The Natural History of Scaphoid Fracture Malunion: A Scoping Review

Abstract

Background  A scaphoid malunion occurs when a scaphoid fracture heals in a nonanatomic position or when the fracture is fixed without correction of the sagittal angular deformity. Although altered carpal mechanics and early osteoarthritis have been suggested as natural sequelae, the natural history and clinical outcomes are debatable.

Purposes  The purpose of this study is to review and summarize the available literature regarding clinical, functional, and radiographic outcomes of patients with scaphoid malunion.

Methods  A systematic search of the MEDLINE/PubMed, EMBASE, Cochrane Library, and Web of Science was performed to identify published studies concerning the clinical and radiological results of scaphoid malunion following either acute scaphoid fracture or surgically treated nonunions.

Results  Five publications with a total of 83 malunions were included in the final synthesis. The diagnosis of malunion was based on computed tomography by calculating the lateral intrascaphoid angle (threshold of 35 or 45 degrees) or height/length ratio (threshold of 0.6). Self-reported and clinical evaluation techniques varied between the studies and direct comparison was not possible between the different outcome measures. Most patients demonstrated arthritic changes; however, correlation with pain and functional results was not always present.

Conclusions  This scoping review confirmed that patients with malunited scaphoids seem to have higher likelihood of post traumatic arthritis. However, clinical implications remain uncertain and better methods for assessing and defining scaphoid deformity are required.

Scaphoid fractures are a common orthopaedic injury accounting for greater than 58% of all carpal bone fractures, with an annual estimated incidence of 12 to 43 fractures per 100,000 people. ^{1 2 3 4 5} Although the majority of these fractures are stable and heal uneventfully with cast immobilization, nonunion and malunion may occur. While we know that untreated scaphoid nonunion will lead to the development of degenerative arthritis, ^{6 7 8} less is known about the consequences of scaphoid malunion.

A scaphoid malunion occurs when an acute displaced scaphoid fracture heals in a flexed or nonanatomic position or when a scaphoid fracture is fixed without correction of the sagittal-plane angular deformity. Unstable and displaced fractures are prone to malunion, usually in flexion. ^{6 9} It is suggested that such angular deformity will result in an alteration in wrist biomechanics and potentially cause persistent pain, loss of motion, diminished grip strength, and an increase in the risk for posttraumatic degenerative arthritis. ^{10 11 12 13 14 15 16 17} Although these associations have been demonstrated in the biomechanical literature, ¹⁶ the clinical impact of malunion is controversial. Studies report conflicting findings with some suggesting malunions causes little–to-no functional loss and affected patients can remain asymptomatic even at long-term follow-up. ^{18 19 20}

Despite the adverse clinical and radiological outcomes that have been attributed to scaphoid malunion, it has received relatively little attention in the literature as a separate and unique pathologic entity. Most frequently, it is simply mentioned as a potential complication of nonunion surgery or in association to ligamentous injuries around the scaphoid. ^{15 18 19 21 22} As a result, there is no clear consensus as to what degree of scaphoid deformity can be tolerated without compromising wrist function or how best to manage a patient who has healed with an asymptomatic malunion.

Given the controversy and general paucity in the current literature regarding scaphoid malunion incidence and its natural history, we attempted to review and synthesize the available literature in an attempt to fill this void. ²³ Because of the scarcity of data and variety of study designs, a typical meta-analysis or even systematic review was not possible. As such, we performed a scoping review; a scoping review is a useful tool for a broad analysis of literature when exploring an under-examined area of research. ²⁴ It differs from a systematic review, which has a specific and targeted question with narrow criteria and parameters, such that the focus is not on the assessment of the quality of studies but rather on the examination of the range and nature of current research on a topic to identify gaps in knowledge. ²⁵ By comprehensively reviewing the literature, we hope to provide evidence-based data to aid decision making in the management of these injuries. Furthermore, we aimed to summarize the existing literature regarding the long-term clinical implications of a scaphoid malunion and to identify radiographic outcomes that are responsible for the creation of a symptomatic malunion.

