From Guideline to Practice :Before and After Surviving Sepsis

Campaign in Taiwan

余忠仁臺大醫學院內科副教授

臺大醫院綜合內科部副主任

Surviving Sepsis Campaign

• Phase I: public awareness (2002)• Phase II: forming guidelines (2004)• Phase III: implementation of

guidelines• Goal: To reduce the mortality of

severe sepsis/septic shock by 25% within 5 years (2009)

Surviving Sepsis CampaignGuidelines for Management of Severe Sepsis and Septic Shock

Dellinger RP, Carlet JM, Masur H, et al. Surviving Sepsis Campaign Management Guidelines Committee. Crit Care Med 2004; 32:858-873

Goal: To reduce the mortality of severe sepsis/septic shock by 25% in 2009

Management of Sepsis

• Infection control– Antibiotics, source control

• Hemodynamic support– Ventilation, infusion, pump

• Metabolic/endocrine support– Steroids, glucose control

• New drug– Activated protein C

49.2%

33.3%

0

10

20

30

40

50

60

Standard Therapy n=133

EGDTn=130

P = 0.01*

*Key difference was in sudden CV collapse, not MODS

28-day Mortality

Early Goal-Directed Therapy

River E. N Engl J Med 2001;345:1368.

Central venous arterial catheterization

CVP or FTc CVP < 8 mm Hg or FTc < 330 msec

CVP 8-12 mm Hg or FTc 330-360 msec

500-ml bolus of NS or LR every 30 min prn

< 65 mm HgMAP

MAP >65

Norepinephrine (adjustable dosage) ± vasopressin 0.04 U/min (fixed dosage)

Cl and ScvO2 or SvO2 Cl < 2.0 and ScvO2 or SvO2 <70%

Cl > 2.0 and ScvO2 or SvO2 >70%

Dobutamine 5 g/kg/min(adjustable dosage)

Goals achievedGoals achieved River E. N Engl J Med 2001;345:1368.

0

5

10

15

20

25

30

35

40

6 ml/kg

12 ml/kg

% M

ort

alit

y

ARDSnet Mechanical Ventilation Protocol

Results: Mortality

(The ARDS Network, NEJM 2000;342:1301)

61%53%

0%

20%

40%

60%

80%

100%

53%63%

0%

20%

40%

60%

80%

100%

Low-dose Steroids Placebo

Patients with Relative Adrenal Insufficiency (ACTH Test Non-

responders) (77%)

Patients Without Relative Adrenal Insufficiency (ACTH Test

Responders) (23%)

P=0.04 P=0.96

N=114 N=36 N=34N=115

28-d

ay M

ort

alit

yLow Dose Steroids and Septic Shock

(Annane et al. JAMA 2002;288:862)

• Relative adrenal insufficiency: max < 9 g/dL, 30-60 min after test

Intensive Glucose ControlInitial Infusion RateBlood Glucose Level (mg/dl) Insulin Infusion Rate(U/hr)110-220 2> 220 4

Blood Glucose Monitoring GuidelinesAccuchecks every hour during insulin infusion until four consecutive values are within 80-110 mg/dl, then every 4 hours. If tube feedings or total parenteral nutrition is held or discontinued, hold infusion and monitor blood glucose levels every 2 hours.

Insulin Infusion Titration GuidelinesBlood Glucose Level (mg/dl) Insulin Bolus and Infusion Rate41-60 Stop infusion61-80 Reduce rate by 0.1-0.5 U/hr81-111 No change unless decreased > 20% from previous result; if > 20%, decrease rate 20%111-120 Increase rate by 0.1-0.5 U/hr121-139 Increase rate by 0.5-1 U/hr> 140 Increase rate by 2 U/hr

The Role of IntensiveInsulin Therapy in the ICU

Berghe G, et al. N Engl J Med 2006;354:449-61

Su

rviv

al (

%)

100

00

Intensive treatment

Conventional treatment

Days after admission

20

40

60

80

100 200

P=0.02

300 400 500

MICU3 days

Su

rviv

al (

%)

100

00

Intensive treatment

Conventional treatment

Days after admission

80

84

88

92

96

50 100 150 200 250

P=0.01

Berghe G, et al. N Engl J Med 2001;345:1359-67

Surgical ICU

X

X

X

X

X

Coagulation Fibrinolysis Inflammation

Coagulationcascade

VIIIa

Va

Prothrombin

Thrombin

Tr Thrombomodulin

PC

PAI-1

aPC PS

Increased fibrinolysis

Neutrophiladhesion

Monocyteadhesion

CD1/MHC

Thrombin

Kaplan-Meier survival curves

Days from start of infusion to deathPROWESS. N Engl J Med 2001; 344:699-709

0 2814 217

80

70

90

100

0p = 0.006 (stratified log-rank test)

