The problem with a (entirely) structural based approach to low back pain….

Is there may actually be no problem.

This sounds like channeling the girl from the Matrix proclaiming there is no spoon, but it’s not far off. Before a finding can be called a problem the base rate of the finding needs to be established. Or put another way, what is the incidence of that finding in the normal, asymptomatic population? Too often radiology reports are read as death sentences to patients when in fact it is natural progression of aging or highly present in the general, pain free population. Recently my intern and I set out to see what those rates are in hopes of further demystifying issues to both patients and practitioners. Here, in wonderful blog format, are our findings.

Red Flags

Henscke 2013 would say there is no sufficient red flag to screen for malignancy, but history of prior cancer was the most likely (I know you’re all surprised by that). Dowie 2013, when screening for fracture, states there is no strong association but old age, corticosteroid use, severe trauma, and presence of contusion should all increase the probability. The normal bowel/bladder dysfunction and progressive neuro symptoms or upper motor neuron signs are still included.

The point of this being many of the pathologies don’t really hold up and there are large flaws within the structuralist model. Koes 2006 would argue that up to 90% of low back pain is nonspecific, or said better…we don’t really know what is causing the pain. If established anatomical abnormalities do not really correlate, then it is even more difficult to state that non established things like innominate rotation (even though they don’t rotate), tight hip flexors (which we first have to decide what is “tight”) or trigger points (which still have no evidence) are correlated with their onset.

  • Herniated discs

We might as well start with the most common finding. The classic study here is Brinjiki 2014 which showed the average incidence in the asymptomatic population going from 30% in 20 year olds to 84% in 80 year olds. Baranto in 2009 looked at a cohort of athletes for 15 years and showed a high prevalence of disc degeneration but no correlation with low back pain. Plus, odds are you’ll get better. Regression to the mean is a wonderful thing when it comes to discs; the bigger the bulge, the more likely it is to heal. Macki 2014 also shows a high likelihood of disc healing over time. That being said, if a patient has bowel/bladder dysfunction or progressive neurological symptoms it still heavily warrants further evaluation.

  • Degenerative Disc Disease

Waris 2007 states that if you have DDD early in life it will probably advance but it is not related to symptomology. Brinjiki 2014 would state that DDD is present in 37% of 20 year olds and up to 96% of 80 year olds. This appears more a natural progression of aging but still if almost 40 percent of 20-year olds have the finding it is bordering on being more abnormal if you do not have some form of degenerative disc disease. Maybe the people with the pristine MRI’s are the weirdos after all. This is also a good time to mention that there can be a high variability in interpretation of results within the same person. Herzog 2016 report the same patient, when undergoing 10 MRIs within 3 weeks with different interpretations by 10 radiologists, only yielded a 56% sensitivity.

  • Facet Arthropathy

Kalichman in 2008 found:

“There is a high prevalence of FJ OA in the community. Prevalence of FJ OA increases with age with the highest prevalence at the L4-L5 spinal level. At low spinal levels women have a higher prevalence of lumbar FJ OA than men. In the present study, we failed to find an association between FJ OA, identified by multidetector CT, at any spinal level and LBP in a community-based study population.”

  • Spondylolisthesis

A quick review gives us Andrade 2015 showing no correlation between spondy and long term back pain. This is further corroborated by another study by Kalichman et al from 2009.  There is also the 45 year follow-up study by Beutler 2003 that shows no association between spondy and low back pain. The diagnosis certainly can raise some patient anxiety as essentially they “broke their back.” This research certainly gives cause to the fact that these athletes have a high likelihood of symptom resolution without advanced intervention.

  • Schmorl’s Nodes

The general consensus here as it relates to being a pain generator was a resounding — meh. If we look at Williams 2007  we see up to 30% of the female population presents with them. There was an association between their presence and LBP but it was nonsignificant when accounting for disc disease.  This is further corroborated by Kyere 2012  and Yin 2015.

  • Modic Changes

This one gets a little sticky. According the Mok 2016 study there was a small association but there is a higher correlation with being a smoker or being obese. What is interesting here is the grading of Modic changes with 1 being neovascularization, 2-fatty infiltrate, and 3-sclerotic tissue. Some of the grade one pathobiology sounds analogous to what is in tendinopathies. Teichtahl 2016 would further add to the disadvantageous nature of being obese as it relates to Modic changes, especially type 2.

