Valvular reflux after deep vein thrombosis: Incidence and time of occurrence

Transcription

1 Valvular reflux after deep vein thrombosis: Incidence and time of occurrence Arie Markel, MD, Richard A. Manzo, BS, CCVT, Robert O. Bergelin, MS, and D. Eugene Strandness, Jr., MD, Seattle, Wash. From December 1986 to December 1990, 268 patients with acute deep vein thrombosis were studied in our laboratory. From this group 107 patients (123 legs with deep vein thrombosis) were placed in our long-term follow-up program. The documentation of va~.xatlar reflux and its site was demonstrated by duplex scanning. The duplex studies were done at intervals of i and 7 days, i month, every 3 months for the first year, and then yearly thereafter. The mean follow-up time for these patients was 341 days. In addition, reflux was evaluated in 502 patients with negative duplex study results and no previous history of deep vein thrombosis or chronic venous insufficiency. In the patients with acute deep veiaa thrombosis, valvular incompetence was noted in 17 limbs (14%) at the time of the initial study. Reflux was absent in 106 limbs (86%). In this last group reflux developed in 17% of the limbs by day 7. By the end of the first month, 37% demonstrated reflux. By the end of the first year, more than two thirds of the involved limbs had developed valvular incompetence. The distribution of reflux at the end of the first year of follow-up was the following: (1) popliteal vein, 58%; (2) superficial femoral vein, 37%; (3) greater saphenous veiaa, 25%; and (4) posterior tibia/vein, 18%. Reflux seems to be more frequent in the segments previously affected with deep vein thrombosis. Among cases where segments were initially affected with thrombi, after 1 year the incidence of reflux was 53%, 44%, 59%, and 33% for the common femoral vein, superficial femoral, popliteal vein, and posterior tibia/ vein, respectively. In contrast, among limbs without initial deep vein thrombosis in these segments none of the proximal limbs and 8% of the posterior tibia/ veins showed reflux after i year. In the patients without deep vein thrombosis, reflux in either the deep or superficial system was present in 60 of 1003 legs (6%). Greater saphenous vein reflux was present in 16 legs (2%); it was limited to this vein in five limbs. (J VASC SURG 1992;15: ) Of the complications caused by deep venous thrombosis (DVT) the postthrombotic syndrome (PTS) is the most common late cause for morbidity and disability. Approximately 24 million Americans have varicose veins, whereas six to seven million have stasis dermatitis, and 400,000 to 500,000 have or have had a postthrombotic ulcer. ~ In most cases PTS is caused by the presence of incompetent valves in the deep and superficial veins. 2 Although chronic venous obstruction may occur, it has been implicated only for the development of venous claudication, a Although approximately two thirds of patients From the Department of Surgery, University of Washington. Presented at the Fotty-fifth Annual Meeting of the Society for Vascular Surgery, Boston, Mass., June 4-5, Supported by National Institutes of Health grant no. HL Reprint requests: D. Eugene Strandness, lr., MD, Head Vascular Surgery Section, Department of Surgery, RF-25, University of Washington, Seattle, WA /6/32970 with proximal venous thrombosis develop symptomatic chronic venous disease, 4-6 the precise interval of time after which this occurs is not well documented. Also it is not clear what the relation is between the area initially involved by thrombi and the subsequent development of valvular incompetence. The important questions that we have attempted to answer are as follows: (1) What is the relative incidence of venous reflux in the different venous segments after an episode of DVT? (2) Are thc initially thrombosed venous segments more prone to develop val~lar incompetence than those veins not involved? (3) Can patent venous channels distal (toward the foot) to proximally occluded veins become incompetent? (4) What is the relation between the presence of partial venous thrombosis and the occurrencc of valve damage? (5) Do limbs without previous DVT develop reflux? During the last 4 years, we have regularly monitored those patients with an initial episode of DVT. The follow-up included clinical interviews as 377

