Transactions of the Royal Society of Tropical Medicine and Hygiene Advance Access published October 30, 2014
ORIGINAL ARTICLE
Trans R Soc Trop Med Hyg doi:10.1093/trstmh/tru144
Predictors of mortality in leptospirosis: an observational study from two hospitals in Kolkata, eastern India Rudra P Goswamia,*, Rama P Goswamib, Ayan Basuc, Santanu Kumar Tripathid, Sanghamitra Chakrabartie and Indrajit Chattopadhyaye a
Department of Medicine, Institute of Post Graduate Medical Education and Research, Kolkata, India; bDepartment of Tropical Medicine, School of Tropical Medicine, Kolkata, West Bengal, India; cDepartment of Tropical Medicine, School of Tropical Medicine, Kolkata, West Bengal, India; dDepartment of Clinical and Experimental Pharmacology, School of Tropical Medicine, Kolkata, West Bengal, India; e Geriatric & General (Internal) Medicine, Glan Clwyd Hospital, Rhyl, Wales, LL18 5UJ, UK
Received 6 April 2014; revised 27 July 2014; accepted 28 July 2014 Background: Leptospirosis has a mortality rate of 5–20%. Poor prognostic factors are older age; oliguria; elevated potassium, creatinine and/or bilirubin levels; and altered mental status. We conducted this retrospective study to analyse the predictors of mortality among Indian patients with leptospirosis. Methods: Clinical, biochemical, demographic and treatment related data (time between onset of symptoms and commencement of leptospira specific antibiotics) of 101 leptospirosis patients were reviewed. Predictors identified by univariate analysis were analysed by multivariable Cox regression for survival analysis. Results: Prominent clinical features were: fever (101/101, 100%), jaundice (62, 62.4%), vomiting (42, 41.6%), oliguria (35, 34.7%), cough (18, 17.8%) and dyspnoea (10, 10.0%). Common complications were acute kidney injury (22, 21.8%), cardiovascular collapse (13, 12.9%), haemorrhages (10, 10.0%), meningitis (7, 6.9%), acute respiratory distress syndrome and pancreatitis (5, 5.0% each). Seventeen patients died (16.8%). Univariate predictors of mortality were older age, delayed antibiotic therapy, higher bilirubin, aspartate aminotransferase, alkaline phosphatase, leucocyte count and aspartate/alanine aminotransferase ratio (AAR). Only AAR (HR 1.208, 95% CI 1.051–1.388) and number of days the patient was symptomatic before access to specific antibiotic therapy (HR 1.304, 95% CI 1.081–1.574) remained significant predictors after Cox regression. Conclusions: Multivariate analysis showed high AAR and delayed antibiotic therapy might be associated with fatality. Keywords: Aspartate/alanine aminotransferase ratio (AAR), Early antibiotic therapy, India, Leptospirosis, Mortality, Prognostic factors
Introduction Leptospirosis is a widespread zoonosis of global distribution caused by pathogenic spirochetes of the genus Leptospira.1 The clinical spectrum of leptospirosis ranges from non-specific febrile episodes to severe forms. Severe disease is estimated to occur in 5–15% of all human infections. The classic description is that of Weil’s syndrome: the triad of jaundice, renal failure and haemorrhage,2 though Adolf Weil’s original description comprised of jaundice, renal failure and splenomegaly.3 Recently there has been a greater awareness worldwide regarding the protean manifestations of leptospirosis. These range from pulmonary haemorrhage and acute respiratory distress syndrome (ARDS) to pancreatitis, meningoencephalitis and haemophagocytic syndrome.1,4,5
Only recently has leptospirosis been identified as an important cause of acute febrile illness especially if the disease is protracted for more than a week. There are also a few reports that establish leptospira as a predominant pathogenic entity in post-inundation state in various cities of India.6 Case fatality rates of leptospirosis indicated by these studies range from 5–20%.1 Attempts have been made in various Western populations to identify the factors that lead to fatal forms of leptospirosis. Early evaluation of disease severity might be useful in improving prognosis. Worldwide, independent prognostic factors of a fatal outcome in leptospirosis have been found to include older age, oliguria, hyperkalaemia, abnormal serum creatinine level, ARDS, pulmonary haemorrhage, elevated bilirubin level, hypotension, arrhythmia and altered mental status.1,2,7 Delay in hospitalization
# The Author 2014. Published by Oxford University Press on behalf of Royal Society of Tropical Medicine and Hygiene. All rights reserved. For permissions, please e-mail: [email protected].
1
Downloaded from http://trstmh.oxfordjournals.org/ at UQ Library on November 9, 2014
*Corresponding author: Present address: Abhyudoy Housing, Flat–18/14, ECTP, Ph–IV, Type–B, EM Bypass, Kolkata–107. Tel: +91 943 353 4335; E-mail: [email protected]
R. P. Goswami et al.
may also be a determinant of poor outcome, though some tend to refute this hypothesis.7,8 Recently, in a small series, the issue of disproportionately raised aspartate aminotransferase (AST) level was addressed as a potential risk factor predicting mortality.4 We conducted this retrospective study to analyse the predictors of mortality in Indian patients with leptospirosis.
