Response to: Avidan and Evers (2011) J Alzheimers Dis 24, 201-216

2 February 2014

The topic of long term cognitive impairment is of concern to patients and clinicians and it is critical that it be addressed with appropriately designed studies. It is disappointing to note that despite a plethora of opinion pieces dominating the literature, there remain no prospective studies investigating dementia following surgery and anesthesia.

It was therefore with interest that I re-visited the paper entitled ‘Review of Clinical Evidence for Persistent Cognitive Decline or Incident Dementia Attributable to Surgery or General Anesthesia’ (J Alzheimers Dis 24, 201-216, 2011), in particular, noting the implication from the title, indicating that this manuscript is a review of clinical research studies in both these areas.

I was surprised to identify a number of points which suggest that the peer-review process could be questioned in this particular instance.

I draw your attention to the following issues:

  1. A review is defined as ‘a formal assessment or examination’. In order for an issue to be assessed or examined, there must be published literature on that subject. There is no prospective publication to date investigating incident dementia following anesthesia and surgery; therefore this cannot be a review on that subject. A systematic review would provide Level I evidence if each of the included articles were prospective randomized controlled trials that met criteria for Level II evidence. There were no prospective randomized controlled trial studies investigating incident dementia included in this review (there are none to date in the literature). A retrospective cohort study is Level III-3 evidence.
  2. Sixteen of the 17 articles reviewed are prospective investigations of postoperative cognitive dysfunction (POCD) following surgery and anesthesia. Not one of these prospective studies was designed to assess dementia at baseline, nor were they designed to investigate either incident or prevalent dementia at follow-up. Given there exists no evidence to date that POCD and dementia post anesthesia and surgery are related, using the terms interchangeably is inaccurate and must be avoided.
  3. Incident dementia is the presence of new cases. In order to identify incident dementia at follow-up, it would be essential to determine baseline dementia status. None of the 16 prospective studies reviewed conducted baseline assessment of dementia. Prevalence of dementia may be determined at follow-up if assessed for, but this was not included in any of the reviewed articles.
  4. The remaining one article reviewed which did consider dementia is a retrospective review by the author.
  5. Persistent cognitive decline suggests long-term (longer than one year) decline following anesthesia and surgery, and yet only five of the 16 prospective studies investigated cognition beyond 2 years, and only 7 of 16 beyond 12 months.
  6. For this to have been a review of ‘incident dementia’ as the title and the conclusions state, the review would have been to examine the current evidence on incident dementia following surgery and general anesthesia. This review does not include any reference which meets these criteria.
  7. Many of the studies reviewed are analyzed using group analyses. While this type of analysis investigates the differences between groups, it does not reflect the incidence of decline in individuals, which is the critical issue researchers should be addressing. Additionally, it is only possible to identify incident dementia by describing and analyzing for individual outcomes. This is not possible when employing group analyses.

This could be a reasonable review of literature investigating POCD. It is not a review of incident dementia. It is not a review of persistent cognitive decline. The title and conclusion are misleading at best and it is disappointing that such statements manage to pass through the peer-review process to publication, enter into the body of literature, and have the potential to mislead a keen audience.

A/Prof Lisbeth Evered1,2 (lis.evered@svhm.org.au), A/Prof David Scott1,2, A/Prof Brendan Silbert1,2

1Centre for Anaesthesia and Cognitive Function, St Vincent’s Hospital, Melbourne, Australia; 2Melbourne Medical School, University of Melbourne, Australia; 3Department of Surgery, University of Melbourne, Australia

Comment Icon |
Bookmark Icon Bookmark Recommend Icon Recommend Follow Icon Follow
| Bookmark | Recommend | Follow

Comments

Letter to the Editor: Response

We thank Professors Evered, Scott, and Silbert for their comments regarding our review article [1] and appreciate that they concur with our conclusion that credible clinical evidence for persistent postoperative cognitive decline (POCD) or incident dementia attributable to surgery is lacking. The chief censure they lay at our doorstep and at that of the peer reviewers of the Journal of Alzheimer’s Disease recalls the words of Northumberland in Shakespeare’s Henry IV:

“Yet the first bringer of unwelcome news

Hath but a losing office, and his tongue

Sounds ever after as a sullen bell,

Remember'd tolling a departing friend.”

The essence of their criticism, expressed in seven points, is that the unwelcome news conveyed in our review article was unwarranted. They assert that we did not have grounds to author the review because the corpus of clinical evidence on this topic is so meager. In essence they posit that it is inappropriate (on technical grounds) to question widely held beliefs or practices by examining whether there is sufficient evidence to support them. Their seven points deserve reply.

