Objective. Heel lancing has hitherto been the method of choice for screening tests of inborn errors of metabolism in neonates. Eutectic mixture of local anesthetic has little, if any, analgesic effect when applied to the heel of pre- and full-term infants. Today there seems to be no practical method that effectively alleviates the pain caused by heel lancing. The aim of this study was to investigate whether other methods, such as venipuncture of the dorsal side of the hand, are less painful and more efficient than heel lancing in neonates.
Methods. One hundred twenty healthy, full-term infants who underwent testing for phenylketonuria (PKU) were included in the study. They were randomized into three groups: a venipuncture group (VP) (n = 50), a heel-lance group with a small lancet (SL) (n = 50), or a large lancet (LL) (n = 20). Pain reactions were assessed by the Neonatal Facial Coding System (NFCS). The latency and duration of the first cry as well as the total screaming time was analyzed. The efficacy of the method was assessed by the time taken to complete the PKU test and number of times the skin had to be punctured.
Results. Nociceptive response. The NFCS scores after the first skin puncture were: VP, 247; SL, 333; and LL, 460. The NFCS scores were significantly lower in the VP than in the groups subjected to heel lancing. During the first 60 seconds after skin puncture, 44% of the infants cried in the VP group, compared with 72% and 85% in the SL and LL groups, respectively.
Efficacy. A successful PKU test with only one skin puncture was performed in 86% of the VP group, but in only 19% in the SL group and 40% in the LL group. The total time needed to obtain the PKU blood sample was 191 seconds (median) in the VP group, 419 seconds in the SL group, and 279 seconds in the LL group.
Conclusion. Venipuncture is more effective and less painful than heel lancing for blood tests in neonates.
Heel lancing has hitherto been the method of choice for screening tests of inborn errors of metabolism in neonates. This is a painful method.1 Different methods, such as the use of topical local anesthesia, have been tried for alleviation of pain. Recently, some data have shown that when a eutectic mixture of local anesthetic (EMLA) cream was applied to the heel it had little, if any, analgesic effect in pre- and full-term infants.2-5 A rapid clearance of a topically applied local anesthetic, as a result of high cutaneous blood flow, may explain why EMLA seems to be of little value when it is applied to the neonatal heel.6 Nonnutritive sucking and the sucking of sucrose solutions have also been tried.7-9 Besides the size of the lancet, the design can be of importance for the magnitude of the pain response. A spring-loaded device has been reported to reduce the pain of the heel prick.10,11 Such methods can reduce pain, but it must be concluded that today there are no effective and practical methods for alleviating the pain caused by heel lancing.
This study aimed to investigate whether venipuncture of the dorsal aspect of the hand is less painful and more efficient than heel lancing in neonates. Therefore, the nociceptive response to venipuncture was compared with the response of a heel prick with a standardized lancet or a large lancet.
The study was designed as randomized and blind. Parental informed consent and approval by the local Ethics Committee was obtained. Full-term infants on the maternity ward at Karolinska Hospital, who should be tested for phenylketonuria (PKU), were consecutively included in the study. Infants with any illness or abnormality were excluded. One hundred twenty neonates were randomly allocated to three groups, according to the procedure to be used in the PKU test. Demographic data are presented in Table1.
One hundred twenty identical envelopes were prepared, containing a written allocation to a group. Thus, 50 infants were allocated to a venipuncture group (VP), 50 infants to a small-lancet group (SL), and 20 to a large-lancet group (LL). Sham heel lancing was also performed in an additional nonrandomized group of 11 infants.
