Patient-Related Factors Associated With Scaphoid Proximal Pole Avascular Necrosis Healing

Abstract

Background:

Avascular necrosis (AVN) of the proximal pole is a well-known complication of scaphoid fractures. Avascular necrosis is poorly understood, including the transition from ischemia to necrosis, optimal treatment, and why some AVN heals but others do not. The primary purpose of this study is to evaluate patient-related factors that are associated with healing outcomes in individuals with proximal pole AVN following scaphoid fractures.

Methods:

This is a retrospective review of all patients diagnosed with scaphoid proximal pole AVN secondary to a fracture from 2018 to 2024 in a single center. Patient baseline characteristics and comorbidities at time of diagnosis were collected. If the patient underwent surgical management, procedural factors were collected. The primary outcome was AVN healing after 4 months of follow-up.

Results:

A total of 62 patients met inclusion criteria. Thirty of 62 (48.4%) went onto proximal pole AVN resolution. Hyperlipidemia (P = .030), advanced age at time of diagnosis (P = .038), and elevated body mass index (BMI) (P = .026) were independent factors associated with lack of AVN healing. For patients who underwent surgical management, there was no significant difference in healing outcomes between use of a nonvascularized, or no graft, and use of a vascularized bone flap (P = .115, P = .886, respectively).

Conclusions:

Hyperlipidemia, elevated BMI, and advanced age are patient factors negatively associated with scaphoid proximal pole AVN healing—key information for accurately assessing patient prognosis. For surgical management, the choice of a vascularized bone, nonvascularized bone graft, or no graft does not significantly impact AVN healing.

Introduction

The scaphoid’s distinctive blood supply in combination with its irregular, twisted, and tubular shape has been extensively studied and carries important implications for the nature of scaphoid fractures and healing potential.^1-3 As the scaphoid’s surface is largely articular, its vascular supply is limited. ¹ The dominant blood supply comes from the dorsal carpal branch of the radial artery, which enters through the dorsal ridge in a retrograde manner.^2,3 This anatomy places the proximal pole at high risk of compromised blood supply following a scaphoid fracture, which often leads to nonunion and avascular necrosis (AVN).^4,5

Avascular necrosis is a well-documented complication of scaphoid fractures, with reported incidence rates of up to 50%.^3,6,7 If not healed, AVN ultimately progresses to carpal collapse and degenerative arthritis significantly affecting patients’ quality of life and increasing their financial burden from chronic pain treatment and arthritis management.^3,7 Despite being a common and challenging complication of scaphoid fractures, AVN is still poorly understood, including the transition from ischemia to permanent necrosis, and optimal treatment strategies. Particularly, we have limited understanding of why some AVN heals while others do not and the patient-specific variables that can influence healing.

Scaphoid nonunion healing has been widely studied, with factors such as AVN known to reduce fracture healing rates. ⁸ However, the variables associated with AVN resolution remain unknown. As the sequelae of scaphoid AVN are so taxing, understanding the factors associated with AVN healing could be beneficial in guiding prognosis and patient counseling. In particular, identifying these demographic factors may help clinicians individualize management, set accurate long-term expectations for each patient, and guide whether treatment should focus on primary scaphoid healing or arthritic symptom prevention. The primary purpose of this study is to evaluate patient-related factors that are related to healing outcomes in individuals with proximal pole AVN following a scaphoid fractures.

Materials and Methods

Study Population

From a collected registry of patients with hand pathology, we retrospectively reviewed all patients diagnosed with a scaphoid fracture from January 2018 to November 2024 at a single center. Approval from the NYU Langone Health institutional review board (IRB i23-00184) was obtained before conducting any study activities. Patients were screened using a combination of International Classification of Diseases, Tenth Revision (ICD-10) and Current Procedural Terminology (CPT) codes. A manual chart review was conducted for all screened patients to determine whether any had been diagnosed with proximal pole AVN through histology, intraoperatively by lack of punctate bleeding, or imaging techniques—computed tomography (CT), magnetic resonance imaging, and radiography (x-ray). Inclusion criteria included patients aged 18 years or older who were diagnosed with scaphoid proximal pole AVN treated at our institution with a minimum follow-up time of 4 months. Four months was determined to be a suitable amount of time, as it is generally accepted that after 3 to 4 months without progression toward healing, the condition is no longer considered a delayed union but a chronic nonunion.^2,9,10 All patients meeting inclusion criteria were included in the study.

Independent and Dependent Variables

Baseline characteristics, including sex, age, race, body mass index (BMI) closest to presentation, smoking status, number of alcoholic drinks per week, hand dominance, and medical comorbidities, were obtained from the medical record. Furthermore, the scaphoid fracture location, AVN diagnosis modality, date of presentation, follow-up time, and the date of first mention of AVN and AVN healing were determined by x-ray and/or CT. If the patient underwent surgical management for acute scaphoid fracture fixation, surgery date, surgical technique, and tourniquet time were collected. Fracture pattern and characteristics were similar between patients who underwent surgical management and those who were treated nonoperatively. Patients were stratified by presence or absence of AVN healing as noted on their last follow-up encounter.

