www.elsevier.com/locate/jpedsurg

Journal of Pediatric Surgery (2009) 44, 112–117

Analysis of outcome in 298 extremely low-birth-weightinfants with patent ductus arteriosusFrederick Alexander a,⁎, Louisa Chiub, Matthew Krohb, Jeffrey Hammelb, John Moorec

aHackensack University Medical Center, Hackensack, NJ 07601, USAbCleveland Clinic, Cleveland, OH 44195, USAcMetro Health Medical Center, Case Western Reserve University, OH 44109, USA

Received 30 September 2008; accepted 7 October 2008

0d

Key words:Extremely low birthweight;

Patent ductus arteriosus;Necrotizing enterocolitis;Indomethacin;Ligation

AbstractPurpose: Indomethacin is accepted therapy for patent ductus arteriosus (PDA) in ELBW infants(b1000 g). We hypothesize that surgical ligation may provide comparatively superior outcomes inselect ELBW infants.Methods: Predischarge outcomes of 298 ELBW infants with echocardiography-proven PDA wereretrospectively compared by treatment provided: no treatment (group 1, n = 54), indomethacin (group2, n = 140), ligation (group 3, n = 46), and ligation after indomethacin failure (group 4, n = 58). χ2 andWilcoxon rank sum tests were used to test for significance. Institutional review board approval wasobtained (IRB/05-00395).Results: Group 3 had significantly lower gestational age (P b .001), birth weight (P = .006), and 5-minuteApgar scores (P = .03) compared with group 2. Group 3 and group 1 had a higher rate of pretreatmentintraventricular hemorrhage (IVH) compared with group 2 (P b .001). By contrast, posttreatmentcomplications including acute renal failure, necrotizing enterocolitis, thrombocytopenia, and IVHoccurred more frequently in groups 2 (P = .004) and 4 (P = .001) compared with group 3. Survival was57.7% in group 1 compared with groups 2, 3, and 4 (82.4%, 86.0%, and 92.7% respectively; P = .001).Preoperative conditions associated with nonsurvival include gestational age (P = .009), birth weight (P =.002), maternal preeclampsia (P = .015), 5-minute Apgar score (P = .013), and sepsis (P = .018).Posttreatment complications associated with nonsurvival include acute renal failure (P = .002),thrombocytopenia (P = .002), and necrotizing enterocolitis (P = .034). Survival was not influenced by anycongenital comorbidity, pre- or posttreatment IVH, diameter of the PDA, or recurrence of the PDA afterindomethacin therapy.Conclusions: (1) Patent ductus arteriosis requires treatment in ELBW infants to maximize survival. (2)Indomethacin and surgical ligation permit equivalent survival in low-risk ELBW infants, butindomethacin results in a high failure and complication rate requiring operative salvage in a number ofpatients. (3) Surgical ligation permits survival of high-risk ELBW infants with a low complication rateand is preferable to indomethacin in ELBW infants with the above risk factors.© 2009 Elsevier Inc. All rights reserved.

Presented at the 39th annual meeting of the American Pediatric Surgical Association, Phoenix, AZ, May 27-June 31, 2008.⁎ Corresponding author. Tel.: +1 201 996 2921; fax: +1 201 996 2992.E-mail addresses: falexander@humed.com, falexander21@gmail.com (F. Alexander).

022-3468/$ – see front matter © 2009 Elsevier Inc. All rights reserved.oi:10.1016/j.jpedsurg.2008.10.019

113Analysis of outcome in ELBW infants with PDA

Patent ductus arteriosis (PDA) may persist after birth in graphy. Infants were then placed into 1 of 4 study groups: (1)

nearly 66% of extremely low-birth- weight (ELBW) infantsweighing 1000 g or less, occurringmost frequently in those withlower gestational weight, hyalinemembrane disease, or a historyof antenatal steroids [1]. Inmany infants, itmay cause left to rightshunting leading to congestive heart failure and, in a significantnumber of infants, to cerebral and mesenteric ischemia [2].

Indomethacin is widely used to promote ductal closure [3]but fails in 10% to 40% of ELBW infants [4,5]. Furthermore,this drug may cause renal dysfunction (acute renal failure[ARF]), necrotizing enterocolitis (NEC), or spontaneousintestinal perforation (SIP) because of alterations in localtissue perfusion as well as intraventricular hemorrhage (IVH)secondary to its antiplatelet effects.

