SGEM#385: If the Bones are Good, the Rest Don’t Matter – Operative vs Non-Operative Management of Scaphoid Fractures

Date: November 30th, 2022 Reference: Johnson et al. One-year outcome of surgery compared with immobilization in a cast for adults with an undisplaced or minimally displaced scaphoid fracture: A meta-analysis of randomized controlled trials. Bone Joint J 2022 Guest Skeptic: Dr.Matt Schmitz is an Orthopaedic Surgeon specializing in Adolescent Sports Medicine and Young Adult Hip Preservation. DISCLAIMER: THE VIEWS AND OPINIONS OF THIS BLOG AND PODCAST DO NOT REPRESENT THE UNITED STATES GOVERNMENT OR THE US MILITARY. Case: A 32-year-old male construction worker presents to the emergency department (ED) after falling on his right dominant hand. He has swelling in his distal radius, snuffbox tenderness, decreased range of motion and is neurovascularly intact distal to the injury. X-rays demonstrate a minimally displaced midwaist fracture of the scaphoid. He’s got a big job coming up in a couple of months and can’t work with a cast. He asks if surgery would be a better option? Background: Fractures of the scaphoid are the most common carpal fractures presenting to the emergency department (ED). Initial x-rays pick up 17% with only 7% more being identified on follow-up x-rays (1,2). The classic history for a scaphoid fracture is a fall on outstretched hand (FOOSH). Clinicians need to be careful in taking the history because other mechanisms that hyperextend the wrist like a motor vehicle collision while holding the steering wheel can also apply enough force to fracture the scaphoid. Physical examination of patients with a FOOSH injury include palpating for snuff box tenderness. In a systematic review and meta-analysis (SRMA) by Carpenter et al they were only able to find six studies with a total of 170 patients found in the world’s literature looking at snuff box tenderness.  The evidence had a substantial amount of heterogeneity (3). The LR- to rule out a scaphoid fracture was 0.15 for snuffbox tenderness which is moderate evidence. However, it had a very wide 95% confidence interval around the point estimate (95% CI; 0.05 to 0.43).  There are many other physical exam maneuvers like thumb compression, vibration pain, clamp sign, ulnar deviation pain, radial deviation pain, scaphoid tubercle pain, and resisted supination pronation. None of these have a LR- low enough (<0.1) to reliably rule out a scaphoid fracture. We mentioned x-rays were unreliable as well to rule-out a scaphoid injury. Other imaging modalities like bone scan, ultrasound and CT scan have been used but found to be lacking in accuracy. The best imaging test is an MRI. Initial X-ray 0.24 (0.07–0.79) Follow-up X-ray 0.67 (0.50–0.89) Bond Scan 0.11 (0.05–0.23) Ultrasound 0.27 (0.13–0.56) CT Scan 0.23 (0.16–0.34) MRI 0.09 (0.04–0.19) Emergency physicians can use clinical decision instruments to help in diagnosing certain conditions. There are many validated instruments for fractures such as the Ottawa Ankle Rule (SGEM#3), Ottawa Knee Rule (SGEM#5) and the Canadian C-Spine Rules (SGEM#232). There is no validated clinical decision instrument to help ED physicians accurately rule in or out a scaphoid fracture (4,5). There is not a diagnostic dilemma in this case. The question is does the scaphoid fracture need to be treated operatively or non-operatively. The vast majority (90%) of scaphoid fractures are non-displaced and treated with cast immobilization (6). Displaced fractures increase the risk of non-union from 14% to 50% (7,8,9). If left with a non-union, they almost always result in secondary osteoarthritis of the wrist (10). Also, delayed unions and nonunions are more difficult to treat (i.e. bigger surgery) so there is a trend in orthopedics to perform urgent surgical fixation of scaphoid fractures as opposed to the traditional casting. Whether someone undergoes surgery is an informed decision made between the patient and the surgeon. However, emergency department patients often ask the EM physician if they need surgery. It is good to stay up on the literature so we can prepare the patient for the conversation with the surgeon. We have seen this recently with the non-operative treatment of acute appendicitis (NOTA) and covered this on the SGEM (SGEM#115, SGEM#256, and SGEM#345) Clinical Question: What is the effectiveness of operative vs nonoperative management of un-displaced and minimally displaced (≤ 2 mm) scaphoid fractures? Reference: Johnson et al. One-year outcome of surgery compared with immobilization in a cast for adults with an undisplaced or minimally displaced scaphoid fracture: A meta-analysis of randomized controlled trials. Bone Joint J 2022 Population: Adult patients older than 16 years of age with a un-displaced or minimally displaced (≤ 2 mm) fracture of the waist of the scaphoid. Exclusions: