Description of the condition
The three bones in each finger are called, going from the palm of the hand towards the end of the fingers, the proximal, middle and distal phalanges. The proximal interphalangeal (PIP) joint is the articulation or joint between the proximal and middle phalanges of each finger. The PIP joint is primarily a stable hinge joint. It has a typical range of movement from zero degrees extension (finger is straightened out) up to 110 degrees of flexion (finger is bent) (Glickel 2005). Anatomically, the PIP joint is complex. The contours of the bony surfaces of the proximal and middle phalanges increase the stability of the PIP joint, allowing resistance to lateral (sideways) and rotational stress (Burton 1973; Glickel 2005; Kuczynski 1968). The soft tissue structures surrounding the PIP joint provide further joint stability (Burton 1973; Sprauge 1975). These structures include the volar (palmar) plate, the cord collateral and accessory collateral ligaments and the extensor expansion (Burton 1973; Freiberg 1999; Sprauge 1975). The volar plate, which is a strong soft tissue attachment to the base of the middle phalanx on the palm side, acts as a static ligament and limits hyperextension (thus preventing the finger from being bent backwards) of the PIP joint (Glickel 2005; Sprauge 1975). The other ligamentous structures combine with the volar plate to provide lateral and dorsal (back of the hand) stability (Freiberg 1999).
Injury to the proximal interphalangeal (PIP) joint of the fingers is one of the most common injuries of the hand (Eaton 1976; Sprauge 1975). The spectrum of injury to this joint can range from incomplete soft tissue disruption to unstable fracture-dislocations (Eaton 1976; Incavo 1989). Damage to the PIP joint is frequently the result of a hyperextension force resulting in soft tissue injury, with or without an associated fracture. Hyperextension is defined as a movement of the middle phalanx in a dorsal (backwards) direction beyond neutral, in respect to the proximal phalanx (Sprauge 1975). Any hyperextension injury to the PIP joint will result in some degree of disruption of the ligament complex. Many injuries are incomplete, with only partial damage to ligaments and soft tissue structures. In these cases joint stability is maintained; however, individuals will present with pain, swelling and limited movement (Bowers 1986; Burton 1973). In more severe cases there can be bony disruption and complete rupture of one or more of the supporting structures of the PIP joint (Eaton 1976; Glickel 2005). PIP joint injuries are generally classified into three main types of injury: type I (hyperextension injury); type II (dorsal dislocation); and type III (fracture - dislocation). Table 1 outlines the pathology of these injuries with respect to the damaged structures (Eaton 1976).
Sporting activities account for the majority of PIP joint hyperextension injuries. For example, Gaine 1998 reported that 92.8% of patients in their study received their injuries during sporting activities; 78% of these injuries being a direct result of ball games. Phair 1989 found that all 74 patients in their study reported a hyperextension mechanism of injury, and 51 patients (69%) had sustained an injury to the PIP joint from a sporting activity.
Description of the intervention
Type I (hyperextension) or type II (dislocation) injuries of the PIP joint are generally treated by either a period of immobilisation or by protected mobilisation (or a combination of the two) of the finger with range of motion exercises. Dislocated PIP joints without fractures are reduced (put back into place) using gentle manipulation beforehand. Unstable injuries, usually involving bony disruption, are generally managed surgically (Burton 1973; Eaton 1976).
The PIP joint is often immobilised using plaster of Paris or a rigid plastic or metal splint, secured with tape or Velcro to one side of the finger. In most cases only the PIP joint is kept still or supported and the joints above and below are free to move.
Common types of protected mobilisation are buddy strapping, where the injured finger is strapped to a neighbouring finger, and dorsal block splinting. Dorsal block splinting involves securing a metal or plastic splint to the back of the finger and allowing the patient to loosen the tape to bend the PIP joint for exercises whilst preventing unwanted straightening.
Time frames for either of these forms of treatment vary from one to six weeks. It is anticipated that usual function is resumed between six and eight weeks post injury (Eaton 1976; Freiberg 1999; Incavo 1989; Sprauge 1975).
How the intervention might work
Sprauge 1975 suggested that immobilising the injured PIP joint in a position of flexion for a period of time is favourable for soft tissue healing and for preventing further dorsal or lateral dislocation of the finger. However, other authors report that immobilisation of the PIP joint results in less positive outcomes with individuals taking a greater amount of time to regain range of movement and function of their affected finger (Phair 1989).
Immediate protected mobilisation of the PIP joint, within the individuals' pain tolerance, has been advocated to prevent finger stiffness (Phair 1989). It is considered that mobilisation enhances cartilage and soft tissue healing while movement assists in minimising the formation of tendon adhesions and PIP joint contractures (Freiberg 1999; Kiefhaber 1998). Protective splinting in the form of buddy strapping is considered to prevent lateral and dorsal disruption of the PIP joint and to minimise pain on extension (Bowers 1986; Freiberg 1999; Gaine 1998).