Statistical Analysis

Patient baseline characteristics were measured descriptively. Continuous and categorical variables were compared by independent sample t-tests and χ² or Fisher exact tests, respectively. For this analysis, patients were stratified into 2 cohorts—patients with AVN that healed and patients with chronic AVN that did not heal. A multivariable binary logistic regression was also used to assess the dependent variable of AVN healing. This regression included all patient characteristics, including age, BMI, sex, medical comorbidities, alcohol or tobacco use, and fracture location. All statistical analyses were performed using SPSS (IBM SPSS Statistics for Mac, Version 29.0. Armonk, New York: IBM Corp; 2022). Statistical significance was set at P < .05.

Results

Initial query of the institution database identified 1940 patients who were diagnosed with a scaphoid fracture. Of these, 140 were diagnosed with proximal pole AVN. These patients were chart reviewed which identified a total of 62 patients who had follow-up greater than 4 months and were included in this study. Thirty of 62 (48.4%) went onto healing of proximal pole AVN following a scaphoid fracture (Table 1) as determined by x-ray and/or CT. The average follow-up time to determine whether healing occurred was 26.93 ± 21.06 months. For the 30 patients who ultimately healed and achieved union, the average time to union was 35.58 ± 45.53 weeks.

Table 1.

Demographics and Surgical Factors of Patients Diagnosed With Scaphoid Proximal Pole Avascular Necrosis.

Demographics/surgical factors	AVN healed	Chronic AVN	Total	P value a
No. of cases	30	32	62	—
Age (mean ± SD)	29.00 ± 11.24	40.59 ± 16.61	34.98 ± 15.31	.002
BMI (mean ± SD)	25.35 ± 3.87	27.9 ± 4.66	26.66 ± 4.45	.022
Sex				.092
Male	26 (86.7%)	22 (68.8%)	48 (77.4%)
Female	4 (13.3%)	10 (31.2%)	14 (22.6%)
Race				.310
White	22 (73.3%)	18 (56.3%)	40 (64.5%)
Black	4 (13.3%)	5 (15.6%)	9 (14.5%)
Asian	2 (6.7%)	4 (12.5%)	6 (9.7%)
Hispanic or Latino	2 (6.7%)	5 (15.6%)	7 (11.3%)
Smoking status				.662
Never smoker	20 (66.7%)	18 (56.3%)	38 (61.3%)
Former smoker	5 (16.7%)	8 (25.0%)	13 (21.0%)
Current smoker	5 (16.7%)	6 (18.8%)	11 (17.7%)
Alcohol use	21 (70.0%)	21 (65.6%)	42 (67.7%)	.713
HTN	3 (10.0%)	10 (31.3%)	13 (21.0%)	.040
HLD	3 (10.0%)	11 (34.4%)	14 (22.6%)	.022
Diabetes mellitus	0 (0.0%)	1 (3.1%)	1 (1.6%)	.329
CKD	0 (0.0%)	2 (6.3%)	2 (3.2%)	.164
Laterality				.779
Right	13 (43.3%)	15 (46.9%)	28 (45.2%)
Left	17 (53.1%)	17 (56.7%)	34 (54.8%)
Fracture on dominant hand	15 (50.0%)	16 (50.0%)	31 (50.0%)	1.00
Fracture location				.581
Proximal pole	9 (30.0%)	10 (31.2%)	19 (30.6%)
Waist	20 (66.7%)	22 (68.8%)	42 (67.8%)
Distal pole	1 (3.3%)	0 (0.0%)	1 (1.6%)
Surgical intervention	25 (83.3%)	19 (59.4%)	44 (71.0%)	.038
Days between presentation and surgery (mean ± SD)	65.62 ± 79.07	92.30 ± 137.66	77.47 ± 108.51	.418
Tourniquet time during surgery, min (mean ± SD)	105.89 ± 28.48	130.94 ± 41.08	117.34 ± 36.53	.041

Abbreviations: AVN = avascular necrosis; BMI = body mass index; HTN = hypertension; HLD = hyperlipidemia; CKD = chronic kidney disease.

^aCategorical variables assessed with either Fisher exact or χ² test based on sample size; continuous variables analyzed with independent samples t-test. P values < .05 were considered significant and are in bold.

