Ethical approval for conducting the study was granted from the Ethics Committee of the Medical Faculty of the University of Rostock as well as the Ethics Committee of the Westphalia-Lippe Medical Association and the Westphalian Wilhelms University of Münster (Ethics ID 2020-497-b-S, approval 25th July 2020). The study was carried out in accordance with the institutional research committee's ethical standards.

This retrospective cohort study enrolled active and former soldiers of the German Armed Forces who were diagnosed with a primary complex impairment of the musculoskeletal system following wounding, injuries or severe orthopedic illnesses and/or a primary psychological disorder indicative of posttraumatic stress disorder (PTSD). Patients underwent an initial assessment followed by a three-week MVR. Patients were categorized into an intervention group receiving LTM and a control group not receiving medication based on their medication status. Subgroups were further delineated based on the primary diagnosis. Regardless of their medication status, all patients included exhibited at least one physical impairment. Medication considered for inclusion was prescribed for preventing or alleviating a chronic illness persisting for at least 3 months, while medication used for a case of need that had not yet occurred or taken without continuity was not considered. Data collection spanned from 2011 to 2020, with assessments carried out at the initiation of rehabilitation, encompassing performance diagnostic evaluation conducted subsequent to initial interviews. The inclusion criterion included soldiers demonstrating a significant risk to fitness for duty or ability to perform previous occupational tasks. The exclusion criteria included patients lacking a requirement for somatic rehabilitative measures, those diagnosed with cancer or mental illnesses other than PTSD, and those lacking performance diagnostic assessments.

A total of 381 patient records were analyzed. In this study, the use of at least three long-term medications was defined as the cutoff for polypharmacy (4). Medication data included the name, dosage and frequency of administration of the active ingredient. Drug-induced interactions were identified using Scholz database (as of 2021/22) and were then categorized as moderate or severe by applying the proprietary traffic light system of the software. Additionally, interactions were classified based on the international consensus of potential clinically significant drug interactions in the elderly population (21). To distinguish between the effects of drugs with differing primary mechanisms, medications were categorized into their respective anatomical groups utilizing ATC. Notably, drugs categorized under ATC group N were selected for comparison against other groups. Somatically focused rehabilitation interventions included individually tailored sports programs for all patients based on performance diagnostic examinations.

The maximum performance per kilogram body weight [p-max. in Watts (W)] and performance at lactate level of 4 mmol/L in capillary blood [p-4-mmol/l-Lac. (W)] were assessed through bicycle ergometry. The bicycle ergometer “Schiller ERG 911 S plus” (Schiller Medizintechnik, Feldkirchen, Germany) was used with load adjustments following the German Association for Sports Medicine and Prevention guidelines, employing a step profile with a 50-Watt increase every three minutes (22). Maximum performance was defined as the point of test termination based on subjective stress limits or other predefined termination criteria. The capillary blood lactate concentration was monitored at three-minute intervals throughout the test (21) with a detailed procedure previously described elsewhere (21). Using the “Winlactat” software (Mesics GmbH, Münster, Germany), both the maximum power output and the power at which the patient surpassed the threshold of 4 mmol/L lactate (p-4-mmol/l-Lac.) were calculated through regression analysis.

The data were checked for normal distribution through Q–Q plots and Shapiro–Wilk tests. Descriptive and inferential nonparametric procedures were utilized for statistical analysis. In the descriptive analysis, sample size (N), standard deviation (SD), median (Mdn) and the interquartile range (IQR) of the spiroergometric measurement data were assessed relative to medication at the time of initial rehabilitation with documented performance data thereafter. Differences in performance within and between medication groups were examined using Mann–Whitney U and Kruskal–Wallis tests with multiplicity adjustment according to Bonferroni correction. The associations between diagnostic performance and medication use were investigated using Spearman bivariate correlation analyses. The correlation strength was interpreted according to Cohen (22). The local significance level was set to α = 5%. Two-sided P-values were considered to be noticeable (“significant”) in case p ≤ 0.05. The software SPSS Statistics Version 27 (IBM, Armonk, USA) and Scholz database (ePrax GmbH, Munich, Germany) were used to perform statistical analyses, whereas GraphPad Prism 6 was used to create the figures.

A total of 314 drugs were identified to be taken on a long-term basis, with an average of 2.9 (±2.0) drugs per patient. The drugs were distributed as shown in Figure 1a.

Upon analyzing performance data across the entire study population, no significant differences were detected in terms of the maximum power output (p-max) or maximum lactate threshold (p-4 mmol Lac.) concerning the use of LTM. However, among patients taking ≥ 3 drugs, a significant reduction in maximum performance was evident compared to those not taking any medication [n = 45, 217 (200–233) W vs. n = 67, 225 (208–267) W, P = 0.041]. Maximum performance correlated negatively with the number of long-term drugs with low strength (Table 2).

A total of 283 long-term drugs were reported in this group, averaging 3.0 (±2.0) medications per patient. The distribution of these drugs is depicted in Figure 1b. 72 patients were identified with medication comprising at least two long-term drugs, resulting in 122 potential drug interactions. Of these interactions, 81 were classified as potentially moderate, and 41 were classified as potentially severe. On average, each PTSD patient experienced 1.6 (±1.4) drug interactions. 49 patients (68.1%) were at high risk of experiencing an interaction due to their medication, whereas 31.9% of the drug combinations posed a low risk. Notably, 76.6% of the entire PTSD group had at least one potentially moderate drug interaction, and 53.2% had at least one potentially severe drug interaction. Additionally, among the 72 patients receiving multiple medications, 17 (23.6%) exhibited potential interactions of both moderate and severe severity. Figure 2 illustrates the potential effects of DDIs and their frequency of occurrence in PTSD patients.

