Nutrition

More has been reported about the relationship between infection and nutrition in the general surgery and critical care literature than the spine literature.^5–9 Often such studies involve polytrauma and burn victims in severe catabolic states. However, the principles involved apply to elective spine surgery as well.

A study by Klein et al.¹⁰ followed three groups of patients and analyzed infections and other complications against markers of nutritional status. Patients were deemed nutritionally replete if they had a serum albumin of at least 3.5 g/dL and an absolute lymphocyte count (a stable immune marker) of at least 1500 cells/mm³. Patients falling below either or both of these cutoffs were considered malnourished.

In a group of 114 patients undergoing elective spinal procedures, a total of 85 were found to be replete prior to surgery, and 29 malnourished. The former group had a total of 2 complications, and the latter 11, a difference that is even more dramatic considering the disparate sizes of the groups. Of note, they found that 40% of patients older than 60 years were malnourished. They found similar results among patients who were operated on for spondylodiscitis, as well as spinal cord injury.

Considering the aforementioned, a reasonable nutrition assessment and management approach would be to check serum albumin and absolute lymphocyte counts preoperatively, especially in older patients. If abnormal, surgery should be deferred until a nutrition consult is obtained and the patient is nutritionally replete.

Antiseptic Shower

The use of antiseptic showers, either with povidone-iodine (Betadine) or chlorhexidine gluconate (CHG), has been advocated by some. A study of 700 surgical patients demonstrated a reduction of bacterial skin colonization with either soap, by a factor of 1.3-fold with iodine and 9-fold with CHG.¹¹ Similar results have been found elsewhere.¹² Although evidence to support a clear reduction in SSIs is lacking,¹³ a bottle of CHG solution sufficient for two preoperative showers costs about $9 US at the time of this writing, and its use is likely of great enough benefit to offset that minor cost. Thus, the practice is recommended.

Mupirocin Nasal Ointment

Staphylococcus aureus is the leading cause of SSIs in clean surgical procedures, including spinal operations. An association has been noted between nasal carriage of S. aureus in patients and the occurrence of SSIs. Twenty-five percent to 30% of the U.S. population are nasal carriers of S. aureus at any given time.¹⁴ A short course of treatment with mupirocin (Bactroban) ointment has been shown to eliminate S. aureus in many of these carriers.

To date, two randomized, controlled trials (RCTs) have been conducted to evaluate the efficacy of preoperative mupirocin ointment usage in reducing SSI rates.^14,15 Both studies showed a trend toward efficacy, but neither was significant. A later analysis showed that pooling of the results showed a nearly significant decrease in the infection rate.¹⁶ However, when all nosocomial S. aureus infections (not just SSIs) among patients with nasal S. aureus were considered, the study from Perl et al. did show a statistically significant decrease in incidence with the use of mupirocin ointment.¹⁴

At this point, it is difficult to advocate the widespread use of mupirocin ointment preoperatively, considering the lack of convincing evidence showing benefit, as well as the cost (about $40 US). However, a rapid screening test for S. aureus that uses polymerase chain reaction (PCR) technology has been developed¹⁷ and has shown excellent sensitivity and specificity. This has allowed treatment targeted only toward carriers, which shows promise.¹⁸ It is reasonable to expect that within the next several years this technology will be more widely accessible. If so, it could allow for the selective treatment of carriers, which would be expected to demonstrate a beneficial effect.

Hair Removal

Possibly one of the most ingrained practices in all of surgery is shaving the skin prior to an operation. Unfortunately, it is probably also detrimental. Removing hair by shaving with a razor has been compared with the use of electric clippers in three RCTs.^19–21 These trials were similar in design and focused on clean operations (general and cardiac procedures), so their results were pooled in a recent Cochrane review.²² This yielded a total of 3193 patients, divided nearly evenly between shaving (1627) and clipping (1566). The infection rate was 2.8% for the former group, and 1.4% for the latter, yielding a relative risk (RR) of 2.02, which surpassed statistical significance.