All infants were fed 1 to 2 hours before testing. When tested, all infants were sleeping lightly or awake and resting quietly. A nurse from the Neonatal Intensive Care Unit (nurse 1), performed all PKU tests. Before the PKU test was performed the test area was warmed for 1 minute by gently placing the child's hand or heel between the warm hands of nurse 1, thereafter the skin was cleaned with a disinfectant. After penetrating the skin with either a venipuncture needle or a heel lance, the infants were left undisturbed for 60 seconds during which time no effort was made to induce blood flow. Thereafter, additional skin punctures or squeezing of the heel was done to obtain enough blood for the PKU test. If, after a single skin puncture, the blood flow was sufficient, no further manipulation took place. Before the venipuncture was performed, the dorsum of the hand was gently squeezed to visualize the vein, thereafter no additional squeezing was made. A record was kept of each procedure by the assisting nurse (nurse 2) that spoke into the recording equipment and also stated when the PKU test was finished. A Microlance needle measuring 0.9 × 40 mm (Becton-Dickinson, Madrid, Spain) was used for the venipuncture. Two devices were used for heel lancing. In the SL group, a CCS Minilancet (Clean Chemical, Borlänge, Sweden) was used. This lance has a very sharp end (almost like a needle) protruding approximately 2 mm. The CCS Minilancet was gently placed on the skin, pressure was applied to the end of the device, and the lance then penetrated to a standardized depth. In the LL group, a Microlance (Becton-Dickinson, Meylan Cedex, France) was used. The lance has a sharp triangular edge of approximately 2.5 mm (Fig 1). With this lance, the skin puncture was performed by hand power and without a mechanical device.
Sham heel lancing was performed by using the above-mentioned method but turning the blunt end of a Microlance (LL) to the heel. The infants were then left undisturbed for 60 seconds. Thereafter the untraumatized heel was squeezed for 120 seconds. The nurse (nurse 1) was instructed to squeeze the heel as if to obtain blood after a heel prick.
The facial reaction was videotaped with a Panasonic VCR NV-M 10 (with a real-time counter; Matsushita, Ltd, Japan). The changes in facial activity were later analyzed by two blinded observers, using a Panasonic AG-5700 with remote control and a color monitor. The infant's cry reaction was recorded on a Sony tape-recorder with a Sennheiser (Wedemark, Ireland) MKE microphone. Each parameter was recorded during the whole procedure.
The changes in facial activity were assessed with the Neonatal Facial Scoring System (NFCS).12,13 The facial actions scored included brow bulge, eyes squeezed shut, deepening of the naso-labial furrow, open lips, a taut cupped tongue, and stretching of the mouth (vertically and horizontally). The two observers analyzing the results of the videotape or audiotape had not been informed of the purpose of the study and were unaware of the group to which the infant had been allocated. Each observer assessed the data independently and could not communicate findings to the other. Before participating, the observers were trained for 8 hours with a test tape. The facial reactions were analyzed during the first 15 seconds after the skin was punctured and again during the first 15 seconds when manipulation took place, after a pause of 60 seconds. The observation period was divided into 15 1-second intervals. The presence or absence of the six variables during each interval was recorded. Thus, the range for each variable was from 0% to 100%. The NFCS score is presented as percent positive (present) scores and the total range would therefore be 0% to 600%. Cry was defined as a high-pitched vocalization. The audiotape was reviewed to determine the latency to cry from the skin puncture, the duration of the first cry, and the total time the infant cried during the procedure. When measuring the total crying time, no pauses for more then 2 seconds were included.
The efficacy of each method was assessed by measuring the total time needed to obtain enough blood (0.4 mL) for the PKU test and the number of skin punctures required.
In the SL group, 2 infants screamed before the puncture was performed (nos. 18 and 19). In the same group, 1 infant in gestational week 35 was excluded (no. 72). These patients were not replaced. A summary of the results are presented in Table2.
Interobserver Reliability of the NFCS Score
Interobserver reliability was 91%, computed for all the participants, and for each variable in each time segment, using the conservative Facial Action Coding System reliability formula.14 This means that the two observers made the same evaluation in 91% about the presence or absence of each variable when analyzing each time segment of the video recording.
