A multivariate analysis of determinants of wound healing in patients after amputation for peripheral vascular disease

EurJ VascSurg 4, 291-295 (1990)

A Multivariate Analysis of Determinants of Wound Healing in Patients After Amputation for Peripheral Vascular Disease T h e o d o r e A. A. van den Broek, B o u d e w i j n J. D w a r s , Jan A. R a u w e r d a and Fred C. Bakker Vascular Laboratory, Department of Vascular Surgery, Free University Hospital, Amsterdam, The Netherlands In a prospective study 53 patients undergoing amputations of the lower limb were evaluated for clinical criteria, laboratory results, pulse volume recordings, Doppler pressures, Photoplethysmographic Skin Perfusion Pressures (PPG/SPP) and angiography. The purpose of the study was to analyse which of these techniques predicts wound healing adequately after amputation. Forty-five patients eventually completed all tests. With the exception of PPG/SPP none of these tests were able to predict skin healing. The technique of PPG/SPP proved very reliable in helping to select the level of amputation, if measured anteriorly (P = 0.0001, r = 0.83). Angiographic scoring also correlated significantly (P = 0.0016) with a successful result. This study suggests that surgeons should not rely on their clinical acumen for the selection of the amputation level. In the absence of a reliable non-invasive test, angiography may well be useful but PPG/SPP will enable the surgeon to amputate on an optimal level and thus reduce complications and improve rehabilitation. Key Words: Amputation selection; Skin perfusion pressure; Photoplethysmography; Angiography.

Introduction In peripheral vascular disease w h e n an amputation of the lower limb is considered, primary healing will not occur in up to 20% of cases, and re-amputation will be needed in over 8%, eventually leading to a mortality of 7 to 15%. 1,2 Although more distal amputations can breakdown, surgeons try to maintain the longest stump possible, because rehabilitation is greatly influenced by the level of amputation in older people. An amputation below the knee (BK) will provide successful prosthetic rehabilitation in 73% of patients over 65 years, but an amputation above the knee (AK) in only 50%. 3 In addition, morbidity is increased because the assessment of amputation level is often wrong leading to prolonged immobilisation and clinical criteria are usually used to decide on the level of amputation. Non-invasive techniques such as pulse volume recording (PVR) 4 and Doppler pressure s have been used in the past with varied success. Recently a non-invasive technique called photoplethysmographic Please address all correspondence to: Th. A. A. van den Broek, Boerhaave General Hospital, Tenierstraat 1, 1071 DX Amsterdam, The Netherlands. 0950-821X/90/030291+05 $03.00/0

© 1990Grune&StrattonLtd

assessment of skin perfusion pressure (PPG/SPP) has been described 6 but has not yet been compared with clinical or other non-invasive methods. The purpose of this prospective study, started in 198 6, was to compare clinical criteria, PVR, Doppler and PPG/SPP for their effectiveness in predicting amputation w o u n d healing. In addition, angiographic determination of inflow to the level of amputation was scored as in m a n y patients the possibilities of vascular reconstruction had been considered.

M a t e r i a l s and M e t h o d s Fifty-three amputations were entered into the study, but in eight cases the patients were not amputated within 3 weeks of evaluation, or did not complete the clinical and non-invasive studies, leaving 45 amputations in 42 patients available for evaluation. Nine of these were AK, twenty-two BK, five knee disarticulations (EX), and nine foot amputations. Male patients predominated (69%), and 44% were diabetic. Delayed healing was defined as

T.A.A.

292

van d e n B r o e k

failure of the skin to heal after more than 3 weeks, and complete closure which took more t h a n 8 weeks.

Clinical data Apart from the history (age, diabetes and rest pain) data were collected concerning infection of the limb. Classical criteria used to assess amputation level such as temperature, colour, capillary refill and pulsations were also scored. The presence of ulceration anywhere on the limb was assessed, and divided into dry necrosis, necrosis with the local red discolouration, purulent discharge, or a known preoperative positive culture. The temperature at each possible level of amputation was scored as colder or warmer t h a n the contralateral limb. A blue or red extremity was compared with the skin of the other limb, as was a slow or fast capillary refill. Pulses were scored as present or absent at the femoral, popliteal or ankle level.

