hbo pada diabetic foot

8/13/2019 Hbo Pada Diabetic Foot

1/6

O R I G I N A L A R T I C L E

Adjunctive Systemic Hyperbaric OxygenTherapy in Treatment of SeverePreva lently Ischemic D iabetic Foot UlcerArandomized studyEzio FAGLIA,MDFABRIZIO FAVALES,MDANTONIO ALDEGHI,MDPATRIZIA CALIA,M DANTONELLA QUARANTIELLO, MD

GIORGIO ORIANI,M DMICHAEL MICHAEL,M DPLETRO CAMPAGNOLI,MDALBERTO MORABITO,P HD

OBJECTIVE To evaluate the effectiveness of systemic hyperbaric oxygen therapy (s-HBOT) in addition to a comprehensive protocol in decreasing major am putation rate in diabeticpatients hospitalized for severe foot ulcer.R E S E A R C H D E S I GN A N D ME T HOD S From August 1993 to August 199 5, 70 dia-betic subjects were consecutively admitted in to ou r diabetologic unit for foot ulcers. All the sub-jects underw ent our diagnostic-therapeutic protocol and were randomized to undergo s-HBOT.Two subjects, one in the arm of the treated group and one in the arm of nontreated group, didnot c omplete the protocol and were therefore excluded from the analysis of the results. Finally,35 subjects received s-HBOT and a nother 33 did no t.RESULTS Of the treated group (mean session = 38.8 8), three subjects (8.6%) under-went major amputation: two below the knee and one above the knee. In the nontreated group,11 subjects (33.3%) underwent major amputation: 7 below the knee and 4 above the knee. Thedifference is statistically significant (P=0.016). The relative risk for the treated grou p was 0.26(95% CI 0.08-0.84). The transcutaneous oxygen tension measured on the dorsum of the footsignificantly increased in subjects treated with hyperbaric oxygen therapy: 14.0 11 .8 mmHg intreated group, 5.05.4 mmHg in nontreated group (P=0.0002). Mu ltivariate analysis of majorampu tation on all the considered variables confirmed the protective role of s-HBOT (odds ratio0.084, P=0.033, 95% Cl 0.008-0.821) and indicated as negative prognostic determinants lowankle-brachial index values (odds ratio 1.715,P=0.013, 95% CI1.121-2.626)and high Wag-ner grade (odds ratio 11.199, P=0.022, 95% CI 1.406 -89.146).C O N C L U S I O N S s-HBOT, in conjunction with an aggressive multidisciplinary thera-peutic protocol, is effective in decreasing major amputations in diabetic patients with severeprevalently ischemic foot ulcers.

Systemic hyperbaric oxygen therapy(s-HBOT) has been used in the treat-ment of diabetic wounds (1) butopinions differ regarding its effectiveness(2,3). The aim of this study has been to evaluate the effectiveness of s-HBOT indecreasing major amputation (thigh orankle) rate in a randomized record ofcases of diabetic patients hospitalized forfoot ulcer.FromtheDiabetology Center(E.F, EF, A.A., RCal.,A.Q.), Niguarda Hospital;theDepartmentofAnesthe-sia and Hyperbaric Medicine (G.O., M.M., PCam.), Galeazzi Institute; and theMedical Statisticsand Bio-metrics Institute (A.M.),Milan University, Milan, Italy.

Address correspondence and reprint requests to Ezio Faglia, Diabetology Center, Niguarda Hospital,Piazza Ospedale Maggiore,320162 Milan, Italy.

Received forpublication 1April 1996 and acceptedinrevised form8August 1996.ABI,ankle-brachialindex;ATA,absoluteatmosphere;BPG,bypass graft; PTA,percutaneoustranslumi-nalangioplasty;ROC, receiver operating curve; s-HBOT, systemic hyperbaric oxygen therapy; TcPo2, trans-cutaneous oxygen tension.

