Arthrodesis of the proximal interphalangeal joint of the finger – a systematic review

Abstract

• Arthrodesis of the proximal interphalangeal (PIP) joint of the finger is an established procedure for advanced osteoarthritis. As there are different techniques of fusion, it seems necessary to evaluate the results.

• Primary outcome of this review was to evaluate different arthrodesis methods of the PIP joint and describe different numbers of non-unions. Secondary outcome was to evaluate time to consolidation. Respective complications, if mentioned, were listed additionally.

• The review process was conducted according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines. The selected databases were PubMed, Medline, Embase, Google Scholar and Cochrane Library. Studies reporting outcomes of the arthrodesis with a defined technique and radiological consolidation were included. Complication rates and types were recorded. In total, 6162 articles could be identified, 159 full-texts were assessed and 64 studies were included. Methodological quality was assessed using Methodological Index for Non-Randomized Studies.

• A total of 1923 arthrodeses of the PIP joint could be identified. Twelve different surgical techniques were described, four of these techniques with compression at the arthrodesis site. The most frequently used techniques were K-wires (n = 743, 14 studies), tension-band (n = 313, 15 studies) and compression screws (n = 233, 12 studies). The lowest rate of described non-unions in compression techniques was 3.9% with the compression screw. The highest non-union rate of 8.6% was achieved by interosseous wiring.

• All the described techniques can achieve the goal of fusing an osteoarthritic joint. There is a tendency in the more recent literature for the use of compression techniques.

Introduction

Osteoarthritis of the proximal interphalangeal (PIP) joint, either primary or secondary, limits the range of motion and causes pain with or without instability, leading to significant global hand function impairment (1). Typical aetiologies leading to secondary osteoarthritis are posttraumatic changes, chronic instability or inflammatory diseases, for example rheumatoid arthritis or scleroderma. Operative treatment options include denervation, different arthroplasties, prosthesis or arthrodesis. The aim of arthrodeses is pain reduction in combination with a sufficient global hand function (2). With distinctive deformation of the joint and/or preexisting instability, there is a tendency to recommend arthrodesis because an unstable prothesis is prone to failure. In these cases, the fusion of the joint provides reliable results.

In posttraumatic osteoarthritis, especially of the radial digits with an instability not exceeding 30°, a prothesis could provide excellent results (3, 4). If more than one joint is affected, especially in patients with rheumatoid arthritis, and only a moderate instability exists, silicone arthroplasty is still the method of choice (5).

Arthrodesis of the PIP joint is an established technique for advanced osteoarthritis or when other reconstruction methods have failed. Different techniques for arthrodesis of the PIP joint have been described and their main difference is if there is compression on the arthrodesis or not (6). There is no clear indication in the current literature as to which technique shows the most promising results in terms of union. Typical major complications of PIP joint arthrodesis are non-union and mal-union; minor complications are superficial infections (61).

The aim of this first systematic review was to clarify the following questions: Do different arthrodesis methods of the PIP joint for primary and secondary causes of osteoarthritis or destruction of the joint show (i) different numbers of non-unions? (primary outcome) and (ii) different times to consolidation? (secondary outcome). The different complications of each technique were additionally included but not further evaluated.

Methods

Search methods

The review process was conducted according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines (7). Two reviewers (MM and HV) independently selected studies for inclusion. Disagreements were solved by discussion with a senior author (MR).

The search was conducted from January 1, 1946, to April 28, 2020, in the following databases: PubMed, Medline, Embase, Google Scholar and Cochrane Library by the main author. We initially searched without any language or publication type restrictions. The search algorithm is shown in Table 1.

Table 1.

The respective search string of the different included databases.

