NSAIDs and nonunion of fractures: Effect may not be causal

[Guest guest]

NSAIDs and nonunion of fractures: Effect may not be causal

Aug 3, 2005 Alice Goodman

Boston, MA - Although some studies have suggested that nonsteroidal

anti-inflammatory drugs (NSAIDs) may delay or inhibit healing of bone

fractures, the latest study to weigh in on this issue suggests that in

fact the association between NSAIDs and nonunion of fracture reflects

an attempt to deal with the pain of the bone nonunion, rather than

NSAIDs being a cause of it [1]. The studythe largest to date, with the

greatest number of nonunionsis reported in the June 15, 2005 issue of

Arthritis & Rheumatism.

" The strongest association between NSAIDs and nonunion was found late

in the course of nonunion, rather than immediately after fracture

occurred, suggesting almost a reverse causalitythat is, that the [pain

of the] nonunion is causing the medication use, not that the medication

use is causing the nonunion, " says senior author Dr

(Brigham and Women's Hospital, Boston, MA).

Largest study to date

The community-based study identified a cohort of 9995 patients age 65

and older with an inpatient or outpatient diagnosis of humeral shaft

fracture selected from a database of more than 500 000 patients

enrolled in Medicare between 1995 and 2000. Nonunions developed in 105

patients (1.1%), while 1032 (10.3%) patients were exposed to

nonselective NSAIDs during the 90-day, postfracture period.

noted that this analysis did not include selective COX-2

inhibitors, because the number of patients exposed to these agents was

too low. The study period covered the 90-day time period postfracture

because the purported risk of NSAID exposure is in the early stages of

fracture healing and the majority of fractures heal within three

months. The need for a procedure at 90 days was considered evidence of

nonunion.

The primary hypothesis of the study was that NSAID use would inhibit

fracture healing. The initial analysis found a strong association

between NSAID use and nonunion. However, when the investigators looked

at the time course of medication use, they found that the only

significant association between NSAID use and nonunion was during the

61-90-day window postfracture (RR 3.9, 95% CI 2.0-6.2). NSAID exposure

during days 0-30 and days 31-60 was not significantly associated with

nonunion, said.

" One would anticipate that if a drug were retarding healing, the use of

the drug at the time of fracture would have the most dramatic effect.

In fact, our findings were in the opposite direction. Patients who took

NSAIDs at a later time period seemed to have the highest risk and the

strongest association with nonunion, " told rheumawire.

The investigative team went on to look at opioids (a class of drug

known to have no effect on fracture healing) and found that late use of

opioids during the 61-90-day postfracture time window was also

significantly associated with nonhealing (RR 2.7, 95% CI 1.5-5.2).

" Our study does not give a clear clinical message about the wisdom of

using NSAIDs postfracture. From a clinical standpoint, we did not prove

that NSAIDs are harmful or helpful in terms of fracture healing. We set

out to study a clinical question, but as we probed more carefully, we

observed a source of protopathic bias from the methods and not

necessarily a real clinical finding, " commented.

The fact that no clear association was found between drug use and

outcome is more likely due the inability to detect a nonunion early

after fracture, he explained. Nonunion is only detected at the time of

a procedure to correct it.

Study limitations

said that the association between NSAIDs and delayed healing of

fracture or nonunion, which has caused concern among both orthopedic

surgeons and rheumatologists, as previously reported on rheumawire, was

based on animal studies and several poorly designed studies in humans.

He acknowledged that his study also had limitations and that the study

cannot rule out a detrimental effect of NSAID use on fracture healing.

The limitations include a lack of detail on exact configuration of

fractures or pain intensity. Although the study documents prescriptions

for NSAID use among the fracture cohort, it is not possible to

ascertain whether patients actually took the prescribed medication. It

is possible that NSAIDs and/or opioids could exacerbate nonhealing by

reducing the pain and encouraging use of the affected limb.

said that the fact that use of both NSAIDs and opioids late in

the course of fracture healing had the strongest association with

nonunion supports the theory that analgesics (NSAIDs and opioids) are

being used to treat impending nonunions, rather than causing them.

Detrimental effect seen in animal studies

" This is a good study, and I accept the findings that in a large

database there is no association between NSAID use and nonunion within

the first 60 days following fracture, " says Dr Einhorn, (Boston

University Medical Center, Boston, MA). Einhorn has coauthored several

animal studies showing a detrimental effect of NSAIDs on fracture

healing and wrote an editorial urging caution about the use of these

drugs in fracture patients [2].

Recent experimental studies by Einhorn and colleagues continue to show

an inhibitory effect of NSAIDs and COX-2 inhibitors on bone healingbut

only when animals were taking the drugs. Once the drugs were stopped,

fracture healing took place and was comparable to healing observed in

normal animals not administered drugs, Einhorn tells rheumawire.

These series of experiments looked at the effect of taking NSAIDs and

COX-2 inhibitors for seven and 21 days; healing at 35 days (a validated

time frame for healing in the rat model) was similar between the groups

that had taken the drugs and a control group of drug-free animals.

" Animal data show an inhibitory effect of NSAIDs and COX-2 inhibitors

on bone healing. However, when the drugs are discontinued after one or

three weeks, the effect is reversible, " Einhorn said.

Extrapolating the animal data to humans, Einhorn felt that a normal

patient with a bone fracture could safely take an NSAID or COX-2

inhibitor for associated pain for a short time with no detrimental

effect on healing.

" There probably is a limiting healing response when these drugs are

taken, but it is probably not clinically relevant as long as the drug

is stopped, " he commented.

A different question is whether patients with comorbid conditions

and/or on medications that impair the healing process can safely take

an NSAID or COX-2 inhibitor for fracture pain. " We don't know the

answer to that question, " Einhorn said.

Sources

1. Bhattacharyya T, Levin R, Vrahas M, and DH. Nonsteroidal

antiinflammatory drugs and nonunion of humeral shaft fractures.

Arthritis Rheum 2005; 53:364-367.

2. Einhorn T. Do inhibitors of cyclooxygenase-2 impair bone healing?

J Bone Miner Res 2002; 17:977-978.