RESEARCH DESIGN ANDMETHODSStudy DesignFrom August 1993toAugust 1995, 70 dia-betic subjects, consecutively hospitalized inour diabetologic unit for foot ulcer, under-went our diagnostic and therapeutic proto-col.All the subjects were random ized for s-HBOT treatment. All patients gave theirinformed consent. One subject randomizedfor s-HBOT refused the treatment; one sub -ject, randomized for the non-s-HBOT, diedof an acute stroke 6 days after admission.Both these subjects were excluded from theanalysis of the results. Of the subjects, 35underwent s-HBOT and 33 did not. Clini-cal characteristics of the study populationare shown in Table 1. After random ization,none of the variables listed in Table 1showed a significant imbalance betweenthe treatment and control arm.Diagnostic and therapeutic protocolOn admission to the hospital, lesions wereclassified according to Wagner (4). In ourclinical practice, diabetic subjects withfull-thickness gangrene (Wagner grade IV)or abscess (Wagner grade III) were admit-ted to hospital. Subjects with less-deepulcers (Wagner grade II) were also admit-ted if the ulcer was large and infected andshowed a defective healing in 30 days ofoutpatient therapy. All patients wereexamined for diabetic retinopathy (fundusoculi by ophthalmologist), albumin excre-tion rate (mg/24 h , the average of three 24-h collections, nephelometry-Behring),renal impairment (creatinine >133umol/1,Jaffe-Boehringer Mannheim), arte-rial hypertension (systolic blood pressure>1 60 mmHg and/or diastolic blood pres-sure >95 mmHg or antihypertensive ther-apy), coronary artery disease (CAD)(CAD-resting electrocardiogram and B-mode echocardiography), obesity (BMI>24 kg/m2 for women, >25 kg/m2 formen), dyslipidemia (total cholesterol>6.20 mmol/1, colorimetry, Boehringer

1338 DIABETES CAR E, VOLUME 19, NUMBER 12, DECEM BER 1996

at Indonesia: ADA Sponsored on January 22, 2014http://care.diabetesjournals.org/Downloaded from
http://care.diabetesjournals.org/http://care.diabetesjournals.org/http://care.diabetesjournals.org/http://care.diabetesjournals.org/http://care.diabetesjournals.org/


2/6

Faglia and A ssociates

Table 1Clinical characteristics ojs HBOT and non s HBOT groups

nMenWomenAge (years)Insulin therapyOral therapyDiabetes duration (years)Wagner grade

11111IV

Prior major amputationPrior minor amputationPrior lesionBackground retinopathyProliferant retinopathyMicroalbuminuriaProteinuriaRenal impairmentIlypertension1lyperlipidemiaObesitySmoking habitCoronary artery diseasePrior strokeInfectionPolimicrobial infectionInfection recoveryBone lisisOstepeniaMonckeberg sclerosisPeripheral angiographyHb A lcat admission (%)Hb A lc at discharge (%)Total hospital stay (days)

s-HBOT group35

27(77 .1 )8 (22.9)

61.7 10.421 (60)14 (40)16 104(11 .5 )9 (25 .7 )

22 (62.8)0 ( )6(17 .1 )9 (25.7)

12 (34.2)13(37.1)12 (34.3)

8 (22.8)4 (11 .4 )

19 (54.2)11(31.4)9 (25.7)

11(31.4)14 (40)

3 (8.6)32 (91.4)20 (57)26 (74.2)11 (31.4)15 (42.8)21 (60)31 (88.5)

9.3 2.57.1 1.5

43 .2 31

non-s-HBOT group33

21 (63.6) 112 (34.4) j65.6 9.122 (66.7) 111(33.3) j19 95(15 .2 )8 (24.2)

20 (60.6)0 ( )

10 (30.3)12 (36.4)13 (39.4)9 (27.3)9 (27.3)7 (21 .2 )9 (27.3)

17(51.6)8 (24.2)9 (27.3)

12 (36.4)15 (45.4)4 (12 .1 )

28 (84.8)17(51.6)17(51.6)

