The top fifty most influential articles on hip fractures

Purpose Hip fractures are one of the most common disabling fractures in elderly people and peri-operative management has advanced considerably over the past decades. The purpose of this study was to evaluate the change of scientific focus by creating a top 50 list of the most influential papers on this topic. Methods The Clarivate Web of Science Search was used to identify the most cited articles. The used search phrase was [(hip OR pertrochanteric OR (femoral neck)) AND fracture AND (surgery OR treatment)]. The number of citations, citation density, study type, study design, published year, fracture type, country, evidence level and published journal were recorded. Results The top 50 articles were published between 1973 and 2014 and cited between 88 and 496 times. The mean citation density increased noticeably after the year 2000, representing the knowledge gain of the last 20 years. The topics surgical treatment (n = 19), risk factor assessment (n = 19), perioperative hemodynamic management (n = 7), additional treatment (n = 4) and general reviews (n = 1) were covered. Twenty-five articles were published from institutions in Europe, 24 from institutions in North America and one from an institution in Asia. Conclusion While studies about surgical treatment options and risk factor assessment have been historically important, there was a rise of articles about additional treatment options for osteoporosis and the optimal postoperative care after the year 2005. The presented lists and map of citation classics give an overview of the most influential studies on hip fractures.


Introduction
Hip fractures are one of the most common fractures in elderly patients [1]. The one year mortality ranges between 14 and 36% [2,3]. In 2000, more than 1.6 million hip fractures occurred globally and accounted for 20% of all fractures in patients over 50 years [4]. It is estimated that the absolute number of annual fractures will be 4.5 million by the year 2050 [5,6]. Hip fractures are among the classic fragility fractures of geriatric patients and more than 90% are caused by low energy trauma (i.e. fall from standing height). Established risk factors are osteoporosis, high age, female sex, smoking and a low BMI [4,[7][8][9]. They can be classified into femoral neck fractures, per-or intertrochanteric fractures and subtrochanteric fractures [10]. While peror intertrochanteric fractures are treated with osteosynthesis devices, femoral neck fractures can be either treated with hemi-and total hip arthroplasty or osteosynthesis [11].
The enormous prevalence of hip fractures accentuates the socio-economic significance and explains the sheer infinite number of published articles [12]. In an era of evidencebased medicine, research studies are not only important for a better understanding but also in clinical decision-making. With the increase in studies published recently, it is becoming difficult to overlook the most current research questions. One way to determine the impact of a published article is to use the citation analysis [13][14][15][16][17][18]. Although the quality of an article does not depend solely on its citation rate, it represents its importance in the field and is widely recognized in the scientific community.
The aim of this study was to identify the top 50 most influential articles on hip fractures. To characterize the change of scientific focus and research questions in recent years, the top ten articles over the last five years (2015-2020) were separately evaluated. It was hypothesized that the literature on hip fracture treatment would change over the decades, as the evidence base and quality of studies were expected to improve over time.

Search strategy
The Clarivate Web of Science search was used to identify the most cited articles regarding hip fractures. The used search phrase was [(hip OR pertrochanteric OR (femoral neck)) AND fracture AND (surgery OR treatment)]. It was performed on 18th March 2020 and no institutional review board or ethical approval was required. The options "All databases", "Basic Search", "All years" and search based on the "topic" were applied. The first 100 articles were thoroughly studied and excluded (1) if hip fractures were not the main topic and (2) if there was no full text available.
A second search was performed for articles published over the last five years (2015-2020). With the exception of the time frame, the same parameters as described above were used. Finally, 60 articles were included for the final analysis. Data extraction was performed according to an adapted PRISMA flowchart (Fig. 1). Because no patients were involved, ethics committee approval was waived for this cross-sectional study.

Qualitative analysis
All articles were organized in descending order according to overall number of citations. Two different tables were created: the first for the top 50 articles of all time (Table 1) and the second for the top ten articles between 2015 and 2020 ( Table 2). The following data were extracted from all articles: overall number of citations, citation density, level of evidence, title, first authors name, senior authors name, publication year, published journal, country, institution (according to corresponding author) and language. The level of evidence was either acquired from the article itself or assigned according to the Practical Guide from Wright [19]. In case of questionable results, evidence levels, study types and study designs were clarified in consensus meetings with the senior author.

