Journal of Vascular Surgery
Volume 33, Issue 4 , Pages 715-720, April 2001

Air plethysmography: The answer in detecting past deep venous thrombosis☆☆☆

Irvine Laboratory for Cardiovascular Investigation and Research, Division of Surgery, Anesthetics, and Intensive Care, Imperial College School of Medicine, St Mary's Hospital. London, United Kingdom

Received 10 March 2000; accepted 23 May 2000.

Article Outline

Abstract 

Purpose: In this study we assessed the accuracy of air plethysmography (APG) as a means of detecting earlier deep venous thrombosis (DVT), in comparison with venography, to develop a preoperative test for patients with varicose veins. Methods: In this retrospective analysis of prospectively acquired data, 202 patients referred with the clinical suspicion of chronic venous obstruction (224 lower limbs) and 41 patients (41 lower limbs) who had symptoms and signs suggestive of DVT, but had deep veins that appeared normal on venography, were studied with both venography and APG. Results: The results of venography were negative for past DVT in 169 legs and confirmed past DVT in 96 limbs. The DVTs were confined to the calf in 19 limbs and were found at popliteal level, more proximal, or both in 77 limbs. A total of 95% of the limbs that had earlier proximal DVT (73 of 77) were identified by means of an APG outflow fraction with occlusion of the superficial veins in the first second (OFs) of less than 28%. This is analogous to the Q wave of the electrocardiogram, which is a means of denoting the presence of myocardial infarction. The specificity rate of the method in the detection of past proximal DVT was 96%, the positive predictive value was 92%, and the negative predictive value was 98%. Conclusion: APG is a practical, inexpensive, easy-to-perform, accurate, noninvasive method for the diagnosis of hemodynamically significant (ie, proximal or extensive calf DVT) chronic venous obstruction that could replace venography. (J Vasc Surg 2001;33:715-20.)

 

It is an advantage to know the status of the deep veins in patients with varicose veins who may be candidates for venous stripping.1 The CEAP classification and grading of chronic venous disease in the lower limbs is gaining wide acceptance, because it allows an accurate description and can provide uniform reporting.2 However, to be fully implemented, it requires knowledge of the status of the deep veins.2, 3, 4

In the past, segmental plethysmography in the form of impedance plethysmography (IPG) and strain gauge plethysmography (SPG) have been used in the noninvasive detection of acute proximal deep venous thrombosis (DVT).5, 6, 7, 8 However, 60% of IPG test results that were positive for DVT were found to normalize at 12 weeks,9 and 95% were found to normalize within 1 year.10 Thus, these techniques were not suitable means of identifying limbs with chronic venous obstruction.

Air plethysmography (APG) is a practical, noninvasive, and relatively inexpensive technique of measuring volume changes in the lower leg (excluding the foot) accurately.11, 12, 13, 14 Most authors have used the outflow fraction in the first second without (OF) or with occlusion of the superficial veins at the knee (OFs)11, 12, 13, 14 to detect outflow obstruction. Their results suggest that OF and particularly OFs are reduced in the presence of old DVT. However, none of these studies was performed with venography in patients with varicose veins. Establishing the normal values of leg volume changes and the venous volume (VV) was essential as a baseline to make a comparative evaluation of outflow measurements obtained in patients with past DVT.15, 16

The aim of our study was to determine the value of APG in the detection of residual chronic outflow obstruction in patients with symptoms of chronic venous insufficiency.

Back to Article Outline

Patients and methods 

In this retrospective analysis of prospectively acquired data, patients who were referred to the vascular laboratory to exclude chronic venous obstruction, either because of a probable history of DVT or extra-anatomical varicose veins suggestive of being collaterals, provided 224 legs. By extra-anatomical varicose veins, we mean veins that run anterolaterally or posteromedially on the thigh and not along the course of the long or short saphenous vein on the calf. All patients were examined with venography as the gold standard investigation at the request of their surgeons (either at the acute DVT or at the course of investigation of their chronic venous insufficiency), and all patients gave their informed consent to be studied by means of APG in addition to venography. Each of these examinations was performed and reported without the knowledge of the result of the other examination. In addition, another group of 41 patients (41 lower limbs) without varicose veins and without any evidence of acute or old DVT on venography was examined with APG. In this group, venography was performed because clinical symptoms and signs suggested acute DVT, but the results of venography were negative for DVT.15, 16 This group was called normal/nonacute DVT. There were 243 patients, 61 men and 182 women, with a mean age of 54 years (range, 19-90 years). Twenty-two patients were studied bilaterally.

