SCIENTIFIC EVALUATION OF AUTOGRAFT ACL RECONSTRUCTION

1
Orthopaedic Interest Group www.mc.vanderbilt.edu/
medschool/OIG
2
How to Read the Literatureand Formulate
Hypothesis
Vanderbilt Medical Student Journal Club September
18, 2003

• Kurt P. Spindler, MD
• Professor Vice Chair, Orthopaedics
• Director, Vanderbilt Sports Medicine Ortho PCC
• Head Team Physician, Vanderbilt University

3
How is this Relevant?

• 1. Time efficient way to read literature
• 2. Teach YOU how to evaluate studies
• 3. Develop clinical practice solid ground
• 4. IMPROVE PATIENT CARE!
• 5. Stimulate better clinical research in future
• 6. Avoid wasting economic resources

4
Is There Evidence in Literature Supporting EBM
Approach for ORTHOPAEDICS?

• 1. JBJS-A, Jan 2003 Editorial (Heckman)
• Introducing levels of evidence
• Five levels
• Four study types
• 2. AJSM, 2002 Abstract Format (Reider)
• Background
• Hypothesis
• STUDY DESIGN list
• Methods / Results / Conclusion
• Clinical relevance

5
Goal TEACH

• Evidence-based framework review clinical
  literature
• How to use WORKSHEET !

6
Basics

• Title
• Author
• Reference
• HYPOTHESIS
• PRIMARY
• SECONDARY

7
Type of Study

• Treatment
• Diagnosis
• Screening
• Prognosis
• Causation

8
Type StudyPreferred Design

• 1 HYPOTHESIS OR
  PREFERRED
• TOPIC RESEARCH EXAMPLES RESEARCH
  DESIGN
• Treatment drug, prevention, RCT
• surg
• Diagnosis dx test Cross-sect
  survey
• Screening value of test Cross-sect survey
• Prognosis disease, injury, Longitudinal
  cohort
• condition
• Causation exposure to . . . Cohort or
  case-control

9
Common Sports Medicine Clinical Questions

• 1. Treatment ACL Graft Ham vs PT
• Operative Approach Endo vs Rear Entry
• Meniscus Repair Inside-Out Sut vs All In
• Grade IV Defect Microfx vs Autologous
• 2. Diagnosis SLAP PE vs MRI
• Meniscus tear PE vs MRI
• ACL tr Lachman vs Ant Draw
• 3. Prognosis Natural hx ACL tear Identify risk
  factors
• Risk OA ACL recon Function and/or OA
• Partial meniscectomy Function and/or OA
• stable knee

10
Traditional Hierarchy ofClinical Treatment
Studies

• 1. RCT (randomized controlled trials) only
  computer or random table acceptable
• 2. Cohort two or more groups selected basis
  differences exposure to agent and f/u
• 3. Case control pts particular
  disease/condition identified matched control
• 4. Cross-sectional data collected single
  timepoint
• 5. Case reports/series medical hxs one or more
  patients with condition/tx reported on

11
Methods

• CONTROL or COMPARISON GROUP in treatment study?
• Control for major known variables that could BIAS
  result?
• Identify GOLD STANDARD in diagnostic study?
• Prospective or retrospective?

12
Why do Treatment Studies Need Control Group?

• 1. Basics Scientific Method!
• 2. If no control group tx is same, better, or
  worse than what?
• 3. Quality of control group one measure of
  validity of results
• 4. Unfortunately majority orthopaedic literature
  lack control group -- case series

13
Data Collection

• 1. Prospective study -- data
• collection planned in advance
• 2. Retrospective study -- review data
• normally collected (examples
• include case control or chart reviews
• 3. Prospective design better control
• confounding variables

14
How to Identify Bias

• Study BIAS
  Example
• Allocation groups Selection Fail randomize
• Intervention Performance Fail control
  confounding variables
• Follow-up Exclusion Not uniform or
• (or Transfer) inadequate (lt70)
• Outcomes Detection Dissimilar evaluation
• independent examiner?
• Validated question-
• naire?