Why it is important to do this review
These injuries may result in chronic pain, stiffness, deformity or premature degenerative arthritis (Freiberg 1999). Because the hand is so vital to everyday activities and work, a poor outcome can be disproportionately disabling. Currently there is uncertainty on what is the best approach to take for these injuries in terms of the extent, type and duration of immobilisation. These point to the need for this systematic review in order to inform practice.
To assess the effects (benefits and harms) of conservative interventions for treating hyperextension injuries of the proximal interphalangeal joints of the fingers presenting within one month of injury. Comparisons included:
- PIP joint unrestricted mobilisation (e.g. no splint) versus immobilisation
- PIP joint protected mobilisation (e.g. dorsal blocking splinting, buddy strapping) versus immobilisation
- PIP joint unrestricted mobilisation versus protected mobilisation
Criteria for considering studies for this review
Types of studies
We considered any randomised controlled and quasi-randomised (method of allocating participants to a treatment which is not strictly random: e.g. by date of birth, hospital record number, alternation) trials of conservative management for hyperextension injuries of the PIP joint in the hand.
Types of participants
We included trials with participants that had an acute (less than one month), conservatively-managed hyperextension injury to the PIP joint. Injury was defined through: history of dislocation, history of hyperextension injury, diagnosis of volar plate/collateral ligament fracture avulsion via x-ray, joint laxity or pain in the PIP joint following trauma to the finger.
We excluded trials focusing on participants presenting with open hyperextension injuries, concomitant tendon injury, fracture (with the exception of ligamentous avulsion fractures) or hyperextension injuries requiring surgical treatment. Trials focusing on participants with underlying rheumatological, neurological or congenital conditions were also excluded.
Types of interventions
Randomised comparisons for the conservative management of PIP joint hyperextension injuries included any two or more of the following interventions: unrestricted movement, buddy strapping, protective splinting, and immobilisation. We also aimed to include trials comparing different time frames for treatment.
We excluded trials that compared management after surgical intervention or different exercise regimens. Comparisons of different exercise regimens were excluded due to the high variation possible between studies, for example, frequency and duration of exercises, as well as active versus passive versus resisted motion.
Types of outcome measures
- Self-reported, preferably validated, functional questionnaires such as the Disability of the Arm, Shoulder and Hand (DASH) and the Michigan Hand questionnaires
- Poor outcome: finger stiffness, joint laxity, joint deformity (cosmetic or clinically assessed), subsequent surgery
- Pain (visual analogue scale, ordinal scale, pain questionnaire)
- Range of motion
- Return to work/sport/previous activity
- Number of follow-up appointments
Timing of outcome assessment
Our focus is on primary outcomes at final follow-up in the individual studies. Ideally, follow-up should be at least six months for type I injuries and at least 12 months for type II injuries.
Search methods for identification of studies
We searched the Cochrane Bone, Joint, Muscle Trauma Group Specialised Register (January 2012), the Cochrane Central Register of Controlled Trials (The Cochrane Library 2012, Issue 1), MEDLINE, MEDLINE In-Process & Other Non-Indexed Citations, MEDLINE Daily Update (1946 to January Week 2 2012), EMBASE (1980 to 2012 Week 03), CINAHL (1950 to 24 January 2012) and PEDro - the Physiotherapy Evidence Database (1929 to March 2012) (see Appendix 1 for details of the search strategies). The subject-specific search has not been combined with a trial filter due to the small number of trials in this area.
We searched the WHO International Clinical Trials Registry Platform (to 8 March 2012) to identify ongoing and recently completed trials, and also Zetoc (1993 to 8 March 2012) to identify additional trials reported in conference proceedings.
Trials were included regardless of language and publication status.
Searching other resources
We also searched reference lists of articles, reviews and relevant textbooks.
Data collection and analysis
Selection of studies
Two review authors (MB and ZA) independently screened the titles, abstracts and descriptors of identified studies for possible inclusion. We obtained full text versions of all studies that were considered potentially eligible by either author, and independently assessed these for inclusion. Any disagreement was resolved by discussion with a third author (JC). Where uncertainty remained about eligibility, we contacted trial authors for clarification. Translations for potential studies published in languages other than English were obtained where possible.
Data extraction and management
Two review authors (JC and ZA) independently extracted trial details and data using a standardised data collection form. The form was tested for consistency on two studies prior to the commencement of the review. Any disagreement was resolved by discussion with a third review author (MB).