In addition to this strong evidence against shaving, two other points can be made. First, there is no good evidence to show that hair removal lowers the infection rate. The step may be omitted entirely. Second, depilatory creams have been associated with a lower infection rate than shaving in several trials²²; this provides another alternative to razors should complete hair removal be desired.

Razors should only be used for hair removal with the clear understanding that their use has been associated with higher infection rates in several large, well-designed trials.

Skin Preparation

The rationale for preparing the skin prior to incision is twofold. First, the mechanical scrubbing of the skin removes dirt, as well as some bacteria and dead skin cells. Second, the prep solution should have an intrinsic bactericidal and/or bacteriostatic effect.

Commonly employed agents contain alcohol (isopropyl or ethyl), CHG, or iodine/iodophors. Alcohol has excellent activity against bacteria and good activity against mycobacteria, fungi, and viruses.²³ However, it cannot be used alone because it has essentially no residual activity once allowed to evaporate. Prior to evaporation it is flammable, which makes it incompatible with electrocautery.

CHG has good to excellent activity against bacteria and viruses. It is fair at eliminating fungi and has little activity against mycobacteria.²³ Its residual activity is excellent; however, it can cause keratitis and ototoxicity with serious consequences.

Triclosan (the active ingredient in dishwashing detergent) and parachlorometaxylenol (PCMX) are less efficacious and are not generally considered suitable for use in skin preparation or as surgical hand scrubs.

Given the clinical limitations of other preparations, CHG and iodophor solutions are most commonly used as surgical skin preps. CHG has been shown to reduce bacterial skin colonization to a greater degree than iodophors (see prior section), but no evidence yet demonstrates a lower SSI rate when using CHG.

Nevertheless, in a large RCT, CHG has been shown to reduce the line infection rate when compared with iodophor prep in the placement of central venous catheters.²⁴ A similar level of evidence does not yet exist for CHG as a surgical skin prep, but it is logical to expect that the superiority of CHG would hold true here as well. Thus, favoring CHG as a skin prep is advisable, provided that there is no risk of the solution entering the eyes or ears.

Handwashing

Several options exist here as well. Iodophor and CHG solutions are available and are commonly used with scrub brushes for a specified period of time. Ten minutes has been traditional, but there is no evidence to support this ritual; the U.S. Centers for Disease Control and Prevention (CDC) recommends a duration of 2 to 5 minutes.²³

In addition, waterless hand cleansers have recently come into more widespread use. Typical of these is Avagard (3M, St. Paul, MN). It consists of ethyl alcohol 61% w/w, CHG 1%, and a mixture of skin conditioners and fragrances.

Another category of hand cleaners includes water-aided, brushless formulations. Triseptin (Healthpoint Ltd., Fort Worth, TX) is one of these. It also contains ethyl alcohol 61% w/w, as well as a proprietary formulation of conditioners and fragrances. A large RCT has demonstrated equivalence of an aqueous alcohol-based hand rub with traditional handwashing techniques²⁵; the results of this trial can safely be extrapolated to the products available on the U.S. market.

Because most SSIs arise from the patient’s skin flora, the use of one type of hand cleanser over another is largely left to individual preference. If there is a difference amongst the different agents available, it is likely quite small and has yet to be proven. Guidelines from the CDC regarding this are available.²³

First, artificial nails should be avoided. They can harbor micro-organisms and predispose gloves to tearing. A series of Serratia marcescens wound infections have been traced to a surgical team member with artificial nails.²⁶ Second, fingernails should be kept short and neat. Third, cleaning under the nails is recommended, as is removal of jewelry on the hands and arms, but the scientific support for this is overshadowed by the clear theoretical basis on which these recommendations are founded.

Surgical Gowns

The U.S. Occupational Safety and Health Administration (OSHA) requires that gowns have a minimum level of strikethrough resistance. All of the commercially available, disposable surgical gowns in the United States meet this standard. Porous cloth gowns have generally been eliminated because they do not offer this protection to the patient or to the surgeon.

Some have recommended that gowns and gloves be changed every 1.5 to 2 hours during lengthy cases.²⁷ Although this recommendation is sensible, the cost of this practice must also be considered.

