Ultrasound-Guided Manipulation does not Prevent Malalignment Over Landmark-Based Fracture Reduction in Distal Radius Fracture (Colles)

Abstract

Introduction:

This systematic review aims to determine the relative risk of distal radius (Colles) fracture (DRF) malalignment between ultrasound (USG)-guided and conventional/landmark guided/blind manipulation and reduction (M&R).

Methods:

We searched 3932 records from major electronic bibliographic databases on USG-guided manipulation of DRF. Studies with randomized, quasi-randomized, and cross-sectional study designs meeting the inclusion criteria were included in this review. USG and landmark-guided DRF manipulations were named cases and controls, respectively. The Newcastle–Ottawa Scale was used to assess the quality of included studies.

Results:

Thirteen and nine studies were analysed for qualitative and quantitative analysis in this review. Nine hundred fifty-one DRF patients (475 cases and 476 controls) from 9 studies with mean ages of 51.52 ± 11.86 (22–92) and 55.82 ± 11.28 (18–98) years for cases and controls were pooled for this review. The pooled relative risk estimate from the studies included in the meta-analysis was 0.90 (0.74–1.09). There was a 10% decrease in the risk of malalignment with USG than the landmark guided M&R of DRF. The I² statistic estimated a heterogeneity of 83%. Sensitivity analysis revealed a relative risk of 1.00 (0.96–1.05).

Conclusion:

The USG-guided manipulation does not prevent malalignment over the landmark-based manipulation of DRF. The risk of bias across the included studies and heterogeneity of 83% mandates further unbiased, high-quality studies to verify the findings of this review.

INTRODUCTION

Distal radius fractures (DRF) account for one-sixth of fractures in accident and emergency.[1] Manipulation and reduction (M&R) followed by splint application is the most acceptable treatment for extra-articular and minimally displaced intra-articular DRF. Nevertheless, the high risk of failure of M&R or a need for operative intervention is the cause of concern in these injuries. Recently, several studies have investigated the role of ultrasound (USG) in guiding M&R in DRF.[1] While the landmark-based technique is an established method for DRF M&R, USG offers a real-time assessment of fracture alignment. The claimed benefits of USG for DRF M&R include increased reduction accuracy and a decrease in emergency room (ER) waiting time; however, it is associated with increased costs and training.[2] The utility of both the portable USG machines as the point of care and standard units is described for DRF M&R.[3,4]

Malik et al. reviewed the accuracy of USG-directed M&R in DRF.[1] However, their review was limited to a narrative synthesis of the included studies.[1] It reviewed two studies with mixed populations of children and adults.[5,6] One study in their review reported 66 USG observations among 23 DRF participants.[7] The conclusions in their review were drawn from studies with a risk of reporting bias (selective outcome reporting and incomplete outcome data) favoring USG-guided M&R. Since their review; we identified one randomized controlled trial (RCT), two cross-sectional studies, and a retrospective cohort study that studied the effects of USG-guided M&R in DRF.[8-11] Therefore, we reviewed the comparative efficacy of USG and landmark-guided M&R in preventing malalignment in DRF.

The primary objective of this review was to determine the relative risk of mal-reduction between USG and landmark-guided M&R in DRF.

METHODS

The procedures followed in this review were per the ethical standards of the responsible committee on human experimentation (institutional and national). We conducted this review following guidelines in The Preferred Reporting Items for the Systematic Review and Meta-Analysis (PRISMA).

Eligibility criteria (inclusion and exclusion criteria)

We included studies published in English that compared USG-guided and landmark-based M&R for DRF in adults. Noncomparative cross-sectional studies analyzing pre and post-M&R X-ray parameters in USG-guided M&R were included in this review. Case series and case reports were excluded from this review. The studies which reported the sensitivity and specificity of USG for the detection of DRF were excluded; however, mixed studies describing the accuracy of reduction with the former were included. The definitions of various datasets in this review are described below:

1. USG-guided M&R was defined as the use of a portable/point of care (POCUS) or standard USG machine to visualize fracture ends during the manipulation of DRF

2. Landmark-based manipulation was defined as a method in which tactile/palpatory feedback of distal radius was used to align the DRF

