Implants for fixation of intertrochanteric femoral fracture: a systematic review and network meta-analysis of randomized controlled trials

Abstract

Background

A variety of implant devices have been used for treatment of intertrochanteric femoral fractures (IFF), but the optimal has long been disputed. We aim to summarize the latest evidence for the effectiveness and safety of implants for IFF.

Methods

This systematic review and network meta-analysis included searches of PubMed, Embase, the Cochrane Library, and Web of Science from January 1, 2000 to August 31, 2024, for randomized controlled trials of implants in older adult patients with intertrochanteric femoral fracture. Non-English studies, pathological fractures, pathological reports, animal studies, conference abstracts, and incomplete primary were deemed ineligible. We performed frequentist random-effect network meta-analyses to summarize the evidence and applied the Confidence in Network Meta-Analysis frameworks to rate the certainty of evidence, calculate the treatment effects, categorize interventions, and present the findings. The study was registered with PROSPERO, CRD 42,024,562,020.

Results

A total of 54 eligible trials were identified, involving 15 implants and enrolling 10,275 participants; all subsequent estimates refer to the comparison with sliding hip screw (SHS). InterTAN nail (ITN) resulted in the largest reduction in non-mechanical major post-surgery complications (OR, 0.55; 95% CI, 0.33 to 0.91; moderate confidence). No significant differences were found in terms of Harris hip score, reoperation rate, and overall mechanical complications with moderate to low-level evidence. In secondary findings, percutaneous compression plate (PCCP) resulted in the lowest occurrence of non-mechanical minor post-surgery complications (OR, 0.12; 95% CI, 0.05 to 0.30; high confidence), and proximal femoral nail anti-rotating (PFNA) (OR, 0.05; 95% CI, 0.02 to 0.11; high confidence) resulted in most reduced non-specific mechanical complications, respectively and. ITN demonstrated the highest risk of operative issues (OR, 3.41; 95% CI, 2.03 to 5.73; moderate confidence).

Conclusions

In older patients with intertrochanteric fractures, ITN proved among the most effective in reducing non-mechanical major post-surgery complications, but had the highest risk of intraoperative complications. No implants demonstrated superior effectiveness over others.

Registration of systematic reviews

CRD 42021245678, PROSPERO.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12891-025-09032-w.

Introduction

Intertrochanteric femoral fractures represent a significant global public health challenge, with an estimated 1.3 million cases occurring each year worldwide [1, 2]. The burden of IFF is compounded by high mortality rates, chronic morbidity, and substantial declines in functional ability and quality of life [3, 4]. Surgery was established as the standard treatment for IFF in the 1950 s, and since then, advancements in minimally invasive surgical techniques, geriatric orthopedic management, and the development of implant devices have continued to evolve [5]. Nevertheless, postoperative complications persist as a considerable challenge, and among these, implant-related complications are primary causes for unplanned readmissions, reoperations, functional impairment, and loss of independent ambulation [6, 7]. The ongoing optimization of implant design remains a critical focus in both research and clinical practice.

The 2022 guideline from the American Academy of Orthopaedic Surgeons and the 2023 guideline from National Institute for Health and Care Excellence both provide a strong recommendation for SHS for patients with stable IFFs and Cephalomedullary Device for patients with unstable IFFs, primarily based on evidence from cohort studies or randomized controlled trials (RCTs) [8, 9]. A 2020 network meta-analysis of 36 RCTs found no significant differences in overall postoperative complications among DHS, CHS, PCCP, Medoff sliding plate, LISS, Gamma nail, PFN, and PFNA, however, the reliability of the results was affected by the limited number of studies included and the inconsistencies in certain aspects (Harris hip score and operative time) [10]. Furthermore, two meta-analyses addressing the comparison of PFNA with ITN, published in 2020 and 2024 respectively, showed the conflicting results, and neither have assessed the quality of supporting evidence [11, 12]. Since then, several new types of implants, such as PFNA-II [13] and Zimmer natural nail (ZNN) [14] have been introduced into clinical practice, yielding inconsistent results. Existing reviews are inadequate in providing timely and relevant support for patients, clinicians, and researchers, highlighting an urgent need for up-to-date and comprehensive synthesis of evidence.

We conducted a systematic review and network meta-analysis (NMA) of available RCTs to summarize the latest evidence on the different implants in patients with IFF.

Methods

The protocol for this systematic review and network meta-analysis was registered with PROSPERO. This study followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses 2020 (PRISMA, Supplementary Information 1, Supplementary Information 2), extension statement for network meta-analyses (PRISMA-NMA) and the Assessing the methodological quality of systematic reviews (AMSTAR, Supplementary Information 3) [15–18].

Search strategy and selection criteria

Two researchers searched PubMed, Embase, the Cochrane Library, and Web of Science for RCTs of implants in patient with IFF from January 1, 2000 to August 31, 2024. The search strategy consisted of three core components connected using the AND operator: (1) RCT (randomized controlled trial), (2) implants (e.g. internal fixators, screw), and (3) intertrochanteric femoral fracture (e.g. intertrochanteric fracture*). This search strategy, which initially targeted PubMed, was later modified for the other three databases, employing a mix of keywords and free text. The language was limited to English, and no restrictions in terms of gender, ethnicity, country, or setting were applied. Additionally, we assessed the references cited in the included literature and two relevant systematic reviews [19, 20] (Supplementary Information 4, Appendix 1).

