Long-term outcomes after capitate fractures: a median 16-year follow-up

Introduction The long-term effects of a capitate fracture are unknown. The aim of this study was to assess both clinical and radiological long-term outcomes after a capitate fracture. Materials and methods From a cohort of 526 consecutive patients with post traumatic radial sided wrist pain, 23 were identified diagnosed with a capitate fracture. 16 of these (11 males and 5 females) with a median age at injury of 17.5 years (range 11–27 years) were eligible for a follow-up after a median of 16 years (range 8 to 17 years). In this study patients were examined using conventional radiographs, computed tomography (CT) and magnetic resonance imaging (MRI) at the time of injury and with CT at the follow-up. At follow-up radiological signs of osteoarthritis were graded in four stages and clinical outcome was evaluated by measuring range of wrist motion and grip and pinch strength. The subjective outcome was assessed using DASH and PRWE questionnaires. Results Five patients had isolated capitate fractures and 11 had concomitant fractures, 10 of which had a simultaneous scaphoid fracture. 14 patients had been treated non-surgically in a cast and two patients were treated surgically. None of the fractures were visible on conventional radiographs at the time of injury. At follow-up all fractures had healed without signs of avascular necrosis. In one patient, CT examination revealed osteoarthritis between the capitate and lunate. This did not cause clinical symptoms. Functional impairments and pain scores were low: median DASH score 0, median PRWE 3 and median VAS pain score 0. We found no impairment in range of motion or grip and pinch strength. Conclusions At a median of 16-year follow-up, patients with a capitate fracture report normal self-assessed hand function as well as good wrist motion and strength. The risk of development of posttraumatic arthritis in the joints around the capitate is low.


Introduction
The capitate is centered within the carpus and thus well protected from injury.Capitate fractures are rare, constituting approximately 8% of all carpal fractures [1].The majority of these fractures are reported to occur in combination with other carpal fractures or with associated ligament injuries [2][3][4][5][6].Capitate fractures can be difficult to diagnose with conventional radiographs because they are often non-or minimally displaced due to stabilization by intercarpal ligaments.Therefore, more advanced imaging, such as MRI or CT, are important [2,7,8].The majority of epidemiological studies [9,10] have used medical charts or conventional radiographs to describe carpal fracture prevalence, which therefore is likely to be underrated.
Four major fracture patterns have been described for the capitate.The most common type is a transverse fracture of the capitate body which is often seen in combination with a perilunate injury with or without a simultaneous scaphoid fracture [6,11].Other fracture patterns are: a transverse fracture of the proximal pole (waist), a coronal oblique fracture and finally a parasagittal fracture.Studies have suggested that the capitate is vulnerable to post-traumatic avascular necrosis because of its blood supply, as the vessels enter dorsally and supply the bone in retrograde fashion [12].However, more recent studies, based on micro-CT angiography examination of the vascularity in cadaveric capitates, show that 70% of the capitates also have a vessel entering the proximal pole.This fact can explain why avascular necrosis (AVN) of the capitate is rare [13].Nonetheless, undiagnosed and untreated, capitate fractures can lead to avascular necrosis [14,15] and nonunion [3,15,16].
There are no prior studies where patients with capitate fractures have been randomized to conservative or surgical treatment.However, there is a general understanding that undislocated isolated fractures can be treated for at least 4 weeks in a cast, whereas an isolated dislocated fractures or capitate fractures combined with other carpal fractures benefit from surgical stabilization [2,17,18].
The most common complication after a capitate fracture is development of nonunion, which often is related to delay in diagnosis [16,[19][20][21].In addition, the capitate is multiarticular except for its dorsal and palmar surfaces and a capitate fracture carries a risk for intraarticular cartilage damage and posttraumatic osteoarthritis [14,22].Knowledge concerning long-term clinical and radiological outcomes remains limited [3].
The aim of this study was to assess the long-term clinical outcomes following capitate fractures, and based on CT examinations, to evaluate malunion and development of osteoarthritis in the joints around the capitate.

