Preventing ICU Infections

In depth

The one-page summary above is what the PDF prints. These sections go further.

How ICU infections start

On any given day about 1 in 31 US inpatients has a healthcare-associated infection (3.2% in CDC's 2015 survey, Magill 2018), most often pneumonia, C. difficile or surgical site infection. NHSN (2026) estimates about 18,100 CLABSIs a year in US ICUs and wards, and hospitals reported more than 17,000 CAUTIs in 2023.

With short-term central lines, skin organisms migrate from the insertion site along the outer surface (CDC 2011); with longer dwell, contaminated hubs and connectors seed the lumen (CDC 2002), and about half of hubs are colonized under standard practice (SHEA 2022). Biofilm then shields bacteria from antibiotics, which is why infected short-term lines usually come out.

Urinary organisms climb the outside of the catheter from the perineum or enter the lumen through a broken closed system. Bacteriuria accrues at 3–7% per catheter day, so duration is the main risk factor (SHEA 2022).

An endotracheal tube holds the cords apart, so secretions pooled above the cuff leak past it and the tube grows biofilm; sedation, lying flat and a weak cough remove the rest of the defenses. VAP's attributable mortality is about 13%, highest in surgical patients and near zero in trauma and medical patients (Melsen 2013).

C. difficile spores on hands, rails and equipment resist alcohol; antibiotics that strip the colonic flora let swallowed spores germinate and release toxins A and B. MRSA, VRE and CRE spread the same way, and every antibiotic course selects for them.

Surveillance definitions and quality measures

NHSN definitions are built for consistent counting, not bedside diagnosis. An infection is device-associated only when the device had been in more than 2 consecutive calendar days and was present on the event date or the day before.

Rates are infections per 1,000 device days. The standardized infection ratio (SIR) divides observed by predicted infections from national NHSN data for similar units, so an SIR below 1.0 beats the prediction; the standardized utilization ratio does the same for device days (SHEA 2022).

Since October 2008 Medicare pays nothing extra for CAUTI, vascular catheter infection or certain surgical site infections (mediastinitis after CABG, some orthopedic and bariatric operations) not present on admission. The HAC Reduction Program scores CLABSI, CAUTI, colon and hysterectomy SSI, MRSA bacteremia, C. difficile and the PSI 90 composite; the worst-scoring quartile loses 1% of Medicare payments (CMS 2026). CLABSI and CAUTI are classic nurse-sensitive indicators, benchmarked in NDNQI.

CLABSI: the evidence behind the bundle

In the Keystone ICU project (Pronovost 2006), a checklist of five practices (hand hygiene, maximal barriers, chlorhexidine prep, avoiding the femoral site, removing unneeded lines) cut the mean rate in 103 Michigan ICUs from 7.7 to 1.4 per 1,000 catheter days, a sustained drop of up to 66%.

In 3SITES (Parienti 2015; 3,471 ICU catheters), bloodstream infection plus symptomatic thrombosis ran 1.5, 3.6 and 4.6 per 1,000 catheter days for subclavian, jugular and femoral lines, but subclavian lines caused more pneumothorax needing a chest tube (1.5% vs 0.5% jugular). SHEA 2022 prefers subclavian in the ICU and CDC 2011 says avoid femoral in adults; ultrasound mainly prevents mechanical complications, and PICCs are not a CLABSI-prevention strategy (SHEA 2022).

Chlorhexidine-alcohol beat povidone-iodine-alcohol for skin prep (CLEAN, Mimoz 2015: 0.28 vs 1.77 infections per 1,000 catheter days). Chlorhexidine sponge dressings cut major catheter-related infection from 1.1% to 0.5%, and 7-day changes were noninferior to 3-day changes for colonization (Timsit 2009). Daily chlorhexidine bathing reduced hospital-acquired bloodstream infection by 28% (Climo 2013), but a 9,340-patient trial found no change in combined device infections and C. difficile (Noto 2015); SHEA 2022 still grades ICU bathing essential, with resistance an open question.

Scrub hubs and connectors with alcoholic chlorhexidine (preferred) or 70% alcohol for at least 5 seconds and let them dry, or use antiseptic caps (SHEA 2022). In RSVP (Rickard 2021; 2,944 patients), 7-day administration sets matched 4-day sets (catheter-related bloodstream infection with central lines 1.78% vs 1.46%); CDC 2011 keeps 24 h for blood products and lipids and changes needleless connectors no more often than every 72 h. SHEA 2022 also grades adequate nurse staffing and fewer float nurses in ICUs as high-quality evidence.

Arterial lines cause about 0.96 bloodstream infections per 1,000 catheter days, femoral about twice radial (O'Horo 2014): use a cap, mask, sterile gloves and a small fenestrated drape for radial lines, maximal barriers for femoral or axillary lines, and change transducer systems every 96 h (CDC 2011).