Zhang 2008 would state that they are normal in the asymptomatic population but the usual “more research is needed” caveat applies. Jensen 2008, in their review, report a higher prevalence of changes in the symptomatic population. That being said, Hutton 2011 would imply that they can be reversed. We may get a “treat the donut not the hole” paper for Modic changes down the road. Heavy slow resistance training may just end up being a silver bullet for multiple ailments.

  • Scoliosis

Scoliosis was the wild card in the mix. Going into this we thought we would find some even decent papers related to scoliosis in the asymptomatic population but they proved harder to find than we thought. Instead we found worst case scenarios with Weinstein 2003 which had a 50-year follow-up looking at untreated scoliosis. They concluded:

“Untreated adults with LIS (Lumbar Idiopathic Scoliosis) are productive and functional at a high level at 50-year follow-up. Untreated LIS causes little physical impairment other than back pain and cosmetic concerns.”

The reason I say worst case is the Cobb angles at skeletal maturity were all greater than 35 degrees at recruitment and at 50-year follow-up were all greater than 49 degrees. It seems to undermine the severity of a small scoliosis being correlated with symptoms. Most research advocates that any curve less than 30 degrees should be treated conservatively. Conservatively being bracing and observation. Looking at the most recent Cochrane Reviews for bracing and exercise for scoliosis leaves much to be desired.

  • Stenosis

Burgstallar 2016 would state there is little correlation between degree of lumbar stenosis and pain. This may contribute to the reason that Forth, in NEJM earlier this year, did not find any benefit to surgery for stenosis. Or, to take it one step further, the recent Cochrane Review on surgical options for stenosis concluded there was no good evidence for surgery.

To introduce some points on psychology that will be expanded upon later, McKillop 2014 would advocate that depression plays a key role in outcomes after surgery for stenosis. This is further supported by Sinikallo 2009 who showed a correlation between psychological disposition and outcomes. Once again it would appear taking a structural predisposition to diagnosis and treatment may not be the most efficacious when it comes to spine care.

  • Muscle Size/Fatty Infiltrate

This one might come as a surprise to a few people. Contrary to the belief gaining a lot of traction in the fitness community, being stronger does not inherently lead to having a pain free back. Teichtahl 2016 correlates DDD, Modic changes, and paraspinal fatty infiltrate to form what the authors call a “whole organ pathology.” In a paper from 2015 the same authors correlate fatty infiltrate, but not muscle cross sectional area with LBP and disability. Kalichman 2016  would state that fatty infiltrate is associated with the onset of facet joint OA but once again, no significant correlation with pain. Now, here’s where it gets fun….

In a systematic review by Suri in 2015 they conclude

“Few lumbar muscle characteristics have limited evidence for an association with future LBP and physical performance outcomes, and the vast majority have limited evidence for having no association with such outcomes.”

If it truly were just raw strength that leads to a healthy back none of us who lift heavy weights should ever have back issues. It may not be a question of strength but more of a question of “are you strong enough to do what you are trying to do?” To further muddy the water, the 2011 study by Niemelainen  states “Substantial asymmetry in paraspinal muscle cross-sectional area in healthy adults questions its value as a marker of low back pain and pathology.” This would seem to place a big knock against the striving for symmetry camp. An interesting study alluding to the asymmetrical nature of some sports would also state that an asymmetry can prove advantageous at times.

So What Does Matter….

  • Decrease Patient Fear and Anxiety

Carragee, all the way back in 2005, showed that there was little correlation between structural findings and low back pain but psychosocial variables strongly predicted both short and long term disability.  Wertli in 2014  in relation to fear avoidance beliefs concluded:

“Evidence suggests that FABs are associated with poor treatment outcome in patients with LBP of less than 6 months, and thus early treatment, including interventions to reduce FABs, may avoid delayed recovery and chronicity. Patients with high FABs are more likely to improve when FABs are addressed in treatments than when these beliefs are ignored, and treatment strategies should be modified if FABs are present.”

In relation to pain catastrophizing Wertli, in another 2014 paper, concluded:

“There is some evidence that catastrophizing as a coping strategy might lead to delayed recovery. The influence of catastrophizing in patients with LBP is not fully established and should be further investigated. Of particular importance is the establishment of cutoff levels for identifying patients at risk.”

