Could clinical scores guide the surgical treatment of ... - Springer Link

Pediatr Surg Int (2012) 28:271–276 DOI 10.1007/s00383-011-3016-z

ORIGINAL ARTICLE

Could clinical scores guide the surgical treatment of necrotizing enterocolitis? Vicente Ibań˜ez • Miguel Couselo • Verońica Marijuań Juan Jose´ Vila • Carlos Garcıá-Sala

•

Accepted: 3 October 2011 / Published online: 15 October 2011 Ó Springer-Verlag 2011

Abstract Purpose Test the diagnostic reliability of the score for neonatal acute physiology-perinatal extension-II (SNAPPE-II) and the metabolic derangement acuity score (MDAS) as predictors of surgery in patients with necrotizing enterocolitis (NEC). Methods The SNAPPE-II and the MDAS were applied to 99 patients with NEC. Both the scores were calculated at the moment of diagnosis (T0) and when surgical assessment was required (T1). The main outcome was the need of surgical revision. Comparison between models was made through their receiver operator characteristics (ROC) curves. Results Thirty-five patients required surgical treatment (group A) and 64 responded to medical therapy (group B). Median SNAPPE-II was 22 versus 5 for group A (U test 621, p = 0.002) at T0; and 22 versus 10 for group A (U test 487, p = 0.01) at T1. Measuring the value of the SNAPPEII as a predictor of surgery, the ROC curve was 0.69 (CI 95%, 0.57–0.80) at T0 and 0.67 (CI 95%, 0.55–0.80) at T1. Median MDAS were 2 for both groups A and B at T0 (U test 890.5, p = 0.113) and 2 versus 1.5 for group A at T1 (U test 570, p = 0.043). The ROC curve for MDAS was 0.59 (CI 95%, 0.47–0.71) at T0 and 0.64 (CI 95%, 0.52–0.77) at T1. Conclusions The diagnostic performance of the SNAPPE-II offers mild results in the moment of the diagnosis of NEC, and at T1. The MDAS is non significant at T0 and obtains moderate results at T1. These results do not encourage using the SNAPPE-II and the MDAS as definite V. Ibań˜ez M. Couselo (&) V. Marijuań J. J. Vila C. Garcıá-Sala Hospital Infantil La Fe, Valencia, Spain e-mail: [email protected]

tools to decide for surgical treatment of the patients affected by NEC. Keywords Necrotizing enterocolitis Score ROC curve Surgical treatment

Introduction Necrotizing enterocolitis (NEC) affects the 3–7% of preterm and low weight birth infants [1]. It is considered the most frequent surgical emergency during the neonatal period [2] and 27–50% of these patients will need surgical treatment. Despite all medical advances in the management and treatment of these infants, the mortality rates remain as high as 12–30% [1]. The classic surgical approach to the patient with perforated NEC consists of laparotomy. Primary peritoneal drainage (PPD) was proposed initially by Ein [3] as a treatment for those unstable patients who could not undergo surgery in an attempt to stabilize their condition but afterwards PPD was used as an initial therapy, especially in the most seriously ill patients [4–7]. The indication for surgical treatment can be controversial since the radiographic signs of perforation might not be present in all the children with a necrotic bowel [8–10] Objective methods of measuring illness severity and mortality can provide with valuable data from patients, help in the decision making process of those in whom the surgical treatment is not clear and allow to compare information. Several authors have applied clinical scores to choose from the therapeutic attitudes commented previously. The score for neonatal acute physiology perinatal extension-II (SNAPPE-II) has been used to test clinical response to peritoneal drainage in neonates with NEC [11]: in a group

123

272

Pediatr Surg Int (2012) 28:271–276

of patients who had been previously treated with PPD, those who received a subsequent laparotomy presented a lower score than those who were not operated. The metabolic derangement acuity score (MDAS) proposed by Tepas [12] has been suggested as a valuable method to encourage PPD or operative debridement in infants with signs of intestinal perforation: patients with lower scores were stated as good candidates to be treated with PPD, whereas patients with higher scores got more benefit from a laparotomy. Till now these models have not proved their value in new samples of patients. This study was conducted to determine the diagnostic performance of SNAPPE-II and MDAS as predictors of surgical treatment in infants with NEC.