Several clinical investigations have suggested that surgicalligation of PDA may be preferable to indomethacin in selectedELBW infants [5,6]. However, other reports have highlightedthe possible dangers [7] and futility of surgical ligation [8] Thepurpose of this study is to determine the relative risks andbenefits of indomethacin vs surgical ligation in ELBW withPDA. We hypothesize that surgical ligation may providecomparatively superior outcomes in select ELBW infants.

1. Methods

Patient files of all ELBW infants with PDA documented byechocardiography from June 1996 to April 2005 weredownloaded from a database operated by the Division ofNeonatology at MetroHealth Medical Center to a CD andwere analyzed. No infants were excluded. Prior approval hadbeen obtained by the institutional review board atMetroHealthMedical Center (IRB/05-00395). In each case, 2-dimensionalechocardiography was used to evaluate a machinery heartmurmur often associated with congestive heart failure and/orpulmonary edema. Echocardiographic measurements includedleft to right shunt and PDA diameter (mm).

The decision to treat or not was made by the neonatal teamusing a protocol that includes indomethacin 0.2 mg/kg at 0,12, and 24 hours as primary treatment of a PDA unlesscontraindicated by IVH, renal dysfunction (ARF), or severethrombocytopenia 50,000 platelets/HPF or less. No prophy-lactic indomethacin was given. Infants who required treatmentwith the above contraindications underwent primary surgicalligation using a left, fourth rib posterior-lateral thoracotomyand ligation with 2 oiled 2.0 silk ties. A tube thoracostomywas used after transplural thoracotomy, and none was usedafter extraplural thoracotomy. The tube was removed thefollowing day. In some cases, no treatment was given whenjudged by the neonatal team to be hazardous or unwarranted.

All patients were reevaluated after initial treatment byechocardiography, and success or failure of treatment wasrecorded. In the event of failure, alternative treatment waschosen consisting of either additional doses ofindomethacin or surgical ligation. All infants requiringsubsequent treatment were again reevaluated by echocardio-

no treatment, (2) indomethacin only, (3) surgical ligationonly, (4) surgical ligation after indomethacin failure.

Treatment groups were compared with respect to demo-graphic variables including gestational age, birth weight,5-minute Apgar scores, day of life at diagnosis, and treatmentprovided; comorbidities including congenital anomalies,maternal complications, and pretreatment conditions includingcongestive heart failure, IVH, NEC or SIP, and ARF or liverfailure; posttreatment complications including NEC or SIP,new or increased IVH, ARF, bleeding, thrombocytopenia, andgastrointestinal bleeding; and surgical complications includingwound infection, pneumothorax, chylothorax, bleeding,laryngeal nerve palsy, and division of pulmonary artery.Finally, groups were compared with respect to days onconventional mechanical ventilation, days on total parenteralnutrition (TPN), and survival. Acute renal failure was definedas a creatinine of greater than 2.0 mg/dL and/or a urine outputof less than 1 mL/kg per hour. Necrotizing enterocolitis wasdefined by radiographic findings of pneumatosis intestinalis orfree intraperitoneal air. Spontaneous intestinal perforation wasdiagnosed by a gasless abdomen with or without free air andabdominal discoloration with signs of sepsis. For the purposesof this study, NEC and SIP were taken together and analyzedas a single entity. Sepsis was defined as a clinical conditionrequiring a minimum of 3 days of IV antibiotic therapy.

Comparisons of treatment groups were performed withrespect to both initial treatment (medical, surgical, or neither)andwith respect to treatment received,which subdivided initialmedical treatment into those who did and did not requiresubsequent surgical treatment. The comparisons of multiplegroups with respect to categorical variables were performedusing χ2 tests or, alternatively, with Fisher's Exact tests, if anyexpected cell counts, assuming no association, were less than5. Comparisons of treatment groups with respect to quantita-tive variables were performed using the Kruskal-Wallis tests.Post hoc pairwise comparisons of treatment groups wereperformed with respect to categorical and quantitativevariables, respectively, using χ2 (or Fisher's) and Wilcoxonrank sum tests. Univariable associations between individualoutcomes (survival, complications, and medical treatmentfailure) and study variables were assessed using logisticregressionmodels to produce estimates of odds ratios and 95%confidence intervals. Associations were also assessed multi-variably using logistic regression an odds ratios adjusting forgestational age, birth weight, and the presence of anycongenital or maternal comorbidity as covariates. P valuesare reported for statistical tests and are considered significantwhen .05 or less. Analyses were performed using R version2.3.1 (R Development Core Tesen, Vienna, Austria).

