The nonstress test: An evaluation of 1,000 patients

OBSTETRICS

The nonstress test: An evaluation of 1,000 patients JEFFREY M. BARRETT, M.D. SHERO~J

L. Si1Jt.LYER,

R.~J

.C.

FRANK H. BOEHM, M.D. Nashville, Tennessee In November, 1978, the fetal he~rt rate nonstress test (NST) was instituted as the primary screening procedure for the evaluation ~f fetal well-being ~t Vanderbilt UniversitY Hospital. The resutts of the first 1,000 patients tested are presented.The stillborn rate within 7 days of a reactive NST was 6.4 per 1,000, with the stillbirths Occurring eilmir in patients with diabetes mellitus or with intrauterine growth retardation (iUGR), A review oj other series in which both the total indications jor nonstressed testing and the risk groups in whiCh stlHbirths occurred within 7 days of an NST reveals that patients with diabetes. mellitus (p < 0.025) and patients with IUGR (p < 0.01) are at greater risk for stillbirth within 7 days of an NST. Weekly nonstress testing, effective in preventing stillbirths in most risk groups, is not adequate in patients with diabetes mellitus or IUGR. (AM. J. OsSTET. GYNECOL. 141:153, 1981.)

FETAL HEART RATE reactivity as an indicator of fetal well-being was first suggested by Hammacher 1 and Kubli and colleagues 2 in the late 1960s. In a little more than a decade the nonstress test (NST), based on fetal heart rate reactivity, has at many centers become the most commonly used screening test for the evaluation of fetal well-being. The reported incidence of stillbirths within 7 days of a reactive NST varies from zero 3 to ten 4 per 1,000; these stillbirth rates are believed to represent significant improvement of outcome in high-risk pregnancies. Since November, 1978, the NST has been used as the primary screening procedure for evaluating fetal wellbeing in suspected high-risk pregnancies at Vanderbilt From the Department of Obstetrics and Gynecoiogy, Division of Maternal/Fetal Medicine, Vanderbilt University Hospital. Presented at the Meeting of the Southern Perinatal Association, january 14-16, 1981, New Orleans, Louisiana. Reprint reqitests: jeffrey M. Barrett, M.D., Department of Obstetrics and Gynecology, Division Of Maternal/Fetal Medicine, Vanderbilt University Hospital, Nashville, Tennessee 37232. 0002-9378/811180153+05$00.50/0 © 1981 The C. V. Mosby Co.

REPEAT BIWEEKLY IN PATIENTS WITH DIABETES MELliTUS DR SUSPECTED lUGR; IN OTHER PATIENTS REPEAT WEEKLY

Fig. l. Revised protocol for antepartum fetal heart rate testing at Vanderbilt University Hospital.

University Hospital. The results of our experience with the first I ,000 patients in whom the NST was the primary screening procedure are presented.

Material and methods At Vanderbilt University Hospital from November I, 1978, to September 30, 1980, 2,510 nonstress tests

153

154

Barrett, Salyer, and Boehm

".q;tunl~( 1 J >,

\111

1:1-:.. ~

t)lJ._)r r ( ;'- nc1 ,l

fable I. Indications for NST Indication

Delivery ,;7 dav.1jrom last NST

rota/

Postdatisnt Suspected IUGR Decreased fetal movement Preeclampsia Diabetes (all classes) Premature labor Chronic hypertension Hydramnios Premature rupture of membranes Placenta previa Vaginal bleeding Rh sensitization Other Total*

;•)

l.'

I

!(I

6~

If()

341

l.(li-<0

*Several patients had more than one risk factor.

fable II. Last NST i\lSI' ~eactive

'lonn·active fatal

~7

days from delivery

No. of patients

No. of stillbirths

625

4 3

3fi

661

Table IV. Last NST >7 days from delivery

Stillbirth rate (per /,000) 6.4 83.0 10.5

7

fable III. Antepartum deaths with last \1ST ~ 7 days from delivery NST Reactive

\Jon reactive

4 4 7

None Negative None

TTT£''0

7 i

1\.L.~~

Class C diabetes Nonimmune hydrops Abdominal trauma

()

2

Reactive

Nonrea_ctive Total

No. of patients

I

328 ll 339

No. of stillbirths

-2 !