9 (27.3)21 (63.6)20 (60.6)26 (78.8)

8.5 2.36.6 1.2

50.8 32

P value

u.zy0.10n fou.oz0.20

0.94

0.250.430.800.440.601.000.131.000.131.000.501.000.790.800.800.080.790.091.000.330.170.130.37

Data are means SD orn (%). P values were determ ined by a two-tailed Fishers exact test for discrete vari-ables and by an unpaired Student's t test for continuous variables. Microalbuminuria was defined as analbumin excretion rate S20 and 200 mg/24 h; hypertension was defined according to World Health Organization criteria or antihyper-tensive treatment; hyperlipidemia was defined by total cholesterol >6.20 mmol/1 and/or HDL cholesterol24 and >25 kg/m 2 for women and men, respec-tively; and bone lisis and ostepenia were determined by radiographic findings.

Mannheim; and/or HDL cholesterol2.25mmol/1, colorimetry, Ames; or hypolipi-demic therapy). On admission and at dis-charge, glycosylated hemoglobin levels(HbA lchigh-pressure liquid chrom atogra-phy, normal values 4.4-6%) were meas-ured. Specimens of the foot lesion, afterdecontamination and debridement fol-lowed by curettage (5), were collected foraerobic and anaerobic culture and for

antimicrobial susceptibility testing toantibiotics. Susceptibility testing to topicalantimicrobial agents was also performedaccording to a standardized protocol setup in our microbiology laboratory (6). X-rays were taken of both feet and legs todiscover medial arterial calcifications andbone abnormalities. The sensorimotorneuropathy (7) was investigated with elec-tromyography in all subjects (consideredpresent when showing abnormalities ofnerve conduction velocity and sensoryaction potential in at least two nerves).

The autonomic neuropathy (8) (present ifthe score was > 4 in the five standardautonomic cardiovascular tests) and thevibration sense (9) (impaired if the vibra-tion perception threshold m easured on themalleolus with biothesiometer was >25V) were investigated when technicallypossible in collaborative patients. Theankle-brachial blood pressure ratio (ankle-brachial index [ABI]) was measured byDoppler continuous wave technique. Thetranscutaneous oxygen tension (TcPo^)was measured on the dorsum of the footon adm ission to hospital and on dischargefor the subjects with salvaged limbs andbefore amputation in subjects undergoingamputation. In Table 2, the assessment ofneuropathy and vasculopathy in the studypopulation is reported. The values of theseparameters showed no significant imbal-ance between the two arms. In the sub-jects undergoing s-HBOT, TcPo,i was alsomeasured during treatment in the hyper-baric chamber.

In all subjects an aggressive and radi-cal debridement was performed by a con-sultant surgeon. After surgical curettagethe wound was cleaned with uncolorcdtopical antimicrobial agents and waddedwith occlusive dressing (10). Dressing,with debridement ifnecessary,was carriedout not less than twice a day when necro-sis or exudate were present, daily whenthe ulcer was clean, and every two daysduring the granulation period. On admis-sion to hospital all patientsafter collect-ing a specimen of the ulcer for cultureexaminationwere given empiricalbroad-spectrum antibiotic therapy, subse-quently modified ifnecessary,according tosusceptibility testing results. The antibiotictherapy was continued during the hospitalstay until the culture exam, repeated eachweek, was negative. After discontinuationof the antibiotic therapy, rcculturing toassess the cure was performed every twodays a total of three times. An optimizedmetabolic control was pursued either withsubcutaneous insulin administrations ororal hypoglycemic agents, according toblood glucose determinations, 7 times/day.For blood glucose levels >22 umol/1, aprocedure of intravenous insulin infusionwas administered according to an algo-rithm based on the assessment of bloodglucose levels every 2 h, until blood glu-cose value < 9.9 umol/1 was reached. Inthe subjects with ABI


