8 Controversies in Cholesteatoma Surgery

Except in the eyes of the most dogmatic surgeons, each technique has a proper role in the management of cholesteatoma. The determining factors are the size of the mastoid, the extent of the disease, the presence of a complication (such as labyrinthine fistula), and the reliability of the patient for follow-up. Patients with sclerotic mastoids can be safely and easily treated with CWD, which results in a compact cavity with little risk of recurrence or postoperative drainage. Patients with large, extensively pneumatized mastoids are better treated with CWU when their disease permits, because postoperative cavity maintenance could be problematic. Extensive cholesteatomas are better treated with CWD because of their greater likelihood for recurrence; small, well-localized cholesteatomas can be adequately managed with CWU. Complications are usually better managed by CWD because of the overriding need to create a safe ear. Recurrences too are usually managed with CWD. Unreliable or migratory patients are better treated with CWD.

Fig. 8.1 Otomicroscopic view of right ear, showing probing of attic cholesteatoma.

Fig. 8.2 Computed tomographic scan of the same patient, showing an attic-antral cholesteatoma, enveloping ossicles.

Fig. 8.3 Canal wall up procedure, dissecting a cholesteatoma sac from the mastoid antrum.

In practice, the two areas that might be most problematic for a CWU approach are the sinus tympani and the anterior epitympanic space. The following case illustrates this approach in practice. A 40-year-old woman complained of intermittent drainage and hearing loss in her right ear. Examination revealed a dry crust over the pars flaccida; gentle debridement revealed a deep retraction pocket with squamous debris and scutum erosion (Fig. 8.1). An audiogram revealed a 40 dB conductive hearing loss. CT revealed soft tissue in the attic and antrum (Fig. 8.2).

At surgery, the lesion was first explored through the meatus. By probing the attic retraction pocket, it was seen that the sac extended back to the epitympanum. A CWU mastoidectomy was created with transmastoid atticotomy and facial recess (Fig. 8.3). The sac was incised in the antrum, its contents were evacuated, and it was delivered forward toward the middle ear. The incus was removed, and the malleus head was cut with a nipper. There was found to be more extensive involvement than indicated by the CT, with disease extending toward the anterior epitympanic space (Fig. 8.4).

At this juncture, the surgeon has to decide whether the cholesteatoma can be completely removed via a CWU approach, or whether the canal wall should be taken down. In this case, exposure of the anterior epitympanic space was compromised by a low-lying tegmen—the posterior bony canal wall could not be thinned any further without perforating it, so it was decided to take the canal wall down (Fig. 8.5) (another option in this case would be a transcanal atticotomy, removing the scutum to gain anterior exposure, and reconstructing it later with cartilage).

Fig. 8.4 Transmastoid atticotomy, with disease remaining in the attic. Vertical height of the tegmen (arrows) limits the view of the anterior epitympanum.

Fig. 8.5 After conversion to canal wall down, there is a clear view of residual disease in the attic.

Reconstruction of the canal wall in mastoid surgery is not a new concept. It may have been introduced by Mosher1 and was later utilized by Wullstein² and others. Different techniques and materials have since been used in an effort to decrease cavity size and prevent recurrent disease. The musculocutaneous flap described by Silvola and Palva,³ using temporalis muscle and fibrous tissue pedicled to the conchomeatal skin, is a widely popular means of reducing cavity size but often retracts and does not completely eliminate the mastoid bowl (the technique is described in Chapter 5). Bone pate, which is readily available in the surgical field, is commonly used for cavity obliteration, but the material may resorb or become infected. It might never completely solidify, or it may contain squamous elements that can lead to recurrence.⁴

Mercke⁵ introduced a technique of mastoid reconstruction reimplanting the posterior canal wall and obliterating the cavity with bone chips. This technique resulted in no recurrent cholesteatoma and reduced rate of residual, dry ear, and intact tympanic membrane (TM), and it improved upon the results obtained with CWU and CWD. Gantz et al,⁶ in 2005, published a series of 127 patients who underwent a similar procedure, with a recurrence rate of 1.5% at 4 years mean follow-up. This technique improved on the recurrence rate when compared with CWU, but wound infections were relatively common, requiring inpatient intravenous (IV) antibiotics. In addition, the procedure did not eliminate the need for a second-look operation, and the ossicular chain was routinely sacrificed at the first operation, with ossicular reconstruction delayed until the second stage.

