Article Text

This article has a correction. Please see:

Download PDFPDF

Nitric oxide and severe sepsis
  1. Goroka Base Hospital
  2. PO Box 392, Goroka
  3. Papua New Guinea
  4. Department of Paediatrics, University of Oxford
  5. John Radcliffe Hospital
  6. Oxford OX3 9DU
    1. Goroka Base Hospital
    2. PO Box 392, Goroka
    3. Papua New Guinea
    4. Department of Paediatrics, University of Oxford
    5. John Radcliffe Hospital
    6. Oxford OX3 9DU

      Statistics from

      Request Permissions

      If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

      Editor,—Duke et al report changes in serum markers of nitric oxide (NO) production in children with severe sepsis.1 Their findings provide additional data concerning activation of the l-arginine-NO pathway during systemic inflammation and also raise some important methodological issues.

      The classical serum markers of NO production are nitrite and nitrate; these are also termed reactive nitrogen intermediates (RNls) and are stable and easily measured in body fluids—see review in Feelisch and Stanler.2 Levels of these markers in plasma, urine, and saliva are profoundly affected by dietary nitrate, especially ground water3 and nitrate rich food,2 4 5 where, after ingestion, levels of nitrate may increase 10-fold in healthy adults.6 In order to deplete the body of dietary nitrate and for serum nitrate and nitrite to accurately reflect total body NO production, at least 48 hours of a nitrate free diet is necessary.2 4 5

      In this study, Duke and colleagues used unselected emergency admissions of septic patients to an intensive care unit as study subjects, with elective cardiac surgery admissions as controls. It was therefore not possible to control for dietary confounders in either group. Moreover, cardiac failure results in induction of NO production.7 These factors might have contributed to the wide variation and the high levels of RNIs seen in both groups upon admission. This is one possible explanation why the authors only found a trend, rather than a significant increase, in RNIs in septic children with organ failure compared with those without. It is possible to circumvent the issue of dietary confounders in the measurement of NO production by measuring metabolites unaffected by diet. For example, one can measure hydroxyarginine,8nitrosothiols,2 nitrosohaemoglobin by electron spin resonance,2 conversion of 15N-arginine to15N-citrulline,4 and inducible NO synthase activity directly.5

      Measurement of NO metabolites is also affected by renal function, as RNIs are retained in renal impairment, so it may be more accurate to express RNIs as an RNI:creatinine ratio.4 5 In addition, because RNI concentrations are affected by hydration state, NO formation may be overestimated in disease states characterised by dehydration,5 as may be the case in septic shock.

      A further point concerns the production of NO by leucocytes. It would appear from the clinical details of the septic patients that a proportion were immunosuppressed, although this is not specifically mentioned. In sepsis, an important source of NO may be peripheral leucocytes. Septic neutropenic patients appear to produce less NO than septic patients with normal or raised neutrophil counts.9If some of the study patients were neutropenic, this may result in a significant fall in the amount of NO produced by these children, thereby contributing to the wide range of serum values seen, and possibly acting as a confounder. It would therefore be interesting to know whether the neutropenic septic patients were among those with the lowest serum RNI levels.

      In conclusion we believe that studies investigating the production of NO in pathophysiological situations should take into account possible confounders, such as dietary nitrate, neutrophil count, renal function, and level of hydration.

      Drs Duke and South comment:

      Drs Burgner and Rockett make some important points regarding the interpretation of metabolites of NO in clinical studies. Currently there can only be a crude assessment of the role of NO in sepsis. We cannot infer from the measurement of serum nitrogen oxides, or any of the other measures suggested by the authors, the activity of NO in regional microcirculatory beds, or whether there is a causal relationship between NO production and sepsis morbidity. Even though we did find significantly higher levels of serum nitrogen oxides in children who developed organ system failure, we did not find a biological gradient between the numbers of organ failing and nitrogen oxide concentrations. Although variations in dietary intake of nitrates may partially account for this, there may be more important reasons. First, the cause of organ failure in sepsis will not be attributable solely to NO overproduction, or to any other individual cytokine. Second, there is likely to be a level of microcirculatory NO activity that is necessary to ensure tissue perfusion, and that is beneficial in sepsis. Third, our study had relatively small numbers of children with multiple organ failure, and low power to detect a difference.

      We have analysed the data using the RNl:creatinine ratio and found no significant association between the ratio and mortality or organ system failure. The median (interquartile range) of area under the curve (AUC) RNl:AUC creatinine over the first 48 hours of intensive care unit admission was 957 (433–1268) for survivors, and 739 (335–791) for those who died (Wilcoxon rank sum test, p=0.19); and 990 (620–738) for those with no organ failure and 738 (338–1330) for those with one or more organ systems failing (p=0.31) In infants and children, age dependent differences in the normal range of serum creatinine make the interpretation of this suggested analysis additionally complicated.

      There were four children with immunodeficiency causing neutropenia (numbers 8, 9, 17, and 18). These children did not have significantly lower serum nitrogen oxides than the other children with sepsis.

      We agree that there are many confounding variables in the interpretation of NO metabolites. More importantly, despite the finding that overall NO is increased, many children with severe sepsis have low cardiac output, systemic vasoconstriction and pulmonary hypertension,1-10 and currently available agents that inhibit NO activity seem likely to reduce cardiac output,1-11 worsen oxygen delivery and exacerbate pulmonary hypertension.1-12


      1. 1-1.
      2. 1-2.
      3. 1-3.
      4. 1-4.
      5. 1-5.
      6. 1-6.
      7. 1-7.
      8. 1-8.
      9. 1-9.
      10. 1-10.
      11. 1-11.
      12. 1-12.