2 378 Markel etal. lournal of VASCULAR SURGERY well as duplex examination at the time of each visit. An analysis of these results is presented here. METHODS Starting in December 1986 all patients referred to our cfinical laboratory with a specific question of DVT in the lower extremities were subjected to a duplex ultrasound study. Patients were not admitted to the follow-up protocol if one or more of the following conditions were present: 1. the patient declined to participate; 2. the patient was extremely ill; 3. marked physical disability made visits impossible; 4. distant home location; 5. a previous episode of DVT in the affected leg; 6. symptoms suggested chronic venous insufficiency in the affected leg, and/or duplex examination suggested an old DVT. After informed consent was obtained, eligible patients with DVT were enrolled in our long-term follow-up study. Patients were monitored at intervals of 1 day, 7 days, 1 month, 3 months, and then every 3 months for the first year, and every year thereafter. Data from patients not enrolled in the follow-up protocol and with a negative duplex study outcome were analyzed to establish the prevalence of valvular incompetence in limbs free of DVT. Recruitment into the study has been continuous, and so not all subjects have been under observation the same length of time. Also, the early follow-up visits have more missed appointments than later visits. In particular, the 1-day, 1-week, and 1-month intervals pose scheduling problems either because they occur during hospitalization where compliance with the research protocol must take second place to the treatment requirements and abilities of the patient, or because the subject has left the hospital but is not yet able to return for a follow-up appointment. An assumption made in analyzing these data is that those subjects attending their scheduled follow-up visits are representative of all study subjects elegible to be seen at those time points. A duplex examination was done, and a questionnaire regarding the presence of risk factors and symptoms was filled out at the initial and each follow-up visit. Venous duplex examination was performed with the Ultramark 8 or 9 model duplex scanner (Advanced Technology Laboratories, Bothell, Wash.), with the patient in the reverse Trendelenburg position at 10 to 15 degrees. All the major deep veins of the lower extremities were examined, including the common lilac vein, external iliac vein, common femoral vein, superficial femoral vein, deep femoral vein, popliteal vein, and both posterior tibial veins. The inferior vena cava and the greater saphenous vein were examined in most of the patients and are included in parts of the analysis. Images of the lilac veins were obtained with a 5 MHz transducer. A 5 or 7.5 MHz probe was used to obtain images of the common femoral vein, superficial femoral vein, and popliteal vein. The superficial femoral vein was examined at three locations in the thigh: proximal, midthigh, and distal, at the entrance to Hunter's canal. A 10 MHz transducer was used to visualize the veins of the calf. Pulsed Doppler signals were obtained with a 5 MHz transducer. The duplex scan was considered to indicate total obstruction in a particular segment if there was no spontaneous or augmentable flow and the vein was not compressible. Partial obstruction was diagnosed when there was residual flow in the involved venous segment. The diagnosis of reflux was based on the detection of reverse flow induced by a Valsalva maneuver and/or limb compression. The results of the questionnaire and the results of the duplex examination were entered in the computer with use of the SPSS/PC 7 data entry program. The data were analyzed with the SPSS data analysis program. Statistical analysis When reflux rates were calculated the following rules were applied: incidence rates were computed with use of the number of cases with a definitive duplex study for the specific segment at baseline and at the follow-up time point under consideration. If the segment examination was not definitive at the initial visit then that segment was dropped from all follow-up incidence calculations. If a particular follow-up duplex study was inconclusive, then the segment was excluded from computation for that time point only. Selection of segments with and without thrombosis also required a definitive duplex study at the initial visit with regard to DVT statds. Confidence intervals around the incidence estimates are "exact" values tabled in Geigy Scientific Tables. 8 Comparison of reflux rates in thrombosed versus patent veins was by cross-tabular analysis and Fisher's Exact Test. RESULTS One thousand eleven patients were referred to the vascular laboratory to rule out DVT between December 1986 and December The findings were

