Cold Spots in Neonatal Incubators Are Hot Spots for Microbial Contamination

Abstract

Thermal stability is essential for the survival and well-being of preterm neonates. This is achieved in neonatal incubators by raising the ambient temperature and humidity to sufficiently high levels. However, potentially pathogenic microorganisms also can thrive in such warm and humid environments. We therefore investigated whether the level of microbial contamination (i.e., the bacterial load) inside neonatal incubators can be predicted on the basis of their average temperature and relative humidity settings, paying special attention to local temperature differences. Swab samples were taken from the warmest and coldest spots found within Caleo incubators, and these were plated to determine the number of microbial CFU per location. In incubators with high average temperature (≥34°C) and relative humidity (≥60%) values, the level of microbial contamination was significantly higher at cold spots than at hot spots. This relates to the fact that the local equilibrium relative humidity at cold spots is sufficiently high to sustain microbial growth. The abundance of staphylococci, which are the main causative agents of late-onset sepsis in preterm neonates, was found to be elevated significantly in cold areas. These findings can be used to improve basic incubator hygiene.

INTRODUCTION

To increase the survival chances of premature infants, incubators have been developed which provide a thermoneutral environment. This environment must be defined in terms of temperature, humidity, convection, and surrounding radiant heat sources. The humidification of the incubators has helped in this respect, as the greatest heat loss by preterm neonates is due to evaporative transepidermal water loss. Increased relative humidity (RH) reduces the rate of evaporation and, as a consequence, not only reduces evaporative heat loss (4) but also helps to maintain the electrolyte and fluid balance of preterm neonates (7). Unfortunately, these conditions also increase the risk of microbial infection, because the moist and warm habitat created within an incubator is intrinsically ideal for microbial growth. A fine balance is sought between the thermal requirements of a newborn and the level of humidification and heating of the incubator. The highest levels of humidification and heating are used for the most immature infants with the lowest birth weights, as they have the most difficulty with maintaining their thermal stability (2, 8, 10). The humidification and heating levels are lowered slowly in the course of the first days and weeks after birth as the neonates' weight and the maturity of their skin barrier increase.

The availability of water is one of the most important factors for microbial growth. Water availability on surface-to-air interfaces is determined by the local equilibrium relative humidity (RH_e). Molds require RH_e values of at least 70%, and RH_e values need to be even higher (80 to 95%) to allow for the growth of yeasts, Gram-positive bacteria, and Gram-negative bacteria; the latter group of bacteria require the highest RH_e values (9). The local RH_e on any position within an incubator can be calculated using only three variables, which are the local temperature on that particular spot, the average air temperature of the incubator, and the average RH inside the incubator (1). Both the average air temperature and RH are controlled and monitored in modern incubators. The local temperature on the walls inside the incubator, however, is not under direct control and is instead dependent on the temperature difference between inside and outside the incubator and the airflow within the incubator.

To investigate whether the local level of microbial contamination can be predicted by determining the local temperature distribution within an incubator, we first mapped the temperature distribution within an incubator with an infrared thermometer. Second, swabs were taken from both cold and hot spots, and these were plated to compare the numbers of microbial CFU on these spots. An important prediction was that the cold spots in incubators with high levels of humidification and heating would reach RH_e values that are compatible with microbial growth. This should not be the case for cold spots in incubators with low levels of humidification. To test this, the sampled incubators were divided according to their average RH and temperature values into group 1, with low RH (≤60%) and temperature (<34°C), and group 2, with high RH (≥60%) and temperature (≥34°C).

MATERIALS AND METHODS

Characterization of incubators.