Materials and Methods

MEDLINE/PubMed, EMBASE, Cochrane Library, and Web of Science were used to perform a thorough review of the literature. All years until September 2017 were searched with the following combinations of keywords: “scaphoid,” “carpal navicular,” “fracture,” “malunion,” “malunited,” “abnormal union,” “flexion deformity,” and “humpback deformity.” All terms were searched as keywords, and when available, MeSH, EMTREE, or other applicable subject terms were searched as well. The reference lists of included studies were also subsequently hand searched to identify additional relevant articles. Non-English publications, theses, case reports, general reviews, and letter to the editor articles were excluded.

Initially, titles and abstracts were screened for relevance and articles concerning the clinical and radiological results of scaphoid malunion following either acute scaphoid fractures or surgically treated nonunions were included. Studies that did not separate outcomes of malunion from nonunions or fractures that united without a residual deformity were excluded. Studies with a follow-up less than 6 months and articles concentrating on scaphoid fractures in pediatric patients were also excluded. For all articles meeting our inclusion and exclusion criteria, the number of scaphoid malunions, measurement criteria used for malunion diagnosis, presence of previous nonunion, type of treatment, and subjective and objective clinical and radiological outcomes were extracted and summarized into a database.

Our initial search yielded 190 citations in total with 113 after the removal of duplicates. After reviewing the title and abstracts, 22 articles were identified for a full text review and assessment using our inclusion and exclusion criteria. Ultimately five studies, with a total of 83 malunions, met the eligibility criteria for inclusion in our scoping review ( Fig. 1 ). ^{15 18 19 20 26} Table 1 describes the information evaluated by these studies.

Fig. 1.

Flow diagram of research results.

Table 1. Summary of all articles included in this scoping review ( n  = 5) .

Study (year)	Study type	Number of malunions	Fracture characteristics	Events preceded malunion	Average follow-up, y, (range)	Male gender (%)	Average age, y, (range)	Imaging modality for diagnosis	Diagnostic criteria for malunion	Average malunion parameters	Comparison group	Outcome	Outcome in malunion group
Jiranek et al (1992) 18	Case series	13	Waist or proximal pole	Nonunion treated with the Russe bone graft (distal radius/iliac crest)	11.5 (7–17)	100	22 (16–43)	CT	LISA > 45	LISA 63 degrees (range: 47–87 degrees)	13 waist/proximal pole nonunions treated with the Russe bone graft LISA < 45 degrees	Total objective score a (union, OA, grip strength, ROM)	Significant difference between the groups ( p  < 0.001)
												Grip strength, ROM, total subjective scores b (function, pain, ROM, satisfaction)	No significant difference between the groups
Lee et al (2015) 19	Prognostic case series	15	Waist (67%) or proximal pole (33%)	Nonunion treated with vascularized/nonvascularized bone graft (iliac crest) + IF	7.0 (5.4–9.2)	96	29.6 (15–46)	CT	H:L > 0.6	H:L 0.69 (range: 0.6–0.83)	10 waist nonunions treated with autogenous iliac bone graft + IF H:L < 0.6	Grip strength, ROM, pain, evaluation questionnaire	No significant difference between the groups
Gillette et al (2017) 26	Retrospective cohort study	6	Not clear	NR	22.6 (12.5–30.1)	82.3	26.7 (15–66)	CT	LISA > 45	LISA 58 degrees (range: 55–75 degrees)	11 scaphoid frx malunions (LISA > 45 degrees) treated by corrective intrascaphoid osteotomy (4) or salvage procedure (7)	Quick DASH, PRWE	No significant difference between the groups
Forward et al (2009) 20	Prospective cohort study	23	Waist	Primary fracture treated with below elbow cast	1 (0.99–1.07)	90.5	31 (15–61)	CT	H:L > 0.6	NR	19 Primary frx treated with below elbow cast, H:L < 0.6	Grip strength, ROM, pain, PEM, and DASH scores	No significant difference between the groups
Amadio et al (1989) 15	Case series	26	Waist or proximal pole	Primary (6), nonunion treated with bone graft (20)	5.3 (1–10.3)	100	20.4 (15–34)	CT	LISA > 35	LISA 46.73 degrees (range: 35–60 degrees)	20 waist or proximal pole fractures treated with cast (13) or by “open methods” (7) LISA < 35 degrees	Grip strength, ROM, total score c	Decreased
												Posttraumatic OA	Increased