1"

Stillbirth ra!R (perl ,()(}(}) ti.J 9!.0 H.!-<

-----..------:...-----or-------

Class A diabetes Class D diabetes IUGR l.U\....11'\.

I

Table V. Antepartum deaths with last NST > 7 days from deliverv Contraction stress test

Riskjartur

NST

l'"tVUC

Negative Positive None

,\·ere performed on 1,000 patients for whom delivcn records \Vere available. Nonstress testing \vas begun as ~arly as 27 weeks and continued weekly until the time >f delivery unless the risk factor was felt to ha\e tT;olved. Indications for testing are listed in Table I. Scv~ral patients had multiple risk factors. The tests wert' performed using conventional technique with the patient in the semi-Fowler position. Monitoring was performed using Corometrics* 110. IOl. Ill, or 112 or Brattlet automode fetal monitors. Standard external apparatus was used to determine uterine and fetal activity. Ultrasound and abdominal fetal electrocardio~raphic apparatus were used to monitor heart rate. *Corometrics Medi<·ai Sysren1s, inc., \ValhugfonL Connecticut. tBrattle I ntrument Corporation. Camhridge. Massachusetts.

Contmctum 1tress trst

NST Reactive Nonreactive

Postdatism Postdatism IUGR

___;_______

9 12 :'19

None Negative None

"--··.~---~---~ ~---. --~--·-·

··-

'The dtlr3.tion -of the tests r~ngepm for a period of at least I:) ~econds. The NST was· also considered reactive if ft>tal heart rate accekr;\ti·ons. as described above, occu-rred in response ro exrernal stimulation. In pa~ients in whom reactive patterns on:uri·ed, ~STs were repeated in I week \.mie~s the risk Tac.tors had resolved or the patient failed to prc>ellt f(H· follow-up. Nonreactive pattern. ~onreactive NST was defined as a studv in which the criteria for a reactiw: NST were not met. Auording to our protocol a nonre'aniv(· NST was followed bv a contrat:tion stres-; test (CST). spon-

;1

Nonstress test 155

Volume 141 Number 2

Table VI. Compilation of published experiences with NST in which both the total indications for NST and the groups in which stiHbirth occurred within 7 days of last study were reported Diabetes mellitus No. Krebs and Petres 7 Pratt et al. 8 Devoe9 Rayburn et al. 10 Brettschneider et al. 11 Mendenhall et al. 3 Our study Total

I

Stillbirths

No.

1* 0 1 0 0

20 48 118 13 6 46 98 349

38 90 42 117

26 38 75 426

Suspected IUGR

I

3 6 (p < 0.025)

I

Chronic hypertension, preeclampsia

Postdatism

Stillbirths

No.

1* 0 1 0 0 3 2 7 (p < 0.01)

50 62 96 203 138 127 203 879

I

Stillbirths

No.

0 0 0 0 0 0 0 0

119 117 79 145 72 103 106 744

I Stillbirths 0 0 0 0 0 0 0 0

*Perinatal death not specified as stillbirth.

taneous or induced, within 24 hours. The results of the CSTs were recorded as negative (no evidence of late deceleration pattern after contractions), positive (evidence of late decelerations following more than 50% of contractions), or suspect( occasional late decelerations). In addition all CSTs were described as either reactive or nonreactive based on the criteria for NSTs. When a negative CST was obtained it was followed in I week with an NST. When a positive or suspect CST was obtained, strong consideration was given to delivery. Statistical evaluation was performed using Fisher's exact test.