Sur

viva

l (%

)

placebo(n = 840)

Drotrecogin alfa (activated)(n = 850)

Kaplan-Meier survival curves

Days from start of infusion to death(NEJM 2001; 344:699, CCM 2005;33:2266)

0 2814 217

80

70

90

100

0

Sur

viva

l (%

)

Placebo (n = 840)

DrotAA (n = 850)

ENHANCE　 DrotAA (n = 2375)

PROWESS

Importance of Establishing Therapeutic Goals in Managing Sepsis

Therapeutic goals Absolute Mortality Rate

Low VT MV, Pplat 30 cmH2O 9%

Appropriate antibiotic within 4 hrs 24%

EGDT 16%

Decreasing lactate to 2 mmol/L in 24hrs 25%

Steroid when indicated 10%

Insulin for glucose 80-110 mg/dL 3% (12 ms)

DrotAA for APACHE II 25 13%

Compliance to Guidelines

• ED of Hospital general Universitario “Gregorio Maranon”, Spain– Determination of blood lactate 12.5%– Blood culture done: 85%– Antibiotic within 3 hrs: 32%– Aggressive fluid therapy: 46.6%– Vasoactive drugs when indicated: 43.3%– CVP monitor: 0%

(de Miguel-Yanes JM. Am J Emerg Med 2006;24:553)

The EGDT protocol utilized at Cooper University Hospital (an adaptation of the protocol by Rivers et al)

(Trzeciak S et al. Chest 2006;129:225)

Implementing EGDT in EDCooper University Hospital

• 22 cases treated with EGDT, 20 completed– Median time to reach each end point was 6

hours – EGDT used more fluids, PRBC requirement in

ED– EGDT reduced ICU PAC utilization

(Trzeciak S et al. Chest 2006;129:225)

Implementation of Sepsis Guideline

• Bundle approach–Surviving sepsis bundle

–STOP sepsis bundle

–Multiple Urgent Sepsis Therapies (MUST) protocol

Surviving Sepsis Bundles• Sepsis Resuscitation Bundle

(to be accomplished ASAP and scored over first 6 hrs)– Serum lactate– Blood culture prior to antibiotic– Broad-spectrum antibiotics administered within 3 hrs for ED admissions and

1 h for non-ED ICU admission– If hypotension (SBP<90, or MAP <70) and/or lactate > 4 mmol/L

• Deliver initial minimum of 20-40 ml/kg of crystalloid (or colloid equivalents)• Apply vasopressors for hypotension not responding to initial fluid resuscitation to

maintain MAP 65 mmHg

– If persistent hypotension despite fluid resuscitation or lactate > 4 mmol/L• Achieve CVP of 8 mmHg• Achieve ScvO2 of 70% or SvO2 of 65%

(2005 surviving Sepsis Campaign and the Institute for Healthcare Improvement)

Surviving Sepsis Bundles

• Sepsis Management Bundle(to be accomplished ASAP and scored over 24 hrs)– Low-dose steroids administered for septic shock in accor

dance with a standardized ICU policy– Drotrecogin alfa (activated) administered in accordance w

ith a standardized ICU policy– Glucose control maintained lower limit of normal, but < 15

0 mg/dL– Inspiratory plateau pressure maintained < 30 cmH2O for

MV patients

(2005 Surviving Sepsis Campaign and the Institute for Healthcare Improvement)

Sepsis Bundles:Compliance vs. non-compliance

(Gao F et al. Crit Care 2005;9:R764)

Compliance to 24 h sepsis bundle

Compliance (%)

Glucose control < 8.3mmol/L 64%

Low dose steroid 43%

Activated protein C 30%

Plateau pressure < 30cmH2O 85%

Total 30%

(Gao F et al. Crit Care 2005;9:R764)

STOP Sepsis BundleED patients First 3 months After 6 months

Achieve EGDT in 6 hrs 8% 26%

Lactate monitor at 6th hr 22% 52%

Appropriate steroid therapy 49% 67%

In 208 sepsis patients treated, all components of the bundle were completed in 24 patients, the mortality rate was 12.5%, compared with a 34.2% mortality rate in whom the protocol was not completed, (P = .008), and the hospital length of stay averaged 8.1 days compared with 11.9 days for the 184 patients in whom the bundle was not completed (P = .06).