Abbreviations: CT, computed tomography; DASH, disabilities of the arm, shoulder, and hand; frx, fracture; H:L, height-to-length ratio; IF, internal fixation; LISA, lateral intrascaphoid angle; NR, not reported; OA, osteoarthritis; PEM, patient evaluation measure; PRWE, patient-rated wrist evaluation; ROM, range of motion.

^a Objective rating scale (range: 35–100): modification of scoring rating system proposed by Cooney et al, ³⁸ based on union type (10–25), osteoarthritis (5–25), range of motion (10–25), and grip strength (10–25).

^b Subjective rating scale (range: 0–100): modification of scoring rating system proposed by Cooney et al, ³⁸ based on patients opinion regarding their function (0–40), pain (0–30), ROM (0–10), and overall satisfaction (0–10).

^c Cooney et al's evaluation scale (range 0–100), modification of Green and Obrien wrist evaluation scale. ³⁸

The average age of the study population was 25.7 ± 4.6 years (range: 15–66 years) with the majority of male (95%) patients.. Forty-eight malunions had been previously treated with bone grafting to achieve union, 29 occurred after primary fracture treated with casting, and the origin of the remaining six was not disclosed. Three articles reported on malunion following scaphoid waist or proximal pole fractures treated surgically. ^{15 18 19} One article included malunions solely around the waist. ²⁰ And finally, one article did not specify the exact fracture location. ²⁶ Given the wide range of study designs, fracture characteristics, and assessment modalities, direct comparisons were not possible between the different outcome measures.

Results

Criteria for Diagnosis of Malunion

All five studies assessed the presence of malunion based on computed tomography (CT) scan, using the most central sagittal cut for measuring the lateral intrascaphoid angle (LISA) ^{15 18 26} or the height-to-length ratio (H:L). ^{19 20} Among the studies that measured LISA, two studies set a minimum angle for malunion at 45 degrees, ^{18 26} and one study at 35 degrees. ¹⁵ The two studies that used the H:L used a threshold of 0.6 for diagnosing malunion. ^{19 20}

Self-Reported Functional Assessment

The reported scores based on functional questionnaires are summarized in Table 2 . Three of the five studies evaluated the functional outcomes using several previously validated self-administered questionnaires for wrist pain and disability. ^{19 20 26} Among them, three studies ^{19 20 26} used the disabilities of the arm, shoulder, and hand (DASH) or the quick DASH questionnaire, ^{27 28} two studies ^{19 20} used the patient evaluation measure (PEM) score, ²⁹ and one study ²⁶ used the patient rated wrist evaluation (PRWE) questionnaire. ^{30 31} A total of 38 patients were evaluated using the DASH and the PEM questionnaires, with a mean score of 4.14 (range: 0–10) and 11.14 (range: 3–32), respectively. ^{19 20} Pain levels were assessed in three studies, each time using a different scale.

Table 2. Overall reported scores of functional questionnaires for scaphoid malunion patients.