Materials and methods
2
Results Among the 101 patients (mean age 40.62 + 15.96 years, range 14–76 years), there were 17 deaths giving a mortality rate of 16.8%. Comparison of symptoms, signs and disease complexes are outlined in Table 1. Fever was the commonest symptom present in all patients followed by myalgia, headache, conjunctival suffusion and arthralgia. There was no significant difference in the incidence of these symptoms between survivors and nonsurvivors. Other prominent clinical features were jaundice (61.4%, 62/101), vomiting (41.6%, 42), oliguria (34.7%, 35), cough (17.8%, 18) and shortness of breath (10.0%, 10). Only jaundice was significantly more common among the non-survivors (88% (15/17) vs 56% (47/84), p¼0.01). Other complications included acute kidney injury (AKI) requiring haemodialysis (21.8%, 22/101), cardiovascular collapse (12.9%, 13), haemorrhagic manifestations (10.0%, 10), clinical meningitis with nuchal rigidity (7.0%, 7), ARDS (5.0%, 5) and pancreatitis (5.0%, 5). Figure 1 provides a pictorial description of these symptoms and complications. Table 2 shows univariate analysis of the demographic features and laboratory parameters. Non-survivors were significantly older (59 years vs 36.9 years, p,0.001), had not been prescribed specific antibiotics for a longer duration (10 days vs 4.5 days, p,0.001), had higher mean bilirubin level (13.26 vs 5.04 mg/dL, p,0.001), higher mean AST level (482 vs 159 IU/L, p,0.001), higher mean ALP level (344 vs 189 IU/L, p¼0.003), higher mean AAR (4.47 vs 1.22, p,0.001) and higher mean total leucocyte count (22 000 vs 10 000 /mL, p¼0.003). These data are pictorially represented in Supplementary Figure 1. Table 3 summarises the results of survival analysis by Cox proportional hazard regression. All variables which were identified as
Downloaded from http://trstmh.oxfordjournals.org/ at UQ Library on November 9, 2014
This retrospective study was performed in two tertiary care teaching hospitals in Kolkata, from July 2010 to October 2012. All patients admitted over this period with suspected leptospirosis were retrospectively included in the study. The investigators’ clinical records of patients managed under their care at inpatient and outpatient departments were analysed anonymously. Hospital records of patients’ clinical data and investigations were analysed with permission from the appropriate authorities. Waiver of approval was granted by the Clinical Research Ethics Committee, Calcutta School of Tropical Medicine, Kolkata. Clinical suspicion of leptospirosis was based on six clinical parameters (fever, jaundice, myalgia, conjunctival suffusion, headache and meningeal signs) and two laboratory parameters (albuminuria or azotaemia).7 In every case, the diagnosis of leptospirosis was made by one of the following laboratory measures: lateral flow immunoassay by rapid immunochromatographic test (ICT); ELISA IgM; or a positive culture. Fifty-nine patients were positive with ICT, on-site leptospira IgG/IgM combo-rapid test (CTK Biotech, Inc., San Diego, CA, USA). The test was done with cassette device using three line (control, IgG and IgM) lateral flow immunoassay. The test showed a positive result by appearance of two or three burgundy lines (one control and either one of IgM or IgG lines or both) and negative by appearance of only one burgundy control line. The manufacturer’s literature cited the sensitivity and specificity of the test as 90% and 99% for IgM and 100% and 99% for IgG, the overall agreement being about 98.6% for IgM and 99% for IgG. Forty patients were positive with IgM ELISA (Panbio, Inverness, Brisbane, Australia). The manufacturer’s recommended cut-off for a positive result is ≥11 Panbio units. The manufacturer reported sensitivity and specificity of the test for detection of recent leptospira infection to be 96.5% and 98.5% respectively. In two cases, blood samples were culture positive (EMJH medium–in-house). We also reviewed demographic data (age, gender and occupation), epidemiological data (time between the onset of clinical signs and hospitalization) and treatment related data (time between onset of symptoms and commencement of leptospira specific antibiotics). The clinical data collected included the presence or absence of jaundice, fever, oliguria (urine output ,0.5 L/24 h), cardiovascular collapse (systolic arterial pressure ,80 mm Hg), cough and dyspnoea (respiratory rate .20/minute). We noted complaints of abdominal pain, nausea, vomiting and diarrhoea. Neurological signs were defined by at least one of the following: drowsiness or alteration of consciousness (Glasgow coma score ,13/15), signs of meningitis (corroborated by documented cerebrospinal fluid abnormalities) or convulsions. We recorded the following hemorrhagic complications: epistaxis, conjunctival bleeding, haematuria, haematoma and gastrointestinal haemorrhage.
The following laboratory values were recorded at the time of admission to the hospital: hemoglobin (Hb), platelet count, total leucocyte count, urea, creatinine, presence or absence of hyperkalemia (serum potassium level .5.5 mmol/L), levels of serum bilirubin, AST, alanine aminotransferase (ALT), alkaline phosphatase (ALP), amylase and lipase and total (TLC) and differential (DLC) leucocyte counts. We calculated the AST/ALT ratio (AAR) for each of the patients. The length of hospital stay for each patient till discharge or death was also noted. Continuous data were presented as mean+SD. Unpaired t-test was used to test for the difference of the means between the survivors and the non-survivors. Categorical data were presented as proportions. The x2 test was used to test the difference of the proportion between the two groups. The onset of the disease was the date when the first symptom started, and the end of the analysis was the date when the patient died or was still alive at two weeks following discharge (the usual time of first follow-up). The event of interest was death from leptospirosis. Patients who were still alive at three weeks after discharge as per outpatient follow-up records were considered as survivors. Multivariable Cox regression model was performed to determine factors affecting patient’s survival, taking into account effects of several potential factors simultaneously and hazard ratios were obtained for significant predictors of mortality. All statistical analyses were done using SPSS ver. 16 (SPSS Inc., Chicago, IL, USA).
Transactions of the Royal Society of Tropical Medicine and Hygiene
Table 1. Clinical signs and symptoms in survivors and non-survivors of leptospirosis in patients in two hospitals in Kolkata, July 2010 to October 2012 Survivors Non-survivors p-value (n¼84) (n¼17)
Arthralgia Headache Myalgia Conjunctival suffusion Renal manifestations Oliguria Acute kidney injury requiring dialysis Cardiovascular collapse Gastrointestinal symptoms Jaundice Diarrhoea Vomiting Painful abdomen Pancreatitis Respiratory symptoms Cough Shortness of breath Acute respiratory distress syndrome Pulmonary rales Neurological manifestations Drowsiness/altered sensorium Seizure Meningitis Haemorrhagic manifestations
47% (40) 52% (9) 70% (59) 76% (13) 88% (74) 100% (17) 50% (42) 76% (13)
NS NS NS NS
30% (26) 19% (16)
52% (9) 35% (6)
NS NS
10% (9)
23% (4)
NS
55% (47) 2% (2) 45% (38) 10% (9) 3% (3)
88% (15) 5% (1) 23% (4) 23% (4) 11% (2)
0.01 NS NS NS NS
17% (15) 8% (7) 3% (3)
17% (3) 17% (3) 11% (2)
NS NS NS
26% (22)
47% (8)
NS
14% (12)
11% (2)
NS
2% (2) 7% (6) 9% (8)
0 5% (1) 11% (2)
NS NS NS
NS: not significant.
significant predictors of mortality in univariate tests (details in Tables 1 and 2) were analysed by the multivariate model. However as AST is the numerator in AAR, only age, days spent without specific antibiotics, presence of jaundice, total leucocyte count, ALP, bilirubin and AAR were analysed in the multivariate Cox regression model after entering all the variables and running a backward stepwise procedure. Only two factors remained significant predictors of mortality. These were the AAR (HR 1.208, 95% CI 1.051–1.388) and the number of symptomatic days before access to specific antibiotic therapy (HR 1.304, 95% CI 1.081–1.574).