Patients
From 2004 to 2008, all patients presenting at the emergency ward at our hospital, with posttraumatic radial sided wrist pain following an injury, were asked to participate in the study [1].The inclusion criterion was posttraumatic pain on the radial side of the wrist, located distal to the radiocarpal joint and proximal to the carpometacarpal.Exclusion criteria were: Injuries being a part of a radiocarpal or intercarpal fracture dislocation or intrinsic ligament disruption and injuries older than 14 days.All patients were examined according to a study specific protocol including conventional radiographs of the wrist.Regardless of the result from the radiographs all patients in the study underwent an MRI of the wrist within three working days after inclusion in the study.In addition, if the MRI revealed a scaphoid or a capitate fracture a computed tomography (CT) was performed.A total of 526 patients were examined [23].From this cohort, 23 patients were identified having a capitate fracture.At the present long-term follow up two patients had died and one patient had moved abroad.Thus, 20 patients where accessible for clinical and radiological follow-up.

Imaging
At inclusion, conventional radiographs of the wrist were taken in dorsovolar and lateral projections with an additional four views of the scaphoid.MRI was performed with a small joint coil using a 0.23 low Tesla field MRI unit (Proview Marconi Medical Systems, Vantaa Finland).The study protocol used coronal short tau inversion recovery (STIR) 3-mm slice thickness; coronal T1 field echo 3-dimensional (FE3D) 2 mm-slice thickness; axial T1 fast spin echo 3.5 mm slice thickness; sagittal T1 FE3D 2-mm slice thickness.A fracture was defined as a cortical and trabecular linear lesion, causing intramedullary hyperintensity on the STIR as well as intramedullary hypointensity on T1 weighted images, extending to the cortices.CT was performed using a 16-slice scanner (Somatotom Sensation 16; Siemens AG, Forchheim, Germany).The patient was positioned prone with the hand extended above the head (superman position).Axial sections of 0.6 mm thick slices were obtained from which 1-2 mm thick reconstructions in the coronal and sagittal planes were done.
At the present long-term follow-up, CT was performed with a 320-slice CT scanner (Aquilon ONE Genesis: Canon Medical Systems, Otawara, Japan) Patients were positioned prone in the superman position.Axial 0.5 mm thick sections were obtained.2 mm thick reconstructions in the coronal and sagittal planes of the capitate were attained.

Clinical evaluation and patient rated outcome measures (PROMs)
The clinical examination was done independently by specialist in hand surgery, author DO, who had not participated in the initial treatment of the patients.Also, the examination was done without knowledge of results from the present CT scan.
Patient rated outcome (PRO) were assessed with the Disability of Arm Shoulder and Hand (DASH) questionnaire [24] and the Patient Reported Wrist Evaluation (PRWE) form [25]. Pain was assessed using VAS score from the PRWE form.
Range of movement was measured using a goniometer.Grip strength was measured using an Exacta Hydraulic Hand Dynamometer (North Coast Medical, Gilroy, CA, USA).Pinch strength was measured with B&L pinch gauge (B & L Engineering, Tustin, CA, USA).The mean value of three measures was used for calculation.
Intracarpal instability was assessed with the Lichtman test [26], Watson shift test [27], and a ballottement test between capitate and lunate.All clinical evaluations were done on both hands.

Radiological assessment
We used the classification system proposed by Kadar and coworkers where capitate fracture types are described depending on localization and pattern.Three main categories are described: capitate body fractures (further subdivided in stellate comminuted, oblique high and low, transverse high and low) avulsion tip fractures, and shear depression fractures [6].
From CT at follow-up, we evaluated fracture healing and malunion.The malunion was assessed in the capitate and in the scaphoid, the height-length ratio, lateral intrascaphoid angle, and dorsal cortical angle, were measured [28].Signs of osteoarthritis were assessed in six defined articulations: capitate-scaphoid, scaphoid and radius, capitate and lunate, capitate and 3rd metacarpal, capitate and trapezoid, capitate and hamate (Fig. 1).Osteoarthritis was classified according to Whites classification as described by Clementson [29] (Table 1).Furthermore, the scapho-lunate (SL) interval and the SL angle were measured.All CT examinations were reviewed independently by two of the authors: JOB, senior specialist in musculoskeletal radiology and DO, senior specialist in hand surgery.