Study (year)	Follow-up, y, (range)	Questionnaire 1 (range)	Questionnaire 2 (range)	Questionnaires 3 (range)
Jiranek et al (1992) 18	11.46 (7–17)	Subjective rating scale a 85 (26–100)	NA	NA
Lee et al (2015) 19	7.03 (5.42–9.17)	PEM b 17.5 (14–32)	DASH c 2.82 (0–6.9)	Mayo wrist score d 90.3 (70–100)
Gillette et al (2017) 26	22.57 (12.5–30.1)	PRWE e 33.5 (0–73)	Quick DASH c 22 (0–50)	NA
Forward et al (2009) 20	1 (0.99–1.07)	PEM b 7 (3–12)	DASH c 5 (1–10)	NA
Amadio et al (1989) 15	5.26 (1–10.33)	Cooney et al f 80.38 (50–100)	NA	NA

Abbreviations: DASH, disabilities of the arm, shoulder and hand questionnaire; NA, not available; PEM, patient evaluation measure; PRWE, patient-rated wrist evaluation; ROM, range of motion.

^a Subjective rating scale (range: 0–100): modification of scoring rating system proposed by Cooney et al, ³⁸ based on patients opinion regarding their function (0–40), pain (0–30), ROM (0–10), and overall satisfaction (0–10).

^bPEM questionnaire (range: 0–100); lower score representing a better outcome.

^cDASH questionnaire (range: 0–100), quick DASH (range 0–100); Lower score representing a better outcome.

^dMayo wrist score: modified Mayo wrist scoring system (range: 0–100); higher score representing a better outcome.

^ePRWE (range: 0–100); lower score representing a better outcome.

^f Cooney et al's evaluation scale (range 0–100), modification of Green and Obrien wrist evaluation scale. ^{38 39}

Clinical Functional Assessment

The average grip strength was reported in four studies and ranged between 76 and 98.9% of the contralateral side, with a mean of 86.16%. ^{15 18 19 20} The relative wrist motion was reported in three studies, based on total wrist motion or flexion and extension alone. ^{18 19 20} The relative extension range of motion was explicitly reported in two studies ^{19 20} and ranged between 67 and 74% of the noninjured side, with a mean of 71.24%. Of four studies that examined and compared grip strength, range of motion (ROM), and pain between the malunited and the control groups, only Amadio et al detected a significant difference ( Table 3 ). ^{15 18 19 20}

Table 3. Overview of reported clinical outcomes measurements for scaphoid malunion patients.

Study (year)	Grip strength (% contralateral)	Pain (range)	Wrist motion (% contralateral)	Extension (degrees)
Jiranek et al (1992) 18	76 (65–100)	NR	78 d (55–94)	NR
Lee et al (2015) 19	98.9 (65–110)	VAS a 1.2 (0–4)	93.7 d (80–100)	67 (45–75)
Gillette et al (2017) 26	NR	NR	NR	NR
Forward et al (2009) 20	95 (88–100)	0 b (0–2)	98 e (97–100)	74 (70–79)
Amadio et al (1989) 15	76.07 (24–132)	0.96 c (0–3)	NR	NR

Abbreviations: NR, not reported; VAS, visual analog scale.

^aVAS: (0, no pain and 10, worst pain imaginable); mean (range).

^bPain on a scale of 0–6, median (range).

^cPain scale: 0, none; 1, mild; 2, moderate; 3, severe; mean (range).

^dFlexion–extension arc as a percentage of the noninjured side.

^eTotal wrist motion as a percentage of the noninjured side.

Posttraumatic Arthritis

Arthritis, based on radiographic examination, was evaluated in two studies ( Table 4 ). Using trispiral CT, Amadio et al found a significant association between increased LISA and presence of posttraumatic arthritis ¹⁵ with incidence as high as 54% in the malunion group compared with 20% in the normal alignment group (after an average follow-up of 5.3 years). However, the criteria for defining arthritis or its location were not specified and the severity was not reported. Jiranek et al compared 13 malunited scaphoid fractures and a group of 13 scaphoid fractures with normal alignment. ¹⁸ Both followed Russe's procedures for symptomatic nonunion. During the follow-up, arthritic changes were recorded and graded according to their location and expansion based on CT. At an average follow-up of 11 years, all the patients with scaphoid malunion had some degree of arthritis as did the patients in the normal alignment group; however, the degree of arthritis severity was higher in the malunited group. The most common site for arthritis was at articulation between the radial styloid and the distal pole of the scaphoid. The next most common sites for arthritis were the scapholunate, scaphotrapezium, and scaphotrapezoid joints. However, there was no significant association between the severity of arthritis and the reported pain level, possibly because the arthritis was mostly mild or moderate.