Venography 

The method of venography performed was the St Mary's modification of Lea Thomas, which has been described.17 The venographic criteria used for the diagnosis of past DVT were those established by Lea Thomas and McAllister.18 Depending on the time of the onset of the DVT, the venographic appearance was categorized: (1) a thrombus that has been present for more than 7 days becomes adherent and obliterates the white line of contrast material between the thrombus and the vein wall and the process may extend to completely occlude the vein; (2) an occluded vein may either remain occluded and bypassed by collaterals or the thrombus may retract toward the side wall of the vein, giving a more clearly defined outline and resulting in a narrowed lumen; (3) the process of adherence and retraction may lead to recanalization, with the vein appearing irregular and narrowed with the absence of normal valves.

Air plethysmography 

The air plethysmograph APG (APG-1000, ACI Medical, San Marcos, Calif) consisted of an air chamber (5-L capacity) that was applied around the calf and was connected to a pressure transducer, an amplifier, and a pen recorder (more recently updated to a computerized system). A 100-mL syringe was incorporated in the circuit and was used for calibration to milliliters of limb volume change.11, 14, 15 The patient was in the supine position, with the leg first elevated (45°) to empty the veins, and then with the heel resting on a 30-cm support. The air chamber was inflated to a pressure of 6 mm Hg, which is the lowest pressure required to keep it on the leg without compressing the veins. After calibration, a 14-cm thigh tourniquet, placed close to the groin, was inflated to 80 mm Hg. The VV increased until it reached a plateau, then the thigh tourniquet was deflated rapidly. The observed rapid decrease in volume is a result of venous outflow.

The ratio of the amount of blood that leaves the leg in the first second (V1) over the total venous volume (VV) multiplied by 100 provides the outflow fraction (OF), which is expressed as a percentage.11, 19 The procedure was repeated with the long saphenous and/or other prominent superficial veins occluded by digital external compression at the level of the knee to obtain the outflow fraction with superficial occlusion (OFs). This was done before the rapid deflation of the thigh tourniquet. For the superficial occlusion, we did not use a tourniquet, because this might also compress the popliteal vein, as suggested by Neglen.20 In a patient with large thighs, one might need to use both hands to occlude all the superficial veins.

Another technical point that merits attention and is anecdotal is that at the time of occluding the superficial vein, the VV increases by approximately 10 mL. This represents blood displaced distally by digital compression in the superficial veins. Therefore when calculating the OFs, one should divide the V1 by the initial venous volume and not by the artificially increased venous volume. In our early experience with APG, we calculated the OFs by dividing the V1 by the artificially increased (because of the digital external compression) VV, and we found low OFs values that suggested past DVT. We then asked for a venogram to be performed on these patients, and the results were negative for past DVT. This puzzled us until we observed that by re-calculating the OFs and dividing the V1 by the initial VV (and not by the artificially increased VV), we got consistent results for APG and venography.

Statistical considerations 

The SPSS statistical package (Chicago, Ill) was used as a means of analyzing the data. Not all the variables were normally distributed; therefore we used nonparametric tests for comparison among the groups.21, 22, 23 Because both venography and APG were performed on the same subjects as a means of detecting the patients with past proximal DVT, we used the McNemar χ2 test for paired samples to compare these two methods.24

Back to Article Outline

Results 

Of the patients who were examined for chronic venous insufficiency (CVI) (224 lower limbs), past DVT was confirmed in 96 patients by means of venography. This was confined to the calf in 19 limbs, and it was popliteal, more proximal, or both in 77 limbs. Healthy deep veins were demonstrated by means of venography in 128 limbs. These limbs, with the 41 limbs that had venography to exclude acute DVT but had normal venogram results, provided us with a total of 169 limbs.