McElveen and Chung⁷ introduced a “reversible canal wall down” technique, in which the bony canal wall is removed then refixated using bone cement. Although only five cases were described, the technique appears to provide good surgical exposure but results in a CWU-type cavity with the same potential for recurrence.

“Soft wall” reconstruction has also been attempted, first by Smith et al,⁸ and later by Takahashi et al,⁹ in which a portion of the posterior canal wall is removed for surgical exposure and later replaced by temporalis fascia and canal skin. Recurrence/residual rates were similar to those for CWD, but retraction of the canal skin into the mastoid bowl occurred in nearly 50% of cases. Dornhoffer¹⁰ advocated reconstruction of the canal wall with cartilage from the cimbum concha, and reported an 18% recurrence rate and excellent hearing results with this technique. Although these results are quite acceptable, rigid techniques appear to yield a lower rate of recurrence.

Several foreign materials have been used to obliterate the mastoid cavity. Hydroxyapatite (HA) granules and implants have shown favorable results, but HA cement may have a high rate of infection.¹¹ HA canal wall implants failed 25% of cases, usually because of middle ear infection rather than recurrent cholesteatoma.¹² Ceravital, a bioactive ceramic glass, has also been successful, but its use is cautioned against in patients with immunologic disorders or diabetes.¹³ Titanium implants have also been used to reconstruct the canal wall.¹⁴Hard materials such as these may eliminate the cavity problem but also may hinder the detection of the recurrent cholesteatoma. Because of the need for meticulous fitting and for vascularized soft tissue coverage, rigid implant materials have not gained widespread use, and concern remains over the potential for infection or extrusion.

We have used an alternative technique of reconstructed canal wall mastoidectomy in which the posterior canal wall is removed en bloc and then reused as a free bone graft to obliterate the sinodural angle (Figs. 8.6, 8.7, and 8.8). The surgical exposure is the same as for CWD, so the chance of leaving disease behind is very low. The attic remains exteriorized, thereby preventing recurrent disease from forming in that location. No foreign material or bone pate is used; therefore there is no increased risk of infection and little risk of reimplanting disease. The procedure still results in the creation of a cavity, but the size of the cavity is lessened, both by the reimplantation of cortical bone from the posterior canal, and by maintaining the mastoid tip. The bony reconstruction blocks the sinodural angle and results in a cavity no deeper than the lateral semicircular canal (Fig. 8.8). Using a conchomeatal flap in conjunction with the bony obliteration results in a round bowl with sloping edges, and a functionally small cavity with little surface area to collect debris.

Fig. 8.6 Attic cholesteatoma, with a low-lying tegmen.

Fig. 8.7 The posterior bony canal wall is removed en bloc after opening the facial recess and drilling troughs superiorly at its attachment to the tegmen, and inferiorly at the mastoid tip.

Fig. 8.8 Reconstructed canal wall. The bony implant is used to fill the sinodural angle (arrows), leaning it forward with its base against the facial ridge. The attic remains open.

Concerns remain about the yield of second-stage surgery and the adequacy of follow-up. In most studies, residual disease will be found in ∼20 to 50% of second-look operations, meaning that the majority of operations will result in a negative exploration. This argues for a selective approach to second-look surgery and makes limited, endoscopically assisted surgery more attractive. Imaging techniques such as CT and magnetic resonance imaging (MRI) have been tried in an effort to avoid routine surgical exploration, but their predictive value is not perfect. CT cannot distinguish between one type of soft tissue and another, so postoperative fibrosis and residual disease might look the same. MRI likewise lacks specificity, although newer protocols such as diffusion-weighted imaging might improve on this. Issues of expense and patient discomfort persist with MRI, however.

The principles of facial nerve monitoring have been outlined in detail in several sources.15^,16 A few devices are commercially available for monitoring the facial nerve during surgery. These devices contain two separate electric circuits, one to stimulate the nerve electrically, and the other to monitorthe activity of the facial muscles.