3. Cases: DRF where USG-guided M&R was done were defined as cases for this review

4. Controls: DRF, where landmark-based M&R was done, were defined as controls for this review

5. Success and failures in cases and controls were defined based on the X-ray criteria or the investigator’s report from the studies included in this review

6. However, we defined an endpoint of one attempt to M&R for DRF. Patients requiring second manipulation or operative fixation were considered failures. Patients needing operative intervention after the first attempt of M&R were also considered failures. Within the one attempt of the M&R procedure, the USG view could be repeated as often as necessary until an acceptable alignment was obtained. Proper alignment is viewed by USG as aligning the proximal and distal bone cortex into as straight a line as possible on the number of views that the study investigators deemed necessary

7. For this review, a trained doctor for USG was defined as one who adequately completed the minimum training sessions defined by the study investigators before participating

8. Cases study was a single-arm study where USG-guided M&R of DRF was executed in real time, and later, the accuracy of fracture alignment was confirmed with X-rays

9. The control study was a single-arm study where landmark-guided M&R of DRF was executed, and the accuracy of fracture alignment was later confirmed with USG and X-ray.

Information sources

Major electronic databases up to August of 2022 were searched to identify studies in the English language on USG for M&R of DRF.

Search strategy

Annexure 1 describes this review’s search strategy for the electronic databases. We hand-searched the references from the included studies on USG for DRF. We excluded reports from trial registries, conference proceedings, books, and dissertations.

Selection process

Review authors SN and JA independently screened all titles and abstracts for potentially eligible studies. They obtained full-text reports where appropriate. The same two review authors independently performed the study selection. We resolved disagreements about the inclusion or exclusion of individual studies and data extraction by discussion between the two authors or with a third author (PR) for the final study selection to ensure a consensus. We did not contact the authors of included studies. The data collected included study design, population, interventions, outcome measures, and results. We did not mask the source and authorship of the trial reports.

Data items (outcome measures)

The following characteristics from the included reports were recorded: (1) success in cases and controls, (2) failures in cases and controls, and (3) demographic outcomes (age, gender, duration of fracture, and classification).

Assessment of risk of bias in included studies

Pairs of the same review authors performed independent ‘Risk of bias assessment of the included trials. In this review, we used the Newcastle–Ottawa scale for observational studies for risk of bias assessment.[12] The Newcastle–Ottawa scale is designed to assess four, one, and three numbered items under the selection, comparability, and outcome domains. The maximum number of stars awarded for a study under each domain can be four, one, and three.

Effect measures

Meta-analysis of the included studies was done using Review Manager 5.4. The random effects model pooled success and failure across included studies as a risk ratio. The estimation of heterogeneity and inconsistency across included studies was done using the Chi-square test and I² test. The pooled estimate of the relative risk of failure of M&R was graphically represented with a forest plot. A sensitivity analysis was done to find the pooled estimate after eliminating studies favoring the USG-guided M&R. Review author (P) performed the statistical assessments.

RESULTS

Study selection

We searched 3932 records up to August of 2022 for studies on USG-mediated M&R of DRF from electronic databases Cochrane Central Register of Controlled Trials (CENTRAL) (8), CINAHL (24), DOAJ (142), Embase (183), Medline (21), Proquest (96), PubMed (190), and Scopus (3268). We isolated 19 studies meeting the inclusion criteria from electronic databases[2-11,13-21] [Table 1]. We could not extract more studies from a hand search of references from the isolated reports. We analyzed 13 studies for qualitative analysis[2-4,7-11,13-17] [Table 1]. Six studies were excluded from the qualitative and quantitative analysis.[5,6,18-21] We analyzed nine studies for meta-analysis.[2-4,8-10,13,15,16] Four further studies were excluded from the meta-analysis.[7,11,14,17] We have described the reasons for excluding studies from qualitative and quantitative analysis as shown in Table 1. The PRISMA flow diagram for this review is presented in Figure 1.

Table 1.

