Digital venous photoplethysmography in the seated position is a reproducible noninvasive measure of lower limb venous function in patients with isolated superficial venous reflux
Article Outline
Background
The value of photoplethysmography (PPG) has been questioned because of a lack of reproducibility. We performed this study to determine whether new digital technology has improved the reproducibility of PPG in the noninvasive assessment of lower limb venous function in patients with isolated superficial venous reflux.
Methods
This was a prospective study of 140 legs in 110 patients (65% female; median age [interquartile range], 45 years [36-59.25 years]; CEAP clinical grade C2/3, n = 114; C4-6, n = 26) who underwent repeated digital PPG measurements of refilling time (RT) in both the sitting and standing position after standard exercise regimens by a single observer. RT was measured in all patients 2 to 5 minutes apart and in a randomly selected subgroup of 30 patients (38 limbs) 1 to 2 weeks apart. RT variability was assessed by using Bland and Altman’s coefficient of repeatability (CR-RT), where the CR-RT was 1.96 times the standard deviation of the mean difference in RT between two tests. Venous duplex scanning of both the deep and superficial veins was also performed, and a reverse flow of greater than 0.5 seconds was considered abnormal. Only patients with isolated superficial venous reflux were included in the study.
Results
The CR-RT of the tests on 140 limbs performed 2 to 5 minutes apart was 10 seconds overall, 3 seconds for RT up to 10 seconds, and 16 seconds for RT between 20 and 40 seconds. The CR-RT of the 38 tests performed 1 to 2 weeks apart was also 10 seconds. No systematic variation due to a nonrandom error was found between the measurements performed either 2 to 5 minutes or 1 to 2 weeks apart.
Conclusions
Digital PPG performed in the seated position in patients with isolated superficial venous reflux provides a reproducible method for the noninvasive assessment of lower limb venous function for both clinical and research purposes. However, the variation in precision of RT with the magnitude of the measurement must be taken into account when results are interpreted in individual patients.
Photoplethysmography (PPG) is a noninvasive test that uses a light-emitting diode and a photoelectric cell to detect changes in skin blood volume. PPG is also known as light reflection rheography and works on the principle that as the volume of blood in the skin changes, so too does the amount of light reflected back to the sensor. During exercise, the amount of blood in the skin of the lower limb decreases as a result of venous emptying secondary to the action of the calf muscle pump. By placement of the PPG probe over the skin of the lower limb, the blood volume of the skin can be semiquantified and relates primarily to the effectiveness of the pump mechanism in clearing the leg of venous blood. The time taken for the skin to refill with blood after exercise is known as the refilling time (RT). In health, RT is related to arterial inflow and, in the presence of venous disease, also to the presence and severity of deep and superficial lower limb venous reflux. Although RT has been shown in certain studies to correlate well with other measures of venous function,1, 2, 3 other investigators have found the value of PPG to be limited by their inability to calibrate the signal4 and obtain reproducible results.2 In addition, its use in differentiating limbs with deep venous disease from those with superficial venous reflux has been limited by the failure of tourniquets to eliminate the contribution of superficial venous reflux to the abnormal test results.
In contrast, quantitative digital PPG is observer independent; it uses computer technology and mathematical models to automatically calibrate the signal and determine the RT. Theoretically, this should eliminate observer error and bias and improve both the validity and reproducibility of the test.5 The aim of this study, therefore, was to assess the intraobserver repeatability of RT determined by quantitative digital PPG in patients with isolated superficial venous reflux awaiting surgery.
Methods
One hundred forty legs in 110 unselected patients (65% female; median age [interquartile range; IQR], 45 years [36-59.25 years]) awaiting superficial venous surgery for clinical grade (CEAP classification)6 C2/3 (n = 114) or C4 to C6 (n = 26) disease were examined. Ethics committee approval and written informed consent were obtained.
Quantitative digital PPG (The Dopplex; Huntleigh Healthcare) was used to measure postexercise RT in the sitting and standing positions. RTs were repeated 2 to 5 minutes (RT1 and RT2) apart in all patients and 1 to 2 weeks apart (RT1 and RT3) in a randomly selected subgroup (38 legs in 30 patients; 58% female; median age [IQR], 52 years [40-59 years]; CEAP C2/3, n = 32; C4-6, n = 6).
