Conservative Versus Operative Treatment of Proximal Humerus Fractures in Older Individuals

An Analysis of Insurance Data

Background: Little is known about the frequency and results of conservative treatment of proximal humerus fractures (PHF) in older individuals.

Method: Billing data of the BARMER health insurance carrier for all patients of age ≥ 65 for the years 2005–2021 were retrospectively analyzed with multivariable Cox regression models, taking account of the patients’ age, sex, and individual comorbidity profiles. The defined primary endpoints were overall survival (OS), major adverse events (MAE), thromboembolic events (TE), and complications of surgery or of trauma. Multivariable p-values for the effect of treatment on all primary endpoints were jointly adjusted using the Bonferroni–Holm method.

Results: Of 81 909 patients, 54% were treated conservatively. Conservative treatment was more common in those who received their diagnosis as outpatients (79.5%, vs. 37.2% for inpatients). Operative treatment was associated with significantly longer overall survival (long-term hazard ratio [HR] 0.89, 95% confidence interval [0.86; 0.91]), fewer MAE (0.90 [0.88; 0.92]), and fewer TE (0.89 [0.87; 0.92]), but more complications due to surgery or trauma (1.66 [1,.4; 1.78]; all p < 0.001). By 6 months after diagnosis, 3.1% of the patients who were initially treated conservatively had undergone surgery. Risk factors for failure of conservative treatment were alcohol abuse, obesity, cancer, diabetes mellitus, Parkinson disease, and osteoporosis.

Conclusion: The conservative treatment of PHF is associated with a lower overall rate of complications due to surgery or trauma, but also with more MAE, more TE, and higher mortality. These findings underline the need for individualized and risk-adjusted treatment recommendations.

Proximal humeral fracture (PHF) accounts for more than 10% of all fractures in patients above 64 years of age (1). Owing to the demographic trend towards an aging society, the incidence of osteoporotic fractures will continue to rise. Considerable increases in morbidity and demand on health care resources can be assumed (2, 3). Although the evidence in support of surgical treatment of PHF is limited, trends indicate a continuous increase, leading regularly to heated debates in the surgical community (4, 5, 6).

Most recent investigations into the treatment reality of PHF in the German-speaking countries originate from large trauma centers, whereas hardly any data have emerged from the outpatient sector (7, 8, 9, 10). The published rates of conservative treatment range between 15 and 25%. Given the fact that conservative treatment, even of displaced PHF, achieves similar shoulder function to surgical treatment, the low rate of conservative treatment has been criticized (11).

Recent evidence deriving from routine practice data suggests that comorbidities—especially in elderly patients—have a distinct impact on the course of treatment after locked plate fixation or reverse total shoulder arthroplasty for PHF (12, 13). The advantages of analyzing routine practice data include large cohorts and extended follow-up periods. Furthermore, one can assess the entire (medical) history, taking account of a broad spectrum of the individual comorbidity profile and of patient-related factors such as age and sex, and thus generate a risk profile, which has been lacking in studies to date (14, 15). The hypothesis of this study was that the majority of PHF in Germany are treated conservatively and that conservative treatment of PHF is associated with similar rates of complications—including major adverse events (MAE)—and similar overall survival (OS) to operative treatment.

Materials and methods

Data pool and patient cohort

The remuneration system in Germany is based on the German Diagnosis-Related Groups (G-DRG) system (12, 13, 16).

For the study at hand, billing data (for inpatients and outpatients) of the BARMER health insurance carrier from 2005 to 2021 were available. All older patients (age ≥ 65 years) coded for inpatient or outpatient PHF (ICD S42.2) between January 2011 and December 2020 were included in the study. The exclusion criteria are presented in eSupplement-Figure 1. Patients with prior PHF within the previous 5 years were excluded. The first coded diagnosis of PHF was defined as the index event of the study. Patients were classified into five treatment groups:

Figure

Landmark analysis for all primary and secondary endpoints

• Conservative treatment
• Locked plate fixation of simple PHF (sLPF; OPS 5–793.31, 5–793.k1)
• Locked plate fixation of multifragmented PHF (LPF; OPS 5–794.21, 5–794.k1)
• Reverse total shoulder arthroplasty (RTSA; OPS 5–824.21)
• All other operative treatment variants

Conservative treatment was defined as the absence of surgical treatment within the first 21 days after diagnosis of PHF. Patient characteristics were determined on the basis of inpatient and outpatient data within the past 24 months before PHF diagnosis. Pharmaceutical treatment before diagnosis, i.e., intake of anticoagulants, vitamin D/calcium, and bisphosphonates, was defined using the Anatomical Therapeutic Chemical classification system (ATC). Patients were observed from 21 days after the first diagnosis of PHF to the end of follow-up. The median observation period was 66.4 months. The end of follow-up was defined as death from any cause, exit from the database, or the end of the study (31 December 2021). This last guaranteed at least 1 year of follow-up.

Primary endpoints

The following were defined as primary endpoints:

• Any surgical or injury-related complications
• Major adverse events, i.e., defined as resuscitation, acute myocardial infarction, stroke, sepsis, acute renal failure, acute liver failure, acute respiratory distress syndrome, or death
• Thromboembolic events or death
• Overall survival after PHF

Secondary endpoints

Minor outpatient complications were analyzed as a secondary endpoint. In addition, risk factors for failure of conservative treatment—defined as conversion to operative treatment after at least 3 weeks—were investigated.

Statistical methods

Detailed information on all statistical analyses, including the definitions of all endpoints and comorbidities, were presented in the short study protocol (eSupplement-Material 1). In short, the association between treatment group and primary endpoints was analyzed using multivariable Cox regression models to account for the patients’ entire comorbidity profile. For surgical or injury-related complications and minor outpatient complications, death was considered as a competing event. In this case the subdistributional hazard ratios were determined using multivariable Fine–Gray models.

Furthermore, a landmark analysis was performed to evaluate possible differences between the short-term and long-term outcomes. Analysis of the endpoints covered either the first 3 months after diagnosis (short-term analysis) or the period starting 3 months after PHF (long-term analysis). All patients with follow-up time shorter than 3 months were excluded from long-term analysis.

All analyses were performed both with a binary variable for the treatment effect (surgery: yes/no) and with detailed treatment group (conservative, sLPF, LPF, RTSA, other). The p-values for comparison of surgical vs conservative treatments from the multivariable regressions for the primary endpoints (four outcomes, short-term and long-term) were jointly adjusted using the Bonferroni–Holm method to control the familywise error rate. Adjusted p-values were marked p_adj. All other p-values are to be interpreted as explorative. Moreover, propensity score (PS) matching was performed as a sensitivity analysis.

Results

Demographics and patient characteristics

In total, 81 909 patients (median age 78 years, 84.2% female) with PHF were included in the study. Of these, 44 224 patients (54.0%) received conservative treatment during the first 21 days after diagnosis (operative treatment: sLPF 3.1%, LPF 17.5%, RTSA 8.2%, other 12.3%, unclear 5.0%). In 39.5% of the patients, PHF was first diagnosed in an outpatient setting. Moreover, surgical intervention occurred less frequently in persons treated as outpatients than in hospitalized patients. In the outpatient sector 79.5% received conservative treatment versus only 37.2% in the inpatient sector. The proportion of patients without operative treatment increased from 51.9% in 2011 to 55.6% in 2021, The proportion of patients treated with RTSA also increased (from 2.2% in 2011 to 13.4% in 2020). Patients with unclear operative treatment, i.e., different coded treatments on the same day, or surgery on both sides of the body on the same day, were excluded from further analysis. Patients who did not receive operative treatment were more often had dementia and were more likely to be treated with anticoagulants (Table).

