210 Needleman studys and ours was parental IQ included as a control variable, yet this is a major determinant of child intelligence and it is related to lead level. Needleman et al. did not control confounding variables in their often reproduced graphs of teacher ratings of behaviour. Using a sample homogeneous in socioeconomic status, we found no lead effects for teacher ratings. Even when control variables are entered by analysis of covariance, regression analysis, and so on, the results are inescapably undercorrected.8As researchers we should be aware of this.
A second very serious problem seldom noted and not handled adequately in these reports is a statistical one. Because of the very large number of outcome variables (over 50) the Needleman report is particularly poor in this regard. With 50 independent outcome measures, the likelihood of finding one or more significant at alpha =0’05 is O. 93. The situation is even more extreme for the electroencephalographic study.9 With 320 outcome measures, independent univariate statistical tests yielded 6 -2% significant at the 0 -05 level, 2 -5% at the 0 -02 level, and 0 -6% at the 0 -01 level. (I cannot usually get that close to chance in classroom demonstrations.) These variables were carried forward into other analyses (same subjects) which yielded statistical significance.
substantive responses to these and other serious in research have not been forthcomng. Needleman this problems has tried to divert attention from them by charging that I have a vested interest in the matter. I did accept sponsorship of the Lead Industries Association for a presentation before the Environmental Protection Agency. Without sponsorship, for which I accepted no fee, these views would not have been presented. Essentially everything I said had been published previously. 1,2
Two interests do prevail. The first concern is a strong commitment of careful research procedures. The second concern is that the attention being given to lead effects is detracting from the truely serious needs of children, particularly our vulnerable poverty-level children. Departments of Psychiatry and Reproductive Biology, Cleveland Metropolitan General Hospital and Highland View Hospital, Cleveland, Ohio 44109, U.S.A.
CLAIRE B. ERNHART
examined. The presence of pain was noted when either of the two tests, described above, was done. Pain was graded as "slight" (the subject winced but would allow repetition of testing) or "significant" (she complained more forcefully of pain and would only allow repeat testing with considerable apprehension). The patellar friction test produced slight pain in only 6 of the 70 knees (9%) and in 1 subject this was bilateral. It did not produce significant pain in any knee. The resisted patellar movement test caused slight pain in 17 knees (24%) and significant pain in 15 (21 %). 12 of the 35 women felt some pain in both knees with this test and a further 8 felt pain in one knee. Only 12 subjects felt no pain at all in either knee. No other abnormal signs were found in the knees apart from crepitus in 2 subjects (painful in 1). These results indicate that there is a significant incidence of pain which can be elicited from normal knees when testing for chondromalacia patellae. The interpretation of the tests described, especially the resisted patellar movement test, must be made with care. Only quite severe pain produced by this test can be regarded as indicating any abnormality. The patellar friction test was associated with fewer false positives and therefore may be more reliable in diagnosing chondromalacia patellae. Department of Orthopaedic Surgery, Manchester Royal Infirmary, Mancheater M139WL
ALMITRINE AND SLEEP APNOEA
SIR,-A Lancet editorial expressed the hope that almitrine might be effective in sleep apnoea syndromes and transient
hypoxaemia during sleep in patients with chronic obstructive pulmonary disease (COPD). Our results do not confirm this hopedfor effect in sleep apnoea. In a single blind study of almitrine given orally at a dosage of 200 mg per day versus placebo to eight patients with the sleep apnoea syndrome there was no reduction in the number of sleep apnoea episodes, whatever their type (central or obstructive) or in whatever sleep stage they occurred. The only significant effect was a reduction of
CLINICAL TESTS FOR CHONDROMALACIA PATELLAE
SIR,-Pain on resisted patellar movement (that is, contracting quadriceps whilst the patella is held forcibly downwards)lo,11and painful grating when the patella is pressed against the femur during knee movement (patellar friction test)12 are two signs used to diagnose chondromalacia patellae. Whilst it has been shown that only 50% of patients with signs and symptoms of chondromalacia actually have demonstrable lesions of the articular cartilage the incidence of these two signs in asymptomatic knees has not been previously reported. Both legs of thirty healthy women (average age 20, range 18- 25), of whom had had any symptoms from either knee, were
HL, Gunnoe CE, Leviton A, et al. Deficits in psychologic and classroom performance in children with elevated lead levels. N Engl J Med 1979; 300: 689-95. 8. Reichardt CS The statistical analysis of data from nonequivalent group designs. In: Cook TD, Campbell DT, eds. Quasi-experimentation: Design and analysis issues for field settings Chicago: Rand McNally, 1979. 9. Burchfiel JL, Duffy FH, Bartels PH, Needleman HL. The combined discriminating power of quantitative electroencephalography and neuropsychologic measures in evaluating central nervous system effects of lead at low levels. In: Needleman HL, ed. Low level lead exposure: The clinical implications of current research. New York: Raven Press, 1980 75-89 10. Devas MB. Chondromalacia of the patella. Clin Orthop 1960; 18: 54. 11. Bentley G. The surgical treatment of chondromalacia patellae. J Bone Joint Surg 1978; 60B: 74. 12. Apley AG. System of orthopaedics and fractures, 5th ed. London and Boston: Butter7 Needleman
worths, 1977; 278. 13. Leslie IJ, Bentley G. Arthroscopy Rheum Dis 1978; 37: 540.
diagnosis of chondromalacia patellae. Ann
B. N. LIVINGSTONE
duration of obstructive and mixed apnoeas
during non-rapid-eye-movement sleep. Yet this reduction did not yield any significant improvement in morning blood gas values. Almitrine is a stimulant of peripheral chemoreceptors,2-4 and the reduction in mean duration of apnoeas that we observed accords with the hypothesis that the termination of sleep apnoeas is due to the arousal provoked by chemoreceptor stimulation by h’ypoxia.516 Yet this drug does not seem to interfere with the mechanisms, responsible for the onset of sleep apnoeas, and does not seem to be of any clinical benefit for sleep apnoeic patients. We are still investigating the effects of almitrine in COPD patients prone to hypoxaemia during sleep. Functional
Exploration of the Nervous System
Neurology Clinic, Hôpital Civil, F-67091 Strasbourg,
J. KRIEGER P. MANGIN D. KURTZ
1. Editorial. To sleep, perchance to breathe ... Lancet 1981; ii: 670. 2. Flandrois R, Guerin JC. Effets de l’almitrine sur le contrôle chémoréflexe de la ventilation chez l’homme normal et l’insuffisant respiratoire chronique. Bull Eur
Physiopathol Resp 1980, 16:
M, Schmitt H. Long-lasting hyperventilation induced by almitrine: evidence for a specific effect on carotid and thoracic chemoreceptors. EurJ Pharmacol 1980,
Roumy M, Leitner LM Stimulant effect of almitrine (S 2620) on the rabbit carotid chemoreceptor afferent activity Bull Eur Physiopathol Resp 1981; 17: 255-59 5. Sullivan CE, Issa FG. Pathophysiological mechanism in obstructive sleep apnea. Sleep
1980; 3: 235-46 6 Bowes G, Townsend ER, Kozar L, Bromley SM, Phillipson EA. Effect of carotid body denervation on arousal response to hypoxia in sleeping dogs. J Appl Physiol 1981. 51: 40-45