Nguyen HB et al SCCM 34th Critical Care Congress: Abstract 44

(Nguyen HB et al. Crit Care Med 2007;35:1105)

STOP Sepsis Bundle• 2 years of intervention

Bundle completed (n=77)

Bundle not completed (n=253)

P value

CVP/ScvO2 by 2 hrs 100% 51.8% <0.01

Antibiotics by 4 hrs 100% 87.4% <0.01

CVP goal met at 6 hrs 100% 35.6% <0.01

SBP/MAP goal met at 6 hrs 100% 63.2% <0.01

ScvO2 goal met at 6 hrs 100% 22.9% <0.01

Appropriate steroids 77.9% 58.5% <0.01

Lactate clearance 79.2% 39.1% <0.01

(Nguyen HB et al. Crit Care Med 2007;35:1105)

STOP Sepsis Bundle

(Nguyen HB et al. Crit Care Med 2007;35:1105)

Before-after standardized hospital order set study

(Scott M et al. Crit Care Med 2006;34:2707)

(Scott M et al. Crit Care Med 2006;34: 2707)

Before-after standardized hospital order set study

Before (n=60) After (n=60) P value

28 day mortality

48.3% 30.0% 0.04

Hospital mortality

48.3% 35.0% 0.139

Hospital LOS (d)

12.19.2 8.97.2 0.038

(Scott M et al. Crit Care Med 2006;34:2707)

Multiple Urgent Sepsis

Therapies (MUST) Protocol

(Shapiro et al. Crit Care Med 2006;34:1025-32)

MUST Protocol• Nov, 2003- Jan, 2004, 116 protocol patients an

d Feb, 2000- Jan, 2001, 51 historical control– Mortality rate 20.3% vs 29.4% (p=0.3)– More fluid (4.0L vs. 2.5L, p<0.001)– Earlier antibiotics (90 vs. 120 mins, p<0.013)– More appropriate coverage (97% vs. 88%, p<0.05)– More vasopressor in the first 6 hrs (80% vs. 45%, p

<0.001)– Tighter glucose control (123 vs. 140, p<0.001)– Frequent assessment of adrenal function (82% vs.

10%, p<0.001)(Shapiro et al. Crit Care Med 2006;34:1025-32)

Economic implications of sepsis protocol

(Shorr AF et al. Crit Care Med 2007;35:1257)

Implementing Sepsis Guideline in Taiwan

• Education

• Practice module– “Bundled” approach– Quality improvement

indicator• Information technology

• Research

• Taiwan Guideline

EDGT Taiwan

• 資料來源：台大醫院急診部李建彰醫師

台大醫院急診部敗血症病患標準治療流程 懷疑感染

至少符合下列兩點1. 體溫高於 38.3 ℃ 或低於 36℃2. 心率大於 903. 呼吸速率大於 20 或 PaCO2 小於 324. 白血球大於 12k 小於 4k 或 band > 10%

重新評估

檢查 lactate 做兩套血液培養

區分敗血症嚴重程度

YesNo

敗血性休克定義：半小時內輸注 0.9NS 500 cc 依舊休克

嚴重敗血症超過一個以上器功能障礙 或 lactate 大於 4

敗血症給予適當廣效性抗生素住

進入早期目標導向治療 (early goal directed therapy)六小時內完成下列目標CVP 12-15 SBP 90-140 MAP 65-90 SvO2 ＞＝ 70

目標一： CVP 12-15 ＊未達目標 0.9NS 繼續全速輸注直到 CVP 8-12 然後維持 100cc/hr, 輸液超過 2000cc 0.9NS, 未達目標 , 考慮進入目標二CVP < 4 可考慮給予 HAES ( 肝硬化患者考慮給予 Albumin ), 使用正壓換氣或有腹壓上升情形 , 應維持 CVP >=15

目標二： SBP 90-140 MAP 65-90考慮給予昇壓劑 Dopamine, Norepinehrine 若一小時仍 SBP<90 考慮給予 Vasopressin (0.04u/min) 及 dexamethasone 2mg IV Q6h ( 如臨床懷疑 adrenal insufficiency 雖無敗血性休克亦應給予 )

目標三： SvO2 ＞＝ 70未達目標三 , 若 Hb<10 輸 PRBC, 若 Hb>10 給予 Dobutamine

Q2h 檢查 lactate 直到 lactate < 2

達成目標 ICU 住院

一小時內建立上腔中心靜脈導管一小時內給予適當廣效性抗生素視情形給予插管及機械換氣血糖以連續 insulin 控制在 90~150考慮給予 H2 blocker for stress ulcer prophylaxis