Vascular laboratory In the vascular laboratory Doppler ultrasound segmental pressures and indices (DUS), segmental pulse volume recordings (PVR) and segmental PPG/SPP measurements were obtained at the AK, BK and foot amputation levels. Pulse volume recordings were made using the technique and standards described earlier 7 and categorised 1 - 4 to predict healing. The absence of a PVR (category 5) should predict failure to heal. With a 8 MHz doppler flow transducer and cuffs of variable width, systolic pressures were measured and indices calculated. For PPG/SPP, we used a set-up with a PPG-detector (Photopulse Model PPG-13 and sensors Model PH77, Medasonics). The photoelectric-cell was taped to a paediatric blood pressure cuff and then secured on the leg with a normal blood pressure cuff. Recordings are made from the output of the photocell, and the pressure in the paediatric cuff. Air is then rapidly inflated into the small cuff, raising the pressure to above systolic. As soon as inflow of red cells into skin capillaries starts, the direction of the PPG-trace changes and SPP can be read from the other (pressure) trace (Fig. 1). Measurements were performed posteriorly and anteriorly on the foot, lower leg and upper leg, at the level of possible incisions for amputations.

I

I

F,o

:L _i

__}_j_j

Fig. 1. Traces of cuff pressure (upper) and PPG (Lower). The actual PPG/SPP in this case is 20 mmMg.

included in the preoperative schedule in 34 of 45 c a s e s - a m o n g which were all the failed amputations except one. Although we were aware that angiography gives only a two-dimensional image, it was scored preoperatively in four regions: 1. the aorto-iliac vessels; 2. the deep and superficial thigh arteries; 3. the three below knee arteries and 4. the plantar arch. For each region the arteries were given: O points if no disease was present; 1 point with a 50% stenosis; 2 points with a 75?/0 stenosis; 3 points with complete occlusion. Since obstructions below the proposed level of amputation should not influence healing, only the scores at levels proximal to the amputation were added to each other. Scores therefore ranged from O to 12 points. For example, if a patient had normal aorto-iliac vessels, this region scored zero. An obstructed superficial femoral artery, but normal deep arteries, scored 1 point because of an estimated flow reduction of 50%. A normal popliteal artery but two obstructed and one normal crural arteries m e a n t approximately 259/o residual flow, and therefore 2 points. A totally obstructed plantar arch scored 3 points. W h e n such a patient was submitted to a below-knee amputation, the scores for the first three regions were added, giving a total of 3 points; for a transmetatarsal amputation this would have been 6 points. All scoring of the clinical and angiographic data was done by one of the vascular surgeons without knowledge of the decision on where to amputate. The final decision was made by one of the other vascular surgeons using his o w n clinical criteria.

Angiography Analysis Since in m a n y patients the possibility of reconstructive arterial surgery had been considered, angiography was Eurl VascSurg Vol 4, June 1990

Statistical analysis was performed using the two-tailed

Multivariate Analysis of Criteria for Amputation Wound Healing

Table 1. Results (percentage or mean and standard error) of preoperative criteria in amputations and in healing and non-healing groups after amputation. 4-1, Significance of difference ( P - v a l u e , threshold = 95 %). Healed

Failed

Significance

Age (years) Diabetes Pain at rest

71.2 4-1.9 39.5% 68.4%

77 4- 2.2 71.5% 71.4%

N.S. N.S. N.S.

Necrosis Local red colour Pus Positive culture

81.6% 86.8% 60.5% 34.2%

100%

N.S. N.S, N.S. N.S.

85.7% 57.1% 42.9%

Low skin temperature Equal/high skin temperature Blue extremity colour Red extremity colour Slow capillary refill Fast capillary refill

26.3% 18.4% 10.5 % 31.6% 15.8% 15,8%

14.3% 0% 14, 3 % 0% 28.6% 0%

N.s. N.S. N.S. N.S. N.s. N.s.