Assessment of risk of bias in included studies
Two review authors (JC and ZA) independently assessed the risk of bias in each included study using The Cochrane Collaboration's 'Risk of bias' tool. We assessed generation of allocation sequence, allocation concealment, blinding (of trial participants, care providers and assessors), incomplete outcome data, selective outcome reporting, and other sources of bias. The risk of bias was rated for each domain and was expressed as 'Yes', implying a low risk of bias; 'No', implying a high risk of bias; or 'Unclear', implying the risk of bias is unclear. Any disagreement was resolved by discussion with a third review author (MB) when required. Journal titles, authors and supporting institutions were not be masked at any time.
Measures of treatment effect
Risk ratios and 95% confidence intervals were calculated for dichotomous data. We planned to calculate mean differences or, where different scales or tools were used to measure the same outcome, standardised mean differences and 95% confidence intervals for continuous data.
Unit of analysis issues
A possible unit of analysis issue may arise if an individual had more than one injured finger and the individual (or both or more digits) is randomised rather than the individual digits. When such an issue arose and appropriate corrections were not made, we presented the data for such trials where the disparity between the units of analysis and randomisation is small. Should data pooling be possible, we planned to perform a sensitivity analysis to examine the effects of excluding incorrectly reported trials from the analysis.
Dealing with missing data
We contacted authors of individual studies for missing data and clarification. Where possible, we performed intention-to-treat analyses. We stated beforehand that we would not impute missing standard deviations.
Assessment of heterogeneity
We planned the following approach for assessing study heterogeneity, which we determined from both clinical and statistical perspectives. Clinical heterogeneity was assessed in terms of important differences between trials in the study populations, interventions and outcome measures. Statistical heterogeneity between pooled trials was to be assessed using a combination of visual inspection of the graphs along with consideration of the chi² test (with statistical significance set at P < 0.10), and the I² statistic.
Assessment of reporting biases
We attempted to reduce reporting biases, particularly publication bias, through our comprehensive search of the literature. This extended across multiple databases using a sensitive search strategy, include clinical trial registers, and a search of Zetoc that identified trials reported at conferences but never published in full. Thus, searches were not constrained by language or publication status.
Where considered appropriate from a clinical perspective, we planned to pool data from comparable trials. Initially we planned to use the fixed-effect model and 95% confidence intervals. We also would have considered using the random-effects model, especially where there was unexplained heterogeneity.
Subgroup analysis and investigation of heterogeneity
Should sufficient data become available in future versions of this review, our planned subgroup analyses are based on age (those younger than 16, those aged between 17 and 64 and those aged older than 65), the type of injury (see Table 1), and time elapsed between injury and medical intervention (treatment started within two weeks versus after two weeks) where possible and appropriate.
We planned to perform sensitivity analyses to examine the effects of missing data and of the inclusion of trials at high risk of bias (such as from lack of allocation concealment). However, it was not possible to perform these analyses in this version of the review.
Description of studies
Results of the search
The search strategy found 520 references, of which 480 were excluded through initial screening of reference titles and abstracts. Of those excluded, 155 were duplicates and 325 were not considered relevant. Of the 40 remaining potentially relevant studies, for which full reports were obtained, 35 were excluded as they were clearly not randomised or quasi-randomised controlled trials. Three studies were included (Norregaard 1987 (3 reports of study); Thomsen 1979; Thomsen 1995) and one (Arora 2004) was excluded. No ongoing trials were identified.
All three included studies were reported in medical journals. Norregaard 1987 was reported in a conference abstract (Norregaard 1986), an English language report (Norregaard 1987) and a Danish language report (Norregaard 1988). A translation from Danish were obtained for Thomsen 1979.
Details of the three individual trials can be found in the Characteristics of included studies. A summary of these is presented below.
In total, there were 366 participants with proximal interphalangeal joint hyperextension injuries randomised into the three trials. Norregaard 1987 included 126 participants (as reported in Norregaard 1986 and Norregaard 1988). However, we based our analysis of this trial on the English language report that reported on 112 participants only, 87 of whom were followed up at six months. Thomsen 1979 randomised 200 patients of which 181 were followed up. This study included participants with multiple finger injuries and follow-up analysis was based on the number of treated fingers (191) rather than participants. Thomsen 1995 included 40 participants of whom 39 were followed up at six months. Overall, 307 participants and 317 fingers were included in the six months follow-up results for the three trials.
All three trials were conducted in Denmark with participants presenting to hospital emergency departments or orthopaedic clinics. All trials are over 15 years old: Norregaard 1987 was completed in 1983; Thomsen 1979 in 1978; and Thomsen 1995 in 1993.
