Bulletin of the World Health Organization

A bedside test for methemoglobinemia, Sri Lanka

Fathima Shihana a, Andrew H Dawson b & Nicholas A Buckley c

a. South Asian Clinical Toxicology Research Collaboration, Faculty of Medicine, University of Peradeniya, Peradeniya, Sri Lanka.
b. NSW Poisons Information Centre, Sydney Children’s Hospital Network, Australia.
c. Sydney Medical School, University of Sydney, Sydney, Australia.

Correspondence to Fathima Shihana (email: fathimashihana@gmail.com).

(Submitted: 13 May 2015 – Revised version received: 03 March 2016 – Accepted: 03 March 2016 – Published online: 14 June 2016.)

Bulletin of the World Health Organization 2016;94:622-625. doi: http://dx.doi.org/10.2471/BLT.15.158147


Propanil is an aniline herbicide that is widely used for rice cultivation. However, it is also used for deliberate self-poisoning. Propanil poisoning is usually manifested by methemoglobinemia and haemolytic anaemia, disorders that lead to reduction in effective oxygen transport and hypoxia.1 In general, the severity and symptoms of propanil poisoning correlate with the methemoglobin level in the blood. The recommended first-line treatment of methemoglobinemia is intravenous administration of methylene blue, which reduces the heme group from methemoglobin to hemoglobin. Methylene blue is inexpensive, accessible and easy to administer. The initial dose is 1–2 mg/kg of body weight with repeated doses titrated against the clinical response. Alternative treatments, such as ascorbic acid or N-acetylcysteine, are less effective, but can be used if methylene blue is unavailable. Exchange blood transfusion is only recommended when methylene blue or alternative treatments are ineffective.2

The gold standard for determining methemoglobin concentration in the blood is by spectrophotometry. However, in rural areas, laboratory services are limited and it is usually not possible to measure methemoglobinemia with spectrophotometers or co-oximeters. As a methemoglobin concentration greater than 15% of total haemoglobin levels gives a brownish colour to blood, concentrations above 15% can be detected visually. Clinicians can use this visual method to diagnose patients, but to measure response to methylene blue treatment they need a more accurate method of quantification. We therefore developed a simple and low-cost bedside test for semiquantitative estimation of methemoglobin levels. The test is described in detail elsewhere.3 One drop of blood is placed on white absorbent paper and the colour of the blood spot compared to a reference chart (Fig. 1),3 corresponding to an estimated level of methemoglobinemia. We describe lessons learnt when using this bedside test in three hospitals in Sri Lanka.

Fig. 1. Semiquantitative estimation of methemoglobin levels in blood
Fig. 1. Semiquantitative estimation of methemoglobin levels in blood
(© Fathima Shihana)

Local setting

In Sri Lanka, propanil poisoning has been recognized as a significant clinical problem. Two cohort studies – from two rural tertiary referral hospitals – have shown propanil self-poisoning case-fatality of 11% (45/412).4,5 In these cohort studies, we identified treatment patterns at odds with national guidelines for the treatment of methemoglobinemia published by the Sri Lankan National Poisons Centre. The guidelines recommended the use of methylene blue as a primary treatment.2,6 We observed relatively low use of methylene blue and higher use of ascorbic acid or exchange transfusion. Although clinicians reported that it was easy to diagnose propanil poisoning, the lack of methemoglobin concentration measures made it difficult to determine whether patients were responding to methylene blue. The lack of an objective measure against which to titrate the dose lead to a recurrence of methemoglobinemia in some patients.4,5 Clinicians also raised two additional concerns in using methylene blue and in managing the subsequent reductions in methemoglobin concentration. First, methylene blue is ineffective in patients with complete glucose-6-phosphate dehydrogenase (G6PD) deficiency and may cause hemolysis.7,8 About one in every 20 people from Sri Lanka has G6PD deficiency.9 Second, it is known that excessive doses of methylene blue can, in theory, cause or worsen methemoglobinemia.3

Relevant changes

In collaboration with clinicians at three tertiary rural hospitals, we introduced the methemoglobin colour chart into all the medical wards in 2008. For ward staff, medical officers and trainee doctors, clinical research staff gave a single 20-minute oral presentation describing the test method,3 and how it could be used within the current national treatment guidelines. We provided each ward with a methemoglobin colour chart. The chart was prepared using a good quality printer in a local photo studio. The accuracy of the reproduction of the colour chart was checked with colour analysis that compared its colours to the original colour values derived in the test’s development.3 Each ward received a poster that described the use of the colour chart and reiterated the national treatment guidelines. After introducing the colour chart in the hospitals, we validated the accuracy of the test in the field by taking an additional sample from 13 patients and analysing the blood using a spectrophotometer.3

Independent of the treating team, clinical research assistants examined all poisoned patients until discharge or death. Clinical outcomes and treatment of each patient was prospectively recorded into the cohort database. We then identified, from the database, data on patients admitted with propanil poisoning. We were able to examine outcomes in 401 patients before (2003–2007) and 262 patients after (2008–2014) the introduction of the test.10

Lessons learnt

After we introduced the test, case-fatality for propanil poisoning fell by two-thirds, from 10% (38/401) to 3% (8/262). Retrospective examination of available patient medical records showed an increase in the use of methylene blue after the test was introduced: from 10% (13/136) to 55% (59/107).10 Records showed that titrated doses of methylene blue were more common than single-dose treatment and such dosage patterns have been sustained. The use of less effective treatments and exchange transfusion were reduced,10 suggesting that clinicians accepted the utility of this test in the management of methemoglobinemia.10

Our report has limitations. It is possible that some of the survivals may have been from unrelated improvements in care at this time. Our quasi-experimental comparison of two different time periods cannot exclude alternative explanations for the changes in number of deaths. However, the change in management and reduction in deaths was considerable and no major change in these outcomes had been observed within the study locations in the six years before the intervention.

Box 1 summarizes the main lessons learnt. The uptake of the test into practice was high, as the test quickly provided results, allowing more informed use of national treatment guidelines. Distribution of the colour chart was cheap, as it could be printed from a freely available file by any good-quality colour printer.11

Box 1. Summary of main lessons learnt

  • Low-cost bedside estimation of methemoglobin levels changed case management and was associated with a reduction in deaths from propanil poisoning.
  • The test helped doctors to provide treatment consistent with national guidelines for management of poisoning.
  • Applied research that addresses local clinical concerns can be translated into practice and better health outcomes.

The simplicity and low cost of the test presented here can facilitate the treatment of methemoglobinemia in resource-poor settings. The test was included in the 2011 edition of the national guidelines: Management of poisoning6 and in the curriculum of local postgraduate training programmes for clinicians and nurses.


Andrew H Dawson is also affiliated with the South Asian Clinical Toxicology Research Collaboration and the Sydney Medical School, University of Sydney, Sydney, Australia. Nicholas A Buckley is also affiliated with the South Asian Clinical Toxicology Research Collaboration and the NSW Poisons Information Centre, Sydney Children’s Hospital Network, Australia.


This study was partially funded by the Welcome Trust/National Health and Medical Research Council international collaborative research grant GR071669MA and other NHMRC grants (1011772, 1055176). Andrew Dawson is supported by an NHRMC Practitioner Fellowship (1059542).

Competing interests:

None declared.