The nerve stimulator allows the surgeon to apply a small electrical current to the nerve during surgery. This provides immediate information about the anatomical location and the functional integrity of the nerve. Electrical stimulation is delivered through a probe. Stimulator probes usually operate in a monopolar mode, dispersing current radially from the tip of the stimulator probe (bipolar probes, which emit current between two probe tips, are also available, but they are less often used because the current must be oriented parallel to the nerve).¹⁷ Special surgical instruments also exist that can deliver an electric stimulus during dissection.¹⁸ Stimulators are usually set to deliver a constant current, and the current level can be selected on the device (constant voltage devices also exist but can deliver excessive current if the tissue resistance is low). Typically, a current level of 1 to 2 mA (mA) will be needed to stimulate the facial nerve through an intact bony covering, whereas a very small current (0.1 mA) will stimulate a bare nerve. Stimulation will spread bidirectionally along the nerve. In the setting of facial nerve injury, the nerve will continue to stimulate distal to the site of the lesion for up to 3 days, even after complete transaction. Stimulation proximal to the site of the lesion will give information as to the degree of injury; brisk stimulation at a low current level implies functional continuity of the nerve, whereas a high stimulation threshold or complete absence of stimulation implies a severe conduction block or a transected nerve. Constant-current stimulation is susceptible to shunting, wherein the current is dissipated by any electrolytic fluid (blood, saline, CSF) in the surgical field; the field should be dried with a suction before attempting to stimulate the nerve.

The nerve monitor most commonly employs electromyography (EMG) to detect contraction of the facial muscles during surgery, although there are motion sensors (“strain gauge sensors”) that have a higher threshold for detecting a response. The monitor produces an audible signal (as well as a visual oscilloscopic tracing) in response to facial muscle contraction. The audible signal provides live feedback to the surgeon during the dissection. The facial nerve monitor typically allows for recording from two or more channels during surgery. The usual recording array uses one pair of needle electrodes in the orbicularis oculi muscle and a second pair in the orbicularis oris, plus a common ground (Fig. 8.9). A second ground electrode for the stimulator circuit is inserted into a distant location, usually the shoulder.

Fig. 8.9 Electrode array for facial nerve monitoring using the Nerve Integrity Monitor (NIM) (Medtronic Corp., Minneapolis, MN). There are two needle electrode pairs (blue and red) for live electromyographic (EMG) monitoring of the orbicularis oculi and orbicularis oris. A common ground electrode (green) completes the circuit. A second ground electrode (white) is used for monopolar stimulation. A sterile stimulator probe can be used to deliver current to the nerve during surgery.

EMG monitoring produces various types of responses, and the surgeon needs to be familiar with these.15^,18 “Burst” responses are brief, nonrepetitive signals that occur in response to brief mechanical stimulation of the nerve (Fig. 8.10A). During surgery, burst responses will occur when one brushes against an exposed nerve with a blunt instrument. They serve as an immediate warning to the surgeon and do not usually forebode permanent injury unless the disturbing activity is forceful or sustained. “Train” responses are repetitive signals that occur in response to events that are longer in duration, such as mechanical traction on the nerve (Fig. 8.10B). The surgeon should cease whatever activity produced the train response to avoid permanent injury. Train responses also occur in response to thermal trauma, such as after saline irrigation. Direct electrical stimulation of the nerve elicits a “pulse” response, a synchronous repetitive signal (Fig. 8.10C). Each recording channel produces a different tone, so that the pulse responses are specific for the upper and lower divisions of the nerve.

Fig. 8.10 Types of facial nerve electromyographic (EMG) responses. (A) A “burst” response is a brief, nonrepetitive muscle contraction and occurs in response to a direct mechanical stimulus. (B) A “train” response is a repetitive, asynchronous response that occurs after mechanical traction or prolonged mechanical stimulation of the nerve. It can also occur after cold irrigation. (C) A “pulse” response is a repetitive, synchronous compound action potential elicited by electrical stimulation of the facial nerve.

In cholesteatoma surgery, facial nerve monitoring is most useful when dissecting disease away from an exposed facial nerve. As previously discussed, spontaneous bony dehiscences of the facial canal are common, and the incidence is higher in the presence of erosive disease. When a bare facial nerve is covered by cholesteatoma matrix or by granulation tissue, it is vulnerable to injury during the removal of the disease. The most common site of dehiscence is the tympanic segment of the nerve, just superior to the stapes (Fig. 8.11). The surgical strategy in these cases is to approach the nerve from posteriorly and superiorly, developing a plane of dissection between the intact bone and the mucosal layer, and lifting the matrix away from the surface of the bare nerve using a sharp instrument. Once this plane is established the disease will usually peel away, leaving the nerve sheath intact and undisturbed. In these cases, the nerve monitor will provide audible feedback if there is blunt or sharp disturbance (burst response) or if there is mechanical traction (train response). With careful surgical technique, no response will be elicited.

