus increased prevalence of fractures, limb deformities, hearing loss and often a blue scleral hue. Currently, 12 subtypes of OI have been defined [1]. Types I–IV are autosomal dominantly inherited and were first defined by Sillence et al. in the 1970s [2,3]. During the last decades, several autosomal recessive forms of OI (types V–XVII) have been

⁎ Corresponding author at: Kløvervænget 10, 6th floor Endocrinology Research Unit, Department of Endocrinology, Odense University Hospital, Sdr. Boulevard 29, DK-5000 Odense C, Denmark. E-mail address: [email protected] (L. Folkestad). 1 HA and FTJ share the first authorship of this article.

http://dx.doi.org/10.1016/j.ijcard.2015.06.001 0167-5273/© 2015 Elsevier Ireland Ltd. All rights reserved.

described. In OI, the severity varies from phenotypes with slightly increased fracture risk to perinatally fatal forms. The prevalence of OI is approximately 22 per 100,000 new-born, with a population prevalence of 11 per 100,000 [3]. In approximately 90% of the autosomal dominant cases, OI is caused by mutation in the COL1A1 or COL1A2 genes [4]. These genes code for the collagen type 1 pro-strands that after folding become normal type-1 collagen [5]. In the autosomal recessive cases, the mutations often affect genes that code for the chaperone proteins that facilitate the folding of the collagen pro-strands [5]. Finally, gonadal mosaicism and de novo mutations are seen [6]. While the bone phenotype in OI is well described, less is known about the effects of decreased content of collagen type 1 and altered collagen structure in other tissues and organs. Heart valves, chordae tendineae, annuli fibrosi and the interventricular septum contain

150

H. Ashournia et al. / International Journal of Cardiology 196 (2015) 149–157

collagen type 1 [7]. Approximately 74% of the collagen content of the mitral valve is collagen type I [8]. Chordae tendineae have a central part comprised of compact collagen [7]. Collagen fibres in ventricular myocardium contribute to the tensile stiffness and maintain the architecture of the myocytes [9]. The aorta and most arteries are rich in both type III and type I collagen [10]. Many cardiac disorders are associated with an accumulation, depletion or restructuring of the collagen matrix [11]. Changes to the connective tissue (collagen type III, V or fibrillin 1) cause aortic disease in Marfan's Syndrome and Ehlers– Danloss' Syndrome [12]. In the Aga2 OI mouse model, Thiele et al. [13] showed that in both the severe (i.e., postnatally fatal) and the mild phenotype (i.e., animals surviving to adulthood), mice had cardiovascular disease. Severely affected mice had enlarged septum, right ventricular hypertrophy and significantly lowered ejection fraction (EF) compared to wild type mice [13]. The type 1 collagen was disoriented and there were fewer and thinner collagen fibrils in the cardiac tissue [13]. Echocardiograms were normal in the mild phenotype, but showed increased QRS amplitude and increased J–T intervals [13]. The OIM mouse model, in which mutations in the COL1A2 gene lead to defective skeletal development, smaller stature and skeletal fragility [9], has also been studied for cardiovascular manifestations [9,10,14]. Thus, Weis et al. [9] found that homozygote OIM mice had decreased collagen area fraction with lower fibre number density compared to wild type mice. The homozygote OIM had thicker sepal and posterior walls, but normal systolic function as evaluated by echocardiography [9]. There were no in vivo data regarding the heterozygote mice, and there were no between-group differences regarding posterior wall thickness or left ventricular external diameter [9]. The objective of this systematic review was to examine whether patients with OI have an increased prevalence of heart disease compared to healthy controls. 2. Methods 2.1. Literature search We searched PubMed for publications from 1962 until 2013 week 29 with the search phrase ‘(osteogenesis imperfecta) AND cardiovascular disease [Mesh Term]’. We also searched Embase for publications from 1974 until 2013 week 29 using the search phrase ‘osteogenesis imperfecta AND cardiovascular disease’, where “cardiovascular disease” was entered as a broader term search word. Using the online resources from the American Society for Bone and Mineral Research, we searched their annual meeting abstract database for presented abstracts using the search words ‘Osteogenesis Imperfecta’. Presented abstracts are available for 2009–2013. The bibliography of selected articles was screened for further articles of interest. Papers in English, French, Italian, Spanish, German, Norwegian and Danish were included in the review. Papers in other languages were only included if an English summary was present. 2.2. Study selection and data collection process All abstracts were screened for eligibility by FJT, HA and LF. All types of publications reporting studies on heart disease in patients with OI were deemed eligible. In case of disagreement between the three reviewers, consensus was reached through dialogue. Case reports and case series were included if they described at least one patient with OI and mentioned a cardiovascular diagnosis. Cross-sectional studies were included if there was a clear statement of the criteria for the OI diagnosis, which cardiovascular diseases were found and how these differed between cases and controls. In all studies, cardiovascular diagnoses were only accepted if they were supported by transthoracic or trans-oesophageal echocardiography, cardiac catheterization,

surgical or autopsy findings. Papers not meeting the inclusion criteria were excluded. LF extracted data from the included articles as shown in Table 1. 2.3. Classification of OI type and heart disease For all studies, data regarding OI type were extracted as stated by the article authors. The review authors did not try to assign an OI type from the information given in the case report or case series. Heart disease was classified as heart failure (including cardiomegaly), coronary artery disease, heart valve disease (stenosis and regurgitation), aortic disease (including aneurysms), congenital malformations or other. Patient age was reported as median [range min–max] or mean ± SD, and gender as a ratio based on the numbers stated in the original articles. 3. Results 3.1. Study selection A flow diagram of search results is shown in Fig. 1. The PubMed search provided 357 citations, EMBASE 37 citations and a search of the abstracts from annual ASBMR meetings resulted in 46 abstracts. After adjusting for duplicates, 387 remained. Of these, 299 studies were discarded after reviewing the abstracts as they did not meet the inclusion criteria or met one or more of the exclusion criteria. The full-texts of the 88 remaining studies were reviewed and a further 21 articles were excluded. One additional study was found through the reference lists of the included articles. Thus, a total of 68 studies were included in this review, comprising 51 case reports of OI [15–64], 8 small OI case series (containing less than 10 patients) [65–72], 4 large OI case series (containing 10 or more patients) [13,73–75] and 5 cross-sectional comparative studies including controls [76–80]. 3.2. Participants In the case reports and case series and all but one cross-sectional study, the OI diagnosis was based on the patient's clinical characteristics. The 51 case reports and 8 small case series described 70 patients with OI and heart disease (see Appendices 1 and 2); amongst the 51 case reports these were two unborn foetuses (gender unclear), 39 men and 10 women with a median age of 37 years [range 28 gestational weeks to 65 years]. The 8 small case series included 16 men and 3 women with a median age of 42 years [range 17–63]. Only 12 of these 70 patients had a clear OI type stated: 13 had mild OI, 3 moderate OI and 2 severe OI. Thirty-five of the patients had a family history of OI. Forty-two patients had blue sclera and two had not, while this information was missing in the remaining 26 patients. Fifty-one patients had Table 1 Data items extracted from the included articles. Data item