  1. We did in fact conduct a formal and comprehensive assessment of the available, relevant literature. It is true that randomized controlled trials (RCTs) with incident dementia as an endpoint were not included in our review, as such studies did not exist then, and still do not exist now. However, we did include RCTs that evaluated cognitive outcomes when patients underwent cardiac surgery with general anesthesia versus percutaneous coronary intervention [1]. These studies found no evidence of persistent cognitive decline in those who underwent surgery with general anesthesia. Subsequent to our review article, this finding has been reinforced by a seminal 280 patient RCT by Sauër et al. [2] that found at “7.5 years follow-up, off-pump coronary artery bypass grafting patients had a similar or perhaps even better cognitive performance compared with percutaneous coronary intervention patients.” In his editorial contextualizing this trial, Selnes [3] stated, "this study by Sauër and associates adds to a long list of previous studies that by now have convincingly demonstrated that surgical interventions for coronary artery disease are not associated with a higher risk of late cognitive decline or Alzheimer’s disease than medical or nonsurgical interventions."
  2. We did not, as alleged, use the terms POCD and dementia interchangeably and we agree wholeheartedly that this would be inaccurate and should be avoided. We explicitly differentiated these concepts repeatedly in our review. For example, we stated, “While delirium and dementia are well-defined clinical syndromes, POCD can only be measured using psychometric tests, and neither the tests nor the definitions of POCD have been standardized.” [1]
  3. We agree that few studies on surgical patients to date have determined baseline dementia status or have investigated it at follow-up. However, it is not true to state that our review did not include any studies that evaluated the incidence of dementia after surgery [4-6].
  4. Regarding our study titled “Long-term cognitive decline in older subjects was not attributable to noncardiac surgery or major illness” [6], Alzheimer Disease Research Center records were reviewed retrospectively, but cognitive assessments and clinical dementia ratings were conducted prospectively and annually. One of the contributions of our study to the field was to show that preoperative cognitive trajectories typically predict postoperative cognitive trajectories, which points to the necessity of assessing preoperative cognitive trajectory in studies focusing on postoperative cognition.
  5. We arbitrarily defined persistent POCD as lasting longer than 6 months in our review, whereas Professors Evered, Scott, and Silbert arbitrarily define it as lasting longer than 1 year. We are not convinced that their arbitrary temporal definition trumps ours of a condition whose very existence is controversial; it is unrecognized in the International Classification of Disease code and has no entry in the Diagnostic and Statistical Manual of Mental Disorders.
  6. We included the best evidence on incident dementia following surgery and anesthesia that existed at the time we wrote the review. We disagree that none of the references met the stipulated criteria [4-6]. It is worth noting that we summarized our findings as follows, “In view of conflicting evidence, it is currently unknown whether surgery and anesthesia are associated with an increased risk of incident dementia. Future human clinical studies should be designed to probe scrupulously the hypothesized causal link between exposure to anesthesia and brain damage that accelerates the onset of Alzheimer’s disease.” [1]
  7. It is unlikely to be illuminating, as suggested, to examine cognitive decline in individuals, as cognitive decline and incident dementia are common in older people regardless of whether or not they undergo surgery. Therefore only rigorous group investigations including suitable non-surgical controls and employing appropriate statistical analyses [7] will have the potential to disambiguate whether or not incident dementia following surgery and anesthesia is coincidental or likely to be attributable to the surgical or anesthetic event.

The appropriate question is not why we elected to conduct a formal review when “there remain no prospective studies investigating dementia following surgery and anesthesia.” The germane question is rather why, despite the lack of such studies, there is “a plethora of opinion pieces dominating the literature (suggesting a causal link between surgery/anesthesia and persistent cognitive decline and dementia).” We contend that it is the authors of the plethora of opinion pieces as well as those who have conducted methodologically flawed research who have had the “potential to mislead a keen audience.” The purpose of our review article was to provide a foil to the pervasive literature by evaluating the quality of the evidence base so that a “keen audience” could reach conclusions on dispassionate grounds. In closing, we refer back to Northumberland and ask that Professors Evered, Scott, and Silbert reconsider their salvo directed at us, the messengers. We are not the ones responsible for laying the evidential foundations on which the shaky edifices of persistent POCD and incident dementia attributable to surgery and anesthesia are based; our sin was simply to reveal that the foundations are made of sand.