Fig. 8.11 Dissection of a recurrent cholesteatoma from bare facial nerve, tympanic segment, right ear, working from superiorly downward, using the stimulator probe as a dissecting instrument.

A second situation for monitoring may occur when one is drilling near the nerve, especially around the mastoid segment or second genu (Fig. 8.12). Although the nerve has a fairly constant anatomical course, variability exists in the medial-to-lateral position of the mastoid segment, approaching the digastric groove. Also, when opening the facial recess, it is desirable to shave away the bone anterior to the facial nerve to gain the best view of the sinus tympani (Fig. 8.13).

A third vulnerable area is anterior to the processus cochleariformis (Fig. 8.12), where the nerve may be exposed in the anterior epitympanic space. In these instances, the monitor may give a burst or train response if the drill or dissecting instrument brushes the nerve sheath. Injury may be avoided if the surgeon is aware of the slightly more lateral position of the nerve in this location.

In cases with distorted anatomy, or a nerve embedded in disease, the nerve stimulator may be used to map the course of the nerve anatomically. A monopolar stimulator, such as the Prass probe (Medtronic-Xomed Corp., Jacksonville, FL), can be used at a higher setting initially (eg, 2 mA) to verify that it is working, then the current setting can be turned down (0.5 mA) to gain more specific information about the position of the nerve. A bare nerve should stimulate briskly at 0.5 mA or less, whereas a nerve with an intact bony sheath will require more current, typically 2 mA or higher.

The stimulator can also be used after a difficult dissection to verify the integrity of the nerve. Stimulation with a supra-threshold current level proximal to the site of concern will produce a brisk response if the nerve is intact. An absent response or a very high threshold may be a reason to explore the nerve proximally and distally and to decompress the bony sheath.

Certain caveats should be noted during monitoring. Bipolar or monopolar electrocoagulation causes an artifactual response. This can be defeated using a muting circuit, but monitoring is suspended when muting is in effect. Irrigation can elicit a train response because of thermal stimulation of the nerve. Although this is not clinically significant, it may take several minutes for the response to die down and for normal monitoring to resume. Brief mechanical trauma using blunt dissecting instruments or drill will cause a burst response—this is moderately significant and should be taken as a warning of impending injury if the activity that caused the response is not stopped. An instantaneous trauma to the nerve, however, may not elicit any response if it is very short in duration; therefore it is possible to have a serious injury to the nerve, such as sharp transection, without any signal at all from the nerve monitor.

Fig. 8.12 Schematic view of the facial nerve (black), right ear. The three areas of vulnerability are indicated with red circles: at the facial recess, in the mastoid tip near the digastric (Di), and anterior to the process (Pr) in the anterior epitympanum. L, lateral semicircular canal; OW, oval window, Sig, sigmoid sinus; Tg, tegmen.

When the patient has preoperative facial nerve paralysis in the presence of cholesteatoma or chronic inflammatory middle ear disease, the facial nerve stimulator/monitor is helpful for assessing the degree and location of injury. However, if the nerve does not respond to stimulation preoperatively, the device will not necessarily signal any further trauma to the nerve that occurs during the course of surgery.

Fig. 8.13 Stimulation of the facial nerve in the facial recess (right ear).

Fig. 8.14 (C)hronic suppurative otitis media of the right ear with acute facial paralysis. (A) Dissection of granulation tissue away from the tympanic segment of the facial nerve through the attic, using the stimulation probe. (B) After completion of the dissection, the nerve is stimulated along its length to test its function.

After dissection of the disease, the exposed area of the nerve is visualized, and the bony covering is carefully removed proximal to the site of involvement (Figs. 8.14A,B). The presence of a response to electrical stimulation indicates that the nerve has functional integrity and will probably recover fully. The absence of a response, on the other hand, implies severe neuropraxia and a poorer prognosis, and the surgeon may want to decompress the nerve more widely to allow for progressive edema.

Postoperative Facial Nerve Paralysis

When the patient wakes up with immediate facial paralysis after chronic ear surgery, the patient is usually brought back to surgery for exploration and possible repair of the injured nerve. In this situation, the nerve monitor may be helpful for mapping the exact site of injury and indicating whether the disruption is complete or partial.