Discussion Leptospirosis is a re-emerging disease. Whether this apparent resurgence is due to increased detection owing to better awareness amongst physicians or true re-emergence of the disease remains to be determined. However, the disease burden has certainly increased worldwide as well as in India. It is increasingly
3
Downloaded from http://trstmh.oxfordjournals.org/ at UQ Library on November 9, 2014
Clinical manifestation
being identified as an important cause of acute lung injury, ARDS, acute liver failure and AKI of unknown origin from studies undertaken in various parts of India.9–11 Though some recent Indian studies addressing epidemiological issues on leptospirosis have emerged, there is a lack of data regarding the predictors of mortality, despite the rather high mortality rate of this disease reported from India and abroad.12–15 The mortality rate from leptospirosis in our study was 16.8%. Worldwide mortality rates from the disease vary from 6% (France),16 through 10% (USA),17 14% (Thailand),1 17% (Brazil),2 18% (French West Indies),7 to 20% (Barbados).18 Though similar Indian data is sparse, reports indicate that the mortality rate from leptospirosis may vary from 6.0–15.4%.15,19 Clinical hepatitis with jaundice was the commonest organ related manifestation (63/101, 62.4%) in our study. This was followed by AKI requiring haemodialysis (22, 21.8%), haemorrhagic manifestations (11, 10.0%), meningitis (7, 6.9%), ARDS and pancreatitis (5, 5.0% each). This is in keeping with previous reports with renal failure and pulmonary involvement ranging from 17–71%.19 Prognostic factors associated with mortality have only been evaluated in a very few studies. In our data the overall predictors of fatality in univariate analysis were age, longer delays before commencing antibiotic therapy, hyperbilirubinemia, higher AST and ALP levels, higher AAR and leucocytosis. In a Polynesian cohort of 71 severe cases of leptospirosis, independent predictors of mortality were oliguria, hypotension and abnormal chest auscultation (presence of crackles or rhonchi).20 Similar results were available from Brazil.21 DuPont et al., in the French West Indies,7 reported the following factors associated with mortality: dyspnoea, oliguria, increased white blood cell count, repolarization abnormalities on electrocardiograms and alveolar infiltrates on chest radiographs. Courtin et al. showed that the mortality rate rose from 4.8 to 11% when patients developed respiratory symptoms.22A Thai cohort showed mortality being related to hypotension, oliguria, hyperkalaemia, and presence of pulmonary rales.1 The finding of age as a significant predictor is not surprising as age is often reported as the most significant host factor related to mortality in leptospirosis.4 The average age in our series was higher (around 40 years) and thus accurate implications on age as a significant predictor cannot be made at this point. Severe pulmonary and haemorrhagic manifestations were rather infrequent (,10%) and hence were not significantly associated with death. Though oliguria has been repeatedly shown to be significantly associated with death in many previous studies, our results differ and we believe it to be justifiable. Our patients were admitted to tertiary care referral centres with fully functional haemodialysis units and we successfully dialysed patients with AKI whenever indicated. We believe this might be the cornerstone of management of leptospirosis and is likely to have led to reduction in mortality from AKI. The only Indian report in relation to mortality found dyspnoea, meningism, oliguria, haemoptysis, abdominal pain, disorientation, muscle tenderness, tachycardia, hyperkalaemia and hyperbilirubinaemia (.15 mg/dL) to be significant predictors of death by univariate analysis. Binary logistic regression revealed lung and central nervous system affection to be associated with death.19 However, they did not factor in the time to initiate
R. P. Goswami et al.
Table 2. Demographic features and laboratory values: univariate analysis of predictors of mortality from leptospirosis in patients in two hospitals in Kolkata, July 2010 to October 2012 Factors (analysis by Student’s t test)
Age in years Days spent without specific antibiotics Urea (mg/dL) Creatinine (mg/dL) Total bilirubin (mg/dL) Aspartate aminotranferase (IU/L) (AST) Alanine aminotransferase (IU/L) (ALT) Alkaline phosphatase (IU/L) AST/ALT Ratio (AAR) Total leucocyte count (/mL) Platelet count (10^5 /mL) Haemoglobin (g/dL) Factors (analysis by x2 test) Hyperkalemia (n, %)
Survivors (n¼84)
Non-survivors (n¼17)
p-value
Mean
SD
Mean
SD
36.91 4.58 70.48 2.67 5.04 159.89 153.70 189.68 1.22 10811.81 1.13 11.13
15.12 2.56 61.35 3.21 6.96 214.58 218.01 169.89 0.62 5339.73 0.81 2.42
59.05 10.00 81.88 4.32 13.26 482.88 197.59 344.06 4.47 22059.56 1.14 10.51
13.27 3.46 29.21 2.33 10.29 651.67 470.84 258.98 3.53 12900.43 0.07 2.71
11
13
5
29
,0.001 ,0.001 NS NS ,0.001 ,0.001 NS 0.003 ,0.001 0.003 NS NS NS
NS: not significant.
antibiotics and the AAR and did not employ Cox regression which is better than binary logistic regression for survival analysis. We demonstrated, following multivariate analysis, that early antibiotic therapy might offer a survival advantage. Delay in referral to a specialized infectious disease hospital has been implicated in a higher probability of lethal outcome in Salvador.8 It has already been shown by previous researchers that early antibiotic therapy reduces duration of fever, renal failure and hospital stay.23,24 Both DuPont et al. and Spichler et al. maintain that delayed hospitalization is not an independent factor predicting a worse outcome.2,7 We would however emphasize that it is early and specific antibiotic therapy, rather than the timing of hospitalization, that provides survival advantage. The HR for delay in
4
specific antibiotic therapy is 1.304 which translates into a 30% increased risk of death. Despite this apparently small effect it is worthwhile to note that Cox regression has not been used in the context of leptospirosis in sufficiently large populations so that estimate of HR is yet unknown. The same argument holds of AAR which has an HR of 1.208 which translates into a 21% increased risk of death. Recently it has been recognised that exaggerated elevation of AST in comparison to ALT might be associated with death. The source of this exaggerated AST response has been thought to be extra-hepatic but the exact localisation is imprecise. Damage to parenchymal cellular membrane leading to severe and extensive destruction of cells in various organs might lead to the unusual
Downloaded from http://trstmh.oxfordjournals.org/ at UQ Library on November 9, 2014
Figure 1. Composite bar diagram showing distribution of symptoms and complications of leptospirosis between survivors and non-survivors in two hospitals in Kolkata, July 2010 to October 2012. This figure is available in black and white in print and in colour at Transactions online.