15
Definition of Bias

• 1. SELECTION or SUSCEPTIBILITY difference in
  comparison groups secondary to incomplete
  randomization
• 2. PERFORMANCE differences in care provided
  apart from intervention being evaluated
• 3. EXCLUSION or TRANSFER differences in
  withdrawal from trial
• 4. DETECTION different evaluation for outcomes
  best independent examiner or blinding examiner or
  validated outcome questionnaire self-administered

16
Sports Medicine Examples Bias

• 1. Selection
• ACL tr pt self-select OR vs Nonop tx
  evaluate OA
• Soccer teams self-select ACL inj prevention
  training,
• then report difference incidence ACL tr
• 2. Performance
• Report outcome of meniscal allograft or
  autologous
• chondrocytes fail control concomitant ACL
  recon or HTO!
• 3. Exclusion or Transfer
• Report conclusions based lt70 f/u outcome
  variable

17
Methods

• 1. Did author demonstrate
• demographics (age, gender,
• etc) equal in groups?
• 2. Length of f/u
• Min ______
• Avg ______

18
Basic Statistics

• A. Continuous (ht, age) ? normality tested OK
• Yes ? parametric tests
• No ? nonparametric tests
• B. Discrete (Yes, No, ) ? nonparametric
• C. See handout for Common Statistical Tests
• D. Stats are acceptable Y / N / unknown
• E. Stats consultation requested Y / N /
• Who __________________________

19
Statistical and Clinical Significance Outcomes

• Absolute
  If ns power ( ) Clinically
• Outcome/Result Difference P for (
  ) diff significant
• a.
• b.
• c.
• d.

20
Statistical vs Clinical Significance

• 1. Primary hypothesis determines sample size by a
  clinically meaningful difference in single
  outcome variable chosen.
• 2. Example -- ACL graft choice detect 1 mm side
  to side difference, n ? 60-70. If differences
  are significant but could be 1 vs 2 vs 3 mm, what
  would change your practice?
• 3. Thus clinical significance both common but
  individual!
• 4. Power set 80 to detect clinically
  meaningful stat difference.
• 5. Other examples . . . .

21
Summary

• A. If no comparison group or control group in
  TREATMENT STUDY ? READ FOR INFORMATION ONLY --
  no evidence-
• based reason or data to change practice pattern
• B. If no comparison of groups that equal
  PreTREATMENT ? RESULTS MAY NOT BE CAUSED BY
  TREATMENT BUT BY
• DIFFERENCES IN GROUPS
• C. If INTERVENTION contains additional proven or
  suspected variables other than 1 hypothesis
  indicating Performance Bias, results NOT
  SPECIFIC TO INTERVENTION ?
• RESULTS UNCLEAR SIGNIFICANCE TO TREATMENT

22
Summary

• D. Inadequate follow-up (lt70) indicates
  EXCLUSION or TRANSFER BIAS ? RESULTS COULD
  CHANGE IF
• ADEQUATE FOLLOW-UP
• E. Are STATS acceptable?
• F. Are absolute values which are statistically
  significant also CLINICALLY SIGNIFICANT?

23
Conclusion

• If A ? F acceptable
• If your patient population similar to study
• If you can perform technique/treatment
• CHANGE YOUR PRACTICE ACCORDINGLY!

24
Requirements Research Study Design

• 1. HYPOTHESIS (Ho)
• Measured outcome strength, stability,
  function
• Comparison or control group
• 2. AIMS
• Specific statistical measurements between
  groups at defined intervals (KT1000 _at_ 2 yr)
• 3. METHODS
• 4. RESULTS and STATISTICS
• 5. BUDGET and TIMELINE
• 6. Discussion strengths, weaknesses, alternative
  approaches

25
Steps Research Design

• First Read the literature EBM
• Second Develop working hypothesis
• Third Write Aims/Methods/Results, etc.
• Fourth Establish Budget/Timelines/Team
• Finally Initiate study!

26
Pearls

• 1. Develop ideas methods, results, statistics
  from best EBM in literature review
• 2. Retrospective review your cases!
• Establish sample size
• Timelines to complete
• Generate methods
• 3. Consult statistician BEFORE begin study!

27
Examples EBM ReviewsClinical Questions Sports
Medicine

• Shoulder
• 1. How do you treat Type 3 AC sprains?
• 2. Which treatment do you prefer midshaft
  clavicle fracture?

