Complications after Nonoperative Management of Hamate Fractures.

Hamate fractures are an infrequent injury and are often missed or have a delayed diagnosis with potential for significant patient morbidity. There is a relative paucity in the literature involving large population studies of hamate fractures and subsequent complications. Gaining a better understanding of complications associated with nonoperative management will help guide the decision for operative intervention.
Methods: The PearlDiver patient records database was used to query for patients who sustained a hamate fracture using ICD-9 and ICD-10 codes. Postinjury complications within 1 year of diagnosis were assessed using ICD-9 and ICD-10 codes for diagnoses of complications.
Results: A total of 1120 patients who sustained a hamate hook or body fracture met inclusion and exclusion criteria and were included in the study. Patients who were managed nonoperatively showed a nonunion rate of 2%, ulnar neuropathy rate of 1.7%, and tendon rupture rate of 0.2%. In a subanalysis using only ICD-10 codes to distinguish between hook and body fractures, hook fractures demonstrated a nonunion rate of 2.2%, ulnar neuropathy rate of 2.7%, and tendon rupture rate of 0%, while body fractures had a nonunion rate of 1.2%, ulnar neuropathy rate of 1%, and tendon rupture rate of 0%.
Conclusion: The current study shows very low rates of nonunion, ulnar neuropathy and tendon rupture after nonoperative management of hamate fractures.

Takeaways

Question: What are complication rates following nonoperative management of hamate fractures in a large population study?

Findings: The current study, utilizing the PearlDiver patient records database, shows very low rates of nonunion, ulnar neuropathy, and tendon rupture after nonoperative management of hamate fractures.

Meaning: The study finds that patients with hamate fractures can be safely treated nonoperatively due to the low risk of major complications and that treatment decisions can be more dependent on variables such as time to return to work or sport.

INTRODUCTION

Hamate fractures are an infrequent injury, accounting for only 2%–4% of carpal bone fractures.[1] Due to the rarity of this injury, hamate fractures are often missed or have a delayed diagnosis with potential for significant patient morbidity. Persistent pain following this injury is a common indication for further workup and subsequent diagnosis of a hamate fracture. Complications include nonunion, posttraumatic arthritis, ulnar neuropathy, or flexor digitorum profundus tendon rupture. Nonunion is the most common complication following a hook fracture and can reach a rate of 50% with conservative treatment.[2] Tendon rupture is a serious complication; its rarity has limited study of incidence to case reports.[3,4]

Hamate fractures can be classified using the Milch classification into fractures of either the body or hook. Fractures of the hook are typically associated with sporting activity that involves direct force onto or across the hook of hamate, such as baseball, golf, and racquet sports.[5-8] Body fractures are often associated with direct blows over the hypothenar eminence, strong dorsopalmar compression, or other forms of trauma.[4,5]

Treatment involves nonoperative management with immobilization or operative management with excision versus open reduction and internal fixation. In a study with 81 patients, Bansal et al reported higher rates of transient ulnar nerve dysfunction after surgical excision of the hook.[9] Lamas-Gomez found that open reduction and internal fixation of hamate hook fractures was a safe and effective treatment to restore normal grip strength and return to preinjury levels in a study of 13 patients.[10] Treatment decisions are typically made with the consideration of nonunion, ulnar neuropathy, and tendon rupture. To our knowledge, literature on nonoperative management has been limited to small case reports and case series. The purpose of this study was to evaluate nonoperative management for hamate fractures in a national database and characterize the subsequent complications, including nonunion, ulnar neuropathy, and tendon rupture. Gaining a better understanding of complication rates and risk factors in a large study cohort may influence management algorithms of such injuries.

MATERIALS AND METHODS

Patient records were queried from PearlDiver (PearlDiver Inc., Fort Wayne, Ind.), a commercially available administrative claims database, using International Classification of Diseases, Ninth Revision and Tenth Revision (ICD-9/ICD-10). This study utilized the Humana dataset, which contains the medical records of 25.4 million patients from 2007 to 2017 who were privately insured, commercially insured, or purchased their Medicare Advantage plans through Humana Health Insurance. Institutional review board exemption was granted for this study because provided data were de-identified and compliant with the Health Insurance Portability and Accountability Act.

