Factors predicting outcome in cervical lymph node tuberculosis: insights from a Tunisian case series

Study design and participants

This is a retrospective study conducted over a 3-year period, from 2019 to 2021. We reviewed clinical data from patients with cervical tuberculous lymphadenitis treated in the Head and Neck Surgery Department at La Rabta Hospital in Tunis. The diagnosis of tuberculous lymphadenitis was based on histopathological findings in all patients.

Initially, all patient records diagnosed with cervical tuberculous lymphadenitis were extracted. Patients who met the exclusion criteria were subsequently removed from the study. The exclusion criteria were as follows:

The minimum follow-up period was 12 months after the completion of treatment.

Evolution assessment

The progression was assessed based on two criteria: the duration of treatment and the need for a second surgical procedure to complete the treatment. These two criteria were used to divide the population into two groups: those with a favorable outcome (G1), defined as a treatment duration of less than 9 months without the need for surgical treatment and a lymph node size of less than 1 cm at the end of treatment, and those with an unfavorable outcome (G2), if the treatment lasted more than 9 months and/or if a second surgery was required.

The limit of 9 months was chosen to assess prognosis because, in some cases, the treatment was maintained for more than 6 months as imaging was not immediately available to control the lymph node size, or in the presence of residual adenopathy.

Recurrence was defined as the reappearance of tuberculous lymphadenitis or an increase in the size of lymphadenopathy after the completion of treatment, confirmed by bacteriological, molecular, cytological, or histopathological evidence. The minimum period after treatment completion required to define recurrence was 6 months of remission.

Statistical methods

Data analysis

The clinical profiles of the enrolled patients were described using descriptive statistics, including mean values, standard deviations, medians, and interquartile ranges for continuous variables, and proportions for categorical variables.

An analytical study was conducted to compare two groups of patients to identify factors potentially associated with favorable or unfavorable disease progression. Epidemiological, clinical, biological, radiological, and therapeutic factors were evaluated.

Univariate analysis

Multivariate analysis

A multivariate analysis was conducted using a logistic regression model (or Cox regression for survival data). Variables with p < 0.2 in univariate analysis were included. The results were expressed as Odds Ratios (OR) with 95% Confidence Intervals (CI).

Statistical analysis was performed using SPSS version 22. A p-value < 0.05 was considered statistically significant.

Descriptive analysis

A total of 161 cases of cervical lymph node tuberculosis were initially diagnosed, but only 102 patients were ultimately included.

The group included 32 men and 70 women, with a sex ratio of 0.46. The median age was 34.5 years (range: 8-83). Sixty-one patients (60%) were under the age of forty. Only 22 patients (21.5%) had a special medical history (hypertension, diabetes, pulmonary tuberculosis, autoimmune diseases: Systemic lupus erythematosus, Behçet’s disease, Crohn’s disease, thyroiditis). No cases of known human immunodeficiency virus (HIV) infection were reported. Thirty-nine patients (38.2%) had a history of consuming raw milk.

Tuberculosis was discovered during the exploration of systemic symptoms in 59 patients (57.8%), while 40 patients (39.2%) presented with neck swelling as the first complaint. Two patients were already under treatment for pulmonary and one patient for osteoarticular tuberculosis, and cervical lymphadenitis was detected on radiological examination.

On physical examination, 86 patients (84.3%) had multiple palpable nodes, mostly unilateral (76 patients: 74.5%). The most affected lymph node levels were II and III (94 patients: 92.1%). The median size of the lymph nodes was 3 cm (range: 1.5 cm to 7.5 cm). In 65 cases (63,7%), lymph node size exceeded 3 cm. The skin was inflammatory in 28 cases (27.4%) and a cutaneous sinus tract was present in 9 cases (8.8%) ( Figure 1).

Figure 1. Clinical Features of Tuberculous Lymphadenopathy.

(a) Lymph node in level IIa (arrow), firm, mobile, painless, covered by healthy skin. (b) Multiple jugulo-carotid and spinal lymph nodes: level Vb node fistulized to the skin. (c) Right parotid region lymph node, soft, covered by healthy skin. (d) Multiple lymph nodes in levels II and III, soft, covered by healthy skin. (e) Multiple lymph nodes in levels II and III, soft, tender, covered by inflamed skin. (f) Lower spinal lymph node, soft and tender, covered by inflamed skin with an adjacent fistula.