CAUTI: indications, care and urine stewardship

Up to half of hospital catheters may lack a valid indication (SHEA 2022). In the ICU, hourly output justifies a catheter only while the number changes treatment (resuscitation, diuresis, vasopressor titration), not for ICU status alone. Acute retention is a scanned post-void residual over 500 mL without symptoms or over 300 mL with them (SHEA 2022). CDC 2009 also lists prolonged immobilization (unstable spine, pelvic fractures), which SHEA 2022 dropped. A catheter placed only because surgery is long comes out at the end of the case (SHEA 2022); other postoperative catheters come out as soon as possible, ideally within 24 h (CDC 2009).

Removal is what works: reminders and stop orders cut CAUTI by 52%, and stop orders shortened catheterization (Meddings 2010 meta-analysis), and a national program lowered non-ICU rates from 2.28 to 1.54 per 1,000 catheter days but not ICU rates (Saint 2016). Nurse-driven protocols remove the catheter once no indication remains, then watch voiding with a bladder scanner and intermittent catheterization for retention.

Secure the catheter against traction, keep the bag below the bladder and off the floor, keep tubing unkinked, empty it into a clean container for each patient without the spigot touching, and replace catheter and bag only for infection, obstruction, leaks or a broken system. Routine meatal hygiene is enough; CDC 2009 advises against antiseptic meatal cleaning (SHEA 2022: unresolved). External catheters avoid the urethra, but their skin injury and infection rates are still being defined (SHEA 2022).

Culture urine only for fever, altered mental status or new sepsis without another source, flank pain, acute hematuria or new pelvic discomfort; cloudy or foul urine and pyuria alone are not indications, and asymptomatic bacteriuria is not screened for or treated except in pregnancy or before endoscopic urologic procedures (SHEA 2022). Take the sample from the disinfected port, never the bag, and send it within about an hour or refrigerate it (up to 24 h).

VAP and ventilator-associated events

NHSN moved adults from VAP to VAE surveillance in 2013 because VAP definitions rest on subjective chest films and signs; most VACs prove to be pneumonia, ARDS, atelectasis or pulmonary edema. Patients must be ventilated at least 4 calendar days, the daily minimum is the lowest setting held over 1 h, and PEEP 0–5 counts as 5, so a patient on PEEP 5 must reach 8 (NHSN 2026).

IVAC needs, within 2 days of the worsening, a temperature over 38 °C or under 36 °C or a WBC ≥12,000 or ≤4,000, plus a new antimicrobial continued ≥4 days. PVAP adds a quantitative culture (endotracheal aspirate ≥10⁵ CFU/mL, BAL ≥10⁴, protected brush ≥10³), purulent secretions (≥25 neutrophils and ≤10 squamous cells per low-power field) with an organism, or a positive pleural fluid, histology, Legionella or respiratory virus test; Candida, coagulase-negative staphylococci and enterococci from airway samples do not count.

Chlorhexidine oral care lowered VAP mainly after cardiac surgery, not in double-blind studies of other patients, with a nonsignificant mortality signal (Klompas 2014); toothbrushing was associated with less hospital-acquired pneumonia (RR 0.67) and lower ICU mortality (RR 0.81) across 15 trials (Ehrenzeller 2024). The older IHI ventilator bundle also listed stress ulcer and DVT prophylaxis and, from 2010, chlorhexidine oral care; keep the prophylaxis for its own indications, but it does not prevent VAP.

At the bedside: cuff pressure 20–30 cm H₂O, suction the mouth and subglottic port before moving the tube or deflating the cuff, and drain ventilator condensate away from the patient (CDC 2003).

C. difficile: testing, precautions and treatment

C. difficile causes more than 450,000 US infections a year, with attributable mortality of 4.5–5.7% in endemic periods and recurrence in 10–30% (SHEA 2022). About 8% of patients carry it on admission, so testing formed stool or laxative diarrhea finds colonization: test only new, unexplained diarrhea of ≥3 unformed stools in 24 h, not within 48 h of a laxative or 7 days of a prior test, and never as a test of cure (IDSA/SHEA 2017, SHEA 2022).

Start contact precautions when diarrhea is recognized, before the result, in a single room with its own toilet, and continue at least 48 h after it resolves, or until discharge if CDI rates stay high (IDSA/SHEA 2017). Gloves are the main hand protection. Soap and water removes spores mechanically: IDSA/SHEA 2017 prefers it in outbreaks and accepts either method otherwise, SHEA 2022 calls it preferred after CDI care but notes overall hand hygiene can fall, and exams expect soap and water.

Dedicate equipment, and clean daily and at discharge with a sporicidal EPA List K product such as 1:10 bleach, which SHEA 2022 supports for outbreaks and high-rate units; UV or hydrogen peroxide systems and probiotics remain unproven. Stewardship is the strongest lever (30–50% fewer CDIs in meta-analyses): limit fluoroquinolones, clindamycin, third- and fourth-generation cephalosporins and carbapenems, and stop unneeded proton pump inhibitors, although that alone is unproven (IDSA/SHEA 2017).