Characteristics of isolated studies from search of electronic databases

Study	Study design/group I or II	Number of arms	Sample size	Included in qualitative analysis	Included in meta- analysis	Success/failure cases*, success/failure controls#	Comments
Ammann et al.[9]	Retrospective cohort/I	2	142	+	+	46/25, 75/67	A subgroup analysis of patients >60 years demonstrated lower necessity of operation in USG-guided manipulation and reduction in DRF
Ang et al.[16]	Ambispective bidirectional/I	2	102	+	+	58/4, 76/25	The study compared baseline participant characteristics between groups for type of fracture, anesthesia delivered, and caregivers
Alonso et al.[11]	Cross-sectional/II	1	36	+	−		This study reported sensitivity and specificity of USG for detection of DRF however other parameters for meta-analysis meeting our endpoints of either one manipulation attempt or need for operative intervention were not reported
Bozkurt et al.[15]	Prospective double blind cross-sectional diagnostic accuracy study/II	1	60	+	+	40/20#	This study was a controlled study where the orthopedic residents performed landmark-guided manipulation and reduction. Nevertheless, USG was obtained before and after manipulation and reduction, and the caregivers performing manipulation and reduction were blinded to the findings of USG
Brahm and Turner[14]	RCT/I	2	47	+	−	17/10, 13/7	The RCT was reported as an abstract and compared both the groups on fracture type, commination, caregivers, and reduction time
Chinnock et al.[5]	Multicenter cohort study		90	−	−		This study included both children and adult DRF cases
Chern et al.[6]	Prospective case series		27	−	−		This study included both children and adult DRF cases
Esmailian et al.[13]	Cross-sectional/II	1	154	+	+	146/9*	This study reported the accuracy of fracture alignment as a surrogate to DRF detection by USG. Since the USG guided the fracture reduction it was classified as a case study
Kodama et al.[2]	Ambispective bidirectional/I	2	65	+	+	41/2, 15/7	The study compared fracture type, X-ray parameters before and after manipulation and reduction, costs, time, and success between groups
Kozaci et al.[18]	Prospective observational		83	−	−		This study reported the sensitivity, specificity, positive and negative predictive value of USG in DRF detection. The study included children and adults
Lau et al.[7]	Cross sectional/II	1	23	+	−		The study recruited healthy wrist controls and recorded 66 USG and radiographic observations from 23 DRF
Lau et al.[19]	Cross sectional	1	100	−	−		This study examined the comfort levels of postgraduate residents before and after USG training in the detection of DRF. It also examined the accuracy of fracture reduction; however, USG images were shown to residents instead of DRF cases
Rezashah et al.[10]	Cross sectional study/II	1	30	+	+	25/5*	This study imaged the fracture site before the application of splint by USG and the caregivers performing manipulation and reduction were not blinded to the findings of USG
Sabzghabaei et al.[3]	Case control/I	2	130	+	+	59/6, 49/16	The study compared baseline and postmanipulation and reduction radiographic indices between both the groups
Socransky et al.[17]	Multicenter prospective cohort study/II	1	119	+	−		This study reported success and failures of the accuracy of DRF alignment on Likert Scale unlike other studies included in the meta-analysis
Smiles et al.[8]	RCT/I	2	100	+	+	25/25, 26/24	The study compared mode of anesthesia and caregivers performing manipulation and reduction between groups
Sivrikaya et al.[20]	Cross sectional		90	−	−		This study reported the sensitivity, specificity, positive and negative predictive value of USG in DRF detection
Wood et al.[4]	Cross sectiona/II	1	35	+	+	24/11#	Blinded clinicians to the USG findings performed landmark-based manipulation and reduction in this study
Unkefer et al.[21]	Case report		1	−	−		This was a case report that described the effects of USG guidance on Colles’ fracture reduction

RCT: Randomized controlled trial, USG: Ultrasound, DRF: Distal radius fractures, ^#USG guided Distal radius fracture manipulation, *Landmark guided distal radius fracture manipulation, +: Included, _: Excluded

Figure 1.

The PRISMA flow diagram for included studies. PRISMA: Preferred reporting items for the systematic review and meta-analysis

Study characteristics

We have described study characteristics as shown in Table 1. The characteristics of excluded studies are described in Table 1.

Risk of bias assessment in the included studies

The risk of bias assessment was done on the Newcastle–Ottawa scale and is presented in Table 2. The risk of bias was inversely proportional to the stars awarded to a study.

Table 2.