We included randomized controlled trials that compared any of the following implants for treating IFF: sliding hip screw (SHS), dynamic hip screw (DHS), percutaneous compression plate (PCCP), Gamma nail (GN), proximal femoral nail (PFN), and proximal femoral nail anti-rotating (PFNA), compression hip screw (CHS), Medoff sliding plate (MSP), Zimmer natural nail (ZNN), intramedullary hip screw (IMHS), the InterTAN nail(ITN), Targon PFT cephalomedullary nail(TPFT), the ACE trochanteric nail(ACE), AMBI hip screw(AMBI), and ENDOVIS nail(ENDOVIS). Studies involving pathological fractures, pathological reports, animal studies, conference abstracts, and incomplete primary data were excluded.

Two reviewers independently screened the eligible studies. First, the retrieved literature from the database was imported into Endnote X21 and duplicate items were removed. Next, the reviewers evaluated the eligible studies based on title and abstract, including any ambiguous studies for additional review. Ultimately, a comprehensive evaluation of the potentially eligible articles was performed according to established inclusion and exclusion criteria.

Data extraction and synthesis

A pre-designed table was employed to systematically extract comprehensive information pertinent to each study, including study characteristics (i.e., country, journal, author, funding, and publication year), population demographics (i.e., baseline age, gender, comorbidities, fracture type and stability, medication, sample size, losses and reasons for losses), and interventions (i.e., name, follow-up duration, surgical time, American Society of Anesthesiologists (ASA) status, and intraoperative blood loss)(Supplementary Information 4, Appendix 2, 3 and 10).

The outcomes included Harris hip score (HHS), mortality, intraoperative complications, non-mechanical complications, mechanical complications, and reoperation rate. We recorded the outcomes by timing, and further categorized—intraoperative complications as operative issue and non-specific intraoperative complications; non-mechanical complications as post-surgery medical complications, major post-surgery complications, and minor post-surgery complications; and mechanical complications as cut out, fixation failure, migration, valgus deformity, and non-specific mechanical complications. We formed a multidisciplinary panel comprised of two orthopedists, one geriatric physician, one rehabilitation physician, one pharmacist, one anesthesiologist, one radiologist, and one nurse to judge the following outcomes. We prioritized the outcomes as crucial——HHS, reoperation rate, major post-surgery complications, and overall mechanical complications; important but not crucial——mortality, intraoperative complications, individual mechanical complications (i.e., cut out, fixation failure, migration, valgus deformity, and non-specific mechanical complications), and non-mechanical postoperative complications (i.e., medical complications, minor post-surgery complications). In case of incomplete data, efforts were made to contact the corresponding author. Any discrepancies were resolved through discussion.

For each study, we collected information on participant dropout rates for each group. We assumed that complete case data included mortality, unexpected withdrawal from the study due to relocation, etc., and adverse events. For outcomes that require assessment of participants at the end of follow-up (such as HHS), we prioritized studies with intention-to-treat (ITT) data. If ITT data were not available for these outcomes and the study authors did not report the denominator numbers for each group, we reduced the denominator numbers for each group to account for reported dropout rates.

We used the Cochrane Randomized Trial Bias Risk Tool to assess the methodological quality of the included studies [21], The domains to be assessed included random sequence generation, allocation concealment, blinding of participants and personnel, blinding of outcome assessment, incomplete outcome data, selective reporting, and other biases. The risk of bias in each domain was classified as ‘low’ ‘high’ or ‘unclear’ (Supplementary Information 4, Appendix 4).

Assessing the confidence of evidence

We utilized the Confidence in Network Meta-Analysis (CINeMA) tool to evaluate the certainty of the evidence [22, 23]. This assessment incorporated six key domains: within-study bias, reporting bias, indirectness, imprecision, heterogeneity, and inconsistency. Each of these areas was assessed for levels of concern categorized as no concerns, some concerns, or major concerns. If any domain exhibited serious or very serious concerns, the overall quality of the evidence was downgraded by one to two levels. Ultimately, the final assessment classified the quality of evidence into four categories: high, medium, low, or very low. Any discrepancies were resolved through consensus. The analysis was carried out using the Revman5.4.1 and CINeMA websites (http://cinema.ispm.unibe.ch/#general).

Statistical analysis

Network meta-analysis was conducted within the frequentist framework using Stata version 16.0 through the network command, with corresponding 95% confidence intervals, odds ratios for dichotomous outcomes, mean differences for continuous outcomes. Random-effects models were chosen as they are considered the most appropriate and conservative approach to account for variability among RCTs. Global consistency was evaluated using a global test, and local inconsistency was assessed using node-splitting methods. There was no significant inconsistency when P value > 0.05 for the comparison between direct and indirect effects. We used the ggplot2 package in R to create cumulative probability plots and ranked the implants based on their surface under the cumulative ranking (SUCRA) values to evaluate their effectiveness in treating IFF (Supplementary Information 4, Appendix 6, 7 and 15). The larger the value, the higher the probability.

Considering the reported subgroup effects of fracture stability in the literature, we conducted a separate analysis for patients with unstable intertrochanteric fractures (UIF) and compared the relative effect of extramedullary (eg, SHS) vs intramedullary implants on the risk of complications. The Cochran’ Q test and τ2 values were used to assess heterogeneity, categorizing them as low (< 0.04), low-moderate (0.04–0.16), moderate-high (0.16–0.36), and high (> 0.36) [24]. In cases of high heterogeneity, regression analysis was conducted based on age, gender, publication year, and surgery time. Transitivity was examined by comparing the similarity of included populations based on average age, gender distribution, fracture type, fracture stability, surgery time, and intraoperative blood loss. Funnel plots were employed to evaluate the potential impact of small-scale studies when at least 10 trials were included in a comparison, with parallel Egger tests conducted [25]. A P value > 0.05 indicated no publication bias.