Database	Search string
Pubmed	(((proximal interphalangeal joint[Title/Abstract]) OR (pij[Title/Abstract]) OR (pip-joint[Title/Abstract]) OR (finger[Title/Abstract]) OR (digital[Title/Abstract]) OR (pipj[Title/Abstract]) OR (proximal interphalangeal[Title/Abstract])) AND ((arthrodesis[Title/Abstract]) OR (fusion[Title/Abstract]))) NOT equine[Title/Abstract]
Embase	(‘proximal interphalangeal joint’/exp OR pij:ab,ti OR ‘pip joint’:ab,ti OR ‘proximal interphalangeal joint’:ti,ab OR ‘digital’:ab,ti OR ‘finger’:ab,ti) AND (‘arthrodesis’:ti,ab OR ‘fusion’:ab,ti) AND [embase]/lim
Cochrane Library	(pij OR pip joint OR pip-joint OR proximal-interpalangeal-joint OR proximal interphalangeal joint OR digital OR finger) AND (arthrodesis OR fusion)
Google Scholar	allintitle: (“pij” OR “pip joint” OR “proximal interphalangeal joint” OR “digital” OR “finger”) AND (“arthrodesis” OR “fusion”)

Selection criteria

Full-text reports (original articles, randomized controlled trials, controlled clinical trials, retrospective or prospective observational studies, case series and technical descriptions) concerning PIP joint arthrodesis were screened.

Reference lists from included studies and reviews were screened for additional studies and included. Studies reporting outcomes of the arthrodesis with a defined technique and radiological consolidation were included. Complication rates and types were recorded. Clinical studies with an evidence level of I–IV were included. As there were studies which compared arthrodeses to other techniques of joint salvage, those reporting of five or less arthrodeses were also included.

Studies lacking original data, studies whose data were not doubtlessly concerning the PIP joint as well as studies whose full-text were not available were excluded. Doctoral theses were also excluded.

The search flowchart according to the PRISMA guidelines is depicted in Fig. 1. Initially, 6162 articles were identified. Thirteen additional records from reference lists were included. After removing 1914 duplicates, 4261 articles remained. By screening titles and abstracts, a further 4102 studies were excluded.

Figure 1.

This PRISMA flowchart shows the numbers of articles identified as well as the inclusion and exclusion steps.

The full text of 159 articles was thoroughly assessed and evaluated for reporting the number of treated PIP joints, the technique used and the primary endpoint of consolidation. The 64 studies depicted in Table 2 were finally included, and data were extracted from these based on the inclusion criteria. Six studies that focused on diseases of connective tissue, for example rheumatoid arthritis, were mentioned separately from other indications.

Table 2.

Studies of arthrodeses of the PIP joint of the finger with different techniques.