The present studies were performed at the Department of Neonatology of the Beatrix Children's Hospital, University Medical Center of Groningen (UMCG), the Netherlands, in a 24-bed level III neonatal intensive care unit. This unit is equipped with Caleo incubators. A Fluke 62 Mini infrared thermometer (Conrad, Oldenzaal, The Netherlands) was used to measure the temperature distribution inside an empty Caleo incubator, which was set at an average RH of 70% and a temperature of 37°C (Fig. 1). Various cold spots were identified with temperatures of up to 6°C lower than the average air temperature (37°C), and certain hot spots were up to 2°C warmer. The temperature distribution subsequently was verified by the inspection of the different condensation profiles at decreased or elevated average RH values (55 to 85% RH; data not shown). Incubators with average air temperatures lower than 37°C differed slightly in how much colder the cold spots were relative to the average air temperature. Smaller temperature differences between inside and outside the incubator (approximately 24°C) led to smaller temperature differences within an incubator. As incubators with lower average temperature values invariably also had lower average RH values, an additive reduction in the local RH_e at the cold spots in such incubators occurs. These findings were taken into account for defining the cutoff values used to group the sampled incubators according to their RH and temperature.

Fig. 1.

Temperature distribution within a Caleo neonatal incubator. (A) Image of an empty incubator. The internal temperature distribution is represented from the top view (B), front and rear views (the part below C3 is optional) (C), and side view (D). C1 to C5 indicate cold sampling locations, and H1 to H4 indicate hot sampling locations.

Sampling protocol.

Neonates are routinely transferred into freshly cleaned incubators at weekly intervals, and samples thus can be taken from the used incubators immediately after the transfer without interfering with the care of the neonate. All sampled incubators had been in use for 4 to 7 days. For privacy reasons, no information concerning the neonates other than the temperature and RH settings of the sampled incubators was recorded. Sampling was performed with a cotton-tipped swab (Greiner-Bio-One, Alphen aan de Rijn, The Netherlands) moistened with 70 μl physiological salt solution (0.85% NaCl). The swabs were rubbed vigorously against the identified cold and hot spots on the inner wall of the incubators. For the first 13 incubators included in our studies, three to four hot spots and four to five cold spots were swabbed per incubator. These numbers were standardized to four hot and five cold spots for 10 additionally included incubators. The surface areas that were swabbed on each of these spots were of comparable size (∼20 cm²), except for hot spot 3 and cold spot 4 (∼10 cm²) (Fig. 1). The CFU numbers for the latter two spots were corrected accordingly. No samples were taken if remnants of bodily excretions (e.g., saliva) or remnants from other types of fluids (e.g., fluids from indwelling catheters) were present at the locations that were to be sampled. Such exclusions took place for both groups in approximately equal numbers. In total, samples were taken from 11 incubators of relatively low humidity (≤60%) and temperature (<34°C), designated group 1, and from 12 incubators of relatively high humidity (≥60%) and temperature (≥34°C), designated group 2.

Assessment of microbial contamination levels.

The swabs were immediately taken to the laboratory after sampling. The tips of cotton-tipped swabs were cut off and deposited in sterile tubes containing 1.6 ml physiological salt solution and three sterile glass beads (4 mm in diameter). These tubes were mixed for 30 s on a vortex mixer, incubated for 10 min at 4°C, and mixed again afterwards. Aliquots of 100 μl of the resulting suspensions subsequently were plated on both blood agar and on Sabouraud agar (Mediaproducts BV, Groningen, The Netherlands). The blood agar plates were incubated at 37°C, while the Sabouraud plates were incubated at room temperature. Isolates subsequently were identified by standard diagnostic microbiological procedures and criteria, including colony morphology and microscopy (Gram staining, cell shape, clustering patterns, and presence of spores) and by performing biochemical tests such as API-20E.

Statistical analyses.

The numbers of CFU recovered from the blood agar plates and the Sabouraud plates were summed up to determine the number of CFU per location. If a particular organism was found to be growing on both types of plates, the summation described above was adjusted such that only the highest number of CFU from either plate of this particular organism was used in the summation. As the numbers of CFU per location are not normally distributed, the nonparametric Mann-Whitney U test was used to assess the significance (P < 0.05) of the differences in the level of microbial contamination.