Materials and methods We studied retrospectively the patients diagnosed of NEC in a 3rd level hospital from January 2002 to May 2008. This study has received the approval of our Institutional Review Board and, as a retrospective one, data have been managed keeping personal information concealed. Records of all the patients with NEC in the neonatal intensive care unit (NICU) or in the Neonatology unit were reviewed, including suspected cases—Bell’s stage I [13]. Patients who were suspected to have an isolated intestinal perforation according to their clinical presentation (less than 1,000 g of weight, initial radiological finding of pneumoperitoneum with no previous impairment, treatment with

ibuprofen or dexamethasone) were not included in the study. Those patients whose clinical features were not definite enough to diagnose clinically either isolated intestinal perforation or NEC were included in the study. Only those cases diagnosed in our hospital were considered, in order to reduce missing data and preserve diagnostic criteria, so patients submitted from other hospitals were excluded. Every patient was diagnosed of NEC by a staff paediatrician in the NICU or Neonatology unit according to clinical gastrointestinal manifestations (gastric residuals, abdominal distension, occult or macroscopic gastrointestinal bleeding) and radiological findings (bowel distension, pneumatosis intestinalis, portal vein gas, intraperitoneal fluid or pneumoperitoneum) (Table 1). The need of surgical therapy was considered the main outcome. The decision to operate on was taken according to the surgeon’s criteria: Table 2 summarises the clinical status of each patient (Bell’s stage [13] and radiological findings) before surgery. Surgical approach consisted always in laparotomy. Before surgery, every patient received the corresponding treatment according to his/her morbidities. Two groups were defined. Group A: patients who required surgery and group B: patients who did not require surgery. The SNAPPE-II and MDAS were applied at two different moments: when the first diagnosis of NEC was made (T0) and in the moment in which the surgeon evaluated the patients due to clinical deterioration (T1). In the neonates whose evolution was satisfactory and who did not require surgical assessment, both the scores were calculated at the diagnosis and 36 h after the diagnosis: this T1 moment was

Table 1 Characteristics of the study group

Group A (surgical treatment (35 patients))

Group B (non-surgical treatment (64 patients))

Mean (SD)

Mean (SD)

Gestational age (weeks)

29.2 (4.25)

32.2 (5.15)

Birth weight (grams) Age at diagnosis (days)

1221.9 (682.02) 12.9 (11.22)

1853.6 (942.37) 17.5 (17.33)

Age at surgery (days)

15.4 (12.6)

–

Delay between diagnosis and surgery (days)

2.4 (2.66)

–

n (%)

n (%)

Bell stage I at T0

–

24 (37.5%)

Bell stage II at T0

22 (63%)

39 (61%)

Bell stage III at T0

13 (37%)

1 (1.5%)

Dilatation

12 (34%)

25 (38%)

Pneumatosis

13 (37%)

40 (62%)

Portal gas

3 (9%)

–

X-ray findings at T0

Intraperitoneal free fluid

4 (11%)

–

Pneumoperitoneum

3 (9%)

–

13 (37%)

4 (6%)

Deaths T0 moment of the diagnosis

123


273

Table 2 Characteristics of the patients of group A before surgery Characteristics

Number of patients (%)

Pneumatosis

9 (25%)

Pneumoperitoneum

12 (34%)

Portal gas

6 (18%)

Fixed bowel loop

5 (14%)

Intraperitoneal free fluid

3 (9%)

Bell stage I

0

Bell stage II

19 (54%)

Bell stage III

16 (46%)

considered in that manner after having calculated the median time elapsed from the diagnosis to the clinical examination by a surgeon in the patients who presented clinical deterioration. In the neonates who required immediate surgical treatment after the diagnosis of NEC, each score (SNAPPE-II or MDAS) was considered to be the same in T0 and T1. Therefore, the whole sample was analysed both at T0 and T1. The diagnostic performance of both models was evaluated through their receiver operator characteristics (ROC) curves (Table 3). Comparison between scores was performed with the Mann–Whitney U test. Statistical analysis was carried out using the statistical package for social sciences (SPSS 13.0). The comparison of the area under curve (AUC) was done with the method proposed by Hanley and McNeil for related curves using the EPIDAT 3.0. Table 3 Parameters of SNAPPE-II and parameters of MDAS Scores