Femoral pulses Popliteal pulses

8 9.5% 13.2%

8 5.7% 2 8.6%

N.S. N.S.

PVR-pulsatility DUS (mmHg) Posterior PPG (mmHg) Anterior PPG (mmHg)

86% 68.64-9.5 49.5:t:4 46.34- 3.4

80% 51.3+27 29.64-2.8 19 4-1.2

N.S. N.S. P=0.01 P < 0.001

Angiographic score

4.1 4- 0.3 2

6.8 4- 0.4

P = 0.OO16

probability Mann-Whitney U test for comparison of differences in samples. The Spearman procedure for determination of rank correlation coefficients was used thereafter to demonstrate a direct correlation between wound healing and the values of the tested criteria. A receiver operator characteristics curve analysis was applied to visually emphasise the sensitivity and specificity of tests with a demonstration of optimal thresholds.

86 80

l

~

293

Healed Failed

68.6

-~51.5

~

/ 68 49.5

46:5

29. 19

PVR

DUS

PPGp

PPGa

ANGxIO

Fig. 2. Mean values of vascular laboratory criteria and angiographic score in healing and failing wounds after amputation. PVR in percentage pulsatile signal: DUS, PPGa (posterior) and PPGa (anterior) in mmHg pressure; (Angiographic score) in points x 10. Significant differences were found with PPGp, PPGa and Angiography.

measurements were performed (mean: 7.4) and found to show a highly significant difference between healing and failing wounds, particularly w h e n measured anteriorly. The optimal correlation coefficient for posterior measurements was r = 0 . 5 2 (at a threshold of 3 9 m m H g , P = 0.003) and anteriorly r = 0 . 8 3 . The angiographic score also proved to be a reasonable test with significant differences between healing and failing amputations (Fig. 2). The values of all vascular laboratory criteria between healing and failing wounds are shown in Fig. 2. Although attention was fixed on skin healing at all amputation sites, the data was also analysed for each

.'zL--LZ'=-~

,oo

251 "65 ~ / /

24 ~./..Y.... _

...... /'--"""

Results •~

Overall healing was 84% in this series. None of the AK amputations failed to heal but there was delayed healing in two patients. There was no correlation between a n y of the clinical criteria (Table 1). A pulsatile PVR at the level of amputation was found in most healing as well as failing wounds, (Table 1) and DUS failed to show a significant difference between healing and failing groups either. With PPG/SPP, absolute values both anteriorly and posteriorly at the level of amputation proved valuable. The simplicity of the technique enabled us to do five to eight measurements as well as the other non-invasive investigations in 2 0 m i n . In 53 patients 394 PPG/SPP

~

........-'"'

5o

1(30

50

o

Specificity (%)

Fig. 3. Receiver operator characteristics curve analysis (sensitivity versus specificity) of threshold values from angiographic score (Ang), doppler segment pressures (D), anterior (Pa) and posterior (Pp) PPG/ SPP studies. The small numbers near the curves represent the thresholds at those points. EurJ Vase Surg Vol 4, June 1990

294

T.A.A.

van den Broek

Table 2. Sensitivity, specificity,accuracy,positivepredictive value, negative predictive value and correlation with significance of wound healing. Angiographicscore using 6.4 as a threshold, DUS with 28 mmHgas a threshold, posterior PPG/SPP with 33 mmHg as a threshold, and anterior PPG/SPP with 21 mHg as a threshold. Method--threshold

Sensitivity (%)

Specificity (%)

Accuracy (%)

PPV (%)

NPV (%)

Correlation R

Significance P

Angiographicscore

83

90

89

63

96

0.55

0.0016

PVR-pulsatility

86

20

77

86

20

0.06

N.S.

DUS

77

50

69

90

27

0.26

N.s.