According to Sorensen, even in the back healthy population there is a correlation between fear, anxiety, and pain.

This is interrelated to the above point, but also showing to be integral to outcomes. There is something to be said for Ryan Holiday’s quote “The obstacle is the way.” Pain and structural abnormalities can be framed as either a threat/something that will hinder progress or as a challenge from which to learn. We would have never learned how to build bridges if there were no spans that needed crossing. The question is now, with treatments that lack efficacy and diagnoses with little correlation to symptomology are we creating chasms where none really exist?

  • Be Mindful of Classical Conditioning

This one I’m adding even though classical conditioning has been shown to work (see Peerdeman). If we set expectations and condition patients that a modality is what cures them, then we never really help patients achieve independence. Sometimes we get so caught up trying to set ourselves apart as clinicians that we forget it is ultimately about the patient achieving their goals. It’s convenient to use a technique that could elicit a short-term conditioning reduction in pain, but circling back to the patient expectations component, we need to guide them around the obstacle, not have them expect us to move it. We also have to be aware of the effect of classical conditioning on ourselves. If we’re using a modality that meets our expectations, we need to make sure that we’re not the ones salivating when the bell rings, hoping for a result that may not be there. Pavlov’s dogs would most likely tell you the bell brings the food every time.

  • The Mind is Primary

Perception increasingly rules when it comes to pain. Christopher Hitchens, when faced with cancer, spoke of the metaphor of fighting the disease. The problem he realized was the cancer, the disease, was his body and he also was his body so he was ultimately fighting himself. When it comes to pain we often take the easy explanation of it being “in your head.” This presents the issue that you are your head. In fact, a patient is the only one really in their head. The patient’s pain is real and we should be mindful to acknowledge that. If you want to have your mind blown on this topic I suggest listening to this you are not so smart podcast on the nature of reality. Or you should go here for the really long; want to read version.

  • Now For the Body….

It would appear all exercise works, but quite possibly some better than others. Searle 2015 would state that resistance exercise is more effective. Saragiatto 2016 would say motor control exercises can have some positive effect. Franke 2015 would say muscle energy techniques are out. They may fall more in to the category of manual therapy (which is another topic unto itself) but it is the lowest dose of exercise. But — Nijs 2012 would still say that all exercise needs to be individually tailored to account for the psychological disposition of the patient.

To Close

Both practitioners and patients are often looking for the answer to “what is wrong” but this may often be the wrong question. The better question may be “is anything wrong?” We know that the vast majority of low back pain is nonspecific and 70% to 90% of acute low back episodes resolve on their own. The best treatment for a patient may often be the advice that there is a very high likelihood that this issue will resolve on its own. If we rush to seeking a specific diagnosis, we may find one that likely has little correlation with what the patient is actually experiencing. “Experiencing” being the operative word as both practitioners and patients are often guilty of narrow framing. We want to focus on what is going on right now instead of stepping back and looking at the big picture. There is little correlation with structural abnormality, and often it is patient perception of their current situation and anxiety that contribute more to the issue. Narrow framing is a concept from behavioral economics, but as health care practitioners we are essentially financial advisors, only the currency we deal in is health. We can play black jack with our patient’s health, looking for short term gains and what is going on right now, or we can invest in our patient’s health, stepping back to look at the big picture. One is appealing and leads to all kinds of weird treatments and diagnoses, the other is boring. There is no doubt more people have comfortably retired on a 401k than they have with repeated weekends gambling.

We all want to think we have the best information but sometimes it is pertinent to step back and realize how little we know. What we need to realize is that we are practicing at a point in time too, and disposed to current trends that often operate under the guise of “best practice.” One of the more fascinating papers we came across while going over this has been Lutz from 2003 who basically did a history review in the diagnosis of low back pain during the 20th century. Their conclusions would make a point that we still have much to learn and we should, quite possibly, look to move away from a structuralist model.

“Two factors seemed to influence this development: 1) a tendency to prefer organic, visible abnormalities as etiologies; and 2) an inclination to trust technical diagnostic results more than clinical judgment.”

What we see may not be all there is, what we listen to may have the greatest influence on long term outcomes. Yes, the authors advocate for clinical judgment being used as part of the process. This should be judgment based on epistemology more than experience as a function of time. It is more about how we come to conclusions than how long we have been coming to conclusions. This would still be guided primarily by research or a method of defining how we came to the judgment.