Variables

SNAPPE-II

Birth weight Mean blood pressure Lowest temperature PO2/FiO2 ratio Lowest serum pH Urine output Multiple seizures Apgar Score at 5 min Small for gestational age \3rd percentile

MDAS

Metabolic acidosis Neutropenia Left shift of segmented neutrophils Hyponatremia Bacteremia Trombocytopenia Hypotension

The values of SNAPPE-II and MDAS were taken from [11, 12]

Results One hundred thirty-six patients were treated for suspected NEC during the 6 years of the study. Thirty-seven patients were excluded because they were diagnosed in other centres, and 99 patients remained for this study. Sample characteristics are showed in Table 1. Forty-one patients were examined by a surgeon due to impairment despite initial medical treatment, and 34 of these were operated on. Sixty-four patients did not require surgery and responded to medical treatment. Seventeen patients died: thirteen from group A and 4 from group B. The four deaths in group B corresponded to infants who had a good response to non-operative treatment of NEC and the causes of death were cardiac failure in two cases, multi-organic failure as a consequence of respiratory insufficiency, and one remained unknown. One of the patients in whom surgical treatment had been indicated (thus belonging to group A) presented a sudden and rapid evolution and died before the surgical intervention. Since the necropsy confirmed the diagnosis of NEC and the decision to operate on was taken before the death of the patient, he was included in group A. In this way, group A had 35 patients although only 34 of them had undergone surgery. Operative findings confirmed NEC in 24 cases (71%) and isolated perforation in 10 (29%). Median SNAPPE-II points for groups A and B were significantly different at T0: 22 versus 5 for group A (U test 621, p = 0.002), and at T1: 22 for group A versus 10 for group B (U test 487, p = 0.01) Table 4. Measuring the value of the SNAPPE-II as a predictor of surgery, the AUC was 0.69 (CI 95% 0.57–0.80) at T0 and 0.67 (CI 95% 0.55–0.80) at T1. A diagnostic test is considered perfect when its area under the ROC curve is 1; it is considered a good test when the area under the ROC curve is higher than 0.8 and it is considered not valid when the confidence interval (CI) of the area under the curve includes 0.5. Median MDAS points were 2 for both groups A and B at T0 (U test 890.5, p = 0.113); and 2 for group A vs 1.5 for group B at T1 (U test 570; p = 0.043). The AUC to test the diagnostic performance of MDAS was 0.59 (CI 95% 0.47–0.71) at T0 and 0.64 (CI 95% 0.52–0.77) at T1 (Figs. 1, 2). Comparison between ROC curves was considered useless at T0, as only the SNAPPE-II showed an AUC which was statistically significant. Using MDAS at this moment to predict the need for surgical therapy did not show better results than tossing it up according to its CI which included the 0.5 value. There were no differences between the SNAPPE-II and the MDAS curves at T1 (p = 0.62).

123

274


Table 4 Scores of SNAPPE-II and MDAS

T0 moment of the diagnosis, T1 moment of the assessment by a surgeon

Group A (surgical treatment (35 patients) median (p25–p75))

Group B (non-surgical treatment (64 patients) median (p25–p75))

p

SNAPPE-II at T0

22 (5–33)

5 (5–17)

SNAPPE-II at T1

22 (10–33)

10 (5–34)

0.002 0.01

MDAS at T0

2 (1–3)

2 (1–2)

0.113

MDAS at T1

2 (2–3)

1.5 (1–3)