1. 本流程參考性治療準則 , 若病患合併有 UGI bleeding, CHF, ESRD 等狀況 應以臨床情況決定輸液治療的量及速度

2. 抽血間隔 : Lactate Q2h , DM 病患 sugar Q2H, 3. Cirrhosis 病患應檢查 albumin

Preliminary Results• Period

– 2006 Jan ~ 2006 May

• Setting

– NTUH ED Critical Area

– Staffed by Visiting Staff / Chief Resident/ Physician assistant

– 9 Rooms with Monitor Devices

– 1 SCVO2 monitor

• Patients

– Randomly Selected patients with septic shock

– Patients with severe sepsis not included in this preliminary trial

Characteristics between Case and Control Groups

Case (N=11) Control (N=33) P value

Age 65.45 +/- 20.6 64.72 +/- 21.5 0.97

Sex (male %) 18.2 % 18.2 % 1.0

Comorbidity

(Charlson Score)

2.54 +/- 1.9 3.21+/- 3.3 0.53

SBP 80.2 +/- 8.35 83.8 +/- 4.97 0.19

Acute Renal Failure 3/11 (27.3% ) 17/33 (51.5% ) 0.29

Acute Respiratory Distress

4 (36.4%) 10 (30.3%) 0.72

Conscious disturbance 5 (45.5%) 9 (27.3%) 0.28

Primary Outcome

Log-Rank test: P=0.109

Days

EGDT group

Traditional group

Mortality: 9.1% vs. 45%

Survival Curve

Secondary Outcome

• Length of hospital stay ( alive )

– EGDT group: 20.1 +/- 25.9

– Traditional therapy: 27.4 +/- 22.9

(Non parametric test: P=0.22)

STOP Sepsis Bundle, Taiwan

資料來源：奇美醫學中心柳營院區加護醫學部侯清正主任

Pre-intervention period: yr 2005

Post-intervention period: yr 2006

過程面指標 Variable Pre-intervention

(N=46)Post-intervention

(N=66)P value

Lactate 執行地點的轉變 (%)

ER 7 68<0.001

ICU 67 29

On CVP 執行地點的轉變 (%)

ER 59 680.138

ICU 28 29

2 小時指標遵從率(%)

CVP 78 89 0.118

SVO2 50 82 <0.001

抗生素於 4 小時內給予 (%)

ER ICU

4672

6777

0.0330.515

6 小時指標遵從率(%)

CVP 8-12mmHgMAP >65mmHgSVO2 >70%3 項總達成率

70802622

79896450

0.2780.272

<0.0010.003

輸液及藥物的給予 輸液及藥物的給予 Variable Pre-intervention

(N=46)Post-intervention

(N=66)P-value

ER (cc) 841.3 ± 548.0 1657.8 ± 1481.1 <0.001

ICU (cc) 878.6 ± 700.5 1855.2 ± 1871.5 <0.001

Total (cc) 1719.9 ± 765.5 3513.0 ± 2349.2 <0.001

升壓劑ERICU

5798

6592

0.5180.369

Steroid (%) 62 69 0.51

Sugar (24 小時內 ) <150mg/dl

46 58 0.251

結果面指標 Variable(28-DAY)

Pre-intervention(N=46)

Post-intervention(N=66)

P value

呼吸器 free day 14.0±13.7 20.5±11.6 0.010

ICU free day 11.8±11.4 17.9±10.0 0.004

住院總 free day 7.5±9.2 11.8±8.9 0.013

死亡率 (%)severe sepsisseptic shockAll patients

503839

291620

0.5740.0310.032

住院總金額 186,568±227,561 107,565±84,736 0.029

敗血性休克 6 小時目標達成時間與其死亡率

P=0.005

9

26

42

0

20

40

60

6小時內達成 6小時後達成 都沒達成

%

Glycemic control protocol, KMUH• 資料來源：高雄醫學大學附設中和紀念醫院胸腔內科許超群醫師

Mortality and ICU LOS

¦~«×

87 88 89 90 91 92 93 94 95

¦Ê¤À¤ñ

0

10

20

30

40

50

Mortality RateICU LOS

Sugar Control Protocol, 三次大改版

ICU 專責主治醫師制

24-hours continuous feeding using an effective feeding protocol 是血糖能良好控制的重要基礎

轉出 ICU 前 , 改由 RI tid/sc 控制血糖 ,血糖值反而不穩定 .