Role of the funding source

The funder had no role in study design, data collection, analysis, and interpretation, or writing of the manuscript and the decision to submit.

Results

Literature selection and study characteristics

A total of articles 5,771 were screened, and 249 full-text articles were evaluated. Four additional trials were identified from the reference lists of relevant papers and reviews. In total, we included 54 RCTs involving 10,275 patients, conducted across 23 countries and regions (Fig. 1). The average age of participants at enrollment was 80.23 years (with 4 RCTs not reporting), and the proportion of males was approximately 29.15% (with 2 RCTs not reporting). Sample sizes ranged from 27 to 1,000 individuals, and the follow-up duration ranged from 10 days to 30 months. Table 1 presents a comprehensive list of included studies along with their characteristics. We classified CHS, DHS, and AMBI as SHS.

Fig. 1.

Flow diagram of preferred reporting items identified, included, and excluded for systematic reviews and meta-analyses (PRISMA)

Table 1.

Characteristics of 54 studies included in the network meta-analysis

Year	Country	Follow-up period	Journal	Funding	Randomised treatments	Number of dropouts	Age (mean ± SD), years	Male %	Outcomes	Surgical time (mean ± SD), min
Schemitsch 20231	Canada	1 year	JAMA Netw Open	Stryker	GN:423	56	78.2	36.6	(1); (2); (3); (4); (5);	59 ± 27
					SHS:427	49	78.8	33.3		64 ± 23
Parker20122	UK	1 year	J Bone Joint Surg Br	No	SHS:300	4	81.4	23	(1); (2); (3); (4); (5);	46 ± 12.3
					TPFT:300	2	82.4	17		49 ± 12.7
Schipper 20043	Netherlands	1 year	J Bone Joint Surg	Stryker Howmedica Mathys Medical Nederland	PFN:211	7	82.2	16	(1); (2); (3); (4); (5); (6);	60 ± 2
					GN:213	1	82.6	15.4		60 ± 2
Papasimos 20054	Greece	1 year	Arch Orthop Trauma Surg	NA	SHS:40		81.4	35	(1); (2); (3); (4); (5);	59.2
					GN:40		82.8	40		51.3
					PFN:40		79.4	42.5		71.2
Adams20015	Scotland	1 year	J Orthop Trauma	the Scottish Orthopaedic Research Trust - into Trauma	GN:203	18	81.2	19.2	(1); (2); (4); (5);	55.4
					SHS:197	15	80.7	24.9		61.3
Chenchik 20146	Israe	1 year	J Orthop	NA	DHS:31	0	83.1 ± 6.7	26	(2); (3); (4); (5);	64 ± 26
					PFN:29		83.1 ± 5.7	21		54.5 ± 22.5
Verettas 20107	Greece	10d	Injure	NA	GN:59	0	79.22 ± 7.99	33.9	(4); (6)	42
					DHS:59	0	81.03 ± 6.38	25.4		45
Utrilla20058	Spain	12–30 months	J Orthop Trauma	None	GN:104	3	80.6 ± 7.5	36.5	(1); (2); (4); (5);	46 ± 11
					SHS:106	4	79.8 ± 7.3	26.4		44 ± 15
ZOU20099	China	1 year	J Int Med Res	NA	PFNA:58		65	21	(1); (2); (4); (5);	52 ± 10
					SHS:63		65	24		93 ± 13
Efstathopoulos200710	Greece	6 − 12 months	Int Orthop	NA	GN:56	5	79.5	33.9	(4);	51 ± 11
					ACE 56		78.1	23.2		54 ± 15
Zehir201411	Turkey	6 months	Eur J Trauma Emerg Surg	NA	SHS:102	0	76.86 ± 6.74	0.382	(1); (2); (4); (5);	56.95 ± 5.20
					PFNA:96	0	77.22 ± 6.82	0.385		44.41 ± 5.17
Pajarinen 200412	Finland	4 months	J Bone Joint Surg	None	SHS:41		79.0 ± 11.5	26.8	(4);	45
					PFN:42		80.2 ± 9.4	19		55
