The Incidence of Tuberculosis after a Measles Outbreak

Tuberculosis is one of humanity's biggest health burdens, especially in developing countries. In 2005, there were 8.8 million new tuberculosis cases, and a total of 1.6 million people worldwide died of tuberculosis [1]. Measles also killed an estimated 345,000 persons worldwide in 2005 [2]. There is a belief that measles activates tuberculosis, either by converting a latent infection to active disease or by aggravating an active one [3]. This is based on several epidemiological reports [4–8], and the plausible mechanism is the suppression of delayed-type hypersensitivity after measles.

In a 1976 literature review, however, Flick [3] found that the relationship between measles and tuberculosis was inconclusive. According to his review, earlier studies had strong deficiencies in data collection and methods, and there were no reports that offered valid support for the concept that measles activates tuberculosis. In addition, several data contradict the hypothesis [5, 8]. Although Flick [3] hoped for some new, more valid observations, there have been few reports regarding the relationship between measles and tuberculosis during the past 30 years. Furthermore, there have been no studies conducted outside Western countries. Therefore, the relationship between measles and tuberculosis remains inconclusive.

The burden of tuberculosis in Korea is still not small. The number of estimated cases of tuberculosis in 2005 in Korea was 45,102 (96 cases per 100,000 person-years) [1]. In addition, in Korea, there was a measles outbreak in 2000–2001. During the outbreak, >55,000 measles cases were reported [9]. We aimed to elucidate whether measles had effects on the incidence of tuberculosis after the measles outbreak by using data based on disease notification.

Methods. By law, Korean physicians are required to report cases of measles or cases of tuberculosis that they encounter. Using the data contained in the Korea Communicable Disease Surveillance System, we collected statistics on measles cases from 1 August 2000 through 31 December 2001, based on the date of onset of symptoms, for the measles case group and for patients enrolled as having had tuberculosis between 1 August 2000 and 31 December 2004. For privacy protection, the database was provided to us with part of the identification information deleted. Comorbidity between measles and tuberculosis was verified by matching patient identification through an electronic record-linkage method. In this step, persons with inconsistent data—those whose date of starting anti-tuberculosis therapy preceded or was within 1 month after the date of onset of measles symptoms—were excluded from the analysis. Demographic data for 2002 were collected, to provide basic characteristics of the general population at the midpoint between 2000 and 2004 from the Korea National Statistical Office (http://www.nso.go.kr/).

To compare the incidence of tuberculosis in the measles case group with that in general Korean population, standardized incidence ratios (SIRs) of tuberculosis were calculated as follows: the sum of the age- and sex-specific observed number of tuberculosis cases in the measles case group divided by the sum of the age- and sex-specific expected number of tuberculosis cases in the measles case group. The age- and sex-specific expected number of tuberculosis cases in the measles case group was extrapolated from the age- and sex-specific incidence of tuberculosis for the general population in 2002, which was based on the data provided by the Korean Tuberculosis Surveillance Center and the Korea National Statistical Office. The Korean Tuberculosis Surveillance Center has the registry of notified cases of tuberculosis, and the estimated number of total population was presented by the Korea National Statistical Office. The age- and sex-specific expected number of tuberculosis cases was calculated by multiplying the age- and sex-specific incidence of tuberculosis for the general population in 2002 (the midpoint of 2000–2004) by the sum of individual observation time in each age- and sex-specific measles case group.

The time interval between measles and tuberculosis was calculated by determining the time between the date of onset of measles symptoms and the date of initiation of anti-tuberculosis treatment. The individual observation time in the measles case group was determined either as the time interval between measles onset and tuberculosis onset (if the patient was included in the measles group) or by determining the time between the measles symptom onset date and the last observation date (if the patient had measles until the last observation day). The last observation date was set as 31 December 2004. Statistical analyses were performed using the SQL procedures in SAS, version 9.1 (SAS Institute), and Excel (Microsoft).

Results. The total number of reported patients with measles (measles case group) from 1 August 2000 through 31 December 2001 was 53,979. Among them, 5 cases were excluded from the analysis because of incomplete data for identification Among those 53,974 patients, 53.2% were male, and the median age was 9 years (range, 0.1–26 years). The total number of reported patients with tuberculosis from 1 August 2000 through 31 December 2004 was 173,609. Among them, 63.6% were male, and the median age was 43 years (range, 0.1–85 years).