Three variables were measured for each patient: (1) the outflow fraction values without (OF) and with superficial occlusion (OFs) in the first second; (2) the venous volumes (VV) for each group (non-DVT, past calf DVT, and past proximal DVT); and (3) the descriptive statistics for the three variables (minimum, median, maximum, mean, and SD) for the two non-DVT groups (n = 169), the patients with past calf DVT (n = 19), and the patients with past proximal DVT (n = 77).

Fig 1 shows a box-plot of the values of the VV in milliliters (median and interquartile range) and the mean and SD in subjects without DVT and patients with past calf DVT and past proximal DVT.

  • View full-size image.
  • Fig. 1. 

    Values of venous volume by group. N, Number of limbs; CVI/non DVT, patients with chronic venous insufficiency who had negative results on venograms; normal/non acute DVT, patients referred to the laboratory to exclude acute DVT; CVI/non DVT and normal/non acute DVT, non DVT group; x, mean; sd, standard deviation.

Fig 2 shows a box-plot of the outflow fraction values (median and interquartile range) without (OF) and with superficial occlusion (OFs) in the first second, with mean and SD, in subjects without DVT and patients with past calf DVT and past proximal DVT.

  • View full-size image.
  • Fig. 2. 

    Outflow fraction without (OF) and with superficial occlusion (OFs) at the first second. N, Number of limbs; CVI/non DVT, patients with chronic venous insufficiency who had negative results on venograms; normal/non acute DVT, patients referred to the laboratory to exclude acute DVT; CVI/non DVT and normal/non acute DVT, non-DVT group; x, mean; sd, standard deviation.

Because there was no significant difference (P < .001) between the CVI/non-DVT and the normal/nonacute DVT groups in all three variables (VV, OF, OFs), we grouped them together, and named the group the non-DVT group. When the three groups of patients (non-DVT, past calf DVT, past proximal DVT) were compared with the Kruskall-Wallis test, the results were significant for all three variables (P < .001; Fig. 1, Fig. 2). When we compared the VV of the non-DVT subjects with that of patients with calf DVT and patients with proximal DVT by using the Mann-Whitney U test, the results were still significant (P < .001). However, there was no significant difference in the VV between the patients with calf DVT and patients with proximal DVT. The median values of the outflow fraction were found to be significantly different in all three groups, when compared with each other by means of the Mann-Whitney U test, both without (OF) and with superficial occlusion (OFs; P < .001).

There was no difference shown by means of the McNemar χ2 test comparing venography with APG between the two methods (χ2 = 0.1; P > .6) for the detection of past proximal DVT (Table).

Air plethysmography versus venography
VenographyAPG OF/SO1 < 28%
DVTNo DVTTotal
DVT73 (95%)4*77
No DVT6†163 (96%)169
Total79167246
*Four patients had OF/SO1 ≥ 28% (28%, 30%, 31%, 38%). Their respective venous volumes were 68, 81, 320, and 300 mL. No patient with chronic venous obstruction had a venous volume less than 50 mL. †Six non-DVT patients had an OF/SO1 between 25% and 27%.

McNemar χ2 test = 0.1; P > .6.

Sensitivity rate = 95%.

Specificity rate = 96%.

Positive predictive value = 92% (73/79).

Negative predictive value = 98% (163/167).

Overall accuracy rate = 95%.

DVT, Deep venous thrombosis; APG, air plethysmography; OF/SO1, outflow fraction with occlusion of the superficial veins in the first second.

Overall, four limbs with past proximal DVT were missed by means of APG, all of which were partially occluding the superficial femoral vein lumen, and only seven of the 19 calf thrombi were detected. The sensitivity rate of the method in the detection of past proximal DVT was 95%, the specificity rate was 96%, the positive predictive value was 92%, and the negative predictive value was 98%.

