What causes symptoms suggestive of tuberculosis in HIV-positive people with negative initial investigations?

OBJECTIVE: To identify the causes of symptoms suggestive of tuberculosis (TB) among people living with the human immunodeficiency virus (PLHIV) in South Africa. METHODS: A consecutive sample of HIV clinic attendees with symptoms suggestive of TB (⩾1 of cough, weight loss, fever or night sweats) at enrolment and at 3 months, and negative initial TB investigations, were systematically evaluated with standard protocols and diagnoses assigned using standard criteria. TB was ‘confirmed’ if Mycobacterium tuberculosis was identified within 6 months of enrolment, and ‘clinical’ if treatment started without microbiological confirmation. RESULTS: Among 103 participants, 50/103 were preantiretroviral therapy (ART) and 53/103 were on ART; respectively 68% vs. 79% were female; the median age was 35 vs. 45 years; the median CD4 count was 311 vs. 508 cells/mm3. Seventy-two (70%) had ⩾5% measured weight loss and 50 (49%) had cough. The most common final diagnoses were weight loss due to severe food insecurity (n = 20, 19%), TB (n = 14, 14%: confirmed n = 7; clinical n = 7), other respiratory tract infection (n = 14, 14%) and post-TB lung disease (n = 9, 9%). The basis for TB diagnosis was imaging (n = 7), bacteriological confirmation from sputum (n = 4), histology, lumbar puncture and other (n = 1 each). CONCLUSION: PLHIV with persistent TB symptoms require further evaluation for TB using all available modalities, and for food insecurity in those with weight loss.


S U M M A R Y O B J E C T I V E :
To identify the causes of symptoms suggestive of tuberculosis (TB) among people living with the human immunodeficiency virus (PLHIV) in South Africa. M E T H O D S : A consecutive sample of HIV clinic attendees with symptoms suggestive of TB (71 of cough, weight loss, fever or night sweats) at enrolment and at 3 months, and negative initial TB investigations, were systematically evaluated with standard protocols and diagnoses assigned using standard criteria. TB was 'confirmed' if Mycobacterium tuberculosis was identified within 6 months of enrolment, and 'clinical' if treatment started without microbiological confirmation. R E S U LT S : Among 103 participants, 50/103 were preantiretroviral therapy (ART) and 53/103 were on ART; respectively 68% vs. 79% were female; the median age was 35 vs. 45 years; the median CD4 count was 311 vs. 508 cells/mm 3 . Seventy-two (70%) had 75% measured weight loss and 50 (49%) had cough. The most common final diagnoses were weight loss due to severe food insecurity (n ¼ 20, 19%), TB (n ¼ 14, 14%: confirmed n ¼ 7; clinical n ¼ 7), other respiratory tract infection (n ¼ 14, 14%) and post-TB lung disease (n ¼ 9, 9%). The basis for TB diagnosis was imaging (n ¼ 7), bacteriological confirmation from sputum (n ¼ 4), histology, lumbar puncture and other (n ¼ 1 each). C O N C L U S I O N : PLHIV with persistent TB symptoms require further evaluation for TB using all available modalities, and for food insecurity in those with weight loss. K E Y W O R D S : South Africa; Xpert w MTB/RIF; TB symptoms; human immunodeficiency virus THE WORLD HEALTH ORGANIZATION (WHO) recommends regular screening of people living with the human immunodeficiency virus (PLHIV) for tuberculosis (TB) using a symptom screen comprising any one of current self-reported cough, fever, weight loss or night sweats (hereafter termed the 'WHO tool'), as an essential part of the HIV care package. 1 Although people attending for HIV care in sub-Saharan Africa are highly symptomatic, 2 most of those reporting WHO tool symptoms have negative TB investigations, [2][3][4] and a proportion continue to report symptoms. Early identification of people with active TB among PLHIV is a priority; however, the evidence underpinning investigation pathways after an initial sputum test is weak. [5][6][7][8][9][10][11][12][13][14] The aim of our study was to determine the causes of persistent or recurrent symptoms suggestive of TB among ambulatory adults attending for HIV care who had negative initial TB investigations.