Table

Patient characteristics at diagnosis

Surgical or injury-related complications (primary endpoint)

One year after PHF, 4.8% [4.6; 5.0] of patients with primary conservative treatment had an injury-related complication. A higher rate of complications was observed for locked plate fixation (sLPF: 8.8% [7.7; 9.9]; LPF: 10.8% [10.3; 11.3]; eSupplement-Table 2). Multivariable landmark analysis showed that surgical treatment, compared with conservative treatment, was associated with a significantly higher risk of injury-related or surgical complications in the short term (HR 1.16 [1.07; 1.25] p_adj <0.001) and in the long term (HR 1.66 [1.54; 1.78]; p_adj <0.001; Figure). Cox regression analysis with a time-dependent treatment variable, carried out as a sensitivity analysis, confirmed the observed association (eSupplement-Figure 2).

Major adverse events, thromboembolic events, and overall survival (primary endpoint)

Five years after PHF, 34.2% [33.8; 34.6%] of the patients were deceased. After adjustment for the persons’ individual risk profile, surgical treatment was associated with significantly better OS in both the short term (HR 0.76 [0.70; 0.83]) and the long term (HR 0.89 [0.86, 0.91]; both p_adj <0.001). Similar effects were observed for MAE (short-term analysis: HR 0.81; long-term analysis: HR 0.89; both p_adj <0.001) and TE (short-term analysis: HR 0.80; long-term analysis: HR 0.90; both p_adj <0.001; Figure). Furthermore, no relevant time dependency of treatment effect was observed for OS, MAE, and TE (eSupplement-Figure 2).

Sensitivity analysis 1:1 propensity score matching

As a sensitivity analysis, 1:1 PS matching was performed, including age, sex, year of diagnosis, and risk profile. The balance of the matching cohort is presented in eSupplement-Table 3. Findings for all endpoints could be confirmed (eSupplement-Figure 3).

Failure of conservative treatment

By 6 months after diagnosis, 3.1% of patients with primary conservative treatment had been converted to operative treatment, with RTSA being the most common secondary procedure (proportion of total converted patients treated using the various surgical options: sLPF 4.8% [3.8; 6.1]; LPF 23.4% [21.2; 25.7]; RTSA 41.5% [38.9; 44.2]; other 23.9% [21.7; 26.2]; unclear 6.3% [5.1; 7.7]). Risk factors for failure of conservative treatment were alcohol abuse, obesity, cancer, diabetes mellitus, Parkinson disease, and osteoporosis. In contrast, dementia, male sex, and age per year were associated with lower probability of secondary surgical treatment (eSupplement-Table 4).

Discussion

The findings of this study show that conservative treatment of PHF was associated with a significantly lower rate of surgical/injury-related complications, but with more MAE, more TE, and higher mortality. The associations found for the primary endpoints were confirmed in a propensity score matching analysis.

RTSA represents an exception: fewer surgical/injury-related complications were observed than with conservative treatment. The data highlight the need for individualized and risk-adapted treatment recommendations for PHF. Risk factors for the failure of conservatively treated PHF were quantified systematically for the first time in this study.

The incidence of PHF shows strong variations depending on population age, sex distribution, time of count, and geographical area (2, 17, 18). Both inpatient and outpatient and both operatively and conservatively treated PHF must be included to prevent false assumptions (17, 19, 20, 21). Detailed age- and sex-adjusted incidences of this cohort confirm the distribution as described above (22). However, with 351.1 (± 7.7) per 100 000 person-years the data are particularly relevant in terms of their completeness: they include all patients independent of treatment modality and are nearly five times as high as some previously reported data (21, 22, 23, 24).

As of now, there is no clear evidence to show which treatment option for PHF is superior in older patients, as only a small number of high-quality studies have been conducted and there are therefore few helpful clinical reference points to aid decision making (25). Additionally, little is known about conservative treatment methods. Only 5% of the PHF literature even concerns itself with this topic (26).

Comparing the functional outcome between operatively and conservatively treated PHF, neither recent systematic reviews nor high-quality prospective randomized controlled trials have shown a significant difference, even for displaced fractures (5, 27). An example is provided by the ProFHER trial, in which operative and conservative treatment were compared in 215 adult probands with PHF. After 2 years’ follow-up, the study showed no significant difference in functionality (assessed using the Oxford Shoulder Score), health-related quality of life (SF-12 scores), or fracture-related complications. However, all medical complications occurred in the operative group (28).