Double-Gloving

The practice of double-gloving has been advocated as a protection to both the medical staff and the patient. For the staff, the risk of transmission of a communicable disease, such as HIV or hepatitis, is likely reduced, as is the risk of gross contamination through a compromised glove. Not only is perforation of two gloves more difficult than one, but in the event of a needlestick, the “squeegee effect” of the second glove has been shown as well.²⁸

Double-gloving may reduce the risk of SSI as well. Perforation rates of greater than 20% have been reported for the primary operator.^29,30 This puts the surgeon’s skin in direct contact with the surgical bed, increasing the chance of bacterial contamination for the patient and creating a hazardous exposure for the surgeon.

Although the superiority of double-gloving in prevention of SSI has not been shown, a perusal of a recent Cochrane review reinforces the theoretical benefit of double-gloving to the patient.³¹ When studying the risk of perforation of the inner glove (which is the only glove for single-gloved surgeons, and the inner for those double-gloving), there is a clear trend across the 18 studies reviewed toward a lower likelihood of perforation for those who were double-gloved.³¹

It is important to remember here the axiom that the absence of proof of a benefit is not proof of absence of said effect. In addition to middling clinical evidence, there is a strong theoretical rationale for double-gloving to protect the surgeon and the patient. The cost is minimal, and the potential benefits to all parties are significant.

Antimicrobial Adhesive Drapes

Antimicrobial adhesive drapes, such as Ioban (3M), are commonly used. These consist of an adhesive-backed plastic film that is impregnated with an iodine-containing compound. They have been shown to reduce bacterial contamination of wounds³² but have not reduced SSI rates in a prospective trial.³³ However, because of their insignificant cost (about $7 US), their use remains at the discretion of the surgeon.

Antibiotic Prophylaxis

The cornerstone for antibiotic prophylaxis is the assumption that every wound will be contaminated by skin flora. The goal is to reduce bacterial contamination to levels that can more easily be eradicated by host defenses. Although debate continues regarding the role of antibiotic prophylaxis in clean soft tissue surgery, there is general agreement that it is indicated whenever bone is incised or hardware is implanted, or when SSI would pose catastrophic risk; these criteria neatly circumscribe spinal surgery.

Timing and route of administration should follow the tenet of providing a tissue concentration of antibiotic that is greater than the minimum bactericidal concentration (MBC) of the organisms most likely to cause infection (i.e., Staphylococcus spp.) at the time of incision and maintaining same until the skin is closed. A first-generation cephalosporin (usually cefazolin) is most often given. Peak serum concentration of cefazolin, when given intravenously, is achieved within 5 minutes. The half-life is 1.5 to 2.5 hours; however, with doses commonly given (1 g), the MBC is generally maintained for 4 hours or more. Based on this, redosing cefazolin every 4 hours during the operation is indicated. This regimen also provides for adequate concentrations of cefazolin in clotted blood that remains in the surgical bed postoperatively. Patients allergic to penicillins may require vancomycin or clindamycin for prophylaxis. Vancomycin is also recommended by the CDC if a cluster of infections due to methicillin-resistant organisms is detected; however, no scientific cutoff has been determined for this, and its routine use as prophylaxis is discouraged.²³

Antibiotic prophylaxis, according to the guidelines of the CDC and others,^23,34 should be continued throughout the operation, but should be terminated not more than 24 hours postoperatively. It is important to remember that the use of prophylactic antibiotics can be associated with infectious complications and Clostridium difficile colitis.^35,36 There is no good evidence to support the practice of “drain prophylaxis” or the continuation of antibiotics until surgical drains have been removed.

Core Body Temperature

Maintenance of normothermia during a surgical procedure is perhaps one of the most important tools in preventing SSI. Hypothermia is easily caused by the combination of general anesthesia, patient exposure, and a cold operating room.

