Achondroplasia: Really rhizomelic?

INTRODUCTION

Achondroplasia is the most common form of short limb dwarfism in humans with an incidence of between one in 10,000–300,000 live births [Horton et al., 2007]. In over 80%, it is caused by a spontaneous point mutation in the fibroblast growth factor receptor 3 (FGFR 3) gene [Yamanaka et al., 2003; Horton et al., 2007] leading to abnormal regulation of chondrogenesis at the growth plate, and reduction in longitudinal bone length [Wright and Irving, 2012; Qi et al., 2014].

The shortening of the limb lengths in achondroplasia is widely described as “rhizomelic” [National Organization for Rare Disorders, 2002; Castriota-Scanderbeg and Dallapiccola, 2005; Grunebaum, 2005; Lovell et al., 2005; Richette et al., 2008; Laederich and Horton, 2010]. This term implies that proximal segments of the limbs are disproportionally shorter than distal segments [Vanhoenacker et al., 2001]. While this appearance may be convincing clinically, possibly due to differences in positioning of skin creases on the patient, the description is not necessarily true, or helpful radiologically. The basis for the terminology is likely to have stemmed from descriptions within numerous case reports published during the 1900's [Milne, 1910; Pritchard, 1911; Pritchard, 1928; Morgan, 1930] whereby, radiological assessment of skeletal dysplasias was not routinely employed, and was still in its infancy.

Nevertheless, publications on the topic have been conflicting. In 1967, Langer et al. [1967] published qualitative and quantitative findings of radiographic manifestations in achondroplasia. The authors reported a subjective rhizomelic pattern of shortening in the upper limbs, but absent or only very mild rhizomelic shortening in the lower limbs when compared to published normal standards [Maresh, 1955].

Further research conducted in 1975 by Nehme et al. [1976], concluded that lower limb rhizomelic shortening was a feature of achondroplasia; as expected average femoral lengths were a greater number of standard deviations shorter than tibial lengths, when compared with published normal standards [Anderson and Green, 1948].

To our knowledge, there are no published studies that formally assess both the radiographic measurements of upper and lower limb proportions in achondroplasia with comparison to a locally derived normal (or average statured) population. We hope that by clarifying the type of shortening present, we can improve the accuracy of the radiographic descriptions of the condition.

The objective of this study, was therefore, to determine whether rhizomelic shortening is a genuine radiographic feature of infants with achondroplasia at the point of diagnosis.

MATERIALS AND METHODS

We retrospectively searched our clinical radiological database over a 10 year period for patient reports containing the words “achondroplasia,” “achondroplastic,” and “achondroplast.”

RESULTS

We identified 22 infants with achondroplasia who had undergone limb radiographs with typical clinical and radiological features of the dysplasia to form our study population. One patient in the achondroplasia group did not have imaging of the whole femur on skeletal survey, and therefore, only an upper limb ratio could be measured for this single case. Hundred and fifty average statured controls were identified over a 3-year period. Full patient demographics are included in Table I.

Within the control group, there was no significant difference between left and right limb length ratios for upper (P = 0.1958) or lower limbs (P = 0.1651). The average difference in limb length measurements for the upper limb was −0.1 ± 0.9 mm (range: −1.4 to 2.5 mm), and for the lower limb was −0.1 ± 1.7 mm (range: −9.1 to 5.6 mm). We therefore, used the combined mean limb length for the left and right for all subsequent assessments.

Lower Limb Length Ratios

No statistical significance was found in mean lower limb length ratios (tibia/femoral length) in the control group versus the achondroplasia group (0.83 ± 0.02 vs. 0.84 ± 0.04, P = 0.4663).

The average Z-score for the lower limb ratios in the achondroplasia group was 0.19 ± 1.85 (range: −3.09 to 4.06; 4 (19%) patients with Z scores ≥2 and ≤−2) versus those for the control group of 0 ± 1.00 (range: −3.93 to 3.22; 5 (3%) patients with Z scores ≥2 and ≤−2). The upper and lower limb ratios with accompanying Z scores are given in Table II.

Table II. Upper and Lower Limb Ratios With Accompanying Z Scores for Control Versus Achondroplasia Patient Groups

Group	Limb ratio, mean value ± sd (range)	Z scores, mean value ± sd (range)
Control group–Upper limb*	0.79 ± 0.02 (0.72–0.84)	0 ± 1.00 (−2.66–2.22)
Achondroplasia–Upper limb	0.87 ± 0.04 (0.79–0.93)	3.42 ± 1.44 (0.19–5.88)
Control group–Lower limb*	0.83 ± 0.02 (0.74–0.91)	0 ± 1.00 (−3.93–3.22)
Achondroplasia–Lower limb	0.84 ± 0.04 (0.76–0.93)	0.15 ± 2.81 (−3.09–4.06)

• * In the control study group, both right and left upper and lower limb measurements were obtained for allpatients. A value for upper and lower limb ratios was then calculated from an average of the bilateral measurements.

Figure 1 demonstrates the upper and lower limb ratios for the control and achondroplasia populations with ± 2sd limits.

DISCUSSION

Patients with achondroplasia demonstrated mesomelic shortening of their limbs with the addition of statistically significant “rhizomelic” shortening within their upper limbs, but not their lower limbs, compared with an average statured control group. The results reveal a potential inaccuracy in the medical literature regarding radiographic descriptions of limb proportions in achondroplasia.