Michael S. Avidan and Alex S. Evers

Department of Anesthesiology, Washington University in St. Louis, St. Louis, MO, USA; Tel.: +1 314 747 4155; Fax: +1 314 747 1716; avidanm@anest.wustl.edu

References

[1] Avidan MS, Evers AS (2011) Review of clinical evidence for persistent cognitive decline or incident dementia attributable to surgery or general anesthesia. J Alzheimers Dis 24, 201-216.

[2] Sauër AM, Nathoe HM, Hendrikse J, Peelen LM, Regieli J, Veldhuijzen DS, Kalkman CJ, Grobbee DE, Doevendans PA, van Dijk D; Octopus Study Group (2013) Cognitive outcomes 7.5 years after angioplasty compared with off-pump coronary bypass surgery. Ann Thorac Surg 96, 1294-1300.

[3] Selnes OA (2013) Invited commentary. Ann Thorac Surg 96, 1300-1301.

[4] Ritchie K, Carrière I, Ritchie CW, Berr C, Artero S, Ancelin ML (2010) Designing prevention programmes to reduce incidence of dementia: prospective cohort study of modifiable risk factors. BMJ 341, c3885.

[5] Yip AG, Brayne C, Matthews FE; MRC Cognitive Function and Ageing Study (2006) Risk factors for incident dementia in England and Wales: The Medical Research Council Cognitive Function and Ageing Study. A population-based nested case-control study. Age Ageing 35, 154-160.

[6] Avidan MS, Searleman AC, Storandt M, Barnett K, Vannucci A, Saager L, Xiong C, Grant EA, Kaiser D, Morris JC, Evers AS (2009) Long-term cognitive decline in older subjects was not attributable to noncardiac surgery or major illness. Anesthesiology 111, 964-970.

[7] Nadelson MR, Sanders RD, Avidan MS (2014) Perioperative cognitive trajectory in adults. Br J Anaesth, doi: 10.1093/bja/aet420.

Response:

We acknowledge and appreciate the response from Professors Avidan and Evers to our letter regarding their review article [1]. While we do not wish to enter an ongoing dispute, we feel compelled to clarify the essential issue which appears to have been misinterpreted. Our main concern, as stated in our letter, was that reviewing incident dementia attributable to surgery or general anesthesia (as the authors claim to have done), would require a body of literature investigating incident dementia following surgery/anesthesia. We all agree this work does not yet exist. As academics, we have no preconceived ideas about what the outcome may be, so to suggest we find the conclusions drawn by the authors’ ‘unwelcome news’ is inappropriate. Rather, we wish to highlight that there is, as yet, insufficient literature on which to base such conclusions.

Additionally, the authors’ response refers several times to manuscripts by Ritchie et al. [2], Yip et al. [3], and their own work [1], as the key articles in their review supporting their conclusions. We have discussed the work by the author in our original letter. I refer the authors to their Tables 1 & 2 of the manuscripts included in their review, and note that neither the Ritchie et al. paper, nor the Yip et al. paper were included. Although these articles are mentioned in the text, they do not contribute to the 17 review articles on which the conclusions are based. Additionally, these papers are population studies, not intervention studies, so no patient underwent anesthesia and surgery as a component of the study protocol.

Finally, our comments regarding individual analysis versus group analysis does not suggest we should be selecting random individuals. We agree with the comments by the author regarding the rigorous study design which is required for this work. Our comments relate to the analysis of outcomes. The clinical importance of this field of research must lie in identifying individuals with cognitive decline and ultimately the ability to prospectively identify vulnerable patients if a risk exists, rather than group outcomes which simply tell us if there is a difference between two groups.

A/Prof Lisbeth Evered1,2 (lis.evered@svhm.org.au), A/Prof David Scott1,2, A/Prof Brendan Silbert1,2

1Centre for Anaesthesia and Cognitive Function, St Vincent’s Hospital, Melbourne, Australia; 2Melbourne Medical School, University of Melbourne, Australia; 3Department of Surgery, University of Melbourne, Australia

References

[1] Avidan MS, Evers AS (2011) Review of clinical evidence for persistent cognitive decline or incident dementia attributable to surgery or general anesthesia. J Alzheimers Dis 24, 201-216.

[2] Ritchie K, Carriere I, Ritchie CW, Berr C, Artero S, Ancelin ML (2010) Designing prevention programmes to reduce incidence of dementia: prospective cohort study of modifiable risk factors. BMJ (Clinical research ed) 341, c3885.

[3] Yip AG, Brayne C, Matthews FE, Function MRCC, Ageing S (2006) Risk factors for incident dementia in England and Wales: The Medical Research Council Cognitive Function and Ageing Study. A population-based nested case-control study. Age Ageing 35, 154-160.