Table 4. Osteoarthritis prevalence and severity.

Study (year)	Number of malunions	Average malunion follow-up, y, (range)	Presence of post traumatic OA (%)	Severity (range)	OA location
Jiranek et al (1992) 18	13	11.5 (7–17)	100	1.77 (1–3) a	RS, SL, STM, STD
Amadio et al (1989) 15	26	5.3 (1–10.3)	54	NR	NR

Abbreviations: NR, not reported; RS, radial styloid; SL, scapholunate; STD, scaphotrapezoid; STM, scaphotrapezium.

^a Post traumatic arthritis on a scale of 0 to 4: 0 = no arthritis; 1 = arthritis affecting the radial styloid process and the distal pole of the scaphoid only; 2 = arthritis affecting the scaphotrapezial and scapholunate joints, radial styloid process and distal pole of the scaphoid; 3 = arthritis affecting the scaphocapitate, capitolunate, scaphotrapezial, and scapholunate joints, the radial styloid process, and the distal pole of the scaphoid; and 4 = general arthritis. ¹⁸

Discussion

The purpose of this review was to investigate the natural history of scaphoid malunion. While it has been recognized as a potential cause for residual functional disability, pain, and posttraumatic arthritis, the literature reports varying results. Although scaphoid fracture malunion is now a well-recognized entity, no clear guidelines direct the treatment for a scaphoid that has united in a deformed position. Many authors believe that occurrence of union alone is not enough to prevent secondary osteoarthritis and correction of the intrascaphoid angular malalignment is necessary to restore carpal kinematics and prevent degenerative changes in the future. ^{11 14 15 16 22 23 32 33} However, while some surgeons have advocated treating symptomatic malunions of the scaphoid with a corrective osteotomy, ^{17 23 34 35 36} others have questioned the correlation between the scaphoid malunion and clinical outcomes and do not recommend an osteotomy of a bone that is known to have difficulty obtaining union. ^{19 20 26 33 37}

Lindstrom and Nystrom retrospectively studied 229 healed scaphoid waist fractures with a minimum follow-up of 7 years and found 5.6% of patients displayed patterns of radiographic radiocarpal arthrosis. ³² Although no radiographic measurements of scaphoid malunion were used, the authors hypothesized that a change in carpal dynamics, due to the deformation and shortening of the scaphoid, was most likely the cause of the posttraumatic osteoarthritis and suggested that consideration for open reduction and internal fixation should be made not only for comminuted and displaced fractures but also in selected cases with obvious deformity and angulation of the scaphoid. ³² This is also supported by Amadio et al who reported a significant association between LISA value and the presence of posttraumatic arthritis, decreased relative grip strength, and decreased functional scores. ³⁸ Two studies in our review evaluated posttraumatic changes and found a higher incidence and severity in fractures that healed with malunion as compared with fractures with union and acceptable alignment. ^{15 18} Posttraumatic arthritis was also found to have a significant effect on pain, ROM, and grip strength. ¹⁵

However, others report more favorable outcomes following scaphoid malunion. ^{18 19 20 26 37} Jiranek et al showed compromised objective results but good subjective scores, ¹⁸ and more recent publications ^{19 20} have failed to recognize any statistically significant difference between the malunited and the normal alignment fractures in term of grip strength, ROM, pain, and self-reported functional questionnaires.

One of the factors that might challenge Amadio et al's findings is their use of the modified scale of Green and O'Brien that has not been validated for the evaluation of scaphoid fractures. ^{29 38 39} Another concern is the inclusion of healed scaphoid nonunions in the same group as malunions of an acute fracture. Since scaphoid nonunions have a known link to the development of degenerative changes, it is impossible to differentiate the extent to which the posttraumatic OA is a result of the nonunion versus the altered kinematics created by the malunion. Among the 26 patients with malunion, 20 (77%) were previously treated for nonunion as opposed to only 7 (35%) among the 20 patients in the normal alignment group. ¹⁵ This uneven distribution, in Amadio et al's study, may supply an alternative explanation for the inferior clinical and radiographical results in the malunion group.