MRSA, VRE, CRE and Candida auris

CDC Appendix A keeps contact precautions for MDROs the infection program judges significant. SHEA 2022 calls them essential for MRSA but lets a hospital modify them after an MRSA-specific risk assessment (never during an outbreak or rising rates), so policies differ. Flag the chart, act on lab alerts and tell the receiving unit at transfer.

In REDUCE MRSA (Huang 2013; 74 ICUs, 74,256 patients), universal decolonization with daily chlorhexidine bathing and twice-daily nasal mupirocin for 5 days lowered MRSA clinical isolates (HR 0.63) and bloodstream infection from any pathogen (HR 0.56) more than screening and isolation, one bloodstream infection prevented per 54 patients. Universal gowns and gloves (BUGG, Harris 2013) did not reduce combined MRSA or VRE acquisition but did lower MRSA acquisition.

CRE needs contact precautions, lab alerts, transfer communication and stewardship (CDC 2025). Candida auris colonizes skin for months or indefinitely, so precautions last the whole stay: contact precautions in a single room, alcohol hand rub (it works) and an EPA List P disinfectant, because quaternary ammonium alone fails (CDC 2026).

Nurses carry much of stewardship: cultures before the first dose, clarified allergy labels, the 48–72 h antibiotic review and stop dates.

Hand hygiene, PPE and isolation in practice

Clean hands at the WHO 5 moments: before touching a patient, before a clean or aseptic procedure, after body fluid exposure risk, after touching a patient and after touching the surroundings. Alcohol rub (≥60%) is preferred, enough to wet all surfaces and rubbed until dry; wash with soap and water for at least 15 s when hands are soiled or after C. difficile or norovirus care. Gloves never replace hand hygiene, and artificial nails are banned in ICUs and operating rooms (SHEA 2022).

PPE goes on gown, mask or respirator, eye protection, gloves, and comes off gloves, eye protection, gown, mask, with hand hygiene last; remove a respirator outside the closed door. An AIIR is negative pressure with at least 6 air changes per hour (12 in new rooms). Influenza needs droplet precautions for 7 days or until 24 h after symptoms resolve, whichever is longer, but intubation, extubation, bronchoscopy, sputum induction, open suctioning and CPR need an N95, gown, gloves, eye protection, an AIIR when feasible and only essential staff (CDC 2026).

Surgical site infection: the ICU nurse's part

SSIs follow 1–3% of inpatient operations and up to 60% are preventable (SHEA 2022). Prophylaxis starts within 60 minutes before incision (120 for vancomycin and fluoroquinolones), is weight-based (cefazolin 2 g, 3 g above 120 kg), is redosed at two half-lives (cefazolin every 4 h) and stops after closure even with drains (SHEA 2022, CDC 2017); cardiac programs following STS 2006 continue up to 48 h. Routine vancomycin prophylaxis is discouraged unless the patient carries MRSA.

Decolonize S. aureus carriers, or everyone, before cardiac and orthopedic surgery (up to 5 days of twice-daily nasal mupirocin plus daily chlorhexidine bathing); in Bode 2010 this cut S. aureus infection from 7.7% to 3.4%, with the largest effect on deep SSIs. Chlorhexidine-alcohol prep beat povidone-iodine in clean-contaminated surgery (SSI 9.5% vs 16.1%, Darouiche 2010). Clip only if hair interferes, keep the patient above 35.5 °C, and hold postoperative glucose at 110–150 mg/dL for 24–48 h (SHEA 2022); STS 2009 uses under 180 after cardiac surgery and CDC 2017 under 200.

Keep the sterile dressing on for 24–48 h unless it leaks (CDC 1999), clean hands before every wound contact, and report spreading erythema, purulent drainage, dehiscence, a sternal click or new fever.

Exam pearls

• Removing the device is the best CLABSI and CAUTI prevention; review the need daily.
• CLABSI insertion: checklist, hand hygiene, ultrasound, maximal barriers, dried chlorhexidine-alcohol, subclavian over femoral.
• CVC dressings: transparent every 7 days, gauze every 2, sooner if damp, loose or soiled; scrub the hub ≥5 s.
• Tubing: continuous sets 96 h to 7 days, blood and lipids within 24 h, propofol 12 h.
• CAUTI: closed system, bag below the bladder and off the floor, no scheduled changes; don't culture or treat asymptomatic bacteriuria.
• VAP: head of bed 30–45°, daily SAT plus SBT, toothbrushing without chlorhexidine, subglottic suction if >48–72 h.
• VAC: PEEP up ≥3 or FiO₂ up ≥0.20 for ≥2 days after ≥2 stable days; IVAC adds fever or abnormal WBC and a new antibiotic ≥4 days.
• C. diff: contact precautions until ≥48 h after diarrhea stops; soap and water, bleach cleaning and own equipment.
• Airborne: N95 and negative pressure; droplet: surgical mask; contact: gown and gloves.
• SIR = observed ÷ predicted infections; the worst HAC quartile loses 1% of Medicare payments.