Hypothermia can cause impaired immune function, including reduced antibody production, decreased chemotaxis, and phagocytosis. Kurz et al. conducted an RCT of 200 patients undergoing major colon surgery.³⁷ Core body temperature was allowed to trend downward in the control group (n = 96; mean, 34.7°C) and maintained with warmed fluids and forced warm air heating blankets in the experimental group (n = 104; mean, 36.6°C). The infection rate in the former was 18.8% compared with 5.8% in the latter, a significant difference. Additionally, the normothermic patients tolerated solid food sooner, were discharged earlier, and had greater collagen deposition in the wound. Several other studies were not nearly as well-designed but showed similar results.^38,39

It is incumbent on the surgical team to assist the anesthesiologist in maintenance of normothermia throughout the case. Anecdotally, it seems easier to maintain temperature than it is to warm up an anesthetized patient who has been allowed to become hypothermic. This can generally be accomplished by keeping the patient covered until skin prep and the operating room warm until the drapes and warm air blankets are applied.

Intraoperative Hyperoxygenation

Providing hyperoxygenation during an operation has recently gained attention as a possible means of reducing SSIs. The theoretical backing is strong; however, the clinical evidence is mixed, and prolonged hyperoxia is known to have detrimental effects.

The rationale for hyperoxygenation is based on the positive role of oxygen in both the immune system and wound healing. Superoxide radicals are generated by neutrophils to carry out nonspecific killing, which is the body’s first line of defense against micro-organisms. An increase in the arterial partial pressure of oxygen, P_ao₂, should lead to a subsequent increase in the tissue partial pressure of oxygen, P_to₂. Furthermore, an increase of P_to₂ should lead to optimal collagen formation via the increased activity of prolyl hydroxylase.

In experimental studies the role of operative hyperoxia has proven uncertain. Greif et al. studied 500 patients undergoing major colorectal procedures.⁴⁰ Oxygen was administered at the fraction of inspired oxygen, F_io₂ = 0.3 (control) or 0.8 (experimental). This was continued throughout the procedure and for 2 hours postoperatively. The infection rate in the hyperoxic group was 5.2%, compared with 11.2% in the control group.

These results have yet to be repeated in a study of similar size and design. In another study of 160 patients, Pryor et al. found the infection rate of the hyperoxic group to be 25%, compared with 11% in the control group.⁴¹ This is the opposite of what the earlier study showed, and what would be predicted theoretically. However, this study’s design has been criticized on several grounds, including the small sample size and some significant differences between the control and experimental groups. Finally, Belda et al., in a study of 300 patients, found results similar to those of Greif et al.⁴²

This is an area of study that shows some promise, and hopefully further studies will be published. Nevertheless, the results are far from consistent at this point, and with the known risks of hyperoxia, it is difficult to justify the routine use of this technique for the prevention of infection.

Surgical Drains

Few subjects spark as much controversy among surgeons as the use of surgical drains. Advocates point out the theoretical basis that drains remove blood (itself an excellent culture medium), remove bacteria, and prevent hematoma formation. Opponents cite local immunosuppression and the role of drains as a conduit for bacteria.

At all events, no convincing evidence proves that drains alter the infection rate, one way or the other, in spinal surgery. There is some argument from other disciplines that drains can increase the infection rate, at least drains of the open (i.e., Penrose) variety.^27,43–45 However, this must be taken with a grain of salt, as direct interdisciplinary comparisons cannot necessarily be made.

General surgeons and others who work in the abdomen and pelvis are afforded several luxuries that are not available to spine surgeons. First, they generally work in well-defined tissue planes. This is not the case with dorsal approaches to the spine, which involve a subperiosteal dissection. Second, their dissections along clean tissue planes allow for optimal hemostasis. In the case of spine surgery, incised bone always bleeds, and muscle often does. And third, a hematoma of 200 mL in the peritoneum or retroperitoneum is a nonevent. This is hardly the case for a hematoma of similar size in a subfascial laminectomy bed.

Because of the inherently low infection rate in spinal surgery, and the likely minimal impact of drains on the infection rate, the number of patients needed for an RCT are substantial. Illustrative of this point is an article by Brown and Brookfield.⁴⁶ Eighty-three patients were enrolled and underwent lumbar procedures “larger than single-level unilateral decompressions.” Drain placement was randomized. However, there were no infections. In fact, the only significant result was a higher temperature on postoperative day 1 in the drained group. The significance of this is unclear. In their discussion, the investigators eloquently summarized the reason definitive evidence for or against the use of drains in spinal surgery is unlikely: “We used data on rates of infection from previously published drain studies to estimate the sample size necessary to achieve a power of 0.80. Based on these numbers, it was determined that 9,539 patients would have to be randomized into two groups to determine a true statistically significant difference. We discontinued the study after enrolling a more realistic sample size of 83 patients.”