We like to throw around phrasing regarding treating the entire patient. It would seem like changing their perception at this point shows the greatest efficacy, maybe we should start there. Maybe, there is no problem.

As always, if you want to discuss the article with us on Twitter, or just recommend a beer for us to try you can find us at:




  1. Henschke N, Maher CG, Ostelo RW, de Vet HC, Macaskill P, Irwig L. Red flags to screen for malignancy in patients with low-back pain. The Cochrane database of systematic reviews. 2013.
  2. Downie A, Williams CM, Henschke N. Red flags to screen for malignancy and fracture in patients with low back pain: systematic review. BMJ (Clinical research ed.). 347:f7095. 2013.
  3. Koes BW, van Tulder MW, Thomas S. Diagnosis and treatment of low back pain. BMJ (Clinical research ed.). 332(7555):1430-4. 2006.
  4. Brinjikji W, Luetmer PH, Comstock B. Systematic literature review of imaging features of spinal degeneration in asymptomatic populations. AJNR. American journal of neuroradiology. 36(4):811-6. 2015.
  5. Baranto A, Hellström M, Cederlund CG, Nyman R, Swärd L. Back pain and MRI changes in the thoraco-lumbar spine of top athletes in four different sports: a 15-year follow-up study. Knee surgery, sports traumatology, arthroscopy : official journal of the ESSKA. 17(9):1125-34. 2009.
  6. Chiu CC, Chuang TY, Chang KH, Wu CH, Lin PW, Hsu WY. The probability of spontaneous regression of lumbar herniated disc: a systematic review. Clinical rehabilitation. 29(2):184-95. 2015.
  7. Macki M, Hernandez-Hermann M, Bydon M, Gokaslan A, McGovern K, Bydon A. Spontaneous regression of sequestrated lumbar disc herniations: Literature review. Clinical neurology and neurosurgery. 120:136-41. 2014.
  8. Waris E, Eskelin M, Hermunen H, Kiviluoto O, Paajanen H. Disc degeneration in low back pain: a 17-year follow-up study using magnetic resonance imaging. Spine. 32(6):681-4. 2007.
  9. Herzog R, Elgort DR, Flanders AE, Moley PJ. Variability in diagnostic error rates of ten MRI centers performing lumbar spine MRI exams on the same patient within a three week period. The spine journal : official journal of the North American Spine Society. 2016.
  10. Kalichman L, Li L, Kim DH. Facet joint osteoarthritis and low back pain in the community-based population. Spine. 33(23):2560-5. 2008.
  11. Andrade NS, Ashton CM, Wray NP, Brown C, Bartanusz V. Systematic review of observational studies reveals no association between low back pain and lumbar spondylolysis with or without isthmic spondylolisthesis. European spine journal : official publication of the European Spine Society, the European Spinal Deformity Society, and the European Section of the Cervical Spine Research Society. 24(6):1289-95. 2015.
  12. Kalichman L, Kim DH, Li L, Guermazi A, Berkin V, Hunter DJ. Spondylolysis and spondylolisthesis: prevalence and association with low back pain in the adult community-based population. Spine. 34(2):199-205. 2009.
  13. Beutler WJ, Fredrickson BE, Murtland A, Sweeney CA, Grant WD, Baker D. The natural history of spondylolysis and spondylolisthesis: 45-year follow-up evaluation. Spine. 28(10):1027-35; discussion 1035. 2003.
  14. Williams FM, Manek NJ, Sambrook PN, Spector TD, Macgregor AJ. Schmorl’s nodes: common, highly heritable, and related to lumbar disc disease. Arthritis and rheumatism. 57(5):855-60. 2007.
  15. Kyere KA, Than KD, Wang AC. Schmorl’s nodes. European spine journal : official publication of the European Spine Society, the European Spinal Deformity Society, and the European Section of the Cervical Spine Research Society. 21(11):2115-21. 2012.
  16. Yin R, Lord EL, Cohen JR. Distribution of Schmorl nodes in the lumbar spine and their relationship with lumbar disk degeneration and range of motion. Spine. 40(1):E49-53. 2015.
