A Randomised Controlled Study Comparing Pulse Pressure Variation (PPV) and Pleth Variability Index (PVI) for Goal-Directed Fluid Therapy Intraoperatively in Patients Undergoing Intracranial (Supratentorial ICSOLs) Surgeries

Fluid management intraoperatively has always been a contentious topic with many grey zones. The primary goal of fluid management is adequate volume replacement to maintain optimum cardiac output so that tissue perfusion is not hampered, and to prevent fluid overload to avoid deleterious effects on vital organs [1]. Studies have proven that liberal fluid resuscitation can lead to increased morbid conditions such as pneumonia, respiratory failure, pulmonary oedema, and prolonged hospital stay [2]. Maintaining fluid in neurosurgical cases with long intraoperative duration and haemodynamic fluctuations with increased risk of bleeding depending upon the tumour being operated on is arduous. The primary aim is to provide safe operative conditions and adequate perfusion to restore neurological functions and enhance early recovery [2].

The static parameters like central venous pressure (CVP) and pulmonary artery occlusion pressure (PAOP) were not precise and had certain limitations [3]. Hence, dynamic parameters based on pulse contour analysis are in use, such as pulse pressure variation (PPV), stroke volume variation (SVV), systolic pressure variation (SPV), and plethysmograph-based Pleth Variability Index (PVI) for assessing fluid responsiveness [4].

The fluid challenge can cause a marked or a minor change in cardiac output, depending upon the patient’s status on the slope of the Frank-Starling curve. So, if we fail to predict the response of cardiac output concerning volume expansion, fluid responsiveness can instead lead to haemodilution, resulting in increased cardiac filling pressures and thereby a state of fluid overload [2, 5, 6].

Primary brain tumour management is associated with the use of diuretics. Administration of optimum fluid therapy in such patients is essential, as hypovolemia and dehydration can lead to hypoperfusion of brain tissue, and excess transfusion can lead to increased intracranial pressure, for which fluid management in neurosurgery cases plays a pivotal role.

Shoemaker et al. [7] conducted the first major GDFT study in 1988 in patients undergoing high-risk surgeries. A combination of crystalloids and inotropes was administered to increase cardiac index and maximise oxygen delivery at the tissue level. The results showed a decrease in mortality and morbidity. The recent approaches for fluid resuscitation aim at conservative fluid management as per the body’s requirement. Fluid responsiveness has been defined as more than 15% increased cardiac output by thermodilution in response to a fluid challenge (500 ml crystalloid or colloid administered over 30 min) [8]. It is not easy to calibrate and standardise these heart–lung interactions during spontaneous ventilation. So dynamic parameters require mechanically ventilated patients with a tidal volume of no more than 8 ml/kg with the exclusion of ventricular dysfunction, arrhythmias, and tachypnoea >17/minute to give a fair picture of the filling status of the patient. They can be quantified objectively to derive useful parameters [9, 10, 11]. Dynamic markers assess the changes in stroke volume caused by ventilation changes and their effect on preload [12]. Pulse pressure variation (PPV) reflects the changes in arterial pulse pressures and thus depicts fluid responsiveness [10, 13, 14]. The respiratory cycle causes changes in the left ventricular stroke volume, which reflects pulse pressure because of the routine heart-lung interactions during a breath cycle. Assuming that arterial compliance remains constant over a respiratory cycle, pulse pressure variation would depict the change produced in stroke volume. In a meta-analysis including 22 studies and 807 patients, PPV predicted fluid responsiveness with an area under the receiver operating characteristic (AUROC) curve of 0.94 and a threshold of 12% [8]. In the studies included, fluid responsiveness was defined using one of the following techniques: transpulmonary thermodilution, pulse contour analysis, or oesophageal Doppler [15, 16]. Initially, PPV could be manually determined as the ratio of the difference between the maximal and minimal values of pulse pressure over the mean of the set of values and expressed as a percentage [15, 17, 16].

