Figure 1

Figure 1. Flowchart of patients included in a study of TB diagnostic delays and treatment outcomes among patients with COVID-19, California, USA, 2020. TB high-burden counties included were Alameda (excluding the city...

Using surveillance records of TB and COVID-19, we used name-based probabilistic matching to find persons with diagnosed TB and COVID-19 in California (5). We abstracted records for 58 patients who had TB disease diagnosed in 2020 and COVID-19 diagnosed within 120 days and who resided in 6 local health jurisdictions with high TB burdens: Los Angeles, San Diego, Santa Clara, Orange, Alameda, and Sacramento Counties (Figure 1). We captured TB and COVID-19 symptom profiles and timing, chest imaging results, TB diagnostic testing, and hospitalizations from TB program, hospital, emergency department, and outpatient records, and from death certificates. We also obtained PCR and antigen-based COVID-19 test results, including negative results, beginning on March 9, 2020. We performed statistical comparisons by using 2-sided χ² or Fisher exact tests for categorical data and Wilcoxon 2-sample tests for continuous data (α = 0.05) (Appendix).

Among 58 patients with COVID-19 and TB, 51 had pulmonary or pleural TB disease. The median time from symptom onset to TB diagnosis was 29.0 (interquartile range [IQR] 5.0‒95.0) days. Twenty-two (43%) patients had a diagnostic delay of >30 days (median 95.0 [IQR 60.0‒117.0] days) between TB symptom onset and first TB clinical consultation (Table 1; Appendix Table). Patients with diagnostic delays had indicators of more severe TB, such as acid-fast bacilli smear–positive sputum, cavitary imaging results, or disseminated pulmonary disease, than patients without diagnostic delays (86% vs. 55%; p = 0.02). Diagnostic delays were marginally more common among persons with COVID-19 diagnosed during periods of elevated incidence, considered the statewide 7-day average COVID-19 incidence rate of >15 cases/100,000 population, than persons diagnosed at periods without elevated incidence (82% vs. 55%; p = 0.05).

Figure 2

Figure 2. Timeline of 8 patients included in a study of TB diagnostic delays and treatment outcomes among patients with COVID-19, California, USA, 2020. Symptom onset is the date the first symptoms...

Among 51 patients with COVID-19 and pulmonary or pleural TB, 8 (16%) had >1 missed opportunity for TB diagnosis. We defined a missed opportunity as a documented clinical encounter in which a person with TB risk factors (e.g., experiencing homelessness or being non–US-born, in a correctional facility, or HIV-positive) had TB-specific symptoms but no TB diagnostic testing. TB-specific symptoms were hemoptysis, weight loss, or cough >3 weeks, or chest imaging of cavity, tree in bud pattern, pleural effusions, nodules, miliary, or upper lobe infiltrate; TB diagnostic testing included acid-fast bacilli smear or Mycobacterium tuberculosis nucleic acid amplification test. The median time between the first missed opportunity and start of TB diagnostic testing was 62.5 (IQR 33.5‒70.5) days. Five (63%) missed opportunities occurred during periods of elevated COVID-19 incidence, and 4 (50%) patients had COVID-19 testing (2 COVID-19–negative and 2 COVID-19–positive) instead of TB testing at the clinical encounter where the missed opportunity occurred (Figure 2).

Among the 58 patients, 51 (88%) were hospitalized >1 time (Table 2). Among 73 hospitalizations (average 1.3 per person), 35 (48%) were related to TB disease alone, 23 (32%) to indistinct (i.e., concurrent) TB and COVID-19 disease episodes, and 15 (21%) to COVID-19 alone. All 6 in-hospital deaths occurred during COVID-19–associated hospitalizations. The median overall hospital stay was 12 (IQR 7‒21) days and was similar across all 3 disease-associated hospitalizations, even when we excluded in-hospital deaths.

Two patients did not start TB treatment because they died before TB diagnosis. Of the remaining 56 patients, 42 (75%) completed TB treatment within 12 months, 5 (9%) completed treatment in >12 months (including 1 case with rifampin resistance), 1 refused treatment, and 8 (14%) died before completing treatment. Overall, 10 (17%) patients died. Local TB programs determined that 3 (30%) deaths were definitely related to TB, 5 (50%) were possibly related, and 2 (20%) were probably not related. Of the 8 deaths definitely or possibly related to TB, 3 (38%) had TB and COVID-19 listed as contributors on the death certificate, 4 (50%) had only COVID-19, and 1 (16%) had neither term listed.

Delays in TB diagnosis or documentation of a missed opportunity to diagnose TB were more frequent during periods of elevated COVID-19 incidence, potentially because of pandemic-related staff and health system disruptions and community transmission mitigation policies (6,7). Approximately 1 in 6 persons in our sample had a documented clinical encounter where TB diagnostic evaluations could have been initiated earlier, which was consistent with literature published before the pandemic (8). Delayed diagnosis could lead to increased TB transmission and worse TB outcomes; in this analysis, delayed diagnosis appeared to be associated with more advanced TB, suggesting more infectiousness.

In our sample, 83% of patients had >1 TB-related hospitalization, which is higher than the prepandemic frequency of TB-associated hospitalization in California, which previously was reported as ≈50% (9). This finding might have been influenced by the slightly older age distribution of this patient cohort (median 57.5 [IQR 42‒76] years) compared with pre-pandemic TB patients (median 56.0 IQR 35‒70 [years]) in California from 2017‒2019 (5). The median duration of TB-related hospital stays did not change compared with historical TB hospitalizations in California (9).

TB treatment completion appeared consistent with the pre–COVID-19 era in which ≈75% of patients completed TB treatment within 12 months (10). As we previously described, the proportion of deaths among TB patients with COVID-19 was higher than for TB patients in the recent pre-pandemic period (5). Most (77%) deaths were definitely or possibly TB-related but TB attribution on death certificates had poor correlation with detailed retrospective review, as has also been historically described (11). Thus, death certificates are unlikely to yield accurate estimates for deaths related to TB and COVID-19 co-infections.

Limitations of this study include use of observational data and lack of a comparison cohort of persons with TB who did not have COVID-19 in 2020 but had the same detailed clinical data. Our small sample size also precluded robust subgroup comparisons. The 6 participating TB programs represented 55% of California’s population (12), 53% of reported COVID-19 (13), and 66% of the state’s reported TB in 2020 (1). Nonetheless, our findings may not be generalizable to all areas of California or to other US regions.

In summary, delays in TB diagnoses continued to occur and the frequency of TB-related hospitalizations was higher for patients diagnosed with both TB and COVID-19 during the pandemic than historically observed in California. Nonetheless, the proportion of TB patients with COVID-19 completing treatment within 12 months was similar to persons with TB in the prepandemic period, suggesting TB programs managed to maintain TB treatment standards despite redirection of staff and resources. Pursuing a diagnostic workup for persons at risk of developing TB disease, even when a viral respiratory pathogen is widely circulating, remains critical for reducing TB-related illness in California.