By using a transparent, double-sided sticky pad, the PPG probe was attached to the skin 10 cm above the medial malleolus and 1 to 2 cm posterior to the subcutaneous border of the tibia, avoiding any prominent varicosities or areas of pigmentation, lipodermatosclerosis, or atrophe blanche. With the patient as motionless as possible, the integrated computer automatically calibrated the signal to adjust for the individual’s skin properties.5 Once a stable baseline was achieved, the exercise was commenced with an built-in metronome. In the sitting tests, patients completed 10 dorsiflexions and plantarflexions over a 15-second period and were then asked to rest as motionless as possible. In the standing tests, patients performed 10 tiptoe exercises while holding onto a fixed support and were then asked to rest non–weight bearing as motionless as possible. The ejection of blood from the skin and the subsequent refilling curve were displayed visually (Fig 1). The computer calculated the RT (seconds) from the curve up to a maximum of 45 seconds. A lower RT indicates worse venous physiology, and an RT of greater than 25 seconds is considered to indicate a normal test result.
Fig 1.
Refilling curve produced by digital photoplethysmography machine.
Before the test, the patients rested for at least 10 minutes, and between each two tests they were rested for 2 to 5 minutes to allow complete venous refilling to take place. Each repeat test was performed at the same time of day and in the same room, although the room temperature was not measured.
All patients underwent lower limb venous duplex scanning of the deep (common femoral, femoral, and popliteal veins) and superficial (saphenofemoral junction [SFJ], great saphenous vein [GSV], saphenopopliteal junction [SPJ], and lesser saphenous vein [LSV]) veins to determine the anatomic sites of venous reflux. Patients with complete deep venous reflux (common femoral, femoral, and popliteal veins) were excluded from the study. In addition, the presence of palpable pedal pulses was used to exclude significant peripheral arterial occlusive disease in all limbs.
The surgical strategy was determined by the operating consultant on the basis of the duplex findings. The planned operations consisted of primary GSV surgery (flush ligation of the SFJ, stripping of the GSV to the knee, and multiple stab avulsions), primary LSV surgery (flush ligation of the SPJ, excision of the proximal 5-10 cm of the LSV, and multiple stab avulsions), a combination of primary GSV and LSV surgery, or redo junctional ligation and trunk vein stripping. Perforator surgery and sclerotherapy were excluded.
Statistical analysis was performed by using SPSS 11.0 (SPSS Inc, Chicago, Ill). Paired RT measurements from the same patient were compared by using a Bland-Altman plot, where the mean of RT1 and RT2 (or RT3) was plotted against the difference between RT1 and RT2 (or RT3). If the two test results happened to be identical, the mean difference would be 0. Random (nonsystematic) variation between the two series of test results would be shown as a random scatter of data points about the horizontal axis: in other words, around the mean of 0. Nonrandom (systematic) bias between two series of test results would be shown by a scatter of points clustered above or below horizontal axis: in other words, around a mean difference greater than or less than 0. The Bland-Altman coefficient of repeatability (CR) is the standard deviation of the difference between measurements RT1 and RT2 (or RT3) multiplied by 1.96, which is the value below which the difference between the two measurements will lie with a probability of 0.95.7 Nonparametric tests were used to compare RT between patient groups.
Results
Seated position
In total, there were 140 legs belonging to 110 patients (65% female; median age [IQR], 45 years [36-59.25 years]) with CEAP clinical grade C2/3 (n = 114) or C4 to C6 (n = 26) disease. The median (IQR) RT for all limbs was 13.5 seconds (8.5-22.0 seconds), with a significantly lower RT in limbs with clinically worse disease (C2/3, 15 seconds; C4-6, 7.5 seconds; P = .000; Kruskal-Wallis test).
Figure 2 is a scatter plot of RT1 against RT2. In 140 legs, there was little systematic variation in RT1 and RT2 obtained 2 to 5 minutes apart, with a mean difference of 0.89 seconds and a CR-RT of 10.8 seconds between the two series of measurements (Fig 3). However, CR-RT increased with RT (Table I); it was 3.5 seconds when RT1 was less than 10 seconds and was 15.9 seconds when RT1 was between 20 and 40 seconds. Looking at the limbs according to the clinical status, the CR-RT in the C2/3 group was 11.5 seconds, compared with a CR-RT of 4.0 seconds for the CEAP 4 to 6 group. The CR-RT of RT1 and RT3 obtained 1 to 2 weeks apart in 38 legs (30 patients; 58% female; median age [IQR], 52 years [40-59 years]; CEAP C2/3, n = 32; C4-6, n = 6) was 12.9 seconds (Fig 4). In this case, CR-RT did not increase with RT (Table II).
Fig 2.
Scatter plot of refilling time (RT) 1 against RT2 in the seated position (n = 140).
Fig 3.