Africa: 'Xpert for people attending HIV/AIDS care: test or review?' (XPHACTOR). 15 XPHACTOR study population, recruitment and procedures XPHACTOR study flow, procedures and algorithm are described in detail in the Online Appendix* (section on 'Main study procedures', Appendix Figure A.1). Briefly, we enrolled a systematic sample (using a predetermined system designed to minimise the risk of researcher selection bias) of adults (aged 718 years) attending four HIV clinics in Gauteng Province, South Africa, irrespective of the presence of symptoms suggestive of TB. Patients taking antituberculosis treatment within the previous 3 months were excluded. Patients were enrolled into 'on antiretroviral therapy (ART)' 'ART-experienced' and 'pre-ART' (in HIV care or newly diagnosed HIV-positive, not taking ART) groups. At the time of the study, ART eligibility comprised CD4 count 6350 cells/mm 3 or WHO clinical stage 73. Research staff screened participants for TB at monthly intervals up to 3 months using a standardised questionnaire that incorporated the WHO tool. The study algorithm defined individuals as a priori at highest risk of active TB if they had any of the following: current cough, fever 7 3 weeks, night sweats 7 4 weeks, body mass index (BMI) ,18.5 kg/ m 2 , CD4 count ,100 cells/mm 3 or weight loss 710%. A spot sputum sample was collected from these individuals if possible for Xpert testing. At the 3-month visit, all participants underwent sputum (induced if necessary) and blood cultures for mycobacteria (BACTEC MGITe 960e and 9240e systems; BD, Sparks, MD, USA). We allowed a broad window period around the 3-month XPHACTOR main study visit until around 6 months to maximise follow-up.

Sub-study eligibility and enrolment
Between October 2013 and April 2014 at the XPHACTOR 3-month visit, consecutive participants who were not on anti-tuberculosis treatment and who had persistent or recurrent symptoms suggestive of TB were invited to participate in this sub-study. Persistent or recurrent TB symptoms were defined as 1) self-report of any of cough, fever or night sweats at enrolment, and self-report of any of the aforementioned symptoms at 3-month visit; or 2) self-report of unintentional weight loss and 75% measured weight loss since XPHACTOR enrolment. Figure 1 shows the sub-study flow and procedures. A chest radiograph (CXR) was requested if there was no film available for the previous 6 weeks, and all were asked to bring samples (stool, early-morning urine and sputum) for mycobacterial culture when they attended for research physician assessment. Further procedures were determined by symptoms ( Figure 1); if cough was reported, the research nurse collected an additional sputum sample for bacterial culture (induced if necessary), two nasopharyngeal swabs and one oropharyngeal swab. Sputum samples were tested using routine bacterial microscopy and culture, and polymerase chain reaction (PCR) for bacteria, including Bordetella pertussis. One nasopharyngeal swab was inserted directly into Regan Lowe transport media for Bordetella spp. culture, and the remaining swabs were placed in Primestore medium for PCR detection of B. pertussis and other pathogens ( Figure 1). All samples were transported within 24 h of collection to the research laboratory. PCR for B. pertussis was performed in accordance with the method described by Tatti et al. 16 An abdominal ultrasound scan was requested for those with weight loss. Participants reporting fever or night sweats were given a digital thermometer to record oral temperature (morning, evening, and if any fever or sweats) for 1 week.
Research physician assessment Around 1 week after enrolment, sub-study participants underwent systematic clinical evaluation, including examination by a research physician who arranged a standard set of investigations according to the participant's symptoms ( Figure 1 and Appendix Figure A.2, Appendix section on 'Sub-study research physician assessment').
First-line evaluation for cough was spirometry if cough 78 weeks or features suggestive of chronic obstructive pulmonary disease (COPD) or asthma; if clinically appropriate, blood samples were collected for C-reactive protein (CRP) testing to help distinguish the likelihood of bacterial infection and, if cardiac failure was suspected, for serum b-natriuretic peptide.