A retrospective observational study design was used to evaluate the epidemiology and trends of complications and treatments from patients who sustained hamate hook and body fractures. Study participation was identified by ICD-9, ICD-10, and CPT codes, and included all patients with hamate fractures. ICD-9 coding incorporated all hamate fractures into a single code, whereas ICD-10 coding subdivided hamate fractures and allowed for the distinction between body and hook fractures. Patients who underwent surgical intervention were identified using CPT codes 25645 [open treatment of carpal bone fracture (excluding carpal scaphoid navicular), each bone] and 25210 (carpectomy; one bone). A preliminary data analysis was performed on this operative cohort, and due to a low sample size and equivocal findings, this group was excluded from the study. Other exclusion criteria included patients with other carpal bone fractures, metacarpal fractures or dislocations, phalanx fractures, and distal radius fractures.

Each cohort was queried for basic demographic information and clinical characteristics such as age, sex, body mass index (BMI), and incidences of specific complications. Complications queried from the database included nonunion, ulnar neuropathy, and tendon rupture. Patients were tracked until 1 year after their injury.

Descriptive and statistical analysis was performed with the programming language R (University of Auckland, New Zealand) analyzing odds-ratios with their respective 95% confidence interval (95% CI), chi-squared distribution, and P values. An alpha value of less than 0.05 was considered statistically significant.

RESULTS

A total of 1120 patients who were managed nonoperatively after sustaining a hamate hook or body fracture were identified (Table 1). In total, 621 of these patients, identified by ICD-10 coding, were subdivided to distinguish and evaluate hamate hook and body fractures separately. In total, with both ICD-9 and ICD-10 coded patients, there was a nonunion rate of 2%, ulnar neuropathy rate of 1.7%, and tendon rupture rate of 0.2%. In the hook fracture cohort, the rate of nonunion was 1.2%, ulnar neuropathy was 1%, and tendon rupture was 0%. The hamate body fracture cohort demonstrated a nonunion rate of 2.2%, ulnar neuropathy rate of 2.7% and tendon rupture rate of 0%. There was no significant difference in the complication rates between the cohorts (P > 0.05).

Table 1.

Complication Rates after Nonoperative Management of Hamate Fractures

	Total(ICD-9 & ICD-10)(n = 1120)	Body(ICD-10 only)(n = 435)	Hook(ICD-10 only)(n = 186)	P
Nonunion	22 (2%)	5 (1.2%)	4 (2.2%)	0.32
Ulnar neuropathy	19 (1.7%)	4 (1%)	5 (2.7%)	0.48
Tendon rupture	2 (0.2%)	0 (0%)	0 (0%)

Demographic information of patients and subsequent complications can be found in Table 2. There was no statistically significant difference in the age, gender, and BMI distribution in the cohorts that did and did not develop complications (P > 0.05).

Table 2.

Demographics of Complications following Nonoperative Management

	Nonoperative Patients (n = 1120)	Nonunion(n = 22)	Ulnar Neuropathy (n = 19)	P
Age				0.11
10–19	162 (14.5)	3 (13.6)	1 (5.3)
20–29	262 (23.4)	4 (18.2)	3 (15.8)
30–39	198 (17.7)	2 (9.1)	1 (5.3)
40–49	147 (13.1)	2 (9.1)	5 (26.3)
50–59	144 (12.9)	7 (31.8)	3 (15.8)
60–69	103 (9.2)	2 (9.1)	5 (26.3)
70–79	98 (8.8)	3 (13.6)	1 (5.3)
Gender				0.41
Men	645 (57.6)	14 (63.6)	12 (63.2)
Women	475 (42.4)	8 (36.4)	7 (36.8)
BMI				0.39
Less than 30	104 (42.3)	2 (18.1)	2 (25)
30–35	71 (28.9)	4 (36.4)	4 (50)
35–40	35 (14.2)	2 (18.1)	1 (12.5)
40–45	36 (14.6)	3 (27.3)	1 (12.5)

DISCUSSION

In this examination of a national administrative claims database to evaluate hamate fractures and complications following nonoperative management, the data showed that there were very low rates of nonunion, ulnar neuropathy, and tendon rupture.

In the current literature, the decision to pursue operative treatment for hamate fractures remains controversial. Several studies report better outcomes with fewer complications with operative management. The recommendation for surgery can stem from decreasing risk of tendon rupture. In a case study with review of the literature regarding hamate body fractures, Cano Gala et al state that surgical treatment in the various modalities has been shown to have better clinical and functional results.[11] Athanasiou et al, in another study evaluating hamate body fractures, also state that open reduction and internal fixation is preferred over conservative treatment or closed reduction with percutaneous pinning, especially for delayed cases and large displaced fragments of the hamate.[12] However, these injuries had associated dorsal dislocation of the metacarpal bones. Eder et al do not recommend conservative treatment due to high complication rates such as tendon subluxation and absence of pain relief.[13] They state that open reduction and internal fixation is advantageous due to anatomical repositioning of the fracture, physiological restoration of articular surfaces, and remedying of co-injuries. Furthermore, regarding hook of hamate fractures, multiple studies report good outcomes and safety with surgical excision.[9, 14, 15] In athletes, operative management allows pain relief and return to play in a shorter time than with conservative treatment. Although, Bansal et al found a higher incidence of transient ulnar nerve dysfunction with surgical excision of the hook of hamate than previously reported in the literature.[9]

