Perry D. Clark says that a steroid injected near his spine to relieve persistent back pain instead left him "way, way worse." Twelve years later, he still suffers from continuous stinging in his legs and feet and occasional bursts of excruciating pain.
"It's like somebody took a hot poker out of a fire and jammed it into my foot for two or three seconds," said Mr. Clark, a retired media professional from Petoskey, Mich.
The outbreak of fungal meningitis that has killed 14 people and sickened 156 more has focused attention on the risk of infection from spinal injections. But the same injections have also long been linked to other rare but devastating complications, including nerve damage, paralysis and strokes.
The Food and Drug Administration is already reviewing how to reduce the risk of "catastrophic neurological injuries" from the injections, said Dr. James P. Rathmell, chief of pain medicine at Massachusetts General Hospital, who is involved in the review. The risk of infections did not even factor into the review, though it will now, he said.
The meningitis outbreak is raising new questions about the steroid spinal injections, which are given to millions of Americans. Use has mushroomed even as clinical trials have found only modest evidence that the injections help. Moreover, the steroids, while approved for uses like relieving inflammation in joints, have not been approved by the F.D.A. for epidural injections, next to the spinal cord.
"Not only were these people killed, but there was no ethical reason to give this treatment," said Dr. William Landau, a professor of neurology at Washington University in St. Louis, referring to those who died of meningitis.
Many pain specialists dispute that conclusion. Doctors are allowed to, and often do, prescribe drugs for unapproved uses, they say, and steroids have been used to treat back pain for decades. They contend the injections can be less risky than narcotics or surgery.
Even Dr. Rathmell, who has been calling attention to the complications, said they occur in only about one in 10,000 cases. "In the right individuals, they are a tremendous help," he said of the injections.
Kenny Alhadeff, the producer of the Broadway musical "Memphis," says he is one of them. Several years ago, he said, he had such severe back pain that "I could barely get into a car." His first injection brought immediate relief. Now, after a few years of periodic injections, he is pain-free.
But some defenders of the practice concede that injections are overused. They are most useful for people with herniated disks and pain radiating into the legs or arms. But a study published in the journal Spine in 2007 found that fewer than half of the injections given were for these conditions.
"We are doing too many of these, and many of those don't meet the proper criteria," said Dr. Laxmaiah Manchikanti, who runs a pain clinic in Paducah, Ky., and is chairman of the American Society of Interventional Pain Physicians. He also said that about 20 percent of doctors who perform the procedures were not adequately trained.
Dr. Manchikanti said his own review of Medicare records found an increase of nearly 160 percent in the number of injections from 2000 to 2010.
The increased use is driven by the aging of the population, the desperation of patients and the desire of physicians to help — and there are financial incentives. Medicare and private insurers pay $100 to several hundred dollars for an injection, and there are pain clinics that do almost nothing but injections.
Dr. Richard Deyo, a professor of family medicine at Oregon Health and Science University, said that despite the increase in injections and other aggressive treatments, surveys and Social Security disability records suggest that "people with back pain are reporting more functional limitations and work limitation, rather than less."
Evidence on the effectiveness varies by the condition being treated, the drug used and the injection technique.
A review last year by Washington State, which was considering whether to pay for such procedures, found that for one set of circumstances, there were seven clinical trials that showed the injections were helpful, another seven that found them no better or even worse than a placebo, and three with unclear results.
The state agency decided that the evidence was strong enough to justify paying for injections under certain circumstances.
The serious complications, while extremely rare, are more noticeable because of the explosive growth in the number of injections. In one anonymous survey, 287 pain physicians reported 78 serious complications, including 13 deaths, among their patients.
The injections are made into the epidural space just outside the spinal column. This is the same site used in numbing the pain of childbirth, though women in labor receive an infusion of a local anesthetic, not an injection of a steroid.
But the needle can sometimes go astray, putting the drug into the spinal fluid or arteries, causing nerve damage, hemorrhages and death to nerves by depriving them of oxygen. Many doctors use imaging and fluorescent dye to position the needle, but even that technique is not foolproof.
Another complication is arachnoiditis, an inflammation of a membrane surrounding the nerves of the spinal cord that is marked by pain, nerve damage and bowel and bladder dysfunction. Mr. Clark who said he has this condition, uses a catheter to urinate.
The F.D.A. review is focusing on developing best practices for injection techniques with the aim of reducing the risk of injury.
Some doctors are turning to steroids that are free of preservatives, which may damage nerves, and particles, which may clog tiny blood vessels feeding the spinal cord.
But such products in general are not made by drug manufacturers, causing doctors to turn to compounding pharmacies, which are lightly regulated. One of them, the New England Compounding Center, supplied a contaminated drug — called methylprednisolone acetate — that is suspected of causing the fungal meningitis outbreak.
Moreover, the particle-free steroids may not provide lasting relief, said Dr. Christopher Gharibo, an associate professor of anesthesiology and orthopedics at New York University.
Last year, the label for the steroid Kenalog, made by Bristol-Myers Squibb, was changed to say that epidural injection was not recommended. But the label for Pfizer's Depo-Medrol, the brand name version of methylprednisolone acetate, does not have such a warning.
A Pfizer spokesman said the company did not condone the epidural use of Depo-Medrol.
Dennis J. Capolongo, who runs an advocacy group called the End Depo Now Campaign, says the lack of warning is inexplicable.
A former photojournalist in Washington, Mr. Capolongo said an epidural injection of Depo-Medrol to treat hip pain in 2001 inflamed his nerves, leaving him hospitalized for weeks and bedridden for two years. The pain, while no longer so intense, is not completely gone.
"There are nights I cry myself to bed," he said.