2. Results

Two hundred ninety-eight ELBW infants were found tohave a PDA during the 106-month study. Infants were

Fig. 1

114 F. Alexander et al.

selected for initial treatment by the neonatal team as follows:no treatment in 54 infants; indomethacin in 198 infants; andprimary surgical ligation in 46 infants. Initial echocardio-graphy demonstrated a left to right shunt in 93% of allinfants. Mean PDA diameter differed significantly betweenall groups and was 1.94 ± 0.9 mm in the nontreatment group,2.38 ± 0.7 mm in the indomethacin group, and 2.68 ± 0.6 mmin the surgical group (P = .001, Fig. 1).

Infants in the surgical group had significantly lowergestational age, birth weight, and 5-minute Apgar scorescompared with the other 2 groups of infants. Predictably,infants in the surgical group and the notreatment group had asignificantly higher incidence of IVH compared with theindomethacin group, and the infants in the notreatment grouphad a significantly higher incidence of thrombocytopeniacompared with the other 2 groups (Table 1).

Repeat echocardiography was performed in 153 patientsafter initial treatment with indomethacin demonstrating45.8% (70/153) recurrence vs 2.2% (1/46) recurrence in

Table 1 Initial treatment

None (n = 54) Ind

Gestational age (wk) 25.7v1.9 26.Birth weight (g) 729.6 ± 169.6 739Apgar (5 min) 6.37 ± 1.7 6.8Day of life diagnosis 5.89 ± 10.9 4.1Structural heart 7 (13%) 9 (4Other congenital anomalies 14 (26%) 36Maternal complications 25 (46%) 113Sepsis 37 (77%) 119Pretreatment Thrombocytopenia 9 (22%) 11IVH 37 (76%) 46NEC 4 (9%) 6 (4Ventilatory Dependant 44 (98%) 171

infants treated by primary surgical ligation (P b .001). Theonly pretreatment variable that was significantly associatedwith recurrence after indomethacin was PDA diameter (norecurrence: mean diameter, 2.09 ± 0.8 mm; recurrence: meandiameter, 2.62 ± 0.6 mm; P b .001; odds ratio [OR], 2.75;95% confidence limit [CL], 1.61). Recurrence after indo-methacin treatment did not affect survival (P = .26; OR, 1.82;95% CL, 0.64) but was significantly associated with thelength of mechanical ventilation (no recurrence, 27.97 ± 24.1days; recurrence, 43.94 ± 33.4 days; P = .009; OR, 1.11;95% CL, 1.03) and the number of days on TPN (norecurrence, 29.9 ± 21.3 days; recurrence, 39.54 ± 27.1 days;P = .043; OR, 1.09; 95%, CL 1.00) (Table 2).

Posttreatment complications in 198 infants after initialtreatment with indomethacin include NEC in 15% (30/198),new or increased IVH in 33% (66/198), ARF in 15% (30/198), thrombocytopenia in 25% (44/198), bleeding in 7%(14/198), and gastrointestinal bleeding in 1% (2/198). Incontrast, postoperative medical complications in 46 infants

omethacin (n = 198) Ligation (n = 46) P

1 ± 1.9 24.8 ± 1.5 .001± 140.5 678.7 ± 153.5 .038

0 ± 1.8 5.91 ± 2.3 .0235 ± 4.9 3.93 ± 4.4 NS.5%) 3 (6.5%) NS(18%) 10 (22%) NS(57%) 28 (61%) NS(71%) 30 (68%) NS(7%) 4 (9.5%) .0025(28%) 31 (69%) b.001%) 4 (9.8%) NS(99%) 39 (95%) NS

Table 2 Recurrence after indomethacin

n = 161 No Yes P

Gestational age (wk) 26.3 ± 1.9 25.7 ± 1.8 NSBirth weight (g) 758.2 ± 139.4 724.4 ± 139.7 NS5-min Apgar 6.77 ± 1.9 6.87 ± 1.7 NSCongenital anomalies 13 17 NSMaternal complications 42 49 NSPDA diameter (mm) 2.09 ± 0.8 2.6 ± 0.6 b.001Survival 63 70 NSCMV days 28.0 ± 24.1 43.9 ± 33.40 .009TPN days 29.91 ± 21.3 39.54 ± 27.1 .04

CMV indicates conventional mechanical ventilation; NS, not significant.

