The progression of Alzheimer's disease might be slowed by draining cerebrospinal fluid from the brain, according to pilot data in AD patients reported in the Archives of Neurology and at the American Neurological Association's annual meeting.

The working theory for this line of research is that slow and continuous CSF drainage would reduce the brain burden of soluble forms of putative toxins such as beta-amyloid (Aβ), tau protein, and oxidative toxins. Based on the generally positive results of the pilot study, a larger multicenter trial is already underway to investigate the long-term safety and efficacy of the therapy.

CSF shunts-which typically drain fluid from the lateral ventricles into the peritoneum via a rubber tube-have been used successfully to treat hydrocephalus for almost fifty years. Shunting CSF from the brains of AD patients was tried in the late 1960s, with the goal of reducing symptoms by improving cerebral blood flow. The results were mixed, and the high incidence of complications prevented full exploration of the technique.

The current revival of the method was prompted by two recent reports (Savolainen et al., 1999; Golomb et al., 2000 ) that hydrocephalus patients who also had AD-and who were given shunts-maintained cognitive function equivalent to that of shunted hydrocephalus patients without AD. The authors also cite a large body of evidence suggesting that a failure to clear circulating soluble Aβ and tau may contribute to the formation of amyloid plaques and neurofibrillary tangles, and, perhaps, neurodegeneration.

Gerald Silverberg of Stanford University and colleagues from a number of other institutions, as well as from the shunt manufacturer Eunoe, Inc., employed ventriculoperitoneal shunts that drained CSF at 40 to 140 mL/day-considerably slower than the rate used to relieve hydrocephalus. Their subjects were 29 patients with clinically diagnosed probable Alzheimer's (mild to moderate; Mini Mental State Exam scores 15 to 24). Patients were randomized to CSF shunt (n = 15) or control (no surgery; n = 14) groups. After one year, 11 shunt patients and 12 controls qualified to be part of the efficacy analysis, which included the Mattis Dementia Rating Scale (MDRS) and the MMSE.

Although the study was too small to demonstrate efficacy, the authors point to trends in the cognitive tests that seemed to indicate that the shunt patients, on average, did not deteriorate cognitively during the year, whereas the control patients, on average, did. Similarly, in the subset of patients from whom biochemical samples were obtained after 1 year, there appeared to be some lowering of CSF Aβ and tau levels in the shunt group relative to the control group.

From these same patients, a group of University of Pittsburgh and Eunoe researchers, led by Domenico Pratico from the University of Pennsylvania, measured the clearance of iPF-VI, a marker of in vivo oxidative stress that appears to be elevated in AD. In their poster presentation at the 2002 American Neurological Association annual meeting in New York, these researchers noted that iPF-VI levels had also decreased in the shunt group during the 1-year study period.

According to the authors of the Archives article, their results justified the full, multicenter trial that is now underway. In an accompanying Archives editorial, however, David Bennett of Rush Alzheimer's Center in Chicago, Illinois, and Michael McDermott of the University of Rochester in New York take a more critical view. They note that there were five "notable" adverse events in the shunt group: two new-onset seizures, a shunt infection, a bowel injury during surgery, and a severe headache in a patient who had previously experienced migraines. (All these subjects recovered and continued the trial.) This level of complications, according to Silverberg and associates, is consistent with previous experience of shunt placement in elderly hydrocephalus patients, but Bennett and McDermott ask whether this is an appropriate comparison group. They point out that shunts are a matter of life or death for hydrocephalus patients, whereas their proposed beneficial effects in Alzheimer's patients are not convincingly demonstrated here, even by the standards of a pilot study. Further, given the hope for therapies that more directly address the underlying causes of neurodegeneration in AD, they wonder whether CSF shunts will ever be a relevant treatment for AD.—Hakon Heimer


  1. Alzheimer’s disease (AD) brain is characterized by progressive and diffuse accumulation of neuritic plaques and neurofibrillary tangles, which are composed of deposits of amyloid-β peptide (Aβ) and abnormally phosphorylated tau protein. In the recent past, a considerable effort has been made to develop methods to retard their deposition or improve their clearance. As a result, Aβ has become a realistic target for developing effective therapies for AD. However, after some promising and exciting data from animal studies with vaccines against Aβ, preliminary negative results in AD patients have tempered the initial enthusiasm.

    In this study, Silverberg et al. treated AD patients for 12 months with a low-flow CSF drainage and found reductions in CSF Aβ (1-42) and tau, which were associated with a stabilization of their cognitive function. These results would be consistent with the hypothesis that improving CSF circulation may result in slowing the progression of AD.

    There are several aspects of this pilot study that need to be taken into account: 1) the sample size was too small to provide a definitive test of the hypothesis; 2) the study was not double-blind; 3) the controls did not receive any surgery, so a placebo effect cannot be ruled out. Despite these considerations, I believe that this study is very exciting and interesting. The results justify a compelling rationale for a randomized, larger and controlled clinical trial, which could give us clear and definitive answers. Considering that today the main pharmacological interventions for AD offer only minor and transient symptomatic benefit, CSF shunting could be considered as a completely novel and promising approach to treat AD.

  2. Equivocal data but intriguing nonetheless. Maybe another evidence for Aβ clearance as therapeutic for amyloid pathology/clinical outcome. Hard to know though whether this is persuasive enough to mount a sufficiently powered trial and blinded perhaps?

    View all comments by Edward Koo

Make a Comment

To make a comment you must login or register.


External Citations

  1. Savolainen et al., 1999;
  2. Golomb et al., 2000

Further Reading

No Available Further Reading

Primary Papers

  1. . Assessment of low-flow CSF drainage as a treatment for AD: results of a randomized pilot study. Neurology. 2002 Oct 22;59(8):1139-45. PubMed.
  2. . Cerebrospinal fluid shunting for Alzheimer's disease?. Neurology. 2002 Oct 22;59(8):1126-7. PubMed.