Risk of bias assessed using the Newcastle–Ottawa Scale

Study*	Selection†	Comparability‡	Exposure§
Amman et al.[9]	⛤⛤	⛤	⛤⛤⛤
Ang et al.[16]	⛤⛤	⛤	⛤⛤⛤
Bozkurt et al.[15]	⛤	-	⛤⛤
Esmailian et al.[13]	⛤	-	⛤⛤
Kodama et al.[2]	⛤⛤	⛤	⛤⛤⛤
Rezashah et al.[10]	⛤	-	⛤⛤
Sabzghabaei et al.[3]	⛤⛤	⛤	⛤⛤⛤
Smiles et al.[8]	⛤⛤	⛤	⛤⛤⛤
Wood et al.[4]	⛤	-	⛤⛤

*Three domains with eight questions were used, ^†Selection had four questions, ^‡Comparability had one question, ^§Exposure had three questions

Selection

All except two studies[8,13] had selection bias for the numbered item of representativeness of the exposed cases and case definition of distal radius fracture. The studies used explicit radiological criteria to define a DRF.[8,13] The participants were not independently validated, and there was subjectivity in recruitment.

Comparability

This domain had numbered items of comparability based on study design or analysis. The design of the studies included in this review is described under the characteristics of included studies. Five studies reported on the domain of comparability between the participants recruited in either of the groups.[2,3,8,9,16] The variables compared between the groups in these studies were age, gender, fractured side, type of fracture, mechanism of injury, the experience of caregivers for fracture manipulation and USG imaging, and radiological parameters before and after M&R. The comparative studies reported no difference between the groups on the described parameters.

Outcome

All studies have numbered a star on ascertainment of exposure included in this review.

The domain of “comparability” did not apply to single-arm studies. The numbered items of selection of controls, the definition of controls, and the same method of ascertainment of exposure for cases and controls under the domains of selection and outcome did not apply to single-arm studies. The numbered item of nonresponse rate under the outcome domain did not apply to all the studies included in this review.

Demographic outcomes

The preset criteria for an acceptable fracture alignment on radiography were reported by eight studies.[2,3,7,10,13-16] The radial height and inclination did not vary as acceptable criteria for fracture alignment; however, studies were heterogeneous for the volar tilt ranging from 10° of dorsiflexion to 20° of volar-flexion. Six studies relied on the orthopedic surgeon/treating physician’s decision for an acceptable fracture alignment;[4,8,9,13,15,16] however, the inter-rater agreement and reliability of reduction success and assessment methods were reported by three studies.[7,10,13]

Patient selection

Four[4,8-10] and six[3,7,13,15-17] studies stratified participants based on USG-trained personnel and convenient sampling availability. Three studies did not describe the sampling method for participant recruitment.[2,11,14] No study described the method of allocation concealment. However, two studies had DRF allocation by the hand surgeons and orthopedic surgeons to the ultrasonography group.[2,7] Seven studies[2,8-10,14-16] (4 emergency physicians, one research assistant, an orthopedic surgeon, and hand surgeons) described the caregivers who recruited patients into the studies, while five did not.[3,7,11,13,17] Three studies described the age of the fracture at the time of presentation, while the others did not.[4,7,16] The choice of anesthesia ranged from hematoma block to procedural sedation. Three studies[7,10,11] used hematoma block exclusively, while the rest were heterogeneous in selecting anesthesia for M&R.

Ultrasound machine and personnel

Five[4,7-8,10,17] and seven studies[2,3,9,11,13,15,16] utilized point-of-care portable USG machines and standard units for imaging the fracture site. The frequency of the USG probe ranged from 5 to 15 MHz for the studies in this review. Seven[3,4,8,13,15,16] and two[2,7] studies were imaged, two (dorsal and radial) and three (dorsal radial and volar) distal radius surfaces to guide the fracture manipulation. Caregivers across included studies received USG training ranging from video demonstrations to certification. We could not draw logical conclusions on the influence of physician experience for achieving better radiological indices between the groups due to extensive heterogeneity in the included studies for the description of caregivers. Emergency physicians and residents executed M&R in eight studies,[3,8-11,13,16,17] while orthopedic surgeons or residents and hand surgeons) execute M&R[2,4,7,15] in four studies. Two studies had mixed caregivers (emergency clinicians and residents).[2,14] There were 4,[4,8,9,14] 2,[15,16] and 2[13,17] studies where orthopedic surgeons, senior emergency physicians, and radiologists assessed primary outcomes. Five studies did not report outcome assessors in this review.[2,3,10,11,14]

RESULTS OF INDIVIDUAL STUDIES

Results of individual studies are presented in Table 1.