Outcomes
The primary outcomes were DASH and PRWE.The secondary outcomes were: wrist range of motion, grip-strength, pinch strength, signs of osteoarthritis on CT examination and evaluation of fracture healing.

Statistics
Data analysis was done using SPSS version 27 for Macintosh (IBM, Armonk, NY, USA).Statistical differences between medians were tested using Mann-Whitney U test.P-value less than 0.05 was considered statistically significant.
Interobserver agreement was calculated for arthritic staging according to Whites classification [30].Interobserver agreement was analyzed with kappa statistics including calculations of CI.The resulting kappa value was 0.84, suggesting good interobserver agreement [31].

Ethics
Ethical approval was obtained from the Swedish Ethical Review Authority (ref.lu-459 03 and 2021 − 00957).The regional committee for radiation safety sanctioned the study.Informed written consent was obtained from all participants.

Results
Of 20 patients asked to participate in this long-term followup, 3 declined and one patient could not be reached.Thus, the participants comprised 11 males and 5 females with a median (range) age of 17,5 years (11 to 27 years) at injury.Median (range) time from injury to follow up was 16 years (8 to 17 years) (Table 2).
The most common trauma mechanism was a fall to an outstretched extended hand in the same plane, as six patients fractured their capitate in a simple fall and seven patients acquired the injury in a fall during sport activity such as  Of the 16 patients that were assessed at follow up all but two had been treated conservatively with a cast for 6 to 12 weeks.Two patients underwent surgery, one due to a concomitant dislocated scaphoid fracture and a Galeazzi-fracture, and one patient because of an associated TFCC injury.

Clinical evaluation
There were no significant differences in wrist range of motion, or grip-and pinch strength between the injured an uninjured wrist (Table 3).
football, floorball, basketball, ice hockey or skateboard.Two patients had a bicycle injury and one patient acquired the fracture in a fall on a staircase.
Five patients had an isolated capitate fracture.Ten had a capitate fracture in combination with a scaphoid fracture, whereof nine were undislocated (dislocation below 0.5 mm) and one was dislocated.One patient had a capitate fracture in combination with fracture in the base of the 3rd and 4th metacarpals.
All fractures were diagnosed on MRI.None of the capitate fractures were visible on conventional radiographs but 4 out of 15 fractures were detected on the CT examination.One of the patients did not undergo initial CT examination.

Discussion
The majority of capitate fractures are seen in combination with a scaphoid fracture whereas isolated capitate fractures are reported to be rare [6,32].In this study 10 out of 16 patients had an associated scaphoid fracture (Figs. 3 and  4).Five patients out of 16 had an isolated capitate fracture which is a higher rate than reported in earlier studies.An explanation for the higher number of isolated fractures compared to prior literature could be that the diagnosis was based on MRI and not on conventional radiographs.Furthermore, fracture classification based on localization of the fracture might differ between imaging modalities.In our One patient with an associated scaphoid fracture that healed with a slight humpback deformity, had signs of midcarpal instability with a positive Lichtman test.

PROM
The median DASH score at follow up was 0 (range 0-33).The median PRWE was 3 (range 0-29).(Table 3) Two patients had DASH and PRWE scores indicating severe disability.One of these had carpal tunnel syndrome at the time of follow-up, while the other patient originally had more complex injury in terms of a capitate fracture in combination with a scaphoid-and Galeazzi-fracture.

