Ventilatory support in cancer patients

Transcription

1 Ventilatory support in cancer patients D. Benoit, MD, PhD Department of Intensive Care Medicine Ghent University Hospital

2

3

4

5 La ventilation non-invasive (VNI) réduit le risque d intubation et la mortalité chez le cancéreux avec une IR Oui Non

6 La ventilation non-invasive (VNI) devrait être utiliser d emblée chez le sujet cancéreux avec une IR ne présentant pas de contre-indications classiques Oui Non

7 Invasive or. non-invasive mechanical ventilation?

8 outside the clearcut indications for NIV : cute pulmonary edema cute COPD excacerbation

9 Invasive or. non-invasive mechanical ventilation?

10 Mortality in ventilated cancer patients uthor, year Total n. of patients Solid tumors Haematol. Malign. Hospital mort. (%) Schuster, Ewer, Peters, Brunet, Sculier, / 70 Shapira, / 76 Epner, Groeger, / 84

11 Reluctance for ICU admission «Patients who require mechanical ventilation longer than 24h are likely to die in the hospital (94% mortality in the largest study). Prognosis should be reassessed at frequent interval with particular attention to the developement of MOF» Uptodate 2010

12 Mortality in ventilated cancer patients uthor, year Total n. of patients Solid tumors Haematol. Malign. Hospital mort. (%) Kress, Massion, Benoit, Maschmeyer, Depuydt, zoulay, Soares, / 68

13 Duration of mechanical ventilation 100% mortality if duration of MV > 5 days Schuster, m J Med 1983 > 6 days Ewer, JM 1986 > 4 days* Denardo, Crit Care Med 1989 > 7 days* Torrecilla, Crit Care Med 1988 > 15 days* Huaringa, Crit Care Med 2000 No restriction > most experts in the field * allo-bmt recipients

14 Duration of mechanical ventilation P =0.94 Benoit, unpublished data 2005 alive dead in hospital mortality 0,00 25,00 50,00 75,00 days of ventilation n=66 n=172

15 Invasive or. non-invasive mechanical ventilation?

16 -30%

17 -30%

18 -60%

19

20

21

22 Matched-cohort analysis (1:2) Variables NIV (n=26) IMV (n=52) P-value age 44 (35-63) 58 (41-69) 0.06 ML 9 (35 %) 13 (25 %) 0.64 ctive disease 7 (27 %) 12 (23 %) 0.78 Leukopenia 6 (23 %) 9 (17 %) 0.55 SPS II Pa02 / Fi02 72 (56-86) 147 (78-201) <0.001 PEEP level 5 (5-8) 5 (5-10) 0.17 Vasopressor need 7 (27) 25 (48) Dialysis 4 (15) 18 (35) 0.08 Hospital mortality 17 (65 %) 34 (65 %) 0.99 Conditional logistic regression analysis: NIV OR 1.08; 95% CI, ) Depuydt, Chest 2004

23

24

25 Matched-cohort analysis (1:2) Variables NIV (n=26) IMV (n=52) P-value age 44 (35-63) 58 (41-69) 0.06 ML 9 (35 %) 13 (25 %) 0.64 ctive disease 7 (27 %) 12 (23 %) 0.78 Leukopenia 6 (23 %) 9 (17 %) 0.55 Too sick too late..? SPS II Pa02 / Fi02 72 (56-86) 147 (78-201) <0.001 PEEP level 5 (5-8) 5 (5-10) 0.17 Vasopressor need 7 (27) 25 (48) Dialysis 4 (15) 18 (35) 0.08 Hospital mortality 17 (65 %) 34 (65 %) 0.99 Conditional logistic regression analysis: NIV OR 1.08; 95% CI, ) Depuydt, Chest 2004

26 Variables live (n=48) Died (n=118) P-value Male gender 23 (48 %) 77 (65 %) ML 8 (17 %) 36 (30 %) 0.08 ctive disease 8 (17 %) 42 (35 %) Leukopenia 7 (15 %) 38 (32 %) 0.02 SPS II 50 ±14 62 ±19 <0.001 Intubation < 24 h 36 (75 %) 84 (72 %) 0.70 Non-invasive ventilation 9 (19 %) 17 (14 %) 0.25 PEEP level 5 (5-8) 5 (5-8) 0.76 Pa02 / Fi (67-274) 112 (76-206) 0.58 Bacteremia 13 (27 %) 16 (14 %) 0.04 * Depuydt, Chest 2004 Soares, Crit care Med /166 (15.6 %) 40/463 (9.0 %)

27

28 Non-invasive ventilation Start «early»

29 NIV or IMV? 1) Can the absolute risk reduction in mortality of 30% in the RCT s be attributed to the ventilation mode only? 2) Explanation why non invasive mechanical ventilation is associated with a lower mortality?

