Physical Function Following Megaprosthesis Surgery in the Lower Limb—A Prospective Cohort Study of 38 Patients

Abstract

Background

Limited knowledge exists on early outcomes postmegaprosthesis surgery. Yet, understanding these outcomes is crucial for patient counseling. This prospective study investigates physical function outcomes in lower limb megaprosthesis surgery patients, employing both objective and subjective measures and explores associations between them.

Methods

A total of 38 patients underwent treatment with a proximal femur, distal femur, or proximal tibia megaprosthesis. Muscle strength tests, range of motion, Timed Up and Go test, and the Musculoskeletal Tumor Society score were conducted at 4, 8, and 12-18 months postsurgery. Repeated measurements analysis was performed along with a comparison between treated limb, untreated limb, and predictive values.

Results

Bilateral muscle strength reduction was observed, with the proximal tibia group showing the most pronounced deficits. None of the groups exhibited changes in strength over time. All groups had decreased joint flexion in the treated limb compared to the untreated limb 12-18 months postsurgery. Timed Up and Go performance improved in all groups, but remained below average compared to reference values. An association was observed between a lower Timed Up and Go Test and higher Musculoskeletal Tumor Society scores, with the latter being lowest in the proximal tibia group.

Conclusions

This study provides comprehensive first-year outcome data following megaprosthesis surgery, offering valuable information to guide patient care. Markedly reduced strength in both treated and untreated limbs compared to predicted normal values was observed. The significant deficit in walking ability was clearly associated with patient-reported outcomes.

Introduction

Limb salvage techniques are the gold standard treatment for the management of malignant bone tumors [1] and an estimated 85% of patients with extremity sarcoma are successfully treated with limb salvage procedures [2]. This is mainly due to advances in multimodal therapies resulting in increased 5-year survival rates, ranging from 60 to 90% [3,4], and surgery still remains the primary treatment modality of extremity sarcoma. To achieve negative margins the surgical procedure is extensive and often the entheses and surrounding soft tissue are compromised leading to a significant reduction in physical function [5,6].

Clinical outcomes in sarcoma patients are frequently assessed using Patient-Reported Outcome Measurements [[7], [8], [9]]. However, according to the World Health Organization's International Classification of Functioning, Disability, and Health [10] physical function should be evaluated from both an objective perspective to assess physiological function (body function) and anatomical structures, as well as from a subjective viewpoint to evaluate task management (activity) and social engagement (participation). This approach offers researchers a more comprehensive assessment of the patients’ postoperative functioning and enables physicians to provide patients with a clearer understanding of their postoperative expectations.

Long-term studies have investigated various aspects of physical function after reconstructive surgery [[11], [12], [13]]. However, few prospective studies have focused on the early postoperative functions [[14], [15], [16]]. This study addresses physical function during the initial year following lower limb megaprosthesis surgery, specifically focusing on muscle strength, range of motion (ROM), Timed Up and Go (TUG) test, and overall activity levels measured with the Musculoskeletal Tumor Society (MSTS) score.

Material and methods

Patients

Between January 2019 and May 2022, a total of 48 patients were treated with a knee or hip megaprosthesis at the Aarhus University Hospital. Patients above 18 years with remaining life expectancy of more than 12 months were eligible for inclusion. Of the 48 treated patients, 40 were eligible for inclusion. Two declined participation, and a total of 38 patients (24 men and 14 women) were enrolled in the study with 21 receiving a megaprosthesis of the proximal femur (PF), 12 of the distal femur (DF), and 5 of the proximal tibia (PT). Table 1 summarizes patient characteristics.

Table 1.

Patient characteristics.