0.043

Discussion Management of NEC still remains a challenge because despite all the advancements in neonatal care and more knowledge regarding its pathophysiology, the mortality and the morbidity of surgery in these patients have not been reduced significantly. Bell’s classification [13] has been widely used to stage NEC but the surgical indications proposed by Bell have changed since they were firstly presented. [14–16]. Pneumoperitoneum is considered a major indication of surgery, and other signs such as persistence of asymmetrical bowel loops, portal vein gas, positive paracentesis or clinical deterioration may warrant surgical therapy [8, 14, 17–20]. The main problem with the radiological signs is that they might have a high positive predictive value (almost 100% for pneumoperitoneum) but a very low sensitivity (less than 50%) [14, 21, 22]. With regard to clinical deterioration, there is always a subjective component that might lead to error. Objective and standard clinical data would be useful to guide therapeutic decisions more accurately. Several laboratory tests have been classically used in NEC (platelet count, neutrophil count, presence of metabolic acidosis, electrolyte levels) [23–26] but up to now individual laboratory parameters have just been useful as an orientation of the extent of NEC [25, 26]. Different indexes including some of these laboratory parameters have been used to asses more accurately the management of NEC [26, 27]. Bonnard studied SNAPPE-II to guide the treatment of very low birth weight newborns with perforated NEC [11]. He suggested that, after the insertion of a PPD, the performance of a salvage laparotomy could be guided by this score. Patients who were less affected (lower scores) went through a laparotomy while those who were in worse conditions (higher scores) did not—attitude which supports Ein’s initial concept of peritoneal drainage [3]. The SNAPPE-II was designed as an admission score for NICUs to characterise mortality risk within the first 12 h. It can be used in patients with all birth weights and gestational ages. It includes several variables that provide a maximum punctuation of 162 and a minimum of 0; the higher punctuation, the higher risk of mortality [28]. Although Bonnard proposed to use the SNAPPE-II to test the

123

Fig. 1 Comparison between the ROC curves of SNAPPE-II and MDAS at the moment of the diagnosis

Fig. 2 Comparison between the ROC curves of SNAPPE-II and MDAS at the moment of the assessment by a surgeon

response to PPD there are no data about its use as a predictor of primary surgical treatment in NEC. The MDAS includes some of the laboratory parameters which are frequently used in the diagnosis and management of NEC [12]. When it was initially created, the


natural history of NEC and intestinal perforation [29] were taken on account and Tepas established three different pathways which would lead to bowel perforation in very low birth weight neonates: sudden perforation with no previous metabolic derangement (which would correspond to neonatal intestinal perforation), bad response to medical treatment and metabolic derangement followed by signs of free intraperitoneal air, and bad response to medical treatment plus metabolic derangement without signs of free intraperitoneal air (the last two would correspond to NEC). The MDAS was presented as a good tool in the management of these patients, treating them with PD when there were not metabolic alterations, and with laparotomy when there were signs of metabolic impairment. Moreover, the metabolic derangement has been recently proposed as a guide to decide the timing of surgical treatment in the patients with NEC and no signs of free peritoneal air in the radiological studies [30]: according to Tepas, when three or more components of metabolic derangement are present, surgical treatment should be considered strongly, although his latest score has not been validated yet for this purpose. Metabolic impairment is an outstanding event in the progression of patients with NEC, and could be useful to distinguish between NEC and isolated intestinal perforation. However, when the clinical features of the patient do not suggest strongly an isolated intestinal perforation (sudden pneumoperitoneum in the X-ray film without previous illness, previous treatment with ibuprofen or dexamethasone) it can very difficult to differentiate one of these two disorders prior to surgical intervention. [29, 31– 33]. Besides, considering that this paper tries to assess the capability of the SNAPPE-II and the MDAS to guide the timing of surgery, those patients who offered diagnostic doubts between NEC and isolated intestinal perforation were included in the study. By applying both scores in two different moments in the evolution of NEC we considered that they might be able to take into account the physiological derangement of those patients in whom the disease progressed. T0 represented the moment of the first diagnosis: it was time when the initial severity of NEC was assessed. T1 was considered firstly in the neonates who, despite a correct non-operative management, did not improve their clinical and radiological situation. In order to achieve a better diagnostic performance those patients who presented a favourable evolution and who were not evaluated by a surgeon were also included in the T1 analysis. The median time which elapsed since the diagnosis of NEC and the clinical examination by a surgeon in the patients of the sample was 36 h; therefore, this was the T1 moment used to calculate the scores in those patients who were not examined by a surgeon. When the decision to operate is uncertain, clinical scores could have great interest. In this study, the diagnostic