Second-line evaluation for cough comprised a trial of appropriate treatment for those with clinical features suggestive of cough due to upper airways disease, angiotensin-converting enzyme (ACE) inhibitors or gastro-oesophageal reflux disease (GORD). All participants were screened using validated tools for depression (Patient Health Questionnaire 9 [PHQ-9]), 17 household food insecurity (household food insecurity access score [HFIAS]), 18 and alcohol misuse (Fast Alcohol Screening Test [FAST] score), 19 and were asked about use of tobacco, snuff and wood-burning stoves. Using a standardised form, the physician abstracted information from clinic records relevant to assigning final diagnoses, such as chronic disease diagnoses, results of recent investigations in particular for TB, and history of HIV, ART and TB.

Sub-study follow-up
Sub-study participants were followed for a further 3 months and screened for TB at each visit by research staff using a standardised questionnaire incorporating the WHO tool, with further investigation for TB in accordance with the XPHACTOR study algorithm ( Figure 1). The research physician reviewed participants at these visits if required to assign final diagnoses.
Causes of TB symptoms in HIV-positive adults physician using pre-set criteria (Appendix Table A), including the case definitions for TB detailed below.
TB case definitions 'Confirmed' TB was defined as a positive result on 1) Xpert (on sputum sample), 2) line-probe assay (LPA) performed on smear-positive or cultured isolate (GenoType MTBDRplus, Hain Lifesciences, Nehren, Germany) or 3) M. tuberculosis culture, from any sample collected within 6 months of sub-study enrolment. 'Clinical TB' was defined as anti-tuberculosis treatment started within 6 months of sub-study enrolment in the absence of microbiological confirmation.
Radiological definitions 'Probable radiological TB' was defined as the presence of 1) any of cavitation, predominantly upper lobe infiltrates, pleural or pericardial effusion, or clear miliary picture on CXR, or 2) any of abdominal lymphadenopathy, splenic microabscesses, pleural or pericardial effusion on ultrasound scan. 'Possible radiological TB' was defined as the presence of any of lymphadenopathy (hilar or mediastinal), pulmonary nodules or other infiltrates. Participants with 'probable' or 'possible' radiological TB features but without bacteriological confirmation who started anti-tuberculosis treatment within 6 months of substudy enrolment were assigned 'clinical' TB.

Ethical approval
The study protocol was approved by the ethics committees of the University of the Witwatersrand, Johannesburg; University of Cape Town, Cape Town, South Africa; and the London School of Hygiene & Tropical Medicine, London, UK. All participants provided written informed consent or, if unable to write, witnessed verbal consent.
Of 103 sub-study participants, 14 (14%) (6 on ART, 8 pre-ART) fulfilled TB case definitions (7 confirmed, 7 clinical). Eight participants started treatment due to compatible imaging (4 ultrasound, 2 abdominal ultrasound and CXR, and 2 CXR), of whom 3 were subsequently bacteriologically confirmed on sputum (1 Xpert þ culture, 2 culture). Four participants started anti-tuberculosis treatment based on a positive sputum result (2 Xpert, 2 culture). One participant started treatment based on histology following fine-needle lymph node aspiration, and one based on lumbar puncture.
The median time from enrolment to start of antituberculosis treatment was 21 days (range 1-137) for 13 participants with a documented anti-tuberculosis treatment start date. One further participant had positive Xpert on sputum 149 days after enrolment but an unknown treatment start date. Among the 8 pre-ART participants, 4 started anti-tuberculosis treatment after ART initiation (3 within 3 months, 1 within 6 months). A further two participants who were enrolled but who did not undergo physician assessment fulfilled case definitions for confirmed TB, of whom one died before anti-tuberculosis treatment was initiated (Figure 2).