The most reasonable alternative, and the place where the spine community should look for answers to this question, is probably the orthopaedic literature, especially that related to total joint arthroplasty. Similarities include the need to dissect through muscle, the necessary incision of bone in every case, similar blood loss, and the implantation of hardware.

In a meta-analysis of RCTs in the orthopaedic literature, 3689 wounds were studied in 3495 patients.⁴⁷ All patients underwent total hip or knee arthroplasty, and the groups were evenly divided with regard to drain placement. Although there was a trend toward lower infection rates in the group with drains, it was not significant. What was significant, however, was an increased transfusion requirement in the drained patients.

Any good evidence for or against drains is offset by the evidence on the other side of the argument. However, general words of caution always apply. Drains are not a substitute for good hemostasis. They should not be left as a matter of routine, but should be used when indicated. And if drains are employed, one should realize that the studies cited here all utilized closed-suction–type drains, and the results likely do not apply to open (i.e., Penrose) drains.

Operating Room Hygiene

In the United States, the design of operating rooms, including size, layout, and air-handling systems, is dictated by the recommendations of the American Institute of Architects. Surgeons have little control over these factors. They can, however, control what happens inside the operating room, and they can empower the operating room staff to exert such control as well. Much of what we do is dictated by expert opinion and tradition. We should not expect any more solid evidence to support this, as the ethical justification for such scientific trials would be suspect, at best. Summarized here are some of the protocols that are violated most frequently.

Traffic should be kept to a minimum during a procedure. Although the surgeons are generally present in a case for the duration, this is not true of the others in the room. Allowing breaks for the scrub and circulating nurses and the anesthetist is a matter of established routine, but it still increases traffic flow into and out of the operating room. To be avoided is the practice of a person entering the room to ask if breaks are desired. This can be accomplished by use of the telephone. Certainly the restocking of surgical and anesthetic supplies should not take place during a case.

The Association of periOperative Registered Nurses (AORN) maintains a list of operating room standards and recommendations.⁴⁸ This set of clear, straightforward guidelines should be familiar to all practicing surgeons. Among other things, they advise against covering sterile fields. This is because the cover can create air currents during removal, which could carry bacteria or contaminants onto the field and also because nonsterile parts of the cover may brush against the field during removal.

One of the most important aspects is maintenance of the sterile field. AORN recommends a boundary of 12 inches between nonsterile personnel and the sterile field; asking for a greater distance in order to achieve the specified 12 inches seems a perfectly reasonable approach.

Glucose Control

The importance of tight glucose control in the postoperative period cannot be overstated. Although most of the evidence here comes from the critical care, general, and cardiothoracic literature, it clearly shows that poor control of serum glucose levels is associated with infectious complications.

In a landmark study, 1548 ICU patients (mostly postsurgical) were randomized to one of two regimens of glucose control.⁴⁹ In the control group, patients were started on an insulin drip if the serum glucose measured 215 mg/dL or higher, and the same was adjusted to maintain serum glucose of 180 to 200 mg/dL. The study group was started on an insulin drip for any reading greater than 110 mg/dL and titrated to maintain 80 to 110 mg/dL.

The amount of insulin administered to the study group was over twice what was given to the controls, and 98.7% of the study group received insulin at some point. (It is noteworthy that this study looked at all ICU patients, and not just diabetics.) Compared with the control group, the patients receiving intensive insulin therapy had a decreased chance of requiring ventilator support or hemodialysis; a 46% reduction in the number of patients who became bacteremic at any point; and a lower mortality rate if the ICU stay extended beyond 5 days.

A dose-response–type relationship between postoperative blood glucose levels and deep sternal wound infections has been reported in several studies of cardiac surgery patients.^50,51 These levels seem to be most important in the first 48 hours postoperatively.