3 Volume 15 Number 2 February 1992 Venous reflux after deep vein thrombosis 379 Table I. Incidence of reflux in any deep vein during follow-up Follow-up interval Legs (n) Reflux (%) 95%-CI CI, Confidence interval. 1 day 40 2 (5) 1-17 I wk 54 9 (17) mo (37) mo (52) mo (53) mo (56) yr (69) yr (63) yr 10 5 (50) positive for DVT in 268 patients (27%). In 702 patients (69%) the results were negative, and in 41 (4%) the findings were inconclusive. Of the patients with positive results, 161 cases were not enrolled in the follow-up study for the reasons previously mentioned, that is: in 58 cases (36%) the patient declined to participate or lived too far from the laboratory; in 52 cases (32%) the patient was ineligible because of a previous episode of DVT in the affected leg; in 51 cases (32%) the patient was extremely ill or had a physical disability making follow-up visits impossible. One hundred seven patients (123 legs) were included in this analysis. The mean age was 52 _ 19 years (range:, 10 to 83 years). There were 58 men and 49 women. Thrombi were present in one leg in 91 patients and in both legs in 16 patients. The length of the follow-up ranged from 1 day to 3 years (mean, 341 days). During the 4-year interval reported here, 48 of the 107 subjects had been withdrawn from followup. Reasons for these losses are the following: death (20 subjects), too ill to continue (12 subjects), moved away (7 subjects), and withdrew consent or lost contact (9 subjects). Of the 20 subjects who died, only two completed 1 year of follow-up; 14 completed 1 month of follow-up or less. Patients with initial DVT and follow-up One htmdred six of the 123 legs with initial DVT entered into the follow-up program did not show reflux in any' deep venous segment at the initial visit. The incidence of reflux at the different follow-up intervals in this group is shown in Table I. The incidence ofreflt~ was 17% after the first week, 37% after the first month, 69% after the first year, and 63% among limbs monitored 2 years. When analyzed by specific venous segments, the findings were as noted in Fig. 1 and Table II. In comparison with the more proximal veins (common femoral vein, superficial femoral vein, and popliteal vein), reflux was less frequent in the posterior tibial veins, being 18% (4 of 22) at 1 year and 11% (2 of 18) at the end of the second year. At 1 year reflux was more frequent in segments previously affected with DVT, than in those without initial thrombosis (Table III). The incidence of reflux was significantly higher in common femoral veins thrombosed at the baseline evaluation compared with common femoral veins patent at the same examination (p = 0.01). A similar although not statistically significant difference was observed in the other veins. At 2 years of follow-up the common femoral vein and posterior tibial vein showed a similar pattern, but the superficial femoral vein showed no difference, and the popliteal vein showed a slightly lower rate. Totally occluded veins more commonly developed reflux than did veins with only partial obstruction. When initial total obstruction was present in the common femoral vein, 9 of 14 limbs (64%) developed reflux at the 1-year follow-up, In contrast 0 of 3 limbs (0%) with baseline partial obstruction in the common femoral vein developed reflux in that segment during the same interval of time. The 1-year incidence of venous valve incompetence for the superficial femoral vein was 6 of 13 (46%) and 1 of 3 (33%), respectively, for those cases with total or partial obstruction at presentation. For the popliteal vein it was 10 of 14 (71%) and 0 of 3 (0%), respectively, during the same follow-up period. All the common femoral veins, popliteal veins, and posterior tibial veins observed at 1 year had recanalized, including all those with initial total obstruction. This was also true of the superficial femoral vein, except for two cases with initial total obstruction, which remained occluded at 1 year. Valve incompetence was detected in the greater saphenous vein in 5 of 20 limbs (25%) after 1 year. Concomitant reflux in the common femoral vein

4 380 2~larkel et al Journal of VASCULAR SURGERY Common femoral vein Superficial femoral vein d Popliteal vein Posterior-tibial vein ~D ld lw lm 3m 6m 9m ly 2y 3y ld lw lm 3m 6m 9m ly 2y 3y Fig. 1. Incidence of reflux in deep veins at the different follow-up intervals; 1 day, 1 week, 1 month, 3 months, 6 months, 9 months, 1 year, 2 years, 3 years. The standard errors of the estimate are shown for each time interval. Table II. Incidence (%) of reflux during follow-up among various venous segments Follow-up interval CFV SFV Pop PTV GSV 1 day i week mo I mo mo mo i 17 I yr yr yr CFV, Common femoral vein; SFV, superficial femoral vein; Pop, popliteal vein; PTV, posterior tibial vein; GSV, greater saphenous vein. through the popliteal vein was present in four of these five cases. Only in one Of the five legs with greater saphenous vein incompetence was the greater saphenous vein thrombosed at the initial study. In regard to symptoms, none of the patients in this study developed skin ulceration during followup. It may be that the duration of fouow-up was too short for this to occur. Patients without DVT (and no follow-up) Seven hundred two patients had a negative duplex study outcome for acute DVT. Two hundred patients with a previous history of DVT and/or present physical findings