30 NIV or IMV? 1) Can the absolute risk reduction in mortality of 30% in the RCTs be attributed to the ventilation mode only? Very small groups in a heterogeneous population High probability of imbalances in baseline charactheristics Per definition not well randomized.

31 Study uthor Journal N Mortality Goal-directe therapy in severe sepsis and septic shock ctivated protein C Prowess trial (apc) RDS network study low tidal volume Low dose corticosteroids in septic shock Rivers N NEJM % vs % (p=0.009) Bernard NEJM % vs % (p=0.005) NIH NEJM % vs % (p=0.007) nnane JM % vs. 63 % (p=0.02) Tight glycemia control in the ICU Van den Berghe NEJM % vs. 8 % (p=0.04)

32 Ventilatory mode Mortality Cause(s) of RF

33 Ventilatory mode Mortality Cause(s) of RF

34 «Successful» NIV trial = surrogate marker for rapidly reversible and therefore per definition succesfull treatment of the underlying cause of RF?

35 Cause of RF and mortality 0% 20% I.Treatable and often rapidly reversible admission diagnosis cute pulmonary edema COPD excacerbation II. Treatable and potentially reversible admission diagnosis 50% Documented or clinically suspected bacterial infection Pneumocystis jerovicii pneumonia III. Difficult to treat and/or often slowly reversible admission diagnosis 80% Invasion of hematological malignancy Viral pneumonia Invasive pulmonary aspergillosis 90% IV. Difficult to treat and often irreversible admission diagnosis No diagnosis, solid tumor invasion, idiopathic fibrosis

36 Non-invasive ventilation Immunocompromised patients with respiratory failure (n=52) NIV (n=26) Standard (n=26) P-value Final diagnosis 17 (65%) 11 (42%) 0.09 Intubation ICU mortality Hospital mortality 12 (46%) 17 (77%) 10 (38%) 18 (69%) 13 (50%) 21 (81%) Hilbert, N Engl J Med 2001

37 Non-invasive ventilation Immunocompromised patients with respiratory failure (n=52) Final diagnosis (n=31) No final diagnosis (n=21) P-value Bacteria 15 4 (26%) - Bacteria + fungi 3 2 (67%) - Fungi 7 2 (28%) - CMV / PCP 6 4 (67%) - ICU mortality 12 (38%) 16 (76%) /15 (27%) 24/37 (65%) /18 (30%) 22/34 (65%) Hilbert, N Engl J Med

38

39

40

41

42

43

44

45 Causes??

46 NIV or IMV? 1) Can the absolute risk reduction in mortality of 30% in the RCT s be attributed to the ventilation mode only? Very small groups in a heterogeneous population No!! Imbalances in baseline charactheristics Discrepancies with results in general ICU population Only effective in well-known indications Weaning protocol / short-acting drugs in IMV group?

47 NIV or IMV? 1) Can the absolute risk reduction in mortality of 30% in the RCT s be attributed to the ventilation mode only? No!! 2) Explanation why non invasive mechanical ventilation is associated with a lower mortality?

48

49

50 How many patients died directly because of nosocomial infection.? Did nosocomial infection indirectly increased mortality by increasing the duration of ventilation? Time of onset.?

51

52

53 NIV or IMV? 1) Can the absolute risk reduction in mortality of 30% in the RCT s be attributed to the ventilation mode only? No!! 2) Explanation why non invasive mechanical ventilation is associated with a lower mortality??

54

55 Ventilatory mode Mortality Cause(s) of RF!

56

57 dmission diagnosis and mortality 0% 20% I.Treatable and often rapidly reversible admission diagnosis cute pulmonary edema COPD excacerbation II. Treatable and potentially reversible admission diagnosis 50% Documented or clinically suspected bacterial infection Pneumocystis jerovicii pneumonia III. Difficult to treat and/or often slowly reversible admission diagnosis 80% Invasion of hematological malignancy Viral pneumonia Invasive pulmonary aspergillosis 90% IV. Difficult to treat and often irreversible admission diagnosis No diagnosis, solid tumor invasion, idiopathic fibrosis

58 Non-invasive ventilation Start «early» for a potentially rapidly reversible complication

59 NIV or IMV? NIV trial: for how long?

60

61

62

63

64 Non-invasive ventilation Start «early» for a potentially rapidly reversible complication for a couple of hours! = % of the population..!

65

66 Conclusion

67 La ventilation non-invasive (VNI) réduit le risque d intubation et la mortalité chez le cancereux avec une IR Oui Non

68 La ventilation non-invasive (VNI) devrait être utiliser d emblé chez le sujet cancereux avec une IR ne présentant pas de contre-indications classiques Oui Non

69 Medicine is learned by the bedside and not in the classroom. Let not your conceptions of disease come from words heard in the lecture room or read from the book. See, and then reason, and compare and control. But see first Sir William Osler 1904