Is DISI Deformity Related to Presence of a Medial Lunate Facet in Patients with Scapholunate Dissociation?

Abstract

Background  Lunate morphology has been suggested to influence carpal kinematics.

Purpose  We investigate a possible relation between presence of a medial lunate facet and dorsal intercalated segment instability (DISI) of the wrist in patients with a scapholunate (SL) dissociation.

Methods  We retrospectively reviewed patients diagnosed with SL dissociation between 2000 and 2017. Lunate morphology was categorized based on radiographs and magnetic resonance imaging (MRI), as type I or II according to Viegas and Galley. DISI was defined as radiolunate angle > 15 degrees and SL instability as SL angle > 60 degrees. SL distance > 3 mm was considered as widening and carpal height ratio < 0.5 was considered as carpal collapse. We used descriptive statistics to report on SL instability and DISI in patients with Viegas type I and type II lunates. We calculated kappa to determine agreement between radiographs and MRI and to determine inter- and intraobserver agreement.

Results  Of 119 patient files, 79 wrists met the inclusion criteria of which 25 were type I lunates and 54 type II. Similar spreading of the data of both groups was found regarding DISI, SL instability, and SL widening based on radiographic classification of the lunate, even after adding MRI findings. In the presence of carpal collapse, capitate-to-triquetrum distance was higher. We found a substantial inter- and intraobserver agreement for lunate classification.

Conclusion  Our results suggest a similar prevalence of DISI deformity or enlarged SL angle in patients with type I or II lunate in presence of SL dissociation. The Viegas classification is a reliable and reproducible classification system.

Level of evidence  This is a Level III, cross-sectional study design.

Lunate morphology has been suggested to alter carpal kinematics and, therefore, to have an impact on progression of dissociative carpal instability. ^{1 2} Viegas described two major types of lunate, based on the anatomy of the midcarpal articulation. ³ In type I lunates, there is no articulation between the lunate and the hamate, and in type II lunates, a medial facet articulation with the hamate is present. Galley et al found a medial lunate facet will increase capitate-to-triquetrum (CT) distance and proposed a lunate classification based on CT distance, to identify the presence of a medial facet. ⁴ The luno-hamate articulation is hypothesized to increase stability against the extension deformity force of the lunate in scapholunate (SL) instability. ²

It has been suggested that the incidence of dorsal intercalated segment instability (DISI) deformity associated with SL ligament lesions, is lower in patients with lunate type II, but observational radiologic studies so not reflect a consensus on this subject so far. ^{1 2}

In this retrospective cross-sectional study, the purpose was to describe the relation between lunate morphology (presence of absence of a medial lunate facet) and SL instability in terms of DISI deformity. We hypothesized a lunate type II can be protective against DISI deformity forces in SL dissociation. Second, we evaluated the interobserver reliability and intraobserver reproducibility of the Viegas classification of the lunate ^{3 5} and calculated correlation between CT distance and SL angle (SLA) or radiolunate angle (RLA).

Materials and Methods

After approval of the Ethical Board of the University Hospital, we retrospectively reviewed all patient files with the diagnosis of SL dissociation, who underwent a modified Brunelli tenodesis (MBT) and SL advanced collapse (SLAC) wrist who underwent a proximal row carpectomy of partial wrist fusion (two- or four-bone fusion) from 2000 to 2017 and found 119 files with these search terms. In this patient cohort, 29 patients were planned for MBT, 26 for proximal row carpectomy, and 64 for partial wrist fusion. Exclusion criteria were a history of previous wrist or carpal fractures (11 patients), inflammatory arthritis (3 patients), previous wrist surgery (7 patients), and congenital hand anomalies (0 patients). Surgery records confirm the presence of a SL ligament lesion in all cases. An absence of preoperative radiographs or inadequate imaging (no proper alignment on radiographs: radius-to-long finger metacarpal angle > 20 degrees in lateral view) was the reason for exclusion of an additional 19 patients, giving us 79 files to include for further analysis.

Radiographs of the patients included in this study were reviewed independently by two of the authors, level of experience 1 and 3 as described by Tang and Giddins ⁶ for lunate morphology, DISI deformity, and SL dissociation. Lunate morphology was determined on posteroanterior radiographs using two different methods. The first method was based on the absence (lunate type I) or presence (lunate type II) of a medial lunate facet ⁵ as illustrated in Fig. 1 . Second, lunate morphology was determined as described by Galley et al, ^{4 7} with a type I lunate in patients with a CT distance of 0 to 2 mm, intermediate lunate type with a CT distance of 2.1 to 3.9 mm, and a type II lunate with a CT distance of 4 mm or more. DISI deformity was determined using RLA on true lateral radiographs. An RLA of +15 degrees or more was considered a DISI deformity. The SLA on true lateral radiographs was used to evaluate SL instability. An SLA of 60 degrees or more was defined as a SL instability. All carpal angles were measured using the tangential method. ^{8 9} Carpal height ratio (CHR) was determined as the ratio of the carpal height to the length of the third metacarpal. ¹⁰

Fig. 1.