NFCS Score After the First Skin Puncture (0 to 15 Seconds)
The NFCS score is given in Fig 2. The median NFCS score was 247 in the VP group, 333 in the SL group, and 460 in the LL group. The difference between VP and LL was highly significant (P < .0005). The difference between VP and SL was also significant (P < .02, Mann-Whitney U test).
NFCS Score During Manipulation for Blood Sampling (60 to 75 Seconds)
The NFCS score is given in Fig 3. The median NFCS score was 142 in the VP group, 420 in the SL group, and 497 in the LL group. All differences between VP and lancet groups were significant (P < .001, Mann-WhitneyU test).
NFCS Score During Sham Procedure (0 to 15 Seconds)
When the blunt end of the lancet was used, the NFCS scores were 0 in 7/11 (range, 0 to 74) of the cases, indicating that this procedure was not painful.
NFCS Score During Sham Procedure (60 to 75 Seconds)
The median NFCS scores during heel squeezing was 87 (range, 0 to 366). In comparing NFCS score during heel squeezing, the results were highly significant when comparing the sham group with the SL and LL groups (Mann-Whitney U test, P < .0001 andP < .0006, respectively). No significant difference was found between the VP and the sham group (Mann-WhitneyU-test, P = .5).
Latency to Cry and Cry Duration After First Skin Puncture
After the first skin puncture, a cry was recorded in 22/50 (44%) infants in the VP group, 34/47 (72%) in the SL group, and 17/20 (85%) in the LL group. No child in the sham group reacted with a cry during the first 60 seconds. The difference between the groups is highly significant (P = .0001, log rank test, χ2). Analysis of the duration of the first cry showed no difference between SL and VP, although a tendency to longer duration was seen in the LL group.
Total Duration of Cry
The total duration of the cry was significantly shorter in the VP group than in the SL group (P < .001, log rank test, χ2). The median values in the respective groups were: 82 seconds, range 1 to 249 (VP); 270 seconds, range 52 to 578 (SL); 228 seconds, range 95 to 578 (LL); and 90, range 1 to 115 (sham). All patients in the LL, SL, and sham group reacted with a cry during the entire PKU sampling procedure compared with 68% in the VP group.
Efficacy of the Procedure
In 86% (43/50) of the cases only one skin puncture was needed to obtain enough blood in the VP group. In the LL group, the first attempt was successful in 40% (8/20) of the cases (VP-LL, P < .01, 2 × 2 Table, χ2) and in the SL group in only 19% (9/47) of the cases (VP-SL, P = .0025, 2 × 2 Table, χ2). The numbers of skin punctures needed are shown in Fig 4.
Time Taken to Complete the Procedure Successfully
The median values in the respective groups were: 191 seconds (VP), 419 seconds (SL), and 279 seconds (LL). The differences between the groups were significant (VP-SL, P = .00002, VP-LL,P < .05, log rank test, χ2).
Hitherto, heel lancing has been the method of choice for sampling blood in neonates. This method causes considerable distress.1,15,16 Rutter and Barker17studied the invasive procedures in a newborn intensive care unit. They found that blood sampling by heel prick was the commonest procedure (55% of the total). One premature infant in gestational week 23 underwent 450 invasive procedures.
Repetitive heel pricks can lead to sensitization and hyperalgesia of skin of the heel and mount a chronic pain response. Because of sensitization, the nociceptive threshold will decrease after repeated heel pricks. Even though EMLA does not alleviate the pain caused by heel lancing it has been found to reverse this hypersensitivity in preterm infants.18 Early experience of pain may be a factor contributing to the development of inappropriate strategies for coping with pain later in life.19,20 Various methods, such as the use of topical local anesthesia, nonnutritive sucking, and the sucking of sucrose solutions have been tried for alleviating pain.2-57-9 Some of these methods have shown that the pain caused by heel lancing can be reduced slightly, but no method offers effective alleviation of pain. The psychological effect and the future implications of the association of pleasurable events, such as breastfeeding, in connection with nociceptive stimulus are not known.