Back to Article Outline

Discussion 

Duplex scanning is a highly reliable method of detecting proximal thrombosis in patients in whom it is clinically suspected.17, 25, 26, 27, 28 However, the detection of the late sequelae of thrombosis may be more difficult,29, 30, 31 especially because 30% to 50% of duplex scans will normalize within 1 year.32, 33, 34

The natural history of DVT is unknown. Whether the veins will recanalize partially or completely and whether the valves will be competent or have reflux is unpredictable.18 The outflow fraction reflects the functional capacity of the venous system to drain the blood collected in the lower limb during the period of venous occlusion. The outflow measurements are analogous to the Q wave of the electrocardiogram, which denotes the presence of a myocardial infarction. Once an episode of an hemodynamically significant (ie, not isolated calf) DVT has taken place, the OF and OFs values became lower than normal and remained that way. The VV is a measure of the venous capacity of the limb, and its value reflects the space available for blood to accumulate. Normally, this lies between 70 to 200 mL. At the time of an acute proximal DVT, the VV is less than 50 mL.15 After the acute phase of a proximal obstruction, as collateral circulation develops, the VV gradually increases and can reach values much higher than 200 mL. It is those collaterals that otherwise confound outflow tests. The confounding effect of collaterals is substantially removed with digital compression right before the outflow test.

Because the VV, OF, and OFs values of the 169 patients who had non-DVT venogram results were similar, they were grouped together for subsequent analysis. They indicate that the range of these values is the same, irrespective of presentation, when the venogram results are normal. Ideally, we should have studied patients with normal venogram results who were also asymptomatic. However, it would be unethical to perform venography on a healthy person. Because the outflow fraction values of healthy subjects (n = 41) and those of patients with chronic venous insufficiency but no evidence of past DVT on venography (n = 128) are not significantly different, we felt justified in using the aforementioned non-DVT patients as our control group.

There was no difference in the VV in the patients with past calf DVT and patients with past proximal DVT (Fig 1). All three values (OF, OFs, VV) of patients with calf DVT had an intermediate position between those of non-DVT subjects and those of patients with proximal DVT. Also, the sample size for the patients with calf DVT was small. Therefore, the reliability of APG as a means of diagnosing past calf DVT needs further investigation.

Volume changes measured with segmental plethysmographic techniques, such as SPG, do not necessarily represent changes in the whole leg.35 The use of APG with an air chamber that includes the lower leg (excluding foot) has the advantage of providing information that is more representative and reproducible.11

APG has only recently been applied for the diagnosis of acute DVT.15, 36 Our results show that APG has a place in chronic cases, also. Christopoulos et al37 demonstrated that in limbs with normal deep veins, the outflow fraction OF is higher than 38%. Hemodynamically significant (proximal or extensive calf) acute DVT15 or limbs with a history of DVT confirmed by means of venography have an OF ranging from 15% to 34%, which is reduced further after digital occlusion of the superficial veins (OFs). Our results confirm this. The possible explanation for this may be provided by the finding of irregularity of the venous wall shown by means of venography in such patients. The wall may become rigid, and at high pressure it may fail to stretch as much as the normal wall. This failure to dilate increases the outflow resistance of the veins. The effect of occluding the superficial veins is a means of detecting the contribution of these veins as collaterals. This maneuver not only assists in the diagnosis of obstruction in acute or past DVT, but also helps the surgeon decide whether to ligate these veins.

The primary difference in the APG test is that it allows for digital compression of the superficial veins, as compared with IPG, in which the clinician's fingers on the leg will alter the impedance measurement. To our knowledge, digital compression has not been used with either IPG or SPG. Because better sensitivity and specificity rates are obtained by using finger compression of superficial veins with APG, the APG method is superior. There is clearly a difference between the measurement systems in that SPG uses a segmental sensor that detects circumferential changes at only 1 limb location, whereas the APG sensing cuff measures volume changes in milliliters of the lower leg (excluding the foot).

The key improvement to the outflow concept is the additional step of digitally compressing the superficial veins in the APG method, which improves sensitivity and specificity rates. Additionally, modifying the measurement of VV by not including the artifactual volume (because of digital occlusion) increases the accuracy of the technique.

Outflow phenomena are typically characterized as an exponential washout equation. The manual calculation of time constants associated with an exponential is cumbersome, so 1 second outflow percentages have been used historically.11, 19 The SPG sensor measures percent change in the limb segment circumference, which is used as a means of calculating volume based on an assumed model. The IPG measures changes in the electrical impedance of the limb, which is assumed to represent changes in blood volume, also based on a model. The APG more directly measures limb volume changes because it is calibrated with the introduction of 100 mL of air with a large syringe. This likely accounts for its increased measurement accuracy.

