Diabetes mellitus and tuberculosis: pattern of tuberculosis, two-month smear conversion and treatment outcomes in Guangzhou, China
There is a high burden of both diabetes (DM) and tuberculosis (TB) in China. We evaluated the association between DM and the pattern of disease, 2-month sputum smear conversion and treatment outcomes of patients with TB in Guangzhou, China.
All patients registered with TB from September 2011 to June 2012 were screened for DM and assessed for treatment outcomes in relation to presence or absence of DM and quality of DM control using patient registers, treatment cards and electronic record systems.
There were 1589 patients with TB of whom 189 (12%) had DM. Among those with DM, there was a significantly higher proportion of men, persons aged 35 years and older and persons with smear-positive pulmonary tuberculosis (PTB) (P < 0.01). In patients with DM and new smear-positive PTB, there was a higher proportion who had positive sputum smears at 2 months (21.7% vs. 5.6%, RR 3.85, 95%CI 2.24–6.63), who were lost-to-follow-up (5.2% vs. 1.7%, RR 3.23, 95%CI 1.08–9.63) and who failed treatment (10.3% vs. 2.3%, RR 4.46, 95%CI 1.96–10.18) compared with patients who had no DM. There was no significant association between these adverse outcomes and DM control as measured by 2 and 6-month fasting blood glucose.
Diabetes mellitus in new smear-positive patients with PTB was associated with failure to sputum smear convert at 2 months and adverse treatment outcomes of loss-to-follow-up and failure. Further research is needed to understand the reasons for these findings and to determine whether the current length of treatment of 6 months is adequate.
Il y a une charge de morbidité élevée à la fois pour le diabète (DS) et la tuberculose (TB) en Chine. Nous avons évalué l'association entre le DS et le profil de la TB, la conversion des frottis de crachats à deux mois et les résultats du traitement des patients TB à Guangzhou, en Chine.
Tous les patients enregistrés avec une TB de septembre 2011 à juin 2012 ont été dépistés pour le DS et évalués pour les résultats du traitement en rapport avec la présence ou l'absence de DS et la qualité du contrôle du DS à partir des registres des patients, des cartes de traitement et des systèmes de registres électroniques.
Il y avait 1 589 patients TB dont 189 (12%) avaient le DS. Parmi ces derniers, il y avait une proportion significativement plus élevée d'hommes, de personnes âgées de 35 ans et plus et de personnes avec une TB pulmonaire (PTB) à frottis positif (P < 0.01). Chez les patients atteints de DS avec une nouvelle TBP à frottis positif, il y avait une proportion plus élevée qui avait des frottis positifs à 2 mois (21.7% vs. 5.6%, RR 3.85, IC 95%: 2.24 à 6.63), qui était perdu au suivi (5.2% vs. 1.7%, RR 3.23, IC 95%: 1.08 à 9.63) et avec un échec du traitement (10.3% vs. 2.3%, RR 4.46, IC 95%: 1.96 à 10.18) par rapport aux patients non atteints de DS. Il n'y avait pas d'association significative entre ces résultats négatifs et le contrôle du DS tels que mesurés par la glycémie à jeun à 2 mois et 6 mois.
Le DS chez les nouveaux patients atteints de TBP à frottis positifs a été associé à un échec de la conversion des frottis d'expectoration à 2 mois et à des résultats négatifs de traitement, à la perte au suivi et à l’échec. Des recherches supplémentaires sont nécessaires pour comprendre les raisons de ces résultats et de déterminer si la durée actuelle de traitement de 6 mois est suffisante.
Hay una alta carga tanto de diabetes (DM) como de tuberculosis (TB) en China. Hemos evaluado la asociación entre DM y el patrón de enfermedad, la conversión del frotis de esputo en dos meses y el resultado del tratamiento de pacientes con TB en Guangzhou, China.
A todos los pacientes registrados con TB entre Septiembre 2011 y Junio 2012 se les realizaron pruebas para DM y se evaluaron los resultados del tratamiento con relación a la presencia o ausencia de DM y la calidad del control de la DM utilizando historias clínicas de los pacientes, registros de tratamiento y sistemas de registros electrónicos.