275

performance of SNAPPE-II and MDAS for surgery offered mild results. The SNAPPE-II obtained poor/fair areas under the ROC curve at T0 and T1: the lower limit of the CI was under 0.60 and the upper limit was 0.80 in both moments. This score was created to predict mortality in patients in NICUs in the first 24 h of life. These neonates present a wide group of pathologic events with a typical early onset, period in which NEC might not have developed yet, so the patients for whom the SNAPPE-II was designed could be slightly different from the group of patients with NEC. Thus, applying the SNAPPE-II to such a specific group of patients may go beyond the original intention of the score. With regard to MDAS, the results were not better than those of the SNAPPE-II. A non-significant result was obtained when the first diagnosis of NEC was made; in that moment (T0) metabolic deterioration might not have been completely developed. Considering the moment when the patients required surgical assessment (T1) the area under the ROC curve was not entirely satisfactory: the CI between 0.52 and 0.77 was significant but did not reach the 0.80. Thus, in the group of children with an initial bad response to medical treatment that should have developed metabolic alterations, the MDAS was not a good test to discriminate who would undergo surgery. Our results do not support the use of the SNAPPE-II and the MDAS to asses accurately the need of surgical therapy in patients with NEC. The SNAPPE-II showed moderate results at the moment of diagnosis and when there was clinical deterioration; the MDAS had similar results only when applied at the moment of clinical impairment. Other scores might provide useful information which could help in the decision making process in NEC. Conflict of interest of interest.

The authors declare that they have no conflict

References 1. Henry MC, Moss RL (2008) Neonatal necrotizing enterocolitis. Semin Pediatr Surg 17:98–109 2. Henry MC, Moss RL (2004) Current issues in the management of necrotizing enterocolitis. Semin Perinatol 28:221–233 3. Ein SH, Marshall DG, Girvan D (1977) Peritoneal drainage under local anesthesia for perforations from necrotizing enterocolitis. J Pediatr Surg 12:963–967 4. Cheu HW, Sukarochana K, Lloyd DA (1988) Peritoneal drainage for necrotizing enterocolitis. J Pediatr Surg 23:557–561 5. Ein SH, Shandling B, Wesson D, Filler R (1990) A 13 year experience with peritoneal drainage under local anesthesia for necrotizing enterocolitis perforation. J Pediatr Surg 25:1034–1037 6. Lessin MS, Luks FI, Wesselhoeft CW, Gilchrist BF, Iannitti D, De Luca FG (1998) Peritoneal drainage as definitive treatment for intestinal perforation in infants with extremely low birth weight (\750 g). J Pediatr Surg 33:370–372

123

276 7. Moss RL, Dimmitt RA, Barnhart DC, Sylvester KG, Brown RL, Powell DM et al (2006) Laparotomy versus peritoneal drainage for necrotizing enterocolitis and perforation. N Eng J Med 354:2225–2234 8. Kosloske AM (1994) Indications for operation in necrotizing enterocolitis revisited. J Pediatr Surg 29:663–666 9. Sharma R, Tepas JJ III, Mollitt DL, Pieper P, Wludyka P (2004) Surgical management of bowel perforations and outcome in very low birth weight infants (B1,200 g). J Pediatr Surg 39:190–194 10. Sharma R, Hudak ML, Tepas JJ III, Wludyka PS, Marvin WJ, Bradshaw JA, Pieper P (2006) Impact of gestational age on the clinical presentation and surgical outcome of necrotizing enterocolitis. J Perinatol 26:342–347 11. Bonnard A, Zamakhshary M, Ein S, Moore A, Kim PCW (2008) The use of the score for neonatal acute physiology-perinatal extension (SNAPPE II) in perforated necrotizing enterocolitis: could it guide therapy in newborns less than 1500 g? J Pediatr Surg 43:1170–1174 12. Tepas JJ III, Sharma R, Hudak ML, Garrison RD, Pieper P (2006) Coming full circle: an evidence based definition of the timing and type of surgical management of very-low-birth-weight (\1000 g) infants with signs of acute intestinal perforation. J Pediatr Surg 41:418–422 13. Bell MJ, Temberg JL (1978) Neonatal necrotizing enterocolitis: therapeutic decisions based upon clinical staging. Ann Surgery 187:1 14. Kosloske AM, Papile L-A, Burstein J (1980) Indications for operation in acute necrotizing enterocolitis of the neonate. Surgery 87:502–508 15. Grosfeld JL, Cheu H, Schlatter M, West KW, Rescorla FJ (1991) Changing trends in neonatal enterocolitis. Ann Surg 214:300–307 16. Gordon PV, Swanson JR, Attridge JT, Clark R (2007) Emerging trends in acquired neonatal intestinal disease: is it time to abandon Bell’s criteria? J Perinatol 27:661–671 17. Buonomo C (1999) The radiology of necrotizing enterocolitis. Radiol Clin North Am 37:1187–1198 18. Molik KA, West KW, Rescorla FJ, Scherer LR, Engum SA, Grosfeld JL (2001) Portal venous air: the poor prognosis persists. J Pediatr Surg 36:1143–1145 19. Sharma R, Tepas JJ III, Hudak ML, Wludyka PS, Mollitt DL, Garrison RD, Bradshaw JA, Sharma M (2005) Portal venous gas and surgical outcome of neonatal necrotizing enterocolitis. J Pediatr Surg 40:371–376 20. Lin PW, Stoll BJ (2006) Necrotising enterocolitis. Lancet 368:1271–1283