3/6

Hyperbaric oxygen in diabetic foot ulcer

Table 2Assessment of peripheral vasculopathy and neuropathy in s-HBOT and non-s-H BOTgroups

nClaudicationABITcPo2(mmHg)Sensorimotor neuropathyImpaired vibration senseAutonomic neuropathy

s-HBOT group354(11 .4 )0.65 0.28

23.25 10.635 (100)24 (85.7)17 (73.9)

non-s-HBOT group33

10 (30.3%)0.64 0.25

21.29 10.731 (93.9)23 (85.2)15 (71.4)

P value0.070.870.450.231.000.70

Data are means SD orn (%). P values were determ ined by a two-tailed Fisher's exact test for discrete vari-ables and by an unpaired Student's t test for continuous variables. ABI is the ankle-brachial blood pressureratio measured by the Doppler technique; sensorimotor neu ropathy was defined by abnormalities in ^ 2nerves at electromiography; im paired vibration sense was defined by a vibration perception thresh old >2 5V at biotesiometry; and autonom ic neur opathy was defined by a score > 4 at five cardiovascular tests.

intra-arterial digital subtraction techniquewas performed if there were no con-traindications (creatinine > 22 1 um/1, orparaproteinemia) (11). In these subjectsthe opp ortunity and possibility of carryingout a percutaneous transluminal angio-plasty (PTA) or a bypass graft (BPG) wasassessed. The presence of focal stenosesinvolving > 50 % of vessel lumen was con-sidered an indication of PTA. The stenosescompletely occluding the lumen or withlength >10 cm were respectively consid-ered as an impossibility or a contraindica-tion forPTA.When there was an impossi-bility of performing PTA, the arteriogramwas evaluated by vascular surgeons tocarry out a BPG. Based on angiographiccriteria bypasses were performed when apatent vessel in continuity with the footwas present. During hospitalization, allpatients were provided with orthopedicdevices to remove mechanical stress andpressure at the site of the ulcer, whilemaintaining ambulation. The orthesis wasmade u p of an Alkaform insole molded inplastic cast and an extra deep special shoewith a rigid sole (Buratto, Italy) allowingthe insertion ofabandaged foot.The limb was considered salvaged

when the plantar support was preservedand the ulcer healed despite minor ampu-tations (toe or forefoot amputation), asthey are lost in presence of major amputa-tion (above or below the knee). The deci-sion to carry out a major amputation wastaken by the consultant surgeon who wasunaware of whether the s-HBOT wasadministered or not.s HBOTIn the grou p random ized for s-HBOT, thepatients breathed pure oxygen in a multi-

place hyperbaric chamber, pressurizedwith air, with a soft helmet. The chosenpressure, in our study, was 2.5 absoluteatmosphere (ATA) in the first phase, toenhance the antibacterial effect and toquickly restore a sufficient tissue partialpressure of oxygen. In the second phase,to stimulate a fibroblastic activity forreparative effect, we applied 2.4-2.2 ATA.Our scheme considers a daily session (90min for each session) in the first phase,and an hebdomadal (5/7) session in thereparative phase.

Statistical analysisThe sample size (12) of 34 patients/armwas decided to detect a reduction of 1/3 ofmajor amputation rate with type-oneerror, a = 0.05, and power 1, (3 = 0.80(two-sided test). The randomizationschedule adopted requires that a patientshould be allocated to the treatment armafter hospital admission by consulting a