28
EBM Type III AC Sprains

• 1. RCT (randomized controlled trials) Nonop vs
  Op
• Bannister G JBJS-B 1989 (n84, 13 mo f/u)
• Larsen E JBJS-A 1986 (n60, 48 mo f/u)
• Results clinical outcomes equal!
• Rehabilitation faster nonoperative
• 2. Meta-Analysis 24 Articles Tx
• Phillips A CORR 1998
• AROM and strength same
• No signif benefit surg outcome analysis
• 3. Strength Evaluation Nonop Tx
• Tibone J AJSM 1992 (retrospective 4.5 yrs)
• No strength deficits

29
EBM Clavicle Fx

• 1. Nowak J Injury 2000 (Sweden)
• 101 fx/100,000
• Male female 21 (71 males, 30 females)
• 75 middle third
• 95 healed without problem
• 2. Anderson K Acta Orthop Scand 1987
• RCT Fig 8 vs Sling n 61
• Clinical and XR EQUAL
• 100 union
• Initial displacement -- UNCHANGED!

30
EBM Clavicle Fx

• 1. Nordqvist A J Orthop Trauma 1998
• Avg 17 yo f/u 225 nonop mid clav Fx
• Sx 82 (185) none, 17 (39) moderate
• Union 97 (218) healed, 3 (7) nonunion
• Position 68 (153) nl, 24 (53) malunited
  (displaced)
• FYI -- 40 malunions, 3 nu clinically rated good

31
Why ACL Reconstruction 2003?

• 1. Primary repair?
• 2. Augmentation?
• 3. Meniscus tears?
• 4. Higher function?
• 5. Prevent arthritis?

32
Why ACL RECONSTRUCTION 2003Evidence Based
Literature

• 1. Primary repair no better nonoperative tx
• Sandberg RCTs. JBJS-A, 1987
• 2. PT ACL recon better than repair and
  augmentation
• Engebretsen. AJSM, 1990
• Grontvedt. JBJS-A, 1996
• 3. Reconstruction significant decrease meniscus
  tear rate from 27 (nonoperative) to 3 (ACL
  recon)
• Anderson. JBJS-A, 1989

33
ACL Reconstruction Techniques

• 1. Which autograft do you choose?
• Reasons . . .
• 2. Approach endoscopic (single- incision) or
  rear entry (two- incision)?
• Reasons . . .

34
Summary Six RCTs in 2000

• 1. Function No difference
• 2. ROM ? PT 1.5 - 3.0 in 33
• 3. KT1000 PT more stable 1-3.4 mm (50)
• 4. Ant or PF pain No difference 83
• 5. Kneeling pain PT greater 3/3 (100)
• 6. Isokinetic Ham weaker 7-11 (50)

35
Operative Technique

• 1. Arthroscopic vs Mini Open
• (patella stays in trochlea)
• -- Raab 93 Cameron 95 -- No difference !
• 2. Endoscopic (single inc) vs
• Rear-Entry (two inc)
• -- Brandsson 99, Reat 97, Garfinkel 93,
• ONeill 96, Gerich 97
• No major differences trend better rear-entry

36
Decision Making Meniscus Surgery

• 1. Op vs Nonop treatment?
• 2. Repair vs Excision?
• 3. Partial vs Complete?
• 4. Technique Repair?
• Inside-out sutures
• All-inside arrows

37
EBM Meniscus

• 1. RCTs (Cochrane database 2001)
• a. Surgery vs nonoperative -- none
• b. Repair vs excision -- none
• c. Partial vs complete equal long-term XRs
• 2. RCTs and Prospective Comparative?
• Inside-Out Suture EQUALS All-Inside Arrows
• a. Albrecht-Olsen.? Kn Surg Spt Traum Arth 1999
• b. Kirkley? in preparation
• c. Spindler? AJSM, in press

CAVEAT All NWB five weeks!!
38
Summary First Steps Research

• 1. Review literature EBM
• 2. Generate hypothesis
• 3. Construct preliminary aims
• 4. Retrospective review clinical cases sample
  size
• CAVEAT Basic science requires lab usually
  university

39
Thank you
40
References

• Wright JG JBJS-Am 2000
• Hurwitz SR JBJS-Am 2000
• McLeod RS Surgery 1996
• Greenhalgh T How to Read a Paper. Br Med J
  2001
• Lang TA and Secic M How to Report Statistics in
  Medicine. ACP 1997
• Spindler K, Johnson R, Reider B ICL AOSSM 2002

41
Orthopaedic Interest Group www.mc.vanderbilt.edu/
medschool/OIG