Case report

Case series

Cross-sectional studies

Patient gender (ratio) Patient age (median/mean)

X X

X X

X X

Osteogenesis imperfecta characteristics OI type (where stated) Scleral hue Fracture history Dentinogenesis imperfecta Hearing loss Diagnostic criteria for OI

X X X X X

X X X X X

X

Cardiovascular disease Diagnosis stated at an individual level Diagnostic modality used Cardiovascular outcome stated in article aim

X X

X

X X X

H. Ashournia et al. / International Journal of Cardiology 196 (2015) 149–157

151

Fig. 1. Search strategy and study selection.

experienced multiple bone fractures. Information on dentinogenesis imperfecta was missing for 58 patients, and information about otosclerosis or pre-senile deafness was missing for 45 patients. The four large case series comprised 233 patients in total. White et al. [73] included 20 patients (15 women), clinically diagnosed and grouped according to the Sillence classification: I (14 patients), IV (3 patients) and unclassifiable (1 patient). The patients were seen after they were referred to a regional metabolic bone disease clinic or for

orthopaedic treatment. The median age of the participants was 41 years [range 13–68]. Hortop et al. [75] included 53 women and 56 men with OI, ranging from 1 to 75 years of age (mean age 27 years). Patients were diagnosed based on their clinical characteristics and grouped according to Sillence into 4 groups (I, II, III or IV). Eightythree patients had OI type I, 16 patients had OI type III and 10 patients had OI type IV. The participants had been referred to one of three clinical centres for genetic counselling and were identified by two of the

152

H. Ashournia et al. / International Journal of Cardiology 196 (2015) 149–157

Table 2 Pooled data from the 70 patients described in the case reports or small case series. Number of cases

% of total cases described

Heart murmur on auscultation Ventricular hypertrophy on ECG Heart failure including cardiomegaly Coronary arteriosclerosis

37 22 1 3

53% 31% 1.4% 4%

Valvular disease Aortic stenosis Aortic regurgitation/insufficiency Mitral valve stenosis Mitral valve regurgitation/insufficiency Tricuspid stenosis Tricuspid regurgitation/insufficiency

0 40 0 24 0 2

NA 57% NA 34% NA 3%

Aortic disease Aortic dissection and aneurysms

13

19%

1 1

1.4% 1.4%

Congenital malformations Persistent foramen ovale Coarctation of the aorta Other Coronary artery dissection Supravalvular aorta stenosis Left atrial rupture

1 1 1

1.4% 1.4% 1.4%

investigators [75]. Vetter et al. [74] included 58 children aged 1– 16 years and followed them for an unspecified number of years. Eighteen children (mean age 9 [range 2–16]) were classified as OI type I, 25 children as OI type III (mean age 9 [range 1–15]) and 15 as “unclassified” (mean age 6 [range 1–10]). No information was given on how the patients were diagnosed. Thiele et al. [13] included 23 children and young adults with OI type III (11 boys and 12 girls) and 23 children and young adults with OI type IV (8 boys and 15 girls) with a mean age of 12 ± 4 years. All participants had progressive deformities and moderately severe forms of OI [13]. In 40 of the 46 patients, a mutation to either the COL1a1 or COL1a2 genes were identified [13]. All the participants were followed through the NIH Clinical Centre and the Children's National Medical Centre [13]. The 5 case–control studies comprised four patient cohorts including a total of 196 patients. Kalath et al. [79] included 31 patients with OI and 50 (11 females) healthy controls. The healthy controls were divided into three age groups, young (b35 years of age), middle-aged (35–55 years of age) and old (N55 years of age). The controls were excluded if they had a history of connective tissue disorder, heart disease or hypertension, obesity, abnormal or technically insufficient echocardiograms. No data were provided on individual OI patients, but seven were categorized as type IV OI and the remainder as type I OI. All between-group comparisons were done comparing age-matched OI and controls. We assumed that the age range of the OI patients was similar to that of the controls. Migliaccio et al. [78] included 40 patients (21 females and 19 males, mean age 40 years) with OI. All were outpatients followed and monitored for bone and other systemic disorders. Thirty-five patients were classified as OI type I, three patients as OI type III and two patients as OI type IV. The controls were 20 healthy men and 20 healthy

women with a mean age of 41 years. Patients and controls were excluded if they had a family history of cardiovascular disease, previous history of systemic arterial hypertension or coronary artery disease, endocrine disorders or metabolic syndrome. Jiménez et al. [80] included 26 patients, of whom 70% had mild OI (not classified according to Sillence). The 19 women and 7 men with OI had a mean age of 38.8 ± 11.95 years, and the 18 women and 7 men who served as controls had a mean age of 32.5 ± 12.13 years. Radunovic et al. [76,77] included 41 male and 58 female patients with OI in both cross-sectional studies. Seventy-seven of the patients were classified as OI type I, 10 as type III and 11 as type IV (one patient did not meet the classification criteria and was not assigned an OI type). All the patients were diagnosed at a National Resource Centre for Rare Disorders. Mean patient age was 43.9 ± 12.3 years. The studies included 52 age- and sex-matched healthy controls without any history of OI, cardiovascular disease, diabetes or hypertension. 3.3. Cardiac examinations Several diagnostic modalities were employed in the patients described in the 59 small case series and case rapports (see Appendices 1 and 2). In 1983, White et al. [73] used echocardiography to visualize the aortic root diameter, the aortic valve and mitral valve. In 1986, Hortop et al. [75] used M mode echocardiogram to measure the cardiac dimensions. In 1987, Kalath et al. [79] used M-mode echocardiography to visualize the aortic root diameter and wall diameter. The participants were lying supine or in a left decubitus position. In 1989, Vetter et al. [74] used 2D echocardiography, taking special care to visualize the mitral and aortic valves. In 2009, Migliaccio et al. [78] performed 2D echocardiography with continuous and pulse-wave Doppler. Right and left ventricular dimensions and left ventricular diastolic and systolic function were measured. In 2010, Jiménez et al. [80] used 2D echocardiography and colour Doppler methods to assess the cardiac function in patients and controls. Right and left ventricular dimensions and left ventricular diastolic and systolic function were reported. In 2012, Thiele et al. [13] used 2D echocardiograms including spectral Doppler to evaluate cardiac function. In 2012, Radunovic et al. [76,77] performed standard echocardiography and diagnoses were made in accordance with the guidelines of the American Society of Echocardiography. Both right and left ventricle measurements were reported in the two papers. All the studies reported on the presence of valvulopathies and their severity. 3.4. Findings: case reports and small case series The pooled data from the 59 case reports and small case series are presented in Table 2 and a detailed presentation of all cases is seen in Appendix 1 and Appendix 2. Many patients had more than one cardiovascular disease. Most patients presented with either heart murmur or dyspnoea as their primary complaint. The most frequent cardiovascular disease amongst the OI patients was valvular regurgitation or valvular insufficiency. Forty patients had aortic valve regurgitation, 24 had mitral