Miedel 200413	Sweden	12 months	J Bone Joint Surg	Stryker Howmedica, Sweden (SGN), and Swemac, Sweden (MSP), the Trygg-Hansa Insurance Company, the Swedish Orthopaedic Association	GN:109	3	84.6	16	(1); (2); (3); (4); (5);	61
					MSP:108	3	82.7	22		65
Ahrengart 200214	Sweden	6 months	Clin Orthop Relat Res	the The Karolinska Institute Foundation, Lund University, Skane County Council and Stryker-Howmedica	GN:210	0		29	(1); (2); (3); (4); (5);	60
					SHS:216	0		28		45
Xu201015	China	1 year	J Int Med Res	NA	PFNA:51	0	78.5 ± 7.97	29	(1); (2); (3); (4); (5);	68.5 ± 9.9
					DHS:55	0	77.9 ± 7.82	29		56.5 ± 11.8
Sharma 201816	India	1 year	Rev Bras Ortop	NA	SHS:29	0	62.27	39.4	(1); (2); (3); (4); (5); (6);	69.7
					PFN:31	0	60.67	37.9		56.9
Parker201717	UK	1 year	Injury	Peterborough Hospitals Hip Fracture Research Fund	SHS:500	6	82.1	23.2	(2); (4); (5);	44 ± 11.9
					Targon PF intramedullary nail:500		82.2	22.4		45 ± 12.8
SADOWSKI200218	Switzerland	1 year	J Bone Joint Surg	None	95° fixed-angle screw-plate (Dynamic Condylar Screw):19	1	77 ± 14	26.3	(1); (2); (4); (5);	166 ± 48
					PFN:20	0	80 ± 13	35		82 ± 53
Barton 201019	UK	1 year	J Bone Joint Surg	None	GN:100	3	83.3	22.7	(1); (2); (4);
					SHS:110	0	83.1	19
Kumar 201220	India	at least 1 year	J Clin Orthop Trauma	None	PFN:25	0			(1); (2); (3); (4); (5); (6);	55 ± 18
					SHS:25	0				87 ± 3.2
Xu201021	China	17.5 months	Injury	NA	PFNA:55	5	76.8 ± 9.6	41.8	(1); (2); (3); (4); (5);	66.6 ± 15.4
					GN:52	6	76.6 ± 8.2	28.8		73.1 ± 20.8
Guerra 201422	Brazil	1 year	Injury	Nonene	SHS:19	0		26.3	(1); (4);
					PFN:12	0		8.3
Cheng201423	China	6–26 months	Chin J Traumatol	NA	PCCP:65	0	72.4 ± 4.18	43.1	(1); (2); (4); (5); (6)	55.2 ± 8.4
					SHS:56	0	78.6 ± 3.92	39.3		88.5 ± 14.7
Jose´200924	Spain	1 year	Arch Orthop Trauma Surg	NA	GN:40	0		15	(1); (4); (5);	85.82
					PCCP:40	0		27.5		86.51
Philip201225	Belgium	1 year	Acta Orthop. Belg	NA	GN:61	6	73 ± 12.5	57.4	(1); (4); (5);	41
					ACE:51		77 ± 14	62.7		51
Parker201726	UK	1 year	Bone Joint J	Peterborough Hospital Hip Fracture Fund	SHS:200	1	83.2	23.5	(1); (2); (3); (4); (5);	42.1 ± 11.1
					Targon PFT cephalomedullary nail:200	1	82	30		38.3 ± 10.2
Hopp201627	Germany	6 months	Acta Orthop. Belg	None	GN:39	1	80.73 ± 8.44	33.3	(1); (2); (4); (5); (6)	64.6 ± 29.22
					the InterTAN nail (ITN):39	1	82.70 ± 7.06	17.9		78.03 ± 34.07
Janzing 200128	Belgium	1 year	J Trauma	NA	SHS:44		83	23	(2); (3); (4);	65 ± 30
					PCCP:39		82	10		49 ± 18
Carulli 201729	Italy	1 year	Clin Cases Miner Bone Metab	NA	PFNA: 71	3	81.62 ± 7.82	40.9	(1); (2); (4); (5);	46.06 ± 10.10
					SHS:69	3	83.41 ± 7.90	36.2		61.21 ± 15.01
Guo201330	China	12–24 months	J Orthop	NA	PCCP:45	0	71.6 ± 7.5	35.6	(1); (4); (6);	53.0 ± 9.4
					PFNA:45	0	74.2 ± 8.8	42.2		66.5 ± 18.1
Ioannis 201431	Greece	12 months	Int Orthop	NA	SHS:35	0	83.1 ± 6.5	20	(4); (5);	75.5 ± 21.9
					GN:36	0	82.9 ± 5.8	22.2		45.7 ± 22.7
Makridis 201032	Greece	12 months	J Orthop Surg Res	NA	intramedullary hip screw (IMHS):110	0	83.5	30.9	(1); (3); (4); (5);	25.4
					ENDOVIS nail:105	0	83.9	31.4		24.8
Kouvidis 201233	Greece	24–56 months	Strat Traum Limb Recon	NA	SHS:79	3	82.53 ± 6.79	34.6	(1); (2); (3); (4); (5);	55.18 ± 11.50
