The Conundrum of Cervical Adenopathy

Two new studies may help identify the pathogen for cervical adenopathy in children, an often frustrating condition with numerous and divergent potential causes.

Most of us are comfortable treating certain presentations, such as tender, erythematous, anterior cervical lymphadenitis, for which the cause should be Staphylococcus aureus or Group A streptococci (GAS). We usually prescribe antibiotics and expect improvement within 10 days. A second presentation with a simple disposition is the nontender, unilateral, submandibular node with a purplish color and thinning of the skin over it. Usually, swelling has been present for weeks but has only recently developed the new color and thinning skin. This latter condition is usually nontuberculous mycobacterium (NTM) and is cause for referral to a surgical colleague.

But we have less certainty when predicting the cause of firm, unilateral, nontender adenopathy persisting after the usual 10-day course of antibiotics for initially tender adenitis, or when the child presents initially with nontender, unilateral, soft, mobile cervical adenopathy. Such nodes will often resolve spontaneously if we simply observe them for another 2–3 weeks. Those that don't are often the result of NTM. During the added observation, we often begin a diagnostic quest for such diverse causes as Epstein-Barr virus, Bartonella species, tuberculosis, NTM, Toxoplasma, or in some areas Histoplasma.

If our diagnostic quest is unrevealing and the node persists beyond 4–6 weeks, we usually refer for surgical excision to obtain histopathologic evaluation, plus cultures for mycobacteria, fungi, and conventional bacteria. Where available, fine-needle aspiration can fulfill the same objective. However, parents may become frustrated with the “tincture of time” approach and the slow pace before deciding on surgical referral. Two recent studies offer hope that we can make more rapid and accurate decisions.

A Dutch group reviewed the sensitivity and specificity of tuberculin skin testing (TST) in identifying NTM in Netherlands-born children (median age 48 months) with cervicofacial lymphadenopathy. Of these, 15 had immigrant parents, but none had traveled to TB-endemic areas (Clin. Infect. Dis. 2006;43:1547–51).

A total of 112 received a diagnosis of an NTM, confirmed by culture, polymerase chain reaction testing, or both. The infections were caused by Mycobacterium avium-intracellulare in 83 patients, by M. haemophilum in 21, and by other NTM species in eight. Nonmycobacterial lymphadenopathies were present in 46 of the children, including Bartonella henselae in 20, streptococcal infections in 14, staphylococcal infections in 11, and tuberculosis in one.

Using a cutoff of 5 mm of induration to define a “positive” NTM result, the TST's overall accuracy in detecting NTM was 0.84 (sensitivity, 70%; specificity, 98%; positive predictive value, 98%; and negative predictive value, 64%). Although 10-mm induration is the usual cutoff for M. tuberculosis, this study's data suggest that a 5–10-mm cutoff can predict NTM in previously healthy children with cervical lymphadenopathy, particularly where endemic TB is not common.

In my practice, I have used 5–10 mm of TST induration to strengthen my sense that NTM was the agent for nontender, unilateral cervical adenopathy of greater than 2 weeks' duration. This study increases my confidence with this approach.

In general, because community-acquired methicillin-resistant S. aureus has become more common, my algorithm is to first use 10 days of clindamycin, for antistaphylococcal and -GAS coverage. If the node persists but is asymptomatic and stable in size, I begin a work-up during a 2–4-week observation period before considering surgical referral. Unlike adults, in whom nonpainful neck masses are cancer until proven otherwise, prepubertal children rarely have this presentation for cancer, and so we can wait to see if slow resolution occurs.

My work-up includes serology, TST, and a chest x-ray (CXR). The CXR might show hilar adenopathy, histoplasmosis, or—rarely—tumor. Serology seems helpful only in the approximately 10% of these cases that turn out to be something other than NTM, such as Toxoplasma, Bartonella, or Histoplasma. Unfortunately, this work-up plus observation usually adds up to 6–8 weeks before surgical referral.

Clinical clues can sometimes hasten this process. If the node develops that rock-hard or adherent feel to palpation, I refer earlier to rule out the rare cancerous node. If the node breaks down to drain, or evolves into the purplish node noted above, prompt referral is warranted because of likely NTM.

Barring those or other new clues, can we more quickly feel confident in earlier surgical referral? The Dutch data suggest that we could postpone serologic evaluation until after a TST and perhaps a CXR. Here's how I would proceed, based on the size of the TST induration:

▸ Less than 5 mm and normal CXR. Proceed with serology and use standard observation before surgical referral.