Pleth variability index and fluid management practices: a multicenter service evaluation

The introduction of a new technology has the potential to modify clinical practices, especially if easy to use, reliable and non-invasive. This observational before/after multicenter service evaluation compares fluid management practices during surgery (with fluids volumes as primary outcome), and clinical outcomes (secondary outcomes) before and after the introduction of the Pleth Variability Index (PVI), a non-invasive fluid responsiveness monitoring. In five centers, 23 anesthesiologists participated during a 2-years period. Eighty-eight procedures were included. Median fluid volumes infused during surgery were similar before and after PVI introduction (respectively, 1000 ml [interquartile range 25–75 [750–1700] and 1000 ml [750–2000]). The follow-up was complete for 60 from these and outcomes were similar. No detectable change in the fluid management was observed after the introduction of a new technology in low to moderate risk surgery. These results suggest that the introduction of a new technology should be associated with an implementation strategy if it is intended to be associated with changes in clinical practice.


Introduction
The management of intraoperative blood volume and the major complications of both hypovolemia and hypervolemia remain major and unresolved issues [1]. It has been suggested that dynamic parameters such as pulse pressure or stroke volume variation may, at least partially, solve this problem [2]. Dynamic parameters assess hemodynamic response to respiratory variations which allows to monitor the fluid-dependency status. They have been well validated for the prediction of the response to a fluid challenge [2]. The Pleth Variability Index (PVI), as a dynamic and non-invasive parameter, could fill a gap for the low to moderate risk patient and procedure, especially when cardiac output monitoring is not considered [2]. To improve fluid management, the optimization of the PVI value has been proposed [3,4] by targeting a PVI value between 10 and 13%. Optimizing this value may help optimize cardiac output without taking the risk of giving too much fluids (i.e. fluids not associated with a significant increase in cardiac output) (Fig. 1). However, the current ability of the introduction of PVI on its own to change the practice of clinicians is not known.
This multicenter service evaluation aims to establish whether the introduction of the PVI in the clinical practice is associated with a modification of the practice during low to moderate risk surgery. Forget et al. BMC Res Notes (2021) 14:293 Main text

Design and outcomes
This before/after multicenter service evaluation compares the fluid volumes used before and after the introduction of the PVI, focusing on fluids volumes (primary outcome), with clinical outcomes as quality indicators (secondary outcomes).
Thus, three consecutive phases were planned: a pilot survey (preparatory phase), a phase 1 (before the introduction of the PVI) and a phase 2 (after).
This work is presented following the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) Statement for cohort studies [5].

Ethics committee consideration
This multicenter service evaluation aims to assess how clinicians change their fluid management after the introduction of the PVI in three kind of surgery: Knee/hip arthroplasty, colorectal surgery. Written informed consent was waived by the ethics committee of the principal investigator (Patrice Forget, affiliated to the UCLouvain in 2012) (26th March 2012, Chairperson: Prof. J-M Maloteaux, Commission d'Ethique Biomédicale Hospitalo-Facultaire, Institution: UCLouvain), because considered as a practices audit, the essential information to be recorded being physicians' practices changes, without any breach in patient confidentiality. The project was registered before any data collection (Clinicaltrials.gov: NCT02271841) and performed in accordance with the ethical standards of the Declaration of Helsinki (1964) and its subsequent amendments.

Settings
The recruitment of a minimum of 3 centers was anticipated in Belgium and/or France who had no significant experience with the technology, to reach approximately 30 participating anesthesiologists. All the centers (n = 10) that worked within the network of the university hospital were considered.

Inclusion / exclusion criteria
Fives sites were recruited to participate in this research; each site may be in use of Masimo technology, but not PVI at the time. More than one anesthesiologist per center was recommended. Into these centers, all anesthesiologists had the opportunity to participate.
The project was limited to surgical procedures like knee/hip arthroplasties and colorectal surgery. To avoid a too high heterogeneity challenging data interpretation, patients under 18 years and ASA (American Society of Anesthesiologists) 4 were excluded.

Equipment
Standard of care monitoring for surgery included a pulsoxymeter, continuous electrocardiogram, non-invasive blood pressure monitoring, gas analyzers including capnography. No limitation was suggested in the use of other type of monitoring.
The use of Masimo Radical 7 devices SET technology with PVI feature was possible after the introduction of the technology, which was preceded by an appropriate training (delivered by PF, proposing a grey zone approach for the decision of fluid loading above a PVI value of more than 10 to 13% during more than 5 min) [3,4]. During this training, the determinants of cardiac output and the influence of intravascular blood volume were discussed. The advantages and disadvantages of the different types of advanced hemodynamic monitoring were presented and discussed with the participants. The information was based on a previously published expert consensus [2].