					Endovis:86	5	81.95 ± 7.21	20		51.22 ± 12.94
Peyser 200734	Israel	1 year	J Bone Joint Surg	None	PCCP:50	0	78.9	32	(1); (2); (4); (5);	53.2
					SHS:53	0	82.5	34		51.1
McCormack201335	Canada	6 months	Injury	NA	SHS:86	5	83	24.4	(1); (2); (4);	51
					MSP:77	4	83.6	23.3		50
Vaquero 201236	Spain	12 months	Injury	aofoundation和Synthes, GmbH, Switzerland	GN:30	15	83.5 ± 7.4	17	(1); (2); (3); (4); (5); (6);	37 ± 10
					PFNA:31	18	83.6 ± 7.5	10		35 ± 10
Matre201337	Norway	12 months	J Bone Joint Surg	Smith & Nephew	the InterTAN nail (ITN):341	53	84.1	24.3	(1); (2); (3); (4); (5);
					SHS:343	54	84.1	25.7
Zhang201338	China	12–30 months	Orthopedics	NA	PFNA:56	3	72.4 ± 8.7	33.9	(1); (2); (3); (4); (5); (6);	53.7 ± 11.3
					the InterTAN nail (ITN):57	2	72.9 ± 7.6	40.4		66.5 ± 15.2
Seyhan 201539	Turkey	1 year	J Orthop	NA	PFNA:43		75.91 ± 13.77	25.6	(2); (4); (6)	72.98 ± 12.48
					the InterTAN nail (ITN):32		75.34 ± 13.52	25.9		73.91 ± 8.87
Brandt 200240	Netherlands	1–8 months	Injury	NA	PCCP:33	0	80.1		(1); (2); (3); (4); (5);	46.6
					SHS:38	0	81.6			69.2
Sanders 201541	Canada	12 months	J Orthop Trauma	NA	SHS:126	13	81.0 ± 0.8	26	(1); (2); (4); (5);
					the InterTAN nail (ITN):123	3	80.6 ± 0.8	29
Kosygan 200242	England	6 months	J Bone Joint Surg	NA	PCCP:52	0	82.7 ± 8.5	15.4	(1); (4); (5);	82.7 ± 8.5
					SHS:56	0	82.8 ± 9	21.4		82.8 ± 9
LI201843	China	18 months	Eur Rev Med Pharmacol Sci	NA	PFNA:40	0	75.6 ± 2.5	50	(4); (5); (6);	38.5 ± 5.7
					SHS:40	0	75.5 ± 2.6	52.5		43.6 ± 9.0
Adeel202044	Pakistan	12 months	J Pak Med Assoc	None	SHS:34	0	60.88 ± 12.49	64.71	(1); (4); (5);	58.71 ± 7.84
					PFN:34	0	59.32 ± 2.39	73.53		35.35 ± 5.48
OVESEN 200645	Denmark	12 months	Hip Int	NA	SHS:73	11	78.5 ± 11.7	28.8	(1); (2); (4); (5);	51 ± 22
					GN:73	8	79.9 ± 10	27.4		65 ± 29
Prakash 202246	India	24 weeks	Cureus	None	SHS:23	0	61.09 ± 11.69	39.13	(1); (4); (6);
					PFN:23	0	65 ± 14.98	47.83
Olsson 200147	Sweden	4 months	J Bone Joint Surg	None	MSP:54	8	84 ± 7.5	25.9	(1); (2); (4); (5);	58
					SHS:60		84 ± 7.3	33.3		55
Sameer 201748	India	12–30 months	J Clin Diagn Res	None	SHS:19	0	71.74	63.1	(1); (2); (4);	104.2 ± 33.72
					PFN:18	0	74	66.7		106.2 ± 26.31
Harrington 201949	USA	1 year	Injure	NA	SHS:52	0	82.1 ± 8.6	21.2	(1); (3); (4); (5);	88 ± 27.5
					IMHS:50	0	83.8 ± 8.5	20		108 ± 26.8
Singh201950	India	1 year	J Clin Orthop Trauma	NA	PFNA:30	3	72.76 ± 9.5	30	(1); (2); (4); (5); (6);	54.66 ± 19.20
					SHS:30	1	69.33 ± 5.7	53.3		71.1 ± 24.81
XU202051	China	12 months	Orthopedics	NA	PFN:66	28	76.0 ± 1.2	37.9	(2); (3); (4); (5);	64.1 ± 1.9
					GN:70		75.4 ± 1.0	38.6		68.6 ± 2.4
Saudan 200252	Switzerland	1 year	J Orthop Trauma	None	SHS:106	4	83.7 ± 10.1	20.8	(1); (2); (4); (5);	65 ± 26
					PFN:100	5	83.7 ± 9.7	24		64 ± 33
Shin201753	South Korea	12.3 months	Injure	None	Zimmer natural nail (ZNN):172		76.22 ± 16.40	63	(2); (6); (4); (5);	70.40 ± 20.10
					PFNA:181		77.71 ± 16.44	70		57.25 ± 12.15
Ekstro¨m200754	Sweden	12 months	J Orthop Trauma	NA	PFN:105	26	82	24	(1); (2); (3); (4); (5);	56 ± 21
					MSP:98	24	82	25		62 ± 29