All participants in the three trials had sustained a hyperextension injury to the proximal interphalangeal joint. Diagnostic criteria varied across the studies and were reported to various levels. Norregaard 1987 included participants with hyperextension trauma presenting with four signs of injury (volar tenderness, pain at maximal flexion, relief from pain in the intermediate position and recurrence of pain at maximal extension). In contrast, Thomsen 1995 did not describe inclusion criteria beyond that of lesions to the volar plate of the proximal interphalangeal joint. Thomsen 1979 admitted participants who sustained hyperextension trauma and volar pain at maximal flexion, pain relief in semi-flexion and pain increasing at maximal extension. Participant demographics of gender and age were not reported in Norregaard 1987. The participants of Thomsen 1979 were aged between 7 and 76 years, and 41% were male. Thomsen 1995 reported an age range of 18 to 79 years and that 49% of participants were male.
The three trials evaluated four different comparisons in all.
Norregaard 1987 compared unrestricted mobilisation of the affected PIP joint versus immobilisation of the affected PIP joint in 15 degrees of flexion, the metacarpophalangeal joint in 80 to 90 degrees flexion and the wrist in 30 degrees of extension for three weeks using an aluminium splint.
Thomsen 1995 investigated limitation of PIP joint protected motion through the application of a double finger elastic bandage (motion ranging between neutral and 25 to 30 degrees of flexion) versus immobilisation with an aluminium splint applied to the dorsum of the finger prohibiting PIP joint movement. The splint was designed to position the finger in 15 degrees of flexion at the PIP joint. These interventions were instituted for an initial two week period, after which the elastic bandage or aluminium splint was removed and normal hand movement was encouraged.
In Thomsen 1979, patients were randomised into four groups: group one, aluminium splint for one week; group two, aluminium splint for three weeks; group three, plaster of Paris cast for one week; group four, plaster of Paris cast for three weeks. The authors describe positioning the PIP joint in semi-flexion but do not report the angle of measurement. Active range of motion was commenced for all groups once the aluminium splint or plaster of Paris had been removed, at either one or three weeks. Thomsen 1979 made two comparisons: immobilisation of the affected PIP joint for one week versus three weeks; and aluminium splint versus plaster of Paris cast. However, results were provided for the first comparison only.
Norregaard 1987 reviewed joint stiffness, laxity and deformity at both six months and three years post injury. However, the authors do not clearly report the methods by which they determined and measured these sequelae. Secondary outcomes of pain and range of movement were acknowledged. Pain was rated using the Moller scale (Moller 1974). The origin, method, validity and reliability of this scale is not reported in the literature. Range of movement was described in terms of an extension or flexion deficit. It is not clear if goniometry was used to assess for deficit or what parameters the authors use to define a range of movement deficit.
Thomsen 1979 presented pooled results for the two groups immobilised for one week and the two groups immobilised for three weeks. They commented on stiffness and deformity but did not report the methods by which they measured these outcomes. Secondary outcomes of pain and deficits in extension and flexion were reported. This trial also monitored discomfort from use of a splint or plaster and recorded sick leave. As with Norregaard 1987, pain was rated using the Moller scale (Moller 1974). The parameters of defining an extension/flexion deficit were not described.
Norregaard 1987 and Thomsen 1979 both appeared to use the same assessment tool that considered pain (during movement, work and hobby activities), volar plate tenderness, stiffness and coldness, thickened joint, extension and flexion deficits, swan neck deformity and button-hole deformity. This assessment tool is not referenced by either study.
Thomsen 1995 assessed range of motion, volar stability and return to work at two weeks. The authors also used an assessment of results scale originally developed by Benke 1979 that is based on movement, symptoms, stability and function. Outcome was rated as excellent, good or poor. The authors used this scale to comment on the primary outcome measure of 'poor outcome'. The only secondary outcome reported in this study is range of movement. This is discussed in terms of achieving PIP joint range of motion 0 to 90 degrees. There was no information provided on how these deficits were measured and if goniometry was used.
Risk of bias in included studies
See Figure 1 and Figure 2. All three trials included in this review were methodologically flawed. Many items described in the 'Risk of bias' tool were judged to be at high risk of bias. Others were judged as being unclear because of insufficient information. More details about the sources of bias in each study can be found in the Characteristics of included studies.
|Figure 1. Risk of bias graph: review authors' judgements about each risk of bias item presented as percentages across all included studies.|
|Figure 2. Risk of bias summary: review authors' judgements about each risk of bias item for each included study.|
Norregaard 1987, which was a quasi-randomised trial, based on the date of patient trauma, was at a high risk of selection bias. Thomsen 1979 used a random numbers table for sequence generation but Thomsen 1995 did not describe their method of sequence generation. There was no details to determine whether allocation was concealed in either trial.
The nature of these trials make it impossible for participants to be blinded to their treatment. It is also not possible to blind those delivering treatment. However, it is unclear if there was any blinding of assessors at follow-up in any of the included studies.