Transactions of the Royal Society of Tropical Medicine and Hygiene
Table 3. Relative risk of death from leptospirosis for each selected factor using Cox regression model in patients in two hospitals in Kolkata, July 2010 to October 2012 Factors
Coefficient of regression SE (B coefficient)
Wald
p-value
Hazard ratio (exp[B])
95% CI for HR
Age in years Jaundice Days spent without specific antibiotics Total bilirubin Alkaline phosphatase AAR Total leucocyte count
0.03 -1.64 0.266
0.244 1.21 0.096
1.585 1.847 7.705
0.208 0.17 0.006
1.03 0.193 1.304
0.98–1.07 0.018–2.071 1.081–1.574
0.013 0.002 0.189 0.001
0.029 0.001 0.071 0.001
0.211 1.941 7.102 0.871
0.646 0.146 0.008 0.351
1.013 1.002 1.208 1.001
0.958–1.072 0.999–1.005 1.051–1.388 0.998–1.002
ASTelevation.4 To check for extra-hepatic source of AST we formulated the index AAR and this was a significant factor in both univariate and multivariate analyses. Since creatinine phosphokinase (CPK) is also a marker for muscle damage this, like the AAR, may also be related to mortality in leptospirosis. However the importance of AAR is based on the ease of calculating it and the comparative low cost–an important factor in a country like India. Though it would have been interesting to see the effect of CPK on mortality, unfortunately while analysing the records retrospectively, it was noted that CPK levels were done in less than half of the cases. However, whenever the CPK estimation was done, this was found to be raised ranging from mild elevation to very high levels. The present study had several limitations like its retrospective nature; inability to incorporate CPK and electrocardiographic findings as reports of all the patients were not available; and inability to incorporate serial chest imaging findings and serial laboratory data analyses.
Conclusions In conclusion we have demonstrated that, despite many apparent clinical and biochemical predictors of mortality in leptospirosis, early diagnosis and treatment remains the cornerstone of successful therapy and survival. The exaggerated AST response in leptospirosis is an ominous sign and the AAR appears to be an important prognostic marker. Despite the new findings, our study is however limited by the retrospective nature of the design and unavailability of autopsy data.
Supplementary data Supplementary data are available at Transactions Online (http:// trstmh.oxfordjournals.org/).
Authors’ contributions: RPG (IPGMER) and RPG (STM) conceived the study; RPG (IPGMER), RPG (STM), SKT, SC and IC designed the study protocol; RPG (IPGMER), RPG (STM) and AB carried out the clinical assessment and data
collection; RPG (IPGMER) carried out analysis and interpretation of these data; RPG (IPGMER) and RPG (STM) drafted the manuscript; RPG (STM), SC and IC critically revised the manuscript for intellectual content. All authors read and approved the final manuscript. RPG (IPGMER) is the guarantor of the paper. Funding: None. Competing interests: None declared. Ethical approval: Waiver of approval was granted by the Clinical Research Ethics Committee, Calcutta School of Tropical Medicine, Kolkata.
References 1 Panaphut T, Domrongkitchaiporn S, Thinkamrop B. Prognostic factors of death in leptospirosis: a prospective cohort study in Khon Kaen, Thailand. Int J Infect Dis 2002;6:52–9. 2 Spichler AS, Vilac¸a PJ, Athanazio DA et al. Predictors of lethality in severe leptospirosis in urban Brazil. Am J Trop Med Hyg 2008;79: 911–4. 3 Weil A. Uber eine eigenthumliche, mit Milz-tumor, Icterus und Nephritis einhergehende, acute Infectionskarankheit. Deutsches Arch Klin Med 1886;39:209. 4 Chang ML, Yang CW, Chen JC et al. Disproportional exaggerated aspartate transaminase is a useful prognostic parameter in late leptospirosis. World J Gastroenterol 2005;11:5553–6. 5 Krishnamurthy S, Mahadevan S, Mandal J, Basu D. Leptospirosis in association with hemophagocytic syndrome: a rare presentation. Indian J Pediatr 2013;80:524–5. 6 Sambasiva RR, Naveen G, P B, Agarwal SK. Leptospirosis in India and the rest of the world. Braz J Infect Dis 2003;7:178–93. 7 DuPont H, Dupont-Perdrizet D, Perie JL et al. Leptospirosis: prognostic factors associated with mortality. Clin Infect Dis 1997;25:720–4. 8 Flannery B, Pereira MM, Velloso L, de F et al. Referral pattern of leptospirosis cases during a large urban epidemic of dengue. Am J Trop Med Hyg 2001;65:657–63.
5
Downloaded from http://trstmh.oxfordjournals.org/ at UQ Library on November 9, 2014
AAR: aspartate aminotranferase/alanine aminotransferase ratio.
R. P. Goswami et al.
9 Bhadade RR, de Souza RA, Harde MJ, Khot A. Clinical characteristics and outcomes of patients with acute lung injury and ARDS. J Postgrad Med 2011;57:286–90.
17 Health CW, Alexander AD, Galton MM. Leptospirosis in the United States: analysis of 483 cases in man 1949–61. N Engl J Med 1965;273:915–22.
10 Deepak NA, Patel ND. Differential diagnosis of acute liver failure in India. Ann Hepatol 2006;5:150–6.
18 Edwards CN, Nicholson GD, Hassell TA et al. Leptospirosis in Barbados. A clinical study. West Indian Med J 1990;39:27–34.
11 Jha V, Chugh KS. Community-acquired acute kidney injury in Asia. Semin Nephrol 2008;28:330–47.
19 Pappachan MJ, Mathew S, Aravindan KP et al. Risk factors for mortality in patients with leptospirosis during an epidemic in northern Kerala. Natl Med J India 2004;17:240–3.
12 Debmandal M, Mandal S, Pal NK. Serologic evidence of human leptospirosis in and around Kolkata, India: a clinico-epidemiological study. Asian Pac J Trop Med 2011;4:1001–6. 13 Chauhan V, Mahesh DM, Panda P et al. Profile of patients of leptospirosis in sub-Himalayan region of North India. J Assoc Physicians India 2010;58:354–6.
20 Doudier B, Garcia S, Quennee V et al. Prognostic factors associated with severe leptospirosis. Clin Microbiol Infect 2006;12: 299–300. 21 Ko AI, Reis MG, Dourado CM et al. Urban epidemic of severe leptospirosis in Brazil. Lancet 1999;354:820–5. 22 Courtin JP, Di Francia M, DuCouedic L et al. Respiratory manifestations of leptospirosis: retrospective study of 91 cases (1978–1994) [in French]. Rev Pneumol Clin 1998;54:382–92.
15 Ittyachen AM, Krishnapillai TV, Nair MC, Rajan AR. Retrospective study of severe cases of leptospirosis admitted in the intensive care unit. J Postgrad Med 2007;53:232–5.
23 Watt G, Padre LP, Tuazon ML et al. Placebo-controlled trial of intravenous penicillin for severe and late leptospirosis. Lancet 1988;327:433–5.
16 Bourrier P, Chennbault JM, Achard J et al. Leptospirosis: retrospective analysis of 99 cases observed in 10 years in the Central West of France [in French]. Me´d Mal Infect 1988;1:4–8.
24 McClain JBL, Ballou WR, Harrisson SM, Steinweg DL. Doxycycline therapy for leptospirosis. Ann Intern Med 1984;100: 696–8.
6
Downloaded from http://trstmh.oxfordjournals.org/ at UQ Library on November 9, 2014
14 Sethi S, Sharma N, Kakkar N et al. Increasing trends of leptospirosis in northern India: a clinico-epidemiological study. PLoS Negl Trop Dis 2010;4:e579.