Non-RCTs, children, displaced >2 mm Intervention: Operative management Comparison: Non-operative management Outcome: Primary Outcome: Patient-reported outcome measure (PROM) of wrist function at 12 months Secondary Outcomes: Pain, grip strength, range of motion (ROM) of the wrist, and complications including radiological evidence of nonunion Type of Study: Systematic review meta-analysis of RCTs Authors’ Conclusions: “We found no difference in functional outcome at 12 months for fractures of the waist of the scaphoid with ≤ 2 mm displacement treated operatively or nonoperatively. The complication rate was higher with operative treatment.” Quality Checklist for Therapeutic Systematic Reviews: The clinical question is sensible and answerable. Yes The search for studies was detailed and exhaustive. Yes The primary studies were of high methodological quality. The assessment of studies were reproducible. Yes The outcomes were clinically relevant. Unsure There was low statistical heterogeneity for the primary outcomes. No The treatment effect was large enough and precise enough to be clinically significant. Unsure Results: Using the PRISMA guidelines and searching multiple databases they identified a total of 456 studies. There were seven trials ultimately included in the meta-analysis. Only four RCTs (n= 537) reported functional outcome at 12 months (primary outcome). The demographics of those four RCTs had a mean age of 32 years and 84% were male. Key Result: No statistical difference between operative and non-operative management. Primary Outcome: Patient-reported outcome measure (PROM) of wrist function at 12 months Non-statistical difference using Hedges’ g (0.15 [95% CI; -0.02 to 0.32]; p = 0.082) Secondary Outcomes: There were some statistical differences favoring operative management using fixed effect meta-analysis at 6 months for grip strength, ROM, and odds ratio for non-union. Complications were higher in the operative group. 1. Small Number of Studies: This was mentioned earlier. While they included seven RCTs only four were used for the meta-analysis of the primary outcome with n=537. 2. Uncertainty: With only a few studies to meta-analyze and different outcome measures there were wide 95% confidence intervals around the point estimate of the observed effect size. Four of the seven RCTs were rated as low-quality using the Cochrane Risk of Bias Tool V2.0. All these issues lead to a fair amount of uncertainty in the magnitude and precision of the primary and secondary outcomes. 3. Primary Outcome Measure (PROM): The primary outcome for each of the four RCTs used in the meta-analysis was different. This makes it hard to compare one study to another. It also contributes to the high I2 test as a measurement of heterogeneity for the primary outcome of PROM. It was 74% using the fixed effect model worse using the random effect mode (84%). Disabilities of the Arm, Shoulder and Hand questionnaire (DASH) Patient Evaluation Measure (PEM) Patient-Rated Wrist Evaluation (PRWE) function subscale, Adapted Green O’Brien score In addition, is function and PROM at 12 months the purpose of operative treatment? Or is it getting people back to work quicker?  Treatment for scaphoid wrist fractures has typically been long immobilization in a thumb spica cast which is very cumbersome for a manual labourer.  So, it depends on what your patient would value and prefer. Grip strength and ROM favoured operative treatment at six months. This was a secondary outcome and considered hypothesis generating that could be explored further. They don’t mention earlier healing time in operatively treated fractures. There is a difference between rate of nonunion and time to union. Let’s not forget about complications. Some were related to “scar” 4cm incision that is not used any longer; others were arthroscopic assisted reductions 4. Hedges’ g: A Hedges’ g is a measurement of effect size. It was first described by Larry Hedges in 1981 (JSTOR). Typically, it is used to determine how much an group (experimental) is different from another group (control). It is very similar to Cohen’s d but is better when sample sizes are smaller (<20). Small effect size is considered 0.2, medium effect size 0.5, and a large effect size >0.8. The Hedges’ g for the primary outcome in this study was 0.15 (95% CI; -0.02 to 0.32); p = 0.082 representing a small effect size that was not statistically different between operative and not-operative group. 5. Bond Article: They seem to ignore the 2001 article by Bond et al that looked at a military population. This study showed a quicker return to full duty and quicker healing times. Comment on Authors’ Conclusion Compared to SGEM Conclusion: A friendly amendment to their conclusions would be that they did not find a “statistically difference” between the two treatments rather than saying “no difference”.