Results of syntheses (quantitative analysis [meta-analysis])

We pooled 951 DRF from nine studies.[2-4,8-10,13,15,16] There were 475 cases 476 controls. There were 3 cases and 2 control single-arm studies. The mean age of cases and controls was 51.52 ± 11.86 (22–92) and 55.82 ± 11.28 (18–98) years. There were 202 and 147 male cases and controls. There were 273 and 294 female cases and controls.

We estimated a 10% decrease in the risk of malalignment of DRF M&R with USG relative risk 0.90 (confidence intervals [CI]: 0.74–1.09) [Figure 2]. The effect estimate crossed the null value. There was a substantial heterogeneity of 83% [Figure 2]. A sensitivity analysis was done after excluding three studies[2,3,16] favoring USG-guided M&R demonstrated a relative risk of 1.00 (CI: 0.96–1.05) [Figure 3]. The heterogeneity after sensitivity analysis and selective exclusion of studies favoring USG M&R was 0% [Figure 3].

Figure 2.

Pooled effect estimates of relative risk between USG versus blind/conventional/landmark-guided distal radius fracture M&R. M&R: Manipulation and reduction, USG: Ultrasound

Figure 3.

Sensitivity analysis after exclusion of studies favoring USG-guided M&R. M&R: Manipulation and reduction. USG: Ultrasound

DISCUSSION

We estimated a benefit of a 10% decrease in the risk of malalignment using USG-directed manipulation. However, we observed that the confidence interval of our pooled estimate of effect ranged from 0.74 to 1.09 [Figure 2]. This uncertainty, ranging from favor to USG-guided M&R by 26% and landmark-guided reduction by 9%, is because of heterogeneity among the included studies in all the aspects [Figure 2]. Commination at the fracture site, angulation of the distal fragment, the type of anesthesia used, and the experience of the caregivers for M&R were the factors that contributed to the clinical heterogeneity in this review. While there were studies[2,3,8,16] that matched cases and controls for the above-described factors, most did not. Failure to randomize, conceal allocation, and blinding of participants, caregivers, and outcome assessors in most studies contributed to the methodical heterogeneity in this review. One study satisfied all the criteria for an RCT, and the findings of this review are similar to it.[8] Two studies studied sensitivity, specificity positive predictive value for the USG detection of DRF; however, reduction success was reported as a surrogate outcome by them.[13,15] They reported outcomes similar to the findings of this review. One had a difference of 1 DRF between cases and controls while the other was a controlled study.[13,15] Studies, as mentioned earlier, carried maximum weightage in MA due to large sample size, narrow CI, and study design.

Three studies[2,3,16] had a similar effect estimate and direction of effect favoring the USG-guided M&R than the other six studies[4,8-10,13,15] in MA. These aforementioned studies[2,3,16] contributed to statistical heterogeneity in this review due to overlapping point estimates and wide CI. Sensitivity analysis performed after eliminating these studies[2,3,16] demonstrated a relative risk of 1.00 (CI: 0.96–1.05). The outcomes of sensitivity analysis from six studies[4,8-10,13,15] with narrow CI crossing the null effect, and overlapping CI represent the actual outcomes of this review. The statistical heterogeneity from these studies could be because of bias introduced due to the inappropriate design of the included studies. Two of these reports had an ambispective study design with a prospective USG and retrospective landmark-guided arm.[2,16] While one study[2] reported no difference between the groups on radiographic indices, the other two[3,16] described improvement in one (volar tilt).

We have described the need for this review despite a review by Malik et al. under the introduction section.[1] Malik et al. narrated the outcomes of individual studies favoring USG and landmark-guided M&R for manipulation success rate, surgical fixation rate, accuracy of fracture detection, and improvement in radiological parameters without pooling effect estimate of failure or success. The authors also concluded a need for properly designed studies for logical conclusions on the efficacy of USG in M&R.

There is a source of publication bias in this review due to the noninclusion of reports from trial registries and studies published in languages other than English. Malik et al. reported the nonavailability of published results as the cause behind the noninclusion of reports from trial registries. Individual studies had selective outcome reporting of the described primary endpoints. However, there was minimal scope for changing the primary research question in most of the studies. We did not investigate gray literature in this review.