ORIGINAL ARTICLE
Trans R Soc Trop Med Hyg doi:10.1093/trstmh/tru144
Predictors of mortality in leptospirosis: an observational study from two hospitals in Kolkata, eastern India Rudra P Goswamia,*, Rama P Goswamib, Ayan Basuc, Santanu Kumar Tripathid, Sanghamitra Chakrabartie and Indrajit Chattopadhyaye a
Department of Medicine, Institute of Post Graduate Medical Education and Research, Kolkata, India; bDepartment of Tropical Medicine, School of Tropical Medicine, Kolkata, West Bengal, India; cDepartment of Tropical Medicine, School of Tropical Medicine, Kolkata, West Bengal, India; dDepartment of Clinical and Experimental Pharmacology, School of Tropical Medicine, Kolkata, West Bengal, India; e Geriatric & General (Internal) Medicine, Glan Clwyd Hospital, Rhyl, Wales, LL18 5UJ, UK
Received 6 April 2014; revised 27 July 2014; accepted 28 July 2014 Background: Leptospirosis has a mortality rate of 5–20%. Poor prognostic factors are older age; oliguria; elevated potassium, creatinine and/or bilirubin levels; and altered mental status. We conducted this retrospective study to analyse the predictors of mortality among Indian patients with leptospirosis. Methods: Clinical, biochemical, demographic and treatment related data (time between onset of symptoms and commencement of leptospira specific antibiotics) of 101 leptospirosis patients were reviewed. Predictors identified by univariate analysis were analysed by multivariable Cox regression for survival analysis. Results: Prominent clinical features were: fever (101/101, 100%), jaundice (62, 62.4%), vomiting (42, 41.6%), oliguria (35, 34.7%), cough (18, 17.8%) and dyspnoea (10, 10.0%). Common complications were acute kidney injury (22, 21.8%), cardiovascular collapse (13, 12.9%), haemorrhages (10, 10.0%), meningitis (7, 6.9%), acute respiratory distress syndrome and pancreatitis (5, 5.0% each). Seventeen patients died (16.8%). Univariate predictors of mortality were older age, delayed antibiotic therapy, higher bilirubin, aspartate aminotransferase, alkaline phosphatase, leucocyte count and aspartate/alanine aminotransferase ratio (AAR). Only AAR (HR 1.208, 95% CI 1.051–1.388) and number of days the patient was symptomatic before access to specific antibiotic therapy (HR 1.304, 95% CI 1.081–1.574) remained significant predictors after Cox regression. Conclusions: Multivariate analysis showed high AAR and delayed antibiotic therapy might be associated with fatality. Keywords: Aspartate/alanine aminotransferase ratio (AAR), Early antibiotic therapy, India, Leptospirosis, Mortality, Prognostic factors
Introduction Leptospirosis is a widespread zoonosis of global distribution caused by pathogenic spirochetes of the genus Leptospira.1 The clinical spectrum of leptospirosis ranges from non-specific febrile episodes to severe forms. Severe disease is estimated to occur in 5–15% of all human infections. The classic description is that of Weil’s syndrome: the triad of jaundice, renal failure and haemorrhage,2 though Adolf Weil’s original description comprised of jaundice, renal failure and splenomegaly.3 Recently there has been a greater awareness worldwide regarding the protean manifestations of leptospirosis. These range from pulmonary haemorrhage and acute respiratory distress syndrome (ARDS) to pancreatitis, meningoencephalitis and haemophagocytic syndrome.1,4,5
Only recently has leptospirosis been identified as an important cause of acute febrile illness especially if the disease is protracted for more than a week. There are also a few reports that establish leptospira as a predominant pathogenic entity in post-inundation state in various cities of India.6 Case fatality rates of leptospirosis indicated by these studies range from 5–20%.1 Attempts have been made in various Western populations to identify the factors that lead to fatal forms of leptospirosis. Early evaluation of disease severity might be useful in improving prognosis. Worldwide, independent prognostic factors of a fatal outcome in leptospirosis have been found to include older age, oliguria, hyperkalaemia, abnormal serum creatinine level, ARDS, pulmonary haemorrhage, elevated bilirubin level, hypotension, arrhythmia and altered mental status.1,2,7 Delay in hospitalization
# The Author 2014. Published by Oxford University Press on behalf of Royal Society of Tropical Medicine and Hygiene. All rights reserved. For permissions, please e-mail: [email protected].
1
Downloaded from http://trstmh.oxfordjournals.org/ at UQ Library on November 9, 2014
*Corresponding author: Present address: Abhyudoy Housing, Flat–18/14, ECTP, Ph–IV, Type–B, EM Bypass, Kolkata–107. Tel: +91 943 353 4335; E-mail: [email protected]
R. P. Goswami et al.
may also be a determinant of poor outcome, though some tend to refute this hypothesis.7,8 Recently, in a small series, the issue of disproportionately raised aspartate aminotransferase (AST) level was addressed as a potential risk factor predicting mortality.4 We conducted this retrospective study to analyse the predictors of mortality in Indian patients with leptospirosis.