Characteristics	Total	PF	DF	PT
	n = 38	n = 21	n = 12	n = 5
Gender, n (male:female)	24:14	15:6	7:5	2:3
Age, y
Median range	64 (19-82)	69 (23-82)	61 (19-79)	50 (31-67)
BMI, kg/m2
Mean SD	27 (5)	27 (5)	27 (6)	27 (6)
Follow-up, days
4 mo, median range	123 (90-153)	122 (101-146)	123 (104-153)	133 (90-143)
8 mo, median range	246 (200-284)	247 (200-276)	244 (217-284)	246 (221-272)
12-18 mo, median range	386 (326-555)	386 (326-555)	392 (362-519)	386 (365-506)
Resection length, cm
Median range	14 (8-27)	15 (10-20)	13 (8-27)	14 (10-21)
Indication for surgery, n
Malignant tumor	10	1	7	2
Benign tumor	5	0	3	2
Metastases	17	15	1	1
Other	6	5	1	0
Revision surgery, n
Infection	4	2	1	1
Polyethylene wear	2	2	0	0
Aseptic loosening	2	2	0	0
Dislocation	2	2	0	0
Nonunion	2	2	0	0
Metal ion elevation	1	0	1	0
Expandable prosthesis revision	1	0	1	0
Surgical muscle loss, n
None	20	7	9	4
Minor	17	13	3	1
Extensive	1	1	0	0
Chemotherapy, n
None	24	11	9	4
Neoadjuvant	0	0	0	0
Adjuvant	10	7	2	1
Neo- and adjuvant	4	3	1	0
Radiation, n
None	32	16	12	4
Neoadjuvant	2	2	0	0
Adjuvant	4	3	0	1
Neo- and adjuvant	0	0	0	0

BMI, body mass index; SD, standard deviation.

Revision surgery was defined as having undergone at least 1 previous arthroplastic procedure at the same anatomical site.

The extent of muscle resection was evaluated intraoperatively based on direct surgical observations. Resections were classified as follows:

• •None: intraosseous tumor location with no muscle involvement,
• •Minor: partial resection of muscle or tendon,
• •Major: extensive soft tissue resection, typically due to tumor involvement of surrounding musculature.

The megaprostheses and procedures utilized included the Modular Universal Tumor and Revision System (MUTARS, Implantcast GmbH, Buxtehude, Germany) for hip surgeries (both total and hemi hip arthroplasty), and the Global Modular Replacement System (Stryker, Kalamazoo, Michigan) for knee surgeries (both DF and PT). The Trevira tube (Implantcast, Buxtehude, Germany) was applied for soft tissue reattachment in all hip surgeries. All Global Modular Replacement System knee reconstructions included a medial gastrocnemius flap and skin transplantation for prosthesis coverage. All patients received standard postoperative physiotherapy.

Three follow-up assessments were conducted at 4, 8, and 12-18 months. Figure 1 illustrates the inclusion and loss to follow-up. All tests were performed by a single examiner.

Figure 1.

Patient flow chart.

Muscle strength

A Biodex System 3 PRO dynamometer (Biodex Medical Systems, Shirley, NY, http://www.biodex.com) was used to determine maximal muscle strength. The highest peak torque (Newton-meters) achieved was used in further analysis. Test protocols were based on the Biodex System 3 Pro manual and a prior study by Harbo et al [17]. Participants started with a warm-up, following test with standardized audio instructions. Visual and manual checks ensured joint and dynamometer alignment. Isokinetic tests included 8 maximal reciprocal contractions with 15-second rests, while isometric tests involved 3 maximal 5-second contractions with 40-second rests. Isokinetic knee assessments used an 80° ROM and a 90°/second velocity, while hip assessments used a 60° ROM and a 60°/second velocity. Isometric strength was evaluated at neutral joint positions (knee at 70°, hip at 40°). Tests were repeated or excluded based on a variation threshold (30% for knee, 40% for hip) to ensure maximal performance and to avoid outliers [18].

A predictive value (PV) was calculated using reference equations from Harbo et al. [17], incorporating sex, body mass, height, and age, to estimate expected muscle strength in each patient.

ROM

ROM measurements were performed using a universal goniometer with 1° accuracy, aligning its fulcrum according to the joint's rotational axis. The outermost positions were recorded with the contralateral leg fixed in a neutral position. The measurements were performed in the following sequential order: hip extension, hip flexion, hip abduction, hip adduction, hip internal rotation, hip external rotation, knee flexion, and knee extension. ROM assessment was conducted with the patient in a supine position, except for hip extension, which was measured with the patient in a full lateral position. The untreated limb was examined first.

TUG test

TUG was conducted once at each follow-up, following established study protocols [19]. Patients began the test seated, rose, walked 3 meters, turned, returned to the starting position, and then resumed a seated position. The entire task was timed to assess functional mobility and balance.

MSTS Score

The MSTS form was completed by the examiner in collaboration with the patient to ensure accuracy. The MSTS score assesses functional outcome based on 6 domains: pain, function, emotional acceptance, support, walking ability, and gait. Each domain is evaluated on a Likert scale from 0 to 5, with higher scores indicating better function. Hence, the maximum total score is 30 [20].