Table 4 Comparative risk of complications after treatmentreceived

115Analysis of outcome in ELBW infants with PDA

after primary surgical ligation were NEC in 7% (3/46), newor increased severity of IVH in 17% (8/46), ARF in 7%(3/46), thrombocytopenia in 17% (8/46), bleeding in 2%(1/46), and gastrointestinal bleeding in 4% (2/46). Post-operative surgical complications in this group includeligation of pulmonary artery in 1 infant; pneumothorax in 3infants; and nil episodes of chylothorax, wound infection,bleeding, and recurrent laryngeal nerve injury (Table 3).

Fifty-eight infants failed indomethacin therapy and weresubsequently referred for surgical ligation comprising afourth treatment group. All treatment groups were thencompared with respect to posttreatment complications(Table 4). Complications were significantly more commonin infants treated with indomethacin compared with thosetreated by primary or salvaged surgical ligation.

Survival by treatment received is given in Table 5. Therewas no significant difference in survival between any of thetreatment groups, although there was a trend towardimproved survival in the group treated with both indometha-cin and surgical ligation compared with indomethacin alone(P = .07). The most important finding is that survival wassignificantly better in all treatment groups compared with thegroup of infants who received no treatment.

Pretreatment variables that were associated with survival(or nonsurvival) are given in Table 6. Infants with lowergestational age and lower birth weight were less likely tosurvive as well as those with congestive heart failure orsepsis. Variables that had no apparent affect on survivalinclude 5-minute Apgar, congenital anomalies, IVH, NEC,or PDA diameter.

Posttreatment variables that affected survival are given inTable 7 and include NEC and thrombocytopenia. As can be

Table 3 Surgical complications in 104 infants after surgicalligation of PDA

Ligation of pulmonary artery 1Pneumothorax 3Chylothorax 0Wound infection 0Bleeding 0Recurrent laryngeal nerve injury 0

seen, complications such as ARF and bleeding were morelikely to affect survival after indomethacin compared withsurgical ligation.

3. Discussion

The ostensible findings of this study are thatindomethacin and surgical ligation yield similar andsignificantly better survival rates compared with notreatment of PDA in ELBW infants. Indomethacinincurred a 45.8% recurrence rate in this study. However,because most infants who recurred were salvaged bysurgical ligation, recurrence did not adversely affectsurvival. Indomethacin was associated with a significantlyhigher posttreatment complications compared with surgicalligation, and complications including NEC, ARF, andthrombocytopenia were in turn associated with a sig-nificantly reduced survival.

Why survival after indomethacin treatment was notmore affected by the relatively high complication rate maybe explained by selection bias inherent in the study. Asindicated, treatment was selected for infants based on aprotocol that tended to exclude indomethacin for thosewith preexisting IVH, ARF, or thrombocytopenia. How-ever, this was certainly not the only selective factor

Table 5 Survival by treatment received

Survival (%) P

No treatment 57.7Indomethacin ⁎ 82.4 b.001Surgical ligation 86.0 .003Both indomethacin and surgery ⁎ 92.7 b.001

⁎ P = .07; indomethacin vs both.

Table 7 Survival by post operative variable

Variable Survival

Yes (n = 226) No (n = 55) P

Recurrence 58 8 NSNECAfter indomethacin 19 8 .034After ligation 1 4 .003

IVHAfter indomethacin 51 12 NSAfter ligation 13 1 NS

ARFAfter indomethacin 16 8 .002After ligation 7 1 NS

ThrombocytopeniaAfter indomethacin 27 14 .002After ligation 8 5 .002

BleedingAfter indomethacin 6 6 .01After ligation 1 1 NS

Gastrointestinal bleedingAfter indomethacin 1 1 .042After ligation 2 1 NS

Table 6 Preoperative variables associated with survival

Variable Survival

Yes (n = 226) No (n = 55) P

Gestational age (wk) 26.1 ± 1.9 24.7 ± 1.8 .009Birth weight (g) 751.4 ± 142.4 627 ± 143 .002Maternal preeclampsia 50 3 .0155-min Apgar 6.71 ± 1.9 6.02 ± 2.0 NSCongenital anomalies 49 7 NSIVH 85 25 NSNEC/SIP 8 4 NSSepsis 132 43 .018Congestive heart failure 10 8 .04PDA diameter 2.36 ± 0.8 2.39 ± .08 NS