Statistical analysis
Statistical analysis was performed with IBM SPSS Statistics software (BM Corp. Released 2018. IBM SPSS Statistics for Windows, Version 25.0. Armonk, NY, USA: IBM Corp.). Descriptive statistics (mean, SD, minimum and maximum) were computed for all metric variables.

Results
The initial search yielded 41.782 results. In-and exclusion criteria were applied and a total of 60 articles were included in the study (Fig. 1). Based on the total citation number, a top 50 list for all time (Table 1) and a top 10 list for the years 2015 to 2020 (Table 2) were created. The top 50 articles were all published between 1973 and 2014 and a sharp increase was seen in the current millennium (Fig. 2). Thirty-eight out of the top 50 studies were published after 2000. In 2008, the most articles were published. All included studies were cited between 88 and 496 times. The mean citation density increased noticeably after the year 2000 (Fig. 2). All articles were published in English language.

Topics
Articles of the top 50 list covered the following topics: (1) surgical treatment (n = 19), (2) risk factor assessment (n = 19), (3) peri-operative haemodynamic management (n = 7), (4) additional treatment (n = 4) and (5) general review (n = 1). The top 10 studies of the last ten years addressed the same topics (Fig. 3). The assessment of risk factors for post-operative mortality included evaluation of early surgery and mobilization as well as individual parameters such as sex, age and comorbidity. Reviews regarding surgical treatment focused on different implant options such as hemiarthroplasty, total hip replacement, dynamic hip screw or cephalomedullary nails. One biomechanical study compared the helical blade with the dynamic hip screw in body donor specimens. The topic of peri-operative haemodynamic management included concerns such as hidden blood loss during surgery, thresholds for transfusion and post-operative thrombosis prophylaxis. Additional treatment options included the investigation of possible beneficial prophylaxis for osteoporosis or pre-and post-operative optimized nutrition supply. Forty-two studies investigated femoral neck fractures, 32 examined pertrochanteric fractures, two evaluated subtrochanteric fractures and in one study, the included hip fractures were not classified.

Study type, study design and level of evidence
Regarding the study types, there were 27 therapeutic, 13 prognostic, eight reviews and one basic science article (Fig. 4). One study combined both basic science and therapeutic concepts. Concerning the study design, eleven articles were conducted as randomized controlled trials, three as non-randomized controlled trials, four as prospective and twelve as retrospective cohort studies. Four were prospective and seven retrospective case series, two reviews, five systematic reviews and meta-analysis, respectively. There was one cross-sectional and one basic science study (Fig. 5). Level IV was the most frequent evidence level, followed by levels II, III, I and V (Fig. 6).

Journals and countries
The top 50 articles were published in 23 different journals. Twenty-five articles were published from institutions in Europe, 24 from institutions in North America and one from an institution in Asia. Considering the different population distribution, North America had 4.2 articles per 100,000,000 inhabitants, Europe 3.4 articles per 100,000,000 inhabitants and Asia 0.02 articles per 100,000,000 inhabitants (Fig. 7).    with risk factor assessment, seven with peri-operative haemodynamic management, four with additional treatment and one article was a general review paper The top ten articles from 2015 to 2020 were also primarily published in Europe (n = 5), North America (n = 4) and Asia (n = 1).