Incomplete outcome data
Participants were lost to follow-up in all three trials. The reporting of loss to follow-up was inconsistent in Norregaard 1987, which was judged at high risk of attrition bias based on the inconsistent reporting and the high loss to follow-up (22% at six months using data from Table 1 of the trial report; see the Characteristics of included studies). Nine per cent of patients randomised to treatment with immobilisation for one week were lost compared with 10% immobilised for three weeks in Thomsen 1979. One participant was lost to follow-up in Thomsen 1995 from the protection motion group.
It is difficult to determine the presence of selective reporting as protocols are not available for any of the trials.
Other potential sources of bias
All three studies failed to provide sufficient information regarding the baseline characteristics of their treatment groups. Two of the three studies (Thomsen 1979; Thomsen 1995) reported an inadequate length of follow-up (six months). Although outcome measures reported in each of the studies are appropriate to these injuries, the methods of assessment vary and it is not clear if these were undertaken in a validated and reliable fashion, if these are subjective or objective measures, nor the parameters for determining the presence of a particular outcome. For example, Thomsen 1979 and Norregaard 1987 both report pain using the Moller scale, but do not describe the nature of this scale. Similarly, all three trials report range of movement as an outcome without reporting how this was assessed (e.g. goniometry). Further bias may have been introduced in the assessment process by the involvement of multiple assessors. Thomsen 1995 reports the involvement of three assessors whilst the other two studies do not provide any information as to the number of assessors.
Effects of interventions
No immobilisation versus immobilisation
Norregaard 1987 found no statistically significant differences between the no immobilisation (unrestricted motion) group and the splint immobilisation group in the reported outcomes at either six months or three years. The risk of attrition bias, in particular relating to the high loss to follow-up and inconsistencies in the reporting of the losses to follow-up raises some questions over the reliability of the results from this quasi-randomised trial. The following analyses use the data from the main publication of this study only. At six months after injury there were no statistically significant differences between the two groups for poor outcome, a primary outcome measure, particularly joint stiffness and coldness (6/43 versus 11/44; RR 0.56, 95% CI 0.23 to 1.37) or thickened joints (8/43 versus 7/44; RR 1.17, 95% CI 0.46 to 2.94) (see Analysis 1.1). One participant in the splint group developed a swan neck deformity. Similar findings applied at three years (see Analysis 1.2): stiffness and coldness (3/34 versus 5/34; RR 0.60, 95% CI 0.16 to 2.31) and thickened joint (8/34 versus 4/34; RR 2.00, 95% CI 0.66 to 6.02). No participant had swan neck or button hole deformity at this time. Similar numbers in each group had 'sequelae', which were mainly 'mild' involving 'minor pain': 9/34 versus 7/34; RR 1.29, 95% CI 0.54 to 3.06). While there were some differences across the three papers for this trial, none of these changed the overall findings: the results for 'sequelae' at two to three years are given in Analysis 1.2.
There was also no difference between the two groups for secondary outcomes of pain and tenderness at six months (see Analysis 1.3): pain during movement (7/43 versus 5/44; RR 1.43, 95% CI 0.49 to 4.17); pain during work (5/43 versus 5/44; RR 1.02, 95% CI 0.32 to 3.28); pain during hobby activities (4/43 versus 4/44; RR 1.02, 95% CI 0.27 to 3.83); and volar plate tenderness (5/43 versus 8/44; RR 0.64, 95% CI 0.23 to 1.80). Similar findings applied at three years for pain and tenderness (see Analysis 1.4): pain during movement (3/34 versus 2/34; RR 1.50, 95% CI 0.27 to 8.42); pain during work (5/34 versus 2/34; RR 2.50, 95% CI 0.52 to 12.01); pain during hobby activities (4/34 versus 1/34; RR 4.00, 95% CI 0.47 to 33.97); and volar plate tenderness (6/34 versus 6/34; RR 1.00, 95% CI 0.36 to 2.79). There were no differences between groups at six months for extension deficit (5/43 versus 3/44; RR 1.71, 95% CI 0.43 to 6.70) or flexion deficit (7/43 versus 8/44; RR 0.90, 95% CI 0.36 to 2.25) (see Analysis 1.5). The same lack of statistically significant differences between two groups applied at three years in both extension deficit (1/34 versus 2/34; RR 0.50, 95% CI 0.05 to 5.26) or flexion deficit (2/34 versus 3/34; RR 0.67, 95% CI 0.12 to 3.74) (see Analysis 1.6).
Protected mobilisation versus immobilisation
Thomsen 1995 found no statistically significant differences in reported outcomes between protected mobilisation versus immobilisation in an aluminium splint at two weeks following injury, nor at six months. There was no formal measure of functional outcome. At two weeks, fewer participants in the protection mobilisation group had instability (3/19 versus 7/20; RR 0.45, 95% CI 0.14 to 1.49; see Analysis 2.1) and more were able to bend their finger (11/19 versus 6/20; RR 1.93, 95% CI 0.89 to 4.17; see Analysis 2.2); neither result reached statistical significance.
