This study was a monocentric, prospective, cross-sectional trial at a tertiary hospital. Ethical approval was provided by the Ethikkommission der Charité—Universitätsmedizin Berlin, Germany (Ethical Committee N° EA4/102/22). All patients were checked for eligibility and gave their written informed consent. The trial was registered prior to patient enrollment at WHO International Clinical Trials Registry Platform and German Clinical Trial Register (DRKS00029603, date of registration 07.07.2022). The study protocol was carried out in accordance with the Declaration of Helsinki and this manuscript adheres to the applicable CONSORT guidelines.

The trial was performed from the 7th of July 2022 to the 13th of September 2022. Adult patients meeting ASA Classification I or II criteria scheduled for elective surgery in general anesthesia were enrolled in this study. Patients with disease or surgery of the upper extremities, shoulders, or thorax, and patients who met any of the following criteria were excluded: systemic vascular disease, recent thrombotic events, local vascular disease of the upper extremities, pre-existing cardiac disease associated with systolic or diastolic dysfunction, venous puncture on the limb to be measured within two weeks before examination, frequent venous punctures in the history or vasoactive concomitant medication.

All procedures and measurements were performed in a tempered (19–21 °C) anesthesia induction room of a central operating theatre area. Patients were placed in a supine position on an operating table with armrests padded underneath and fixable at respective degree settings in the shoulder joint during the whole procedure (Fig. 1).

Figure showing the positioning of the arm in relation to the shoulder joint. a 0° position, 1 shows the measuring point of the cubital vein, and 2 is the measuring point of the cephalic vein. b 30° retroflexion position. c maximum retroflexion position (median 75° [Interquartile range 65–90])

All measurements were performed on the cubital vein (CuV) and cephalic vein (CeV) on the nondominant arm using a high-frequency linear ultrasound probe (8–18 MHz, ultrasound device: Vivid iq, GE Healthcare, Chicago, IL, USA). Optimal measurement positions were identified in the fossa cubiti and at the middle of the forearm by ultrasound (Fig. 1a). To ensure comparability, the position was marked and all subsequent measurements were taken at these markings. To avoid potential changes in vein geometry and size, the probe was placed gently on the skin to obtain an image without applying additional pressure. Each set of measurements included vein circumference (CI) and vein diameter in the out-of-plane probe position as well as vein diameter in the in-plane probe position to ensure accurate 2D measurements. CI was measured by tracing the edges and all diameters were measured from anterior to posterior (Supplemental Fig. 1).

All measurements were performed under the following patients’ conditions in consecutive order: 1) patient awake without tourniquet (awake-NT), 2) patient awake with tourniquet (awake-AT), 3) patient in GA without tourniquet (GA-NT), and 4) patient in GA with tourniquet (GA-AT). Within each condition, different levels of retroflexion (0°, 30°, and max°; Fig. 1a-c) were applied in a randomized order.

The patients waited in their respective positions for 60 s before the corresponding measurement was taken. After each measurement, patients rested in a supine position without a tourniquet and in 0° arm retroflexion for 2 min to ensure consistency in the starting position for each subsequent condition and arm position during measurement.

The maximum possible retroflexion (max°) was determined in advance by slowly rotating the patient´s outstretched arm passively backward from the horizontal axis. As soon as the maximum possible retroflexion was reached or the patient indicated paresthesia or pain, the arm was slightly returned from retroflexion until there was no more discomfort. The thus determined degree of retroflexion [°] corresponded to the max° position for each patient.

A tourniquet was applied using a 13.5 cm wide blood pressure cuff, placing it 2–3 cm above the cubital fossa and inflating it to 60 mmHg. This setting was selected because it demonstrated the greatest impact on vein enlargement and decreased compressibility when compared to other tourniquet techniques [17]. The optimal result was achieved when the cuff was inflated to a pressure of 60 mmHg and applied for 30–60 s. Following this period, no further increase in vein size was observed [10].

According to these previous findings, our patients remained in each condition for 60 s before measurements were performed and the strong effect of tourniquet application was confirmed in our work.

Apart from the study-related procedures described above, no further measures were taken on the nondominant arm throughout the whole study period. Monitoring for anesthesia and IV access were established in the dominant arm before the study measurement. Peripheral oxygen saturation as well as heart rate were measured continuously. Blood pressure was measured initially before awake measurements and with the beginning of GA induction in 2.5-min intervals on the dominant arm. Induction and maintenance of general anesthesia were performed according to institutional standard operating procedures. Induction was performed with an opioid followed by a bolus of propofol, maintenance with propofol or sevoflurane.

For analyzing measures to maximize vein size within a study cohort, a sample size calculation was performed in advance with a two-sided significance level of 5% and a power of 80%. Based on the previously reported effects of tourniquet application which range from 0.14–0.87 mm, with a maximum standard deviation of 1.3 mm, a mean change of 0.505 mm in vein diameter was assumed [12]. The calculation of the sample size yielded 54 patients for our study. To account for a potential dropout rate of 10%, the study was planned with a final total of 60 patients.

Data processing and analysis were performed using IBM SPSS Statistics (version 25; IBM, Armonk, NY, USA). Data are presented as median (Interquartile range [IQR]). Differences between measurements were evaluated using a paired Friedman test with Bonferroni-correction for multiple comparisons. Significance was set at P < 0.05, also for Bonferroni-correction after mathematical adjustment was applied. The effect size for non-parametric testing was estimated using a z-statistic by calculating correlation coefficient r, with > 0.1 representing a small, > 0.3 a medium, and > 0.5 a large strength of association, according to the recommendations of Cohen [18]. To analyze the accuracy of the measurements, the intra-class correlation coefficient was calculated between in-plane and out-of-plane measurements.

Patient demographics and hemodynamic/anesthesia data were reported as qualitative data and percentage or median [IQR], as applicable. Demographic as well as hemodynamic data and ventilation pressure settings were reported for transparency of the study cohort. No further statistical analyses were performed on this data.

Only patients with complete data sets were included in the final data analyses.