or history suggesting chronic venous insufficiency were excluded. Patients with a history of varicose veins were excluded as well. In the 502 remaining patients (1003 limbs), 60 limbs (6%) had reflux. The sites of valve incompetence in the 60 legs were as follows: (1) common femoral vein, 30 of 60 (50%); (2) superficial femoral vein, 29 of 60 (48%); (3) popliteal vein, 25 of 60 (42%); (4) posterior tibial vein, 20 of 60 (33%); (5) greater saphenous vein, 16 of 60 (27%). In 5 of the 16 legs (31%) with greater saphenous vein incompetence the reflux was limited to the greater saphenous vein, that is, isolated superficial venous reflux. DISCUSSION The importance of valvular incompetence as a cause of the PTS is well known. Chronic venous

5 Volume 15 Number 2 February 1992 Venous reflux after deep vein thrombosis 381 Table III. Incidence of reflux during follow-up among vein segments initially thrombosed or initially patent I yr 2yr Initially thrombosed ~ Initially patent Initially thrombosed ~ Initially patent Vein segment n (%) n (%) n (%) n (%) CFV 9/17 (53) 0/8 (0) 7/14 (50) 3/8 (37) SFV 7/16 (44) 0/3 (0) 3/15 (20) 1/5 (20) PV 10/17 (59) 0/1 (0) 6/17 (35) 1/2 (50) PTV 3/9 (33) 1/13 (8) 2/7 (29) 0/11 (0) ~Includes partial and total obstruction. obstruction appears to be the cause of venous claudication but is not an essential feature for the development of pain, edema, and stasis pigmentation and ulceration. 2 With the introduction of duplex scanning it became possible to study each vein separately to document patency and valvular incompetence. This method has been demonstrated to be accurate and reliable for the diagnosis of venous thrombosis.i The method is also suitable for documenting the presence of valvulaj: reflux of the deep and superficial system.ll, 12 In the current study we have excluded a significant number of patients from follow-up because of severe illness or disability. Longer term surveillance is not possible in patients with these conditions. Although our follow-up patient group was relatively healthier than those cases that were excluded, they were by no means free of serious disease. Even with our exdusion criteria a significant proportion of patients were enrolled in the study that were either chronically ill or had disabling advanced disease. Cancer, for example, was present in 34 of 107 (32%) patients in the follow-up group and in a similar proportion of those cases excluded from the study: 51 of 161 (32%). Also, a relatively high number of follow-up patients died or withdrew because of severe illness. We believe that our follow-up group is generally representative of patients that are discharged from the hospital after their DVT event. Although earlier studies have shown that after a period of months or years a high proportion of patients with DVT will develop reflux, 4,11 the exact time course is not clear. The present study shows that reflux can develop very soon after an episode of DVT. By the end of the first week 17% of the limbs have developed valve incompetence at some level. After the first month 40% of the legs have reflux. This increase continues progressively during the next months, and after the first year approximately two thirds of the patients have valvular incompetence. This pattern of progression is also present when each one of the major venous segments is considered separately. It is clear that the proximal veins have a higher incidence of reflux than the posterior tibial veins. This is probably due to the lower incidence of thrombosis in these distal leg veins. Past studies have questioned the relationship between DVT and subsequent venous insufficiency. 13 The present study shows that valve incompetence develops most frequently in those segments that have been initially affected by thrombi. The development of reflux in a specific segment will be clue in most cases to the damage caused to the valve by the thrombotic process. Furthermore, when a venous segment was totally occluded initially, the subsequent incidence of reflux in this segment was much higher than in those veins presenting initially with partial obstruction. Although the numbers are limited, our data show that reflux in the distal deep veins seems to be infrequent when only the proximal veins were thrombosed at the baseline visit. In those cases where the posterior tibial veins were not thrombosed initially but DVT was present more proximally, reflux occurred in only 8% of the cases examined at 1 year. To evaluate reflux, patency must be present. With the exception of two cases of superficial femoral vein, all the veins that were initially completely occluded had recanalized after 12 months. This means that a high proportion of veins that recanalized did not show reflux, that is, approximately 40% to 50% of the segments examined at 1 year in our study. Whether reflux occurs in the "normal" leg without previous DVT has also been the subject of debate. In a study by Raju 132 of 14 (15%) of healthy volunteers had reflux. Lindner et al.s found the contralateral leg to have the same or even higher