As no study has identified which factors are responsible for symptomatic as opposed to asymptomatic malunions, we can assume that additional variables, such as the initial mechanism of injury, the applied force during the fall, cartilage damage, presence of associated ligamentous injury, and daily loads to individual may contribute to the clinical presentation. ^{14 26}

Our study has several limitations. The included case series have small sample sizes, include a heterogeneous patient population and have mostly retrospective study designs. However, one of the primary limitations identified in this study relate to the methodology of defining and quantifying scaphoid malunions. While some articles choose H:L greater than 0.65 as their cut-off value, ⁴⁰ others define malunion as H:L above 0.6. ^{19 20} Based on LISA, malunion is suggested when the calculated angle exceed 45 degrees ^{18 26 40} while a cut-off value of 35 degrees was historically suggested by Amadio et al.

There is not only a lack of agreement regarding the definition of malunion but also no consensus regarding the precise and universally accepted techniques to assess the degree of scaphoid deformity to establish the definition of a malunion. The H:L has been shown to be the most reliable and reproducible measure; however, studies continue to rely on the inferior measurements, such as the DCA and LISA. ^{18 20 26 41 42 43} Table 5 summarizes the available measurement techniques. Although the H:L is reported as the most reliable, it too has been shown to be a difficult measure to achieve high-inter and -intrarater reliability. ⁴⁴

Table 5. Options for scaphoid fracture malunion assessment.

Name	Plane	Technique
Lateral intrascaphoid angle (LISA)	Sagittal view	Angle created between lines drawn perpendicularly to the proximal and distal articular surface of the scaphoid
Anteroposterior intrascaphoid angle (ISA)	Coronal view	Angle created between lines drawn perpendicularly to the proximal and distal articular surface of the scaphoid
Height-to-length ratio (H:L)	Sagittal view	The ratio is calculated by dividing the scaphoid height by its length. A parallel baseline line is drawn along the palmar aspect of the scaphoid. The length is measured along the baseline from the most proximal to the most distal part of the scaphoid. The maximum height of the scaphoid is measured on a line perpendicular to the baseline
Dorsal cortical angle (DCA)	Sagittal view	Angle created between tangential lines drawn along the dorsal cortex of the proximal and distal half of the scaphoid

Difficulty in achieving consistent measures also relates to the image chosen to make the measurements, as scaphoid height and length varies based on the selected CT slice chosen for assessment. ²⁰ Ring et al also reported that the choice of the reconstruction plane may add additional variability to measurements of scaphoid deformity. ⁴²

Another limitation to our stated purpose of assessing the natural history of scaphoid malunion is that many of the studies also included scaphoid nonunions, which were surgically treated, but ultimately healed with a humpback deformity. Sequalae of scaphoid nonunion advanced collapse arthritis (depending on the chronicity of the nonunion), soft tissue damage, vascular impairment, and residual scar after surgical procedure may all serve as confounders in an attempt to attribute less favorable results exclusively to the residual deformity.

The lack of standardized outcome measurements and consistent follow-up imaging are further limitations to our study.

Based on current literature, a malunited scaphoid does seem to associate with a higher likelihood of posttraumatic arthritis; however, the true incidence and clinical implications of malunion are uncertain. Influenced by scanning plane, position of the wrist, subjective selection of the most central sagittal slide, and manual measurement of length and height on the chosen CT slide, assessing scaphoid deformity is fraught with pitfalls. Given the above, consensus on the definition of a malunion should be reached and measurement techniques should be standardized and improved. Future studies should concentrate on clinical and radiological outcomes following primary scaphoid malunions.

Until then, the true impact of malunion will remain unclear and debatable.