Univariate analysis showed no significant difference in patient factors, including sex, race, scaphoid fracture location (proximal pole, waist, or distal pole), number of alcoholic drinks a week, and smoking status between those with healed AVN and those with chronic unhealed AVN (Table 1). However, a diagnosis of hypertension (HTN) or hyperlipidemia (HLD) was significantly associated with chronic AVN (P = .040, P = .022, respectively). Furthermore, patients who failed to heal had an average age of 40.59 years, compared with patients whose AVN healed had an average age of 29.00 years (P = .002). Similarly, patients who failed to go onto AVN healing had a greater average BMI (27.9 kg/m²) compared with the healed group (25.25 kg/m²) (P = .022).

A multivariable binary logistic regression confirmed the relationship between AVN healing and these patient-specific factors (Table 2). Increased age and increased BMI were related to a decreased odds of AVN healing (P = .038, P = .026, respectively). Patients with a diagnosis of HLD were found to have 85% lower odds to experience AVN healing (P = .030). However, HTN was not significantly associated with AVN healing (P = .060) when put into the multivariable logistic regression.

Table 2.

Multivariable Regression Evaluating the Impact of Patient Factors on Scaphoid Proximal Pole Avascular Necrosis Healing.

Demographics/surgical factors	Odds ratio	95% confidence interval	P value a
Age	0.95	0.90-0.99	.038
BMI	0.84	0.72-0.98	.026
HTN	0.25	0.056-1.16	.060
HLD	0.15	0.028-0.83	.030
Tourniquet time during surgery, min	0.98	0.95-1.00	.062

Abbreviations: BMI = body mass index; HTN = hypertension; HLD = hyperlipidemia.

^aP values < .05 were considered significant and are in bold.

A separate analysis was conducted on the 44 patients who underwent initial operative treatment for their scaphoid fractures. All patients in this cohort had similar scaphoid fracture patterns and characteristics. For patients who underwent open reduction internal fixation (ORIF) of scaphoid fracture with proximal pole AVN, time between initial presentation and surgery was not significantly associated with healing; however, tourniquet time was associated with AVN healing. Proximal pole AVN that did not heal had 25.05 more minutes of tourniquet time compared with AVN that healed (P = .041). However, on further investigation with a multivariable logistic regression, longer tourniquet time did not have any significant relationship with AVN healing (P = .062).

In addition, different scaphoid fracture fixation techniques were used and thus evaluated. Nine (20.5%) patients underwent ORIF without any bone grafting, whereas 35 (79.9%) patients underwent bone grafting (Table 3). Compared with conservative management, surgical intervention was associated with AVN healing (P = .038). Further classifying surgical intervention, 14 patients had vascularized bone flaps, whereas 21 patients had a nonvascularized bone graft. Compared with using vascularized bone, using a nonvascularized graft or no graft at all showed no significant difference in AVN healing (P = .886, P = .155, respectively) (Table 4).

Table 3.

Surgical Intervention Technique Used to Treat Proximal Pole Scaphoid Avascular Necrosis.

Surgical technique	AVN healed	Chronic AVN	Total	P value a
No surgery	5 (16.7%)	13 (40.6%)	18 (29.0%)	.033
No graft	3 (10.0%)	6 (18.8%)	9 (14.5%)
Graft	22 (73.3%)	13 (40.6%)	35 (56.5%)

Abbreviation: AVN = avascular necrosis.

^aVariables assessed with χ² test. P values < .05 were considered significant and are in bold.

Table 4.

Univariate Regression Evaluating Comparing Surgical Technique’s Impact on Scaphoid Proximal Pole Avascular Necrosis.

Surgical technique	Odds ratio	95% confidence interval	P value a
Vascularized graft (reference)	—	—	—
No surgery	0.21	0.05-0.96	.044
No graft	0.90	0.22-3.68	.886
Nonvascularized graft	0.28	0.03-1.62	.155

^aP values < .05 were considered significant and are in bold.

Discussion

Scaphoid fractures are among the most common hand injuries—accounting for up to 11% of all hand fractures. ¹¹ Hand surgeons frequently see this pathology in clinical practice. Scaphoid nonunion often develops AVN which occurs in up to 100% of proximal pole fractures.^12-14 Although prevalent, there is a paucity of literature on AVN healing. In this study, we sought to understand patient factors associated with scaphoid proximal pole AVN healing.

Our study found that certain comorbidities are significantly correlated with poorer AVN healing outcomes. Namely, patients diagnosed with HLD had 85% lower odds of experiencing AVN healing compared with those without HLD. Vessel calcification compromises blood flow necessary for the scaphoid to heal. ¹⁵ Angiogenesis and quality blood supply are known to be critical factors for healing, and both HTN and HLD impair the ability of vessels to bring oxygen, nutrients, and inflammatory cells to the necrotic site to initiate healing.^16,17 Furthermore, as osteoblasts and adipocytes originate from the same mesenchymal stem cell, previous studies have found that excessive blood lipid influences these stem cells to differentiate toward adipogenesis instead of osteogenesis, causing poor bone regeneration.^15,18,19 A similar process is seen with HTN. Hypertension’s association with AVN healing approached but did not reach significance. However, HTN has previously been linked to osteoporosis and fractures, although the mechanism driving this is poorly understood.^20,21 Yet, the proposed mechanism can also be used to understand why it may be associated with poor AVN outcomes. Hypertension drives oxidative stress which in turn affects bone metabolism and causes more bone resorption than regeneration. ²¹ Although there have not been dedicated studies understanding this pathophysiology effect on scaphoid AVN, we theorize that these mechanisms are contributors to poor AVN healing outcomes.