Data analyses

Gaussian distribution was evaluated using histograms and q-q plots. Normally distributed data are summarized by means and standard deviations, while non-normal data are summarized by medians and range. A repeated measures mixed effect model was used to analyze muscle strength, TUG, ROM, and MSTS scores over time, with time as a fixed effect and patient as a random effect. Tukey’s multiple comparisons test was applied if differences were observed. Model performance was validated through q-q plots and residual histograms [21]. Spearman's Rank Correlation assessed correlations between variables.

Sample size calculations, based on prior study data [17], indicated a minimum of 9 patients per group to detect a 25% strength reduction with 0.05 significance and 0.80 power.

Study data were collected and managed using REDCap [22,23] electronic data capture tools hosted at Department of Clinical Medicine, Aarhus University, Denmark.

The analyses were conducted using Stata version 17.0 and GraphPad Prism version 10.1.1. This study adheres to the reporting guidelines in STROBE.

Results

Muscle strength

PF patients exhibited an overall strength deficit in the treated limb compared to both the untreated limb and PV in all strength tests at all 3 time points (Fig. 2 and Supplementary Data A. At 12–18-month follow-up, a strength deficit of 20%-34% and 34%-51% was observed in the treated limb compared to the untreated limb and PV, respectively. Additionally, a 7%-27% strength deficit was observed in the untreated limb compared to PV.

DF patients exhibited an overall strength deficit in the treated limb compared to both the untreated limb and PV, with a tendency to smaller deficits in the hip flexion and hip extension test (Fig. 3 and Supplementary Data A. At 12–18-month follow-up, a 38%-57% knee strength deficit was observed in the treated limb compared to the untreated limb. Similarly, a deficit of 25%-60% was noted compared to PV. Both treated and untreated limbs exhibited a 33% deficit in isometric hip flexion compared to PV.

Figure 3.

Mean peak torque at 4, 8, and 12–18-month follow-up after megaprosthesis treatment in the DF, comparing a predictive model (red), the untreated limb (green), and the treated limb (blue). Nm, Newton meter.

PT patients exhibited an overall tendency of strength deficits in the treated limb compared to both the untreated limb and PV, except for hip strength at 8 and 12–18-month follow-up (Fig. 4 and Supplementary Data A. A 43%-70% knee strength deficit was observed in the treated limb compared to the untreated limb at 12-18 months. A 56%-77% knee strength deficit and a 34%-46% hip strength deficit were observed compared to PV. The untreated limb exhibited strength deficits of 22%-43% compared to PV.

Figure 4.

Mean peak torque at 4, 8, and 12–18-month follow-up after megaprosthesis treatment in the PT, comparing a predictive model (red), the untreated limb (green), and the treated limb (blue). The missing data point for isokinetic knee flexion at 4 months is due to patients being unable to complete the evaluation satisfactorily within the set variation limits. Nm, newton meter.

Although mean strength improvements in the treated limb ranged from 7 to 66% when comparing 4 months to 12-18 months, no statistically significant improvement over time could be demonstrated (Figure 2, Figure 3, Figure 4 and Supplementary Data B).

Figure 2.

Mean peak torque at 4, 8, and 12–18-month follow-up after megaprosthesis treatment in the PF, comparing a predictive model (red), the untreated limb (green), and the treated limb (blue). Nm, Newton meter.

A significantly poorer outcome was observed in patients with minor muscle resections compared to no muscle resection (Fig. 5), and in patients who had revision surgeries compared to primary surgeries (Fig. 6), in the PF group. The same subgroup analysis was unfeasible in the DF and PT groups due to the small size of the patient cohort.

Figure 5.

Mean strength deficits between patients undergoing no muscle resection and minor muscle resection are represented as mean values with 95% CIs. Stars indicate statistically significant differences between the groups.

Figure 6.

Mean strength deficits between patients undergoing primary surgery and revision surgery are represented as mean values with 95% CIs. Stars indicate statistically significant differences between the groups.

ROM

In the PF group, there was an increased mean ROM for hip internal and external rotation, along with a decrease in mean ROM for hip extension, when comparing the treated limb to the untreated limb at all 3 time points (Fig. 5 and Supplementary Data C-E). At 12–18-month follow-up, the treated limb in the PF group showed decreased mean ROM in hip flexion (97° vs 110°) and hip extension (4° vs 8°), and an increase in hip internal rotation (41° vs 28°) and external rotation (59° vs 38°). In the DF group, a decrease in knee flexion (114° vs 129°) and hip internal rotation (25° vs 31°) was observed in the treated limb. In the PT group, a decrease in hip extension (10° vs 12°), hip internal rotation (29° vs 36°), and knee flexion (110° vs 130°) was observed in the treated limb.