Table 3 Characteristics of participants in the included cross-sectional studies. Study

Year

No of patients (n)

No of controls (n)

Age cases (years)

Age controls (years)

Sex cases (F/M)

Sex controls (F/M)

Type I (n)

Kalath et al. [79] Migliaccio et al. [78] Jiménez et al. [80] Radunovic et al. [76,77]

1987 2009 2010 2012

31 40 26 99

50 40 25 52

NA 40 38.8 ± 11.95 43.9 ± 12.3

NA 41 32.5 ± 12.13 43.7 ± 13.9

NA 21/19 19/7 57/42

NA 20/20 18/7 30/22

24 35 NA 77

NA = Not applicable. ⁎ Indicates significant difference between cases and controls p b 0.05.

H. Ashournia et al. / International Journal of Cardiology 196 (2015) 149–157

valve regurgitation and 2 patients presented with tricuspid regurgitation. Several patients had a combination of valvulopathies. Almost 1 in 5 patients had experienced aortic dissection or aortic aneurysm. 3.5. Findings: large case series Using old-fashioned echocardiography, White et al. [73] visualized the aortic valve in 19 patients and found aortic regurgitation in one patient. In 17 of the other patients, the aortic cusps were described as thin. In 10 patients the mitral valve cusps were thinner than expected, and one patient had mitral prolapse. The aortic root diameter was enlarged with a mean diameter of 3.19 ± 0.55 cm. The end diastolic diameters were normal in all patients but one patient had aortic regurgitation. In 1986, Hortop et al. [75] noted aortic regurgitation in two patients and mild aortic stenosis in one patient from a total of 66 patients. Aortic root dilatation was found in 8 out 66 individuals [75]. The left atrial dimensions were somewhat less than expected in a healthy population (88% of mean expected size ± 11%), and left ventricular dimensions were slightly greater than expected (103% of mean expected size ± 12%) [75]. In 1989, Vetter et al. [74] found congenital heart malformations in one child with OI type I (aortic stenosis), four children with OI type III (2 atrial septal defects and 2 mitral valve prolapse) and one of the unclassified patients (Fallot's tetralogy). Two patients developed mitral valve prolapse during the study. Using normal values published by Lange et al. [81], Vetter et al. [74] calculated a standard deviation score and found that 95% of patients with OI type I were within ±2 SD of the normal values for aortic, left atrial and left ventricular end diastolic diameter, and septal and posterior wall thickness. In the patients with OI type III, 28% of the aortic diameters, 40% of the septal thickness values and 68% of the posterior wall thickness values were above + 2 SD limits. Using modern echocardiography in 2012, Thiele et al. [13] found that, in both the type III and IV groups, 78% of the children and young adults (18 and 23 patients respectively) had one or more valvular or cardiac chamber abnormalities [13]. Thirty-one patients had mild tricuspid valve regurgitation, 12 patients had mitral valve regurgitation (11 mild and 1 moderate), 4 patients had pulmonary valve regurgitation and 3 had aortic valve regurgitation. No ventricular abnormalities were found, but 5 patients had a dilated left atrium and 3 had a left-to-right atrial shunt.

153

also reported mild tricuspid valve regurgitation in 88.9% of patients and 94.2% of controls, while 2% of the patients and none of the controls had moderate tricuspid regurgitation [76]. The authors did not offer an explanation for this, but 11 patients had a mild restrictive ventilation pattern and 2 patients had a mild obstructive pattern that can lead to pulmonary hypertension and moderate tricuspid regurgitation. Mild pulmonary valve regurgitation was found in 36.4% of OI patients and 46.2% of controls (p N 0.05) [76]. Migliaccio et al. [78] found 7.5% patients with mitral valve prolapse, and 10% patients with left ventricular false apical tendon (fibrous or fibromuscular bands that stretch across the left ventricle from the septum to the free wall). Valvular regurgitation, without specific valvular structural alterations, was found in 95% of OI patients (25% mitral regurgitation, 10% mitral and aortic regurgitation, 30% mitral, tricuspid and aortic regurgitation) and in 2.5% of controls (p b 0.001) [78]. Jiménez et al. [80] found no significant difference between patients with OI and controls, reporting valulopathy in 12% of OI patients and 20% of controls (p = 0.44). Migliaccio et al. [78] found increased right ventricular dimensions even without correcting for differences in body surface area. Jiménez et al. [80] found smaller end-diastolic left ventricular diameter in patients with OI. Both Migliaccio et al. and Radunovic et al. found a tendency towards a similar difference, however this was not significant (p N 0.05) [77,78]. Radunovic et al. [76] found the right ventricular dimensions in OI patients to be significantly larger than in healthy controls when corrected for body surface area. Only Jiménez et al. [80] reported data regarding systolic function, but found no significant differences between cases and controls. Migliaccio et al. [78] found diastolic dysfunction in 95% of OI patients and 2% of controls (p b 0.001), while Jiménez et al. [80] found no significant between-group differences. 4. Discussion We included 68 papers in this systematic review, comprising a total of 499 patients. The most commonly reported heart diseases amongst the patients with OI were valvulopathies and increased aortic diameter. Similar results were seen in the large case series and cross-sectional studies. 4.1. Patient classification and epidemiology

3.6. Findings: case–control studies The data from the cross-sectional studies are summarized in Tables 3 and 4. Radunovic et al. [77] found the aortic root diameter to be significantly different between the two groups (20.2 mm ± 2.6 mm in OI patients vs. 19.2 mm ± 2.3 mm, p b 0.05 in controls). Kalath et al. [79] found increased aortic root stiffness in both the young and the middle-aged groups. Radunovic et al. [77] found mild mitral valve regurgitation in 57.5% of OI cases and 61.5% of the controls, however moderate mitral regurgitation was found in 7.1% of OI cases and none of the controls. Aortic valve regurgitation was found in 10.1% of patients with mild OI, 10.1% of patients with moderate OI and none of the controls. The authors

Few of the 70 patients described in the case reports and small case series were characterized according to Sillence by the primary investigators. Standardized reporting in future case series would help analysis comparing the different OI groups. The data presented in the reports did not allow us to classify the patients according to Sillence. Indeed, any post publication classification of phenotype would introduce a bias in the analysis due to the risk of misclassification of the patients. Furthermore, the case reports and small case series covered 45 years of medical publications, and diagnostic criteria and modalities for cardiovascular disease have changed during this period. Disease patterns, access to medical care and treatment options are likely to differ between countries and have also changed over the last 45 years. Indeed, the risk of cardiovascular death increased by 32% between 1990 and 2010 [82].