PPG/SPP posterior

74

67

71

93

29

O.38

0.012

PPG/SPP anterior

97

86

91

97

86

O.83

0.0001

level, For distal amputations (foot and toe) SPP-PPG was significanty correlated with healing. At the toe the anterior surface was better (P=O.O01), whereas for the foot the posterior aspect ( P = O.015) was superior. For major amputation sites the anterior SPP-PPG was again the best test. In order to find the optimal separation value for the quantitative methods (angiographic score, DUS and PPG/SPP) at all sites together, we performed a receiver operator characteristic curve analysis on these data (Fig. 3). The best separation for angiographic score was between 5.5 and 6.5 for posterior PPG/SPP at 39 mmHg, and for anterior PPG/SPP between 21 and 25 mmHg. The curve from DUS values demonstrates either a good sensitivity, or a good specificity, but never both. The sensitivity and specificity of non-invasive and angiographic criteria are shown in Table 2. An additional observation evolved from the data on PPG/SPP. If we offered PPG/SPP to the surgeon as the gold standard for amputation level selection in ten patients an amputation at a level lower than that actually selected could have been performed. However, in at least three of these patients a proximal amputation level was chosen because these patients were thought not to be suitable for rehabilitation.

Discussion

The prediction of amputation healing remains a controversial problem. The clinical criteria currently used have often lead to high complication rates, and failed once more in our study. Although others found the lowest palpable pulse to be relevant 8 the accuracy of this physical sign has been questioned by others 9 which is confirmed by our results. The first non-invasive technique used to predict amputation levels was PVR, which unfortunately has a predictive value of only 30 to 50%, 4' 5 a finding also confirmed by our study. DUS is at present the most frequently EurJ VascSurg Vol 4, June 1990

employed technique for predicting healing but in comparison with other methods, [the receiver operator characteristics curve analysis (Fig. 3)], DUS does not offer an acceptable threshold. This confirms other studies in finding high positive predictive values (9 7%) but low negative predictive values of 68%. 4'5,1° These values were worse when only distal amputations were taken into account. 11 Other techniques, such as transcutaneous oxygen tension measurement 12 are not often available in the vascular laboratory and it takes 2 0 m i n for each single measurement. Without oxygen inhalation low values have proved a poor predictor of amputatio n level quite apart from being time consuming. ~3 For these reasons this technique is not routinely used. Skin perfusion pressure or skin bood flow measured scintigraphically are known to be superior techniques in amputation level selection ~4 and in diabetic patients the perfusion pressure method may be used as well. 15 These scintigraphic techniques however are invasive, somewhat cumbersome and time consuming. The use of PPG/SPP to select amputation level is a new method which has not previously been compared with other methods. Although PPG instruments in the arterial mode respond very well to small reduction in perfusion pressure, the low flow found at potential amputation sites are better assessed in venous (DC) mode, as used in our series. In addition, the calibration problem of PPG is avoided by using perfusion pressure, instead of the transducer signal itself for determination of thresholds. 16 Scintigraphic SPP proved to give a reliable level of selection with thresholds at 20 to 3 0 m m H g . 14,15 The same thresholds evolve with PPG/ SPP: at 21 mmHg an optimal sensitivity is achieved (97%), whereas specificity is maximal at 2 5 m m H g (100%). It is not surprising that anterior PPG/SPP predicts healing better than posterior measurements as skin circulation is more critical in the anterior areas especially at the below knee level. In our study good results were also obtained with angiographic scoring. Of course this parameter assesses arterial flow at a level quite different from PPG/SPP: the

Multivariate Analysis of Criteria for Amputation Wound Healing

level of l a r g e a r t e r i e s , s o m e d i s t a n c e a w a y f r o m t h e h e a l i n g site. S P P m e a s u r e m e n t o n l y e v a l u a t e s s k i n , w h i c h is indirectly influenced by the transport capabilities of large a r t e r i e s . T h e s i g n i f i c a n c e of t h e s e d i f f e r e n c e s is i m p o r t a n t as s k i n b l o o d flow d e t e r m i n e s h e a l i n g o f t h e w o u n d w h e r e a s v a s c u l a r i s a t i o n of t h e m u s c l e s will a l m o s t cert a i n l y h a v e a b e a r i n g o n t o l e r a n c e of t h e p r o s t h e s i s , muscle function and rehabilitation. Our results suggest t h a t for f u t u r e a s s e s s m e n t of h e a l i n g a n d r e h a b i l i t a t i o n a c o m b i n a t i o n of p a r a m e t e r s will p r o v e n e c e s s a r y , e.g. a n g i o g r a p h y a n d SPP.

