Clinical and epidemiological aspects of cases of tuberculosis associated with diabetes in Salvador, Bahia, Brazil

In Brazil, the growing prevalence of chronic diseases, such as diabetes mellitus (DM), affects individuals of working age, increases the burden on social security systems, and perpetuates a vicious cycle of poverty and social exclusion. There are an estimated 6,317,621 diagnosed cases of DM in Brazil, corresponding to 5.2% of the population¹. The greatest economic impact is on healthcare, with increasing costs of treatment of the disease and its complications.

There is evidence that DM increases the risk of lower respiratory tract infections and the incidence of tuberculosis (TB) among patients with diabetes mellitus has been reported to be 2-3 times higher than that of the general population^2,3. This increased risk is the result of a reduction in phagocytic activity and in the production of T-cells and interferon-gamma, due to high levels of hemoglobin A1c in diabetic patients^2,4.

Brazil is among the 22 countries with the highest number of cases of TB. In 2014, the incidence rate of all forms of TB was 33.5/100,000 inhabitants⁵ and this was higher among immunocompromised individuals, such as those co-infected with the human immunodeficiency virus (HIV) and patients with DM^6. These co-morbidities favor the permanence of high incidence rates of TB and hamper TB control^2,3,6. Studies comparing the clinical characteristics of patients with these two comorbidities are scarce. The aim of the present study was to describe and compare the clinical and epidemiological features of patients with TB, with and without diabetes.

This cross-sectional study included diabetic and non-diabetic patients with TB, with no previous history of TB treatment or of multidrug resistance. All patients were undergoing treatment provided by healthcare services in Salvador, Bahia, Brazil, between August 2008 and December 2010.

New cases of active pulmonary TB were defined as individuals over 14 years of age living in Salvador and presenting with respiratory symptoms, who were investigated for TB at any of a number of health care units. Individuals had no previous history of TB, and all agreed to participate. Cases were defined as individuals diagnosed with pulmonary TB by the attending chest physician. All potential cases underwent a smear microscopy and culture for Mycobacterium tuberculosis, and individuals without a positive result or who did not live in Salvador were excluded.

DM was diagnosed based on the measurement of blood glucose levels using a glucose meter and lancing device, irrespective of the time of the last meal. A blood sample was obtained by pricking the lateral surface of a fingertip with the lancet, after cleaning the area. The blood sample was then applied to a reagent test strip, previously inserted into an Accu-Chek Advantage II meter.

According to the diagnostic criteria set by the American Diabetes Association⁷, DM is defined as fasting glucose levels ≥ 126mg/dl or the presence of classic symptoms of hyperglycemia or hyperglycemic crisis, with glucose levels of 200mg/dl. Fasting is defined as no caloric intake for at least 8 hours. The clinical criteria for diagnosis were polydipsia; polyphagia; and weight loss, pain and wounds that were difficult to heal. Patients with a previous diagnosis of diabetes and those who reported use of medication for DM were considered diabetic, regardless of the results of tests for glycemia.

Patients with active pulmonary TB and blood glucose levels above normal were informed of their results and referred to healthcare professionals at the Tuberculosis Control Program for the necessary therapeutic adjustments.

A standardized questionnaire was administered by trained nursing technicians, under nurse supervisions. This instrument was designed to collect data on the patient’s clinical and sociodemographic status, lifestyle, the presence or absence of symptoms of DM, the presence or absence of a Bacillus Calmette Guerin (BCG) immunization scar, and the number of scars. Furthermore, records were reviewed in an attempt to collect data of any previous history of diabetes.

A descriptive analysis of patient clinical and sociodemographic characteristics was performed. The STATA software program, version 12.0, was used for the analysis. The chi-square test and Student’s t-test were used to compare the results, which were expressed as means and standard deviations of the mean, to estimate the p-value of continuous variables that do not have a normal distribution, with a log-transformation for base 10.

The Internal Review Board of the Institute of Public Health, Federal University of Bahia, approved the study under reference (Process number 2007/012). All participants signed an informed consent form, according to the requirements of the standardized guidelines established by the Brazilian Ministry of Health’s Resolutions 196/96 and 304/2000. The committee determined that the study met ethical, legal, and regulatory norms and standards for research involving human subjects in Brazil, as well as international norms and standards.

Of the 323 cases of tuberculosis included in the study, 47 (14.6%) patients were classified as diabetic. Comparing the sociodemographic variables between diabetic and non-diabetic patients with TB, the former group were older (44.7% of the patients were ≥ 50 years of age), had a higher educational level (34% high school or above), and a lower family income (48.9% with < R$600 Brazilian reals per month). These differences were statistically significant. Conversely, no differences were found between the two groups in the number of male subjects, single subjects, or homeowners (Table 1).