Radiology
At follow-up, and demonstrated by CT examination, all fractures were healed.One patient did not do a CT-examination at follow up due to pregnancy.No patient demonstrated radiological signs of avascular necrosis of the capitate.
Only one patient demonstrated slight posttraumatic osteoarthritis which was located between the capitate and the lunate (Fig. 2).In this patient the capitate fracture had healed with a slight volar compression and the associated scaphoid fracture with humpback deformity.There was a scapholunate angle of 82 degrees.Clinically the patient had signs of midcarpal instability with a positive Lichtman test.The patient also had a moderate loss of grip strength (66% strength compared to the uninjured side) and pinch strength (64% compared to the uninjured side).Despite  [40].We speculate that the early ossification of the capitate is the reason for the low median age of fractures in our cohort.
It has been emphasized, that early diagnosis is important, as delayed treatment of capitate fractures can lead to avascular necrosis, nonunion, and post-traumatic osteoarthritis [3,14].Several studies advocate early MRI and/or CT examination in patients with post traumatic wrist pain but normal conventional radiographs [1,41,42].Our study supports this recommendation since none of the capitate fractures were visible on conventional radiographs.However, only 4 of 16 capitate fractures were detected on CT.One explanation for the low detection rate on CT is that a 16-slice CT scanner was used.Based on our study, MRI is recommended when clinical suspicion of a capitate fracture exists, but CT results are normal.Newer MR scanners with higher signal-to-noise ratios are likely to enhance image quality.Additionally, dual-energy and photon-counting CT, now becoming standard, may significantly improve the diagnosis of capitate fractures.Future studies are needed to assess to role of MRI and CT in the diagnostic work-up of patients with suspected capitate fractures.Gelberman et al. [11] suggested that the capitate is vulnerable to post traumatic AVN due to its retrograde vascular supply.However, a recent study based on micro-CT angiography have shown than the capitate is also supplied by a more proximal vascular system.This suggests that individuals without a proximal vascular supply could have in increased risk for development of AVN [12,14].However, there are only few cases described with AVN after an isolated capitate fracture [2,14] which is in accordance with our series where we did not find any case with AVN.All patients in our study were diagnosed and treated shortly of trauma.If left untreated, the most common fracture type in our study, the low transverse fracture, could gradually dislocate causing the head of the capitate to rotate with wrist movement.This could cause interruption of the vascular supply with the possibility to cause AVN or nonunion [14,15,43,44] Delayed union, malunion and nonunion have been reported following capitate fractures [13].Rand and colleagues found two non-unions in 11 fractures whereas Kadar and colleagues, only found one non-union out of 23 fractures corresponding to a 4% non-union rate [3,6].In addition, a literature review from 1999 described a total of 11 capitate nonunions [16].Our long-term results support that early treated capitate fractures with no or minor dislocation carries low risk for developing pseudoarthrosis, malunion or avascular necrosis.In clinical practice nonunion of a prior non diagnosed capitate fracture is very rare.This supports the notion that minimally displaced capitate fractures, without other associated injuries, only need to be immobilized in a cast for the duration of pain.Our long-term results support that capitate fractures with no or study, the diagnose and fracture classification are based on MRI.The oblique low (37.5%) and transverse low (37.5%)fracture were the most common fracture types.However, the small number of fractures in our study does not allow for statistical analyze of fracture types in relation to age or type of trauma.
The capitate is protected from trauma by the surrounding bones (and strong ligament attachments).However, this study shows that a capitate fracture alone or in combination with a scaphoid fracture, are most commonly caused by a simple fall in the same level, onto an extended wrist.One interesting question is why a fall to an outstretched hand result in a scaphoid and capitate fracture in a majority of cases whereas some only sustain an isolated capitate fracture.Prior biomechanical studies have suggested that depending on the position of the hand and wrist at the time of injury, the type and presentation of the capitate fracture varies [33].Among patients with capitate and scaphoid fractures, Fenton [34] suggested that the scaphoid fracture occurs first and secondarily, if the injury force is not fully dissipated on