At six months, there was no difference between groups in the number of participants who were considered to have had an excellent outcome (16/19 versus 16/20; RR 1.05, 95% CI 0.79 to 1.41; see Analysis 2.3). The remaining participants (3 versus 4) had a good outcome and thus none of the participants were considered to have had a poor outcome.
The authors reported that all participants were able to return to their normal job within two to three weeks post injury.
Immobilisation 1 week versus 3 weeks
Thomsen 1979 used two methods of immobilisation (plaster cast and aluminium splint) to restrict movement for one week or for three weeks. They presented "complaints and findings" results for 191 fingers (181 participants) split according to the period of immobilisation: one week versus three weeks. Thomsen 1979 did not provide details of the distribution of the participants with multiple injuries to their PIP joints: but these were equally distributed with five extra fingers in each duration-of-immobilisation group. There were no statistically significant differences in the reported outcomes between the two duration of immobilisation groups at six month follow-up. Similar results were found for poor outcome (see Analysis 3.1) as represented by complaints of joint stiffness and coldness (23/96 versus 28/95 fingers; RR 0.81, 95% CI 0.51 to 1.30) and joint thickening (19/96 versus 15/95 fingers; RR 1.25, 95% CI 0.68 to 2.32). Likewise, there was little difference between the two groups in pain or tenderness (see Analysis 3.2): pain during movement (20/96 versus 16/95 fingers; RR 1.24, 95% CI 0.68 to 2.24; pain during work (17/96 versus 13/95 fingers; RR 1.29, 95% CI 0.67 to 2.51); pain during hobby activities (19/96 versus 15/95 fingers; RR 1.25, 95% CI 0.68 to 2.32); and volar plate tenderness (1/96 versus 3/95; RR 0.33, 95% CI 0.03 to 3.12). Few people in each group had decreased range of motion at six months (decreased extension: 3/96 versus 1/95 fingers; RR 2.97, 95% CI 0.31 to 28.04; decreased flexion: 3/96 versus 3/95 fingers; RR 0.99, 95% CI 0.20 to 4.78) (see Analysis 3.3). The authors reviewed the number of sick days taken by participants, regardless of immobilisation method, and concluded that number of sick days increased when the PIP joint was immobilised for a longer period of time.
Aluminium splint versus plaster of Paris
Thomsen 1979 restricted movement for one or three weeks using either an aluminium splint or a plaster cast. The authors reported there was no difference between two methods of splinting and presented pooled data according to the duration of immobilisation. However, they reported the cost of application of an aluminium splint to be less than the cost of a plaster cast (9.53 Danish kroner versus 33.73 Danish kroner), mainly reflecting differences in staffing time involved.
Despite a comprehensive search for appropriate trials, we were only able to identify three trials that adequately met our selection criteria for inclusion in this review.
Summary of main results
Three trials of 366 participants with hyperextension injuries to the proximal interphalangeal joints of the hand managed with conservative interventions were identified.
Norregaard 1987, which compared unrestricted mobility with splint immobilisation, found no statistically significant differences in outcome between the two groups at six months or three years post injury. Reported outcomes included finger movement, pain, joint tenderness and stiffness.
Thomsen 1995 found no statistically significant differences in reported outcomes between protected mobilisation versus immobilisation in an aluminium splint at two weeks following injury, nor at six months. Reported outcomes included finger movement, instability, and overall outcome.
Thomsen 1979 found no statistically significant differences between immobilisation for one week versus three weeks in the reported outcomes, relating to poor outcome, pain and tenderness and range of motion at six months follow-up. Thomsen 1979 reported, without providing data, no difference in outcome at six months post intervention between aluminium splint versus plaster of Paris aside from greater application (mainly staffing) costs for plaster of Paris.
There was no statistically significant difference in outcomes reported by any of the three studies included in this review and it is therefore not possible to determine if one form of treatment (unrestricted mobilisation, protected mobilisation, immobilisation) is better than any other when deciding how to conservatively manage PIP joint hyperextension injuries.
Overall completeness and applicability of evidence
All studies report on one of this review's primary outcome measures: 'poor outcome', including finger stiffness, joint laxity and joint deformity. The other primary outcome measure, self-reported functional outcomes, was not acknowledged in any of the three studies.
All three studies adequately reported the inclusion and exclusion criteria. While it is likely that these trial populations are representative, the lack of detailed information on the baseline characteristics of the trial participants on all three trials reduces the confidence on the applicability of their findings.