  17. Mok FP, Samartzis D, Karppinen J, Fong DY, Luk KD, Cheung KM. Modic changes of the lumbar spine: prevalence, risk factors, and association with disc degeneration and low back pain in a large-scale population-based cohort. The spine journal : official journal of the North American Spine Society. 16(1):32-41. 2016.
  18. Dudli S, Fields AJ, Samartzis D, Karppinen J, Lotz JC. Pathobiology of Modic changes. European spine journal : official publication of the European Spine Society, the European Spinal Deformity Society, and the European Section of the Cervical Spine Research Society. 25(11):3723-3734. 2016.
  19. Teichtahl AJ, Urquhart DM, Wang Y. Modic changes in the lumbar spine and their association with body composition, fat distribution and intervertebral disc height – a 3.0 T-MRI study. BMC musculoskeletal disorders. 17:92. 2016.
  20. Zhang YH, Zhao CQ, Jiang LS, Chen X D, Dai LY. Modic changes: a systematic review of the literature. European spine journal : official publication of the European Spine Society, the European Spinal Deformity Society, and the European Section of the Cervical Spine Research Society. 17(10):1289-99. 2008.
  21. Jensen TS, Karppinen J, Sorensen JS, Niinimäki J, Leboeuf-Yde C. Vertebral endplate signal changes (Modic change): a systematic literature review of prevalence and association with non-specific low back pain. European spine journal : official publication of the European Spine Society, the European Spinal Deformity Society, and the European Section of the Cervical Spine Research Society. 17(11):1407-22. 2008.
  22. Hutton MJ, Bayer JH, Powell JM. Modic vertebral body changes: the natural history as assessed by consecutive magnetic resonance imaging. Spine. 36(26):2304-7. 2011.
  23. Docking S, Rosengarten S, Daffy J, Cook J. Treat the donut, not the hole: The pathological Achilles and patellar tendon has sufficient amounts normal tendon structure. Journal of Science and Medicine in Sport. 2014;18. doi:10.1016/j.jsams.2014.11.015.
  24. Kongsgaard M, Qvortrup K, Larsen J. Fibril morphology and tendon mechanical properties in patellar tendinopathy: effects of heavy slow resistance training. The American journal of sports medicine. 38(4):749-56. 2010.
  25. Weinstein SL, Dolan LA, Spratt KF, Peterson KK, Spoonamore MJ, Ponseti IV. Health and function of patients with untreated idiopathic scoliosis: a 50-year natural history study. JAMA. 289(5):559-67. 2003.
  26. Negrini S, Minozzi S, Bettany-Saltikov J. Braces for Idiopathic Scoliosis in Adolescents. Spine. 41(23):1813-1825. 2016.
  27. Romano M, Minozzi S, Zaina F. Exercises for adolescent idiopathic scoliosis: a Cochrane systematic review. Spine. 38(14):E883-93. 2013.
  28. Burgstaller JM, Schüffler PJ, Buhmann JM. Is There an Association Between Pain and Magnetic Resonance Imaging Parameters in Patients With Lumbar Spinal Stenosis? Spine. 41(17):E1053-62. 2016.
  29. Försth P, Ólafsson G, Carlsson T. A Randomized, Controlled Trial of Fusion Surgery for Lumbar Spinal Stenosis. The New England journal of medicine. 374(15):1413-23. 2016.
  30. Machado GC, Ferreira PH, Yoo RI. Surgical options for lumbar spinal stenosis. The Cochrane database of systematic reviews. 11:CD012421. 2016.
  31. McKillop AB, Carroll LJ, Battié MC. Depression as a prognostic factor of lumbar spinal stenosis: a systematic review. The spine journal : official journal of the North American Spine Society. 14(5):837-46. 2014.
  32. Sinikallio S, Aalto T, Koivumaa-Honkanen H. Life dissatisfaction is associated with a poorer surgery outcome and depression among lumbar spinal stenosis patients: a 2-year prospective study. European spine journal : official publication of the European Spine Society, the European Spinal Deformity Society, and the European Section of the Cervical Spine Research Society. 18(8):1187-93. 2009.
  33. Teichtahl AJ, Urquhart DM, Wang Y. Lumbar disc degeneration is associated with modic change and high paraspinal fat content – a 3.0T magnetic resonance imaging study. BMC musculoskeletal disorders. 17(1):439. 2016.