PVI, a measurement only available with the Masimo SET pulse oximetry (Radical 7, Masimo Corp, Irvine, CA, USA), is the first commercially available parameter to calculate respiratory variation using photoplethysmography. Data were collected noninvasively via pulse oximeter sensor [18, 19, 20]. The pulse oximeter uses red and infrared light to measure oxygen saturation [14]. A constant amount of light (DC) from the pulse oximeter is absorbed by the skin, other tissues, and nonpulsatile blood. In contrast, a variable amount (AC) is absorbed by pulsatile arterial blood flow [19]. AC and DC signals are extracted from the amplitude of the plethysmograph waveform. PVI reflects respiratory variations in the pulsatility index (PI) [20, 13, 21]. The dynamic change in PI during the respiratory cycle results in the PVI, which was automatically calculated using the formula below: PVI=100×(PImax−PImin)/Pimax

Chen et al. [22] studied the effects of stroke volume variation, pulse pressure variation, and Pleth Variability Index in predicting fluid responsiveness during different y pressure in prone position. Fischer et al. [23] performed a study on responsiveness to an intravenous fluid challenge in patients after cardiac surgery that compared arterial pulse pressure variation and the digital plethysmographic variability index.

But the efficacy of these dynamic parameters is yet to be clarified and established in the domain of neurosurgery cases, where no significant research has been done. Our study aims to compare the efficacy of PPV and PVI to guide GDFT in neurosurgery cases intraoperatively.

The study population consisted of patients who underwent elective supratentorial neurosurgeries, fulfilling the inclusion and exclusion criteria.

This study is based on the previous meta-analysis findings [24], where results showed that both PVI- and PPV-guided GDFT strategies were equivalent for the primary outcome of length stay in hospital (median [interquartile range] days) 2.5 [2.0–3.3] vs. 3.0 [2.0–5.0], P = 0.230, respectively. The sample size was calculated for 80% power (1–beta) with absolute precision of 5% (two-sided significance level or 1–α of 95%). The loss of follow-up was calculated at 10%, and a sample size of 80 was calculated, with 40 in group 1 (PPV) and 40 in group II (PVI). After identifying the prospective participants, they were allocated into groups I (PPV) and II (PVI). A statistician not involved in the study randomised the groups using the computer-generated random list. The allocation concealment was sequentially numbered, and the group was concealed in opaque envelopes. The envelope was opened when the patient reached the operation theatre, and the groups were allotted accordingly.

After giving written informed consent, the patients were all monitored under standard ASA monitors in the operation room. Before induction, ABG was done. Anaesthesia was induced with a sleep dose of thiopentone, fentanyl (2–3 μg/kg), and inhaled sevoflurane. To facilitate endotracheal intubation, vecuronium bromide (0.1 mg/kg) was administered. After that, intermittent positive pressure ventilation (IPPV) was instituted in a volume-controlled mode with a tidal volume of 6–8 ml/kg adjusted to obtain a PaCO2 of 30 35 mmHg. Intraoperatively, anaesthesia was maintained with 1–2 vol % sevoflurane and intermittent doses of fentanyl and vecuronium, as assessed by haemodynamic parameters and train of four to maintain a TOF of 1, respectively. After induction of anaesthesia, an ultrasound-guided central venous catheter was placed in the right internal jugular vein. An arterial catheter was placed in the left radial artery. Oxygen saturation was measured continuously using the Masimo Rainbow SET^R monitoring system (Radical 7, Masimo Corp, Irvine, CA, USA). A pulse oximeter to measure the PVI was placed on the patient’s index finger without an arterial cannula, covered with an impermeable black shield to prevent optical interference (only in the PVI group). A forced-air warming system (Bair Hugger Warming System, Augustine Medical, Eden Prairie, MN, USA) was applied to prevent hypothermia in the operating room and ICU. The hand with the pulse oximeter probe was kept warm. In addition, the waveforms of standard haemodynamic variables – heart rate (HR), arterial pressure, end-tidal carbon dioxide (EtCO2), and end-tidal gas concentration – were recorded using an operation theatre (OT) monitor (Beneview T8, Mindray, Hamburg, Germany).