Type II (n)

Type III (n)

Type IV (n)

Other (n)

Height cases (cm)

Height controls (cm)

Weight cases (kg)

Weight controls (kg)

Body surface Area cases (m2)

Body surface area controls (m2)

0 0 NA 0

0 3 NA 10

7 2 NA 11

0 0 NA 0

NA NA 144 ± 17.42 155.5 ± 20.9

NA NA 162 ± 8.27⁎ 173.3 ± 8.7⁎

NA NA 54.25 ± 13.81 65.7 ± 16.5

NA NA 65.08 ± 16.71⁎ 72.7 ± 14.1⁎

NA NA 1.43 ± 0.24 1.65 ± 0.31

NA NA 1.69 ± 0.22⁎ 1.86 ± 0.20⁎

154

H. Ashournia et al. / International Journal of Cardiology 196 (2015) 149–157

Table 4 Echocardiography findings.

Kalath et al. [79] (mean ± sd)

Migliaccio et al. [78] (mean ± SEM) Jiménez et al. [80] (mean ± sd) Radunovic et al. [77] (mean ± sd) Radunovic et al. [76] (mean ± sd)

Measurement

LV End-diastolic diameter (mm)

RV end-diastolic diameter (mm)

Interventricular septum thickness (mm)

LV posterior wall thickness

Left atrium diameter (mm)

Ejection fraction (%)

Aortic root diameter (mm)

Cases (stiff-stiff subgroup)

NA

NA

NA

NA

NA

NA

NA

Controls

NA

NA

NA

NA

NA

NA

NA

Cases Controls Cases Controls Cases Controls Cases Controls

47.0 ± 0.6 49.0 ± 0.4 43.36 ± 5.75 47.48 ± 5.94* 47.4 ± 6.1 48.1 ± 5.8 NA NA

29.0 ± 0.5 28.0 ± 0.5 NA NA NA NA NA NA

9.3 ± 0.1 10.0 ± 0.1 NA NA 10.01 ± 2.8 8.7 ± 1.5* NA NA

9.2 ± 0.1 9.0 ± 0.1 NA NA 8.9 ± 1.9 8.1 ± 1.3* NA NA

35.0 ± 0.7 35.0 ± 0.3 33.7 ± 5.1 33.5 ± 6.2 NA NA NA NA

65 ± 0.01 66 ± 0.04 NA NA NA NA NA NA

22.0 ± 0.1 22.7 ± 0.5 27.0 ± 4.5 25.0 ± 3.9 20. 2 ± 2.6* 19.2 ± 2.3 NA NA

* Indicates significant difference between cases and controls p b 0.05.

4.2. Bias by including families Reports of rare hereditary diseases may be biased by including family members, as patients within the same family and carrying the same mutation will have a tendency towards the same phenotype. There is also a risk of genetic anticipation when including more than one generation from one family, and thus increasing the severity of a phenotype over time. None of the large case series mentioned whether relatives had been included in the study. In contrast, Hortop et al. [75] tried to minimize the influence of one family's phenotype by including only one member of each family in the analysis. Kalath et al. [79] grouped their patients and controls by age, but included all patients from nine families in the analysis. 4.3. Blinding of echocardiography Blinding of echocardiographic procedures is difficult and was only done by Radunovic et al. [76,77], who performed a post hoc analysis of the images recorded through three heart cycles for each participant for every standard projection. In the other studies, echocardiography was interpreted by the investigator during the procedure and may in theory have been influenced by the investigator's expectations regarding, for example, patients with severe OI. In the only study, in which two independent investigators performed the echocardiography [78], inter- and intra-observer coefficients of variance were 2%. 4.4. Selection bias None of the studies reviewed were population-based. Instead, patients were recruited to the cross-sectional studies from special clinics or programmes. Thus, in the study by Thiele et al. [13] participants were identified through a NICHD, however, no further information was given on how the participants were selected. Similarly, Vetter et al. [74] included little information about the selection of their participants or how these patients were included in their study. Radunovic et al. [76,77] included 99 patients participated in their study, but 259 OI patients (159 aged 25 years or more) registered with the National Resource Centre for Rare Disorders. Thus, selection bias may exist in the published studies. 4.5. Publication bias There is risk of positive publication bias in the case reports and small case series. Descriptions of patients with a rare disease (OI) and a complicating disease (cardiovascular disease) are more likely to be published [83]. It is therefore difficult to give much weight to a single case report or small case series. Although a pattern emerges from the pooled data, as shown in Table 2, it should be kept in mind

that even if there was no real association between OI and cardiovascular disease, patients with OI would have a risk of valvular heart disease of about 2.5% [84]. 4.6. Risk of type-1 error As the articles reported many variables (mean, range, min–max), 1 in 20 variables would be expected to significantly differ between cases and controls by chance alone (with alpha at 0.05). None of the studies reported Bonferroni corrected-p-values. This may give false positive results. 4.7. Risk of type-2 error Each of the case–control studies included between 26 and 99 patients and are thus prone to type 2 error (a false negative result). Using the data regarding aortic root diameter [80], a sample size calculation (with the between-group mean difference and standard derivation regarded as the true difference) indicated that 71 participants in each group for the between-group difference reported to be significant. It is difficult to achieve sufficient statistical power to detect all between-group differences in studies on rare diseases. 4.8. Limitations in study selection and search It cannot be ruled out that we, by our search strategy, have excluded articles that would add to the data gathered in this review. We have used the broadest terms possible to minimize this risk. We also hand search the reference lists and abstracts published from a large scientific meeting. This strategy have previously been found useful and is used as an example in the PRIMSA guidelines [85]. 5. Causation or association? Although no proof regarding causality can be made in observational studies, the Bradford Hill criteria [86] can be used to evaluate whether OI is a likely cause of cardiovascular disease. 5.1. Strength of the association There seems to be an association between OI and cardiovascular pathology, but the association is statistically weak and causality cannot be ruled out. Kalath et al. [79] found increased stiffness in the circumferential direction of aorta. When aortic root diameter was corrected for differences in body surface area between cases and controls, it was found to be significantly increased [77,80]. This finding is further underlined by Hortop et al. [75] who found that many patients had an aortic root diameter greater than the predicted value, and above the 99th percentile