table of random numbers at the hospitaldata elaboration center. Fishers exact testwas used for comparison of discrete vari-ables between the two arms. The relativerisk and the Wolf CIs were estimated forthe comparison of major am putation rates.The students t test for unpaired data wasused for the comparison of continuousvariables (13). Multivariate logistic regres-sion (Stata Statistical Software, Stata,1995) of major amputation has been per-formed considering the covariates listed inTables 1,2, and 5 and found to be signifi-cant in the univariate analysis. The relativeodds ratio and 95% CIs were calculated.The global validity of the multivariate sta-tistical model was tested using logisticreceiver operating curve (ROC).RESULTS The subjects who under-went s-HBOT attended an average of 38 8 sessions. Two subjects showed symp-toms of barotraumatic otitis, which didnot cause the interruption of treatment. Inthe s-HBOT group, three subjects under-went a major amputation (8.6%): oneabove the knee and two below the knee.In the non-s-HBOT group, 11 subjectsunderwent major amputation (33.3%): 4above the knee and 7 below the knee. Thedifference is statistically significant (P =0.016) (Table 3).The relative risk for thesubjects treated with s-HBOT is 0.26(95% CI 0.08-0.84). Major amputationwas performed in the patients of the s-HBOT group after 57.6 2 4 days (range:31-78) from hospital admission, and after72.8 59 days (range: 26-176) in thepatients of the non-s-HBOT group. InTable 3 the major amputation rate relatedto Wagner grade, and the minor amputa-

Table 3Major and minor amputation rates of s-HBOT and non-s-HBOT groups

nMajor amputationsAmputated limbsSalvaged limbs

Major amputation/Wagner gradeIIIIIIV

Minor amputationsForefootTo eNo amputation

s-HBOT group35

3 (8.6)32 (91.4)

0/4 ()1/4 (25)2/22(9.1)5 (14.3)

16 (45.7)11(31.4)

non-s-HBOT group33

11(33.3) 122 (66.7) ]

0/5 ()0/8 ()11/20(55)4(12.1)8 (24.2)10 (30.3)

Pvalue

n m U.U1D

0.330.002

0.61

Data aren (%). P values were determined by a two-tailed Fisher's exact test.

1340 DIABETES CAR E, VOLUME 19, NUMBER 12, DECEM BER 1996



4/6

Faglia and Associates

Table 4TcPo2 valuesof s HBOT and non s HBOT groups at admission and at discharge;comparison ofincre sebetween the two groups

s-HBOT group non-s-H BOT group P valueAt admissionAt dischargeVariation

23.237.314.0

35 1 0 . 1 6 . 1 1 .

718

21.326.3

5.0

33

10.13.5.4

75

0.46 2

Data are means SD and are given as TcPo 2(mmHg). P values were determined by an unpaired Studentst test [Satterthwaite (13) degrees of freedom: 48.25].

tion rate between the two arms was alsocompared.The comp arison between the values ofTcPo2on admission and at discharge orbefore amputation are shown in Table 4.The increase of TcPo2 in the s-HBOTgroup is 9.05 mmHg higher than that ofthe non-s-HBOT group (P =0.0002). Themean jump of TcPo2 level in hyperbaricchamber was 493.5 152.1 mmHg. Vas-cular procedures were performed in 26subjects (17 PTAs, 9 BPGs). The PTA wasperformed as follows: 6 in the iliac orfemoral arteries (3 in the s-HBOT and 3 inthe non-s-HBOT group) and 11 in thepopliteal or infrapopliteal arteries (6 in thes-HBOT and 5 in non-s-HBO T group). Allthe BPG were carried out using saphenousvein in situ: in four the site of distal anas-tomosis was the popliteal or infrapoplitealarteries (two in the s-HBOT and three inthe non-s-HBOT group) and in five thedorsalis pedis artery (two in the s-HBOTand two in the non-s-HBOT group). Theoutcome of vascular procedures is shownin Table 5.

Multivariate analysis, carried out onthe variables found associated with majoramputation in the univariate analysis,showed the protective role of s-HBOT(odds ratio 0.084, P = 0.033, 95% CI0.008-0.821) and indicated as negativeprognostic determinants Wagner grade(odds ratio 11.199, P = 0.022, 95% CI1.406-89.146 ) and ABI value reduction(odds ratio 1.715, P = 0.013, 95% CI1.121-2.626). The calculated odds ratioforABIvalues express the increased risk ofmajor am putation per 0.1 unit decrease ofthe variableitself.The area unde r the ROCcurve calculated from the multivariatelogistic model is 0.9501 (Fig. 1).CO NC LU SI O N S Diabe tic u lcersfrequently do not heal because ofacombi-nation of hypoxia and infection (14). Sys-temic hyperbaric oxygen greatly increases