Reference	Year	LoE	MINORS score	Technique	Arthrodeses, n	Non-union, n	Consolidation (t)	Finger	IMM (t)	PROM (type, data)	Complications (Y/N)
Al-Qattan (9)1	2016	IV	11	Interosseus + K-wire	5	0	5 weeks	5× DII	NR	NR	N
Allende & Engelem (10)2	1980	IV	8	Tension-band	16	0	4–6 weeks	5× DII, 3× DIII, 4× DIV, 3× DV	Splint for discomfort	NR	Y – Lateral deviation; infection
Arata et al. (11)	2003	IV	9	Bioabsorbale rod	1	0	7.9 weeks	NR	3–4 weeks	NR	N
Ayres et al. (12)	1988	IV	10	Herbert screw	51	1	6 weeks	NR	2 weeks	NR	Y – 4× fracture dorsal cortex, 2× pain
Bansky & Racz (13)	2005	IV	3	Plate	2	0	NR	2× DII	NR	NR	N
Baruch & Kahanovich (14)3	1980	IV	4	Angulated bone peg	5	0	NR	NR	3 weeks	NR	N
Biskop (15)4	1985	IV	11	Tension-band	25	0	12 weeks	7× DII, 5× DIII, 9× DIV, 4× DV	–	NR	Y – 2× inflammation
Breyer et al. (16)	2015	III	10	Tension-band	24	2	9.4 weeks	NR	2–3 weeks	NR	Y – 5× superficial infection
				Compression screw	29	1	9.8 weeks	NR	2–3 weeks	NR	Y – 1× superficial infection, 1× deep infection
Buechler & Aiken (17)	1987	IV	10	Bone graft and plate	25	2	45–90 days	5× DII, 13 DIII, 6 DIV, 1 DV	NR	NR	Y – 1× infection
Buck-Gramcko & Oehme (18)5	1988	III	6	Interosseus + K-wire	84	NR	7 weeks	NR	NR	TAM	Y – 22× superficial infection, 3× osteoporotic fracture, three hardware failure
				Lag screw	6	NR	8.1 weeks	NR	NR	TAM	Y – 4× fracture dorsal cortex, 2× rotation, 1× tissue defect
				Tension-band	20	NR	8.2 weeks	NR	NR	TAM
				K-wires	8	NR	10.6 weeks	NR	NR	TAM
Burton et al. (19)	1986	IV	12	K-wires	34	0	9.2 weeks	NR	3–4 weeks	NR	Y – 2× delayed union, 1× arterial spasm
Carroll & Hill (20)	1969	IV	6	Cup/cone + K-wire	230	9	6–8 weeks	NR	6–8 weeks	NR	Y – 4× rotational error
Faithfull & Herbert (21)	1984	IV	4	Herbert screw	5	0	NR	NR	–	NR	N
Goth & Konigsberger (22)	1996	IV	9	Lag screw	23	0	7.5 weeks	NR	2 weeks	PS-100°	N
Harrison & Nicolle (23)	1974	IV	2	Harrison–Nicolle peg	35	1	NR	NR	2 weeks	NR	Y – 1× infection
Herzog (24)	1961	IV	5	Bone peg	11	0	8–12 weeks	NR	5 weeks	NR	N
Hoffmann & Rossack (25)	1975	IV	6	External fixation	10	0	NR	NR	5 days	NR	N
Høgh & Jensen (26)6	1982	IV	9	Interosseus + K-wire	23	NR	8 weeks	NR	6 weeks	NR	Y – 1× infection, 1× pain with amputation
Hohendorff et al. (27)	2016	IV	9	Tension-band	16	1	NR	5× DII 4× DIII, 1× DIV, 6× DV	6 weeks	Pain VAS, DASH, PS	Y – 1× infection
Jones et al. (28)	2011	III	8	K-wires	2	1	9 months	1× DII, 1× DIV	NR	MHOQ	Y – 2× malunion
				Tension-band	10	3	10 weeks	4× DII, 4× DIII, 2× DIV	NR
				Plate	1	1		1× DV	NR
Khuri (29)	1986	IV	8	Tension-band	10	0	6–8 weeks	4× DII, 2× DIV, 4× DV	7–10 days	NR	N
Kowalski & Manske (30)	1988	IV	10	K-wires	6	0	6–12 weeks	2× DII, 2× DIV, 2× DV	6 weeks	NR	N
Kvasnička (31)	2019	IV	8	External fixation	2	0	6.9 weeks	NR	NR	NR	N
Leibovic et al. (32)	1994	III	10	K-wires	99	21	10 weeks	49× DII, 63× DIII, 51× DIV, 61× DV	NR	NR	Y – 3x superficial infection, 1× osteomyelitis, 2× CRPS
				Tension-band	66	3	11 weeks
				Herbert screw	35	0	9 weeks
				Plate	4	2	12 weeks
Leonard & Capen (33)	1979	IV	8	External fixation	21	4	8 weeks	3× DII, 2× DIII, 6× DIV, 10× DV	8 weeks	NR	NR
Lewis et al. (34)	1986	IV	8	Cup/cone	6	0	NR	NR	8 weeks	NR	N
Lister (35)	1978	IV	8	Interosseus + K-wire	16	2	9.6 weeks	NR	–	NR	N
Martin (36)7	1981	III	10	Lag screw	89	9	7.1 weeks	NR	NR	NR	Y – 80× superficial infection, six fracture of dorsal cortex, 8× osteomyelitis, two breakage of wire
				K-wires	84	23	9.25 weeks	NR	NR	NR
				Tension-band	19	1	10.3 weeks	NR	NR	NR
				Interosseus + K-wire	4	0	10.7 weeks	NR	NR	NR
				Plate	1	0	6.2 weeks	NR	NR	NR