Despite the age of these trials, being more than 15 years since publication, these methods of conservative management continue to be applicable in clinical practice. Trials report participants presenting to emergency departments and orthopaedic clinics, as is commonly the case. However none of the three trials describe the methods used to evaluate the outcome measures on which the authors base their results. Additionally, the studies did not report information on exercise prescription. It is therefore difficult to extrapolate these outcomes to clinical practice and makes it almost impossible to replicate the studies in order to compare results.
Quality of the evidence
The quality of evidence available for the purpose of this review is very low. As well as being quasi-randomised, Norregaard 1987 was inconsistently reported and was at high risk of attrition bias. Thomsen 1979 was the only study to report an adequate method of sequence generation, but neither this trial nor Thomsen 1995 provided information on allocation concealment. Whilst we recognise that for the three included trials it was not possible to blind treatment providers or the patients to the interventions included, blinded assessors could have been used to review and record clinical outcomes.
Potential biases in the review process
In an attempt to reduce the possibility of bias in the review process, we undertook a comprehensive search of electronic databases. We acknowledge that all three studies are more than 15 years old. Although we found more recent publications on management of PIP joint hyperextension injuries, none were randomised controlled trials. There may be unpublished trials that have been reported in abstracts only. For future updates, we will attempt to identify non-published studies by liaising with content area specialists and search conference proceedings. As part of this review, contact was made with two authors to establish both the suitability of the trial for inclusion in the review and for clarification on the randomisation process.
Agreements and disagreements with other studies or reviews
We did not find any other reviews that looked at conservative management of hyperextension injuries.
Implications for practice
There is insufficient evidence from trials testing the need for, and the extent and duration of, immobilisation to inform on the key conservative management decisions for treating hyperextension injuries of the proximal interphalangeal joints.
Implications for research
Hyperextension injuries of the proximal interphalangeal joints are a common presentation in hand clinics (Eaton 1976; Sprauge 1975). Currently, there is no consensus on optimum management. Further randomised trials comparing intervention options, exercise prescription and outcomes would be useful in determining the appropriate level of intervention for these injuries. Outcomes should include self-reported functional ability and pain, and objective measures of deformity, motion, grip strength.
A randomised controlled trial comparing thermoplastic splinting and early active mobilisation versus buddy strapping and early active mobilisation is recommended. As well as short term follow-up at three and six months to monitor recovery and capture time to return to function and employment, longer term follow-up at 12 and 24 months post injury is needed. Direct and indirect costs should also be included in the study to determine the cost-effectiveness of each intervention. Such trials should meet the CONSORT criteria for design and reporting of non-pharmacological studies (Boutron 2008).
The authors would like to thank Roma Bhopal, Lesley Gillespie, Nigel Hanchard, Helen Handoll and Adam Watts for helpful comments about the protocol and the review, as well as acknowledging the assistance of Lindsey Elstub and Amy Kavanagh of the Cochrane Bone, Joint and Muscle Trauma Group in preparing drafts of both the protocol and the review.
We would like to thank Anne Moseley at The George Institute for Public Health for running searches through PEDro.
We would like to thank Steve Powell, Clinical Effectiveness Department, St George's Healthcare NHS Trust for his assistance in obtaining full text articles.
We would also like to thank Dr Annette Bluemle and Jeppe Schroll for their assistance in translating articles.
We thank Mr John Jakobsen for providing further details of his trial.
Our thanks also to the Therapies Department, St George's Healthcare NHS Trust and to the staff of St George's University Library for their support and assistance.