116 F. Alexander et al.

affecting the results of this study. Infants treated initiallyby surgical ligation had significantly lower gestational age,birth weight, and lower 5-minute AGPAR scores inaddition to larger PDA diameter in comparison to infantstreated with indomethacin. Thus, it is apparent that surgicalligation was preferentially chosen for infants with a greaterperceived risk of mortality and recurrence. And, althoughthe results may validate the use of surgical ligationespecially in high-risk infants, the process of selection nodoubt contributed to the nonsensical findings of equivalentsurvival between indomethacin and surgical therapydespite significant medical complications associated withthe former.

Another contributing factor to selection bias may havebeen the use of surgical ligation as salvage therapy. In thisstudy, a significant number of infants (58/198) underwentsurgical ligation after treatment failure with indomethacin.In fact, the group of infants treated with both indomethacinand surgical ligation had the absolute best survival of all4 groups (92.7%), which nearly reached statistical sig-nificance (P = .07) when compared with the indomethacin-only group. Thus, it is conceivable that surgical ligationused to salvage medical failure may have altered survivaldata by shifting the higher risk patients away fromindomethacin therapy.

Finally, one more plausible explanation for the relativelyhigh survival rate associated with indomethacin despitecomplications is clustering. Infants who were least likely tosurvive because of sepsis and congestive heart failure mayhave been prone to other pretreatment conditions such asIVH, ARF, and thrombocytopenia, which by protocol wereindications for surgical ligation as opposed to indometha-cin. None of these 3 pretreatment conditions were bythemselves significantly associated with decreased survival.However, as a group, they may have had some effect on thedistribution of patients. In any event, the point is well madethat higher-risk ELBW infants were more likely to undergosurgical ligation either as initial treatment or salvagetherapy with equivocal or slightly better outcomes thanindomethacin alone.

Of course, the foregoing conclusion presumes a low rateof surgical complications and postoperative mortality. In thisstudy, medical complications were frequent occurring in73.9% (34/46) of infants having primary surgical ligation,69% (40/58) of infants having salvage ligation, and 90.7%(127/140) after indomethacin therapy. In contrast, surgicalcomplications were relatively uncommon, occurring in only3.8% (4/104) infants: one requiring a second surgicalintervention and none leading to death.

In the absence of a randomized prospective study, thequestion may be fairly asked what information derivedfrom this study may help to inform decision making in thetreatment of PDA in ELBW infants. As discussed, selectionbias limits conclusions that may be drawn about the relativemerits pertaining to survival of indomethacin vs surgicalligation as primary treatment of PDA in ELBW infants. Atthe same time, it is apparent from the study that surgicalligation plays an important role in supporting survival ofELBW infants with PDA. One obvious example is thefinding that salvage therapy conveys an important survivaladvantage to those infants who fail indomethacin therapy.Secondly, of the many variables identified in the study thatwere associated with decreased survival, the most strikingwas no treatment of the PDA. Infants in this group were20% to 30% less likely to survive than infants treated byeither indomethacin or surgical ligation (P b .001) despitesharing a similar profile with respect to other risk factorsincluding gestational age, birth weight, 5-minute Apgarscores, sepsis, and NEC. Thus, it seems reasonable toconclude that many of these infants would have benefitedfrom either medical or surgical treatment, or both.

117Analysis of outcome in ELBW infants with PDA

One final issue is the 45.8% recurrence rate incurred afterindomethacin therapy in this study. This is completely in linewith other studies [4,5] and, as previously discussed, is notassociated with higher mortality in this study because it wasnearly always salvaged by surgical ligation after indometha-cin therapy (P = .48). Higher mortality in turn is notassociated with a larger PDA diameter (P = .72). As evidenceof this, the group of infants who received no treatment hadthe highest mortality of all groups while at the same timehaving the lowest mean PDA diameter. This underscores animportant finding that mortality is related not to PDAdiameter but to persistence of the PDA irrespective of size.At the same time, recurrence in this study was significantlyrelated to a larger PDA diameter (P b .001). Infants with amean PDA diameter of 2.62 ± 0.63 mm had a 100%recurrence compared to those with a diameter of 2.09 ±0.83 mm who had a 92.8% success rate after indomethacintherapy. Considering the potential harmful effects ofindomethacin because of medical complications, it wouldseem reasonable to take into account the size of the PDAmeasured by echocardiography as well as the overallcondition of the infant including gestational age and weight,5-minute Apgar scores, sepsis, and congestive heart failurebefore deciding upon the primary treatment.