The outflow measurements correlate well with invasive quantification of the severity of venous obstruction, obtained by measuring the arm/foot pressure differential with cannulation of a vein in the arm and foot in a supine position.11, 38

In conclusion, APG (OFs < 28%) is a reliable, portable, inexpensive, easy-to-perform, and accurate noninvasive method for the diagnosis of hemodynamically significant (ie, proximal or extensive calf DVT) chronic venous obstruction, even when the deep veins have recanalized. If the results are substantiated by further prospective studies, the outflow fraction of APG may become a routine test for preoperative screening of patients who have varicose veins and a history or clinical signs suggestive of past DVT.

Back to Article Outline

References 

  1. Cockett FB. Management of the postphlebitic limb. Br J Hosp Med. 1971;6:767–770
  2. Porter JM, Moneta GL. International consensus committee on chronic venous disease. Reporting standards in venous disease: an update. J Vasc Surg. 1995;21:635–645
  3. Nicolaides AN. The investigation of chronic venous insufficiency: a consensus statement. Circulation. 2000;102:E126–E163
  4. van Bemmelen PS, Mattos MA, Hodgson KJ, Barkmeier LD, Ramsey DE, Faught WE, et al.  Does air-plethysmography correlate with duplex scanning in patients with chronic venous insufficiency?. J Vasc Surg. 1993;18:796–807
  5. Hallbook T, Gothlin J. Strain gauge plethysmography and phlebography in diagnosis of deep venous thrombosis. Acta Chir Scand. 1971;137:37–52
  6. Sakaguchi S, Ishitobi K, Kameda T. Functional segmental plethysmography with mercury strain-gauge. Angiology. 1972;23:127–135
  7. Hull R, Van Aken WG, Hirsh J, Gallus AS, Hoicka G, Turpie AGG, et al.  Impedance plethysmography using the occlusive cuff technique in the diagnosis of venous thrombosis. Circulation. 1976;53:696–700
  8. Anderson DR, Lensing AWA, Wells PS, Levine MN, Weitz JI, Hirsh J. Limitations in the diagnosis of clinically suspected deep-vein thrombosis. Ann Intern Med. 1993;118:25–30
  9. Jay R, Hull R, Carter C, Ockelford P, Buller H, Turpie AG, et al.  Outcome of abnormal impedance plethysmography results in patients with proximal-vein thrombosis: frequency of return to normal. Thromb Res. 1984;36:259–263
  10. Huisman MV, Buller HR, Ten Cate JW. Utility of impedance plethysmography in the diagnosis of recurrent deep-vein thrombosis. Arch Intern Med. 1988;148:681–683
  11. Christopoulos D, Nicolaides AN, Belcaro G. Air plethysmography. Phlebology Digest. 1992;4:4–11
  12. Gillespie DL, Cordts PR, Hartono C, Woodson J, Obi-Tabot E, LaMorte WW, et al.  The role of air plethysmography in monitoring results of venous surgery. J Vasc Surg. 1992;16:674–678
  13. Cordts PR, Hartono C, LaMorte WW, Menzonian JO. Physiologic similarities between extremities with varicose veins and with chronic venous insufficiency utilizing air plethysmography. Am J Surg. 1992;164:260–264
  14. Comerota AJ, Harada RN, Eze AR, Katz ML. Air plethysmography: a clinical review. Int Angiol. 1995;14:45–52
  15. Kalodiki E, Nicolaides AN. Air-plethysmography for the detection of acute DVT; new criteria. Vasc Surg. 1997;31:123–129
  16. Kalodiki E. Diagnosis and prophylaxis of deep venous thrombosis [PhD thesis]. London: Univ of London; 1996;
  17. Kalodiki E, Marston R, Volteas N, Leon M, Labropoulos N, Fisher CM, et al.  The combination of liquid crystal thermography and duplex scanning in the diagnosis of deep vein thrombosis. Eur J Vasc Surg. 1992;6:311–316
  18. Lea Thomas M, McAllister V. The radiological progression of deep venous thrombus. Radiology. 1971;99:37–40
