The Cafeteria Test

1
The Cafeteria Test

• Good policy and a good research protocol can
  be explained to your surgeon colleague in the
  noisy hospital cafeteria, between two beepers
  beeping, and over a cup of coffee

2
Pregnancy Fetotoxicity- NeonatologyWhere Myths
Persist

• For instance  Efavirenz is teratogenic 

3
Efavirenz and pregnancy

• Teratogenicity/Developmental Toxicity Animal
  StudiesMalformations were observed in three of
  20 infants born to pregnant cynomolgus monkeys
  receiving efavirenz from gestational days 20 to
  150 at a dose of 30 mg/kg twice daily (resulting
  in plasma concentrations comparable to systemic
  human therapeutic exposure). The malformations
  included anencephaly and unilateral anophthalmia
  in one microphthalmia in another and cleft
  palate in the third. Primate teratogenicity
  studies have not been conducted for delavirdine
  or nevirapine.

4
Antiretroviral Pregnancy Register

• Overall, the prevalence of birth defects in 2427
  live births was 3.0 percent. The CI of 2.4-3.8
  includes the prevalence in patients not exposed
  to ARVs (2.7) indicating no significant
  difference
• For abc, lpv, efv, nvp, tfv, rtv, sufficient
  births have accrued to exclude a two-fold
  increase, and for 3-TC and zdv, a 1.5-fold
  increase can be excluded
• Regarding efv 7 defects in 281 first trimester
  exposures, i.e. 2.5 (confidence interval 1.0
  5.1), none of which resembled the CNS lesions in
  monkeys
• 3 cases of myelomeningocoele reported in the
  literature

www.apregistry.com, accessed December 15, 2007.
Last interim report June 2007
5
 Ciprofloxacin is contra-indicated 

• Bone problems in puppies, not reproduced in other
  species
• In dogs, some quinolones do, others dont. The
  most powerful of all nalidixic acid, a precursor
  drug used widely in the 1970ies
• Follow-up of these children, now adults No bone
  problems.
• Inadvertent exposure Nothing
• Inspite of this, the pediatric c.-i. to cipro
  still exists

6
Spurious  contra-indications 

• A luxury for the rich
• A burden for the poor

7
SMART (Strategies for Manage-ment of
anti-retroviral therapies)
Patients with gt 350 CD4 cells, mostly on
treatment, willing to be randomized to

1. Drug conservation (DC or STI) arm Treatment
   stopped if CD4 gt 350, started at lt 250
2. Virologic suppression (VS or CT) arm Keep VL lt
   50 at all times

8
SMART As Planned

• NIH, CPCRA
• 6000 patients, planned follow-up of 5 to 9 years
• 900 endpoints expected in 2009, with power to
  detect difference of approximately 20 between
  arms

9
SMART As It Turned out

• Recruitment suspended on January 11, 2006, after
  5472 patients had been recruited, and 164
  endpoints observed
• Excess of endpoints in the STI (DC) arm

10
Primary endpoint of OI/death
STI group CT group
Curves diverge after 4 months
11
SMART Confirmed Events

• Event STI CT RR P
• N /100py N /100 py
• AIDS/death 118 3.3 46 1.3 2.63 lt10-4
• Death 55 1.5 30 0.8 1.84 0.007

12
SMART Type of AIDS Events

• Event STI CT
• Esophageal candidiasis 24 7
• PCP 8 2
• Recurrent bacterial pneumonia 7 2
• Kaposis 6 1
• Persistent herpes simplex 3 2
• Lymphoma 4 1
• Disseminated zoster 3 1
• TB 2 2
• All others 7 2
• Patients with any event 54 17

13
Few Deaths are AIDS-Related

• Type STI CT
• Cancer (excluding AIDS-related) 11 5
• Cardiovascular 7 4
• Substance abuse 3 5
• Accident, suicide, violence 3 4
• AIDS-related opp. disease 4 3
• Other infections 3 1
• Various other causes 9 5
• Unknown 15 3
• Total 55 30

14
SMART (Supposedly) Drug-Related AEs Are Also
Worse in STI Group

• Event STI CT RR P
• N /100py N /100 py
• MI, major CAD,stroke, RF, LF 63 1.8 38 1.0 1.68
  0.01

15
What if Treatment Was Re-started at Higher CD4
Counts ?

• STI CT RR Delta NNP
• Last CD4lt 250 7.6 10.5 0.72 250-349 4.2 2.3 1.83
  1.9 32 350-499 2.7 1.4 1.93 1.3 47gt499 2.0 1.2
  1.67 0.8 76
• Number of patients with endpoints (AIDS or
  death) per 100 patient-years of follow-up.
• NNP number of patient years of treatment
  needed to prevent 1 event, considering that
  patients in the STI group were treated during 33
  percent of days, compared to 94 percent in the CT
  group

16
The Cost-Effectiveness of HAART

• Numbers from Switzerland
• Before HAART, approx. 800 AIDS/Deaths per year
• After HAART, approximately 100
• Approximately 5000 patients are being treated
• 5000/700 or approximately 7 years worth of
  treatment to prevent one event

17
I would like to stop. Can I do so safely?
18

• STI CT RR Delta NNP
• Last CD4lt 250 7.6 10.5 0.72 250-349 4.2 2.3 1.83
  1.9 32 350-499 2.7 1.4 1.93 1.3 47gt499 2.0 1.2
  1.67 0.8 76

The Future of STIs (1) At higher CD4 counts
500 to stop, 400 to start again
19
Summary Harmful to Stop ?

• SMART participants started HAART at CD4 counts of
  approx. 200 (median, large range)
• Intermittent treatment is inferior to continuous
  treatment
• Holds true at all CD4 counts
• At high CD4 counts, NNP exceeds 50 years of
  treatment to prevent one event

SMART does not show that intermittent treatment
is inferior to no treatment
20
When To Stop ?(in women who started ART for PMCT)

• Women who had an indication for treatment (CD4 lt
  350) SMART says Dont stop !
• Women who did not have an indication Treat for
  one year, then stop.
• If treatment started without CD4 count measured
  Stop after 1 year. Measure CD4 counts after
  three months, treat if lt 350.