TUG test

At the 4-month follow-up, the PT group exhibited the longest completion time, with a mean difference of 6 seconds (95% confidence interval (CI): −13 to 2) compared to the PF group and 9 seconds (95% CI: −17 to −0.4) compared to the DF group (Fig. 5 and Supplementary Data F). At the 12–18-month follow-up, all 3 groups demonstrated equal TUG completion times.

MSTS Score

At the 4-month follow-up, MSTS scores exhibited similarities across the 3 groups (Fig. 6 and Supplementary Data G). Subsequently, both the PF and DF groups displayed enhanced mean scores at the 12–18-month follow-up, whereas the PT group did not demonstrate such improvement.

A positive correlation was observed across muscle strength variables in both treated and untreated limbs (Table 2). A negative correlation between TUG and MSTS scores was also identified (rho = −0.66, 95% CI: −0.78 to −0.49) (Fig. 7).

Table 2.

Spearman's rank correlation analyses of muscle strength variables presented with Rho and 95% CIs.

Variables		1	2	3	4	5	6	7	8	9	10	11	12
1	Untreated limb, isokinetic knee extension		0.6-0.9	0.6-0.8	0.4-0.7	0.7-0.9	0.6-0.8	0.3-0.6	0.5-0.8	0.2-0.6	0.6-0.8	0.6-0.8	0.6-0.9
2	Untreated limb, isokinetic knee flexion	0.766		0.5-0.8	0.5-0.8	0.6-0.8	0.4-0.7	0.06-0.5	0.5-0.8	0.1-0.5	0.5-0.8	0.5-0.8	0.5-0.8
3	Untreated limb, isometric knee extension	0.749	0.640		0.3-0.7	0.4-0.7	0.5-0.8	0.01-0.5	0.2-0.6	0.02-0.5	0.3-0.7	0.3-0.7	0.5-0.8
4	Untreated limb, isokinetic hip extension	0.607	0.664	0.525		0.5-0.8	0.4-0.7	0.4-0.7	0.4-0.8	0.5-0.8	0.6-0.8	0.4-0.7	0.3-0.7
5	Untreated limb, isokinetic hip flexion	0.811	0.711	0.599	0.697		0.7-0.9	0.4-0.8	0.6-0.8	0.4-0.7	0.6-0.9	0.7-0.9	0.6-0.8
6	Untreated limb, isometric hip flexion	0.718	0.585	0.635	0.614	0.807		0.4-0.7	0.4-0.8	0.4-0.7	0.4-0.8	0.6-0.8	0.6-0.8
7	Treated limb, isokinetic knee extension	0.474	0.312	0.260	0.599	0.625	0.565		0.6-0.9	0.9-1.0	0.4-0.8	0.3-0.7	0.2-0.7
8	Treated limb, isokinetic knee flexion	0.670	0.654	0.431	0.605	0.730	0.605	0.777		0.6-0.9	0.7-0.9	0.6-0.8	0.5-0.8
9	Treated limb, isometric knee extension	0.414	0.339	0.262	0.636	0.595	0.587	0.926	0.768		0.5-0.8	0.3-0.7	0.3-0.7
10	Treated limb, isokinetic hip extension	0.735	0.649	0.496	0.720	0.769	0.634	0.614	0.798	0.641		0.8-0.9	0.6-0.9
11	Treated limb, isokinetic hip flexion	0.741	0.668	0.531	0.596	0.826	0.712	0.501	0.718	0.522	0.868		0.8-0.9
12	Treated limb, isometric hip flexion	0.772	0.635	0.645	0.509	0.734	0.716	0.472	0.663	0.518	0.778	0.886

Figure 7.

Correlation between MSTS score and TUG, with rho = −0.66 and 95% CI = −0.78 to −0.49.

Supplementary data H provides detailed correlations between TUG and MSTS scores, as well as ROM and muscle strength in the PF, DF, and PT groups.