A Cochrane review analyzed quality of life and functionality at 1 year after operatively and conservatively treated PHF. No significant, clinically relevant difference was found. However, there were signs of a higher overall risk of MAE after surgery (11). While it is not possible to make a statement on functionality, the available data underline that LPF, compared with conservative treatment, is an independent risk factor for surgical/injury-related complications in the short and the long term. RTSA stands out as an exception. In contrast to Handoll et al., after adjustment for patients’ risk profiles, our data show better OS and fewer MAE and TE following surgical treatment.

In summary, the above-mentioned pros and cons must be considered whenever deciding which treatment option to pursue. The individual risk factors listed in this article can provide a decision aid. For example, if surgery is needed RTSA could be used more liberally. Some 3.1% of patients initially treated conservatively underwent surgery within 6 months, with LPF in 28.2%, RTSA in 41.5%, and other procedures in 30.2% of cases. Again, the reasons for conversion remain elusive, as the small body of literature does not quantify the most important complications of conservative treatment, such as shoulder arthritis, osteonecrosis, non-union, malunion, secondary fracture displacement, or shoulder instability (29). In consideration of the increased complication rates for salvage RTSA after failed LPF, primary RTSA after failure of conservative treatment seems reasonable and was the option most frequently pursued in the cohort observed here (24).

With regard to the decision between operative and conservative treatment of PHF in older patients, some authors state a guiding principle that patients with low demands and poor health are as a rule more suitable for conservative treatment (30). The literature to date contains no studies in which risk factors for failure of conservative treatment were analyzed in an adequately high number of patients. Hence, the comorbidities listed in eSupplement-Table 4, such as diabetes mellitus (HR 1.11), Parkinson disease (HR 1.39), and osteoporosis (HR 1.25) are presented as risk factors for the first time in the literature and may aid decision-making. However, they are also known to be risk factors for failure of operative treatment (13, 24, 29). It is therefore hard to use these factors solely for the severity and impact of surgical complications. Direct comparison can be made with regard to MAE, TE, and mortality. While higher mortality rates were described in other studies, the high rate of MAE and TE are novel additional risk factors that need to be given particular consideration during conservative treatment (31).

Limitations and strengths of the study

The data were collected originally for billing purposes, not primarily for research. Hence, there may have been a trend in coding in favor of higher compensation categories, which would influence the results. With high probability, however, all treatment groups were affected in similar manner. Moreover, there may have been discrepancies between the medications prescribed and those actually taken. Furthermore, neither the fracture classification nor the reason for treatment decision can be derived from our data. However, in the LPF group approximately 18% were coded as simple fracture, which agrees with Launonen et al. (2). Additionally, the presented data permit no causal conclusions regarding complications and mortality cannot be made.

Strengths of the study are the large cohort size, the long follow-up period, the representative population cross-section, and the completeness of the data, including age, sex, treatment modality, and course of secondary disease.

Conclusions

The literature to date shows no functional superiority of any specific treatment method for PHF in older patients. Analysis and assessment of treatment risks may improve the outcome. Conservative treatment of PHF is associated with a lower overall number of surgical/injury-related complications, but more MAE, more TE, and higher mortality. However, RTSA stands out by virtue of having fewer surgical/injury-related complications than conservative treatment. Therefore, more liberal use of RTSA can be recommended. Our data highlight the need for individualized and risk-adapted recommendations for the treatment of PHF.

Ethics committee approval

The study was conducted in accordance with the Declaration of Helsinki and was approved by Westphalia–Lippe Ethics Committee (no. 2022–300-f-S). Given the anonymity of insurance data, no prior written informed consent for the analyzed data had to be obtained.

Funding

This work was supported by the funding program Innovative Medizinische Forschung (Innovative Medical Research; grant number I-KÖ122113) of the Faculty of Medicine, University of Münster, Germany.