Data and analyses
- Top of page
- Authors' conclusions
- Data and analyses
- Contributions of authors
- Declarations of interest
- Sources of support
- Differences between protocol and review
- Index terms
Appendix 1. Search strategies
The Cochrane Library (Wiley Online Library)
#1 (proximal interphalangeal):ti,ab (77)
#2 (PIPs OR PIPj* OR PIP joint*):ti,ab (307)
#3 (#1 OR #2) (358)
#4 MeSH descriptor Dislocations, this term only (179)
#5 MeSH descriptor Fractures, Bone, this term only (976)
#6 MeSH descriptor Finger Injuries, this term only (82)
#7 (fracture* OR dislocat* OR injur*):ti (8773)
#8 ((injur* OR fracture* OR dislocat*) NEAR/3 (ligament* OR volar plate OR interphalangeal OR proximal)):ti,ab (431)
#9 (hyperexten* OR sublux*):ti,ab (175)
#10 (#4 OR #5 OR #6 OR #7 OR #8 OR #9) (9589)
#11 (#3 AND #10) (25)
MEDLINE (Ovid Online)
1 "proximal interphalangeal".ti,ab. (1910)
2 (PIPs or PIPj* or "PIP joint*").ti,ab. (1323)
3 VOLAR PLATE/in (11)
4 1 or 2 or 3 (2796)
5 DISLOCATIONS/th (3101)
6 FRACTURES, BONE/th (6275)
7 FINGER INJURIES/th (812)
8 (fracture* or dislocat* or injur*).ti. (277899)
9 ((injur* or fracture* or dislocat*) adj3 (ligament* or "volar plate" or interphalangeal or proximal)).ti,ab. (10489)
10 (hyperexten* or sublux*).ti,ab. (9309)
11 5 or 6 or 7 or 8 or 9 or 10 (290636)
12 4 and 11 (579)
13 IMMOBILIZATION/ (10879)
14 ("early motion" or "early active motion" or mobili*).ti,ab. (135541)
15 exp EXERCISE THERAPY/ (24123)
16 exercis*.ti,ab. (170086)
17 ((buddy or neighbour* or adjacent) adj2 (strap* or tape* or taping or wrap*)).ti,ab. (46)
18 SPLINTS/ (7069)
19 BANDAGES/ (12730)
20 splint*.ti,ab. (9851)
21 14 or 15 or 16 or 17 or 18 or 19 or 20 (337734)
22 12 and 21 (202)
EMBASE (Ovid Online)
1 PROXIMAL INTERPHALANGEAL JOINT/ (1128)
2 "proximal interphalangeal joint*".ti,ab. (1517)
3 (PIPs or PIPj* or "PIP joint*").ti,ab. (1493)
4 1 or 2 or 3 (3189)
5 DISLOCATION/th or FINGER DISLOCATION/ (1706)
6 FINGER FRACTURE/ (603)
7 JOINT INJURY/th or JOINT FRACTURE/th (130)
8 FINGER INJURY/th (471)
9 ((injur* or fracture* or dislocat*) adj3 (ligament* or "volar plate" or interphalangeal or proximal)).ti,ab. (11783)
10 (fracture* or dislocat* or injur*).ti. (306032)
11 (hyperexten* or sublux*).ti,ab. (9870)
12 5 or 6 or 7 or 8 or 9 or 10 or 11 (318155)
13 4 and 12 (603)
14 MOBILIZATION/ (12918)
15 ("early motion" or "early active motion" or mobili*).ti,ab. (148937)
16 exp EXERCISE/ or MOVEMENT THERAPY/ (164981)
17 exercis*.ti,ab. (199501)
18 ((buddy or neighbour* or adjacent) adj2 (strap* or tape* or taping or wrap*)).ti,ab. (49)
19 SPLINTS/ or SPLINTING/ or BANDAGE/ (17806)
20 splint*.ti,ab. (10286)
21 14 or 15 or 16 or 17 or 18 or 19 or 20 (434718)
22 13 and 21 (206)
CINAHL (Ebsco Interface)
1. TX (proximal N3 interphalangeal N3 joint*) (220)
2. TX (PIP or PIPj*) (315)
3. TX (PIP N3 joint*) (82)
4. TX (volar N1 plate*) (81)
5. S1 or S2 or S3 or S4 (544)
6. (MH "Finger Dislocation") (240)
7. (MH "Finger Injuries") (26)
8. MH "Hand Fractures") (29)
9. TI (fractur* or dislocat* or injur*) (62074)
10. TX (injur* or fracture* or dislocat*) N3 (ligament* or interphalanageal or proximal) (6336)
11. TX (hyperexten* OR sublux*) (2823)
12. S6 or S7 or S8 or S9 or S10 or S11 (68529)
13. S5 and S12 (164)
14. (MH "Joint Mobilization") (1)
15. TX ("early motion"or "early active motion" or mobili*) (21335)
16. (MH "Therapeutic Exercise+") (30452)
17. TX (exercis*) (89288)
18. TX (buddy or neighbour or adjacent) N2 (strap* or tape* or taping or wrap*) (13)
19. (MH "Splints") (429)
20. (MH "Taping and Strapping") (1539)
21. S14 or S15 or S16 or S17 or S18 or S19 or S20 (138485)
22. S13 and S21 (32)
/in: injuries subheading
/th: therapy subheading
.ti: textword in title
.ti,ab: textword in title and abstract
Contributions of authors
All authors were involved in the preparation of the protocol, the conception of which was based on previous work undertaken by MB. All authors contributed to the work and writing involved in the review. Data entry and analysis were completed by SM. JC, MB and ZA completed the first draft of the review with feedback from SM.
Declarations of interest
Sources of support
- St George's NHS Healthcare Trust, UK.
- No sources of support supplied
Differences between protocol and review
In our risk of bias assessment, we did not assess patient reported and clinician rated outcome measures separately in our assessment of assessor blinding and completeness of outcome data. Another difference was that we also assessed risk of performance bias associated with blinding of participants and care providers.
Medical Subject Headings (MeSH)
MeSH check words
* Indicates the major publication for the study