123

Pediatr Surg Int (2012) 28:271–276 21. Tam AL, Camberos A, Applebaum H (2002) Surgical decision making in necrotizing enterocolitis and focal intestinal perforation: predictive value of radiologic findings. J Pediatr Surg 37:1688–1691 22. Faingold R, Daneman A, Tomlinson G, Babyn PS, Manson DE, Mohanta A et al (2005) Necrotizing enterocolitis: assessment of bowel viability with color Doppler US. Radiology 235:587–594 23. O’Neill JA Jr, Stahlman MT, Meng HC (1975) Necrotizing enterocolitis in the newborn: operative indications. Ann Surg 182:274–279 24. Buras R, Guzzetta P, Avery G, Naulty C (1986) Acidosis and hepatic portal venous gas: indications for surgery in necrotizing enterocolitis. Pediatrics 78:273–277 25. Gupta SK, Burke G, Herson VC (1994) Necrotizing enterocolitis: laboratory indicators of surgical disease. J Pediatr Surg 29:1472–1475 26. Ragazzi S, Pierro A, Peters M, Fasoli L, Eaton S (2003) Early full blood count and severity of disease in neonates with necrotizing enterocolitis. Pediatr Surg Int 19:376–379 27. Dykes EH, Gilmour WH, Azmy AF (1985) Prediction of outcome following necrotizing enterocolitis in a neonatal surgical unit. J Pediatr Surg 20:3–5 28. Richardson DK, Corcoran JD, Escobar GJ, Lee SK (2001) SNAPII and SNAPPE-II: simplified newborn illness severity and mortality risk scores. J Pediatr 138:92–100 29. Cass DL, Brandt ML, Patel DL, Nuchtern JG, Minifee PK, Wesson DE (2000) Peritoneal drainage as definitive treatment for neonates with isolated intestinal perforation. J Pediatr Surg 35:1531–1536 30. Tepas JJ III, Sharma R, Leaphart CL, Celso BG, Pieper P, Esquivia-Lee V (2010) Timing of surgical intervention in necrotizing enterocolitis can be determined by trajectory of metabolic derangement. J Pediatr Surg 45:310–314 31. Hwang H, Murphy JJ, Gow KW, Magee JF, Bekhit E, Jamieson D (2003) Are localized intestinal perforations distinct from necrotizing enterocolitis? J Pediatr Surg 38(5):763–767 32. Blakely ML, Tyson JE, Lally KP, McDonald S, Stoll BJ, Stevenson DK et al (2006) Laparotomy vs. peritoneal drainage for necrotizing enterocolitis or isolated intestinal perforation in extremely low birth weight infants: outcomes through 18 months adjusted age. Pediatrics 117:e680–e687 33. Chiu B, Pillai SB, Almond PS, Beth Madonna M, Reynolds M, Luck SR, Arensman RM (2006) To drain or not to drain: a single institution experience with neonatal intestinal perforation. J Perinat Med 34:338–341