Conflict of interest statement

The authors declare that no conflict of interest exists.

Manuscript received on 7 May 2023, revised version accepted on 19 March 2024.

Corresponding author
Prof. Dr. med. J. Christoph Katthagen

Klinik für Unfall-, Hand- und Wiederherstellungschirurgie
Universitätsklinikum Münster
Albert-Schweitzer-Campus 1, Gebäude W1

48149 Münster, Germany

Cite this as:
Katthagen JC, Raschke MJ, Fischhuber K, Ikingl J, Marschall U, Sußiek J, Faldum A, Stolberg-Stolberg J, Köppe J: Conservative versus operative treatment of proximal humerus fractures in older individuals—an analysis of insurance data. Dtsch Arztebl Int 2024; 121: 454–60. DOI: 10.3238/arztebl.m2024.0059

References

1. Baron JA, Karagas M, Barrett J, et al.: Basic epidemiology of fractures of the upper and lower limb among Americans over 65 years of age. Epidemiology 1996; 7: 612–8

2. Launonen AP, Lepola V, Saranko A, Flinkkila T, Laitinen M, Mattila VM: Epidemiology of proximal humerus fractures. Arch Osteoporos 2015; 10: 209

3. Curtis EM, Moon RJ, Harvey NC, Cooper C: The impact of fragility fracture and approaches to osteoporosis risk assessment worldwide. Bone 2017; 104: 29–38

4. Jo YH, Lee KH, Lee BG: Surgical trends in elderly patients with proximal humeral fractures in South Korea: a population-based study. BMC Musculoskelet Disord 2019; 20: 136

5. Beks RB, Ochen Y, Frima H, et al.: Operative versus nonoperative treatment of proximal humeral fractures: a systematic review, meta-analysis, and comparison of observational studies and randomized controlled trials. J Shoulder Elbow Surg 2018; 27: 1526–34

6. Brorson S, Palm H: Proximal Humeral Fractures: The choice of treatment. In: Falaschi P, Marsh D (eds.): Orthogeriatrics: The management of older patients with fragility fractures. Cham: Springer International Publishing 2021; 143–53

7. Ott N, Muller C, Jacobs A, et al.: Outcome of geriatric proximal humeral fractures: a comparison between reverse shoulder arthroplasty versus open reduction and internal fixation. OTA Int 2022; 5: e188

8. von Dercks N, Hepp P, Theopold J, Henkelmann R, Hackl D, Kossack N: Health care for inpatients with a proximal humeral fracture – an analysis of health insurance data. Z Orthop Unfall 2022; 161: 439–46

9. Dey Hazra RO, Blach RM, Ellwein A, Katthagen JC, Lill H, Jensen G: Latest trends in the current treatment of proximal humeral fractures—an analysis of 1162 cases at a level-1 trauma centre with a special focus on shoulder surgery. Z Orthop Unfall 2022; 160: 287–98

10. Katthagen JC, Huber M, Grabowski S, Ellwein A, Jensen G, Lill H: Failure and revision rates of proximal humeral fracture treatment with the use of a standardized treatment algorithm at a level-1 trauma center. J Orthop Traumatol 2017; 18: 265–74

11. Handoll HH, Elliott J, Thillemann TM, Aluko P, Brorson S: Interventions for treating proximal humeral fractures in adults. Cochrane Database Syst Rev 2022; 6: CD000434

12. Koeppe J, Stolberg-Stolberg J, Rischen R, Faldum A, Raschke MJ, Katthagen JC: In-hospital complications are more likely to occurrence after reverse shoulder arthroplasty than after locked plating for proximal humeral fractures. Clin Orthop Relat Res 2021; 479: 2284–92

13. Stolberg-Stolberg J, Köppe J, Rischen R, et al.: The surgical treatment of proximal humeral fractures in elderly patients. Dtsch Arztebl Int 2021; 118: 817–23

14. Olsson C, Petersson CJ: Clinical importance of comorbidity in patients with a proximal humerus fracture. Clin Orthop Relat Res 2006; 442: 93–9

