1338

A. Betrosian M. Papanikolaou M. Balla G. Georgiadis

ARDS due to Yersinia enterocolitica sepsis in a patient with thalassemia major Accepted: 31 August 1999 Sir: We read with interest the article by Mavrommati et al. [1] regarding the acute respiratory distress syndrome (ARDS) as a rare complication of Yersinia enterocolitica sepsis in a patient with thalassemia major. Interestingly, the case they describe mirrors our own experience. A 31-year-old woman with homozygous beta-thalassemia was admitted to the emergency department of fever of 38.5 C for 5 days, abdominal pain, and diarrhea. She received conventional therapy with no response. She had been splenectomized and had been on desferrioxamine chelation since childhood. On admission she was febrile but oriented to time and place; her blood pressure was 100/60 mmHg, heart rate 120 beats/min, and respiratory rate 18 breaths/min. Physical examination was remarkable for a palpable liver (span of 6 cm) and tender abdomen. Laboratory values revealed hemoglobin 8.5 g/dl, white blood count 35 ” 103/mm3, with 90 % neutrophils, alanine aminotransferase 135 IU/l, aspartate aminotransferase 202 IU/l, total bilirubin 1.8 mg/dl, lactate dehydrogenase 483 IU/l. She was treated empirically with intravenous amoxycillin (3 g/day) and ciprofloxacin (400 mg/day). Twenty-four hours later her status deteriorated with dyspnea, tachypnea, and severe respiratory distress. Blood gas analysis on 100 % oxygen inspiration revealed pH 7.285, partial pressure of oxygen 83.2 mmHg, partial pressure of carbon dioxide 50, bicarbonate 23.2 mmol/l, and oxygen saturation 94.5 %. Chest auscultation revealed bilateral crackles on inspiration, while chest X-ray showed diffuse pulmonary infiltrates (Fig. 1). The patient was intubated and mechanically ventilated. Right heart catheterization via a Swan-Ganz catheter revealed central venous pressure 11 mmHg, mean pulmonary arterial pressure 28 mmHg, pulmonary capillary wedge pressure 14 mmHg, and cardiac index 4.6 l/min per m2. With the diagnosis of ARDS complicating abdominal sepsis, she was admitted to the operating room. At laparotomy, a

Fig. 1 Pulmonary infiltrates in the patient with yersinia sepsis

gangrenous appendix and mesenteric lymph nodes were present. A blood culture drawn on admission revealed Yersinia enterocolitica serotype O : 3, sensitive to fluoroquinolones, aminoglycosides, and ureidopenicillins. Imipenem/cilastatin (2 g/ day) and amikacin (1 g/day) were added to ciprofloxacin, while amoxycillin was withheld. Despite temporal improvement, the patient's status deteriorated with recurrent septic episodes and shock, worsening respiratory function, barotrauma, and finally multiple organ failure. She died 45 days after admission. Pulmonary complications of Yersinia sepsis include infiltrates [2], empyema [3], abscess [4], and recently ARDS [1]. However, the last has been previously reported in a 58-year-old alcoholic patient [3]. In our case, the development of the syndrome early in the course of the disease, prior to laparotomy, is strong evidence of a direct association between Yersinia sepsis and ARDS. Despite immediate treatment, combined antibiotic therapy, and intensive care support, the patient's immune deficiency might have led to a lethal outcome. Yersinia enterocolitica sepsis carries an increasing mortality in patients with blood disorders, splenectomy, or with debilitating diseases. The large number of thalassemic patients in Greece makes necessary in cases of Yersinia sepsis an early diagnosis, combined antibiotic treatment, and supportive therapy in the intensive care unit.

References 1. Mavrommati P, Ladis V, Lagona E et al (1999) ARDS in a patient with homozygous beta-thalassemia due to yersiniosis. Intensive Care Med 25: 226±229

2. Taylor BG, Zafarzai MZ, Humphreys DW et al (1977) Nodular pulmonary infiltrates and septic arthritis associated with Yersinia enterocolitica bacteremia. Am Rev Respir Dis 116: 525±529 3. Clarridge J, Roberts C, Peters J et al (1983) Sepsis and empyema caused by Yersinia enterocolitica. J Clin Microbiol 17: 936±938 4. Foberg U, Fryden A, Kihlstrom E et al (1986) Yersinia enterocolitica septicemia: clinical and microbiological aspects. Scand J Infect Dis 24: 269±279

)

A. P. Betrosian ( ) × M. Papanikolaou × M. Balla × G. Georgiadis Intensive Care Unit, Hippokration General Hospital, 114 Vas. Sofias Str., Athens 11527, Greece

P. Mavrommati T. Hatzis

Reply Accepted: 31 August 1999 Sir: We would like to thank Dr. Betrosian and colleagues for their interest in our article. They described another case of ARDS following Yersinia enterocolitica sepsis in a young woman with beta-thalassemia. Such cases demonstrate the need for more careful evaluation of iron-overloaded patients as soon as they present mild gastrointestinal illness. Although the patient's immunological condition is one of the most

1339

important factors for the prognosis of yersinia infection in patients with beta-thalassemia, it is possible that intravenous antibiotic treatment including fluoroquinolones ± alone or in combination ± as soon as the first symptoms of yersiniosis occur may prevent yersinia sepsis and complications such as life-threatening ARDS, especially to splenectomized patients.