If the disruption is complete, the nerve edges should be freshened and the nerve reapproximated without tension using fine sutures (eg, 9–0 Prolene, Ethicon, Inc., Somerville, NJ). If the ends of the nerve cannot be brought together without tension, a nerve graft should be placed, using the great auricular nerve or sural nerve as a donor. If there is partial disruption, judgment must be used as to whether to allow the nerve to heal spontaneously or to repair the nerve primarily. If the injury involves less than a 50% cross-sectional area, or if the electrical response is preserved, the frayed edges of the nerve can be brought together with a collagen sheath, and partial recovery of function can be expected over a 6- to 12-month period.

Other Surgical Scenarios

If a middle ear tumor is suspected, the nerve monitor plays a valuable role because the anatomy may be altered and facial nerve involvement may not be predicted preoperatively. Middle ear adenoma (Fig. 8.15) may mimic inflammatory ear disease and may erode the bony covering and compress or distort the nerve anywhere along its course. Salivary choristoma (Fig. 8.16) is composed of heterotopic salivary tissue that can occur anywhere along the facial nerve. Facial nerve paralysis has been reported after biopsy of this unusual lesion, and the facial nerve monitor may help to prevent injury.¹⁹

Glomus tympanicum tumors arise along the course of the Jacobson nerve within the middle ear cleft; glomus jugulare (Fig. 8.17) and glomus vagale tumors may grow into the middle ear space via the jugular foramen. Removing these tumors usually requires a facial recess approach; jugulare tumors often require facial nerve transposition or a facial bridge technique to the hypotympanum.

Facial nerve schwannomas can occur along the entire length of the facial nerve, and while these rarely occur solely in the middle ear, middle ear involvement is common. Removal of a facial nerve schwannoma will usually result in facial nerve paralysis, and primary grafting should be planned. If the diagnosis is not made preoperatively, the facial nerve stimulator can be helpful for confirming the fact that the tumor is of facial nerve origin.

Fig. 8.15 Adenoma of the left middle ear and mastoid. The nerve stimulator probe is used to dissect the tumor away from the medial wall of the antrum approaching the facial nerve from posteriorly.

Fig. 8.16 Salivary choristoma of the middle ear, a rare entity consisting of heterotopic salivary tissue. The facial stimulator, seen here, can be used to be sure that the lesion does not stimulate before it is dissected. C, chorda tympani nerve.

On the other hand, it is easy to argue for the potential benefits of facial nerve monitoring. When properly used, the monitor gives valuable information about the position of the nerve, provides feedback to the surgeon during delicate dissection on an exposed nerve, and usually warns of impending injury. These benefits would be difficult to prove “scientifically” because the variables encountered in surgery do not lend themselves easily to quantitative measurement in an outcome study, but most surgeons would agree that the information provided by monitoring could help prevent a facial nerve complication. Because the incidence of facial nerve injury is very low, a very large study would have to be constructed to have sufficient statistical power. Facial paralysis will often result in medicolegal action, even in the absence of negligence or wrongdoing by the surgeon,20 and failing to use a nerve monitor may be perceived as a departure from current “best practices.”

Fig. 8.17 Glomus jugulare tumor, right ear. A canal wall down mastoidectomy with facial bridge technique has been performed. The tumor (Glo) has been exposed in the retrofacial area, above the jugular bulb (Jug). L, lateral semicircular canal; VII, facial nerve.

In real-world current practice, the facial nerve monitor is probably used by most surgeons in most cases of cholesteatoma surgery. Although some surgeons will employ it selectively, monitoring should probably be set up and used routinely to be most useful because it is impossible to predict when it will really be needed.21 In a policy statement,²² the American Academy of Otolaryngology–Head and Neck Surgery (AAO-HNS) has recognized “the proven efficacy of neurophysiologic monitoring of the facial nerve which may minimize the risk of injury to the nerve during surgical procedures in which the nerve is vulnerable.” But they also state that these “guidelines are not a substitute for the experience and judgment of a physician . . . and in no sense do they represent a standard of care.” This statement strikes a balance between the opinions of surgeons’ pro and con.