DISCUSSION
In this representative sample of HIV clinic attendees in South Africa reporting persistent or recurrent WHO tool symptoms 3 months after a negative initial investigation for TB, among those able to produce sputum, 14/103 (14%) had TB. Half started anti-tuberculosis treatment based on imaging, mainly abdominal ultrasound, which illustrated the limitations of sputum-based diagnostics for detecting extrapulmonary TB. With an estimated 40% shortfall globally between notified cases and estimated incidence of TB in 2016, and with South Africa one of the 10 countries accounting for most of this gap, 20 we recommend using multiple diagnostic modalities, particularly imaging, to help identify these missing TB patients.
infectious (TB, 46%; lower respiratory tract infection, 31%). 21 Munyati et al. also identified a high proportion of non-communicable disease diagnoses, in particular post-tuberculous disease, asthma and heart failure. 21 We also found post-TB chronic lung disease to be a relatively common diagnosis; better criteria to distinguish it from active TB and optimal management are needed. [22][23][24] Our data support Chakaya et al.'s call for large-scale epidemiological studies of post-TB lung disease. 22 Severe food insecurity was the most common cause of weight loss. Food insecurity has not previously been evaluated as a possible cause for weight loss in the context of TB screening, although it is well described as a barrier to adherence to ART. 25 We only assigned this diagnosis after searching for other, more likely diagnoses, and chose severe (rather than moderate) food insecurity as a more specific marker. Clinicians should consider screening for food insecurity among people with weight loss, particularly if not associated with other symptoms, and ensure patients are linked to social support where available. Forty per cent of our study participants screened positive for significant depression, and almost one fifth had harmful alcohol use, comparable with estimates of 31% and 7-31%, respectively, from a systematic review in sub-Saharan Africa of HIV-positive people on ART by Nakimuli-Mpungu et al. 26 In their pooled analysis, individuals with significant depression were less likely to adhere to ART. 26 Screening for depression with provision of appropriate care should be part of the HIV care package in lower-income settings to help optimise ART adherence and treatment outcomes.
The WHO tool was developed for use in resourcelimited settings to provide a simple clinical algorithm to reliably rule out TB before providing isoniazid preventive therapy to PLHIV. As the tool was designed to maximise sensitivity (78.9%) and minimise the negative likelihood ratio for TB, it has low specificity (49.6%). At a TB prevalence of 5% in PLHIV, it has a negative predictive value of 97.7%, but a very low positive predictive value (8%). 27 Individuals who screen positive, the majority of whom will not have TB, require further evaluation for TB using Xpert, which has been recommended as the initial diagnostic test. 1 This poses a huge challenge in resource-constrained settings when it is used, as recommended, for active TB case finding in PLHIV at every clinical encounter. 1 Simple, low-cost strategies to prioritise those with WHO tool symptoms for TB investigation, such as 'second-step' clinical algorithms, 15,28 or point-of-care CRP testing (also suggested as an alternative TB screening tool), 29 are potential solutions that merit further evaluation.
Strengths of our study included our systematic physician evaluation of a representative sample of HIV clinic attendees in a clinically relevant manner with a standardised set of investigations and longitudinal follow-up of participants. We cannot rule out that additional diagnosis of TB and other specific diagnoses might have been made if further investigations had been undertaken. Weight loss was commonly reported by our study participants, but we restricted our study to those with measured weight loss to make this criterion more objective.

CONCLUSIONS
TB, post-TB chronic lung disease and food insecurity were the main diagnoses for symptoms suggestive of TB in our population of HIV clinic attendees who had previously undergone systematic screening and investigation for TB, and we were able to assign diagnoses for more than 90% of participants. Our study highlights the need to continue to investigate for TB using multiple modalities among HIV-positive people with persistent symptoms, as well as evaluation for food insecurity, and for further studies to guide the identification and management of the sequelae of pulmonary TB.