H. Ashournia et al. / International Journal of Cardiology 196 (2015) 149–157

155

Right ventricular basal diameter/body surface area (cm/m2)

Right ventricular mid-cavity diameter/body surface area (cm/m2)

Right ventricular longitudinal diameter/body surface area (cm/m2)

Aortic pressure-strain elastic modulus (Ep) (g/cm2)

Aortic circumferential elastic modulus (Eθ) (g/cm2)

NA

NA

NA

NA

NA

NA

NA NA NA NA NA NA 9.0 ± 2.4 8.6 ± 1.7

NA NA NA NA NA NA 1.9 ± 0.5 1.7 ± 0.3*

NA NA NA NA NA NA 1.7 ± 0.5 1.5 ± 0.5*

Young group 9557 ± 5% Young group 507 ± 21% NA NA NA NA NA NA NA NA

Young group 6838 ± 57% Young group 3060 ± 8% NA NA NA NA NA NA NA NA

for the general population. Both Radunovic et al. [76,77] and Migliaccio et al. [78] found higher incidence of mitral and tricuspid insufficiencies in OI patients but this was not confirmed by Jaminez et al. [80]. Similarly, left ventricle dimensions were larger in OI patients than controls in two studies [77,80], but not in a third [78]. The lack of a clear definition of diastolic dysfunction (which was sometimes referred to as increased ‘stiffness’ of the heart) may explain some of the differences between the studies. Although Migliaccio et al. [78] found diastolic dysfunction in 95% of patients and only 2% of the controls, Jiménez et al. [80] did not show this, even though E-wave velocity was non-significantly decreased in OI patients (p = 0.07). 5.2. Consistency of the observed association Although there were slight differences between the three patient cohorts in the cross-sectional studies [76–78,80], the results point in the same direction suggesting increased prevalence of cardiovascular disease in patients with OI. The between-study variations may be due to case-mix and variations in reporting that make it difficult to draw strong conclusions. Similar findings of valvulopathies and proximal aortic disease were found in both the cross-sectional studies (with controls) and the case series (without controls). 5.3. Specificity of the observed association The data do not show an increased absolute risk of cardiovascular disease amongst patients with OI. None of the OI participants in the case–control studies apparently had symptoms that needed treatment, while 53% of participants in the case reports had a heart murmur and many others had dyspnoea and other cardiovascular symptoms. One could argue that patients with a symptomatic cardiovascular disease will have sought medical help prior to entering a population-based study. This may lead to ‘overly healthy’ OI patients in the populationbased studies, making any differences between cases and controls less evident. Remembering that the absence of evidence is not the evidence of absence. 5.4. Temporal relationship of the association Patients with a disease that increases the risk of fractures and causes varying degrees of loss of function may have lower physical activity levels than otherwise healthy controls [87], and low physical activity itself can lead to cardiovascular disease [88]. Both Jiménez et al. [80] and Radunovic et al. [76,77] reported elevated body mass index (BMI) in OI patients (26 kg/m2 and 27 kg/m2 respectively) but normal BMI in controls (25 kg/m2 and 24 kg/m2 respectively). Furthermore, Radunovic et al. [76,77] found significantly higher systolic and diastolic blood pressure in OI cases than in controls. OI patients may have a higher risk of

Middle age group 2389 ± 53% Middle age group 741 ± 19%

Middle age group 12,760 ± 50% Middle age group 3489 ± 6%

cardiovascular disease due to higher blood pressure, obesity and lower physical activity, and that this risk outweighs the effects of the decreased levels of collagen in the heart. Cardiovascular pathology was found in the two case series including children, however, thus the cardiovascular findings in OI are not likely to be due to lifestyle factors alone. 5.5. Biological gradient In mouse models (OIM and Aga2), more severe cardiovascular pathology has been found in more severe phenotypes [13,14], suggesting a biological gradient that is dependent on the amount of affected collagen. The current literature in humans also supports a biological gradient. No direct between-group comparisons were made by Vetter et al. [74], but markedly more patients with OI type III had cardiovascular findings above the +2SD for age- and gender-matched controls, indicating a relationship between the severity of the collagen default and the cardiovascular phenotype. Radunovic et al. [77] found that the left ventricular diameter, interventricular septum, posterior wall thickness and left ventricular mass were significantly different in patients with type III OI compared to type I or type IV. But a biological gradient was not found by Hortop et al. [75] who found the widest aortic root diameter in patients with OI types I, III and IV from two separate families but found no differences in prevalence of above-normal aortic root diameter between patients with types I and IV compared to type III patients (p = 0.18). 5.6. Plausibility As patients with OI have decreased or altered biosynthesis of collagen type 1, it is biologically plausible that they will also have altered concentration of collagen type 1 in other tissues than bone. The cardiovascular connective tissue in the heart valves, aortic wall and the heart chambers has a high concentration of type 1 collagen [7,8]. The Aga2 severe phenotype mice have disordered matrix collagen in cardiac tissue, with fewer and thinner collagen fibrils [13]. This is correlated to the poorer cardiac function of these mice, underlining a biologically plausible cause of cardiovascular disease in patients with OI. 6. Conclusion This systematic literature search identified 51 case reports of patients with OI, 8 small OI case series, 4 large case series and 5 crosssectional studies comparing the incidence of heart disease in patients with OI and healthy controls. These studies comprised a total of 499 patients and covered 45 years of medical literature. Our review findings support the hypothesis that patients with OI have increased risk of heart disease compared to healthy controls. It is biologically plausible