Furthermore, we found elevated BMI to be an independent factor associated with chronic AVN. Obesity has been studied previously regarding fracture healing with mixed results. A review article concluded that obesity could be a risk factor for nonunion in weightbearing bones, however, not in upper extremity bones. ²² At the cellular level, studies have shown obesity impairs revascularization.^18,23 As the scaphoid proximal pole receives approximately 80% of its blood supply primarily from a single branch of the radial artery, we believe that obesity-related impairments in revascularization could have a disproportionately greater impact on its healing compared with other bones in the upper extremity. ¹²

Advanced age is well documented to lower the potential of a fracture to heal. ¹⁶ Although there are no studies focused on AVN healing, some studies are dedicated to factors associated with AVN development. Schwitzer et al ²⁴ and Zi-Sheng Ai et al ²⁵ both studied factors associated with the development of AVN of the femoral head secondary to femoral neck fracture. Both studies found that age was independently associated with AVN. In addition, Patterson et al ²⁶ found that advanced age decreases the odds of scaphoid nonunion healing. Our study agrees with and adds to the general body of knowledge that age and bone healing are associated.

Smoking is recognized to inhibit angiogenesis and delay the fracture healing process. ¹⁶ Rachunek-Medved et al ²⁷ studied active smoking’s impact on scaphoid nonunion healing and discovered that smokers were approximately 77.4% less likely to achieve scaphoid union post fracture compared with nonsmokers. Furthermore, Konstantinidis et al ²⁸ found that smoking had a negative impact on of scaphoid nonunion healing complicated by proximal pole AVN. Animal models have shown nicotine inhibits osteogenesis and angiogenesis impacting bone healing. ²⁹ We hypothesized that smoking would be associated with poor AVN healing outcomes, in agreement with prior studies. However, we found no association between smoking status and proximal pole AVN healing. We theorize that the nonsignificance in our study is attributed to the limited sample size in the smoking group. With similar reasoning, we cannot comment on diabetes mellitus and chronic kidney disease and their association with AVN healing as the sample size is too small. With a larger cohort, a true effect can be more accurately assessed in a future study.

In a separate analysis of patients who underwent surgical intervention, tourniquet time was found to be significant in a t-test but did not retain significance in regression. As the tourniquet time included any additional procedures that were performed, there was substantial variability in tourniquet time which may have prevented the regression model to detect significance. Our study found that surgical fixation was significantly associated with AVN healing; however, there were no significant healing differences between graft use nor graft types. However, bone graft during these cases was not taken from a consistent donor site. In addition, the use of vascularized, nonvascularized, or no bone graft was surgeon preference. With a small cohort, the analysis may be underpowered to detect a true difference. Therefore, further investigation with a larger sample size of patients undergoing ORIF with no other interventions and consistent graft donor sites may help clarify the impact of tourniquet time and other surgical technique on AVN outcomes.

There are several limitations that should be acknowledged. This is a retrospective study and thus carries the risk of bias innate to retrospective reviews. Furthermore, the study has a limited sample size based on the nature of scaphoid pathology. Scaphoid fractures have a false-negative rate from 2% to 20% on radiographs and symptoms often are minimal or vague wrist pain.^11,30,31 Therefore, scaphoid fractures and proximal pole AVN, though prevalent, are often missed. It is our intention to continue following patient outcomes and include a greater sample size in future studies. Last, patients were treated by different health care providers, which introduce treatment variability. However, our primary outcome was binary and all imaging was read by trained musculoskeletal radiologists and reviewed by the authors, contributing to the consistency and reliability of this study. Future studies may continue to investigate operative factors and surgical techniques that may influence scaphoid AVN healing to continue to grow this body of literature.

Conclusion

In conclusion, our findings suggest that HTN, HLD, elevated BMI, and advanced age are all patient factors negatively associated with scaphoid proximal pole AVN healing. This suggests that patient factors impacting blood vessel integrity and osteogenesis could influence bony healing. The long-term consequences of proximal pole AVN can lead to decreased range of motion, diminished grip strength, and wrist pain, diminishing patient quality of life.^32,33 Understanding patient factors that are associated with scaphoid proximal pole AVN healing is critical for prognostication and patient counseling. We believe our study provides initial understanding of factors influencing scaphoid AVN healing, laying a foundation for further contributions to this field.