References

[1] Koch J, Hensley G, Roy L, et al. Prevalence of spontaneous closure ofthe ductus arteriosis in neonates at a birth weight of 1,000g or less.Pediatrics 2006;117(4):1113-21.

[2] Dollberg S, Luska A, Reichman B. Patent ductus arteriosis,indomethacin and necrotizing enterocolitis in very low birth weightinfants: a population-based study. J Pediatr Gastroenterol Nutr 2005;40(2):184-8.

[3] Amin SB, Handley C, Carter-Pokaras O. Indomethacin use for themanagement of patent ductus arteriosis in pre-terms: a web-basedsurvey of practice attitudes among neonatal fellowship programdirectors in the United States. Pediatr Cardiol 2007;28(3):193-200.

[4] Shaffer CL, Gal P, Ransom JL, et al. Affect of age and birth weight onindomethacin pharmacodynamics in neonates treated for patent ductusarteriosis. Crit Care Med 2002;30(2):343-8.

[5] Little DC, Tratt TC, Blalock SE, et al. Patent ductus arteriosis in micropreemies and full-term infants: the relative merits of surgical ligationversus indomethacin treatment. J Pediatr Surg 2003;38(3):492-6.

[6] Ivey HH, Kattwinkel J, Park TS, et al. Failure of indomethacin to closepersistent ductus arteriosis in infants weighing under 1,000g. Br Heart J1979;41(3):304-7.

[7] Koehne PS, Bein G, Alexi-Meskhishvili V, et al. Patent ductus arteriosisin very low birth weight infants: complications of pharmacological andsurgical treatment. J Perinat Med 2001;29(4):327-34.

[8] Raval MV, Laughon MM, Bose CL, et al. Patent ductus arteriosisligation in premature infants: who really benefits, and at what cost.J Pediatr Surg 2007;42:69-75.

Discussion

Speaker7 Very nice paper. Our cardiologists are reallyunwilling to tell us how big the ductus is and when theydo, they are usually wrong. They do quantitate flow, butfrom what I understand those kinds of measurements arenot terribly accurate either. I would caution you about the2-mm mark.

Dr Chiu7 We did look at small, medium, and large, and thenmeasurements as well. We were able to identify a differencebetween 2 and 3 mm. Again, I agree. I think it is probablydependent on the cardiologist you work with as well.

Speaker7 I was very interested in your zero complicationrates with regard to the injury of the recurrent laryngealnerve. We in our institution looked at this prospectivelyand found that a significant number of the very smallbabies had some damage when they were inspected by ourENT surgeons. I would like to hear your comment aboutthat. Thank you.

Dr Chiu7 We did not go as far as to have ENT evaluate, butour surgeons had very good success.

De Clinton Cavett, MD (ROANOKE, Va)7 Congratulationson a very nice paper. Again, as last year and this year, weare beginning to separate out some of these very sicksurgical ductus babies. I agree with our South Carolinacolleague. The ductus is very dynamic, and one particularmeasurement of millimeters, even an hour later, is goingto be different. A static measurement of diameter ofductus regardless of the observer is going to be different.These ductuses vary even from an immediate preoperativeechocardiogram to the actual observation in the operatingroom. Although it would be nice to have an objectivemeasurement of the diameter, I am afraid it is notnecessarily going to be accurate one day to the next.

We are finding in our own institution, as we put out

experience together, that this dynamic problem is an issue,and we are beginning to separate our patients into thosewho are compromised respiratory by the ductus, and alsothose who are having simply cardiovascular and hemo-dynamic events from the ductus, whether they have aorticsteal, mesenteric steal, and renal steal rather than simplyrespiratory problems from the ductus. It is going to beinteresting to try and separate these two. You are getting acombination of the two problems but also a cardiovas-cular problem simply from the steal.