Materials and methods
2
Results Among the 101 patients (mean age 40.62 + 15.96 years, range 14–76 years), there were 17 deaths giving a mortality rate of 16.8%. Comparison of symptoms, signs and disease complexes are outlined in Table 1. Fever was the commonest symptom present in all patients followed by myalgia, headache, conjunctival suffusion and arthralgia. There was no significant difference in the incidence of these symptoms between survivors and nonsurvivors. Other prominent clinical features were jaundice (61.4%, 62/101), vomiting (41.6%, 42), oliguria (34.7%, 35), cough (17.8%, 18) and shortness of breath (10.0%, 10). Only jaundice was significantly more common among the non-survivors (88% (15/17) vs 56% (47/84), p¼0.01). Other complications included acute kidney injury (AKI) requiring haemodialysis (21.8%, 22/101), cardiovascular collapse (12.9%, 13), haemorrhagic manifestations (10.0%, 10), clinical meningitis with nuchal rigidity (7.0%, 7), ARDS (5.0%, 5) and pancreatitis (5.0%, 5). Figure 1 provides a pictorial description of these symptoms and complications. Table 2 shows univariate analysis of the demographic features and laboratory parameters. Non-survivors were significantly older (59 years vs 36.9 years, p,0.001), had not been prescribed specific antibiotics for a longer duration (10 days vs 4.5 days, p,0.001), had higher mean bilirubin level (13.26 vs 5.04 mg/dL, p,0.001), higher mean AST level (482 vs 159 IU/L, p,0.001), higher mean ALP level (344 vs 189 IU/L, p¼0.003), higher mean AAR (4.47 vs 1.22, p,0.001) and higher mean total leucocyte count (22 000 vs 10 000 /mL, p¼0.003). These data are pictorially represented in Supplementary Figure 1. Table 3 summarises the results of survival analysis by Cox proportional hazard regression. All variables which were identified as
Downloaded from http://trstmh.oxfordjournals.org/ at UQ Library on November 9, 2014
This retrospective study was performed in two tertiary care teaching hospitals in Kolkata, from July 2010 to October 2012. All patients admitted over this period with suspected leptospirosis were retrospectively included in the study. The investigators’ clinical records of patients managed under their care at inpatient and outpatient departments were analysed anonymously. Hospital records of patients’ clinical data and investigations were analysed with permission from the appropriate authorities. Waiver of approval was granted by the Clinical Research Ethics Committee, Calcutta School of Tropical Medicine, Kolkata. Clinical suspicion of leptospirosis was based on six clinical parameters (fever, jaundice, myalgia, conjunctival suffusion, headache and meningeal signs) and two laboratory parameters (albuminuria or azotaemia).7 In every case, the diagnosis of leptospirosis was made by one of the following laboratory measures: lateral flow immunoassay by rapid immunochromatographic test (ICT); ELISA IgM; or a positive culture. Fifty-nine patients were positive with ICT, on-site leptospira IgG/IgM combo-rapid test (CTK Biotech, Inc., San Diego, CA, USA). The test was done with cassette device using three line (control, IgG and IgM) lateral flow immunoassay. The test showed a positive result by appearance of two or three burgundy lines (one control and either one of IgM or IgG lines or both) and negative by appearance of only one burgundy control line. The manufacturer’s literature cited the sensitivity and specificity of the test as 90% and 99% for IgM and 100% and 99% for IgG, the overall agreement being about 98.6% for IgM and 99% for IgG. Forty patients were positive with IgM ELISA (Panbio, Inverness, Brisbane, Australia). The manufacturer’s recommended cut-off for a positive result is ≥11 Panbio units. The manufacturer reported sensitivity and specificity of the test for detection of recent leptospira infection to be 96.5% and 98.5% respectively. In two cases, blood samples were culture positive (EMJH medium–in-house). We also reviewed demographic data (age, gender and occupation), epidemiological data (time between the onset of clinical signs and hospitalization) and treatment related data (time between onset of symptoms and commencement of leptospira specific antibiotics). The clinical data collected included the presence or absence of jaundice, fever, oliguria (urine output ,0.5 L/24 h), cardiovascular collapse (systolic arterial pressure ,80 mm Hg), cough and dyspnoea (respiratory rate .20/minute). We noted complaints of abdominal pain, nausea, vomiting and diarrhoea. Neurological signs were defined by at least one of the following: drowsiness or alteration of consciousness (Glasgow coma score ,13/15), signs of meningitis (corroborated by documented cerebrospinal fluid abnormalities) or convulsions. We recorded the following hemorrhagic complications: epistaxis, conjunctival bleeding, haematuria, haematoma and gastrointestinal haemorrhage.
The following laboratory values were recorded at the time of admission to the hospital: hemoglobin (Hb), platelet count, total leucocyte count, urea, creatinine, presence or absence of hyperkalemia (serum potassium level .5.5 mmol/L), levels of serum bilirubin, AST, alanine aminotransferase (ALT), alkaline phosphatase (ALP), amylase and lipase and total (TLC) and differential (DLC) leucocyte counts. We calculated the AST/ALT ratio (AAR) for each of the patients. The length of hospital stay for each patient till discharge or death was also noted. Continuous data were presented as mean+SD. Unpaired t-test was used to test for the difference of the means between the survivors and the non-survivors. Categorical data were presented as proportions. The x2 test was used to test the difference of the proportion between the two groups. The onset of the disease was the date when the first symptom started, and the end of the analysis was the date when the patient died or was still alive at two weeks following discharge (the usual time of first follow-up). The event of interest was death from leptospirosis. Patients who were still alive at three weeks after discharge as per outpatient follow-up records were considered as survivors. Multivariable Cox regression model was performed to determine factors affecting patient’s survival, taking into account effects of several potential factors simultaneously and hazard ratios were obtained for significant predictors of mortality. All statistical analyses were done using SPSS ver. 16 (SPSS Inc., Chicago, IL, USA).
Transactions of the Royal Society of Tropical Medicine and Hygiene
Table 1. Clinical signs and symptoms in survivors and non-survivors of leptospirosis in patients in two hospitals in Kolkata, July 2010 to October 2012 Survivors Non-survivors p-value (n¼84) (n¼17)
Arthralgia Headache Myalgia Conjunctival suffusion Renal manifestations Oliguria Acute kidney injury requiring dialysis Cardiovascular collapse Gastrointestinal symptoms Jaundice Diarrhoea Vomiting Painful abdomen Pancreatitis Respiratory symptoms Cough Shortness of breath Acute respiratory distress syndrome Pulmonary rales Neurological manifestations Drowsiness/altered sensorium Seizure Meningitis Haemorrhagic manifestations
47% (40) 52% (9) 70% (59) 76% (13) 88% (74) 100% (17) 50% (42) 76% (13)
NS NS NS NS
30% (26) 19% (16)
52% (9) 35% (6)
NS NS
10% (9)
23% (4)
NS
55% (47) 2% (2) 45% (38) 10% (9) 3% (3)
88% (15) 5% (1) 23% (4) 23% (4) 11% (2)
0.01 NS NS NS NS
17% (15) 8% (7) 3% (3)
17% (3) 17% (3) 11% (2)
NS NS NS
26% (22)
47% (8)
NS
14% (12)
11% (2)
NS
2% (2) 7% (6) 9% (8)
0 5% (1) 11% (2)
NS NS NS
NS: not significant.
significant predictors of mortality in univariate tests (details in Tables 1 and 2) were analysed by the multivariate model. However as AST is the numerator in AAR, only age, days spent without specific antibiotics, presence of jaundice, total leucocyte count, ALP, bilirubin and AAR were analysed in the multivariate Cox regression model after entering all the variables and running a backward stepwise procedure. Only two factors remained significant predictors of mortality. These were the AAR (HR 1.208, 95% CI 1.051–1.388) and the number of symptomatic days before access to specific antibiotic therapy (HR 1.304, 95% CI 1.081–1.574).