The patients in both the groups (PPV and PVI) received a baseline fluid of 2 ml/kg/hr RL infusion. The fluid was administered as maintenance to maintain a stable perfusion pressure, and the additional boluses were issued whenever the dynamic parameters indicated. The fluid challenge consisted of a 250 ml bolus of crystalloid infused over 10 min. Patients in the PVI group received a fluid challenge if PVI was higher than 15% for more than 5 min, while patients in the PPV group received a fluid challenge if PPV was higher than 13% for more than 5 min [14, 15, 25]. Boluses were repeated until the threshold values reached both groups. Phenylephrine in aliquots of 25 mcg was titrated if mean arterial blood pressure remained below 65 mmHg despite preload optimisation. Additional volumes were required for administering antibiotics and analgesics, which were added to the total infused volume. Management of acute haemodynamic instability associated with haemorrhage was left to the attending anaesthesiologist’s discretion. If estimated blood loss exceeded 500 ml in groups, colloid or packed red blood cells (PRBC) were infused depending on the patient’s haemoglobin (Hb) to keep Hb around 9g/dl. PPV, PVI, BP, and HR recordings were noted every 10 minutes for 30 minutes of surgery and then every 30 minutes. A fall in BP >20% of baseline SBP and persisting despite average PPV or PVI was treated with vasopressor such as phenylephrine (25 mcg bolus). Patients were extubated when fully awake and meeting extubating criteria. ABG was taken at the end of the surgery, and vitals were noted. Intraoperatively, ABG was taken depending on the patient’s condition and requirement.

After screening through the eligibility criteria as per the inclusion and exclusion criteria, 72 patients needed to be enrolled in the study with a dropout of 10%, amounting to 8 patients, hence a total of 80 patients. Cases were registered from 15 December 2020 after CTRI approval. But only 74 patients were taken for analysis, as three from Group I were eliminated from the study. One patient was removed from the study due to excess blood loss and volatile PI values. Two patients from Group I were eliminated due to intraoperative inotropic support.

Similarly, two patients from Group II were eliminated due to technical error; there was a failure to read the PVI values in the Masimo monitor. One patient was eliminated from the study due to excessive blood loss causing haemodynamic instability and the use of inotropic support. As enumerated in the study protocol, after randomisation the patients in Group I received fluid therapy intraoperatively as guided by the PPV values. The patients in Group II received fluid therapy as per the PVI values. The groups had homogenous distributions by age, sex, and BMI with a P-value of 0.48, 0.20, and 0.29, respectively, belonging to ASA I, II, or III category with supratentorial tumours planned for decompressive craniotomy. ABG was taken before induction to record the baseline lactate levels as an indication of adequate tissue perfusion and to see whether the dynamic parameters were sensitive enough to predict a deficit in intravascular volume in the perioperative period. Every 10 minutes, vitals and urine output were noted hourly to check haemodynamic stability and perfusion. Comparison between both the groups was made based on the postoperative lactate values, length of stay in the critical care unit (CCU), and number of emesis and hypotension episodes (>20% fall of MAP from baseline). The observation depicted that both the parameters were comparable, with no significant difference in the postoperative lactate values with a P-value of 0.18 (Table 1). Similarly, the mean total fluid administered, mean blood loss, length of CCU stay, and emetic and hypotension episodes also showed no significant differences among the groups with P-values of 0.41, 0.78, 0.25, 0.30, and 0.67, respectively (Table 1). The mean urine output of both the groups was nearly equal, with a P-value of 0.79, thus indicating adequate perfusion.

Neurosurgery cases require appropriate fluid management intraoperatively. It is to be ensured that no apparent fluid overload leads to tissue oedema in the brain or hypovolemia, affecting tissue perfusion and leading to hypoxemia and increased serum lactate levels. Goal-directed fluid therapy prevents delayed recovery in the postoperative period, emesis, and hypotension episodes intraoperatively and in the recovery room.

Cardiac output as a tool guides fluid therapy during the peri-operative period. With advances in haemodynamic monitoring, static parameters have been replaced with dynamic parameters such as PPV, PVI, etc., taking heart-lung interactions into account [8, 15, 25].