McGlynn et al. (37)	1988	IV	8	K-wires	28	0	6–12 weeks	NR	11× none, 14× 3–6 weeks, 3× 8–10 weeks	NR	Y – 2× superficial infection
Mikolyzk & Stern (38)	2011	IV	7	Interosseus + Steinmann Pin	5	0	NR	1× DII, 3× DII, 1× DV	6 weeks	NR	N
Moberg (39)	1960	III	6	K-wires	15	1	NR	NR	NR	NR	NR
				Bone peg	50	2	6 weeks	NR	6 weeks	NR
Netscher & Hamilton (40)	2012	IV	6	Bone peg + K-wires	1	0	6 weeks	1× DIV	6 weeks	NR	N
Newman et al. (41)	2018	IV	2	Compression screw	6	0	NR	NR	NR	NR	N
Novoa-Parra et al. (42)	2018	IV	10	Compression screw	6	0	8 weeks	3× DIV, 3× DV	–	DASH, pain VAS	N
Ono et al. (43)	2019	IV	8	Plate	1	1	8 weeks	1× DII	NR	ROM, Grip	N
Pellegrini & Burton (44)	1990	IV	7	K-wires	10	1	NR	5× DII, 2× DIII, 3× DIV	NR	Grip, pinch	N
Pfeiffer & Nigst (45)	1970	IV	6	Lag screw	7	0	6 weeks	NR	–	NR	N
Popova & Yankov (46)	1980	IV	6	Staples	10	0	6–8 weeks	NR	4 weeks	NR	N
Pribyl et al. (47)	1996	IV	9	K-wires	39	0	9 weeks	9× DII, 6× DIII, 14× DIV, 10× DV	8 weeks	NR	Y – 2× superficial infection, 1× mal-union
Prokes & Lutonsky (48)	2005	IV	7	External fixation	9	0	6.7 weeks	NR	–	NR	N
Reill & Renne (49)	1973	IV	5	Lag screw	20	0	NR	NR	NR	NR	N
Robertson (50)	1964	IV	2	Interosseus wiring	11	0	NR	NR	6	NR	N
Sabbagh et al. (51)	2001	IV	10	Harrison–Nicolle peg	20	1	NR	NR	2 weeks	NR	Y – 1× superficial infection, 1× deep infection
Sanderson et al. (52)	1991	IV	9	Harrison–Nicolle peg	36	2	NR	NR	2 weeks	NR	N
Savvidou & Kutz (53)	2013	IV	12	X-Fuse	2	0	8–12 weeks	2× DV	3 weeks	DASH	N
Seitz et al. (54)	1994	IV	8	External fixation	7	0	4–6 weeks	NR	–	NR	N
Stahl & Rozen (55)7	2001	IV	9	Tension-band	41	0	7–14 weeks	12× DII, 10× DIII, 8× DIV, 11× DV	4–6 days	NR	Y – 2× superficial infection, 3× pain
Strzyzewski et al. (56)	1971	IV	7	External fixation	10	0	NR	NR	6 weeks	NR	N
Tan et al. (57)	2017	III	12	Interosseus + K-wire	2	1	12 weeks	1× DIV, 1× DV	1 weeks	ROM, DASH	NR
				Lag screw	1	0	13 weeks	1× DIV	1 weeks
Taylor & Spencer (58)	1994	IV	8	Harrison–Nicolle peg	9	3	NR	2× DII, 2× DIII, 3× DIV, 2× DV	NR	NR	Y – 1× mal-union
Teoh et al. (59)	1994	IV	8	Lag screw	9	0	8.2 weeks	2× DII, 1× DIII, 2× DIV, 4× DV	–	NR	N
Uhl & Schneider (60)	1992	IV	8	Tension-band	32	1	12 weeks	3× DII, 6× DIII, 10× DIV, 13× DV	1 week	NR	Y – 1× infection
Vitale et al. (61)4,8	2015	III	10	Tension-band	9	1	NR	NR	NR	Grip, pinch, pain VAS, MHOQ	Y – 1× pain, 1× tendon adhesion
				Plate	3	0	NR	NR	NR
				Compression screw	2	0	NR	NR	NR
Reference	Year	LoE	MINORS score	Technique	Arthrodeses, n	Non-union, n	Consolidation (t)	Finger	IMM (t)	PROM (type, data)	Complications (Y/N)
Vorderwinkler et al. (62)4	2011	III	8	Tension-band	6	0	NR	NR	NR	NR	N
				External fixation	1	0	NR	NR	NR	NR	N
Wexler et al. (63)	1977	IV	8	External fixation	31	2	4–6 weeks	NR	1 week	NR	Y – 5× infection
Wright & McMurtry (64)	1983	IV	10	Plate	35	0	6 weeks	NR	–	NR	N
Wuestner et al. (65)	1986	IV	6	PDS peg	2	0	6 weeks	1× DII, 1× DIV	2–3 weeks	NR	NR
Zolotov (66)	2004	IV	6	Tension-band	6	0	NR	1× DII, 1× DIII, 2× DIV, 2× DV	2–3 weeks	NR	N
Indication rheumatoid inflammatory diseases
Belsky et al. (67)9	1982	IV	7	K-wires	50	0	NR	NR	NR	NR	N
Bracey et al. (68)10	1980	IV	12	Plate	24	0	NR	NR	–	NR	N
Gilbart et al. (69)11	2004	IV	13	Tension-band	13	0	NR	1× DII, 3× DIII, 4× DIV, 5× DV	NR	NR	Y – 3× local irritation
Granowitz & Vainio (70)10	1966	IV	9	K-wires	122	8	6 weeks	19× DII, 29× DIII, 43× DIV, 31× DV	–	NR	Y – 3× infection
Jones et al. (71)11	1987	IV	10	Interosseus + K-wire	53	3	6–8 weeks	NR	6 weeks	NR	N
Lipscomb et al. (72)11	1969	IV	6	K-wires	16	0	5–8 weeks	3× DII, 4× DIII, 4× DIV, 5× DV	NR	NR	Y – 5× superficial infection