Discussion
In the present study, we evaluated the most influential articles on hip fractures. The most cited papers were analyzed and a list of "citation classics" was compiled. The number one article was cited a total of 496 times and dealt with the prophylaxis of post-operative embolism in hip fracture surgery. The number of citations is comparable to the data evaluating the topic "spine fractures" [16] and "arthroscopy" [15] but stands in huge contrast to the citation numbers found on fragility fractures [18] or hip and knee arthroplasty, in which the top paper was cited 2495 times [13]. The majority of the top 50 papers focused on the optimal choice of surgical treatment (n = 19), followed by risk factor assessment (n = 19) and the peri-operative haemodynamic management (n = 7). The number of study topics and the enormous variation among them are not surprising, as hip fractures are the most common fractures in elderly patients, with high morbidity and mortality rates [20,21]. Handling of these patients requires a multidisciplinary approach which includes various specialties such as orthopaedic trauma surgeons, anesthesiologists, geriatric physicians and physiotherapists. A similar collective study on spine fractures identified only two major topics, osteoporosis and pedicle screws. Eleven articles were conducted as randomized controlled trials, three as non-randomized controlled trials, four as prospective and twelve as retrospective cohort studies. Four were prospective and seven retrospective case series, two reviews, five systematic reviews and meta-analysis, respectively Most studies were conducted in Europe and North America and were written in English language. This emphasizes the huge role of these continents in scientific research and can further be supported by previous studies, in which the same countries were predominant [13,14,17]. A world map with the geographic areas that published research on hip fractures can be found in Fig. 8. The majority of presented articles were retrospective cohort studies (n = 12), followed by randomized controlled trials (n = 11) and retrospective case series (n = 7). Randomized control trials, the gold standard of scientific research, are almost evenly distributed among studies of the optimal choice of surgical treatment and peri-operative haemodynamic management. The large number might be due to the high incidence of hip fractures, which favours big trials. This contrasts similar studies on hip arthroscopy or fragility fractures, where randomized controlled trials were the minority and case series were prevailing [14,18]. Lefaivre et al. could show that among the top 100 articles published in the field of orthopaedics, there was not a single randomized controlled trial [22]. This can be explained by the study design and the wide range of orthopaedic subspecialties and their individual level of knowledge. In the article about spine fractures, no absolute numbers are given about the type of study [16].
There has been an increase in studies and mean citation densities since 2000 (Fig. 2). This finding is consistent with previous studies on spine fractures and represents the knowledge gain of the last 20 years [16]. Data in the present study clearly demonstrates the increase of studies on haemodynamic management and additional treatment in the last 20 years, as physicians became aware of the importance of individual and adjuvant treatment (i.e. osteoporosis treatment) after surgery, similar to the study from Donnally et al. concerning spine fractures [16]. Hip fractures are life-threating events due to the various pre-existing conditions of this geriatric and multimorbid patient population. The surgery itself is not solely responsible for the patient's survival and outcome [23]. This may be further supported by the increasing age of patients and the increasing understanding of the molecular biology parameters of osteoporosis. This implication is further accompanied by a higher evidence level over the last 20 years. All level I studies of the top 50 articles were performed between the years 2003 and 2010. However, level IV was still predominant overall. Present distribution is similar to recent articles on spine fractures, hip arthroplasty or hip arthroscopy [13,14,16].
There are some intrinsic problems with this kind of study and citation analysis. As previously described, it does not account for self-citation and the author's preference to cite articles in the journal in which they seek to publish their own work [13-15, 17, 24]. Furthermore, there is a clear time effect in citation analysis. The most recent articles are at disadvantage, because there is not enough time for citations to accumulate. To compensate for that, we included the top ten articles of the last five years. Another possible weakness is the "Snow-Ball" effect, which suggests that authors are likely to cite a study that a previous publication has cited without questioning the quality and accuracy of this study   [25]. The total citation count was used to determine the ranking, because we wanted to show the exponential growth medical research has experienced during the last decades. However, we added the citation density in Tables 1 and 2 for a better understanding of the article´s impact. Finally, we recognize that in many contexts, the value of a contribution cannot be quantified simply by the number of citations a publication receives.
We created a top 50 list of citation classics and elucidated the change of research focus during the last 20 years. The top ten list from 2015 to 2020 was added to highlight the most recent topics. In conclusion, the most cited studies on hip fractures focused primarily on surgical treatment options and risk factor assessment. After the year 2005, studies about additional treatment options for osteoporosis and an optimal post-operative care gained in importance. The presented tables with references can serve as a guide for a comprehensive understanding of the historical and current literature pertaining to hip fractures.
Author contribution All authors contributed to the study conception and design. Data collection, formal analysis, visualization and interpretation were performed by Gilbert Manuel Schwarz and Madeleine Willegger. Stefan Hajdu and Reinhard Windhager were responsible for proofreading and resources. The first draft of the manuscript was written by Gilbert Manuel Schwarz and all authors commented on previous versions of the manuscript. All authors read and approved the final manuscript. Data availability The authors confirm that the data supporting the findings of this study are available within the article. Additional data is available from the corresponding author (Madeleine Willegger) on request.

Declarations
Ethics approval No institutional ethical approval was required because the study is a systematic review of the literature.

Consent to participate
No consent to participate was necessary because the present study is a systematic review of the literature.

Consent to publish
No consent to publish was necessary because the present study is a systematic review.

Competing interests
The authors declare no competing interests.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.