NA: not available

(1) mortality; (2) reoperation rate; (3) intraoperative complications;(4) postoperative complications; (5) mechanical complications; (6) Harris hip score (HHS)

Risk of bias, confidence of evidence, transitivity, and consistency

Out of 54 experiments, 3 studies (5%) were found to have a high risk of bias. Local inconsistencies were found during mortality, reoperation rate, post-surgery medical complications, overall mechanical complications, but there is no strong statistical evidence to suggest global inconsistency (Supplementary Information 4, Appendix 5). The study population was similar in terms of age, gender distribution, fracture type, fracture stability, surgery time, and intraoperative blood loss. After evaluating the level of evidence using CINeMA, most paired comparison results were found to have moderate confidence (Supplementary Information 4, Appendix 8). The tau² results showed low heterogeneity, except for HHS, mechanical complication-fixed failure, overall mechanical complications, and non-specific intraoperative complications. Furthermore, we did not observe asymmetry in the funnel plot (Supplementary Information 4, Appendix 13).

Crucial outcomes

Evaluated by HHS, the network meta-analysis enrolled 13 RCTs with 1,227 subjects, and all pairwise comparisons were not statistically significant (Fig. 2).

Fig. 2.

Network meta-analysis of different implants for HHS. HHS= Harris hip score; SHS=sliding hip screw; PFNA=proximal femoral nail anti-rotating; PFN=proximal femoral nail; PCCP=percutaneous compression plate; ITN=InterTAN nail; GN=Gamma nail; ZNN=Zimmer natural nail; OR=odds ratio; CI=Confidence interval. A Network geometry indicating number of participants in each arm (size of points) and number of comparisons between arms (thickness of lines). B Forest plot represents the direct and indirect compared with SHS; Effect sizes are presented as ORs with 95% CI. C Ranking plot according to surface under the cumulative ranking curve (SUCRA) analysis in the network meta-analysis for HHS. The position of each line on the graph corresponds to the ranking probability of each intervention. D Funnel plot of HHS

For overall mechanical complications, a total of 44 RCTs were included, involving 7,216 patients. The NMA results showed that there was no difference between all intervention measures and SHS (Fig. 3).

Fig. 3.

Network meta-analysis of different implants for overall mechanical complications. SHS=sliding hip screw; PFNA=proximal femoral nail anti-rotating; PFN=proximal femoral nail; PCCP=percutaneous compression plate; ITN=InterTAN nail; GN=Gamma nail; ZNN=Zimmer natural nail; MSP=Medoff sliding plate; IMHS=intramedullary hip screw; ENDOVIS=ENDOVIS nail; ACE=the ACE trochanteric nail; TPFT=Targon PFT cephalomedullary nail; OR=odds ratio; CI=Confidence interval. A Network geometry indicating number of participants in each arm (size of points) and number of comparisons between arms (thickness of lines). B Forest plot represents the direct and indirect compared with SHS; Effect sizes are presented as ORs with 95% CI. C Ranking plot according to surface under the cumulative ranking curve (SUCRA) analysis in the network meta-analysis for overall mechanical complications. The position of each line on the graph corresponds to the ranking probability of each intervention. D Funnel plot of overall mechanical complications

A network meta-analysis of non-mechanical major post-surgery complications included 37 randomized controlled trials with a total of 6,639 patients. The results showed that ITN significantly reduced the incidence of complications (OR, 0.55; 95% CI, 0.33 to 0.91; moderate confidence), while other implants had no difference from SHS (Fig. 4).

Fig. 4.

Network meta-analysis of different implants for non-mechanical major post-surgery complications. SHS=sliding hip screw; PFNA=proximal femoral nail anti-rotating; PFN=proximal femoral nail; PCCP=percutaneous compression plate; MSP=Medoff sliding plate; ITN=InterTAN nail; IMHS=intramedullary hip screw; GN=Gamma nail; ZNN=Zimmer natural nail; ENDOVIS=ENDOVIS nail; TPFT=Targon PFT cephalomedullary nail; OR=odds ratio; CI=Confidence interval. A Network geometry indicating number of participants in each arm (size of points) and number of comparisons between arms (thickness of lines). B Forest plot represents the direct and indirect compared with SHS; Effect sizes are presented as ORs with 95% CI. C Ranking plot according to surface under the cumulative ranking curve (SUCRA) analysis in the network meta-analysis for non-mechanical major post-surgery complications. The position of each line on the graph corresponds to the ranking probability of each intervention. D Funnel plot of non-mechanical major post-surgery complications.

Regarding the reoperation rate, we analyzed 40 trials involving 6,859 patients. The network meta-analysis showed that there was no difference between all implants compared to SHS (Fig. 5).

Fig. 5.

Network meta-analysis of different implants for reoperation rate. SHS=sliding hip screw; PFNA=proximal femoral nail anti-rotating; PFN=proximal femoral nail; PCCP=percutaneous compression plate; MSP=Medoff sliding plate; ITN=InterTAN nail; GN=Gamma nail; ZNN=Zimmer natural nail; ENDOVIS=ENDOVIS nail; TPFT=Targon PFT cephalomedullary nail; OR=odds ratio; CI=Confidence interval. A Network geometry indicating number of participants in each arm (size of points) and number of comparisons between arms (thickness of lines). B Forest plot represents the direct and indirect compared with SHS; Effect sizes are presented as ORs with 95% CI. C Ranking plot according to surface under the cumulative ranking curve (SUCRA) analysis in the network meta-analysis for reoperation rate. The position of each line on the graph corresponds to the ranking probability of each intervention.D. Funnel plot of reoperation rate

Important but not crucial

For mortality, the network meta-analysis comprised 43 trials, involved 6,044 patients. The results showed that there was no difference between all interventions compared to SHS (Supplementary Information 4, Append 6 Table 6.1).

Through evaluating intraoperative complications – operative issue, the network meta-analysis included 9 RCTs involving 2,368 patients, ITN (OR, 3.41; 95% CI, 2.03 to 5.73; moderate confidence) was associated with a higher risk of inducing intraoperative fractures, while the remaining implants did not show differences compared with SHS (Supplementary Information 4, Append 6 Table 6.3).

For non-specific intraoperative complications, the network meta-analysis included 9 studies involving 1,477 participants, which indicated that compared with SHS, which showed that all implants were no difference compared with SHS (Supplementary Information 4, Append 6 Table 6.4).

Through analyzing non-mechanical post-surgery medical complications, 28 trials were included, involving 4,569 patients. The results showed that none of the implants were no difference compared with SHS. In the analysis of non-mechanical minor post-surgery complications, 40 trials were included, involving 6,445 patients, and the results showed that PCCP effectively reduced the incidence of complications (OR, 0.12; 95% CI, 0.05 to 0.30; high confidence), while no difference were seen between other implants and SHS (Supplementary Information 4, Append 6 Table 6.5 and Table 6.7).