PTSD patients on medication exhibited significantly poorer maximum performance unmedicated patients (Figure 3). The presence of medication was inversely correlated with maximum performance, albeit weakly (Table 2). Furthermore, analysis of LTM intake indicated that patients using more than three LTMs performed significantly worse than controls (Figure 3 and Table 2). This decline in maximum performance was observed to be proportional to an increase in the number of drugs taken.

Classifying active substances according to the ATC criteria revealed that combining substances from two different ATC groups significantly worsened the maximum performance compared to no medication. However, no additional significant differences in maximum power were observed when using active substances from three or more groups (Figure 3).

The occurrence of polypharmacy was indicated to be a potential influencing factor, with formulations containing three or more active ingredients showing a significant effect (Figure 3). The maximum performance correlated negatively with the presence of polypharmacy in a low manner (Table 2). Since polypharmacy is associated with an increased risk of DDI, potential DDIs were identified and categorized. Among individuals with PTSD, those exposed to a DDI demonstrated significantly inferior maximal performance compared to controls taking fewer medications. Moderate to severe interactions did not yield statistically significant disparities in maximal performance (Figure 3). No significant difference in performance was observed with p-4 mmol-Lac compared to other medication categories tested.

A total of 31 drugs, with an average intake of 2.4 (±1.4) drugs per patient, were identified among individuals with somatic diseases and LTM (Figure 1c).

There were no significant differences in p-max or p-4 mmol-Lac between the two groups in terms of maximum performance on bicycle ergometry. compared to the no-medication group (Table 3). Moreover, there were no statistically significant differences in p-max or p-4 mmol-Lac. those taking either 1–3 or >3 long-term medications were compared to those in the no-medication group.

As shown in Table 3, the comparison of the maximum performance achieved between PTSD patients and somatically impaired patients revealed no significant difference. In addition, maximum performance did not differ significantly in the presence of medication. Among patients without medication, those with PTSD achieved significantly greater maximum performance values compared to the somatic group at baseline.

The comparison of performance at a 4 mmol/L lactate concentration in capillary blood indicated no significant difference, which did not change depending on the LTM.

We observed a notable decline in maximum performance among PTSD patients taking LTM. Notably, patients with greater medication complexity, such as those on multiple drugs or experiencing polypharmacy and DDIs, exhibited diminished performance levels. This suggests a perceived hindrance to rehabilitation efforts.

Recent findings in the literature highlighted significantly decreased physical fitness in depressive patients compared to healthy controls, which could not be explained by BMI, age, sex, or physical activity (23). However, our study revealed a distinct pattern among PTSD patients on LTM. Specifically, these individuals displayed markedly lower maximum performance levels and exhibited lower blood lactate concentrations than PTSD controls. Our findings indicate that factors such as age, weight, height, WHtR, and BMI did not significantly influence the observed decline in maximum performance. This result aligns with recently published literature in this area (23). These findings suggest that neither the presence of PTSD nor the anthropometric characteristics examined are primary determinants of reduced maximum performance.

As we focused on medication, we attributed the observed differences in performance primarily to a combination of medication-related factors, as illustrated in Figures 2, 3. These factors will be thoroughly examined and discussed.

In addition to the observed findings in PTSD patients, there was no significant decrease in maximum performance among patients with somatic impairments (Table 3), suggesting no medication-induced effects on performance. While pain can discourage physical activity and pharmacological interventions may improve mobility to some extent, weakness and fatigue present obstacles to mobilization (41, 42).

Consequently, the impact of LTM on the physical performance of our patients may neither be detrimental nor advantageous and may vary depending on impairment severity and musculoskeletal impairment.

The study showed that PTSD patients achieved significantly greater maximum performance values than patients in the somatic group did compared to primarily somatically impaired patients without existing LTM (Table 3), but there was no significant difference at the 4 mmol/L lactate threshold. Anthropometric data comparison revealed that somatically impaired patients were significantly younger and lighter than PTSD patients were (Table 1). This is plausible considering that orthopedic complaints are associated with physical impairment, leading to restricted p-max in somatically impaired patients, while PTSD patients do not exhibit primary physical limitations.

To the best of our knowledge, this is the first retrospective cohort study to identify LTM and assess the potential relationship between medication and physical performance in middle-aged patients undergoing rehabilitation. Nevertheless, the present study has limitations.

First, data were retrospectively extracted from paper-based patient files with medication examined in terms of existing LTM, number of medications and the presence of polypharmacy and drug interactions. As there was no patient contact at any time during the data collection, it was not possible to carry out a comprehensive medication review. However, the factors mentioned are possible influencing factors that are strongly associated with a negative effect on the overall performance outcome (24, 33, 35, 43, 44). Future studies should include patient interviews that systematically examine influencing factors to include and further assess any deviations.

Secondly, the German armed forces' healthcare system involves multiple interfaces, potentially leading to communication challenges and diverse drug-related issues. Involving public pharmacies may introduce risks of medication errors and inaccurate documentation, as outlined in the Swiss cheese model (42). Loss of information may contribute to misjudgments regarding the actual medication regime. This limitation serves a crucial purpose in highlighting the necessity for medication.

Thirdly, self-procuring over-the-counter medication adds complexity to medication management. For example, the use of herbal agents and synthetic sleeping pills without physician consultation represents unreported cases, further complicating medication assessment.

An optimum individual performance capacity is required to be effective during military service. Understanding the diverse physiological characteristics and sex-specific pharmacology may offer further insights into the relationship between medication and performance, particularly in individual cases. Due to the limited number of female patients, such considerations could not be assessed.