Near-infrared spectroscopy: what we know and what we need to know--a systematic review of the congenital heart disease literature

Abstract

Objectives: Neurologic dysfunction is a problem in patients with congenital heart disease. Near-infrared spectroscopy may provide a real-time window into cerebral oxygenation. Enthusiasm for near-infrared spectroscopy has increased hopes of reducing neurologic dysfunction. However, potential gains need to be evaluated relative to cost before routine implementation. Responding to data in ways that seem intuitively beneficial can be risky when the long-term impact is unknown. Thus, we performed a systematic review of the literature on near-infrared spectroscopy in congenital heart disease.

Methods: A literature search from 1950 to April 2007 for near-infrared spectroscopy in congenital heart disease was undertaken. We identified 54 manuscripts and\13 reviews.

Results: There were 47 case series, 4 randomized trials, and 3 retrospective studies. Two studies had postdischarge follow-up, one incorporating neurologic testing. Neither of these studies demonstrated a benefit. One retrospective study, which included near-infrared spectroscopy and other intraoperative measures of cerebral perfusion, demonstrated a decrease in neurologic dysfunction using this combination of monitors. Three small studies were able to correlate near-infrared spectroscopy with other clinical and radiologic findings.

Conclusions: Many centers, and even entire countries, have adopted near-infrared spectroscopy as standard of care. The available data suggest that multimodality monitoring, including near-infrared spectroscopy, may be a useful adjunct. The current literature on the use of near-infrared spectroscopy alone, however, does not demonstrate improvement in neurologic outcome. The data correlating near-infrared spectroscopy findings with indirect measures of neurologic outcome or mortality are limited. Although near-infrared spectroscopy has promise for measuring regional tissue oxygen saturation, the lack of data demonstrating improved outcomes limits the support for widespread implementation.