Discussion Leptospirosis is a re-emerging disease. Whether this apparent resurgence is due to increased detection owing to better awareness amongst physicians or true re-emergence of the disease remains to be determined. However, the disease burden has certainly increased worldwide as well as in India. It is increasingly
3
Downloaded from http://trstmh.oxfordjournals.org/ at UQ Library on November 9, 2014
Clinical manifestation
being identified as an important cause of acute lung injury, ARDS, acute liver failure and AKI of unknown origin from studies undertaken in various parts of India.9–11 Though some recent Indian studies addressing epidemiological issues on leptospirosis have emerged, there is a lack of data regarding the predictors of mortality, despite the rather high mortality rate of this disease reported from India and abroad.12–15 The mortality rate from leptospirosis in our study was 16.8%. Worldwide mortality rates from the disease vary from 6% (France),16 through 10% (USA),17 14% (Thailand),1 17% (Brazil),2 18% (French West Indies),7 to 20% (Barbados).18 Though similar Indian data is sparse, reports indicate that the mortality rate from leptospirosis may vary from 6.0–15.4%.15,19 Clinical hepatitis with jaundice was the commonest organ related manifestation (63/101, 62.4%) in our study. This was followed by AKI requiring haemodialysis (22, 21.8%), haemorrhagic manifestations (11, 10.0%), meningitis (7, 6.9%), ARDS and pancreatitis (5, 5.0% each). This is in keeping with previous reports with renal failure and pulmonary involvement ranging from 17–71%.19 Prognostic factors associated with mortality have only been evaluated in a very few studies. In our data the overall predictors of fatality in univariate analysis were age, longer delays before commencing antibiotic therapy, hyperbilirubinemia, higher AST and ALP levels, higher AAR and leucocytosis. In a Polynesian cohort of 71 severe cases of leptospirosis, independent predictors of mortality were oliguria, hypotension and abnormal chest auscultation (presence of crackles or rhonchi).20 Similar results were available from Brazil.21 DuPont et al., in the French West Indies,7 reported the following factors associated with mortality: dyspnoea, oliguria, increased white blood cell count, repolarization abnormalities on electrocardiograms and alveolar infiltrates on chest radiographs. Courtin et al. showed that the mortality rate rose from 4.8 to 11% when patients developed respiratory symptoms.22A Thai cohort showed mortality being related to hypotension, oliguria, hyperkalaemia, and presence of pulmonary rales.1 The finding of age as a significant predictor is not surprising as age is often reported as the most significant host factor related to mortality in leptospirosis.4 The average age in our series was higher (around 40 years) and thus accurate implications on age as a significant predictor cannot be made at this point. Severe pulmonary and haemorrhagic manifestations were rather infrequent (,10%) and hence were not significantly associated with death. Though oliguria has been repeatedly shown to be significantly associated with death in many previous studies, our results differ and we believe it to be justifiable. Our patients were admitted to tertiary care referral centres with fully functional haemodialysis units and we successfully dialysed patients with AKI whenever indicated. We believe this might be the cornerstone of management of leptospirosis and is likely to have led to reduction in mortality from AKI. The only Indian report in relation to mortality found dyspnoea, meningism, oliguria, haemoptysis, abdominal pain, disorientation, muscle tenderness, tachycardia, hyperkalaemia and hyperbilirubinaemia (.15 mg/dL) to be significant predictors of death by univariate analysis. Binary logistic regression revealed lung and central nervous system affection to be associated with death.19 However, they did not factor in the time to initiate
R. P. Goswami et al.
Table 2. Demographic features and laboratory values: univariate analysis of predictors of mortality from leptospirosis in patients in two hospitals in Kolkata, July 2010 to October 2012 Factors (analysis by Student’s t test)
Age in years Days spent without specific antibiotics Urea (mg/dL) Creatinine (mg/dL) Total bilirubin (mg/dL) Aspartate aminotranferase (IU/L) (AST) Alanine aminotransferase (IU/L) (ALT) Alkaline phosphatase (IU/L) AST/ALT Ratio (AAR) Total leucocyte count (/mL) Platelet count (10^5 /mL) Haemoglobin (g/dL) Factors (analysis by x2 test) Hyperkalemia (n, %)
Survivors (n¼84)
Non-survivors (n¼17)
p-value
Mean
SD
Mean
SD
36.91 4.58 70.48 2.67 5.04 159.89 153.70 189.68 1.22 10811.81 1.13 11.13
15.12 2.56 61.35 3.21 6.96 214.58 218.01 169.89 0.62 5339.73 0.81 2.42
59.05 10.00 81.88 4.32 13.26 482.88 197.59 344.06 4.47 22059.56 1.14 10.51
13.27 3.46 29.21 2.33 10.29 651.67 470.84 258.98 3.53 12900.43 0.07 2.71
11
13
5
29
,0.001 ,0.001 NS NS ,0.001 ,0.001 NS 0.003 ,0.001 0.003 NS NS NS
NS: not significant.
antibiotics and the AAR and did not employ Cox regression which is better than binary logistic regression for survival analysis. We demonstrated, following multivariate analysis, that early antibiotic therapy might offer a survival advantage. Delay in referral to a specialized infectious disease hospital has been implicated in a higher probability of lethal outcome in Salvador.8 It has already been shown by previous researchers that early antibiotic therapy reduces duration of fever, renal failure and hospital stay.23,24 Both DuPont et al. and Spichler et al. maintain that delayed hospitalization is not an independent factor predicting a worse outcome.2,7 We would however emphasize that it is early and specific antibiotic therapy, rather than the timing of hospitalization, that provides survival advantage. The HR for delay in
4
specific antibiotic therapy is 1.304 which translates into a 30% increased risk of death. Despite this apparently small effect it is worthwhile to note that Cox regression has not been used in the context of leptospirosis in sufficiently large populations so that estimate of HR is yet unknown. The same argument holds of AAR which has an HR of 1.208 which translates into a 21% increased risk of death. Recently it has been recognised that exaggerated elevation of AST in comparison to ALT might be associated with death. The source of this exaggerated AST response has been thought to be extra-hepatic but the exact localisation is imprecise. Damage to parenchymal cellular membrane leading to severe and extensive destruction of cells in various organs might lead to the unusual
Downloaded from http://trstmh.oxfordjournals.org/ at UQ Library on November 9, 2014
Figure 1. Composite bar diagram showing distribution of symptoms and complications of leptospirosis between survivors and non-survivors in two hospitals in Kolkata, July 2010 to October 2012. This figure is available in black and white in print and in colour at Transactions online.
