Research news and notes

1. Spine Patient Outcomes Research Trial (SPORT) study—is lumbar discectomy useful? 

The studies with the greatest impact on neurosurgical practice, especially as seen by nonneurosurgeons, are often not published in neurosurgical journals. The results of the Spine Patient Outcomes Research Trial (SPORT) were published in JAMA on November 06 [4], [5]. The project had 2 parts—a randomized trial of surgical vs nonoperative treatment for lumbar disk herniation and radiculopathy and an observational cohort study.

The randomized study [5] divided 501 patients from 13 practices over 4 years (2000-2004), all deemed to be potential surgical candidates, into surgical and nonoperative groups and followed them up for more than 2 years. Patients were included if they had symptoms of lumbar radiculopathy for at least 6 weeks and had some form of nonoperative treatment in that period. They had imaging (usually magnetic resonance imaging) and no strong indication for surgery, such as neurologic deficits, tumor, or marked spinal instability. Contraindications to surgery or unwillingness to consider surgery were also exclusion factors. Surgery in this study meant a standard open discectomy. Nonoperative treatments were varied.

The study aimed to perform an intent-to-treat analysis. This attempt was thwarted by a massive crossover of patients. Only 50% of patients assigned to surgery were operated within 3 months, and 30% of those assigned to nonoperative treatment received surgery in the same period. Eventually, 43% of the nonoperative group had surgery. Generally, patients with worse symptoms or those who felt they were getting worse got surgery, and those who had better functional status and felt that they were improving chose not to have surgery, regardless of the assignment. Intent-to-treat analysis revealed only minor differences in outcome between the groups, as expected. Analysis of patients “as treated” revealed a marked advantage for surgery on a variety of measures of satisfaction, pain, and function. The authors commented that the as-treated analysis does not protect from confounding factors and is no longer a true randomized study. Placebo effects of surgery and differences between patients who choose surgery and those who do not may limit the ability to interpret the results.

I believe that the confounding runs even deeper. It was not surprising that patients crossed over, with the help of their doctors, because in the surgeons minds, the decision to operate or not was clinical and individual and not amenable to randomization. The surgeons felt they knew the indications. They did not operate on improving patients, despite the randomization, and they operated on those who were getting worse. The surgeons took a large variety of factors into account and individualized the care for each patient. Given the large personal expertise of each surgeon, the study was not really randomized.

A second parallel study in the SPORT project [4] looked at the outcomes of those patients who refused to be randomized. In this observational cohort, 743 patients were enrolled initially. Of those 521 chose surgery, and only 4% of those remained nonoperative at 2 years of follow up. Of the 222 who initially chose nonoperative treatment, 22% had surgery by the end of 2 years, leaving 191 who remained nonoperative. Despite the crossover from the nonoperative to surgical group, the surgical group demonstrated substantially better outcomes by a variety of self reported outcome measures (satisfaction with symptoms, self rated improvement, etc). Importantly, both groups showed marked improvement over the 2-year follow-up. Both components of SPORT appear to demonstrate efficacy and benefit of surgery for herniated lumbar disk and radiculopathy. However, in reality, they add very little information not already known. Ultimately, most of the patients in both arms of the study had surgery whenever the patient and their surgeon felt they really needed and wanted it. These patients were generally happy with their choice. We know this from our practice. Will this large and expensive study change any current practice? Is it able to prove anything? According to the authors, the study was initiated because the large regional differences within the United States and lower discectomy rates internationally raised questions about the appropriateness of the indications for some of these surgeries. Was this study ever able to answer the question posed? Could the selection of patients and procedures by established excellent spine surgeons participating in this study have shed light on questionable practices by other surgeons in other institutions?

2. Glioma stem cells resist radiation 

Glioblastoma remains deadly because the tumor recurs after surgery, radiation therapy, and chemotherapy no matter how aggressive the modalities. Although they are considered diffuse tumors and the infiltrative nature is commonly considered to be the root of the difficulty to eradicate the tumors, in reality, most glioblastomas recur locally, in a limited number of sites at the margin of the original, and appear as focal masses. This fact may indicate that the recurrence starts from a small number of cells. In a recent letter to Nature, Bao et al [1] described how a population of glioma stem cells resists the damage of radiation. The authors first describe identification of cancer stem cells within gliomas. The concept of cancer stem cells is increasingly prominent in theories of cancer formation and growth. These cells tend to be more robust, less abnormal, with better repair potential and a better capacity to proliferate compared to “regular” cancer cells. In this study, the authors noted that after irradiation, the proportion of CD133+ cancer stem cells increased 4-fold, compared with the total number of tumor cells. Because CD133 expression was not induced in committed tumor cells, the result meant increased radioresistance of the stem cells. Further experiments demonstrated that the CD133+ cells retained tumor forming capacity after radiation, much exceeding the capacity of “regular” glioma cells. They were able to repair DNA better than tumor cells and were capable of “checkpoint activation,” a process by which DNA replication is stopped to allow repair. This latter effect appears to be a key factor in the overall radioresistance of the glioma stem cells.

These findings and, in general, the understanding of the glioma stem cells, may finally improve the treatment of glial tumors. Therapy may target the “checkpoint” mechanism directly. Alternatively, treatment may be directed toward differentiating or “normalizing” the tumor stem cells. This approach is described in another article in the same issue of Nature [3]. The authors described how bone morphogenic proteins (BMP) suppress glial stem cells. These proteins are active in the differentiation of normal human neural stem cells. Transient exposure to BMP4 in vitro abolished the capacity of transplanted glioblastoma cells to establish tumors in rodents; BMPs activated receptors on the tumor cells that reduced their proliferation and increased expression of neural markers, without killing the cells. The size of the CD133+ population decreased overall. The 2 articles together point to a future when a combined approach to glioblastoma treatment may include radiation combined with a treatment to induce differentiation of the residual glioma stem cells.

3. Human embryonic stem cells derived from blastomeres 

An abundant source of pluripotent human stem cells is the basis for all research into repair and replacement of tissues damaged by injury or disease. Embryonic stem cells are, at this time, the only source of stem cells that can develop into any human tissue. To create human stem cell lines, an embryo had to be destroyed, precipitating an ethical and religious argument. Advances in technology can make the argument obsolete. At the blastocyst stage, a single cell can be removed from the early embryo without apparently affecting its viability. In the 23 November issue of Nature, Klimanskaya et al [2] report successful derivation of human embryonic stem cells from a single-cell biopsy from an early embryo. Such biopsies are routinely used in the field of in vitro fertilization for preimplantation diagnosis. The embryos that undergo such tests and are implanted can develop normally. These tests are currently widely accepted and performed. When grown in a dish with other similar cells, 28 of 53 biopsied cells proliferated and formed vesicles, and 2 continued to proliferate indefinitely, with normal kariotypes, markers of pluripotent stem cells, and the ability to form all 3 embryonic germ layers. Apparently, not all the cells in the early embryo can form a stable stem cell line, but some can. This may be enough to produce many new stem cell lines for research. A little more than 5 years after the US government placed restrictions on new embryonic stem cell lines, the paradigm shifts again.