¹ 1x only Kwire; ²Unclear description of complications, might have occurred ad thumb; ³1x additional Kwires; ⁴Personal communication with author; ⁵Non-union not differentiated between PIP- and other arthrodesis as well as techniques; ⁶Non-union and complication not differentiated between PIP- and other arthrodeses; ⁷Complications not differentiated between PIP- and other arthrodesis; ⁸Complications not differentiated between different methods; ⁹Psoriac arthritis; ¹⁰Rheumatoid arthritis; ¹¹Systematic sclerosis.

DASH, disabilities of Arm, Shoulder and Hand Questionnaire; DII, index finger; DIII, long finger; DIV, ring finger; DV, little finger; IMM, immobilization; LoE, level of evidence; MHOQ, Michigan Hand Outcomes Questionnaire; nr, not reported; PROM, patient-reported outcome measure; PS, palm spacing; ROM, range of movement; TAM, total active movement; VAS, visual analogue scale.

Data extraction

Data were extracted from the included studies by two authors independently (MM and HV) according to a predefined data extraction sheet. The level of evidence, quality and risk of bias assessed with the standardized critical appraisal instrument, Methodological Index for Non-Randomized Studies (MINORS) score, where applicable, were recorded (8). The methodological quality score MINORS shows a mean of 8 with a global ideal score of 16. Fifty-five articles had level IV evidence, and nine articles had level III evidence. Nearly all studies were retrospective data analysis. We extracted the number of PIP joint arthrodesis, the technique used, time of immobilization, number of non-unions, time to radiological consolidation, and the incidence and type of complications. All patients regardless of their age with arthrodesis were included in this review.

Results

Included studies

A total of 1923 arthrodeses of the PIP joint could be extracted from the included papers (Table 2). The main indications for the arthrodesis of the PIP joint were primary or secondary osteoarthritis, joint infection or traumatic destruction. Included are six studies that consisted only of patients with rheumatic disease, for example rheumatoid arthritis or systemic sclerosis. These results are presented separately in Table 2.