Enrolment
At enrolment, research staff administered a standardised questionnaire, which incorporated the World Health Organization (WHO) tool, collected details of tuberculosis (TB) and human immunodeficiency virus (HIV) treatment, and basic demographic and socioeconomic information. Further investigation was prioritised according to the XPHACTOR ('Xpert for people attending HIV/AIDS care: test or review?') algorithm with an immediate spot sputum sample sent for Xpert w MTB/RIF (Cepheid, Sunnyvale, CA, USA) testing for individuals at a priori highest risk of active TB: 1) all assigned 'high priority' (any of the following: current cough, fever 7 3 weeks, body mass index [BMI] ,18.5 kg/m 2 , CD4 count , 100310 6 /l, measured weight loss 710% in preceding 6 months or other feature raising high clinical suspicion of TB); 2) those in the pre-ART group with CD4 count of ,200310 6 /l at enrolment; 3) all those in the HIV testing and counselling (HTC) group (whose CD4 count was unknown) at enrolment. For all other participants, a spot sputum sample was collected at enrolment and frozen at À808C within 24 h for testing with Xpert at the end of the study ( Figure  A.1). This enabled comparison of sensitivity and specificity of the XPHACTOR study algorithm to detect TB cases against sensitivity and specificity if Xpert had been performed immediately for all with any WHO tool symptom.

Follow-up
Participants were reviewed monthly to 3 months, with repeat WHO symptom screen and a spot sputum requested for Xpert if 'high priority' by the study algorithm at that visit, with the exception of those in the 'on ART' group who were asymptomatic at enrolment, who were telephoned at 1 and 2 months to update locator information but were not asked about TB symptoms. At the 3-month visit, sputum (induced if necessary) and blood were collected for mycobacterial culture on liquid media (BACTEC MGITe [Mycobacterium Growth Indicator Tubes; BD, Sparks, MD, USA] 960e and 9240e systems) from all study participants, regardless of symptoms ( Figure A.1). We allowed a broad window period around the scheduled 3-month visit until around 6 months to maximise study follow-up.
Participants who submitted an Xpert sample were reviewed within 1 week. If Xpert-positive, antituberculosis treatment was initiated; if negative, research staff repeated WHO symptom screening and facilitated the Xpert-negative algorithm, which comprised chest radiograph (CXR), spot sputum for TB culture and/or antibiotic trial as clinically appropriate (Figure A.1).
Investigation results were returned to clinic staff, who were responsible for management decisions. Clinic records were reviewed at the end of the study to ascertain any additional relevant investigations and/or TB diagnoses. Deaths were identified through reports from participant-nominated contacts, clinic staff, and by accessing the Department of Home Affairs vital statistics database using participants' South African identification (ID) numbers.

Sub-study research physician assessment
The research physician administered a standardised questionnaire which had targeted questions to systematically identify the cause for WHO tool symptom(s) reported. All participants were asked about past medical history, current medications, and investigations and treatment undertaken to date. For example, those with cough were asked about cough duration and triggers, associated symptoms and any preceding respiratory tract infection. Results for all previous investigations undertaken for the study were reviewed by the research physician, for example, CXR and abdominal ultrasound scan; Xpert and mycobacterial culture on sputum; and sputum, nasopharyngeal and oropharyngeal samples for respiratory pathogens. Participants with pleural effusions or lymphadenopathy were referred to the clinic physician for consideration for aspiration or fine-needle aspiration, as appropriate. Symptomspecific evaluation undertaken by the research physician is summarised below.

EVALUATION OF COUGH
First-line evaluation Blood samples were collected from all participants with cough for full blood count and, where appropriate, C-reactive protein (CRP), to help distinguish the likelihood of bacterial respiratory infection and, if febrile (.38.38C), aerobic and anaerobic bacterial blood cultures. If bacterial infection was suspected, oral antibiotics or hospital admission were facilitated when clinically appropriate.