References 1 FINCH DRA, MACDOUGALM, TIBBSDJ, MORRIS PJ. Amputation for vascular disease: the experience of a peripheral vascular unit. Br J Surg 1980;67:233-237. 2 KEAGYBA, SCHWARTZJA, KOTBM, BURNHAMS], JOHNSONG. Lower extremity amputation: the control series. ] Vasc Surg 1986; 4: 321326. 3 STEINBERGFU, SUNWOOKI, ROETTGERF. Prosthetic rehabilitation of geriatric amputee patients: a follow-up study. Arch Phys Med Rehabil 1985;66:742-745. 4: GIBBONSGW,. WHEELOCKFC, SIEMBIEDAC, HOAR S, ROWBOTHAMJL, PERSSONAB. Noninvasive prediction of amputation level in diabetic patients. Arch Surg 1979;114:1253-1257. 5 NICHOLASGG, MYERSJL, DEMUTHWE. The role of vascular laboratory criteria in the selection of patients for lower extremity amputation. Ann Surg 1982;195:469-673. 6 BROEKTAA VAN DEN, DWARSBJ, RAEWERDAJA, BAKKERFC. Photo-

295

plethysmographic selection of amputation level in peripheral vascular disease. ]Vasc Surg 1988 ; 8:10-13. 7 RAINESJK, DARLINGRC, BUTH], BREWSTERDC, AUSTENWG. Vascular laboratory criteria for the management of peripheral vascular disease of the lower extremities. Surgery 1976; 79: 21-29. 80'DwYER KJ, EDWARDSMH. The association between lowest palpable pulse and wound healing in below knee amputations. Ann Royal Coll Surg England 1985 ;67:232-234. 9 MYERSKA, SCOTTDF, DEVINETJ et al. Palpation of the femoral and popliteal pulses: a study of the accuracy as assessed by agreement between multiple observers. Eur J Vasc Surg l 987; 1 : 245-249. 10 BARNESRW, THORNmLLB, NIX L, RITTaERSSE, TURLEYG. Prediction of amputation wound healing. Arch Surg 1981 ; 116: 80-83. 11 WELCH GH, LEIBERMANDP, POLLOCKJG, ANGERSONW. Failure of doppler ankle pressure to predict healing of conservative forefoot amputations. Br J Surtl 1982 ; 72 : 888-891. 12 CINA C, KATSAMOURISA, MEGERMANNJ, BREWSTERDC, STRAYHORN EC, ROBISONJG, ABBOTTWM. Utility of transcutaneous oxygen tension measurements in peripheral arterial occlusive disease. ] Vasc Surg 1984;1:362-371. 13 McCOLLUMPT, SPENCEVA, WALKERWF. Oxygen induced inhalation changes in the skin as measured by transcutaneous oximetry. Br J Surg 1986; 73 : 882-885. 14 HOLSTEIN PE. Skin perfusion pressure measured by radioisotope washout for predicting wound healing in lower limb amputation for arterial occlusive disease. Acta Ortop Scand 1985;56:suppl 213 (Thesis). 15 DWARSnJ, RAUWERDAJA, BROEKTAA VAN DEN, HOLLANDERW DEN, HEIDENDALGAK, RII GL VAN.A modified scintigraphic technique for amputation level selection in diabetics. Eur J Nucl Med 1989; 15 : 38-61. 16 KVERNEBOK, MEGERMANJ, HAMILTONG, ABBOTtWM. Response of skin photoplethysmography, laser doppler flowmetry and transcutaneous oxygen tensiometry to stenosis-induced reductions in limb blood flow. Eur l Vasc Surg 1989;3:113-120. Accepted 29 September 1989

Eur J Vasc Surg Vol 4, June 1990