Había 1589 pacientes con TB, de los cuales 189 (12%) tenían DM. Entre aquellos con DM, había una proporción significativamente mayor de hombres, personas con 35 años o más y personas con tuberculosis pulmonar con frotis positivo (TBP) (P < 0.01). En pacientes con DM y una nueva baciloscopia positiva para TBP, había una mayor proporción que tenían frotis positivos de esputo a los 2 meses (21.7% vs. 5.6%, RR 3.85, IC 95% 2.24–6.63), que fueron perdidos durante el seguimiento (5.2% vs. 1.7%, RR 3.23, IC 95% 1.08–9.63) y cuyo tratamiento fracasaba (10.3% vs. 2.3%, RR 4.46, IC 95% 1.96–10.18) comparado con pacientes que no tenían DM. No había una asociación significativa entre estos resultados adversos y el control de la DM, medido mediante pruebas de glucosa en ayunas a los 2 y 6 meses.
La DM en pacientes con una TBP con baciloscopia positiva estaba asociada con fallos en la conversión del frotis de esputo a los 2 meses y resultados adversos del tratamiento. Se requieren más estudios para explicar estos hallazgos y determinar si es adecuada la duración actual del tratamiento de 6 meses.
China is one of the high burden tuberculosis (TB) countries in the world, with the second largest number of patients with TB estimated at 0.9–1.2 million per annum (WHO 2011). China follows the DOTS model for TB control and over the last 20 years has had much success in reducing the burden of disease caused by TB. Treatment success rates in new smear-positive pulmonary patients with TB (PTB) are consistently above 90% (WHO 2011). As a consequence of population growth, ageing, changed lifestyle and urbanisation, China is also witnessing an escalating epidemic of diabetes mellitus (DM) (Danaei et al. 2011; International Diabetes Federation 2011). Available data from a recent study in a nationally representative sample show the age-standardised prevalence of DM and pre-diabetes to be 9.7% and 15.5%, respectively, which extrapolates to 92 million adults with DM and 148 million with pre-diabetes in the country (Yang et al.2010).
Diabetes mellitus is a well-known risk factor for TB. It increases the risk of developing active TB by a factor of 2–3 compared with the normal population (Stevenson et al. 2007; Jeon & Murray 2008; Dooley & Chaisson 2009; Ruslami et al. 2010). DM patients with TB also appear to experience worse treatment outcomes than patients with no DM: studies show delays in sputum culture conversion, an increased risk of death and an increased risk of recurrent disease after successful completion of treatment (Wang et al. 2009; Baker et al. 2011; Jiménez-Corona et al. 2013). However, these findings are controversial, as other studies show no effect of DM on some of these TB treatment outcomes (Singla et al. 2006; Alisjahbana et al. 2007; Dooley & Chaisson 2009).
The double burden of disease is a serious and growing challenge for Chinese health systems, and DM poses a threat for TB control in China. In line with guidelines from the China National TB Programme, it is recommended that all patients with TB, except those with multidrug resistant TB (MDR–TB, i.e., resistant to isoniazid and rifampicin), use standardised short regimens for treatment (Ministry of Health & Chinese Centre on TB Control & Prevention (CDC) 2008). There is little information about how this short course regimen performs in patients with DM–TB and whether patients with good or poor control of DM experience different treatment outcomes.
Bidirectional screening of TB and DM has been implemented in several sites in China since September 2011 (Li et al. 2012; Lin et al. 2012). One of these sites is Guangzhou where the current study was conducted. The objectives of this study were to describe in patients with TB in two clinics in Guangzhou in relation to the presence or absence of DM and the quality of DM control: (i) demographic characteristics, types and categories of TB, (ii) sputum smear conversion at 2 months, and (iii) treatment outcomes of new patients with PTB.
Setting and patients
This was a cross-sectional and retrospective cohort study involving a record review of patients with TB registered and treated in two TB clinics in Guangzhou Chest Hospital. Guangzhou, the capital of Guangdong province in the south of China, has 12.7 million inhabitants, of whom 37.5% are migrants. The HIV prevalence in Guangzhou is 36/100 000. Guangzhou Chest Hospital has two TB clinics, both of which have been implementing screening of DM in the routine setting since September 2011 (Li et al. 2012). There are 47 staff (16 doctors and 31 nurses) in Number 1 TB clinic and 29 staff (nine doctors and 20 nurses) in Number 2 TB clinic. Only ambulatory patients are treated as outpatients in the two clinics.