In the analysis of cut out, implant failure, migration, and valgus deformity, 25, 22, 12, and 16 RCTs were respectively included, involving 5,091, 3,444, 1,834, and 3,179 patients. The results showed that there was no difference between all implants and SHS (Supplementary Information 4, Append 6 Table 6.9, Table 6.10, Table 6.11 and Table 6.12).

Finally, the analysis of non-specific mechanical complications included 19 studies and 4,170 patients, and the results showed that PFNA (OR, 0.05; 95% CI, 0.02 to 0.11; high confidence) and ITN (OR, 0.25; 95% CI, 0.03 to 0.49; high confidence) effectively reduced the incidence of complications (Supplementary Information 4, Append 6 Table 6.1).

Regression analysis and separate analysis

We conducted four additional meta-regression analyses to explore the potential moderating effects of the following variables on HHS, mechanical complications-fixation failure, overall mechanical complications, and intraoperative complications-operative issue: age, gender, operative time, and publication year (Supplementary Information 4, Appendix 9). The results showed that age (P = 0.04) and publication year (P = 0.04) were potential moderator for HHS, the duration of surgery is a potential modulator of the overall mechanical complications (P = 0.02), and age (P = 0.01) and gender (P = 0.02) were potential moderator for intraoperative complications-operative issue.

For the analysis of intraoperative, non-mechanical, and mechanical complications in patients with UIF, 8, 14, and 12 randomized controlled trials were respectively included, involving 1,289, 1,604, and 1,377 patients (Supplementary Information 4, Appendix 11). The NMA results indicated that all the implants showed no difference. The top-ranked by SUCRA was SHS, PFN (OR, 0.48; 95% CI, 0.21 to 1.07), and ITN (OR, 0.09; 95% CI, 0.00 to 2.48), with 85.3%, 86.1%, and 80.0% respectively. When the implants were categorized into intramedullary fixation and extramedullary fixation for reanalysis, for non-mechanical and mechanical complications, all the implants demonstrated no difference compared with SHS. However, for intraoperative complications, PFN (OR, 8.38; 95% CI, 1.29 to 54.45) could induce more intraoperative complications.

Discussion

Principal findings

This network meta-analysis comprehensively evaluated and compared the efficacy of 15 currently utilized implants in the treatment of IFF, incorporating data from 54 RCTs involving 10,275 patients. Compared to the SHS, PCCP and PFNA proved the most effective in minimizing non-mechanical major post-surgery complications and non-specific mechanical complications, with high confidence of evidence. ITN emerged as the most effective implant for reducing postoperative complications, but was associated with increased risk of intraoperative complications.

The key finding of this study is that ITN significantly reduces the incidence of major non-mechanical postoperative complications, such as wound infection and fracture nonunion, in elderly patients with hip fractures. This advantage may be partly explained by the minimally invasive nature of the ITN procedure, which reduces soft tissue dissection and intraoperative blood loss [26], thereby preserving local blood supply and limiting inflammation. Additionally, shorter operative and fluoroscopy times [27], may further decrease tissue trauma and infection risk. The ability to allow early mobilization [28], may also enhance circulation and promote healing, consistent with previous findings [29]. In contrast, the pooled results showed ITN is associated with an increased risk of operative complications, potentially due to localized bone stress concentration resulting from excessive compression during insertion [11]. As a newer generation of implants, the learning curve and unfamiliarity with the procedure may contribute to this risk, which can be mitigated through enhanced surgical experience and skills.

Percutaneous compression plate is theoretically regarded as an integration of the advantages of IMN and SHS, offering both minimally invasive implantation and enhanced stability. Previous systematic reviews and meta-analyses have demonstrated that PCCP significantly reduces intraoperative blood loss and non-orthopedic postoperative complications (such as deep vein thrombosis and nosocomial infections), compared to SHS [30]; and reduced hospital stays, blood transfusions, and implant-related complications compared to intramedullary nail [31]. Note that the studies included in these reviews “have several methodological issues and some are of relatively poor quality.” In our research, we did not observe or measure these advantages of PCCP; however, we did find that PCCP is most effective in reducing minor surgical complications, primarily superficial surgical site infections (80 out of 101 events). This benefit may be attributed to the minimal soft tissue trauma associated with PCCP, achieved through the use of two small percutaneous portals and small-diameter gradual drilling [10].

For both Harris Hip Score (HHS) and mortality, the pooled analysis revealed no significant differences across the various implants, which is consistent with findings from previous meta-analyses [32, 33].This lack of differentiation is likely due to the nature of these outcome measures. HHS, while widely adopted, is a general functional score that may not sensitively reflect implant-specific benefits, particularly in elderly patients, whose recovery is influenced by baseline frailty, comorbidities, and pre-injury mobility.Moreover, the absence of preoperative HHS scores in most included studies limits our ability to assess the degree of functional improvement, further reducing the interpretability of HHS-based comparisons. Similarly, mortality is more affected by overall health status and perioperative risk than by implant design.Although some implants differ in design, materials, and biomechanical properties, such differences may not be captured by global measures like HHS or mortality.Future original studies should include more specific and sensitive outcome measures, such as fracture site collapse, anatomical reduction, and recovery of ambulatory function, to more accurately assess the clinical effects of different implants.