Statistical analysis
Based on previously published data in similar procedures, with low expected variability, a difference of 250 ml ± 250 ml could be anticipated [4]. This small difference was chosen to unmask any practice change, small or large in the fluid management, independently on any clinical significance. The type of surgery was considered for low variability, rather than being able to measure as well as hypervolemia (on the right) may be associated with postoperative complications. Fluid management associated with optimal cardiac output (green curve) may correlate with the lower rate of perioperative complications improvement in outcome. The clinical results were considered as secondary outcomes, forming part of a service evaluation. For this, we used the Dindo-Clavien and the Postoperative Morbidity Survey classifications [6,7].
A sample size calculation showed that at least 22 procedures recorded before and after (44 in total) would be sufficient to exclude the null hypothesis (being the absence of a difference of 250 ml ± 250 ml) with an alpha of 0.05 and a power of 90% [4]. We decided not to restrict the number of procedures, up to end 2018.
Data were collected and managed using REDCap (REsearch Electronic Data Capture) tools [8]. The forms were designed by the research team and tested with dummy data before use. Non-response bias was minimized as much as possible through automatic reminders and personal communication from the lead investigator. There was no incentive other than to receive the material for free and be part of the research project.
As normal distribution was rejected in most cases, Mann-Whitney U test was used for the comparisons. Data are given with mean ± SD or median [interquartile range 25-75] [IQR25-75], and 95% confidence interval (95%CI) as appropriate. A p-value of less than 0.05 was considered significant. No subgroup analysis, sensitivity nor other inferential analysis was planned. Statistica version 7.0 (STATSOFT, Tusla, USA) was used for all the analyses.

Pilot survey
A pilot survey was sent to 19 anesthesiologists to document the current practices (Additional file 1: Tables S1, S2, Figure S1). These nineteen were all from the 23 ultimately included who accepted to respond to this pilot questionnaire. This confirmed that an advanced and/or invasive monitoring was frequently used by most of the practitioners.

Participants
After the pilot survey, 75 anesthesiologists were approached, in five centers, by the local investigator and 23 completed the project (Additional file 1: Fig. S1). The reasons for not including the others were not documented. These practitioners progressively started the first phase (before the introduction of the new monitoring), including consecutive patients from February 2015 up to October 2017. After the introduction of the PVI, the second phase of the data collection occurred for all the centers during the year 2018.

Patients and procedures
In the five centers, 88 patients and procedures were followed ( Table 1). The follow-up was complete for 60 patients from the 88. Ventilation practices were slightly different (more frequently administered lower tidal volumes) but this was not statistically significant ( Table 1).
The total volume administered during surgery was similar when comparing the period before PVI introduction with the period after (primary outcome) (median [IQR25-75]: respectively, 1000 ml [750-1700] and 1000 ml ). The other outcomes were also similar ( Table 2, Additional file 1: Figure S2) (p > 0.05 for all the comparisons).

Discussion
No detectable change occurred before and after the introduction of the PVI. Particularly, no change was observed in the average amounts of fluids used during prosthesis and colonic surgeries after this introduction of a fluid responsiveness monitoring. However, this does not permit to exclude any change in the practice that may occur in particular patients, but not visible in aggregated analyses.
Moreover, the changes in practices may be limited by several factors, notably a probable evolution in the ventilatory practices, toward lower tidal volumes and slightly higher frequency, even if not statistically significant. Interestingly, the pilot survey showed that most of the practitioners declared that the institution does not have any written protocol, care guide or statement concerning hemodynamic management. This may introduce the question whether a new monitoring may help if not introduced in a protocol. An observation of this is the fact that, in a meta-analysis on goal-directed therapy (GDT), the effect of the intervention was particularly evident in older series, and namely in non-ERAS (enhanced recovery after surgery) programs [8]. One may speculate that the existence of a protocol (in this case, ERAS program) may have an important impact interacting with the added value of a new monitoring. These patients may also present a lower risk profile, and then benefiting less of a GDT. On the other hand, the introduction of a quality improvement program has been recently shown as complex, and not necessarily directly associated with improved outcomes [9]. The relative low introduction rate of the cardiac-output monitoring in the EPOCH trial reinforced our observation that invasive technologies would not be seen suitable in most low to moderate risk procedures [1].
The question remains whether a non-invasive technology may replace more invasive ones for low to moderate risk surgery. This study was not designed for, but the potential exists and would merit to be further explored. Exploratory analyses showed that 38/54 procedures (70%) before the introduction of the PVI did not reported any use of an invasive monitoring vs. 27/34 (79%) (p > 0.05) suggesting a shift toward less invasive monitoring.
In conclusion, this work shows that no detectable change in the fluid management was observed after the introduction of a new technology in low to moderate risk surgery.

Limitations
The limitations of this work are linked to the small sample size, small tidal volumes, laparoscopic procedures, low risk surgery, and the lack of generalizability in other contexts, like in higher risk surgery, but highlight the fact that even in low variability procedures, there is a need for implementation of protocols and a place for non-invasive monitoring.