The diagnosis of TB in the two clinics is made in line with China TB Control Programme Guidelines (Ministry of Health & Chinese Centre on TB Control & Prevention (CDC) 2008). Patients are categorised as either new or previously treated cases according to the patient's history and are further classified by type of TB into smear-positive PTB, smear-negative PTB and extrapulmonary TB (EPTB). Each person with suspected TB is investigated by sputum smear examination. If sputum specimens are smear positive, the patient is diagnosed as smear-positive PTB; if sputum smears are negative and chest radiography is compatible with active PTB, the patient is diagnosed as smear-negative PTB after discussion by clinical and radiographic doctors; if the patient has a local extrapulmonary lesion with symptoms compatible with TB, the purified protein derivative test is positive and other diseases are excluded, the diagnosis is EPTB. The tests for sputum smear examination are free; Mycobacterium tuberculosis culture is not free. Both TB clinics provide free treatment for patients with PTB using standardised regimens, but patients with EPTB have to pay for therapy. Treatment for new PTB patients consists of an initial phase of rifampicin, isoniazid, pyrazinamide and ethambutol given daily for 2 months followed by a continuation phase of rifampicin and isoniazid daily for 4 months. Retreatment patients are given a standard retreatment regimen (Ministry of Health & Chinese Centre on TB Control & Prevention (CDC) 2008). Patients with EPTB are treated for varying lengths of time, depending on disease severity, and treatment is not standardised. Monitoring the response to therapy is done according to national guidelines, using TB patient registers, TB treatment cards and electronic systems to record and report on patient information. In all patients with PTB, sputum smears are done at the end of the initial phase of therapy at 2 months, during the continuation phase of therapy and at the end of therapy at 6 months.
The diagnosis of DM for patients with TB follows national guidelines which stipulate that a fasting blood glucose (FBG) is carried out using venous plasma and a biochemical analyser with a cut-off threshold in line with that recommended by WHO (2006). Screening is based on (i) determining whether patients are already known to have DM and (ii) in those with no known DM, measuring FBG (Li et al. 2012). For a diagnosis of DM to be made, a patient has to have two FBG measurements ≥7.0 mm (126 mg/dl), with one measurement done in the TB clinic and one later in the DM clinic after referral from the TB clinic. All patients with previously diagnosed DM and newly diagnosed DM are enrolled into DM care where the diabetes doctors decide on the type of therapy. Care for the dually affected patient is thus given at the TB clinic and also in the DM clinic.
All patients registered with TB in the two clinics between 1 September 2011 and 30 June 2012 (10 months) were included in the study.
Data were obtained from TB patient registers, TB treatment cards and the electronic data base for DM in Number 1 or Number 2 clinic. Data variables included (i) TB status: TB registration number, date of registration, age, sex, type of TB (smear-positive PTB, smear-negative PTB, EPTB), category of TB (new or previously treated); for smear-positive new PTB, sputum smear status (positive, negative or not done) at 2 months, (ii) TB treatment outcomes in new patients with PTB, with favourable outcomes defined as cure and treatment completed and unfavourable outcomes defined as loss-to-follow-up, dead, failure and transfer-out); DM status at time of registration and DM control at 2 and 6 months of TB treatment. Data were collected into a paper proforma and EXCEL electronic tool by five staff members of the TB clinic Number 1 and Number 2 in Guangzhou Chest Hospital and by two staff members of the China Center for TB control. Data were double entered into EpiData and validated.
Analysis and statistics
Descriptive analysis was performed. Comparisons were made between patterns of TB, treatment outcomes and 2-month sputum smear conversion in patients with and without DM and in relation to quality of DM control determined by FBG at baseline, 2 months and 6 months. DM control was based on fasting status from the previous night and arbitrarily set at – DM good control, FBG ≤ 7.0 mm; DM poor control, FBG 7–10 mm; DM bad control, FBG > 10 mm. Comparisons were made using the chi-square test, relative risks (RR) and 95% confidence intervals (95% CI) where appropriate. Where there were univariate associations of risk factors for DM and TB, these were further analysed using multivariate logistic regression. Levels of significance were set at 5%.
Ethics approval to use, report and publish the collected data was obtained from the Union Ethics Advisory Group, Paris, France.
Demographic characteristics and pattern of TB in patients with and without DM
There were 1589 patients with TB (469 migrants, 30%), including those with newly diagnosed and previously treated disease, who were diagnosed, registered and initiated on treatment in No 1 and No 2 clinic, Guangzhou Chest Hospital. Of these, 189 (12%) had DM. Demographic characteristics and patterns of TB in patients with and without DM are shown in Table 1. On univariate analysis, there was a significantly higher proportion of men, people 35 years and older, people with smear-positive PTB and persons with previously treated disease among those with DM than those without. However, after adjusting for age, there were no differences with regard to previously treated disease.