There are two previous studies, published regarding differences in radiographic limb proportions between an achondroplastic and an average statured population [Langer et al., 1967; Nehme et al., 1976]. In one study [Langer et al., 1967], lower limb proportions in achondroplasts of various ages (34 subjects aged 0–14 years old, of which 6 were aged under 1 years of age) were compared with average statured published standards, not a locally derived control population. Although they reported rhizomelic shortening of the lower limbs, we were concerned that published standards may not be the most accurate comparison between different populations. In addition, the authors commented upon subjective marked rhizomelic shortening in the upper limbs, however, numerical data for upper limb ratios were not calculated.

The second study [Nehme et al., 1976] analyzed differences in limb length measurements rather than ratios between an achondroplasia population, and published average statured standards. The authors of this paper concluded the presence of lower limb rhizomelic shortening, due to the fact that femoral lengths were a greater number of standard deviations (average sd = 8; sd range = −10.1 to −6.4) below the published standards than tibial lengths (average sd = 7; sd range = −9.5 to −4.1) in 18 patients aged between 0–18 years old. A test of statistical significance was not performed, and upper limb ratios were not calculated as radiographs were not obtained of this region.

Our study included both upper and lower limb ratios within a locally derived control group, and an achondroplastic population. All measurements for our achondroplasia patients were calculated by two observers blinded to each other's findings, whereas, previous studies have only employed one observer. We therefore, consider our method to be a more accurate way of assessing for differences between the populations.

A further strength to our study is that it includes information regarding patient genetic analysis. Interestingly within our results, we identified two patients with rare FGFR3 mutations who were not outliers within the study population. This suggests that our findings regarding limb length proportions within our achondroplasia population need not necessarily be confined to those with “typical” FGFR3 mutations. Alternatively, within the study group, there were two patients with upper limb length ratios lying further than two standard deviations from the mean who had not undergone genetic analysis for their diagnosis. Although one may argue that the diagnosis of achondroplasia may be disputed in these cases without genetic analysis, in the presence of typical clinical features, and supportive radiographic findings, genetic testing is not always undertaken to confirm a clinical diagnosis in the United Kingdom. It is possible that in certain cases, there may not be rhizomelic shortening of the upper limbs in infants at all.

Our findings help to correct a widely published medical myth relating to the radiographic description of achondroplasia. This study may be of use to radiologists, less experienced in assessing skeletal dysplasias, who could potentially misinterpret “rhizomelic limb shortening” as a pre-requisite in diagnosing achondroplasia. Conversely, the presence of true, severe rhizomelic shortening may suggest diagnoses other than achondroplasia, which may be overlooked if there is too much emphasis placed upon rhizomelic shortening in achondroplasia. We have not demonstrated whether this information can be used in altering patient management and may not assist experienced clinicians, and radiologists in diagnosis.

There were several potential limitations to our study. The relatively small sample size of patients with achondroplasia led to wide confidence intervals within which limb length ratios may lie. Nevertheless, the significant differences between mean values of upper limb ratios do still lend validity to our conclusions. In addition, we analyzed our study group at one point in time, namely at diagnosis. Our study did not investigate whether the limb length proportions vary over time or with increasing age.

The use of post-mortem skeletal imaging as an average statured (“normal”) control population could be also argued as not reflective of “live” age-matched control subjects. We attempted to counter this by correlating all control subjects with their autopsy results to confirm none of the causes of death were associated with altered bone growth. Each post-mortem skeletal survey was also double reported by two consultant paediatric radiologists to exclude underlying radiographic bone abnormalities. Given that very few healthy live children have complete imaging of both upper and lower limbs, and the ethical implications in subjecting normal patients to increased radiation burden for research purposes, analysis of post-mortem imaging was felt to be the most realistic method by which to derive a control population. The use of published standards may have been an alternative measure (as adopted by previous studies), however, published “normal” limb ratios [Robinow and Chumlea, 1982] may not be applicable to the 21st century paediatric population.

We also confirmed cases of achondroplasia by radiographic rather than genetic testing. In practice, although genetic testing would usually be offered it is not considered essential in reaching a diagnosis owing to the constellation of multiple characteristic features of the condition, and therefore, our case inclusion reflects those of normal practice.

Furthermore, differences in skeletal survey imaging for dysplasia diagnoses nationwide [Offiah and Hall, 2003] (as oppose to those for post-mortem skeletal surveys) meant that not all imaging for our achondroplasia population was standardized (for example, the laterality of upper and lower limbs, use of radiographic equipment, and number of images differed between subjects). This factor resulted in one patient without full imaging of either femur, meaning a lower limb ratio for this case could not be derived. For this reason, two observers were employed in measuring limb lengths in the achondroplasia group to prove that despite such variation, measurement results were consistent.

Future work on the subject of achondroplasia and limb proportions may include analysis of a larger population sample, and exploration for reasons behind the differences in upper and lower limb proportions. It is already speculated that the effects of FGFR3 mutations are responsible for relative overgrowth of the fibula [Matsui et al., 2001; Stanley et al., 2002; Lee et al., 2007] and this may hold true for the differences in limb proportions also.

To conclude, patients with achondroplasia demonstrate mesomelic limb shortening, with the addition of rhizomelic shortening of the upper, but not the lower limbs. The results reveal a potential inaccuracy in the medical literature regarding radiographic descriptions of limb proportions in achondroplasia.