reactive NST and positive spontaneous CST. Immediate cesarean section was performed, but the neonate was given an Apgar score of 0 at the time of delivery. The third patient was in an automobile accident at what was thought to be 27 weeks' gestation. Because of the gestational age a decision was made against performing a CST. Two days later she was delivered of a stillborn infant ¥veighing 980 gm; cause of the stillbirth ¥las abruptio placentae. Delivery later tha_n 7 days after last NST. Reactive NST. Three hundred and twenty-eight patients had reactive NSTs more than 7 days from delivery (Table IV). Two stillbirths occurred at 9 and 12 davs. both in oatients in whom tht> ori~rinal -o------ indications ------- for the NST was postdatism (Table V). In both cases an antepartum review of the pregnancy data was performed after the first NST, and it was believed that the pregnancy was not actually post date. For this reason weekly testing was discontinued. Nonreactive NST. Eleven patients had nonreactive NSTs more than 7 days from delivery (Table IV). There was one stillbirth, which occurred in a patient whose indication for NST was suspected IUGR (Table V). This NST was originally read as being reactive, and a CST was not performed. A review after the fetal death revealed that the NST was actually nonreactive. An antepartum review of the patients' pregnancy data was performed after the second NST, and it was felt that IUGR was not present. For this reason weekly NSTs were not continued at Vanderbilt University Hospital. The fetus was stillborn 39 days later and was found to be severely growth retarded, weighing 1,400 gm at term. The other 10 patients had negative CSTs. 1

Results A total of 2,510 NSTs were performed on 1,000 patients, of whom 661 gave birth within 7 days of the last study and 339 later than 7 days from the last study. Indications for testing are shown in Table I. Delivery within 7 days of last NST. Reactive NST. Six hundred and twenty-five patients had reactive NSTs within 7 days of delivery, with four stillbirths occurring in this group (Table II). Two of these patients had diabetes mellitus and two had suspected intrauterine growth retardation (IUGR) (Table III). In neither of the diabetic mothers was glucose regulation a problem between the time of the last NST and the stillbirth. All deaths occurred 4 or more days from the last test. One patient with class D diabetes also had a negative spontaneous CST. Nonreactive NST. Thirty-six patients had nonreactive NSTs within 7 days of delivery, with three stillbirths occurring in this group (Tables II and III). In one patient with class D diabetes fetal death occurred 7 days after a nonreactive NST, with a negative CST (Table III). In the second patient the fetus had massive nonimmune hydrops and was found to have a non-

,

.1

-

-

----

---

Comment The initial experience with nonstressed fetal heart rate testing at Vanderbilt University Hospital, resulting

156

Barrett, Salyer, and Boehm

.•.;cpt(·J;d)l~l

\!q

in a stillbirth rate of 6A per I .000 within 7 days ot a nonreactive NST. is similar to that found in severai other series.•-u It is clear that a fetus with a nonreactiH· NST is at much higher risk of stillbirth than a fetu, with a reactive NST (p < 0.01). The indications fm antepartum fetal heart rate testing, which in our experience carried the greatn risk for stillbirth within 7 da\" ol a reactive NST or nonreactive :\ST with negative CST, were diabetes mellitus and suspected llTGR. .\ review of published reports of experiences with Lhe NST reveals that, although there are many such reports, the indications for testing in those fetuses who were stillborn within 7 davs of an "1ST have not been described regularly. 'l'able \'I i~ a contpilation oi the series in which both the total indications for antepartum fetal heart rate testing within 7 days of deliven and the risk groups in which stillbirths occurred withil! 7 days of an ~JS~I' have been gi\'cn. ()nly the n1o~t common indications for NST have been listed. This table again shows that diabetes mellitus (p < 0.0251 and suspected IUGR (p < 0.0 I) ar(· the risk factors with tht· greatest chance of producing stillbirth \Vithin 7 days of an ~ST. ln addition, in our experience a negative CST does not appear to be more effective than an NST in predicting stillbirth within 7 days of the last studv in patients with diabetes mellitus fwo of our diabetic patients were delivered of stillborn infants within i davs of a CST, one with a reactive fetal heart rate pattel n and the other with a nonreacti\'e pattern. Our condu~ion from these data is that antepartum fetal heart rate testing should be performed more frequently than every 7 days in patients with diabetes mellitus or suspected IUGR. An interesting finding was that patients who ga\t· birth later than 7 days after the last NST was pertormed had an incidence of stillbirth of ~.8 per l ,000, which is similar to that of the group whose last test was within 7 davs of delivery. Two of these stillbirths occurred in patients whose original indication for NST was postdatism. In both of these patients testing was

REFERENCES I. Hammacher, K.: The clinical significance of cardiotot·og-

raphy, in Huntingford, P.J., Huter, K. A., and Saling, E., editors: Perinatal Medicine, New York, 1969, Academic Press, Inc., pp. 80-93. 2. Kubi, F. W., Kaeser, 0., '\nd Hinselmann, M.: Diagnostic management of chronic placenta insufficiency, in Pedle, A., and Finzi, C ... editors: The Foetoplacental Unit, ~~~s~~;.~am, 1969, Excerpta Media Foundation, pp. ,')"'.:J-.J.J~.