Forty-seven patients with tuberculosis, including 8 (17.0%) proven by positive results of smear or culture, contracted both measles and tuberculosis. The incidence of tuberculosis in the measles case group was 47 cases per 210,451.8 person-years. The median age of patients in the measles case group was 13 years, and 24 (51.1%) were male. The median time interval between measles onset and tuberculosis onset was 829 days (range, 31–1462 days).

The overall incidence of tuberculosis in the measles case group, adjusted by sex and age, was significantly lower than that in general population (SIR, 0.73; 95% CI, 0.53–0.97). The lower incidence of tuberculosis in the measles case group was more evident in the male subgroup (incidence ratio, 0.66; 95% CI, 0.42–0.98), but in females, we could not find any statistical significance (incidence ratio, 0.81; 95% CI, 0.52–1.22). Among patients aged 0–9 years, the incidence ratio was 1.29 (95% CI, 0.42–3.01); among patients aged 10–19 years, the incidence ratio was 0.72 (95% CI, 0.52–0.98). When we limited the definition of tuberculosis to disease developing within 1 year (SIR, 0.53; 95% CI, 0.23–1.03) or within 3 months (SIR, 1.44; 95% CI, 0.39–3.70) after the onset of measles symptoms, the statistical significance disappeared (table 1).

View larger version:

• In this window
• In a new window

• Download as PowerPoint Slide

Table 1

Standardized incidence ratios of tuberculosis in the measles case group.

Discussion. In this study, the overall incidence of tuberculosis in the measles case group with a mean duration of 3.9 years was ∼30% less than that in general population, which contradicts the classic hypothesis that measles can activate tuberculosis [4–8]. However, we cannot conclude that measles reduces the risk of tuberculosis: any plausible biological mechanisms by which measles could reduce the risk of tuberculosis have not yet been elucidated. The complexity that derives from the fact that tuberculosis has 2 phases—acquisition of infection and progression from infection to active disease—also causes difficulties in interpreting the results.

There could be a preventive effect against tuberculosis in the measles case group. Most patients with measles were children, so we supposed that those parents who accessed the health care system because of their children's measles might have had more opportunity to obtain information or advice on health, including prevention of tuberculosis thorough bacille Calmette-Guárin vaccination. In fact, tuberculosis after measles is a concern of physicians [10]. However, we have little information to verify those health-related behaviors among the patients in the measles case group in the study.

Conversely, measles could speed up the progression from infection to active tuberculosis, which might support the classic hypothesis. If there were a pool of persons infected with tuberculosis and if measles accelerated the progression from infection to active disease, there might be fewer cases of infection that had not progressed to active disease immediately after the measles epidemic passed. However, the number might gradually increase again as newly infected cases were found. In fact, our study showed that the SIR between 1 and 3 months after diagnosis of measles was the highest, at 1.44; the SIR between 3 months and 1 year dropped to 0.36; and after 1 year, the SIR again increased over time. However, we could not find definite statistical significance.

There are limitations to our study, because the data collected were based solely on notification about communicable disease. First, it is unrealistic to expect that all of cases of measles or tuberculosis are reported, and some cases might not have been detected because of absent or mild symptoms. Second, there could be problems in defining cases of tuberculosis, particularly in children. Only 17% of tuberculosis cases were microbiologically proven in our study. Also, we had information on the date of initiation of anti-tuberculosis treatment but nothing on the date of onset of tuberculosis symptoms. Thus, there is a possibility that cases with delayed tuberculosis diagnosis, regardless of measles development, were included. Another weak point of our interpretation was that we could not adjust for factors influencing the infection or development of tuberculosis, such as family history, socioeconomic or nutritional status, bacille Calmette-Guárin vaccination status [11], and underlying diseases [12], because of lack of clinical information with the epidemiological data.

In conclusion, despite these limitations, we did not find a positive relationship between measles and tuberculosis after the 2000–2001 measles outbreak in Korea.

• Received June 2, 2007.
• Accepted November 8, 2007.

• © 2008 by the Infectious Diseases Society of America