Table 1. Characteristics and pattern of tuberculosis in patients with and without diabetes mellitus in Guangzhou, Chinaa
|Male||145 (76.7)||897 (64.1)||1.73||1.25–2.38||0.0006|
|Female||44 (23.3)||503 (35.9)||Reference|| || |
|<15||0 (0.0)||4 (0.3)||N/A|| || |
|15–34 years||16 (8.5)||737 (52.6)||Reference|| || |
|35–54 years||90 (47.6)||439 (31.4)||8.01||4.76–13.47||<0.001|
|> = 55 years||83 (43.9)||220 (15.7)||12.89||7.68–21.65||<0.001|
|Type of TB||Smear-negative PTB||70 (37.0)||842 (60.1)||Reference|| || |
|Smear-positive PTB||118 (62.4)||548 (39.1)||2.31||1.75–3.05||<0.001|
|EPTB||1 (0.5)||10 (0.7)||1.18||0.18–7.78||0.5875|
|Category of TBb||New||163 (86.2)||1290 (92.1)||Reference|| || |
|Previously treated||26 (13.8)||110 (7.9)||1.7||1.17–2.48||0.0065|
Sputum smear conversion at the end of 2 months of anti-TB treatment
Of 580 new smear-positive patients with PTB, 535 (92%) had a sputum smear examination at the end of 2 months of treatment: 45 tested smear positive. More than 20% of patients with DM had smear-positive sputum at 2 months which was a significantly greater proportion than that observed in patients without DM (Table 2). Although the prevalence of smear-positive sputum at 2 months was higher in patients with poor or bad DM control (25.5%) than in those who had good DM control (12.9%), these differences did not reach statistical significance (Table 3).
Table 2. Patients with new smear-positive PTB who were sputum smear positive for acid-fast bacilli at 2 months of anti-TB treatment in relation to the presence or absence of diabetes mellitus, Guangzhou, Chinaa
|With DM||92||20 (21.7)||3.85||2.24–6.63||<0.001|
|Without DM||443||25 (5.6)||Reference|| || |
|All patients with TB||535||45 (8.4)|| || || |
Table 3. Patients with DM and new smear-positive PTB who were sputum smear positive for acid-fast bacilli at 2 months of anti-TB treatment in relation to diabetes control, Guangzhou, Chinaa
|DM good control||31||4 (12.9)||Reference|| || |
|DM poor control||30||9 (30.0)||2.33||0.80–6.75||0.1031|
|DM bad control||25||5 (20.0)||1.55||0.46–5.17||0.4928|
|DM poor + bad control||55||14 (25.5)||1.97||0.71–5.47||0.2693|
|All patientsb||86||18 (20.9)|| || || |
Treatment outcomes in TB patients with and without DM
Treatment outcomes in patients with and without DM are shown in Table 4. Patients with DM had significantly higher rates of loss-to-follow-up and failure than those who did not have DM, otherwise there were no differences with other outcomes such as treatment success, death or transfer-out. Rates of failure were no different in the DM patients with TB in relation to whether DM control was good, poor or bad at 6 months (Table 5).
Table 4. Treatment outcomes in new smear-positive PTB patients with and without diabetes mellitus in Guangzhou, Chinaa
|With DM||97||77 (79.4)||5 (5.2)||3 (3.1)||10 (10.3)||2 (2.1)|
|Without DM||483||416 (86.1)||8 (1.7)||46 (9.5)||11 (2.3)||2 (0.4)|
|Risk ratio (95%CI)|| ||0.92 (0.83–1.03)||3.23 (1.08–9.63)||0.38 (0.12–1.18)||4.46 (1.96–10.18)||5.29 (0.76–37.01)|
Table 5. Patients with DM and new smear-positive PTB who failed treatment in relation to diabetes control, Guangzhou, Chinaa
|DM good control||29||4 (13.8)||Reference|| || |
|DM poor control||33||4 (12.1)||0.88||0.24–3.20||1.0000|
|DM bad control||15||2 (13.3)||0.97||0.20–4.69||1.0000|
|DM poor + bad control||48||6 (12.5)||0.91||0.28–2.94||1.0000|
|Totalb||77||10 (13.0)|| || || |
This is the first comprehensive study in a large city in China to assess treatment outcomes in patients with TB who were screened at the time of registration for DM. TB patients with DM, compared with those without DM, were more likely to be male, older and have smear-positive pulmonary and previously treated TB. In terms of treatment outcomes, a higher proportion of new patients with PTB and DM had positive sputum smears at 2 months and higher rates of loss-to-follow-up and failure by the end of treatment. We found no association between these adverse outcomes and DM control as measured by FBG at 2 and 6 months.