Biological tests revealed anemia in 25 patients (24.5%) and lymphopenia in 22 patients (21.5%). HIV testing was performed in only three patients, and all tests were negative.

The tuberculin skin test was performed in 88 patients (86.2%). An induration or erythema diameter greater than 10 mm was observed in 75 patients.

All patients had cervical ultrasound (US) and/or computed tomography (CT) of the neck and thorax. The most prevalent radiological features were hypo echogenicity on US (53.9%) and necrosis on CT (40.1%) ( Figure 2).

Figure 2. Multiple Necrotic Cervical Lymphadenopathies on CT Scan.

Imaging revealed the involvement of extra cervical lymph nodes in 22 patients (21.5%), particularly in the mediastinum (15 patients), and associated pulmonary tuberculosis in three patients.

Fine needle aspiration (FNA) was performed in 85 patients (83.3%). Epithelioid granulomas with necrosis were observed in 33 patients ( Table 1 bis).

Microbiological testing was performed in 21 patients (20.6%). It revealed the presence of Acid-fast Bacilli on direct microscopic examination in 6 patients, but mycobacterium tuberculosis cultures were negative in all cases.

Molecular tests using the Xpert/MT PCR technique were carried out in 21 patients (20.6%) and were positive for Mycobacterium tuberculosis in 16 patients. Intermediate resistance to rifampicin was found in 7 patients.

Surgery associated with histopathological study was performed in all patients. The surgical procedures consisted of lymphadenectomy in 58 patients (56.8%), lymph node dissection in 9 patients (8.9%), and drainage of abscess with curettage/or fistula excision in 35 patients (34.3%) ( Figure 3).

Figure 3. Surgical Procedures.

• 3A: Lymph Node Dissection. • 3B: Lymphadenectomy.

A four-drug regimen (isoniazid, rifampicin, pyrazinamide, ethambutol) was systematically administered for two months. The following two-drug regimen (mainly Isoniazid + Rifampicin) was prescribed.

Ethambutol was maintained for more than two months, with an average duration of 4 months (range: 2–12 months) in 65 patients (63.7%).

Eleven patients received second-line antituberculosis drugs, consisting of fluoroquinolones.

By the second month of treatment, compared to the initial size of the lymphadenopathy, clinical improvement was observed in 51 patients (50%), stable lymph node size in 32 patients (31.3%), and a paradoxical reaction in 15 patients (14.7%).

The outcome was considered favorable in 56 patients (G1) and unfavorable in 46 patients (G2).

In G2, two patients required surgery with a total treatment duration of less than 9 months, 22 patients received anti-tuberculous treatment for more than 9 months, and the remaining 22 patients underwent surgery along with prolonged treatment.

Tuberculosis recurrence was observed in 12 patients (11.7%), with a median delay of 12 months (range: 7–63 months). These patients were predominantly G2 (11 patients: 91.6%).

Statistical analysis

Univariate analysis (Table 1 Underlying data⁷) revealed that the epidemiological characteristics and medical history were not significantly associated with any specific disease progression.

Analysis of physical examination data showed that pathological skin conditions (inflammatory skin/fistula) were significantly associated with unfavorable disease progression. No biological criteria were found to influence the outcome. Regarding imaging data, significant association was found between lymph node necrosis on CT and unfavorable disease progression. When the initial procedure involved abscess drainage, patients were more likely to experience an unfavorable outcome. However, if the initial procedure was lymph node dissection, patients were twice as likely to have a favorable outcome. The presence of intermediate resistance to rifampicin did not influence disease progression. Prolonged ethambutol prescription beyond the first two months of treatment was significantly associated with a favorable disease outcome.

Patients showing improvement by the second month were more likely to achieve a favorable outcome. Multivariate analysis ( Tables 2 and 3) revealed that pathological skin conditions, necrosis on CT, abscess drainage, occurrence of a paradoxical reaction, and lack of improvement by the end of the second month of treatment were independently associated with an unfavorable outcome. On the other hand, healthy skin adjacent to the lymph nodes, initial lymph node dissection, favorable evolution at the end of the second month of treatment, and prolonged ethambutol prescription beyond two months were factors independently associated with a favorable disease outcome.

All data supporting these findings are publicly available through Zenodo (DOI: 10.5281/zenodo.15351592)⁷

Descriptive analysis

Among the 102 study participants, most of the patients (61, or 60%) were under the age of 40 years. This finding is consistent with several published reports.^2,8,9 Furthermore, like other studies, a female predominance was observed, in contrast to the clear male predominance in pulmonary tuberculosis.^2,8 The reasons for this female prevalence remain unexplained.

