Reducing Complications in the ICU

Abstract & Commentary

Synopsis: Grade A Evidence supports the increased risk for DVT and VTE in critically ill patients and, thus, the use of prophylactic measures against these complications in such patients.

Source: Saint S, Matthay MA. Am J Med 1998; 105:515-523.

Saint and matthay comprehensively reviewed the recent medical literature on the epidemiology of three important complications associated with ICU care: venous thromboembolism, stress-related upper gastrointestinal bleeding, and vascular catheter-related infections. They sought to derive evidence-based recommendations from reported studies pertaining to these complications. Relevant publications were located via a MEDLINE search and cross-citation, focusing on reports published in the last 10 years; on meta-analyses; and on rigorously designed and carried out clinical trials. The levels of evidence provided by each reviewed study (e.g., Grade A = evidence based on results from at least one randomized control trial primarily in ICU patients, etc.) were recorded and used in developing Saint and Matthay’s recommendations. A summary of their recommendations follows.

Venous Thromboembolism (VTE)

Patients in the ICU are at greater risk for developing VTE than are other patients. Those at increased risk for developing VTE in the ICU include those with age older than 40 years, prior eposodes of VTE; underlying malignancy; prolonged immobility or paralysis; major surgery; congestive heart failure; fractures (especially of hip, pelvis, or leg); and stroke. Patients with hypercoagulable states, such as protein C resistance or dysfibrinogenemia, are at especially increased risk. Several studies document the fact that relatively few of the patients who fall into these increased risk groups receive VTE prophylaxis while in the ICU.

Measures to prevent VTE are effective, although data come mainly from studies in surgical patients. Low-dose unfractionated heparin (e.g., 5000-7000 U subcutaneously every 8-12 hours) is effective in reducing the incidence of deep venous thrombosis (DVT) in a variety of patient groups. Low molecular weight heparin is also effective in preventing DVT in surgical and stroke patients, but few studies have been done in medical patients; it is considerably more expensive than unfractionated heparin. Intermittent pneumatic compression is also effective for several groups of surgical patients that have been studied, and is a reasonable alternative for patients who cannot receive anticoagulants.

Saint and Matthay conclude that Grade A evidence supports the increased risk for DVT and VTE in critically ill patients and, thus, the use of prophylactic measures against these complications in such patients. They conclude that either unfractionated or low molecular weight heparin can be used. The evidence that intermittent pneumatic compression may be substituted for either form of heparin is less compelling, especially in medical patients.

Stress-Related Upper Gastrointestinal Bleeding

Upper gastrointestinal (GI) bleeding is common in patients admitted to ICUs. Those at most risk are patients with respiratory failure requiring mechanical ventilation and those with coagulopathy (e.g., platelet count < 50,000/mL or prolonged prothrombin or partial thromboplastin time). Patients with extensive burns and those with multiple traumatic injuries may also be at increased risk.

From their study of available evidence, Saint and Matthay conclude that prophylaxes with H2-receptor antagonists prevent clinically important upper GI bleeding but have not been shown to affect mortality. They recommend prophylaxes with H2-receptor antagonists for patients with coagulopathy or on ventilators whose anticipated duration of ventilatory support will be less than four days. For patients whose anticipated duration of ventilatory support will be more than four days, they recommend sucralfate because of the potential for increased propensity to ventilator-associated pneumonia with the use of H2-receptor antagonists in such patients. Saint and Matthay point out, however, that the strength of the evidence here is less than with the prevention of DVT and VTE.

Catheter-Related Vascular Infection

Risk factors for catheter-related vascular infections include longer duration of catheterization, catheter location (e.g., femoral or internal jugular vs subclavian sites), the use of transparent dressings (which may increase the risk), the absence of systemic antibiotic therapy, and the use of less stringent barrier precautions during placement. It is uncertain whether multilumen catheters increase the risk for catheter-related infections.

Chlorhexidine gluconate is the agent of choice for skin disinfection prior to catheter insertion. Central venous triple-lumen catheters impregnated with antibacterial agents are probably preferable to nonimpregnated catheters when the line will be in place for less than seven days and the patient remains at increased risk for infection despite rigorous adherence to appropriate infection control practices. Central venous lines, pulmonary arterial catheters, and arterial lines need not be changed routinely at any fixed interval so long as there is no evidence for local catheter-related infection, fever without a clear source, or catheter malfunction. Transparent dressings may increase the risk for local infection but provide other benefits that may justify their use. Saint and Matthay recommend that pulmonary artery catheters not be used unless the patient is likely to benefit from the availability of central hemodynamic data, although they state that the evidence here is not from properly designed randomized control trials.


This is a useful examination of a sometimes confusing body of literature on three important ICU-related complications. An accompanying article (Gould MK, et al. Am J Med 1998;105:551-553) points out that Saint and Matthay did not perform a formal meta-analysis or use the usual format of a systematic review, and that their paper is more a comprehensive narrative review of the subject under consideration. Nonetheless, it is helpful to clinicians to have the current literature summarized in a structured manner and to have recommendations presented along with summaries of the data on which they are based.

Venous thromboembolism, stress-related upper GI bleeding, and catheter-related infections are common, important complications with which everyone working in the ICU environment needs to be concerned. There is abundant evidence that patient care in many settings falls far short of the state of the art in this area. The recommendations of Saint and Matthay are reasonable, practical, and based on a rigorous examination of the available literature. Patient outcomes in the ICU would likely improve if these recommendations were more widely followed.