RESULTS

The warm humidified air that is blown into the Caleo incubators from hot air vents does not reach all parts of the interior walls equally well (Fig. 1). Cold spots were found predominantly at the front and back sides and in the middle of the incubator on the doors and canopy. Hot spots mapped close to the hot air vents and on the canopy above the hot air vents. Group 1 represented the incubators in which the RH_e values of the cold spots were predicted not to exceed 80%. The cutoff RH_e value of 80% was chosen because it represents the growth limit of staphylococci, which are very xerotolerant human commensals (6). Group 2 represented the incubators in which the RH_e values of the cold spots were predicted to exceed 80%.

Local RH_e values can be calculated by first calculating the saturation vapor pressure (e_s) within an incubator using the following equation, which is commonly known as the Magnus equation (1): e_s = C × exp[(A × T)/(B + T)], where T is the average air temperature in Kelvin and the values of the coefficients A, B, and C are 17.625 and 243.04 K and 6.1094 mbar, respectively. The local saturation vapor pressure at a particular location [e_s(L)] is calculated in the same way using the local temperature at that location instead. The actual vapor pressure (e) of the air within an incubator is subsequently calculated as e = RH × e_s/100, where RH is the average relative humidity inside the incubator. Local RH_e values then can be calculated using the e_s(L) of those locations: RH_e = e/e_s(L) × 100.

The number of CFU obtained from each sampled location was taken as a measure for the level of microbial contamination. No significant differences (P = 0.275) were found between the cold and hot spots from incubators that had RH values of 60% or lower in combination with average air temperatures lower than 34°C (group 1) (Fig. 2). However, significant differences (P = 0.002) were found between the cold and hot spots from the incubators that had RH values of 60% or higher in combination with average air temperatures higher than 34°C (group 2). These differences were a common feature of group 2 incubators, and they did not relate to a limited number of highly contaminated incubators. Specifically, the six spots with the highest contamination levels (>50 CFU) among the cold spots from group 2 incubators were identified in five different incubators (Fig. 2). Even if the incubators were not divided into groups 1 and 2, the differences in the CFU at the hot and cold spots still remained significant (P = 0.009). Interestingly, certain microbial species, especially Gram-positive bacteria, were found to be more prevalent at the cold spots of group 2 incubators than at the hot and cold spots of group 1 incubators and the hot spots of group 2 incubators (Table 1). A direct comparison of the numbers of CFU found per species per location was difficult, as the numbers of CFU per location were not normally distributed. Therefore, the mean of the log₁₀(CFU + 1) value of organisms from a particular type (e.g., staphylococci) per spot from a particular location type (group 1 or 2 incubator, hot or cold spot) is shown in Table 1.

Fig. 2.

Microbial contamination levels on hot and cold spots in group 1 and 2 incubators, which have low and high average RH values, respectively. The numbers of CFU are indicated. The numbers of sampled spots and the relative frequencies of CFU are indicated. Furthermore, the numbers of incubators contributing to each CFU class specified in the pie charts is indicated in parentheses.

Table 1.

Prevalence of microbial groups at different locations in neonatal incubators

Microbial group	Mean log(CFU + 1) (% presencea) in:
	Group 1 hot spots	Group 1 cold spots	Group 2 hot spots	Group 2 cold spotsb
Staphylococci	0.04 (10.3)	0.08 (14.0)	0.04 (10.0)	0.36(35.5)
Micrococci	0.09 (20.5)	0.07 (16.0)	0.07 (10.0)	0.06 (12.9)
Sporulating bacteria	0.01 (2.6)	0.01 (2.0)	0.03 (7.5)	0.03 (4.8)
Other Gram-positive bacteria	0.06 (10.3)	0.03 (8.0)	0.01 (2.5)	0.12(17.7)
Gram-negative bacteria	0.03 (7.7)	0.01 (2.0)	0.04 (7.5)	0.11 (9.7)
Yeast	0.02 (5.1)	0.04 (4.0)	0.04 (5.0)	0.02 (4.8)
Molds	0.02 (5.1)	0.03 (4.0)	0 (0)	0.02 (4.8)

^aPercentage of locations colonized with 1 or more CFU of the indicated microbial group.

^bValues in boldface are significantly higher than those of other groups.