Some patients had autoimmune diseases, including Crohn’s disease. Certain treatments, such as tumor necrosis factor (TNF) antagonists, are now known to affect the antituberculosis immune response.^10,11

The HIV-positive population was not represented in our study, as individuals with known HIV infection and tuberculous lymphadenitis are followed up in the infectious disease department at our institution. Moreover, only a few patients were tested for HIV. No conclusions can be drawn regarding HIV infection from our series, although many studies have reported a low HIV positivity rate in patients with tuberculous lymphadenopathy.^2,9,12–14 However, although Tunisia has a low incidence of HIV—albeit recently increasing¹⁵—and a low HIV positivity rate among patients with tuberculous lymphadenopathy in published series, we recommend HIV testing in cases of tuberculous lymph node involvement, particularly in the presence of an unfavorable outcome.

In our study, a significant percentage of participants (38.2%) reported raw milk consumption, similar to other studies, suggesting that Mycobacterium Bovis strains may contribute to tuberculous lymphadenitis in certain regions.^8,16,17

Most patients had unilateral and multiple lymph node involvement. In their studies, Mathiasen et al. and Qian et al. reported unilateral involvement in most cases.^14,18 However, cervical tuberculous lymphadenitis can be unilateral or bilateral, and all lymphatic levels of the neck can be affected.

Tuberculosis can present with various cytopathological patterns on fine-needle aspiration (FNA): granulomas with necrosis, granulomas without necrosis, and necrosis alone, with the first pattern being the most common and most highly suggestive of tuberculosis.¹⁹ In our series, 32.3% of cases showed epithelioid granulomas with necrosis on FNA; however, all patients were ultimately treated based on histopathological findings indicative of tuberculosis. This approach could be explained by the low positivity rate of bacteriological tests, the concern of misdiagnosing a malignancy, and, in some cases, the need for surgery in patients with abscesses or sinus tract formation.

Despite their high sensitivity, cytology and histopathology lack specificity for extrapulmonary tuberculosis, as granulomas can also be present in other diseases, such as nontuberculous mycobacterial infections.²⁰ FNA is considered the first-line of investigation for tuberculous lymphadenitis because it is cost-effective, safe and rapid, particularly in endemic countries with limited resources.^19–22 However, diagnosing tuberculosis solely on cytological or histopathological criteria should be approached with caution, and careful correlation with clinical and radiological data is essential to establish an accurate diagnosis.²³

The isolation of Mycobacterium tuberculosis remains the gold standard for a confirmatory diagnosis of tuberculous lymphadenitis, either through culture or by PCR testing of a sample from an affected lymph node.²⁴

The culture positivity rate was low in our study. In fact, the rate of culture positivity in lymph node tuberculosis ranges from 18% to 62%.²⁴ The Xpert® MTB/RIF assay was the only molecular test performed, conducted in 20.6% of cases, likely due to availability issues. This targeted PCR technique, by detecting the Mycobacterium tuberculosis complex, provides a rapid diagnosis with greater sensitivity than culture and high specificity. It also detects the rifampicin resistance gene, which was found in 7 of our patients (6.8%). Other multiplex PCRs, detecting more resistance mutations on target genes, exist but are unavailable in hospital laboratories in Tunisia. Next-generation sequencing, which enables the simultaneous sequencing of multiple genes, is becoming increasingly available in many countries. The widespread adoption of these techniques in resource-constrained countries could be crucial for implementing more effective anti-tuberculosis therapy.^20,25

Treatment of active tuberculosis involves a well-established two-phase approach: an initial two-month, four-drug regimen (Isoniazid, Rifampicin, Pyrazinamide, Ethambutol) followed by a maintenance phase with a two-drug regimen (Isoniazid, Rifampicin). The World Health Organization (WHO) recommends a treatment duration of six months for lymph node tuberculosis caused by sensitive strains. However, the treatment duration may extend beyond this period, with many studies reporting durations exceeding six months.^24,26,27 There remains controversy regarding the treatment duration for tuberculous lymphadenitis, as the course of the disease is unpredictable and there are no clear criteria for assessing cure.²⁷