♦ Endoscopes in Chronic Ear Surgery

Endoscopes have proved to be a useful adjunct to surgery of the middle ear and mastoid.²³ The advantage provided by the endoscope is the ability to place the surgeon’s eye deep within the surgical cavity, and to be able to look around corners. This provides an advantage in areas such as the facial recess and anterior epitympanic space, which are difficult to visualize directly using the microscope. The disadvantages of the endoscope are the need to switch from the microscope to a different optical system (using a separate video setup or the naked eye), the presence of an object within the surgical field (with the potential to injure the stapes or facial nerve), the generation of heat (another potential source of injury), the loss of three-dimensionality, the loss of a free hand (or the need for an assistant to hold the scope), and the loss of visualization while operating with blood in the surgical field. These disadvantages make it unlikely that the endoscope will replace the microscope in ear surgery, but the enhanced visualization makes the endoscope useful nonetheless.

Fig. 8.18 Endoscopic view of left ear showing residual cholesteatoma in the attic and posterior tympanic sinus that was not seen with the microscope alone. C, chorda tympani; VII, facial nerve; S, stapes; Chol, cholesteatoma.

Smaller endoscopes provide a view of recesses within the middle ear space. A 2.7 mm, 30 degree scope can be inserted into the middle ear via a tympanomeatal flap, allowing the surgeon to look in the sinus tympani and epitympanum and reducing the possibility of leaving disease behind. This should translate into a lower rate of residual disease, and some authors have claimed improved results.24^,25 The ability to peer around corners may also permit less bone removal during mastoid surgery. It is possible that the endoscope may preclude the need for a facial recess approach for sinus tympani disease, allowing the surgeon to look in the sinus tympani and epitympanum and reducing the possibility of leaving disease behind (Fig. 8.18).²⁶

The endoscope may also allow for a less invasive approach for second-stage surgery.^27,28 Passing the endoscope through a small stab incision in the postauricular scar allows the surgeon to rule out residual disease in the mastoid cavity without the need for a large postauricular incision (but of course if disease is found, a formal revision operation will be necessary).

An endoscope-only technique has been devised for surgery of the middle ear and attic.²⁹ Using a 4 mm endoscope in the ear canal gives a wider field of view than the microscope can, and allows a complete transcanal atticotomy to be performed. For disease confined to the attic, this may provide thorough surgical removal through a more limited approach.

♦ Adequacy of Follow-Up

Adequacy of follow-up is an important factor in determining success or failure after cholesteatoma surgery. Follow-up is necessary to determine whether a patient has recurrent or residual disease, to clean and to manage problem cavities, and to assess postoperative hearing status in the short and long term. In clinical practice, however, many factors affect the reliability of follow-up, such as the patient’s distance from the surgeon, (changeable) insurance status, relationship with a referring physician, or moving out of the area. Without regular follow-up, the long-term results of cholesteatoma management are unknowable. For surgeons with out-of-town referral sources, this can be an even greater problem.

The tendency of cholesteatoma to recur lends particular importance to this issue. In a patient whose reliability is in question, CWD surgery is often the safest option. CWD has an intrinsically low recurrence rate, and residual disease can be spotted on exam and often dealt with in an office setting. Further, CWD exteriorizes the mastoid, so that complications (ie, spread of infection to the labyrinth, facial nerve, or intracranial structures) are unlikely to occur.

Several studies have taken up the issue as to what constitutes adequate length of follow-up. The recurrence rate for cholesteatoma increases with the length of follow-up in both children and adults.^30–32 Because cholesteatoma is thought to be more aggressive disease in children, the importance of follow-up is greater in pediatric patients. Recurrent cholesteatoma can form at any time, even years after the original surgical procedure (in contrast to residual disease, which is theoretically present at the conclusion of the surgery). Long-term follow-up studies have reported recurrent disease up to 15 years later, and giant cholesteatomas have been discovered after even longer intervals.³³ Although the likelihood of recurrence diminishes with time, there is no end point, practically speaking, at which follow-up is no longer needed.

The reporting of results can be influenced by the adequacy of follow-up. This can affect the surgeon’s perceived success rate in treating patients. Most studies traditionally used a standard rate calculation, where the total number of observed recurrences is divided by the total number of years of follow-up for the study. This method, however, fails to account for patients who drop from follow-up, or who have insufficient length of follow-up.³⁴ The use of Kaplan-Meier analysis adjusts for those who are “censored” from the study, and gives a truer estimate of disease recurrence overall.³⁵

♦ Conclusions

Stay updated, free articles. Join our Telegram channel

Join