A large concern with conservative treatment of hamate fractures is the risk of nonunion, which may be as high as 50%.[6] Kadar et al, in a retrospective study including 51 patients who sustained hook of hamate fractures, found a nonunion occurrence of 24% in patients with nonoperative treatment, whereas it did not occur in those that were treated operatively.[16] However, in an investigation of eight patients with acute and subacute hamate hook fractures, Whalen et al[17] found that seven showed documented healing of their fractures and were asymptomatic. The one patient who developed painful nonunion was not compliant with treatment. No clear treatment algorithm exists in the current literature. Recommendations have been based on expert opinion, case reports, or small case series with no statistical significance. Conservative treatment measures for hamate fractures typically involve applying a volar forearm splint or short arm cast for 6–8 weeks.

This large population study shows low complication risk in nonoperative management of hamate fractures. Within the first year after sustaining a hamate fracture, complications (including nonunion, tendon rupture, and ulnar neuropathy) were rare. This study serves to provide a broad view of the complication rates for these rare injuries and may help guide the decision for treatment for these patients. The rarity of severe complications, such as tendon rupture, suggests that the decision to pursue surgical intervention can be more dependent on variables such as time to return to work or sport rather than concerns for avoiding the above complications. However, as is common with database studies, important details are unknown that may have influenced treatment and subsequent complications, such as the inherent stability of the fracture, time to treatment, or the presence of CMC subluxation. The presence of pain or the effect on sports activity was also not evaluated in this study.

Due to low patient numbers that were managed operatively, it was not possible to perform a valid analysis of those who had surgical intervention for their hamate fracture. A larger, multicenter study should be considered to produce reliable data regarding operative treatment for hamate fractures.

Based on the demographic information found in this study, the data suggest that age, gender, and BMI did not play a significant role in the development of complications. Men were the predominant gender in all cohorts, which aligns with previous studies that show hamate fractures are most common with men and male athletes in particular.[5]

LIMITATIONS

There are several limitations to the present study. A significant limitation inherent to large administrative database studies is the lack of noteworthy patient details that would be potentially pertinent to the study, such as severity of injury, treatment protocol, techniques, time to union, or functional results. Furthermore, the validity of the study and results are limited by the accuracy of procedural coding within the database. Miscoding and noncoding by providers is a potential source of error. Additionally, the database relies largely on the Humana network and may not represent a true cross-sectional representation of patients sustaining hamate fractures in the United States, with variable results depending on the administrative database used.[18] There is a limitation on the generalizability of the results due to Humana data being heavily favored towards Southern and Midwestern regions. The study did not analyze other important demographic and epidemiologic factors such as time to care, duration of symptoms, sports status, comorbidities, co-injuries, or mechanism of injury. There was no method to reliably distinguish patients that sustained hamate hook versus body fractures in the patients found by ICD-9 codes, which may have an influence on the rates of complications in the separate cohorts. However, our ICD-10 data demonstrated similarly low rates of complications in the subanalysis. Due to the lack of a filter for body mass index in the database and having to rely on ICD codes, there are missing data points regarding BMI in our study population and therefore may not represent the true BMI distribution among these patients. Furthermore, selection bias may also be present as patients enter and exit the database. These limitations are inherent to almost all retrospective database studies.

CONCLUSIONS

This study utilizing a national database illustrates the epidemiology and complication rates for nonoperatively-treated hamate fractures. Nonoperative management of hamate fractures demonstrated low and acceptable complication rates of nonunion, ulnar neuropathy, and tendon rupture. Male patients made up the majority of all cohorts. The study was limited by factors inherent to national database studies, including a lack of specific patient care details, overrepresentation of particular regions in the country, and potential errors with procedural coding. The rare occurrence of these injuries, with most studies comprising of case reports and small studies, necessitates expansion of our knowledge base. The study finds that patients with hamate fractures can be safely treated nonoperatively due to the low risk of major complications and that treatment decisions may be more dependent on variables such as time to return to work or sport. Further studies are necessary to identify optimal indications for operative intervention.