In Tunisia, the official national practice guidelines recommend a treatment duration of six months, consistent with WHO guidelines. A Maghrebian recommendation published in 2015, considering the high incidence of Mycobacterium Bovis among cattle and the high consumption of raw milk products, suggested that the treatment duration could be extended to 12 months. It also recommended a prolonged prescription of ethambutol from 1 to 4 months following the initial four-drug regimen for patients with persistent tuberculous lymphadenitis after initial surgery, as Mycobacterium Bovis is naturally resistant to pyrazinamide.²⁸

Mycobacterium Bovis, present in unpasteurized milk and dairy products, is thought to reach the superior and anterior cervical lymph nodes through micro-ulcerations in the buccal mucosa.^16,17

In their Tunisian series, Ghariani et al. identified Mycobacterium Bovis as the strain responsible for tuberculous lymphadenopathy in 76% of bacteriologically diagnosed cases (79 cases).¹³ However, while the prevalence of Mycobacterium Bovis infection is likely high, the exact rate remains unknown in Tunisia. Improving molecular diagnostic tools would likely resolve this issue. The control of bovine tuberculosis, which is endemic in Tunisia, is crucial to reducing the incidence of human tuberculous lymphadenitis secondary to Mycobacterium Bovis strains.¹⁶

Statistical analysis

In our study, neither age nor gender was predictive of any particular outcome. Similarly, comorbidities did not appear to influence the progression of the disease. Seok et al. reported a significant correlation between younger age and residual lymph node enlargement (defined as lymphadenopathy greater than 10 mm in diameter with enhancement patterns of tuberculosis after six months of treatment).²⁷

Furthermore, no significant association was found between the initial size of the lymph node and disease progression. This contrasts with the literature, which associates a size greater than 3 cm with unfavorable outcomes.^29–31 This discrepancy could be explained by measurement errors or sampling issues. An initial lymph node larger than 3 cm may not be associated with poor progression if it was surgically removed during the initial diagnostic procedure.

Pathological skin changes were associated with an unfavorable progression. In contrast, Soriano et al. found that skin changes did not influence prognosis.²⁶

Regarding radiological data, our study revealed that necrosis on CT scans increased the risk of unfavorable disease outcomes. Zhang et al. also found that a larger area of necrosis was associated with a higher likelihood of poor treatment prognosis. They suggested that necrosis reflects high bacterial virulence, weak immunity, or severe allergic responses, and that its increase is associated with poor response to drug therapy.³²

A statistically significant correlation was observed between surgical drainage of a cold abscess and unfavorable disease progression. This could be explained by the fact that cold abscesses may form distally from the original persistent lymphadenopathy. However, initial neck dissection was associated with a favorable outcome.

Liu et al. suggested that surgical removal of cervical lymph node tuberculosis could shorten the time for anti-tuberculosis drug treatment.³³ Tahiri et al. found that biopsy with subtotal excision was associated with an unfavorable outcome.³⁴

Surgical treatment remains debated in tuberculous lymphadenopathy. Although often difficult and risky, many studies support the theory that surgical treatment should be combined with medical treatment to manage lymph node tuberculosis.^30,33,34 Lekhbal et al. reported an association between lymphadenopathy greater than or equal to 3 cm and the need for surgery in the treatment of cervical lymph node tuberculosis.²⁹

Further studies are needed to evaluate the effectiveness of surgical treatment and establish standardized surgical protocols for lymph node tuberculosis.

Prolonged ethambutol therapy beyond the first two months was statistically significantly associated with a favorable disease outcome. This supports the hypothesis that Mycobacterium Bovis strains are frequently involved in tuberculous lymphadenopathy in Tunisia.^13,16

In our study, improvement at the end of the second month of treatment was significantly associated with a favorable progression. The two-month evaluation may predict the overall outcome and help clinicians manage the disease by conducting further investigations or modifying the therapeutic approach.

Limitations

The main limitation of the present study was its retrospective design. Additionally, the study was conducted during the COVID-19 pandemic, which led to the postponement or omission of many complementary exams.

Furthermore, our study was based on a sample that may not be representative of the entire population affected by cervical lymph node tuberculosis, especially since HIV-positive patients are followed in another department. Both microbiological and molecular diagnoses were insufficient. The surgical procedures performed for initial diagnosis were not uniform across all patients, making it difficult to compare outcomes.

In our study, a 9-month period was used to define the outcome; however, in most published studies focusing on outcomes, this period was set at 6 months.