A Change of Heart: Limitations of National Healthcare Safety Network (NHSN) and Society of Thoracic Surgeons (STS) Surveillance Strategies in Identifying Surgical Site Infections after Heart Transplant Surgery

Abstract

Surgical Site Infections (SSI) surveillance after heart transplantation is critical for preventing SSI. We analyzed the accuracy of three SSI surveillance systems: National-Healthcare-Safety-Network (NHSN), Society-of-Thoracic-Surgeons (STS), and detailed manual surveillance (TXID). NHSN and STS showed low sensitivity compared to TXID, highlighting the need for more accurate and efficient surveillance strategies.

Introduction

Surgical site infections (SSI) are a severe complication of orthotopic heart transplantation (OHT), with rates ranging 4.8–12.4%.^1–4 These infections are associated with increased hospital length of stay, morbidity, mortality, and healthcare costs.^1–5 Accurate SSI surveillance is crucial for preventing infections and improving patient outcomes. Most OHT programs adopt the National Healthcare Safety Network (NHSN) or the Society of Thoracic Surgeons (STS) systems for SSI surveillance.^6,7 An alternative strategy is manual SSI surveillance through detailed case review and adjudication performed by Transplant Infectious Diseases (TXID) physicians. Each surveillance system has different methodologies (Table 1), leading to variability in SSI detection rates. This study aimed to evaluate the sensitivity (SN) and specificity (SP) of NHSN and STS surveillance in identifying SSI after OHT procedures compared to the manual TXID system.

Table 1.

Surveillance definition for surgical site infections in heart transplant surgeries among three surveillance systems based on depth of infection at the time of study period (1/1/19–12/31/20)

Category	National Healthcare Safety Network (NHSN)	Society of Thoracic Surgeons (STS)	Transplant ID Surveillance (TXID)
Surveillance window	30 days for superficial and deep incisional SSI. 90 days for deep incisional and organ/space SSI.	30 days for all SSI for the period 01/01/19–06/30/20. From 07/01/20 to 12/31/20 a 90-day surveillance window was adopted for deep sternal wound infections.	90 days for all SSI
SSI category evaluated	Primary	Primary	Primary and secondary
Interval surgeries following index transplant surgery	SSI attributed to interval surgeries.	SSI attributed to index transplant surgery, irrespective of interval surgeries.	SSI attributed to index transplant surgery, irrespective of interval surgeries.
Superficial incisional SSI	Occurs within 30 days. Involves only skin and subcutaneous tissue of the incision.	Occurs within 30 days. Involves only skin and subcutaneous tissue of the incision.	Occurs within 90 days. Involves only skin and subcutaneous tissue of the incision.
Deep incisional SSI	Occurs within 90 days. Involves deep soft tissues (e.g., fascial and muscle layers).	Occurs within 30 days (90-day surveillance began on 07/01/20). Involves deep soft tissues (e.g., fascial and muscle layers).	Occurs within 90 days. Involves deep soft tissues (e.g., fascial and muscle layers).
Organ/Space SSI	Occurs within 90 days. Involves any part of the anatomy (organs or spaces) other than the incision that was opened or manipulated during the operation	Occurs within 90 days. Involves any part of the anatomy (organs or spaces) other than the incision that was opened or manipulated during the operation	Occurs within 90 days. Involves any part of the anatomy (organs or spaces) other than the incision that was opened or manipulated during the operation.

Methods

This study included all adult (≥18 years) single-organ OHT surgeries performed 1/1/19–12/31/20 at an academic medical center. Re-do OHT surgeries were included. During this period three different surveillance systems were implemented to detect SSI: NHSN, STS, and TXID. The primary objective of this study was to evaluate the sensitivity (SN) and specificity (SP) of NHSN and STS surveillance systems in identifying SSI after OHT procedures compared to the manual TXID surveillance system, which was considered the referent standard. The study was approved by the Health System’s Institutional Review Board (IRB number: Pro00104110). Study data were collected and managed using REDCap electronic data capture tools hosted at Duke University.⁸

SSI definitions recommended by NHSN, STS, and TXID surveillance systems were adopted as detailed in Table 1.⁹ Superficial incisional, deep incisional, and organ space SSIs were evaluated. Primary (e.g. sternotomy site) and secondary (e.g. ECMO cannulation sites) SSI were evaluated by TXID, while NHSN and STS evaluation was limited to primary SSI. Details on the TXID adjudication process for SSI are described in our previous publication.¹

Descriptive statistics analyzed the accuracy of NHSN and STS definitions for detecting SSI following OHT surgeries compared to TXID surveillance. Sensitivity, specificity, positive predictive value, and negative predictive value were calculated.

Results

The study period included 162 adult OHT procedures. TXID surveillance identified 14 SSIs following 162 (8.6%) OHT procedures. (Figure 1) These SSI occurred at a median of 27.5 (IQR 16.0–44.0) days after OHT. Among 14 SSIs identified by TXID, 12 involved the primary surgical incision and 2 the secondary surgical incisions.