In the PF group, TUG exhibited weak association with most variables, but moderate correlations with hip extension (negative) and knee extension (positive). In the DF group, TUG had weak correlations overall, but a moderate negative association with hip adduction. In the PT group, TUG showed moderate to strong correlations with ROM (negative: knee/hip flexion, hip rotations; positive: knee extension). TUG was moderately associated with isokinetic knee flexion (positive) and isometric knee extension (negative).

MSTS scores showed weak associations with ROM and muscle strength in the PF group. In the DF group, MSTS had weak associations overall, except for moderate correlations with hip flexion (positive) and isokinetic knee/hip flexion (negative). In the PT group, MSTS scores were weakly associated with most variables but showed moderate positive correlations with hip and knee flexion and a strong positive correlation with hip external rotation.

Discussion

The life expectancy for cancer patients, even with metastatic disease, continues to improve due to advancements in novel treatments and improved diagnostics. Consequently, assessing physical function postsurgery has become a crucial metric when reporting outcomes in orthopaedic oncology. While patient-reported outcome measures are widely employed as a standard method for evaluating and documenting various outcomes related to patient health and well-being, there is a growing advocacy for incorporating more objective measures [10].

This study investigated muscle strength, ROM, TUG, and MSTS scores in PF, DF, and PT megaprosthesis patients at 4, 8, and 12-18 months post-treatment. No statistically significant increase in muscle strength in the treated limb was observed over time. Compared to PVs, bilateral muscle strength reduction was observed, with a more pronounced reduction noted in the treated limb, with the PT group exhibiting the most significant deficits. All groups exhibited decreased joint flexion in the treated limb compared to the untreated limb, and the PF group showed reduced extension and increased hip rotation. TUG performance improved over time in all groups, with comparable results at 12-18 months. At 4 months, MSTS scores were similar, but the PT group had a less substantial increase over time. A significant positive correlation was identified among all muscle strength tests for both treated and untreated limbs, along with a notable negative correlation between TUG and MSTS scores. Overall, TUG was negatively correlated with ROM, except for a positive correlation with knee extension. Spearman’s rho was strongest in the PT group. TUG correlation with muscle strength was overall weak. MSTS had an overall positive correlation with ROM.

The predominant approach in lower limb megaprosthesis surgery research comprises retrospective or cross-sectional studies [24]. This study represents one of the first prospective cohort studies examining physical function outcomes in megaprosthesis patients. Our finding of bilateral strength reduction compared to PV aligns with Fernandes et al’s study [11], with 2 to 12 years postoperative follow-up. Direct group comparisons are challenged, given that the data from the study by Fernandes et al were combined from all 3 surgical sites. Comparing our 12–18-month follow-up results with Bernthal et al’s long-term study (2.5 to 28 years) reveals more significant knee extension deficits in PT arthroplasty (84 vs 60%) and comparable deficits in PF (35% vs 34%) and DF arthroplasty (53% vs 57%) [12].

In contrast, Bernthal et al found smaller deficits in knee flexion compared to our PF (11% vs 20%) and DF (2% vs 38%) arthroplasty, though PT deficits were similar (43% vs 43%) [12].

Bernthal et al attributes pronounced PT deficits to patella tendon enthesis removal and reconstruction with a medial gastrocnemius flap. Over time, an increase in knee extension deficit in PT arthroplasties and a decrease in knee flexion strength deficit in PF and DF arthroplasties could be expected.

To our knowledge, no other studies have conducted such versatile examinations as this study. Given the moderate to very strong positive correlations across all strength tests, similar deficits are likely in long-term studies.

Primary tumor surgeries, especially for malignant cases, require more muscle excision to ensure clear margins and reduce recurrence, unlike surgeries for benign or metastatic tumors, which prioritize muscle preservation to maintain function. Patients undergoing revision surgeries with minor muscle resections had worse outcomes than those with primary surgeries and no resections. Revision surgeries often involve further muscle resection and lead to increased scar tissue, contributing to poorer outcomes. Notably, no prior studies have been found on this topic.

The lack of improvement in muscle strength over time is a noteworthy and concerning finding. Several factors may contribute, including reduced physical conditioning before surgery due to pain, immobility, or tumor-related disuse. In malignant cases, neoadjuvant therapies may also negatively affect muscle function. Additionally, the extent of soft tissue resection during megaprosthesis implantation can lead to permanent functional deficits. Postoperative rehabilitation may be limited by patient age, comorbidities, or reduced participation.