the scaphoid, fracture of the neck of the capitate takes place.Based on biomechanical studies, were the wrist was forces in extension and ulnar deviation, Mayfield proposed a classification of perilunate dislocations and proposed that ligamentous injuries develop in a sequential fascion [35] Furthermore, Sterin and Siegel found in a cadaver specimen that fracture of the capitate could be produced in forced extension by the dorsal lip of the radius impacting on the capitate whereas the scaphoid fractured in tension created at the midcarpal joint level by forced extension [36].Future biomechanical studies are needed for better understanding of the trauma mechanisms behind both isolated capitate fractures and capitate fractures combined with other injuries.However, it seems reasonable to assume that a combined scaphoid and capitate fracture indicates a more severe trauma requiring a closer follow-up in order to detect or exclude associated ligament injuries [37].We found, one patient with radiologic osteoarthritis between the capitate and lunate (Fig. 2) The patient had a malunited scaphoid fracture, DISI configuration but no scapho-lunate dissociation.He also had a positive Lichtmans test.Given the patient status and what is known from biomechanical studies we believe that the injury is best described as a greater arc injury.However, this patient did not experience any subjective problems from the wrist.
Carpal fractures in children aged under 16 years are rare and among them, scaphoid fractures are most common [32,38].The median age of our patients with capitate fractures was 17.5 years which is younger than reported in a prior study [6] and also younger than reported for scaphoid fractures [39].The capitate is the first carpal bone to start ossification which is completed at a mean age of 15.3 years in minor dislocation, with or without a concomitant scaphoid fracture, treated with a cast at the time of injury carries low risk for developing pseudoarthrosis, malunion or avascular necrosis.In clinical practice nonunion of a prior non diagnosed capitate fracture is very rare.We speculate that minimally displaced capitate fractures without other injuries can be treated in a cast.A CT scan at 3-4 weeks can assess healing; if bridging callus is present, the fracture is stable and can be mobilized.However, capitate fractures with a concomitant scaphoid fracture may be more unstable and future studies are needed to assess if such injuries should be treated with longer cast immobilization.
Existing classifications systems for arthritis are solely based on conventional radiographs [30,45] One previous study modified the White´s classification in order to include the additional information obtained from CT [29].It allowed a more precise assessment of joint space narrowing as well as information on periarticular arthritic changes.Only one of our patients demonstrated signs of osteoarthrosis at the follow-up after a median of 16 years.This is in contrast to fractures of the scaphoid distal pole where osteoarthrosis in the STT (Scaphotrapeziotrapezoid) joint is more common [30].It is well known that radiological signs of osteoarthritis not always leads to clinical symptoms.In accordance, our patient with radiological signs of arthritis between the capitate and lunate did not experience any clinical symptoms.
The strengths of our study are the inclusion of consecutive patients from a large prospective cohort and the use of conventional radiographs, CT and MR at diagnosis and CT for assessment at follow-up.However, some limitations are evident.Due to a focus on scaphoid fractures the inclusion criteria were posttraumatic radial sided carpal pain.Capitate fractures which present with central or ulnar pain may have been missed.Secondly, the number of capitate fractures is low, and it was not possible to recognize significant differences on outcome measures between fracture types.Finally, our modified classification system for osteoarthritis based on CT limits direct comparison with results from other studies.

Conclusion
I this study, we found that patients with a capitate fracture were young, had acquired the fracture due to a fall in the same plane, and that the majority had a concomitant scaphoid fracture.Furthermore, the capitate fractures were often undislocated, and not visible on conventional radiographs.The risk of associated ligament injury and subsequent instability is low.The majority of capitate fractures can be treated in a cast and achieve excellent clinical result without risk of developing secondary arthritis or avascular necrosis.

Table 3
Results at long-term follow up a Associated Galeazzi fracture b At follow up signs of osteoarthritis and DISI c Patient was diagnosed with a severe carpal tunnel syndrome prior to assessment in the study these findings, the patient rated hand function was excellent (DASH 2.5 and PRWE 1.5).