Bland-Altman plot of the difference in refilling time (RT) taken 2 to 5 minutes apart (RT1 − RT2) against mean RT in the seated position (n = 140). The mean difference of RT1 − RT2 was 0.89 seconds (SD, 5.4 seconds); the RT coefficient of repeatability was 10.6 seconds.
Table I. Coefficient of repeatability for first measurement (RT1) less than 10 seconds, 10 to 20 seconds, and 20 to 40 seconds
| RT1 (s) | n | Mean difference in RT (SD) | Coefficient of repeatability |
|---|---|---|---|
| <10 | 55 | −0.35 (1.79) | 3.50 |
| 10-20 | 44 | 1.0 (3.91) | 7.66 |
| 20-40 | 20 | 0.3 (8.11) | 15.9 |
Fig 4.
Difference between refilling time (RT) 1 to 2 weeks apart (RT1 − RT3) against mean RT for the seated position (n = 38). The mean difference for RT1 − RT3 was −0.76 seconds (SD, 6.59 seconds); the RT coefficient of repeatability was 12.9 seconds.
Table II. Coefficient of repeatability for first measurement (RT1) less than 10 seconds, 10 to 20 seconds, and 20 to 40 seconds in the seated position 1 to 2 weeks apart (n = 38)
| RT 1 (s) | n | Mean difference in RT (SD) | Coefficient of repeatability |
|---|---|---|---|
| <10 | 12 | −2.40 (6.08) | 11.9 |
| 10-20 | 13 | −1.85 (3.55) | 6.96 |
| 20-40 | 20 | 2.00 (4.69) | 9.19 |
Standing position
The standing results from 72 limbs had to be excluded because of an inability to either stabilize the baseline before the test or to generate sufficient venous outflow for the limb during exercise at 1 or more time points. In a further 31 limbs, the test was incomplete. For the remaining 37 limbs, Fig 5 is a plot of RT1 against RT2. There was little systematic variation between RT1 and RT2, with a CR of 7.2 seconds (Fig 6). In 27 of these limbs, the test was also repeated 1 to 2 weeks later. This showed greater variability, with a CR-RT of 12.8 seconds (Fig 7). Fig 3, Fig 4, Fig 6, Fig 7 show a cluster of points about 45 seconds due to the fact that this is the upper limit of measurement of the machine.
Fig 5.
Refilling time (RT) 1 against RT2, 2 to 5 minutes apart, for the standing position (n = 37).
Fig 6.
Difference between refilling time (RT) 1 and RT2 against the mean of RT1 and RT2, 2 to 5 minutes apart for the standing position (n = 37). The mean difference for RT1 − RT2 was 0.46 seconds (SD, 3.69 seconds); the RT coefficient of repeatability was 7.23 seconds.
Fig 7.
Difference between refilling time (RT) 1 and RT3 against the mean of RT1 and RT2, 1 to 2 weeks apart, for the standing position (n = 27). The mean difference for RT1 − RT3 was −1.3 seconds (SD, 6.53 seconds); the RT coefficient of repeatability was 12.8 seconds.
Relationship between seated venous RT and anatomic distribution of venous reflux based on venous duplex scan
In the seated position, the median (IQR) RT in limbs with SFJ reflux alone with tributary varicosities (n = 14) was 17.5 seconds (14.0-34.0 seconds), compared with 12.75 seconds (9-24 seconds) in limbs with SFJ and segmental GSV reflux (either thigh or calf GSV; n = 62) and 11.0 seconds (8.0-22.0 seconds) in limbs with complete GSV reflux from groin to ankle (n = 43). In the limbs with SPJ and complete LSV reflux (n = 12), RT was 11.25 seconds (9.25-16.88 seconds). Although there seemed to be a trend toward worsening (lower) RT with more superficial venous segments of reflux, this was not statistically significant (P = .148; Kruskal-Wallis test). In the remaining limbs (n = 9), the anatomic arrangement of superficial venous reflux was a combination of GSV and LSV segments and SFJ and SPJ reflux.
Discussion
These data show that reproducible measures of RT can be obtained in the sitting position in nearly all patients with isolated superficial venous reflux by using quantitative digital PPG. However, perhaps not surprisingly, there is a marked increase in variability with increasing RT, resulting in a lack of precision in patients with higher (longer) RTs with less severe venous disease. Whether this is important depends on circumstances and will clearly inform the choice of test in both the clinical and research settings.
Photoplethysmography is one of several investigations that have been developed to assess the hemodynamic function of the lower limb venous system. When combined with assessments of venous anatomy as derived by venous duplex scan, these objective measures help to identify the severity of underlying venous disease, in addition to history and clinical examination. However, the very fact that numerous methods exist highlights that each has its own limitations.