Compared to Fernandes et al, similar ranges of motion were observed for knee flexion and extension in both DF and PT reconstruction at 12–18-month follow-up. However, Fernandes et al noted better flexion in the untreated limb compared to our study. Mean hip flexion and extension after PF reconstruction were lower in our study than in Fernandes et al's. The reduced flexion and extension suggest potential improvement over time in both limbs, except for the treated limb after knee reconstruction.

Studies on adult patients receiving megaprosthesis often lack TUG measures. Morri et al [13] conducted a prospective cohort study on tumor knee endoprosthesis patients, performing TUG tests at 3, 6, and 12 months postoperatively. Results showed mean TUG times of 14, 9, and 7 seconds, with a 48% improvement over time. The lower median age in Morri et al's study (19 years, range 9-66) may contribute to the observed improvement compared to our study. In our study, the mean TUG time for all groups at 12–18-month follow-up is below average [25].

At 12–18-month follow-up, mean MSTS scores in our study (19 for PF, 20 for DF, and 15 for PT) are lower than the minimum values reported by P.J. Wilson et al based on a systematic review of 28 studies [24]. The review indicated mean MSTS score ranges of 21-28 for PF, 22-27 for DF, and 20-26 for PT, with follow-up periods ranging from 24 to 184 months. The longer follow-up in Wilson's review makes direct comparison unfeasible but suggests a potential increase in MSTS scores over time. However, MSTS scores might not adequately reflect the true functional capacity of patients with outcomes surpassing postoperative expectations, primarily due to ceiling effects [26,27].

Limitations

The small sample size limits the study’s generalizability, statistical power, and may explain the wide range of strength improvement (7%-66%) and the lack of significant gains in treated limbs. The cohort's heterogeneity introduces confounding variables and varying treatment responses, complicating interpretation and generalization. Missing data, common in sarcoma patients due to illness severity or physical limitations, were adjusted for in the analysis.

Intraoperative classification of muscle resection may involve inter-surgeon variability. Similarly, defining revision surgery as at least one prior arthroplasty procedure may not fully capture the surgical complexity or history.

Additionally, demanding tests may lead to patient withdrawal, potentially biasing results to appear more favorable.

Conclusions

Our study provides a comprehensive analysis of both subjective and objective outcome measures the first year following megaprosthesis surgery of the PF, DF, or PT. Marked strength deficits were observed in the treated limb across all regions, with knee strength reduced by up to 70% and hip strength by up to 46% compared to predicted normal values. Even the untreated limb showed deficits of up to 43%, indicating a broader functional impact. These objective deficits were mirrored by impaired functional mobility and closely associated with patient-reported outcomes. While TUG effectively captured overall mobility, dynamometry provided more specific insights into muscle deficits that can be targeted in rehabilitation programs.

Declaration of generative AI and AI-assisted technologies in the writing process

During the preparation of this work the authors used ChatGPT in order to improve readability and language. After using this tool/service, the authors reviewed and edited the content as needed and takes full responsibility for the content of the publication.

Ethics

The study adheres to the principles of the Helsinki Declaration and has obtained approval from the local ethics committee (case no. 1-10-72-268-18) and the Danish Data Protection Agency (case no. 1-16-02-363-18). Written consent was obtained from all patients before their inclusion.

Funding

The study received support from an unrestricted grant provided by Fisher Medical, Denmark and Holms Mindelegat.

Data availability

Anonymized data are available upon request.

Conflicts of interest

Sarah Stammose Freund received salary from an unrestricted grant provided by Fisher Medical, Denmark (ID: AGR-2017-731-2766) and Holms Mindelegat (ID: 20006-1936).

Thomas Baad-Hansen received research support from Fischer Medical as a principal investigator and is a board member/committee appointments at EMSOS.

The other authors declare there are no conflicts of interest.

For full disclosure statements refer to https://doi.org/10.1016/j.artd.2025.101754.

CRediT authorship contribution statement

Sarah Stammose Freund: Writing – original draft, Visualization, Validation, Resources, Project administration, Methodology, Investigation, Funding acquisition, Formal analysis, Data curation, Conceptualization. Michael Melchior Bendtsen: Writing – review & editing, Resources, Conceptualization. Bjarne Hauge Hansen: Writing – review & editing, Resources, Conceptualization. Henning Andersen: Writing – review & editing, Validation, Supervision, Resources, Methodology, Conceptualization. Thomas Baad-Hansen: Writing – review & editing, Supervision, Resources, Project administration, Methodology, Conceptualization.