有些病患在轉入 ICU 一段時間後血糖才開始高 , 故即使一開始的血糖值不高 ,所有的重症病患均應持續監控血糖 .

Infusion Insulin 之劑量變動極大 ,但血糖值均能在可接受範圍內

血糖值正常時仍需每日監控 ,每日一次

胰島素之使用隨血糖值調整 ,有時會斷斷續續使用

轉入 MICU 12 小時後 , 血糖獲得良好控制

24 h 後 , 血糖極少 > 200 mg/dL

Efficacy and Safety of Insulin Infusion Protocol

279 patients admitted to MICU during

Feb 1 to April 30, 2006

242 patients

222 patients

170 patients

116 patients54 patients

Readmission: 16 patientsMissing data: 19 patients

Oral feeding: 20 patients

ICU stay < 48 h: 52 patients

No insulin infusionduring ICU stay

With insulin infusionduring ICU stay

170 位病患轉入 MICU 時之血糖值分佈

116 位病患於 ICU 住院期間曾使用 insulin infusion therapy, 其中僅有 36.6% 於轉入 ICU 時即開始使用 insulin.

116 位病患於 ICU 住院期間曾使用 insulin infusion therapy, 其在 ICU 住院時 , 約六成的時間需使用 insulin.

ICU 住院 < 48 小時之病患僅有 28.8% 曾使用 insulin infusion therapy, ICU 住院 > 48 小時之病患則有 68.2% 曾使用 insulin infusion therapy,

Assessment of Adrenal function

NTUH Experiencecortisol 1

(<15) (n=16)cortisol 2

(15~34) (n=39)cortisol 3 (>34)

(n=19)P value

Cortisol (μg/dl) 8.583.69 22.025.81 62.0631.36

ACTH (pg/ml) 28.2532.94 39.6144.71 36.2924.39

28 days mortality No. (%)

2 (12.50%) 9 (23.08%) 12 (63.16%) 2vs3: 0.003

1vs3: 0.002

Inotropic agent n. Mean (SD)

1.06 (1.06) 0.85 (0.67) 1.42 (0.69) 2vs3: 0.003

Steroid supply No. (%)

13 (81.25%) 11 (28.21%) 7 (36.84%)

28 days mortality No. (%)

2 (15.38%) 5 (45.45%) 4 (57.14%)

NTUH experience

  Responder Nonresponder (n=16) (n=9)

28 days mortality 3 (31.3%) 1 (11.1%)

Taper vasopressor 10 (62.5%) 7 (77.7%)

Use steroid 8 7

28 days mortality 3 (37.5%) 0 (0%)

Taper vasopressor 5 (62.5%) 6 (85.7%)

Not use steroid 8 2

28 days mortality 0 (0%) 1 (50%)

Taper vasopressor 5 (62.5%) 1 (50%)

NTUH, adrenal function assessment

• Septic shock, vasopresser >24 h, MICU

Jun~Sep, 2003(n=98)

Apr~Jun, 2005(n=119)

ACTH/Cortisol level

59 (60.2%) 78 (65.5%)

ACTH test 25 (25.5%) 20 (16.8%)

Use of rhAPC, Taiwan2005 2006 2007

Jan 6 4 11

Feb 6 7 9

Mar 4 15 10

Apr 6 16 5

May 4 15 13

Jun 10 6

Jul 6 8

Aug 7 7

Sep 5 5

Oct 5 2

Nov 7 6

Dec 5 7

Total 71 98 48

6.5% absolutereduction in

mortality

0

35

20

15

10

5

30

25

Mor

talit

y(%

)

Historical (n=63)

drotrecogin alfa(activated) (n=1

7)

30%

23.5%

NTUH historical control and EVBR study: 28-day all-cause mortality

• High risk of death • APACHE II score 25, or• Sepsis-induced multiple organ failure, or• Septic shock, or• Sepsis induced acute respiratory distress syndrome

• Treatment should begin as soon as possible once been identified

• No absolute contraindication • Weigh relative contraindications

Recombinant human Activated Protein C

Grade B

Dellinger, et. al. Crit Care Med 2004, 32: 858-873.

Industry and guideline

(Eichacker et al. N Engl J Med 2006;355:1640)

Risk vs, Benefits in the real world

(Eichacker & Natanson. Intensive Care Med 2007;33:396)

Implementing Sepsis Guideline in Taiwan

• Education

• Practice module– “Bundled” approach– Quality improvement

indicator• Information technology

• Research

• Taiwan Guideline