3. Mendenhall, H. W., O'Leary, J. A., and Phillips, K. 0.: The non-stress test: The value of a single acceleration in

1

;

i -,

l~~l"iJ

)i-,_,;·.l·i. -fc,\lwr 1tl

discontinued because cllltepartull! rt'C\ alu;1tions ... u~pnted. b11t :d.tcr Ill<' ''T•H•d lc.'ol .t reevaluation ol the pregnant\ da.Lt "a' !witt'\ ul t" shm' that i L'GR was not actuaiiv prco-l:ltL : ilc1 dnn: weekll, testing was discontinued. From rbc findings in our patient population iu whom ddin:n· .ts~umptioth m<.n be I

"1-

•

l

i

,

ntd.Ul'. I'll'~!. IISK l
I

(!(}

I

,l

ITSOI\'t': Ul
a

p.ttlt'!H

who has an '\ST for an indication -;uch <~'- dt·ttu· :1 letu' j, diagno;,nl ..t~ ht'lng po"it d;Hc 'Jl is suspected ol havmg It 'GR, \iS l' should 1)\' congcsb that tilt· diagnoses were \'ITOtH·ous. Based on 1he results of our experience wit IJ the l\'ST as the priman screening studv tor ktal well-being. \\t~ have revised our protocul for antep~llturn i-ctal lt:~ting (Fig. IJ. It rhc NST is reactive. wt:ckh 1\ST, are petformed except in patients with diabetes 111elhtus or SU>pccted ll'CR: r.hese p4tienb are tested biwet:'kly. In p;~tient;, ill \\hot !I risk LH tor> resoh <" we "ill consider discontinuing weekh k;,ting unles, there 1' ;1 rtTurrencc ol thdt risk factor r>r ,1 ne11 ri:-;k factor dcvdops. However. 111 the case ul ;nn ktus ;dH> is onu·labeled as being post datt· or j, oll'>!wcted oll~a1lllg ILGR, en·n it <1 later review
J OssTET. (;y:-a:coL. 136:87. 1980. -t. Evertson, L R., and-Paul, R. H.: Antepartum feral heart rate testing: The nonstress test, AM. JOssn:T. GYNF..COL. 132:895; 1978. 5. Keegan, K. A., and l'aul,R. f-1:: Antepart\lrn le·tal· heart rate. testing. IV. The nonstress test as_ a primary approach. AM. J. OBSTET. GYNECOL.1H!75, Hf8{), 6. Druzin, M. D., Gratacos; J., and Paui,R. H.: Anter:artum f.,,~ I fw:'.rt rat .. tf'stimr. VI. Predictive reliability--of "nor~~~~;, t;s~~ in the prev;ntion of antepartum death., AM. J. 0BSTET. GYNECOL 137:746, 1980. evaluating fetal risk. AM.

Volume 141 :>lumber 2

7. Krebs, H., and Petres, R. E.: Clinical application of a scoring system for evaluation of antepartum fetal heart rate monitoring, AM. J. 0BSTET. GYNECOL. 130!7 65, 1978. 8. Pratt, D., Diamond, F., Yen, H., Bieniarz, J., and Burd, L.: Fetal stress and nonstress tests: An analysis and comparison of their ability to identify fetal outcome, Obstet. Gynecol. 54:419. 1979. 9. Devoe, L.: Clinical implications of prospective antepar-

Nonstress test

157

tum fetal heart rate testing, AM. j. 0BSTET. GYNECOL. 137:983, 1980. 10. Rayburn, W., Greene, J., and Donaldson, M.: Nonstress testing and perinatal outcome, J. Reprod. Med. %4: 191, 1980. ll. Brettschneider, I•., Goldstein, P., Baer, D., Kimball, A., and London, R.: Fetal acceleration determinations and perinatal outcome, J. Reprod. Med. 25: I 0, 1980.