Of particular interest is the significant increase (P < 0.0005) in the numbers of CFU from staphylococcal species on the cold spots of group 2 incubators (Table 1), since at the UMCG and as reported by others (11), most of the infections in preterm neonates are caused by staphylococci. The abundance of Gram-negative bacteria on cold spots of group 2 incubators also was found to be slightly elevated, although not to a statistically significant level (P = 0.233). Of all the Gram-negative bacteria, a Methylobacterium species, which forms pink colonies on Sabouraud agar, stood out, as it was found a few times in high numbers of CFU. Micrococcus luteus and other Micrococcus species were not found at increased frequency at the cold spots of group 2 incubators. Similarly, no significant differences were observed for the numbers of Gram-positive sporulating bacteria, yeasts, or molds. The number of other Gram-positive bacteria not represented by the afore-mentioned groups also increased significantly (P = 0.017) on the cold spots of group 2 incubators. Particularly noteworthy among these other Gram-positive bacteria were the enterococci, as these bacteria were found exclusively on the cold spots of group 2 incubators.

DISCUSSION

The present studies show that the levels of microbial contamination are significantly elevated only on cold spots that occur within neonatal incubators with relatively high average temperature and RH settings. This observation, which was made in an operational clinical setting, is fully consistent with our predictions based on previous laboratory studies with an environmental control chamber containing precisely controlled temperature/RH_e gradients (5).

In particular, the level of staphylococcal contamination was increased significantly on the cold spots of group 2 incubators. Staphylococci are known for their ability to grow at relatively low RH_e values, and these bacteria consequently are among the few organisms that can grow at cold spots which reach RH_e values only slightly above 80% (80 to 90%) (6). These bacteria furthermore are very desiccation resistant (3) and are ubiquitously present as part of the human microbiota. Consequently, there is a high probability to detect staphylococci at locations frequented by humans. The increased level of staphylococcal contamination within group 2 neonatal incubators is a reason for concern, since the vast majority of infections in preterm neonates are caused by coagulase-negative staphylococci and, at lower rates, by Staphylococcus aureus (11, 12). Furthermore, the slightly increased numbers of Gram-negative bacteria at the cold sites of neonatal incubators with high humidity levels call for attention, because infections with Gram-negative bacteria are known to have the highest neonatal death rates (11).

Simple improvements in hygienic measures can be readily deduced from our observations. The occurrence of condensation on the interior surfaces of neonatal incubators should be prevented as much as possible. When condensation does occur and when a clear view is needed to handle the neonate, the condensate preferably should be wiped away with a dry disposable microfiber cloth (13). Clearly, this should not be done with a reusable piece of cloth lying in the incubator. After wiping away the condensate, the nurse's hands should be disinfected before handling the neonate to prevent any bacterial transmission to the neonate. It also might be advisable to change incubators more often than once a week for neonates who are nursed in incubators with high temperature and RH values (group 2) if this is not too stressful for these neonates. Incubators from group 2 also could be placed in wards with elevated room temperature values to prevent the formation of relatively cold spots, thus preventing microbial growth from occurring in the first place. Lastly, improved incubator design, for example by placing small heating elements at vital spots, might be a similarly effective strategy.

In conclusion, we have demonstrated that the level of microbial contamination is significantly elevated in incubators with high average air temperature and high relative humidity values, especially in the areas where the local temperature is lower than the average temperature of the air inside the incubator. Preterm neonates with a gestational age of less than 30 weeks require elevated temperature (≥34°C) and humidity (≥60%) settings during the first weeks of life. As the risk of late-onset sepsis increases for more premature neonates (11), our findings make it conceivable that this is at least partially related to the environmental conditions in their incubators. The identification of local cold spots as so-called hotspots for microbial growth, coupled with appropriate hygienic measures, may contribute to reducing the frequency of late-onset sepsis in neonates. We believe that this hypothesis calls for a larger prospective follow-up study on the correlations between relative humidity, microbial load, and infection rates in preterm neonates, which also should take into account relevant information on the patient population. If such a prospective study confirms our present hypothesis, which we consider likely in view of the reproducibility of our data, appropriate standards for the evaluation of microbial loads in neonatal incubators need to be established.