A trained research physician performed spirometry in accordance with American Thoracic Society (ATS) and European Respiratory Society (ERS) standards 1 for any participant with cough 7 8 weeks, or features suggestive of chronic obstructive pulmonary disease (COPD) or asthma, unless respiratory clinic spirometry results were already available. The Advanced Medical Engineering spirometer (AME, Cape Town, South Africa) was used, with calibration checks performed in accordance with the manufacturer's recommendations. Up to eight seated readings were taken, and post-bronchodilator (5 mg nebulised salbutamol) spirometry performed if a spirometry Causes of TB symptoms in HIV-positive adults i abnormality was found. If ATS/ERS within-and between-manoeuvre acceptability criteria were not met, we reported usable curves (good start and satisfactory exhalation). Post-bronchodilator spirometry data were used to confirm airflow obstruction, defined using Global Lung Initiative (GLI) 2012 equations (forced expiratory volume in 1 s [FEV 1 ]/ forced vital capacity [FVC] , lower limit of normal at the 5 th centile). 2 Post-bronchodilator increase in FEV 1 .12% of predicted and .200 ml was used to confirm asthma. Participants were referred to the clinic physician for further management if spirometry confirmed asthma or COPD, or if spirometry was normal but obstructive airways disease likely, or if another spirometry abnormality was identified. Response to any treatment provided was assessed at 4-12 weeks.
If participants had clinical features suggestive of cardiac failure, serum b-natriuretic peptide (BNP) was measured and if levels were .100 pg/ml, further Figure A.1 XPHACTOR main study flow and procedures. * XPHACTOR algorithm at enrolment: high priority (any of current cough, fever 7 3 weeks, BMI , 18.5 kg/m 2 , CD4 , 100310 6 /l, measured weight loss 7 10% in preceding 6 months or other feature raising high clinical suspicion of TB); medium priority (any of fever ,3 weeks, night sweats, measured weight loss ,10% in preceding 6 months); low priority ¼ no TB symptoms. † Samples tested using Xpert at the end of the study to enable comparison of sensitivity and specificity of the XPHACTOR study algorithm to detect TB cases against sensitivity and specificity if Xpert had been performed immediately for all with any WHO tool symptom. ‡ XPHACTOR algorithm at monthly follow-up: high priority (any of current cough, fever 73 weeks, night sweats 74 weeks, BMI , 18.5 kg/m 2 , CD4 , 100x10 6 /l, measured weight loss 710% in preceding 6 months or other feature raising high clinical suspicion of TB); medium priority (any of fever ,3 weeks, night sweats ,4 weeks, measured weight loss ,10% in preceding 6 months); low priority ¼ no TB symptoms. § Screened by research nurse between October 2013 and April 2014. Eligible if not on anti-tuberculosis treatment and persistent or recurrent TB symptoms, defined as: 1) any of cough, fever or night sweats reported at enrolment and any of aforementioned symptoms reported at 3-month visit; OR 2) 75% measured weight loss at 3-month visit and reported unintentional weight loss. XPHACTOR ¼ Xpert for people attending HIV/AIDS care: test or review? HIV ¼ human immunodeficiency virus; ART ¼antiretroviral therapy; BMI ¼ body mass index; TB ¼ tuberculosis; WHO ¼ World Health Organization.
ii The International Journal of Tuberculosis and Lung Disease cardiological evaluation and management was facilitated. If participants had symptoms compatible with Pneumocystis jiroveci pneumonia (fever/exertional dyspnoea/tachypnoea) and CD4 count ,200 cells/ll, exercise oximetry was undertaken and the participant referred to the responsible clinic physician for further management.
Any suspicious CXR features were discussed with the clinic physician to facilitate further appropriate evaluation or treatment as deemed appropriate, for example, pleural aspiration, computed tomography (CT) imaging, endoscopy or bronchoscopy or presumptive anti-tuberculosis treatment.

Second-line evaluation of cough
If diagnosis for cough was not identified by first-line evaluation, a trial of appropriate treatment was arranged for those with clinical features suggestive of cough due to upper airways disease, angiotensinconverting enzyme (ACE) inhibitors or gastrooesophageal reflux (GORD). This comprised corticosteroid nasal spray and/or antihistamine, or ear, nose, throat referral if upper airways disease was suspected; switching ACE inhibitor to a suitable alternative; lifestyle advice and trial of a proton pump inhibitor if GORD was suspected. Response to treatment was reviewed at 4-12 weeks. Smoking cessation advice was given to all current smokers, and improvement in cough evaluated at 4-12 weeks postcessation, if applicable. If no likely cause for cough was identified after second-line evaluation, referral to a respiratory physician was facilitated.