156

H. Ashournia et al. / International Journal of Cardiology 196 (2015) 149–157

that patients with OI may have an increased risk of developing heart disease and in particular valve disease. However, there is a need for further studies to assess prevalence and risk of heart disease in OI patients. Such cross-sectional and longitudinal studies should be powered to allow adjustment for confounders for cardiovascular disease and should use standardized methods of evaluation. Genetic studies would also be valuable to investigate possible correlations between genotype and phenotype. Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.ijcard.2015.06.001. Conflict of interest The authors report no relationships that could be construed as a conflict of interest. Funding This work was done without external funding Acknowledgements We would like to acknowledge the help of the staff at the Medical Research Library for their efficient retrieval of literature from journals far and wide from the last 65 years. Without them this work would not be possible. We also thank Claire Gudex for reviewing an earlier version of this manuscript. References [1] N. Bishop, Characterising and treating osteogenesis imperfecta, Early Hum. Dev. 86 (2010) 743–746. [2] D.O. Sillence, D.L. Rimoin, Classification of osteogenesis imperfect, Lancet 1 (1978) 1041–1042. [3] P.E. Andersen Jr., M. Hauge, Osteogenesis imperfecta: a genetic, radiological, and epidemiological study, Clin. Genet. 36 (1989) 250–255. [4] F.S. Van Dijk, G. Pals, R.R. Van Rijn, P.G. Nikkels, J.M. Cobben, Classification of osteogenesis imperfecta revisited, Eur. J. Med. Genet. 53 (2010) 1–5. [5] P.H. Byers, S.M. Pyott, Recessively inherited forms of osteogenesis imperfecta, Annu. Rev. Genet. 46 (2012) 475–497. [6] R.D. Steiner, J. Adsit, D. Basel, COL1A1/2-related osteogenesis imperfecta, in: R.A. Pagon, M.P. Adam, T.D. Bird, C.R. Dolan, C.T. Fong, K. Stephens (Eds.),GeneReviews, 1993 (Seattle (WA)). [7] C. Millington-Sanders, A. Meir, L. Lawrence, C. Stolinski, Structure of chordae tendineae in the left ventricle of the human heart, J. Anat. 192 (Pt 4) (1998) 573–581. [8] W.G. Cole, D. Chan, A.J. Hickey, D.E. Wilcken, Collagen composition of normal and myxomatous human mitral heart valves, Biochem. J. 219 (1984) 451–460. [9] S.M. Weis, J.L. Emery, K.D. Becker, D.J. McBride Jr., J.H. Omens, A.D. McCulloch, Myocardial mechanics and collagen structure in the osteogenesis imperfecta murine (oim), Circ. Res. 87 (2000) 663–669. [10] A.G. Vouyouka, B.J. Pfeiffer, T.K. Liem, T.A. Taylor, J. Mudaliar, C.L. Phillips, The role of type I collagen in aortic wall strength with a homotrimeric, J. Vasc. Surg. 33 (2001) 1263–1270. [11] H. Ju, I.M. Dixon, Extracellular matrix and cardiovascular diseases, Can. J. Cardiol. 12 (1996) 1259–1267. [12] B. Callewaert, F. Malfait, B. Loeys, A. De Paepe, Ehlers–Danlos syndromes and Marfan syndrome, Best Pract. Res. Clin. Rheumatol. 22 (2008) 165–189. [13] F. Thiele, C.M. Cohrs, A. Flor, T.S. Lisse, G.K. Przemeck, M. Horsch, et al., Cardiopulmonary dysfunction in the osteogenesis imperfecta mouse model Aga2 and human patients are caused by bone-independent mechanisms, Hum. Mol. Genet. 21 (2012) 3535–3545. [14] B.J. Pfeiffer, C.L. Franklin, F.H. Hsieh, R.A. Bank, C.L. Phillips, Alpha 2(I) collagen deficient oim mice have altered biomechanical integrity, collagen content, and collagen crosslinking of their thoracic aorta, Matrix Biol. 24 (2005) 451–458. [15] A. Alfirevic, S. Insler, Deep hypothermic circulatory arrest in a patient with osteogenesis imperfecta, J. Cardiothorac. Vasc. Anesth. 21 (2007) 245–249. [16] T. Aoki, S. Kuraoka, S. Ohtani, Y. Kuroda, Aortic valve replacement in a woman with osteogenesis imperfecta, Ann. Thorac. Cardiovasc. Surg. 8 (2002) 51–53. [17] S.S. Ashraf, N. Shaukat, M. Masood, T.J. Lyons, D.J. Keenan, Type I aortic dissection in a patient with osteogenesis imperfecta, Eur. J. Cardiothorac. Surg. 7 (1993) 665–666. [18] B. Badmanaban, A. Sachithanandan, S.W. MacGowan, Aortic valve replacement in osteogenesis imperfecta—technical and practical considerations for a successful outcome, J. Card. Surg. 18 (2003) 554–556. [19] J.M. Bennett, J. Gourassas, M.S. Stevens, Double valve replacement in a patient with osteogenesis imperfecta, Eur. J. Cardiothorac. Surg. 1 (1987) 46–48.