Illustration of a lunate type I ( A ) and type II ( B ) according to Viegas. ⁴

We used descriptive statistics to report on RLA, SLA, and SL distance (SLD) in the different types of lunate morphology. We calculated a kappa-coefficient to evaluate the inter- and intraobserver variability of the radiological measurements. The kappa-coefficient is interpreted as slight (0–0.20), fair (0.21–0.40), moderate (0.41–0.60), substantial (0.61–0.80), and almost perfect (0.81–0.99) agreement. To describe the relation between lunate type or CT distance and SLD and carpal height, we calculated a Pearson's correlation coefficient.

Results

Patient Demographics

Evaluation of 119 patients records from year 2000 to 2017 yielded 79 wrists meeting our inclusion criteria. Mean age was 48.4 years (range 22–75) and there were 59 men and 20 women. Differences in lunate morphology between both sexes are illustrated in Table 1 .

Table 1. Patient demographics.

	Lunate I	Lunate II
Age	45.8	48.7
Sex
Male	22	37
Female	3	17

Lunate Morphology

When lunates were classified using the Viegas method ^{4 11} based on the presence or absence of the medial facet, 25 lunates type I and 54 lunates type II were found, compared with 5 lunates type I, 38 intermediate, and 36 lunates type II using the Galley et al method. ⁴ We found a substantial interobserver agreement with kappa 0.79 for the Viegas method and kappa 0.82 for the Galley et al method. Intraobserver reproducibility was nearly perfect, with kappa 0.87 for the Viegas and 0.88 for the Galley et al method. In 21 patients, magnetic resonance imaging (MRI) was available. Comparing lunate morphology using the Viegas medial facet method of plain radiography and MRI, reveals only two cases where medial facet of the lunate was missed on plain radiograph when compared with MRI. This is illustrated in Table 2 .

Table 2. Lunate classification based on presence or absence of a medial facet.

	Lunate I	Lunate II
Rx	5	16
MRI	3	18

Abbreviations: MRI, magnetic resonance imaging; Rx, radiography.

Note: Comparison between analysis on plain radiographs and MRI. MRI was available in 21 cases, medial lunate facet was missed on plain radiograph in only 2 cases.

Carpal Instability

Because majority of lunates was classified as intermediate using the Galley et al method (38 out of 79), we used the Viegas medial facet method to investigate the relationship between DISI or SL instability and lunate type. Of the 79 patients included, 73.4% ( n  = 58) had a DISI deformity, 18 (31%) of them had a type I lunate and 40 a type II lunate. In the group without DISI deformity, 7 patients (33%) had a type I lunate and 14 had a type II lunate. We noticed a static SL dissociation in 91.1% ( n  = 72), 22 (30%) of them had a lunate type I and 50 had a lunate type II. The other 7 patients had an SLA within normal values, 3 (43%) of them had a lunate type I. One patient had a SLAC wrist, with marginal values for SLA (58 degrees) and RLA and the other six had a dynamic SL instability diagnosed on stress radiographs. This is illustrated in Table 3 . We analyzed the relation between lunate type and SLD or CHR, as illustrated in Table 4 . In the analysis of the relation between CHR and lunate type, only 54 wrists were included. Twenty-five had to be excluded because the third metacarpal was not completely visualized on the radiographs. A box and whisker plot was composed ( Fig. 2 ) to illustrate similar values for spread of SLA, RLA, SLD, and CHR when comparing patients with lunate type I to patients with lunate type II. Table 5 further completes the comparison of these groups.

Table 3. Relation between lunate morphology and carpal stability.

	Lunate I	Lunate II
RLA < 15°	7 (7/25 = 28%)	14 (14/54 = 26%)
RLA > 15°	18 (18/25 = 72%)	40 (40/54 = 74%)
SLA < 60°	3 (3/25 = 12%)	4 (4/54 = 7%)
SLA > 60°	22 (22/25 = 88%)	50 (50/54 = 93%)

Abbreviations: RLA, radiolunate angle; SLA, scapholunate angle.

Table 4. Relation between lunate morphology and scapholunate distance (SLD) and carpal height ratio (CHR).

	Lunate I	Lunate II
SLD < 3 mm	9 (9/25 = 36%)	14 (14/54 = 26%)
SLD > 3 mm	16 (16/25 = 64%)	40 (40/54 = 74%)
CHR < 0.5	8 (8/19 = 42%)	24 (24/35 = 69%)
CHR > 0.5	11 (11/19 = 58%)	11 (11/35 = 31%)

Fig. 2.

Box and whisker plot of data spreading. A similar spreading when comparing patients with lunate Viegas type I to patients with lunate Viegas type II is illustrated.

Table 5. Illustration of SLA, RLA, SLD, and CHR when comparing Viegas lunate type I to Viegas lunate type II.