Evaluation of 75% unintentional weight loss Evaluation aimed to identify a broad spectrum of causes of weight loss, including TB, endocrine 2 Evaluation for TB undertaken for all study participants. * At least one sputum for Xpert or mycobacterial culture; 11 participants had both. † Pleural effusion (n ¼ 1). ‡ Pleural effusion (n ¼ 5), upper lobe infiltrates (n ¼ 4), cardiomegaly with ultrasound confirmed pericardial effusion and pleural effusion (n ¼ 1), cardiomegaly with ultrasound confirmed pericardial effusion (n ¼ 1) cavitation and pleural effusion (n ¼ 1). § Three of the four participants with M. tuberculosis on sputum culture had previously had sputum negative for M. tuberculosis on culture and/or Xpert before sub-study enrolment. ¶ Participant had previously had negative sputum for mycobacterial culture at enrolment to sub-study. CXR ¼ chest radiograph; TB ¼ tuberculosis. iv The International Journal of Tuberculosis and Lung Disease disorders, malignancy, depression, inadequate access to food and drug misuse. Blood samples were collected from all participants with weight loss for renal, liver and thyroid function, full blood count and glycated haemoglobin (HbA1c; to identify type 2 diabetes mellitus), and clinically significant results were reported to the clinic physician for further management. If participants reported diarrhoea, stool samples were collected for microscopy, bacterial culture, parasitology and for Clostridium difficile, if antibiotics had been taken in the preceding 12 weeks.
The Patient Health Questionnaire-9 (PHQ-9) 3 was used to screen all sub-study participants for depression. This score categorises depression as 0-4 (none or minimal), 5-9 (mild), 10-14 (moderate), 15-19 (moderately severe) and 20-27 (severe). Participants with scores of 710 were evaluated further by the research physician, and referred to the clinic physician or psychology service if deemed clinically depressed. The Fast Alcohol Screening Test (FAST) 4 was used to screen for hazardous alcohol consumption (FAST score 7 3) and all participants were asked about drug misuse. If FAST score was 73, brief intervention was provided and, if appropriate, participants were referred to drug services.
We used the Household Food Insecurity Access Scale (HFIAS) 5 to measure food access, categorising the scale as 1) 'food secure', 2) 'mildly food insecure access', 3) 'moderately food insecure access' or 4) 'severely food insecure access'. A HFIAS score of 4 was deemed a cause of unintentional weight loss. Participants with food insecurity were referred to the clinic dietician.
Clinical features suggestive of possible malignancy were discussed with the clinic physician to facilitate further appropriate evaluation or treatment as deemed appropriate.
Evaluation of fever or night sweats Self-recorded participant temperature measurements were reviewed. If a likely focus of infection was identified, then relevant samples were submitted for appropriate microbiological evaluation such as midstream urine (bacterial culture), stool (bacterial culture, parasitology, Clostridium difficile), sputum (bacterial culture), blood (malaria film) or swabs (bacterial culture). If deemed clinically appropriate, antibiotics were provided and the participant reviewed to assess response to treatment.
If no likely focus infection was identified, blood was collected for renal, liver and thyroid function, full blood count, CRP, HbA1c, aerobic and anaerobic culture if documented fever .38.38C; and urine for microscopy and culture. Abdominal ultrasound was arranged and CXR, if either no recent film or no film since onset of symptoms.
A clinical diagnosis was made if no other more likely cause was identified, and symptoms were suggestive of perimenopausal vasomotor symptoms in females aged 745 years; in younger cases, blood samples were collected to determine follicle-stimulating hormone (FSH) levels.
If, during in-patient treatment, specialist referral or further evaluation such as lumbar puncture, CT imaging (abdomen, chest, sinus), fine-needle aspiration, bone-marrow aspiration or pleural aspiration were deemed necessary, this was facilitated by the clinic physician.
vi The International Journal of Tuberculosis and Lung Disease