References 1. Gayraud M, Scavizzi M et al (1993) Antibiotic treatment of Yersinia enterocolitica septicemia: a retrospective review of 43 cases. Clin Infect Dis 17: 405±410 2. Hoogkamp-Korstanje JAA, de Koning J, Heesemann J (1988) Persistence of Yersinia enterocolitica in man. Infection 16: 81±85

)

P. Mavrommati × T. Hatzis ( ) Critical Care Department, Aghia Sophia Children's Hospital, Thevon & Levadias Street, Athens 115 27, Greece, Tel.: + 30 (1) 7 48 86 86, Fax: + 30 (1) 7 79 76 49

R. M. Calcroft G. Joynt

Metoclopramide improves gastric motility Accepted: 6 September 1999 Sir: Dr. Jooste and colleagues have shown that metoclopramide improves gastric emptying in a miscellaneous group of critically ill patients [1]. In common clinical practice, however, prokinetic drugs are considered a therapeutic option only in patients who demonstrate gastric intolerance, as demonstrated by phenomena such as large nasogastric aspirates, reflux, or vomiting. While we acknowledge that the relationship between large gastric aspirates and decreased gastric motility is still speculative [2], it would be much more relevant to assess the efficacy of metoclopramide in critically ill patients who have been demonstrated clinically to have impaired gastric emptying. The likelihood of these patients having markedly impaired gastric emptying of a magnitude sufficient to cause clinically detectable gastric intolerance is

uncertain, although we acknowledge they may have impaired gastric emptying relative to normals. Factors previously shown to be associated with impaired gastric emptying include being critically ill (Simplified Acute Physiology Score 9.5 ± 3), older age and female sex, opiate use [3], dopamine [4], and other drugs such as anticholinergics. Apart from opiate and inotrope use, the authors give little information regarding these factors. The paper by Dr. Jooste et al. therefore does not address the relevant clinical question: Does metoclopramide improve gastric emptying in patients with gastric stasis or gastric feed intolerance? Assuming that blood paracetamol levels accurately reflect gastric emptying, the magnitude of the difference between the two groups is not substantial when compared with the control group's standard deviation [metoclopramide group 1387.2 ± 560.4 (mean ± SD); saline group 994 ± 623.6]. While the authors have shown a statistically significant increase in the AUC 120 (area under the curve over 120 min) in patients given metoclopramide, we would question the clinical significance of this result. Lastly, in patients with documented gastric intolerance, we have noticed that the use of 1.5 g paracetamol is frequently associated (25 % of cases studied) with detectable serum levels of paracetamol 24 h after the first dose. We would be interested to know whether the authors experienced this problem, and, if so, how did they compensate for this phenomenon (which we believe is merely a consequence of prolonged gastric emptying). In conclusion, we believe that ªthe administration of intravenous metoclopramide improved gastric emptying in a heterogeneous group of critically ill patientsº may have been shown in this study, that ªmetoclopramide is a useful prokinetic drug in this patient populationª has yet to be demonstrated.

References 1. Jooste CA, Mustoe J, Collee G (1999) Metoclopramide improves gastric motility in critically ill patients. Intensive Care Med 25: 464±468 2. McClave SA, Snider HL, Lowen CC, McLaughlin AJ, Green LM, McCombes RJ et al (1992) Use of residual volume as a marker for entereal feeding intolerance: prospective blinded comparison with physical examination and radiographic findings. JPEN J Parenter Enteral Nutr 16: 99±105

3. Heyland DK, Tougas G, King D, Cook DJ (1996) Impaired gastric emptying in mechanically ventilated critically ill patients. Intensive Care Med 22: 1339±1344 4. Tarling MM, Toner CC, Withington PS, Baxter MK, Whelpton R, Goldhill DR (1997) A model of gastric emptying using paracetamol adsorption in intensive care patients. Intensive Care Med 23: 256±260

)

R. M. Calcroft ( ) × G. Joynt Prince of Wales Hospital Shatin, Hong Kong Tel.: + 852 (269) 62689 Fax: + 852 (269) 62689 email: [email protected]

C. A. Jooste J. Mustoe G. Collee

Reply Accepted: 6 September 1999 Sir: Thank you for the opportunity to reply to Drs. Calcroft and Joynt. The purpose of our study was to ascertain the prokinetic efficacy of metoclopramide in critically ill patients, who frequently suffer gastric impairment [1, 2]. Calcroft and Joynt suggest that impaired gastric emptying may be unimportant when it is clinically undetectable. However, clinical tests for impaired gastric emptying are insensitive [3]. The implicit suggestion that subclinically impaired emptying is irrelevant is speculative. We omitted the data on age and sex as, in a crossover study, these are not confounding variables. No patient received drugs known to influence gastric emptying other than those mentioned in the paper. It is contended in the letter that the difference in paracetamol absorption between the saline and metoclopramide groups is not substantial. This assertion does not take into account the sequence of randomisation. In those patients who received saline first, paracetamol absorption was less than the quoted mean of 994 mg ´ min/l in all cases. It was only in those who had already received metoclopramide that the absorption exceeded this mean following saline. As we speculated in the discussion, this may reflect a residual metoclopramide effect. We reiterate our conclusion that metoclopramide is a useful prokinetic in this pa-