[20] R.E. Bonita, I.S. Cohen, B.A. Berko, Valvular heart disease in osteogenesis imperfecta: presentation of a case and review of the literature, Echocardiogr. 27 (2010) 69–73. [21] J.N. Browell Jr., E.H. Drake, Aortic valve lesions associated with osteogenesis imperfecta, Henry Ford Hosp. Med. J. 14 (1966) 245–247. [22] P. Byra, S. Chillag, S. Petit, Osteogenesis imperfecta and aortic dissection, Am. J. Med. Sci. 336 (2008) 70–72. [23] G.S. Chrysant, S.D. Cassivi, C.F. Carey, T.M. Sundt, Double valve replacement in a patient with osteogenesis imperfecta, J. Heart Valve Dis. 11 (2002) 751–754. [24] R.J. Cusimano, Repeat cardiac operation in a patient with osteogenesis imperfecta, Ann. Thorac. Surg. 61 (1996) 1294. [25] M. Desnos, H. Beneton, A. Albou, F. Herreman, J. Fouchard, Valvular insufficiency in Lobstein's disease. A review of the literature apropos of a patient operated on for aortic insufficiency, Arch. Mal. Coeur Vaiss. 73 (1980) 191–197. [26] P.A. Gerlach, J. Rosensweig, K.B. Ramanathan, Successful aortic valve replacement in osteogenesis imperfecta: with special emphasis on peri-operative management, Can. J. Cardiol. 3 (1987) 132–135. [27] S. Hajsadeghi, S.R. Jafarian Kerman, H. Pouraliakbar, R. Mohammadi, A huge coronary artery aneurysm in osteogenesis imperfecta: a case report, Acta Med. Iran. 50 (2012) 785–788. [28] B.A. Heckman, I. Steinberg, Congenital heart disease (mitral regurgitation) in osteogenesis imperfecta, Am J. Roentgenol. Radium Ther. Nucl. Med. 103 (1968) 601–607. [29] R.L. Heppner, H.I. Babitt, J.W. Bianchine, J.R. Warbasse, Aortic regurgitation and aneurysm of sinus of Valsalva associated with osteogenesis imperfecta, Am. J. Cardiol. 31 (1973) 654–657. [30] J. Huang, M. Dinh, N. Kuchle, J. Zhou, Anesthetic management for combined mitral valve replacement and aortic valve repair in a patient with osteogenesis imperfecta, Ann. Card. Anaesth. 14 (2011) 115–118. [31] H. Ichikawa, S. Ishikawa, A. Otaki, T. Takahashi, Y. Sato, T. Koyano, et al., Left ventricular rupture following aortic and mitral valve replacement in a patient with osteogenesis imperfecta: a case report, Kyobu geka Jpn. J. Thorac. Surg. 49 (1996) 294–296. [32] P.A. Isotalo, M.M. Guindi, P. Bedard, M.P. Brais, J.P. Veinot, Aortic dissection: a rare complication of osteogenesis imperfecta, Can. J. Cardiol. 15 (1999) 1139–1142. [33] M.B. Izzat, S. Wan, I.Y. Wan, K.S. Khaw, A.P. Yim, Ministernotomy for aortic valve replacement in a patient with osteogenesis imperfecta, Ann. Thorac. Surg. 67 (1999) 1171–1173. [34] W. Jansen, V. Hombach, A. Osterspey, H. Lubbing, G. Arnold, D.W. Behrenbeck, et al., Valve replacement for mitral in incompetence in a patient with osteogenesis imperfecta (author's transl), Z. Kardiol. 71 (1982) 357–364. [35] R. Jeyamalar, R. Hashim, P. Kannan, Aortic valve replacement in osteogenesis imperfecta tarda—a case report, Singap. Med. J. 30 (1989) 316–317. [36] J. Jornod, M. Rothlin, E. Uehlinger, Cardiovascular anomalies in Lobstein's disease, Schweiz. Med. Wochenschr. 98 (1968) 795–798. [37] M. Kastrup, C. von Heymann, H. Hotz, W.F. Konertz, S. Ziemer, W.J. Kox, et al., Recombinant factor VIIa after aortic valve replacement in a patient with osteogenesis imperfecta, Ann. Thorac. Surg. 74 (2002) 910–912. [38] M. Khashu, G. Pelligra, G.G. Sandor, A.J. Singh, Right-sided cardiac involvement in osteogenesis imperfecta, J. Heart Valve Dis. 15 (2006) 588–590. [39] D. Koentges, F. van de Werf, J. Stalpaert, P. Goddeeris, H. de Geest, Aortic and mitral valve replacement in osteogenesis imperfecta. Report of a case, Acta Cardiol. 41 (1986) 147–153. [40] W. Krahwinkel, R. Kober, D. Horstkotte, F. Borchard, J. Winter, B.E. Strauer, et al., Acute mitral insufficiency in osteogenesis imperfecta, Dtsch. Med. Wochenschr. 119 (1994) 257–260. [41] J.R. Lichtenstein, Osteogenesis imperfecta with floppy mitral valve syndrome, Birth Defects Original Article Series, 111975. 349. [42] A. Lijoi, S. Cisico, E. Caputo, F. Scarano, E. Parodi, G.C. Passerone, Left ventricular rupture after mitral valve replacement in a patient with osteogenesis imperfecta tarda, 26Texas Heart Institute journal/from the Texas Heart Institute of St Luke's Episcopal Hospital, Texas Children's Hospital, 1999. 295–297. [43] Y. Maekawa, T. Hayashi, T. Fujito, K. Hoshi, H. Kamishirado, Y. Iwasaki, et al., Successful surgical treatment of aortic regurgitation due to annuloaortic ectasia and mitral regurgitation caused by tendon rupture in a case of osteogenesis imperfecta, J. Cardiol. 29 (Suppl. 2) (1997) 89–94. [44] P.G. McGlinchey, M.S. Spence, P.P. McKeown, H.C. Mulholland, M.M. Khan, Coexistence of supravalvular aortic stenosis and osteogenesis imperfecta, Ulster Med. J. 70 (2001) 155–157. [45] M.F. McNeeley, B.N. Dontchos, M.A. Laflamme, M. Hubka, C.T. Sadro, Aortic dissection in osteogenesis imperfecta: case report and review of the literature, Emerg. Radiol. 19 (2012) 553–556. [46] R. Melamed, M.M. Aygen, A. Lowenstein, Osteogenesis imperfecta with mitral insufficiency due to ballooning of the mitral valve. A case report, Isr. J. Med. Sci. 12 (1976) 1325–1328. [47] N. Miranda, O.A. Araji, M.A. Gutierrez-Martin, J.C. Tellez, Mitral valve repair in a patient with mitral regurgitation and osteogenesis imperfecta tarda, Interact. Cardiovasc. Thorac. Surg. 13 (2011) 655–656. [48] T. Myrmel, O. Christensen, P. Lunde, Cardiac manifestations in osteogenesis imperfecta. A case report and therapeutic implications, Tidsskr. Nor. Laegeforen. 117 (1997) 519–520. [49] H. Ohteki, S. Ohtsubo, J. Sakurai, N. Koga, K. Kohchi, T. Itoh, Aortic regurgitation and aneurysm of sinus of valsalva associated with osteogenesis imperfecta, Thorac. Cardiovasc. Surg. 39 (1991) 294–295. [50] J.M. Passmore, W.E. Walker, F. Fuentes, Successful aortocoronary bypass in osteogenesis imperfecta, J. Am. Coll. Cardiol. 9 (1987) 960–963. [51] A. Podesta, R. Crivellari, V. Dottori, E. Parodi, G. Passerone, Anaemia, osteogenesis imperfecta and valve diseases. The preoperative treatment with epoetin-alpha to

H. Ashournia et al. / International Journal of Cardiology 196 (2015) 149–157

[52]

[53] [54] [55] [56]

[57]

[58]

[59] [60] [61] [62]

[63] [64] [65]

[66] [67]

[68] [69]

[70]