Figure 1.

Panel A. Venn diagram demonstrating the overlap between surgical site infections detected by the three surveillance systems: transplant infectious disease (TXID, referent-standard), Society of Thoracic Surgeons (STS), and National Healthcare Surveillance Network (NHSN). Panel B. Reasons for National Healthcare Surveillance Network (NHSN) missing 10 surgical site infections (SSI) following orthotopic heart transplant (OHT) and Society of Thoracic Surgeons (STS) missing 9 SSI following OHT. Of note, OHT SSI could be missed for more than one reason.

NHSN identified 7 SSIs, representing an SSI incidence of 4.3%. Among these, TXID deemed 3 cases to be false positives. In contrast, STS surveillance identified 5 SSIs, or SSI incidence of 3.1% with no false positives. NHSN and STS both had false-negatives. Specifically, NHSN did not detect 10 SSIs. NHSN false-negatives resulted mainly from: i) attributing SSI to intervening surgeries and not to index transplant procedure (50.0%); ii) not including the transplant procedure in the NHSN dataset (30.0%); and iii) not evaluating secondary SSI (20.0%). STS missed 9 SSIs. Short surveillance window, lack of evaluation of secondary SSI, and unexplained exclusion of transplant procedures from the STS dataset accounted for 55.5%, 22.2%, and 22.2% of STS false negatives, respectively (Table 2).

For NHSN SN, SP, positive predictive value (PPV) and negative predictive value (NPV) were 28.6%, 97.9%, 57.1%, and 93.6%, respectively. STS surveillance had higher SN (35.7%), SP (100%), PPV (100%), and NPV (94.3%).

Discussion

This study highlights the low sensitivity of NHSN and STS surveillance systems for identifying SSI following OHT surgery compared to the manual, time-intensive TXID surveillance.

Both NHSN and STS are semi-automated surveillance systems. However, both demonstrated poor sensitivity, 28.6% and 35.7%, respectively. The low sensitivity of NHSN and STS may be related to several factors, including the short surveillance window for SSI, the attribution of SSI to intervening surgeries following the index transplant surgery, and the exclusion of secondary SSI.

TXID surveillance continued for 90 days following OHT surgery with the vast majority of SSI identified between 16 and 44 days after transplant.¹ Conversely, STS and NHSN surveillance was shorter. Specifically, STS surveillance was limited to 30 days in the study period 01/01/19–06/30/20 and was expanded to 90 days for deep sternal infections only in July 2020. NHSN surveillance spanned from 30 days for superficial SSI to 90 days for deep and organ/space SSI during the study period. We expect that this issue will continue, since the updated 2024 NHSN SSI guidelines recommend limiting the OHT SSI surveillance period to 30 days.⁹

In addition, NHSN and STS rely on limited clinical elements in conjunction with administrative data, which often misses infections that occur during complex post-operative courses involving re-operations, such as chest closures or washouts. These complexities often shift the SSI attribution away from the original OHT surgery, contributing to under-reporting. Moreover, NSHN and STS did not capture secondary SSIs, which of themselves may lead to admission, surgery, and additional antibiotic therapy.

TXID surveillance is more sensitive in the diagnosis of SSI, but it is a labor-intensive exercise performed by highly trained transplant infectious diseases physicians. We estimated that TXID surveillance required ≥30 minutes per patient. The improved SSI detection of TXID surveillance is likely a result of more comprehensive SSI definitions and of its ability to capture the full scope of post-transplant recovery. This finding highlights the trade-off between the ease of automated surveillance systems and the manual methods.

Despite the strengths of TXID, the manual nature and expertise requirement of this system makes it a challenge for widespread implementation, particularly in hospitals with limited infection control resources. Future studies should explore revision of SSI definitions, as well as ways to improve the sensitivity of existing surveillance systems.

Conclusion

The study demonstrates that current SSI surveillance systems such as NHSN and STS have low sensitivity in detecting SSI following OHT. This is likely due to the complexity of post-transplant hospital courses and the limited surveillance windows used by these systems. In contrast, though time- and resource-intensive, comprehensive manual surveillance conducted by TXID physicians captures SSI that semiautomated systems fail to detect. Modifying existing surveillance definitions to account for these complexities is essential for improving the detection of SSI in OHT patients. Future research should focus on refining surveillance methodologies to ensure they adequately capture post-transplant infections, thereby supporting more targeted infection prevention strategies.

Highlights.

• TXID detected nearly 75% more SSIs than NHSN or STS systems.

• NHSN and STS surveillance had low sensitivity for OHT SSIs.

• Manual TXID surveillance identified secondary SSIs overlooked by NHSN and STS.

• Short surveillance windows contributed to NHSN and STS underreporting SSIs.

• Enhanced SSI definitions are vital for improved post-transplant infection detection.