Background/Purpose:

Evidences for perioperative management of disease modifying anti-rheumatic drugs (DMARDs) and biologic agents (BA) are sparse, and limited mainly to methotrexate & specific surgeries (orthopedics). Such data may not be generalizable to other surgeries or DMARDs/BA. The use of administrative database is difficult here due to lack of validated methods to predict stopping of DMARDs/BA before surgery. Using novel techniques to predict stopping of DMARDs/BA, we used data from Veterans Affairs (VA) to compare infection risks of RA patients who stopped versus continued DMARDs/BA perioperatively over a 10-year period from 2000-2009.

Methods:

We identified 6548 RA patients in VA administrative databases using validated algorithms & included only those on 1 DMARD or BA in the perioperative period. Those on multiple DMARDs/BA were excluded to simplify result interpretation.

We predicted drug stoppages by calculating x = medication stop date closest to the surgery – next start date. y = surgery date-stop date was used to determine if the drug was stopped before or after surgery. To validate this method, two investigators independently reviewed clinic notes from the Houston VA facility for actual start or stop dates before or after surgery.  A third investigator reviewed and resolved conflicting chart review results. ROC analyses were performed to obtain optimal x and y values to distinguish if DMARDs/BA were stopped and if it occurred before or after surgery.

The primary endpoints were wound infections within 30 days of surgery, according to the modified 1992 US Centers for Disease Control and Prevention criteria for postoperative infection, and other infections including pneumonia, UTI and sepsis. Propensity scores were used to match factors that may influence infection rates such as comorbidity scores, chronic steroid use, smoking, diabetes mellitus, etc.

Results:

In the validation part of the study, ROC analyses found that x≥33 days best predicted stoppage of DMARD/BA (AUC=0.954) and y≥-11 best predicted that DMARD/BA was stopped before surgery (AUC=0.846).

Risk of post-op general infection or wound infection in RA patients who stopped DMARDs/BA before surgery were not significantly different compared with those who did not stop these agents. Those who stopped BA after surgery had significantly higher odds of post-op wound (OR 13.7, p=0.014) and post-op general infections (OR 9.2, p=0.005) compared to those who did not stop BA. Similarly stopping DMARDs after surgery was associated with increased risk of post op wound infection (OR 3.08, p=0.000) and post op general infection (OR 1.68, p=0.024) compared with not stopping treatment. Treatment was stopped postoperatively likely because of post-operative infection. 

Conclusion:

Using our novel technique of identifying DMARDs/BA discontinuation, we showed that there was no significant difference in post-op infection risk whether stopping anti-rheumatic treatment preoperatively or not.  Our results grouped all types of surgeries and different DMARDs/BA. Further analyses looking at different types of surgeries and individual DMARDs/BA will be helpful to evaluate possible differences in infection risks between individual DMARDs/BA in different types of surgeries.

---

« Back to 2013 ACR/ARHP Annual Meeting

ACR Meeting Abstracts - https://acrabstracts.org/abstract/perioperative-use-of-anti-rheumatic-agents-does-not-increase-early-postoperative-infection-risks-a-veteran-affairs-administrative-database-study/