6 382 Markd et al. Journal of VASCULAR SURGERY incidence of reflux than the leg with the initial DVT. These findings raise the question as to the mechanisms behind the development of vamilar incompetence. 6 Our study shows that the prevalence of reflux was only 6% in the patients without DVT. This is much lower than described by others, even though this population is not a normal control group since they were suspected of having acute DVT. Why some patients without a history of DVT develop chronic venous insufficiency is not clear. It is possible, as has been proposed by Browse et al.,6 that these patients may have had unrecognized subclinical episodes of DVT. It has been postulated that reflux in the greater saphenous vein could occur in isolation. In the study by Raju, 13 all the cases with greater saphenous vein incompetence also had reflux in deeper veins. The author concluded that greater saphenous vein reflux does not occur alone. In the group of patients without DVT in the present study, 31% of the cases with greater saphenous vein reflux had isolated involvementl This demonstrates that greater saphenous vein reflux can occur alone and is not necessarily related to the presence of involvement of the deep venous system. It is also clear that greater saphenous vein reflux is not always a result of thrombosis of this venous segment. When initial and 1-year follow-up examinations were compared, thrombosis of the greater saphenous vein only appeared to be a factor in 20% of instances when reflux developed. A more plausible explanation for the occurrence of reflux in the greater saphenous vein may be the incompetence that develops in the common femoral vein when it has been occluded and subsequently lysed. In fact, after 1-year follow-up, four of five cases with greater saphenous vein valve incompetence showed concomitant reflux involving the common femoral vein. Severe symptoms were rare in our study. No leg developed skin ulceration. This may be due to the relatively short duration of the follow-up, although studies with follow-up periods longer than ours also found a very low incidence of skin ulceration. 4,s This differs notably from older studies or retrospective studies of patients with chronic venous insufficiency, 1.,ls which showed a high proportion of ulcers developing at an early time. Recent studies have shown that reflux in the distal veins (popllteal and posterior tibial) correlates well with the severity of symptoms, especially ulceration. 16'17 Our study shows that distal vein reflux is more frequent when the initial DVT involved these veins, but can also occur when more proximal veins are affected. Continued follow-up of the same patients will chart the development of more severe complications, including hyperpigmentation and ulceration. EspeciaUy important is the follow-up of those patients with popliteal and posterior tibial vein reflux, who seem to be predisposed for the more severe manifestations of the PTS. REFERENCES 1. Coon WW, Willis PV, Keller JB. Venous thromboembolism and other venous diseases in the Tecumseh community health study. Circulation 1973;48: Killewich LA, Martin R, Cramer M, Beach KW, Strandness DE Jr. An objective assessment of the physiologic changes in the postthrombotic syndrome. Arch Surg 1985;120: Strandness DE Jr. Deep venous thrombosis and the postthrombotic syndrome. In: Strandness DE Jr, ed. Duplex scanning in vascular disorders. 1st ed. New York: Raven Press, 1990: Strandness DE Jr, Langlois Y, Cramer M, Randiett A, ThMe BL. Long-term sequelae of acute venous thrombosis. JAMA 1983;250: Lindner DJ, Edwards JM, Phinney ES, Taylor LM, Porter JM. Long-term hemodynamic and cliulcal sequelae of lower extremity deep vein thrombosis. J VAsc SUV, G 1986;4: Browse NL, Clemenson G, Lea Thomas M. Is the postphlebitic leg always postphlebitic? Relation between phlebographic appearances of deep-vein thrombosis and late se quelae. BMJ 1980;281: SPSS/PC+ version 2.0. SPSS Inc. Chicago, Geigy Scientific Tables. vol 2, 8th ed. Ciba-Geigy, Basel, Siegel S. Nonparametric statistics for the behavioral sciences. New York: McGraw-Hill, 1956: Killewich LA, Bedford GR, Beach KW, Strandness DE Jr. Diagnosis of deep venous thrombosis. A prospective study comparing duplex scanning to contrast venography. Circulation 1989;79: Killewich LA, Bedford GR, Beach KW, Strandness DE Jr. Spontaneous lysis of deep venous thrombi: rate and outcome. I VAsc SURG 1989;9: Szendro G, Nicolaides AN, Zukowski AJ, et al. Duplex scanning in the assessment of deep venous incompetence. J VASe SURe 1986;4: Raju S. Venous Insufficiency of the lower limb and stasis ulceration. Ann Surg 1983;I97: Bauer G. Phlebographic diagnosis of spontaneous thrombosis-the history of phlebography. Acta Chir Scand 1942(suppl);74:l-i06. O'Donneil TF, Browse NL, Buruand KG, Lea Thomas M. The socioeconomic effects of an iliofemoral venous thrombosis. J Surg Res 1977;22: Shull KC, Nicolaides AN, Femandes d Fernandes J, et al. Significance of popliteal reflux in relation to ambulatory venous pressure and ulceration. Arch Surg I979;114: Gooley N, Sumner DS. Relationship of venous reflux to the site of venous valvular incompetence: implications for venous reconstructive surgery. J VASC SURG 1988;7:50-7. Submitted June 10, 1991; accepted Aug. 8, 1991.