increase haematocrit and haemoglobin levels in patients with high risk of perioperative bleeding, Minerva Cardioangiol. 48 (2000) 323–327. J. Rochemaure, M. Safar, J.L. Guermonprez, P. Milliez, A. Meyer, A case of cervical aorta in a patient with Lobstein's disease, Coeur et medecine interne 11 (1972) 771–773. M.E. Rogerson, J.D. Buchanan, C.M. Morgans, Left atrial rupture in osteogenesis imperfecta, Br. Heart J. 56 (1986) 187–189. T. Schwarz, M.S. Gotsman, Mitral valve prolapse in osteogenesis imperfecta, Isr. J. Med. Sci. 17 (1981) 1087–1088. J. Sigmund, W. Sperl, F.M. Fink, H. Stoss, Aortic coarctation and osteogenesis imperfecta, Padiatrie und Padologie 21 (1986) 343–349. J.C. Tardif, K. Taylor, N.G. Pandian, S. Schwartz, H. Rastegar, Right ventricular outflow tract and pulmonary artery obstruction by postoperative mediastinal hematoma: delineation by multiplane transesophageal echocardiography, J. Am. Soc. Echocardiogr. 7 (1994) 400–404. S. Uemura, Y. Harada, H. Kasegawa, H. Simura, S. Murakami, N. Morooka, Successful mitral valve replacement in osteogenesis imperfecta—a case report, [Zasshi] [Journal] Nihon Kyobu Geka Gakkai 39 (1991) 2074–2077. H. Wakiyama, Y. Okada, A. Kitamura, Y. Imai, Aortic root replacement with a stentless bioprosthesis in osteogenesis imperfecta, J. Heart Valve Dis. 17 (2008) 197–199. D.D. Waters, D.W. Clark, P.N. Symbas, R.C. Schlant, Aortic and mitral valve replacement in a patient with osteogenesis imperfecta, Chest 72 (1977) 363–364. R.S. Wong, F.M. Follis, B.K. Shively, J.A. Wernly, Osteogenesis imperfecta and cardiovascular diseases, Ann. Thorac. Surg. 60 (1995) 1439–1443. G.H. Almassi, G.R. Hughes, J. Bartlett, Combined valve replacement and coronary bypass grafting in osteogenesis imperfecta, Ann. Thorac. Surg. 60 (1995) 1395–1397. Y. Moriyama, T. Nishida, H. Toyohira, H. Saigenji, S. Shimokawa, A. Taira, et al., Acute aortic dissection in a patient with osteogenesis imperfecta, Ann. Thorac. Surg. 60 (1995) 1397–1399. M. Stefani, Aortic lesions in a case of osteogenesis imperfecta, Boll. Soc. Ital. Cardiol. 22 (1977) 867–869. G.E. Thibault, Clinical problem-solving. The heart of the matter, N. Engl. J. Med. 329 (1993) 1406–1410. R. Zegdi, N. D'Attellis, P. Fornes, J.F. Fuzellier, J.P. Carteaux, J.N. Fabiani, et al., Aortic valve surgery in osteogenesis imperfecta: report of two cases and review of the literature, J. Heart Valve Dis. 7 (1998) 510–514. S.J. Wood, J. Thomas, M.V. Braimbridge, Mitral valve disease and open heart surgery in osteogenesis imperfecta tarda, Br. Heart J. 35 (1973) 103–106. B. Weisinger, E. Glassman, F.C. Spencer, A. Berger, Successful aortic valve replacement for aortic regurgitation associated with osteogenesis imperfecta, Br. Heart J. 37 (1975) 475–477. D. Stein, F.E. Kloster, Valvular heart disease in osteogenesis imperfecta, Am. Heart J. 94 (1977) 637–641. S. Ohuchi, J. Koizumi, H. Kin, S. Ohsawa, H. Izumoto, K. Ishihara, et al., Valvular heart surgery in osteogenesis imperfecta, Kyobu geka Jpn. J. Thorac. Surg. 55 (2002) 1011–1013. M.Q. Najib, H.V. Schaff, J. Ganji, H.R. Lee, R.L. Click, D.C. Miller, et al., Valvular heart disease in patients with osteogenesis imperfecta, J. Card. Surg. 28 (2013) 139–143.

157

[71] I.M. Cohen, W.V. Vieweg, J.S. Alpert, J.A. Kaufman, A.D. Hagan, Osteogenesis imperfecta tarda. Cardiovascular pathology, West. J. Med. 126 (1977) 228–231. [72] C. Pijoan de Beristain, Association of aortic insufficiency with osteogenesis imperfecta in 2 members of the same family, Rev. Esp. Cardiol. 26 (1973) 405–410. [73] N.J. White, C.G. Winearls, R. Smith, Cardiovascular abnormalities in osteogenesis imperfecta, Am. Heart J. 106 (1983) 1416–1420. [74] U. Vetter, B. Maierhofer, M. Muller, D. Lang, W.M. Teller, R. Brenner, et al., Osteogenesis imperfecta in childhood: cardiac and renal manifestations, Eur. J. Pediatr. 149 (1989) 184–187. [75] J. Hortop, P. Tsipouras, J.A. Hanley, B.J. Maron, J.R. Shapiro, Cardiovascular involvement in osteogenesis imperfecta, Circulation 73 (1986) 54–61. [76] Z. Radunovic, L.L. Wekre, K. Steine, Right ventricular and pulmonary arterial dimensions in adults with osteogenesis imperfecta, Am. J. Cardiol. 109 (2012) 1807–1813. [77] Z. Radunovic, L.L. Wekre, L.M. Diep, K. Steine, Cardiovascular abnormalities in adults with osteogenesis imperfecta, Am. Heart J. 161 (2011) 523–529. [78] S. Migliaccio, G. Barbaro, R. Fornari, G. Di Lorenzo, M. Celli, C. Lubrano, et al., Impairment of diastolic function in adult patients affected by osteogenesis imperfecta clinically asymptomatic for cardiac disease: casuality or causality? Int. J. Cardiol. 131 (2009) 200–203. [79] S. Kalath, P. Tsipouras, F.H. Silver, Increased aortic root stiffness associated with osteogenesis imperfecta, Ann. Biomed. Eng. 15 (1987) 91–99. [80] V. Bonilla Jimenez, J. Saavedra Falero, M.T. Alberca Vela, P. Diaz Guardiola, P. Iglesias Bolano, I. Pavon, Cardiac abnormalities in osteogenesis imperfecta. Case–control echocardiographic study, Med. Clin. 135 (2010) 681–684. [81] L. Lange, V. Fabecic-Sabadi, G. Bein, Comparative review of normal echocardiographic values from the premature infant to the adolescent, Herz 8 (1983) 105–121. [82] R. Lozano, M. Naghavi, K. Foreman, S. Lim, K. Shibuya, V. Aboyans, et al., Global and regional mortality from 235 causes of death for 20 age groups in 1990 and 2010: a systematic analysis for the Global Burden of Disease Study 2010, Lancet 380 (2012) 2095–2128. [83] K. Dwan, D.G. Altman, J.A. Arnaiz, J. Bloom, A.W. Chan, E. Cronin, et al., Systematic review of the empirical evidence of study publication bias and outcome reporting bias, PLoS ONE 3 (2008) e3081. [84] D.M. Brinkley, E.V. Gelfand, Valvular heart disease: classic teaching and emerging paradigms, Am. J. Med. 126 (2013) 1035–1042. [85] A. Liberati, D.G. Altman, J. Tetzlaff, C. Mulrow, P.C. Gotzsche, J.P. Ioannidis, et al., The PRISMA statement for reporting systematic reviews and meta-analyses of studies that evaluate health care interventions: explanation and elaboration, J. Clin. Epidemiol. 62 (2009) e1–e34. [86] A.B. Hill, The environment and disease: association or causation? Proc. R. Soc. Med. 58 (1965) 295–300. [87] L.L. Wekre, K.F. Froslie, L. Haugen, J.A. Falch, A population-based study of demographical variables and ability to perform activities of daily living in adults with osteogenesis imperfecta, Disabil. Rehabil. 32 (2010) 579–587. [88] S.G. Wannamethee, A.G. Shaper, Physical activity in the prevention of cardiovascular disease: an epidemiological perspective, Sports Med. 31 (2001) 101–114.