Neonatal rectal colonization with Malassezia furfur

GJ GRoss, NE MAcD oNALD, AMR MAcKENZIE. Neonatal rectal colonization with Malasseziafur:fur. Can J Infect Dis 199 1;3 (1):9-13. Malasseziafurfur and Candida albicans are fungal pathogens which have been recognized \vith increasing frequency as agents of mortality and serious morbiclity in neonatal intensive care unit patients. A longituclinal study of oral, rectal and umbilical colonization by these organisms of newborns aclmiltecl to a neonatal intensive care unit \vithin 24 h of birth was undertaken . Of 71 infants followed for a minin1Um of 10 days. 24 were colonized \villi Mjw:fur and 12 \vith C a lbicans during the fi rst 10 clays of life. The lower gastrointestinal tract was found to be the most common colonization site for both organisms. Statistically s ignificant (P<0.05) inverse associations were clemonstratecl between gestational age and risk of colonization with either m·ganism at any site. and between birthweight or gestational age and risk of rectal colonization \vith either organism. Antibiotics were associated \vith a relative risk colonization of 4.06 (P-0.06) \vith either organism at any site . It is concluclecl that Mfurfur and C albicans are common colonizing organisms in a neonatal intensive care unit setting and are most frequenlly harboured in the lower gastrointestinal tract. Mjurjur, recenUy in1plicatecl as a systemic pathogen in this population. has not been previously recognized as a gastrointestinal commensal organism. The relationship between colonization and invasive fungal disease. and potential roles for preventive strategies, remain lo be elucidated.

GROSS et a/ M ALASSEZIA FURFUR AND CANDIDA ALB/CANS ARE FUNGAL pathogens which have been recognized with increasing frequency during the past decade as important causes of neonatal morbidity, particularly in low birthweight infants (l-6).M furfur , a dimorphic lipophilic yeast.was previously recognized as the causative agent for tinea versicolor, a chronic infection of the skin.Since 1981.a number of cases of Mfurfur sepsis occurring in the presence of parenteral lipid infusions via central venous catheters have been reported in both neonates and adults (l -3.7 .8) .Although the mortality from Mfurfur appears to b e low, this is not the case for C albicans infection in very low birthweight infants.Very low birthweight infants have been shown to have a 2 to 4% risk of acquiring disseminated candidiasis (4)(5)(6).with reported mortality rates ranging from 26 to 70% (4.5) .
The purposes of this study were: to compare the rates.sites and timing of Mfurfur colonization in neonatal intensive care unit infants to those of C albicans colonization; and to define subgroups of patients at increased risk of colonization with either organism or both.

PATIENTS AND METHODS
All neonates admitted within 24 h of birth to the neonatal intensive care unit at the Children's Hospital of Eastern Ontario between October 1, 1985 and June 1, 1986 were eligible for the study.The hospital has no obstetrical unit.and receives newborns via transfer from a large number of referring institutions in the Ottawa region.Of 150 eligible patients, 100 neonates were studied.The 50 who were not recruited did not differ significantly from the study group in terms of sex ratio .gestational age or birthweight; however, the nonstudy group did have a much higher mortality (36% versus 1 %) and a shorter mean duration of stay in the unit (15.4±21.0days versus 29.9±20.7 days.P<0.01).Exclusion of eligible patients in the study usually resulted from an expectation that their stay in the neonatal intensive care unit would be brief.due to either an excellent or a very poor prognosis: occasionally, exclusion was due to oversight.
Criteria for discontinuation of a neonate from the study included death, discharge or transfer from the neonatal intensive care unit, initiation of antifungal therapy, and age greater than 70 days.
Cultures were taken from the mouth, rectum and umbilical areas on days 1, 3, 5, 7, 10 and 14, and weekly thereafter.Oral and rectal cultures were obtained by careful insertion of sterile cotton swabs into each cavity.Swabs were plated on Sabouraud agar (Oxoid Canada Inc) with and without a thin layer of sterile olive oil on the surface.Plates were examined daily for four days.Criteria for identification of Mfurfur colonies were growth on olive oil-supplemented Sabouraud agar but not on unsupplemented agar, ap-propriate colonial characteristics and microscopic appearance (9) .Colonies were identified as C albicans if there was growth on unsupplemented Sab ourau d agar.colonial ch aracteristics and microscopy were appropriate, and growth on ox gall agar yielded chlamydospores (10).A selective enrichment medium was made up as follows: Sabouraud broth (Oxoid Canada Inc), cefoperazone 50 mg/L, vancomycin 25 mg/ L and gentamicin 20 mg/ L. This medium was distributed in tubes in 2 mL volu m es, and 0.1 mL olive oil was added to each tube.Swabs were added to the medium and vortexed.M furfur grew on the underside of the olive oil layer, and C albicans grew uniformly in the aqueous layer.
x 2 tests were used to examine the association between colonization at any time during the first 10 days of life and various suspected risk factors.The Cramer's V (phi) coefficient was used to measure strength of association (11).This coefficient can be interpreted as a Pearson correlation coefficient for nominal and ordinal variables.Statistical analyses were performed using the SPSS-X Information Analysis System ( 12) and BMDP Statistical Software (13).