7 Volume 15 Number 2 February 1992 Venous reflux after dee? vein thrombosis 383 DISCUSSION ProE Browse (London, England). Once again the group from Seattle have shown us the power of duplex ultrasonography in revealing changes within blood vessels, and once again they have revealed the complexities of the venous system and how difficult a system it is to understand. I must confess that after reading this manuscript I was confused, not by the facts presented, which you have already heard, but by the absence of some facts I think are important. For example, they begin with 123 legs containing thrombns, but we are not told, even when the data are subdivided into specific venous segments, how many remained occluded and how many recanalizcd. Wc are told how many of the rccanalizcd segments showed reflux, but not told how many showed only orthogradc flow. Can we assume that every recanalized vein was incompetent? If not, were there competent valves in previously thromboscd segments, o1: was competence maintained by valves above and below the rccanalized segment? Conversely, when a segment was incompetent, were the healthy vessels above and below it competent or incompetent? What was the effect of the thrombosis on the adjacent segments? Was the incompetence sometimes a secondary event? These concerns lead me to ask about the significance of an incompetent segment. The authors began by analyzing incompetent legs by which they mean a leg containing an incompetent vessel, but does an incompetent vessel in a whole leg have any physiologic significance? It would be interesting if the authors could relate the ch~inges they have observed to a test of overall calf pump function. After all we have known that thrombosis causes incompetence, since the papers of Edwards and Edwards in the 1930's, and although this paper tells us how quickly it happens, what we really need to know is whether incompetence in one segment matters at all and if it does, how it affects function. I must also comment on the numbers in this study. It begins with a follow-up of 40% of all the cases of thrombosis first seen, but thereafter the numbers reviewed steadily fall so that by 1, 2, and 3 years, only 26%, 19%, and finally 8% of the initial group of 40% were being studied. Do these small numbers give a true representation of the whole group? Last, and perhaps the most important point of all, can the ultrasonographer be absolutely certain that the vein seen to be patent 3 months after a thrombosis is the same vein that was seen to be thrombosed at the initial examination? Collateral veins can be indistinguishable from normal veins, and many veins, especially the popliteal and superficial femoral vein, are often double, and only one of them may have been involved in the thrombosis. I would like to hear what criteria the investigators used to ensure that they kaew they were looking at the same vein at each of the follow-up examinations. Dr. Arie Markel. To answer the first question about how many of the segments remained occluded and how many of them were patent, we have the data for each one of the follow-up intervals, but regarding the 1-year follow-up, which relates to the most important part of the data I presented here, I can say that for these major venous segments that I presented including the common femoral, superficial femoral, popliteal and posterior tibial vein, at 1 year follow-up all the segments with the exception of two were patent. It means they have recanalized after 1-year follow-up. Only in two cases of superficial femoral veintwo of nine that did not have reflux at 1 year and were initially thrombosed- only two of nine were still occluded. If we exclude these two cases from analysis, the incidence of reflux will then increase even higher in the superficial femoral vein from 44% to 50%. So, we can say that basically the data we presented at 1-year follow-up show most of the segments, more than 90%, have recanalized. In regard to the question of what is the importance of showing reflux in the whole leg when it may be that only an isolated segment has reflux, it is difficult to answer if this has relevance from the point