The strengths of this study are that a large number of patients with TB were included, consecutively registered and enrolled to treatment within a programme setting, and all new patients with smear-positive disease were treated with the same regimen for 6 months with outcomes monitored and recorded in a standard way. The study report also followed STROBE guidelines (Von Elm et al. 2007). Limitations relate to the operational nature of the study with a high proportion of patients being migrants and a transfer-out rate approaching 10%. We have no data on baseline sputum smear grade in new patients with PTB. It has been previously observed that a high baseline sputum smear grade is associated with failure to smear convert at 2 months (Telzak et al. 1997; Dominguez-Castellano et al. 2003; Singla et al. 2003; Kuaban et al. 2009), and we do not know whether this was different between patients with and without DM. We also have no data on duration of antidiabetic treatment and therefore do not know whether it influences DM control and treatment outcome.
A previous study in China found that the TB relapse rate was higher in patients with DM–TB than those who did not have DM (Zhang et al. 2009). Our study showing that patients with DM–TB had a higher prevalence of previously treated TB is in accordance with these findings. Higher relapse rates have also been documented in Japan, the Democratic Republic of Congo, Saudi Arabia, Tunisia and Mexico (Wada 2000; Mboussa et al. 2003; Singla et al. 2006; Maalej et al. 2009; Zhang et al. 2009; Jiménez-Corona et al. 2013).
Previous studies have examined sputum culture conversion in patients with DM–TB at 2 or 3 months of treatment with nine studies showing RR that varied from 0.79 to 3.25 (Baker et al.2011). In accordance with programme guidelines, we assessed sputum smear conversion at 2 months, finding a higher prevalence of smear-positive sputum among patients with DM–TB compared with those who did not have DM. This is similar to findings recently reported from Mexico (Jiménez-Corona et al. 2013).
In contrast to a systematic review which found an increased risk of death in patients with DM–TB (Baker et al. 2011), we found no association of the effect of DM on case fatality rates, maybe because there were only four deaths in total in our study. The main finding was a high prevalence of treatment failure in patients with DM–TB, which was also found in the prospective study from Mexico (Jiménez-Corona et al. 2013). Failure to smear convert at 2 months is a warning for subsequent treatment failure, and treatment failure is a risk factor for MDR–TB (Becerra et al. 2000; Oxlade et al. 2010). This observation is therefore of concern. A systematic review of just four studies found no association between DM–TB and drug-resistant recurrent disease (Baker et al.2011). However, in a previous study in China, there was a higher frequency of MDR–TB in patients with DM compared with those who did not have DM (Zhang et al. 2009), so this question is still unresolved and requires further prospective research. Reasons for delayed smear conversion and treatment failure are speculative but include drug–drug interactions between oral hypoglycaemic drugs and rifampicin and impaired immunity as a result of DM (Stevenson et al. 2007; Dooley et al. 2009).
This study has implications for TB treatment in patients with DM–TB. There is growing evidence to suggest that treatment failure is increased in patients with DM–TB, and this is coupled with a failure of these patients to smear convert at 2 months at the end of the initial phase of therapy. This phenomenon and the underlying reasons need further assessment. Potential solutions are to increase the length of the initial phase or the entire length of anti-TB treatment, although this would have major policy implications for TB control programmes or to control DM as aggressively as possible at the start of anti-TB treatment.
In conclusion, we have found that DM in new smear-positive patients with PTB was associated with failure to sputum smear convert at 2 months and adverse treatment outcomes of loss-to-follow-up and failure. Further research is needed to understand the reasons for these findings and to determine whether the current length of treatment at 6 months is adequate.
This research was supported through an operational research course, which was jointly developed and run by the Centre for Operational Research, International Union Against Tuberculosis and Lung Disease, France, The Union South-East Asia Regional Office and the Operational Research Unit (LUXOR), Medecins sans Frontieres, Brussels-Luxembourg. Additional support for running the course was provided by the Center for International Health, University of Bergen, Norway. Funding for the course was from an anonymous donor and the Department for International Development, UK.