Surgical techniques

Twelve different surgical techniques were described. Four of these techniques with compression at the arthrodesis site: interosseus wiring with/without K-wire, tension-band, cannulated screw as well as a lag screw – combined a total of 805 arthrodeses. The plate, external fixation and K-wire might hold some applied compression during the arthrodesis but do not hold any compression potential themselves. The most frequently used techniques were, with the number of arthrodesis in descending order, K-wires (n = 743, 14 studies), tension-band (n = 313, 15 studies) and compression screws (n = 233, 12 studies). The included studies cover a time span of 74 years of publication, and that there is an obvious trend towards techniques with compression of the arthrodesis, especially with compression screws.

Non-unions and mean consolidation times

Non-unions were reported in all studies. Two studies included other finger joints besides the PIP and did not report the exact numbers of non-unions concerning the fused joint. In these cases, the studies were only included for the consolidation time, for they reported that explicitly. The lowest non-union rate in compression techniques was 3.9% with the compression screw. Interestingly, the non-union rate for the peg fixations (without compression) was even lower 3.6%. The highest non-union rate showed the interosseous wiring with 8.6% (Table 3).

Table 3.

Amount of non-union joint arthrodeses because of osteoarthritis by technique – only studies which described the number of non-unions of the PIP joint with respective technique were included.

Technique	Studies (n)	Individuals (n)	Non-union (n(%))
Tension-band	14	293	12 (4.1)
K-wires	13	735	64 (8.7)
Compression screw	12	282	11 (3.9)
Interosseus wiring	8	105	9 (8.6)
Pin fixation	9	102	6 (5.9)
Peg fixation	8	165	6 (3.6)
Plate	6	93	4 (4.3)
Total	70	1775	112 (6.3)

Table 4 depicts the mean consolidation times. Further information on how non-uniions were stratified by technique is presented in the Supplementary information and the results are presented in supplementary figures 1 and 2.

Table 4.

Consolidation time by technique – only studies included with joints affected by osteoarthritis. The table depicts the consolidation times (mean ± s.d.) in weeks. Again, there were no statistically significant differences between any analyzed technique in comparison to K-wires. Also, we made a comparison of compression vs non-compression techniques of the mean consolidation time, without statistically significant difference (P = 0.830).

Technique	Studies (n)	Individuals (n)	Consolidation times (weeks)
Tension-band	10	263	9.5 ± 2.2
K-wires	11	668	8.6 ± 1.5
Compression screw	9	255	7.7 ± 1.3
Interosseus wiring	7	187	8.5 ± 2.4
Pin fixation	7	82	6.9 ± 1.7
Peg fixation	3	63	7.3 ± 2.3
Plate	3	64	9.2 ± 3.0
Total	50	1582	8.2 ± 2.0

Complications

Four studies did not describe complications. All others either stated that they had no complications or did not describe them in detail. Most complications besides the non-unions were infections (mostly superficial), pain caused by the implant or mal-unions. The consequences of these complications, that is, if revisionary surgery had to be performed or if superficial infections could be treated by antibiotics, were not reported.

Discussion

A wide range of different surgical techniques for achieving fusion of the PIP joint have been published. Moberg already stated in 1960 that ‘the prime requisite of a good digital arthrodesis is a painless and stable union in proper position occurring in a reasonable space of time’ (39). Nevertheless, a proper comparison, although needed, proves to be difficult because of the variable quality of published studies, different indications for joint fusion, varying definitions of consolidation (radiological vs clinical) as well as lacking important data in large but older studies, where a personal communication with the author is no longer possible (73).

The two main groups of joint fusion techniques which can be differentiated, are techniques with and without compression of the arthrodesis site respectively (6). The most important advantage of the compression is the assumed shorter consolidation time because of higher primary stability, consolidation by primam intentionem with fewer non-unions as well as early functional occupational therapy (60, 64). In this systematic review, the assumption that techniques with compression are more reliable, as demonstrated by Leibovic in 1994, could not be clearly proven (32). One possible reason might be that the compression techniques are surgically more difficult and might tend to non-union if there are no ideal operative results. For example, there is the possibility that a tension-band fusion does not apply the compression to the whole arthrodesis site and therefore renders it unstable. The compression screw however might be easier and more forgiving to implant than tension-band or intraosseus wiring. That might be the reason why the superiority of this implant in contrast to K-wires is evident in different studies in the literature (32, 36).