Facial expression and crying have been shown to distinguish between pain and no-pain situations. The NFCS was introduced by Grunau and Craig12,13,21 in 1987 and has been shown to reflect to changes in pain intensity and is therefore useful for evaluating pain-relieving interventions. The high interobserver agreement found in this study supports a high reliability of the scale, as also does the low NFCS score in the sham group.
The data in the present study clearly suggest that venipuncture is less painful than heel lancing. The 60-second interval, after the first skin puncture, was imposed to distinguish the nociceptive response between a single skin puncture from the response of manipulation for blood collection. When assessing pain during the first time interval (0 to 15 seconds), the LL group scored highest, indicating that this lancet was more traumatizing and more painful than the smaller one (SL). During the manipulation for blood sampling (>60 seconds), the NFCS score indicates that the blood sampling was less painful in the VP group than in the SL and LL groups. Very high NFCS scores were observed during squeezing of the heel. Squeezing thus seems to be a very painful procedure. NFCS scores varied during squeezing of the heel in the sham group. The squeezing is not a standardized painful stimulation and an interindividual variation could explain the range of the response. However, most participants did not react with a cry during the first 60 seconds. This finding resembles the data from a previous study,3 indicating that this is not a painful stimulus. The total duration of crying was slightly longer in the SL group than in the LL group, which suggests that this method, because of the prolonged squeezing and multiple skin punctures, is not necessarily better, although the NFCS scores were lower.
The results clearly demonstrate that venipuncture is more effective than heel lancing when blood sampling is performed in neonates. In the majority of cases in the VP group, one skin puncture was sufficient for obtaining enough blood for PKU, whereas this was true in only 40% in the LL group and 19% in the SL group. This, of course, indicates that VP is more effective than heel lancing, but also that the more traumatizing LL is more effective than the SL. The same pattern was found when measuring the time needed to obtain blood for the PKU test.
One reason why heel lancing has been the method of choice could be that it is thought to require less teaching time than venipuncture. The authors of the present study do not agree with this idea. Both theoretical and practical training is required to obtain blood with a heel lance. In this study, a trained neonatal intensive care unit nurse performed all tests. She did not receive special training only in venipuncture, but in both methods. However, further studies looking into the effect of different skillfulness among nurses would be interesting.
Whether venipuncture requires more or less training than heel lancing seems unimportant when compared with the obvious benefit for the child, both in the short-term and long-term perspective.
Venipuncture is more effective and less painful than heel lancing for blood tests in neonates.
We thank Celeste Johnston and Veronica Berggren (nurse 1) for their help.
- Received April 7, 1997.
- Accepted September 10, 1997.
Reprint requests to (B.A.L.) Department of Pediatric Anesthesia and Intensive Care, Pain Treatment Service, KS/St. Göran's Children's Hospital, S 112 81 Stockholm, Sweden.
- EMLA =
- eutectic mixture of local anesthetic •
- PKU =
- phenylketonuria •
- VP =
- venipuncture group •
- SL =
- small-lancet group •
- LL =
- large-lancet group •
- NFCS =
- Neonatal Facial Coding System
- Larsson BA,
- Jylli L,
- Lagercrantz H,
- Olsson GL
- ↵Ramaioli F, Amici D, Guzinska K, Ceriana P, Gasparoni A. EMLA Cream and the Premature Infant. Presented at the 11th ESRA Congress; 1993; Dublin, Ireland
- Field T,
- Goldson E
- Blass EM,
- Hoffmeyer LB
- Harpin VA,
- Rutter N
- ↵Ekman P, Friesen WV. Manual for the Facial Coding System. Palo Alto, CA: Consulting Psychologist's Press; 1978
- ↵Johnston CC, O'Shaugnnessy D. Acoustical attributes of pain cries; distinguishing features. In: Dubner R, Gebbart GF, Bond MR, eds. Advances in Pain Research and Therapy, V. New York, NY: Raven Press; 1988;336–340
- Copyright © 1998 American Academy of Pediatrics