There were no differences between the two groups in the following factors: previous history of tuberculosis, TB contact, previous treatment for tuberculosis, presence of fever or cough with sputum, hemoptysis, night sweats, weight loss, fever, coughing, or any other symptom, or the presence of a BCG immunization scar (Table 2). There was a positive and statistically significant association with slow wound healing in the group of diabetic patients [11 (23.4%) vs 12 (4.4%), p-value = 0.01)]; fatigue or leg pain [29 (61.7%) vs 118 (42.9%), p-value = 0.02)], and other symptoms (dyspnea\fatigue\asthenia [3 (6.4%) vs 11 (4%)]; body ache\leg pain\foot pain [8 (17%) vs 20 (7.3%)]; headache [3 (6.4%) vs 7 (2.5%) p-value = 0.04] (Table 2).

Furthermore, the log mean of duration of fever was significantly higher in the group with diabetes (25.3 ± 27.1 days) compared to the non-diabetic group (16.6 ± 22.7 days p-value = 0.01) (Table 3). The mean time of cough duration (p-value= 0.13) and the mean time until the onset of TB symptoms (p-value=0.15) were comparable in both groups (Table 3).

These results show that patients with TB with DM are older, have a higher level of schooling, and a lower income compared to non-diabetic patients with TB. The findings regarding age are consistent with previous findings that patients with TB with DM tend to belong to an older age group^3,8.

With the aging population, the prevalence of DM tends to increase and add to the burden already caused by TB, characterizing confluent epidemics, occurring in developing countries, such as Brazil. The association among TB, DM, and older age affects the presentation of TB, and is characterized by more severe clinical signs^3,9,10.

This study demonstrated that the combination of DM and TB occurred more frequently among patients from lower income families. It appears that an individual’s social circumstances increase susceptibility to DM in the same way that tuberculosis is associated with the poorest living conditions^9,10. The association between poor social status, TB¹¹ and DM^2,12 reflects the increased vulnerability of this group in which two chronic diseases with specific follow-up requirements are superimposed.

TB requires regular monitoring by healthcare services for six months, while DM requires continuous care over a lifetime, with regular daily medication, adequate diet, physical activity, and blood glucose control, as well as other requirements^7. It is important to address this aspect when planning health strategies for this group of patients.

To the best of our knowledge, there are no studies specifically reporting the duration of fever in patients with TB with DM. Most studies have focused on the probability of fever or frequency of fever at diagnosis. This symptom was described as three times higher among patients with TB and DM (OR 2.9, CI 95% 1.5-5.7)^9,13. The present finding of a longer duration of fever is compatible with clinical changes related to DM. Diabetic patients with TB have a higher bacterial load, which takes longer to become negative in follow-up smears after the commencement of treatment because lung infiltrates and cavities are more common in these patients, and which may also explain the longer duration of fever in this group^14,15. Fever is a reaction that depends on endogenous pyrogens. In the case of comorbidity between TB and DM, the toxins produced and released by the bacilli may account for the differences. A higher (44.6%) frequency of fever has been reported among patients with DM compared to those without DM (26.6%) (OR 2.2, CI 95%:1.2 – 4.0)¹⁰.

Regarding other clinical manifestations, some symptoms, such as fatigue and weight loss, are common to both diseases; therefore, it is difficult to determine the precise onset of symptoms of TB in individuals who are also diabetic. Furthermore, DM potentially contributes to increased immune pathology and poor control in TB infection, and alters the natural history of this infection. DM may also exacerbate the clinical manifestations of TB; Previous reports indicate that diabetic patients with TB are more likely to complain of physical symptoms, such as dyspnea, fatigue, asthenia, body pain, and headache^3,9,13.

Over time, DM causes functional and structural changes in the circulation, explaining the differences in healing time between diabetic and non-diabetic patients with TB. The primary determinant of these complications is sustained hyperglycemia, which causes biochemical and structural abnormalities in the blood vessels and peripheral nerves. Endothelial damage appears to be the factor triggering the pathogenesis of microvascular complications related to these abnormalities, representing possible symptoms of complications of DM^3,10. Higher proportion of TB severity were observed among patients with TB and DM from Brazil, after adjustment for confounding factors⁹.

These results are relevant because the differences identified between the two groups of TB patients support the need for an assessment of patients with both conditions (TB and DM). Therefore, the association between TB and DM should be prioritized when diagnosing and treating TB. More information is required on the clinical presentation of these comorbidities in Brazil. Taking in account the clinical impact of this comorbidity it is important that clinical cases are identified early and that diagnosis and treatment are not delayed.