Transactions of the Royal Society of Tropical Medicine and Hygiene
Table 3. Relative risk of death from leptospirosis for each selected factor using Cox regression model in patients in two hospitals in Kolkata, July 2010 to October 2012 Factors
Coefficient of regression SE (B coefficient)
Wald
p-value
Hazard ratio (exp[B])
95% CI for HR
Age in years Jaundice Days spent without specific antibiotics Total bilirubin Alkaline phosphatase AAR Total leucocyte count
0.03 -1.64 0.266
0.244 1.21 0.096
1.585 1.847 7.705
0.208 0.17 0.006
1.03 0.193 1.304
0.98–1.07 0.018–2.071 1.081–1.574
0.013 0.002 0.189 0.001
0.029 0.001 0.071 0.001
0.211 1.941 7.102 0.871
0.646 0.146 0.008 0.351
1.013 1.002 1.208 1.001
0.958–1.072 0.999–1.005 1.051–1.388 0.998–1.002
ASTelevation.4 To check for extra-hepatic source of AST we formulated the index AAR and this was a significant factor in both univariate and multivariate analyses. Since creatinine phosphokinase (CPK) is also a marker for muscle damage this, like the AAR, may also be related to mortality in leptospirosis. However the importance of AAR is based on the ease of calculating it and the comparative low cost–an important factor in a country like India. Though it would have been interesting to see the effect of CPK on mortality, unfortunately while analysing the records retrospectively, it was noted that CPK levels were done in less than half of the cases. However, whenever the CPK estimation was done, this was found to be raised ranging from mild elevation to very high levels. The present study had several limitations like its retrospective nature; inability to incorporate CPK and electrocardiographic findings as reports of all the patients were not available; and inability to incorporate serial chest imaging findings and serial laboratory data analyses.
Conclusions In conclusion we have demonstrated that, despite many apparent clinical and biochemical predictors of mortality in leptospirosis, early diagnosis and treatment remains the cornerstone of successful therapy and survival. The exaggerated AST response in leptospirosis is an ominous sign and the AAR appears to be an important prognostic marker. Despite the new findings, our study is however limited by the retrospective nature of the design and unavailability of autopsy data.
Supplementary data Supplementary data are available at Transactions Online (http:// trstmh.oxfordjournals.org/).
Authors’ contributions: RPG (IPGMER) and RPG (STM) conceived the study; RPG (IPGMER), RPG (STM), SKT, SC and IC designed the study protocol; RPG (IPGMER), RPG (STM) and AB carried out the clinical assessment and data
collection; RPG (IPGMER) carried out analysis and interpretation of these data; RPG (IPGMER) and RPG (STM) drafted the manuscript; RPG (STM), SC and IC critically revised the manuscript for intellectual content. All authors read and approved the final manuscript. RPG (IPGMER) is the guarantor of the paper. Funding: None. Competing interests: None declared. Ethical approval: Waiver of approval was granted by the Clinical Research Ethics Committee, Calcutta School of Tropical Medicine, Kolkata.
References 1 Panaphut T, Domrongkitchaiporn S, Thinkamrop B. Prognostic factors of death in leptospirosis: a prospective cohort study in Khon Kaen, Thailand. Int J Infect Dis 2002;6:52–9. 2 Spichler AS, Vilac¸a PJ, Athanazio DA et al. Predictors of lethality in severe leptospirosis in urban Brazil. Am J Trop Med Hyg 2008;79: 911–4. 3 Weil A. Uber eine eigenthumliche, mit Milz-tumor, Icterus und Nephritis einhergehende, acute Infectionskarankheit. Deutsches Arch Klin Med 1886;39:209. 4 Chang ML, Yang CW, Chen JC et al. Disproportional exaggerated aspartate transaminase is a useful prognostic parameter in late leptospirosis. World J Gastroenterol 2005;11:5553–6. 5 Krishnamurthy S, Mahadevan S, Mandal J, Basu D. Leptospirosis in association with hemophagocytic syndrome: a rare presentation. Indian J Pediatr 2013;80:524–5. 6 Sambasiva RR, Naveen G, P B, Agarwal SK. Leptospirosis in India and the rest of the world. Braz J Infect Dis 2003;7:178–93. 7 DuPont H, Dupont-Perdrizet D, Perie JL et al. Leptospirosis: prognostic factors associated with mortality. Clin Infect Dis 1997;25:720–4. 8 Flannery B, Pereira MM, Velloso L, de F et al. Referral pattern of leptospirosis cases during a large urban epidemic of dengue. Am J Trop Med Hyg 2001;65:657–63.
5
Downloaded from http://trstmh.oxfordjournals.org/ at UQ Library on November 9, 2014
AAR: aspartate aminotranferase/alanine aminotransferase ratio.
R. P. Goswami et al.
9 Bhadade RR, de Souza RA, Harde MJ, Khot A. Clinical characteristics and outcomes of patients with acute lung injury and ARDS. J Postgrad Med 2011;57:286–90.
17 Health CW, Alexander AD, Galton MM. Leptospirosis in the United States: analysis of 483 cases in man 1949–61. N Engl J Med 1965;273:915–22.
10 Deepak NA, Patel ND. Differential diagnosis of acute liver failure in India. Ann Hepatol 2006;5:150–6.
18 Edwards CN, Nicholson GD, Hassell TA et al. Leptospirosis in Barbados. A clinical study. West Indian Med J 1990;39:27–34.
11 Jha V, Chugh KS. Community-acquired acute kidney injury in Asia. Semin Nephrol 2008;28:330–47.
19 Pappachan MJ, Mathew S, Aravindan KP et al. Risk factors for mortality in patients with leptospirosis during an epidemic in northern Kerala. Natl Med J India 2004;17:240–3.
12 Debmandal M, Mandal S, Pal NK. Serologic evidence of human leptospirosis in and around Kolkata, India: a clinico-epidemiological study. Asian Pac J Trop Med 2011;4:1001–6. 13 Chauhan V, Mahesh DM, Panda P et al. Profile of patients of leptospirosis in sub-Himalayan region of North India. J Assoc Physicians India 2010;58:354–6.
20 Doudier B, Garcia S, Quennee V et al. Prognostic factors associated with severe leptospirosis. Clin Microbiol Infect 2006;12: 299–300. 21 Ko AI, Reis MG, Dourado CM et al. Urban epidemic of severe leptospirosis in Brazil. Lancet 1999;354:820–5. 22 Courtin JP, Di Francia M, DuCouedic L et al. Respiratory manifestations of leptospirosis: retrospective study of 91 cases (1978–1994) [in French]. Rev Pneumol Clin 1998;54:382–92.
15 Ittyachen AM, Krishnapillai TV, Nair MC, Rajan AR. Retrospective study of severe cases of leptospirosis admitted in the intensive care unit. J Postgrad Med 2007;53:232–5.
23 Watt G, Padre LP, Tuazon ML et al. Placebo-controlled trial of intravenous penicillin for severe and late leptospirosis. Lancet 1988;327:433–5.
16 Bourrier P, Chennbault JM, Achard J et al. Leptospirosis: retrospective analysis of 99 cases observed in 10 years in the Central West of France [in French]. Me´d Mal Infect 1988;1:4–8.
24 McClain JBL, Ballou WR, Harrisson SM, Steinweg DL. Doxycycline therapy for leptospirosis. Ann Intern Med 1984;100: 696–8.
6
Downloaded from http://trstmh.oxfordjournals.org/ at UQ Library on November 9, 2014
14 Sethi S, Sharma N, Kakkar N et al. Increasing trends of leptospirosis in northern India: a clinico-epidemiological study. PLoS Negl Trop Dis 2010;4:e579.