Mechanical complications as an outcome measure have rarely been examined specifically or in isolation in prior individual RCTs due to their relatively low incidence, but they indeed can significantly impact patient functional recovery, as well as rates of readmission and reoperation [34]. Nevertheless, our pooled analysis incorporating 44 RCTs involving 7,212 patients still failed to identify statistically significant differences. This suggests that, despite the claims of certain implants regarding their theoretically superior biomechanical properties, these advantages do not well translate into improved clinical outcomes. This phenomenon can likely be explained by the complexities of surgical management of hip fracture in such vulnerable population, where factors such as bone quality, precision of reduction, fixation skills, and the surgeon’s experience serve as critical determinants [35]. Furthermore, the postoperative rehabilitation protocols and patient adherence significantly influence outcomes [36], underscoring the increasing importance of personalized approaches in future clinical practice.

Unstable intertrochanteric fractures represent a critical focus in research, often serving as the primary study population or as a subgroup for refined analysis in previous RCTs due to their higher surgical demands and typically poorer prognoses [37]. Similarly, the biomechanical and clinical advantages of implants are frequently highlighted in the context of unstable intertrochanteric fractures [38, 39]. Current guidelines recommendations support Cephalomedullary device for unstable fractures [9, 40]. However, our pooled analysis did not demonstrate any advantages of intramedullary fixation devices over SHS, nor were there significant differences between the intramedullary fixation devices, in terms of outcomes such as non-mechanical, and mechanical complications. A 2018 network meta-analysis that included 12 RCTs compared the total complications following the use of the PFNA, ITN, Gamma nail, and SHS for treatment of UIF [19]. In their pooled analyses, the older Gamma nail proved the most likely to reduce overall complications. These results collectively indicate that, despite considerable advancements in design and materials over the past three decades, there has been, to a large extent, “little substantive” improvement in their clinical performance of these implants, if any. Future implant designs should consider innovative directions or conceptual frameworks, particularly if our results are substantiated.

Strengths and limitations

Strengths of this systematic review and NMA include the most comprehensive synthesis of evidence to data on efficacy and safety of current implants for older patients with intertrochanteric fractures, capturing all recent publication. By involving a multidisciplinary team in definition and classification of complications based on the occurrence timing, severity and clinical importance and utilizing rigorous and contemporary evaluation approach to assess the evidence certainty of results enhance the granularity of the findings and ensure their clinical relevance. However, our study has several limitations. First, the absence of individual patient data pooling reduced the prediction of synthesis, especially for the subgroup effects. Second, the inclusion of implants from various manufacturers introduces inherent variations in design specifics. Third, the reliance on postoperative functional data, with only one RCT reporting preoperative baseline HHS, diminishes the internal validity of the findings and their generalizability to patients with varying functional statuses. Fourth, the small sample sizes of the majority of trials included in our analysis may lead to a reduced statistical precision, compromising the capacity to detect meaningful effects. Finally, while we did consider the impact of fracture stability in our analysis of unstable intertrochanteric fractures, the overall paucity of evidence for various fracture types—63.0% of the studies failed to account for fracture stability—limited our ability to fully explore whether treatment effects vary across different fracture classifications. Moreover, the included RCTs did not report surgeon experience or learning curves, which may have influenced intraoperative complication rates. This limits interpretation of the results, and future studies should account for this factor.

Conclusions

ITN clearly reduces the incidence of non-mechanical major post-surgery complications and non-specific mechanical complications, however, it is associated with an increased risk of intraoperative complications. No differences were observed between SHS and others implants regarding Harris hip score, reoperation rate, and overall mechanical complications. Most comparisons evaluating the efficacy of these implants are supported by moderate confidence of evidence. Further large-scale, longitudinal, randomized controlled trials are necessary to provide more robust and consistent evidence.

Abbreviations

IFF

Intertrochanteric femoral fractures

OR

Odds Ratio

MD

Mean Difference

CI

Confidence Interval

NMA

Network meta-analysis

RCTs

Randomized controlled trials

SUCRA

Surface under the cumulative ranking

CINeMA

Confidence in Network Meta-Analysis

ITT

Intention-to-treat

UIF

Unstable intertrochanteric fractures

SH

Sliding hip screw

DHS

Dynamic hip screw

PCCP

Percutaneous compression plate

GN

Gamma nail

PFN

Proximal femoral nail

PFNA

Proximal femoral nail anti-rotating

CHS

Compression hip screw

MSP

Medoff sliding plate

ZNN

Zimmer natural nail

IMHS

Intramedullary hip screw

ITN

the InterTAN nail

TPFT

Targon PFT cephalomedullary nail

ACE

ACE trochanteric nail

AMBI

AMBI hip screw

ENDOVIS

ENDOVIS nail

Authors’ contributions

Study design: Yanbin Zhu., Xiuting Li. Literature search: Shanshan Zhang, Yihao Ge, Zhaodong Bi. Full-text review and data extraction: Jiheng Xiao, Cici Bai, Yuqing Li. Quality assessment: Miao Tian, Yanbin Zhu. Statistical analyses: Jiheng Xiao, Yihao Ge. The drafting and revision of manuscript: Shanshan Zhang, Yihao Ge, Zhaodong Bi.

Funding

This study was supported by CHINA ZHONGGUANCUN PRECISION MEDICINE SCIENCE AND TECHNOLOGY FOUNDATION (2024JZYX04677) and the Key Research and Development Project of Hebei Province(21377731D).

Data availability

The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.

Declarations

Ethics approval and consent to participate

This is a systematic review and network meta-analysis without the need for ethics approval and consent.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.