	SLA		RLA		SLD		CHR
Lunate type	I	II	I	II	I	II	I	II
Mean	80.84	77.56	21.44	19.21	3.68	4.05	0.50	0.50
Stand deviation	13.68	13.62	11.52	9.31	1.43	1.69	0.05	0.10
CI	75–86	74–81	17–26	17–22	3.1–4.2	3.6–4.5	0.48–0.53	0.47–0.53

Abbreviations: CHR, carpal height ratio; CI, confidence interval; RLA, radiolunate angle; SLA, scapholunate angle; SLD, scapholunate distance.

Using the Galley et al method, excluding the intermediate group, we found no correlation between CT distance and DISI deformity or SL instability (Pearson's correlation 0.04 for RLA and 0.14 for SLA). We found no correlation between CT distance and SLD or CHR (Pearson's correlation 0.13 for SLD and 0.1 for CHR), but when comparing the group with carpal collapse (CHR < 0.5) with the group without carpal collapse (CHR 0.5 and more), we found a higher CT distance in the collapse group ( p  < 0.05).

Discussion

The possible relationship between lunate morphology and carpal stability remains controversial in the literature. We found similar data spreading in both groups of lunate type when comparing SL instability or DISI deformity in patients with SL dissociation and/or SLAC. When comparing our patient group to previously published results, ^{1 10} we found similar patient demographics as in the cohort of Pang et al. ¹ They suggested a difference in age or sex distribution could influence carpal stability due to greater ligamentous laxity of the intrinsic and extrinsic ligaments. However, in our patient group, we did not find any correlation between age or sex and SLA or RLA.

The reported incidence of type II lunate varies from 27 to 73%. ^{11 12 13} This wide variation might be due to differences in research method. Sagerman et al ¹⁴ found radiographs to be 64 to 72% accurate in identifying type II lunate compared with direct visualization in a cadaver-based study.

Dyankova and Marinov ¹⁵ suggest type II lunates are very difficult or impossible to identify on plain radiographs when the medial facet of the lunate is less than 3 mm. Direct visualization of the medial lunate facet gives the most accurate classification of lunate morphology, but because of the retrospective character of this study, we were only able to classify the lunate morphology by radiological imaging (radiographs and/or MRI). The accuracy of diagnosing a medial lunate facet of imaging compared with direct visualization was not a subject of this study.

McLean et al ¹⁶ proposed MRI scan as the most accurate method to identify even a small cartilaginous medial facet in a noninvasive way. In our patient cohort, MRI was available in 22 patients. Of the 18 lunates defined type II on MRI, 16 lunates type II were correctly defined on plain radiographs. This represents a very good agreement of 88%, consistent with previous studies on the accuracy of radiographs for determination of lunate type. ¹⁶ We confirmed interobserver reliability and intraobserver reproducibility of lunate classification using plain radiographs with a kappa coefficient of 0.79 and 0.87, respectively, representing a substantial agreement.

Galley et al ⁴ described a method to classify lunate morphology based on CT distance. Shin and colleagues, ² however, found that a wide SL diastasis with proximal migration of the capitate into the widened SL interval can lead to an increased CT distance. This way a false type II lunate classification could arise. A significant difference was found when comparing CT distance of those wrists with carpal collapse (CHR < 0.5) to those with normal carpal height, confirming this hypothesis. This has to be interpreted with some care, taking our small sample size into account. We found no correlation between the SLA or RLA and CT distance (Pearson's correlation 0.04 for RLA and 0.14 for SLA).

This study investigates the relation between lunate morphology and carpal collapse and SL instability in patients with SL dissociation and/or SLAC in a large patient cohort. Radiological evaluation by two reviewers allows us to conclude on interobserver reliability of lunate classification. MRI was available in some cases, allowing us to evaluate the agreement between MRI and plain radiographs.

We acknowledge the limitations to our study. Most importantly it is a retrospective study on a sample of convenience. The retrospective design of the study implies a selection bias and lack of power analysis that could be avoided in a prospective study.

It is based on static imaging allowing us solely to conclude on static dissociations and we are maybe missing out on a group of dynamic instabilities. The time between the SL injury and radiographs is not specified. This could introduce a selection bias as these may be progressive lesions over time. MRI was not available in all patients and no additional direct visualization of the lunate was possible in this retrospective study, possibly misclassifying some of the lunates. No comparison with contralateral wrist was made, possibly misinterpreting some of the carpal relations.

In conclusion, results of this retrospective review suggest that an articulation between the lunate and hamate, as seen in patients with lunate type II morphology, does not have a large influence on carpal stability in patients with SL ligament lesions. This is not consistent to the findings of the previous report of Shin and colleagues. ² We found the Viegas classification a reliable and reproducible classification system and discourage the use of the Galley et al classification because of the relation between CT distance and carpal height.

Wrist instabilities are complex kinematic problems, not to be explained solely by lunate morphology. Further research is necessary to understand the role of the medial facet of the lunate in carpal kinematics.

Funding Statement

Funding None.