Direct ambulatory venous pressure measurement is widely believed to be the best overall assessment of lower limb venous function and, as such, is frequently the gold standard against which other tests are validated.4 However, its usefulness is severely limited by its invasive nature, thus making it a poor option for repeated measures in outcome studies.
More recently, air plethysmography has been shown to give accurate and reproducible information about overall lower limb venous function.8 It is noninvasive, but the equipment needed is bulky and the test is time consuming, thus leaving it less attractive for clinical studies outside the vascular laboratory. In addition, there are differing opinions as to the variability of air plethysmography parameters on the basis of several reproducibility studies.8, 9, 10, 11
Postexercise RTs obtained by PPG have previously been shown to correlate well with those obtained during simultaneous direct pressure measurements.1, 3 The introduction of new digital machines that are compact and easy to use has also removed the problem of calibration, which was the major limitation to its general use for quantification of venous function.4 Because it is noninvasive and well tolerated by patients, it is an attractive test for outcome assessment after treatment of chronic venous disease, such as in clinical trials. It has, however, been criticized for poor discrimination between limbs with differing severity of venous disease.3 In particular, when tourniquets are used to occlude superficial veins to determine the contribution of superficial venous reflux in an abnormal test, this frequently fails to be repeatable. Use of the test in the standing position has previously been shown to improve discrimination.3 Although we found this standing method to be repeatable, there were many limbs in which no reading was possible because of the inability to achieve a stable baseline before exercise or a minimum amplitude during the exercise itself, despite the use of a fixed support for the patients to hold onto. This severely limits the use of the standing position, and we would recommend the test to be performed in the seated position.
Photoplethysmography RT has also been shown to correlate poorly with the degree of deep venous disease as measured by absolute venous pressure measurement.3 This is explained by the fact that PPG RT reflects regional rather than global lower limb venous function, and thus it is related more closely to superficial rather than deep venous incompetence. It is therefore more suited to assessment of treatments that target superficial venous reflux. Indeed, our study was limited to patients with isolated superficial venous disease, because these individuals make up the majority of the venous operative workload in our unit. This test is therefore of use in both a research and clinical setting—whenever repeated assessment of venous function in patients with isolated superficial venous disease is required. Careful review of the English-language literature demonstrates that to date, however, few outcome studies have been performed using PPG as a quantitative assessment, so it is unclear what degree of improvement in RT would be expected after such an intervention. The results of ongoing research in this area are awaited.
In conclusion, therefore, digital PPG in the seated position for the functional assessment of the lower limbs of patients with isolated superficial venous disease is repeatable. Long-term follow-up data are awaited to see what changes in RT are seen after interventions to treat superficial venous reflux and, thus, determine the role of digital PPG as a method of venous outcome assessment.
Author contributions
The authors thank Tim Marshall, Medical Statistician, University of Birmingham, for statistical advice and analysis.
References
- The use of photoplethysmography in the assessment of venous insufficiency (a comparison to venous pressure measurements) . Surgery . 1979;86:434–441
- . Venous photoplethysmography (relationship between transducer position and regional distribution of venous insufficiency) . J Vasc Surg . 1990;11:436–440
- . Photoplethysmography in the assessment of venous insufficiency . J Vasc Surg . 1987;5:405–412
- . Investigation of chronic venous insufficiency (a consensus statement) . Circulation . 2000;102:e126–e163
- . Value of quantitative photoplethysmography for functional vascular diagnostics . Skin Pharmacol Appl Skin Physiol . 2001;14:316–323
- . Classification and grading chronic venous disease in the lower limbs. A consensus statement . Eur J Vasc Endovasc Surg . 1996;12:487–492
- . An introduction to medical statistics . 3rd ed.. Oxford: Oxford University Press; 2003;
- . Air plethysmography and the effect of elastic compression on venous hemodynamics of the leg . J Vasc Surg . 1987;5:148–159
- . Variability and reliability of air plethysmographic measurements for the evaluation of chronic venous disease . J Vasc Surg . 1997;28:638–642
- . Variability of venous haemodynamics with daily activity . J Vasc Surg . 1994;19:361–365
- . Venous assessment using air plethysmography (a comparison with clinical examination, ambulatory venous pressure measurement and duplex scanning) . Br J Surg . 1993;80:967–970
Competition of interest: none.
PII: S0741-5214(05)01854-9
doi:10.1016/j.jvs.2005.10.039
© 2006 The Society for Vascular Surgery. Published by Elsevier Inc. All rights reserved.