We conclude that the outcomes of sensitivity analysis represent an actual effect estimate and, the USG-guided manipulation of DRF offers no benefits over the landmark-guided reductions. There are more costs to the treatment and reduction times in a busy emergency room.[2] Two studies imparted a substantive effect on this review due to the study design and the sample size.[8,13] Both the studies did not find a difference between USG-guided and landmark-based M&R. Therefore, we advocate future RCT with methodological vigor to settle the controversy between the methods of M&R in DRF.

Research quality and ethics statement

This study was a systematic review of the published research available in the public domain so it was not required to undergo Ethics committee approval. The authors followed applicable EQUATOR Network (https://www.equator-network.org/) guidelines during the conduct of this research project.

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.

ANNEXURE

Annexure 1

1. Cochrane central search strategy

   The Cochrane Library (Cochrane Central Register of Controlled Trials (CENTRAL) (2022, Issue 8)) was searched on August 27, 2022 for Mesh terms ‘radius fractures and “ultrasonography.” The search yielded 8 records. Screening of titles identified 5 potentially eligible studies. [#1 MeSH descriptor: [Radius Fractures] explode all trees and #2 MeSH descriptor: [Ultrasonography] explode all trees.]

2. CINAHL search strategy:

   We searched CINAHL (March 2008–August 2022), through EBSCO HDAS.nice.UK on August 21, 2022, with CINAHL subject headings “ultrasonography” and “radius fractures” and obtained 24 results. Screening of the titles identified 5 potentially eligible studies.

3. DOAJ search strategy:

   We searched the Directory of open access journals (DOAJ) on August 19, 2022, for broad terms “Radius” and “Ultrasound” in all subjects and which yielded 142 results. Screening of titles identified 3 potentially eligible studies.

4. Embase search strategy:

   We searched EMBASE (1974 to August 2022) through HDAS.nice.uk (Ovid) on August 22, 2022, through multi-field search using terms “radius” and “fracture” or “radius fracture” or “fracture radius” or “distal radius” and “ultrasonography” in all fields which yielded 183 results. Screening of the titles identified 7 potentially eligible studies

5. Medline search strategy:

   We searched Medline (February 1948 to August 2022), through EBSCO (HDAS.nice.UK) on August 21, 2022, with Mesh terms “ultrasonography” (Major and unexpanded) and “radius fracture” (Major unexpanded) and obtained 21 results. We screened the titles extracted 4 potentially eligible studies.

6. PubMed search strategy:

   We searched PubMed (January 1952 to August 2022) on August 21, 2022, with Mesh terms “radius” and “bone fracture” or “fracture radius” or “fractures radius” and “ultrasonography” without any filters and obtained 190 results. Screening of the titles yielded 12 potentially eligible studies meeting the inclusion criteria. [((((radius [MeSH Terms]) AND (bone fracture[MeSH Terms])) OR (fracture, radius[MeSH Terms])) OR (fractures, radius[MeSH Terms])) AND (“ultrasonography”[MeSH Terms])].

7. Proquest search strategy:

   We searched Proquest (1990 to August 2022), on August 21, 2022, using Mesh terms MESH.EXACT (“Radius Fractures”) AND MESH.EXACT (“Ultrasonography”) and obtained 96 results. Screening the titles identified 6 potentially eligible studies.

8. Scopus search strategy:

   We conducted a Scopus advanced search in title abstract and keywords (1930 to August 2022) on August 21, 2022, with terms “ultrasonography” or “ultrasound” or “sonography” and “radius” or “radial” and “fracture” limited to subject area Medicine, journal, conference proceedings, and English language. We searched document type article with keyword “human” and “humans.” We obtained 3268 results. Screening of the titles yielded 14 potentially eligible studies. [j ultrasonography OR ultrasound OR sonography AND radius OR radial AND fracture AND (LIMIT-TO (DOCTYPE, “ar”)) AND (LIMIT-TO (SUBJAREA, “MEDI”)) AND (LIMIT-TO (EXACTKEYWORD, “Human”) OR LIMIT-TO (EXACTKEYWORD, “Humans”)) AND (LIMIT-TO (LANGUAGE, “English”)) AND (LIMIT-TO (SRCTYPE, “j”) OR LIMIT-TO (SRCTYPE, “p”))]