tissue oxygen levels even though treatmentis only partial throughout the day: oxygentension values by TcPo2 remain elevatedfor several hours after exposure (15). Theoutcome of breathing o 2hyperbarism is anincrease in the diffusion of oxygen physi-cally dissolved inplasma.The supplyingofoxygen to the cells carried out by thecounter-gradient diffusion of concentrationsustains, in a hypoxic situation, m itochon-drial breathing and cell survival, so pre-venting necrotic development of the tissue.Experiments on animals have proved thistheory (16). The index of oxygenation andhema tic level ofthelactates, after occlusionof the femoral artery or the abdominalaorta, show that an inhalation of 100%oxygen un der pressure of 3ATAm aintainsan elevated level of oxygen p artial pressureand reduces the lactates, in contrast withwhat happens in a normobaric condition(17). s-HBOT generates a vascular con-striction as a reflex mechanism defendingagainst hyperoxia (18). In the hypoxic tis-sue this mechanism is able to opposethe compensative vasodilatation of thehypoxia, and it leads to a reduction of theedema, which is often present in the dia-betic foot (19), withasubsequent improve-ment of the microcirculatory flow (20).

s-HBOT has significant direct andindirect effects on the infection. s-HBOThas a direct antibacterial effect on the

anaerobic microorganisms. Both the pro-duction of toxins and the growth of thesebacteria are completely inhibited b y a highlevel of oxygen (21). Moreover, by main-taining an 0 2partial pressure higher than30 m mH g at tissue level, the h yperbarismallows the macrophagic killing activity o?dependent (22). The defective woundhealing appears to be an important factorcontributing toward limb loss in diabeticsubjects (23): a restoration of tissue oxy-gen tension by s-HBOT also assists angio-genesis and advancement into the woundspace (24).

On this basis, for some time now wehave introduced s-HBOT in our therapeu-tic protocol (25). However, although otherstudies have reported increased limb sal-vage with s-HBOT (26,27), support forthis treatment has been judged question-able (3), and even in a recent review the s-HBOT was not included in the armamen-tarium of diabetic foot care (28). This wasdue to the fact that the results on the effec-tiveness of s-HBOT were based on non-randomized studies (3). This study showsthat s-HBOT is effective in decreasingmajor amputations in diabetic subjectswith foot ulcers. This result was obtainedin patients with severe foot ulcer; Wagnergrade IV was the most frequent in ourstudy population. Although in our studypopulation there were many subjects withneuropathy and infection, it is likely thatthe arterial insufficiency, which was con-sidered the most typical characteristic ofWagner grade IV, is the predominant fac-tor leading to major ampu tation. This con-sideration is confirmed by the indicationof low ABI value and high Wagner gradeas negative independent prognostic deter-minants for major amputation in the mul-tivariate analysis. The ROC curve showsan excellent predictive value of this modelin accordance with a wide range of oddsratio. Our data suggest that s-HBOT is

Table 5Outcome of vascular procedures in s H BOT and non s HB OT groups

nPeripheral BPGMajor amputationPercutaneous angioplastyMajor amputationTotal vascular p roceduresTotal major amputations

s-HBOT group35

4/35(11.4)1/4 (25)

9/35 (25.7)1/9 11.1)

13/35(37.1)2/13 (15.4)

non-s-HBOT group33

5/33(15.1)1/5 (20)8/33 (24.2)3/8 (37.5)

13/33 (39.4)4/13 (30.8)

P value

0.731.001.000.291.000.60

Data are n(%). P values were determined by a two-tailed Fishers exact test.

DIABETES CAR E, VOLUME 19, NUMBER 12, DECEM BER 1996 1341



5/6

yperbaric oxygen in diabetic foot ulcer

re unde r ROC cur ve = 0.9 501

1.00 -

75

5

25

Q O B O O J

hbo pada diabetic foot

Documents