of view of later complication. We can only speculate and say that the presence of isolated common femoral or superficial femoral vein reflux could in the long term be the cause for distal reflux. We know that distal reflux has a good correlation with severe symptoms. You asked about the changing of numbers, the numbers in the follow-up were lower than the numbers at the beginning. Many reasons exist for this. The first reason is that we only considered the legs or the segments that did not have reflux at the initial study. The second reason was that despite the fact that we used severe disease to exclude patients, in the first year of the study many patients that were too ill, were included for the study, and during the follow-up time approximately 20% of the patients died. Some of them had disease progress during the follow-up interval, and this also reduced the number. With this still in mind, I think that the numbers represented here can be compared favorably with previous studies. We studied a total of 60 legs at 3-month follow-up, approximately 50 at 6 months, and approximately 32 at 1-year follow-up. To answer the last question, the term "absolute" is a very difficult one. I would say that in most cases we can be confident that the vein we are examining is the common femoral or popliteal vein and not a collateral or one of the bifid veins. Dr. George Johnson (Chapel Hill, N.C.). At the end of 2 years, two thirds of the involved limbs demonstrate reflux by duplex scanning, mostly in the major deep venous system. The manuscript does not state, at least to my reading, if they considered these all to be initial thrombosis or were some recurring? This would certainly make a difference in the time study that was done. Two years after complete thrombosis of a major deep

8 384 Markel et al. Journal of VASCULAR SURGERY vein, 35% had reflux, apparently at the site of recanalization. Does this support that it is better to remain occluded than it is to recanalize? I suppose the data you presented support your thesis, "the mechanism of reflux is recanalizatlon rather than dilation of collateral veins." Thus, except in a few instances, I assume proximal obstruction was not a cause of more distal reflux. As George Bernard Shaw once said of science, this paper "opens more questions than it answers." Dr. Markel. If I understand correctly, the specific question was again related to the incidence of reflux in those segments without DVT and again the point of the presence of occlusion, which I already answered. In those legs that did not have DVT, the proportion of reflux during follow-up was much lower than in those segments that had DVT initially. Still there an important number of cases exist where a segment that did not have DVT at the initial study still developed reflux during follow-up. We do not have to forget that is the segment that did not have DVT at the beginning, but the other segment in the same leg had DVT. All these legs considered have DVT in some segments. Dr. Robert Zwolak (Hanover, N.H.). The group from Seattle has recently described the variability in the definition of reflux in venous segments and the method used to define reflux. Given the variability, I would like to know whether the examiners performing these studies were blinded to the presence or absence of reflux in the previous studies or were blinded to either the patients with DVT or control patients when they did the tests. Dr. Markel. The material for this study is taken from the patient group that usually comes to the laboratory with a question of DVT, so the findings during this first visit are the findings that we presented in this study. We tried to eliminate those cases where the venous insufficiency was already known. This was one of the reasons for exclusion of the patients. If DVT was present in the past, these patients were excluded. We tried to eliminate those patients already having problems and to include only those with any history previously. Dr. Hank Arellano (Denver, Colo.). I have a question regarding the low incidence of vahaxlar incompetence in previously occluded fibial veins. You mentioned using the Valsalva maneuver and I understand that for the proximal veins, but what was your method of checking for valvular incompetence in the tibial veins? Dr. Markel. Usually in our laboratory, we use both the Valsalva maneuver and the manual compression maneuver either by distal compression and release or by proximal compression for the diagnosis of reflux in each one of the segments. The incidence of reflux in the posterior tibial vein in those cases with thrombosis was lower than in other segments, but the incidence was still much higher than in those segments without DVT. If I remember well, at 1-year follow-up it was 33% in those with initial DVT and only 8% in those without DVT initially.