RESU LTS
Of 100 neonates enrolled in the study, 37 yielded cultures positive for Mfurfur, and 16 grew C albicans.There were no documented cases of invasive disease caused by either organism in the neonatal intensive care unit throughout the period covered by the study.
Seventy-one infants were followed for a minimum of 10 days: statistical analysis was restricted to the first 10 days of life in this group.Of these 71, 24 were colonized with M furfur and 12 with C a lbicans during the first 10 days of life.Figure 1 presents the cumulative patient colonization data by site for M furfur , demonstrating that the rectal site was the most common.Colonization with either M fuifur or C albicans at any site did not disappear during the course of follow-upbut other sites often subsequently became colonized.
Gestational age, birthweight, gender and antecedent antibiotic use were assessed as possible risk factors for fungal colonization.Gestational age proved to be the strongest single predictor of colonization.Statistically significant inverse correlations were observed between gestational age category and risk of colonization with either organism at any site and between gestational age category and risk of rectal colonization with either organism (Table 1).Furthermore, gestational age showed an inverse correlation with rectal colonization with M furfur (phi=0.236);55.6% of babies of 29 weeks or less of gestation were colonized.compared with 19.0% of those of at least 36 weeks' gestation, with the intermediate group yielding 31.7% positive cultures.
Although birthweight could be expected to confound gestational age as a predictor for colonization, the only statistically significant relationship established was the  2).This would suggest that lack of maturity, rather than low birthweight per se, is of greater importance in establishing colonization.Antecedent use of antibiotics (ampicillin and gentamicin in standard neonatal dosages) increased the risk of subsequent colonization with either organism to an extent approaching statistical significance (P=0.06).The relative risk of colonization for babies who received antibiotics was 4.06 compared to those who did not.Failure to reach statistical significance was most likely, due to the fact that only eight of 71 babies cUd not receive antibiotics at any time during the first 10 days.a very small group for comparison.
Gender was not shown to be a s ignificant precUctor of colonization.

DISCUSSION
Previous stu cUes of Mfu.rfur have focused on the skin as a major source of lipid substrate for U1e growth of this organism.
AccorcUngly.Faergemann and Fredriksson (14) showed increasing frequency of skin colonization in children aged five to 15 years.but were unable to culture Mfu.rfu.r from those aged one year and less , inducting newboms.Powell and colleagues (l) cUd culture Mfu.rfur from the skin of eight of 25 hospitalized premature infants.In a more recent study involving 361 infants.the same group found M furfur skin colonization in 36.8% of neonatal intensive care unit patients, and identified statistically significant associations with prematurity.low birthweight.length of hospital stay and a number of environmental factors (15).The present data incUcate a lower prevalence of surface colonization (8.5% at 10 days).with no significant relationship to gestational age; this difference can be explained on the basis of the longitucUnal nature of the authors' study compared with Powell et al's cross-sectional approach, in which colonized infants had spent a mean of 51.3 days in the neonatal intensive care unit (1).Bell and colleagues ( 16) have also reported an   17) also reported high skin colonization rates in their neonatal intensive care unit patients, as well as the detection of M furfur in eight of 25 consecutively examined central venous catheters from these patients.
To the authors' knowledge, the lower gastrointestinal tract has not been previously reported as a potential reservoir for Mfurfur in neonates.Data from the present study incUcate that this site is rapidly colonized by M furfur during the first 10 days of hospitalization, and that by the lOth day the prevalence of colonization exceeds 30% (Figure 2).
Terasaka and co-workers (18) analyzed the fatty acid content of meconium from healthy, full term newboms.They found the samples to contain an average of 137 1-1g fatty acids/g meconium; palmitic (C16).stearic (C18) GRoss et a/   and oleic (Cl8) acids were the predominant constituents identified, accounting for 75% of the total in most samples.This would suit the requirement of M furfur for Cl2 to C24 fatty acids (19).even in the absence of enteral feeding.Furthermore, although the meconium of premature infants has apparently not been analyzed in similar fashion.it is evident that lipid absorption is less efficient in preterm than in full term infants due to a number of mechanisms.including reduced pancreatic lipase levels and bile acid pool (20) .This could explain the present fmding of increasing colonization with decreasing gestational age, though the actual fatty acid content of meconium and stool from premature infants remains to be elucidated.The use of broad spectrum antibiotics has long been associated with an increase in candida colonization (6,21.22)and systemic infection (4)(5)(6) .The present data suggest that this association involves M furfur as well as C albicans.
The use of prophylactic oral nystatin in infants at risk of developing systemic candida infections has been advocated (23).In at least two studies a significant reduction in rates of C albicans colonization resulted from this practice (24 ,25).Nevertheless, well documented cases of systemic yeast infection in infants on prophylactic nystatin have been reported (5).To the authors• knowledge, the role of oral nystatin in preventing M furfur colonization and infection in infants has not been evaluated.Since the gastrointestinal tract appears to be an important reservoir for this organism in premature infants, dietary or pharmacological manipulation of the fatty acid content of stools might provide alternative modes of prevention of infection.
The authors conclude that M furfur is a common colonizing organism in their neonatal intensive care unit patients.In contrast with previous reports, they found Mfurjur colonization to be more prevalent than C albicans.Both organisms are most frequently harboured in the lower gastrointestinal tract, a previously unrecognized site for M furfur.Investigations aimed at devising preventive strategies for systemic Mfurfur infection in neonatal intensive care unit patients will have to take into account the important role of the lower gastrointestinal tract as a reservoir for this organism.
CAN J INFECT D1s VoL 3 No 1 JANUARY !FEBRUARY 1992 Neonatal rectal colonization with M furfur

TABLE 1
Relationship between gestational age and neonatal colonization of the mouth , umbilicus or rectum with Malassezia furfur or Candida a/bicans during the first 10 days of life

TABLE 2
Relationship between birthweight and neonatal rectal colonization with either Malassezia furfur or Candida a/bicans during the first 10 days of life