Nevertheless, in the studies included in this systematic review, there is a trend towards techniques with compression over the course of time, especially towards compression screws (41, 42). With further development of the implants, the diameter of the screws got progressively smaller, as 8 mm diameter screws are commercially available now. Thus, these days they can be used in small bones too.

Newer implants like the Apex IP fusion device so far lack any evidence that they are easier to implant or provide a better outcome, maybe because they have not been available in the market long enough (42).

The most reported complications besides the primary outcome of non-unions were infection, mostly superficial. As there is typically very little soft tissue around the PIP joint, protruding implants, like a tension-band, can cause irritation and subsequently a superficial infection. This emphasizes the need for a proper handling of soft tissues (36).

Rheumatoid arthritis and connective tissue diseases

Rheumatoid inflammatory diseases commonly affect the joint, especially the PIP joint, which may lead to contractures and deviations that are both disabling as well as cosmetically unacceptable (74). These diseases could affect the quality of the bones and therefore the stability of arthrodeses as well as the healing of soft tissues. The referenced papers by Gilbart et al. (69), Jones et al. (71)and Lipscomb et al. (72) relate to patients with systemic sclerosis. From a pragmatic point of view, one might state that if something works for this challenging group of patients it will probably work for a patient with osteoarthritis. Interestingly and somewhat counterintuitively, Lipscomb et al. (72) found quicker healing compared to other studies dealing with posttraumatic osteoarthritis.

Biomechanical properties and primary stability

The primary stability of different fusion techniques or implants could provide an interesting insight into the ability of the implant itself to withstand the forces of early function therapy as well as a short or even no immobilization. There are only few papers that have tried to compare the results of different biomechanical studies (75, 76). Therefore, it seems reasonable to conduct a biomechanical study for comparing the different implants and techniques of interphalangeal arthrodesis so that postsurgical treatment can be standardized.

Strengths and limitations

There are several limitations of the existing literature as well as of this study. In order to do a reasonable meta-analysis and statistical evaluation of the different techniques, randomized controlled trials (RCT) are required. On the topic of arthrodesis of the PIP joint, there is no RCT published at all. Therefore, we did a qualitative systematic review with only descriptive data pooling of the different studies with respect to their published technique for greater clearness instead of a meta-analysis. Another limitation is that the literature review for this systematic review showed that there are predominantly studies with an evidence level IV, with a heterogenous MINORS score but a satisfying mean of 8. As the risk of bias as depicted in the MINORS score exists, one might suspect that the published technique makes the apparent effect appear better than it is. There were nine evidence level III studies which could be included. Nevertheless, these results imply a lack of good quality data to statistically compare the different techniques and to achieve recommendations.

Especially the complications of different techniques, which we extracted from the studies, were reported very heterogeneously with no clear evidence on how to avoid them or of their consequences.

Strengths of this systematic review is its novelty and uniqueness, since there are no systematic reviews with a high quality, like PRISMA methodology. It includes a very long-time span of nearly 74 years and covers the most extensive databases. A very large number of abstracts were screened to achieve the most complete systematic review.

Conclusion

The compression screw shows superior results with respect to non-unions in comparison to K-wires. There is a tendency of more published techniques with compression in the last 10 years which might implicate a shift towards compression techniques. Given the limited evidence of the available studies on arthrodesis of the PIP joint, there is a lack of clear indications for other special techniques. The three most often used techniques are K-wires, tension-band and compression screws. The K-wires still have their place in acute trauma with soft tissue defects or replantation. Only large multi-center RCTs can answer the question on which technique for arthrodesis of the PIP joint is the best.

ICMJE Conflict of Interest Statement

The authors declare that there is no conflict of interest that could be perceived as prejudicing the impartiality of the work reported here.

Funding Statement

This work did not receive any specific grant from any funding agency in the public, commercial or not-for-profit sector.

Author contribution statement

H-C Vonderlind and M Ruettermann: both authors contributed equally to this manuscript.