  19. Shi Z, Boccalon H, Elias A, Garcia-Serrano A. Detection of deep vein thrombosis with a computerized strain gauge plethysmograph. Int Angiol. 1992;11:160–164
  20. Neglen P, Raju S. Detection of outflow obstruction in chronic venous insufficiency. J Vasc Surg. 1993;17:583–589
  21. Strike PW. Probability distributions II; continuous variables. In:  Strike PW editors. Statistical methods in laboratory medicine. Oxford: Butterworth-Heinemann; 1991;p. 96–128
  22. Siegel S, Castellan JN. Two independent samples. In:  Siegel S,  Castellan NJ editor. Nonparametric statistics for the behavioral sciences. Boston: McGraw-Hill; 1988;p. 102–167
  23. Siegel S, Castellan JN. The case of k independent samples. In:  Siegel S,  Castellan NJ editor. Nonparametric statistics for the behavioral sciences. Boston: McGraw-Hill; 1988;p. 190–223
  24. Strike PW. Inference II; analysis of 2 × 2 tables. In:  Strike PW editors. Statistical methods in laboratory medicine. Oxford: Butterworth-Heinemann; 1991;p. 477–522
  25. Elias E, Le Corff G, Bouvier JL, Benichou M, Serrandimigni A. Value of real-time B-mode ultrasound imaging in the diagnosis of deep vein thrombosis of the lower limbs. Int Angiol. 1987;6:175–179
  26. Cranley JJ, Higgins RF, Berry RE, Ford CR, Comerota AJ, Griffen LH. Near parity in the final diagnosis of deep venous thrombosis by duplex scan and phlebography. Phlebology. 1989;4:71–74
  27. Comerota AJ, Katz ML, Greenwald LL, Leefmans E, Czeredarczuk M, White JV. Venous duplex imaging: should it replace hemodynamic tests for deep venous thrombosis?. J Vasc Surg. 1990;11:53–60
  28. Rose SC, Schiller VL. Ultrasound best modality for imaging lower DVT. Diagn Imaging (San Franc). 1992;14:78–85
  29. Wysokinski W, Beynes G, Blockmans D, Verhaeghe R. Assessment of the patency of deep leg veins with duplex. Int Angiol. 1991;10:69–71
  30. Aitken AG, Godden DJ. Real time ultrasound diagnosis of deep vein thrombosis, a comparison with venography. Clin Radiol. 1987;38:309–313
  31. Foley WD, Middleton WD, Lawson TL, Erickson S, Quiroz FA, Macrander S. Color Doppler ultrasound imaging of lower-extremity venous disease. Am J Radiol. 1989;152:371–376
  32. Cronan JJ, Leen V. Recurrent deep venous thrombosis: limitations of US. Radiology. 1989;170:739–742
  33. Prandoni P, Cogo A, Bernardi E, Villalta S, Polistena P, Simioni P, et al.  A simple ultrasound approach for detection of recurrent proximal-vein thrombosis. Circulation. 1993;88:1730–1735
  34. O'Shaughnessy AM, Fitzgerald DE. Natural history of proximal deep vein thrombosis assessed by duplex scanning. Int Angiol. 1997;16:45–49
  35. Allan JC. Volume changes in the lower limb in response to postural alterations and muscular exercise. S Afr J Surg. 1964;2:75–89
  36. Hosoi Y, Yasuhara H, Miyata T, Komiyama T, Onozuka A, Shigematsu H. Comparison of near-infrared spectroscopy with air plethysmography in detection of deep vein thrombosis. Int Angiol. 1999;18:287–293
  37. Christopoulos D, Nicolaides AN, Duffy P, Georgiou I. Non-invasive diagnosis and quantification of outflow obstruction in venous disease. J Cardiovasc Surg. 1989;30:72–73
  38. Raju S. New approaches to the diagnosis and treatment of venous obstruction. J Vasc Surg. 1986;4:42–54

☆ Competition of interest: nil.

☆☆ Reprint requests: Dr Evi Kalodiki MD, BA, DIC, PhD, Senior Vascular Research Fellow, Irvine Laboratory, Academic Surgical Unit, St Mary's Hospital, Praed St, London W2 1NY, UK (e-mail: e.kalodiki@ic.ac.uk ).

PII: S0741-5214(01)54536-X

doi:10.1067/mva.2001.111743

Journal of Vascular Surgery
Volume 33, Issue 4 , Pages 715-720, April 2001

Article Tools

* Image Usage & Resolution