  34. Teichtahl AJ, Urquhart DM, Wang Y. Fat infiltration of paraspinal muscles is associated with low back pain, disability, and structural abnormalities in community-based adults. The spine journal : official journal of the North American Spine Society. 15(7):1593-601. 2015.
  35. Kalichman L, Klindukhov A, Li L, Linov L. Indices of Paraspinal Muscles Degeneration: Reliability and Association With Facet Joint Osteoarthritis: Feasibility Study. Clinical spine surgery. 29(9):465-470. 2016.
  36. Suri P, Fry AL, Gellhorn AC. Do Muscle Characteristics on Lumbar Spine Magnetic Resonance Imaging or Computed Tomography Predict Future Low Back Pain, Physical Function, or Performance? A Systematic Review. PM & R : the journal of injury, function, and rehabilitation. 7(12):1269-81. 2015.
  37. Niemeläinen R, Briand MM, Battié MC. Substantial asymmetry in paraspinal muscle cross-sectional area in healthy adults questions its value as a marker of low back pain and pathology. Spine. 36(25):2152-7. 2011.
  38. Kountouris A, Portus M, Cook J. Cricket fast bowlers without low back pain have larger quadratus lumborum asymmetry than injured bowlers. Clinical journal of sport medicine : official journal of the Canadian Academy of Sport Medicine. 23(4):300-4. 2013.
  39. Carragee EJ, Alamin TF, Miller JL, Carragee JM. Discographic, MRI and psychosocial determinants of low back pain disability and remission: a prospective study in subjects with benign persistent back pain. The spine journal : official journal of the North American Spine Society. 5(1):24-35. 2005.
  40. Wertli MM, Rasmussen-Barr E, Held U, Weiser S, Bachmann LM, Brunner F. Fear-avoidance beliefs-a moderator of treatment efficacy in patients with low back pain: a systematic review. The spine journal : official journal of the North American Spine Society. 14(11):2658-78. 2014.
  41. Wertli MM, Eugster R, Held U, Steurer J, Kofmehl R, Weiser S. Catastrophizing-a prognostic factor for outcome in patients with low back pain: a systematic review. The spine journal : official journal of the North American Spine Society. 14(11):2639-57. 2014.
  42. Sorensen CJ, George SZ, Callaghan JP, Van Dillen LR. Psychological Factors Are Related to Pain Intensity in Back-Healthy People Who Develop Clinically Relevant Pain During Prolonged Standing: A Preliminary Study. PM & R : the journal of injury, function, and rehabilitation. 8(11):1031-1038. 2016.
  43. Peerdeman KJ, van Laarhoven AI, Keij SM. Relieving patients’ pain with expectation interventions: a meta-analysis. Pain. 157(6):1179-91. 2016.
  44. Zeidan F, Lobanov OV, Kraft RA, Coghill RC. Brain mechanisms supporting violated expectations of pain. Pain. 156(9):1772-85. 2015.
  45. Searle A, Spink M, Ho A, Chuter V. Exercise interventions for the treatment of chronic low back pain: a systematic review and meta-analysis of randomised controlled trials. Clinical rehabilitation. 29(12):1155-67. 2015.
  46. Saragiotto BT, Maher CG, Yamato TP. Motor control exercise for chronic non-specific low-back pain. The Cochrane database of systematic reviews. 2016.
  47. Franke H, Fryer G, Ostelo RW, Kamper SJ. Muscle energy technique for non-specific low-back pain. The Cochrane database of systematic reviews. 2015.
  48. Nijs J, Kosek E, Van Oosterwijck J, Meeus M. Dysfunctional endogenous analgesia during exercise in patients with chronic pain: to exercise or not to exercise? Pain physician. 15(3 Suppl):ES205-13. 2012.
  49. Cherkin DC, Deyo RA, Street JH, Barlow W. Predicting poor outcomes for back pain seen in primary care using patients’ own criteria. Spine. 21(24):2900-7. 1996.
  50. Coste J, Delecoeuillerie G, Cohen de Lara A, Le Parc JM, Paolaggi JB. Clinical course and prognostic factors in acute low back pain: an inception cohort study in primary care practice. BMJ (Clinical research ed.). 308(6928):577-80. 1994.
  51. Lutz GK, Butzlaff M, Schultz-Venrath U. Looking back on back pain: trial and error of diagnoses in the 20th century. Spine. 28(16):1899-905. 2003.