Cannesson et al. [20, 15] tested the ability of PVI to predict fluid responsiveness in the operating room. Following volume expansion, there were increases in mean arterial pressure [21] A PVI value of more than 15% before volume expansion discriminated between responders and non-responders with 81% sensitivity and 100% specificity. This was the first study to directly assess the ability of PVI to predict fluid responsiveness in mechanically ventilated patients under GA [25].

The randomised clinical trials by Coeckelenbergh et al. [24] and Kim et al. [14] are the studies published where PVI and PPV have been compared intraoperatively to determine their efficacy for goal-directed fluid therapy and fluid responsiveness in low- to moderate-risk abdominal surgeries and spine surgeries in prone position respectively. Both PPV- and PVI-guided GDFT strategies were equivalent for the primary outcome of length of stay (2.5 vs. 3.5), respectively. Fluids infused, post-op complications, and all other results were comparable in both groups. Kim et al. [14] showed that 27 patients were fluid responders in the prone position. They are under the ROC curves for PPV and PVI and were 0.781 (95% CI: 0.646–0.883, P<0.001) and 0.756 (95% CI: 0.618–0.863, P<0.001).

Similarly, Sundaram et al. [13] concluded that the acid-base balance with the lactate levels in the postoperative period was equivalent between CVP (static parameter) and PPV (dynamic parameter) groups, thus reflecting equivalent perfusion in all the neurosurgical patients. Although the baseline lactate recorded was a bit higher than the average population, in both groups, it could be due to the interplay of other factors like pre-op dehydration and prolonged duration of surgery [8].

The findings of our study agree to some extent with the following studies. Wu et al. [16] compared two-stroke volume variation-based fluid protocols in neurosurgical procedures. They reported that the more liberal fluid protocol (targeting lower stroke volume variation) was associated with lower postoperative serum lactate variation. Sundaram et al. [13] compared PPV-directed therapy to CVP-guided therapy and found that the patients in the PPV-guided therapy group received more fluids and had a more stable haemodynamic profile than the CVP group. Hence, the lactate levels did not show a rising trend compared to preoperative levels, indicating better peripheral perfusion.

The area under the ROC curve depicts that 57% of area lies under the curve for PPV and 42% lies under the curve in PVI Group. This may lead to the conclusion that both the parameters, PPV and PVI are not reliable markers in terms of sensitivity and specificity(Fig 1 & Fig 2) in predicting fluid resuscitation for GDFT.

Figure 1.

Figure 2.

FLOW DIAGRAM OF THE STUDY

Furthermore, we said that the higher fluid conditions in the GDFT group were not associated with brain swelling from a surgeon’s perspective, even though we did not use any quantifying scale. Luo et al. [26] monitored stroke volume variation-guided fluid therapy in brain surgeries in a randomised controlled study. Unlike our study, Luo and colleagues had given less intraoperative fluids in their study group compared to the control group. Two differences between our study and Luo and colleagues’ study might account for dissimilar findings: Firstly, the fluid protocols in the control group of Luo et al. [26] studies were not well clarified. Secondly, their study group used a higher baseline fluid infusion rate (3 ml/kg/hr) than our baseline infusion rate of 2 ml/kg/hr.

Our study reflects the ROC curve for postoperative lactate values for both the groups with an area of 0.577 for PPV and 0.423 for PVI (Figure 1, Figure 2). This states that both the markers are ineffective in terms of sensitivity or specificity. Although PPV and PVI as parameters are comparable in intraoperative management and postoperative sequelae, they alone cannot determine goal-directed fluid therapy in neurosurgery cases due to the interplay of multiple factors. Our study echoes similar findings even though both are dynamic parameters with different interpretations of methodologies.

Our study chalks down the following conclusions based on the above results:

PPV and PVI, despite being novel dynamic parameters, depend on many other factors such as surgical blood loss, vasomotor tone, and preoperative hydration status; these must be determined in order to optimise fluid therapy in neurosurgery patients.