15. Garcia-Reza A, Dominguez-Prado DM, Iglesias-Nunez C, et al.: Analysis of predictors of mortality after surgical and non-surgical management in proximal humerus fractures. J Orthop Traumatol 2021; 22: 43

16. Koeppe J, Katthagen JC, Rischen R, et al.: Male sex is associated with higher mortality and increased risk for complications after surgical treatment of proximal humeral fractures. J Clin Med 2021; 10: 2500

17. Iglesias-Rodriguez S, Dominguez-Prado DM, Garcia-Reza A, et al: Epidemiology of proximal humerus fractures. J Orthop Surg Res 2021; 16: 402

18. Pesce V, Vicenti G, Picca G, Rifino F, Carrozzo M, Moretti B: A review of gender differences in proximal humerus fractures. Italian Sex Gender Specif Med 2016; 2: 141–5.

19. Bagley SC, Altman RB: Computing disease incidence, prevalence and comorbidity from electronic medical records. J Biomed Inform 2016; 63: 108–11

20. Hemmann P, Ziegler P, Konrads C, et al.: Trends in fracture development of the upper extremity in Germany–a population-based description of the past 15 years. J Orthop Surg Res 2020; 15: 65

21. Rupp M, Walter N, Pfeifer C, et al.: The incidence of fractures among the adult population of Germany—an analysis from 2009 through 2019. Dtsch Arztebl Int 2021; 118: 665–9

22. Koeppe J, Stolberg-Stolberg J, Fischhuber K, et al.: The incidence of proximal humerus fracture—an analysis of insurance data. Dtsch Arztebl Int 2023; 120: 555–6

23. Klug A, Gramlich Y, Wincheringer D, Schmidt-Horlohe K, Hoffmann R: Trends in surgical management of proximal humeral fractures in adults: a nationwide study of records in Germany from 2007 to 2016. Arch Orthop Trauma Surg 2019; 139: 1713–21

24. Koeppe J, Stolberg-Stolberg J, Rischen R, et al.: Increased complication rates for salvage RTSA after failed locked plate fixation compared to primary RTSA in the treatment of proximal humeral fractures in elderly patients. J Shoulder Elbow Surg 2023; 32: 1574–83

25. Razaeian S, Wiese B, Zhang D, Harb A, Krettek C, Hawi N: Nonsensus in the treatment of proximal humerus fractures: uncontrolled, blinded, comparative behavioural analysis between Homo chirurgicus maccidentus and Macaca sylvanus. BMJ 2020; 371: m4429

26. Slobogean GP, Johal H, Lefaivre KA, et al: A scoping review of the proximal humerus fracture literature. BMC Musculoskelet Disord 2015; 16: 112

27. Xie L, Ding F, Zhao Z, Chen Y, Xing D: Operative versus non-operative treatment in complex proximal humeral fractures: a meta-analysis of randomized controlled trials. Springerplus 2015; 4: 728

28. Handoll H, Brealey S, Rangan A, et al: The ProFHER (PROximal Fracture of the Humerus: Evaluation by Randomisation) trial – a pragmatic multicentre randomised controlled trial evaluating the clinical effectiveness and cost-effectiveness of surgical compared with non-surgical treatment for proximal fracture of the humerus in adults. Health Technol Assess 2015; 19: 1–280

29. Brorson S, Alispahic N, Bahrs C, Joeris A, Steinitz A, Audige L: Complications after non-surgical management of proximal humeral fractures: a systematic review of terms and definitions. BMC Musculoskelet Disord 2019; 20: 91

30. Khoriati AA, Antonios T, Bakti N, Mohanlal P, Singh B: Outcomes following non operative management for proximal humerus fractures. J Clin Orthop Trauma 2019; 10: 462–7

31. Walter N, Szymski D, Kurtz SM, et al.: Proximal humerus fractures – epidemiology, comparison of mortality rates after surgical versus non-surgical treatment, and analysis of risk factors based on Medicare registry data. Bone Joint Res 2023; 12: 103–12